Clinical Review

From the Department of Neurology, Oregon Health & Science University, Portland, OR.

 

Abstract

• Objective: To review the clinical characteristics, epidemiology, and management of the most common neuropsychiatric symptoms (NPS) in Parkinson’s disease (PD).
• Methods: Literature review.
• Results: PD has traditionally been considered a disease of impaired motor function. However, neuropsychiatric complications, such as fatigue, depression, anxiety, psychosis, impulse control disorders, and apathy, frequently complicate the course of the illness. Although the development of new medication options in recent years has had a positive benefit on the management of these troublesome symptoms, responses are frequently suboptimal. The development of valid instruments to measure neuropsychiatric symptoms has been vital in research efforts to bridge the gaps in our understanding. Further elucidation of neuropsychiatric pathophysiologies will help to define treatment targets and has the potential to expand our therapeutic armamentarium.
• Conclusion: While NPS affect patients with established disease, recent investigations have demonstrated risk of symptoms in those with early untreated stages of PD; therefore, better understanding of NPS should be the goal of practitioners treating the entire continuum of PD.

Parkinson’s disease (PD) has traditionally been considered a disease of impaired motor function, but increased recognition of nonmotor symptoms and in particular neuropsychiatric symptoms, such as fatigue, depression, anxiety, psychosis, impulse control disorders, and apathy, offer new opportunities for better care of patients. While neuropsychiatric symptoms affect patients with established disease, recent investigations have clearly demonstrated risk of symptoms in those with early untreated stages of PD; therefore, better understanding of neuropsychiatric symptoms should be the goal of practitioners treating the entire continuum of PD. This review will focus on the clinical characteristics, epidemiology, and management of the most common neuropsychiatric symptoms in PD.

Impulse Control Disorders

The recognition that dopaminergic drugs were successful at treating many symptoms of PD was followed by the disturbing realization that impulse control disorders could be an unfortunate side effect in a substantial minority. Impulse control disorders as defined by DSM-IV [1] are disinhibited behaviors that are maladaptive and recurrent, causing personal and relationship consequences. The impulse control disorders that became associated with PD and medication intake, particularly dopamine agonist use, included gambling, hobbyism, punding (stereotyped, seemingly purposeless behaviors), excessive sexual behavior, shopping, hoarding, and less commonly, compulsive eating. The prevalence estimates of these behavioral disturbances range from 6% to 15.5%, compared with < 2% in the general population [2,3]. The addiction-like dopamine dysregulation syndrome, whereby patients self-medicate with high doses of levodopa and short-acting dopamine agonists beyond what is needed for motor control, can lead to significant impairment of the therapeutic alliance in addition to other patient personal relations. With the advent of surgical options to treat PD and its medication complications, it was observed that stimulation of the subthalamic nucleus could be associated with the spectrum of impulse control disorders [4].

Epidemiology/Risk Factors

In a recent systematic review of the literature of impulse control disorders in PD [5], the authors determined that dopaminergic therapy caused compulsive or impulsive behaviors in approximately 10% of PD patients in the course of their treatment, with pathologic gambling and hypersexuality most frequently experienced. Multiple impulse control disorders are not uncommon and may coexist in one-quarter of patients with compulsivity. There appeared to be more disordered behavior with higher comparable doses of agonists. The authors concluded that impulse control disorder symptoms tended to occur with initiation or dose increases of direct D₂/D₃ agonists, such as pramipexole and ropinirole. Importantly, impulse control disorder behavior improved if not resolved with discontinuation or reduction of dosage of the agonist, even if a compensatory levodopa dosage is added or increased. Perhaps not surprisingly, it was observed that if patients had a preexisting impulse control disorder prior to PD or the initiation of treatment, there was a high likelihood of worsening of symptoms. This small subgroup is estimated at about 1% of PD subjects, which corresponds to the prevalence of impulse control disorders in the general population. Other identified potential risk factors for impulse control disorder development include male gender, young age at onset, a personal or family history of addiction, novelty or risk seeking personality, and a concurrent diagnosis of depression [3]. In a recent study of early PD patients, the risk of developing an impulse control disorder became important once treatment with dopaminergic drugs began and continued for a year or more [6].

Pathogenesis

The pathogenesis is not fully understood, however, mesolimbic dopamine alterations are strongly suspected. It has been long speculated that the high doses of dopamine needed to replete the relatively depleted dorsal striatum overdose the “intact” ventral striatum and cause this neuropsychiatric disorder [7–9]. The additional cognitive impairments in PD, which can include problems with attention, working memory, planning, forethought and decision-making, are faculties that can markedly increase susceptibility to impulse control disorder [8].

The role of serotonin deficiency in the PD brain and its part in inhibiting the patient’s ability to delay rewards adds to the complexity of impulse control disorder pathogenesis. Dorsal raphe nuclei disease in PD results in loss of serotonin innervation to substantial portions of the prefrontal and motor cortices in addition to basal ganglia substructures like the striatum, pallidum and subthalamic nucleus [10]. Together with dopamine, serotonin may work to regulate risk-sensitive decision making, response inhibition, waiting for future rewards, and overall impulse control. Its relative loss therefore also likely contributes to tipping the balance towards impulse dyscontrol [11,12]. The role of other neurotransmitters such as opiate systems involved in the process of acquisition and maintenance of addictive behaviors like dopamine dysregulation syndrome remains to be fully understood.

Treatment

The most successful strategy to address this problem is to reduce or eliminate the offending medication, usually the dopamine agonist. This may be associated with worsening apathy, anxiety or depression; however, substituting levodopa can be a successful strategy in many cases [13]. Zonisamide was described to be possibly effective in a trial of 15 subjects; however, the open label nature of this evidence must be considered as with other case reports using valproate, donepezil, and selective serotonin reuptake inhibitors (SSRIs) [14–16].

Fatigue

An easy to understand operational definition of fatigue is that it is a state of extreme tiredness, weakness, or exhaustion, either physical or mental or both. Fatigue is not uncommon in the general population [17] but is increasingly recognized to occur in numerous disease conditions and is frequently encountered in PD and multiple sclerosis. The latter is of special significance in the consideration of the neurotransmission of fatigue, as it is not thought to be a disease of dopamine deficiency. The pathophysiology remains unclear, and it may differ depending on whether the fatigue is experienced as more physical or mental, or rather motor versus nonmotor as some authors propose.

Fatigue has been conceptualized as central or peripheral in character. Peripheral fatigue is best understood as muscular fatigue caused by repetitive muscular contraction or reduced force generation [18]. Central fatigue however, is divided into mental or physical fatigue. Mental fatigue can occur after sustained attentive or emotional activity. It may alternatively be provoked after boring repetitive tasks or lack of intellectually stimulating activity. Physical fatigue is the sense of body exhaustion or energy to perform physical tasks even though the ability to carry them out exists.

Epidemiology

As recognition of the problem of fatigue increased in the last 2 decades, the realization that one-third to one-half of patients experience it at some point has improved opportunities for recognition and treatment [19]. Fatigue may be the presenting symptom in one-third of patients prior to actual motor symptom onset [20]. Half of untreated PD patients in a biomarker cohort study reported fatigue [6]. Unfortunately, it is also described by patients as one of the most disabling symptoms, causing significant impact on quality of life [19]. Fatigue in PD is associated with higher rates of depressive symptoms, but occurs with higher prevalence in nondepressed patients [21]. Poor ability to initiate and sustain activity due to fatigue is different from depression, excessive sleepiness, or impaired motor function [22,23].

Pathophysiology

The pathophysiology of fatigue remains somewhat unclear, though physical fatigue is likely a significant part of the problem and related to dopamine deficiency based on studies of time and force generation of keyboard strikes in PD subjects before and after L-dopa administration. These subjects had declines in force and increased physical fatigue which improved after L-dopa [24]. In other studies using transcranial magnetic stimulation to study changes in cortical excitability, the degree of physical fatigue correlated with abnormalities in motor evoked potentials during fatiguing exercising. These studies also support the hypothesis that fatigue is a motor symptom [25,26]. In the ELLDOPA study, fatigue worsened more in PD subjects treated with placebo [27]. Other imaging studies have suggested suggested nondopaminergic mechanisms including serotonergic pathway abnormalities [28], thus the question behind the etiology and solution for all cases of fatigue remains to be settled.

Diagnosis

The diagnosis is fatigue may be challenging as it may mask as depression or apathy. There are a number of fatigue rating scales available; however, the validated Parkinson’s Fatigue Scale (PFS) supersedes many of the problems of using a generic scale which could overlap motor questions and potentially be confounding [29,30].

Treatment

Most important is awareness and vigilance for the symptoms of fatigue, depression, and apathy and effort to distinguish between them. It may require structured interviews or assessment tools to properly diagnose the problem. Treatment is less clear in that few studies have clearly indicated the best treatment options. In placebo-controlled trials, methylphenidate did improve fatigue as did levodopa [31]. Modafinil, a hypocretin modulator and a drug first approved by the FDA for treatment of narcolepsy, has demonstrated mixed results in recent years. It may reduce physical fatigue and reduce excessive daytime sleepiness but likely does not reduce subjective symptoms of fatigue [32]. L-dopa can significantly reduce fatigue in many patients, which would argue that it often is a motor symptom [33,24]. In a post-hoc analysis of the ADAGIO delayed start study, patients taking rasagiline 1 mg/day and 2 mg/day (the latter dose exceeds the usual clinical dosing) showed significantly less worsening of symptoms on the PFS compared to placebo over time [34]. It is important to realize that once motor symptoms are optimally treated with dopaminergic medications, while many patients will feel significant relief from fatigue some patients will continue to feel symptomatic.

Apathy

The definition of apathy has become more complicated and refined, incorporating findings from the study of brain disease and behavioral analysis. Marin’s classic elaboration of apathy as lack of motivation not attributable to diminished level of consciousness, cognitive impairment, or emotional distress has been built upon by Levy and Dubois [35–37]. They suggest apathy may be better thought of as an observable behavioral syndrome characterized by a quantitative reduction of self-generated voluntary and purposeful behaviors. They suggest 3 apathetic subtypes: emotional, cognitive, and auto-activational, which reflect different disease states accounting for failure of normal goal-directed behavior.

Epidemiology

Prevalence estimates for apathy in PD vary. This is likely due to the varying recruitment criteria among studies, with some including patients with comorbid depression and dementia and others containing only “pure apathy.” Other reports may have had referral bias issues, as community-based studies report lower prevalences in general. In a group of newly diagnosed PD patients, using more restrictive criteria (apathy subscale of the neuropsychiatric inventory and the diagnostic consensus criteria for apathy validated in PD), Pedersen reported a prevalence of apathy of 14.3% [38]. In a 4-year prospective longitudinal cohort study, an annual incidence rate of 12.3% was reported, with apathy developing in 60% of the cohort by the study’s conclusion [39].

Apathy has been associated with longer disease duration, male gender [40], higher daily levodopa doses [41], more severe parkinsonism [38], and lower education status, though the latter feature remains under debate. Early cognitive deficits appear to be a risk factor for development of apathy [42]. The patterns of cognitive dysfunction and apathy remain unsettled in the literature.

Pathology

The pathology of atrophy remains unexplained and is unlikely to be reduced to a simple atrophy of one nucleus or the tone of one circuit. However, in a small neuroimaging study, severity of apathy correlated with atrophy of the bilateral nucleus accumbens [43], and it is notable that one major input to the nucleus accumbens is the amygdala. According to Braak staging, by stage 4 significant involvement of the amygdala by Lewy bodies has occurred. Others have found changes in grey matter density that could correlate with deficits of the prefrontal-basal ganglia circuitry to produce dysfunction of segregated frontal-subcortical loops. These may correlate with the “autoactivation” deficit pattern of apathy in which patients have a lack of self-initiated actions, even thoughts, though appear more normal when giving externally prompted responses [37,44].

Assessment

Clinically, the relationship between apathy and depression can be hard to disentangle, especially since many studies have found an association between them, especially with regards to apathy and anhedonia. Depression may feature negative self thoughts and sadness while apathy is notable for lack of initiation and effort. Viewed over a longer period of time, apathy tended to worsen in a linear fashion, where depression tended to fluctuate with improvements and exacerbations.

The Movement Disorders Society task force has recommended the Lille Apathy Rating Scale (LARS) for assessment of apathy; English and French versions have been validated in PD patients. It uses a semi-structured interview format assessing 4 dimensions of apathy: self awareness, intellectual curiosity, emotion, and action initiation [45–47].

The impact of apathy cannot be underestimated as this poor show of motivation or effort leads to lack of engagement in old activities or interest in new ones. Spouses may misinterpret this change in behavior as laziness or deliberate social withdrawal, or perhaps entitlement. It is not surprising that apathy routinely shows up on quality of life (QoL) questionnaires as highly impacting patients and families. In one study, apathy was the nonmotor symptom most likely to cause caregiver distress in PD [40,48–50].

Treatment

No approved drugs exist for treatment of apathy. However, clinical experience often confirms that dopaminergic modulation can be helpful in the treatment of apathy as indirect evidence suggests. A meta-analysis of controlled trials using pramipexole and Part I of the Unified Parkinson's disease rating scale (UPDRS) (secondary measure) showed the medication improved scores on this measure of motivation and mood in non-depressed subjects [51] with PD. Rare patients undergoing subthalamic deep brain stimulation have been reported to experience new and sometimes severe apathy after surgery [52]. This was posited at least in part to be the result of reduction of dopaminergic medication due to surgery.

Nondopaminergic pharmacotherapy of apathy is in its infancy. A recent controlled trial of rivastigmine in 31 French subjects with moderate to severe apathy based on LARS showed that 6 months of treatment at 9.5 mg/day improved average scores from –11.5 to –20 compared with placebo. While quality of life did not improve, caregiver burden did. The investigators found in this group of subjects that apathy was a possible herald for early dementia in PD [53].

A post-hoc analysis of the ADAGIO study (rasagiline or placebo in PD patients taking antidepressants) found that rasagiline use was associated with a nonsignificant slowing of apathy development during the trial [54].

Psychosis

Psychotic symptoms are a common occurrence in drug-treated patients, with visual hallucinations occurring in up to 30%, though over a 20-year period up to three-quarters of patients may develop visual hallucinations.After visual, the most common type of hallucination is auditory, followed by the other affected senses such as tactile, olfactory, or even taste [57]. Delusions, which tend to be paranoid in nature, occur in about 5% of patients [55–57]. The presence of psychotic symptoms is associated with poorer quality of life [58].

Symptomatology

The visual hallucinations of PD are usually quite stereotyped, and have been described as “minor” and “non-minor”[59]. Minor hallucinations refer to transient peripheral field stimuli that disappear when brought into central focus, “something flashed by,” a sense of a living being nearby, “a presence in the room,” or illusions whereby objects are transformed, eg, a bush in the yard is a deer.

Auditory hallucinations tend to be vague or indistinct sounds, like music in another room as opposed to voices speaking directly to the patient as might be experienced in a primary psychotic disorder. Tactile forms often involve insects or other animals crawling on the skin. Olfactory hallucinations may take the form of smelling perfume, toxic odors from room vents, etc.

Early in the experience, the visual hallucinations may be amusing in that they consistently remain nonthreatening, similar day to day, and sometimes oddly provide an aspect of comfort or companionship to the patient. More commonly, the hallucinations are bothersome to the patient because the experience indicates to the patient that there is something wrong with their mind. Visual hallucinations often begin in low-stimulus environments, often in the evening or other low-light conditions, but as the problem advances they can occur at any time of day. While visual hallucinations may initially occur for only seconds at a time many days apart, the frequency and duration can increase until they occur hours at a time every day and are accompanied by multiple other visual hallucinations, delusions, and confusion [60].

Delusions tend to be more distressing to patients and caregivers because they are often paranoid in nature. The patient is more likely to act out due to the anxiety the paranoia creates. For example, she may change passwords to online accounts due to a belief that unknown assailants are after her finances. He may go to great lengths trying to prove his wife is cheating.

Risk Factors

While the primary risk factor for psychotic symptom development is dementia [57], it occurs in nondemented patients. Other associations include reduced visual acuity [56], visual processing impairment [61–65], use of dopamine agonists, REM behavior disorder, duration of PD, axial rigidity subtype of disease [61,66–68]. The pathophysiology of psychosis in PD is likely complex and remains currently unexplained. The role of excess dopamine has been described above, but there is also data suggesting cholinergic deficits in the cortex may also contribute. Excess serotonin (increased 5HT2A receptor subtypes) in the temporal lobe within the visual processing pathway has been postulated to be of significance [69,70]. Hypometabolism in visual association areas of the brain in subjects with visual hallucinations has been demonstrated in PET and functional MRI studies [64,71]. This is similar to findings in patients with dementia with Lewy bodies [72].

This review focuses on the primary forms of PD-related psychosis, which occur with a clear sensorium and generally longer exposure to dopaminergic medication. It is important to distinguish 2 other common scenarios in which hallucinations or delusions may occur. In the common toxic-metabolic delirium, a clouded sensorium with attention deficits may be the only clue to the etiology of new onset confusion with visual hallucinations. It is highly likely that resolution of the underlying medical problem will lead to resolution of the new onset psychosis and encephalopathy. In a second scenario, hallucinations precede or occur very shortly after the onset of initiation of dopaminergic medication. This differs from the classic syndrome described earlier, in particular when visual hallucinations precede any initiation of medication, and likely represents the distinction between a diagnosis of Lewy body disease and PD [60].

Treatment

Management of psychosis is approachable, but often the outcome is unsatisfactory and associated with trade-offs in motor control. It is unfortunately true that psychotic symptoms are often associated with increased caregiver burden and are a cause of increased nursing home placements [73]. When considering the workup of psychotic symptoms, the differential diagnosis includes delirium, dream enactment (REM behavior disorder), or less commonly, Bonnet syndrome.

A delirium may be precipitated by a difficult to diagnose infection; new-onset confusion and psychotic symptoms may be the heralding presentation. Urinary tract or upper respiratory tract are common vulnerable sources of infection. Once infection is ruled out, the next practical step is to review the patient’s medication list and manage centrally acting drugs that could be contributing to the altered sensorium. A recent prescription of opioids for a dental treatment or a new muscle relaxant may be a culprit, though it is not that usual. A bladder anticholinergic could be suspect and is worth eliminating especially if its addition coincided with the appearance of the psychotic symptoms. Once the non-dopaminergic medications have been reduced/eliminated, then the PD medications should be considered. The general approach is to eliminate the medications that provide the least benefit while being more likely to contribute to psychotic symptoms. Anticholinergic medications, dopamine agonists, selegiline should all be uppermost in that consideration until one is left with L-dopa and COMT inhibitors (the latter function to increase levodopa availability). Then COMT inhibitors and levodopa can be reduced; however, at any point motor control can suffer with the loss of symptomatic therapies [74].

Clozapine is effective against psychotic symptoms in PD, at doses much lower than used in schizophrenia (300-600mg/day). The average dose in the US randomized controlled clinical trial was 25 mg/day, with no associated motor worsening. Patients in the United States are required by the FDA to be placed in a computer-based registry and monitored for agranulocytosis for the duration of clozapine therapy. This rare adverse event is not dose related. Orthostasis can occur at these low doses however. Fortunately the metabolic syndrome is not associated with this range of administration [75,76].

Quetiapine was not found to be effective in 3 blinded randomized controlled trials despite its rather common use for this purpose. It was not associated with motor worsening, however.

Other neuroleptic medications have not resulted in widespread use, because trials have been open label, or outcomes demonstrated motor worsening. Cholinesterase inhibitors have been the subject of a few positive case series, however results appear to be sporadic, the effect size is relatively small, and side effects of this medication class are common [77–79]. It is clear that there is an unmet need for a medication for psychotic symptoms. Clozapine is effective but onerous in its monitoring requirements. Practically speaking, there are relatively few PD patients who take advantage of it because of its feasibility challenges. Yet the problem of psychotic symptoms is a significant one that imposes important challenges to the patient and caregiver, and may limit the number of medications that the patient needs in order to optimize quality of life.

Pimavanserin, a novel medication which acts as a selective serotonin inverse agonist, is in the early application stages for FDA approval for treatment of psychotic symptoms in PD [80]. In its pivotal phase III controlled trial, the drug reduced not only positive symptoms (hallucinations/delusions) without causing motoric worsening, but also reduced caregiver burden. Pimavanserin improved certain sleep features without causing daytime sedation. If this drug meets final approval, it may present an exciting option for many patients for whom treatment was previously limited.

Depression

A study of early PD suggested that depression is often unrecognized and frequently untreated [1]. Indeed it is not unusual for depression to predate the diagnosis of PD by an average of 4 to 6 years [81]. Expanding to the larger PD population, it is generally accepted that about 30% to 50% of PD patients experience clinically significant depression, and once diagnosed may have a long term course, or may recur [82,83]. This is important as untreated depression is an important cause of poor quality of life in early PD. In addition depression can exacerbate motor disability, lead to earlier motor treatment with medication, and increase caregiver stress [83–85].

Diagnosis

Diagnosing depression in PD is more challenging as somatic, cognitive, and vegetative symptoms of altered mood can be imitated features in PD, such as facial masking, fatigue, sleep changes, weight loss, and working memory dysfunction (Table). Therefore, nonsomatic features tend to be more valuable to the clinician to query, such as the hedonic state (does the patient seek pleasurable activity?), mood, pessimism, state of hope, sense of capability[86,87]. Diagnosis of major and minor depression as well as dysthymia in PD using DSM-IV criteria has been validated [88]. To avoid underdiagnosis of this important commonly coexisting comorbidity, being inclusive with somatic overlapping features is still more important than trying to distinguish which disease state it belongs to.

A number of clinimetric rating scales for depression have been used and their advantages have been largely related to their objective nature (quantifiable); thus, they tend to be most useful in epidemiologic research studies or for larger scale screening purposes. Examples include the the Beck Depression Inventory, the Geriatric Depression Scale, and the Hamilton Depression Scale, all of which have been shown to be valid tools in PD (with the exception of the UPDRS Depression). It is important to note that they do not substitute for a diagnostic clinical interview [89].

Suicide is not common in PD, however suicidal ideation is estimated at about 11% in PD patients [90], and while there was concern initially after deep brain stimulation procedures began that suicide incidence was increased, evidence does not support this [91].

Pathophysiology

The pathophysiology of depression in PD is largely unknown however is thought to be less causally due to psychosocial factors and more etiologically driven by brainstem monoamine and serotonergic dysfunction [92]. Nonetheless, similar to other chronic conditions, PD patients can certainly develop fear of disability, guilt about impact on others, or other reactive mood changes. Overall, rates of depression are higher in PD compared with patients with similar conditions matched for disability [93].

Treatment

First, the clinician must determine if depression is a result of short-term fluctuations, chronic undertreatment of motor disease, or longer-term mood phenomenon. One important pattern to recognize are mood fluctuations, which can parallel motor OFF-ON cycling. It can be valuable to distinguish this as “subsyndromic” depression or anxiety (sometimes referred to as “OFF dysphoria”), as it can respond to improvement in antiparkinsonian medication dosing patterns that reduce fluctuations[94–96]. Similarly, elevating chronic motor undertreatment to goal therapy can result in mood normalization.

If symptoms persist despite optimization of motor/nonmotor fluctuations or chronic undertreatment and are severe enough to warrant treatment, then therapies used can range from nonpharmacologic education, support, and mental health referrals, as well as pharmacologic support in the form of medications.

A frequent but uncontrolled observation was that when undertreatment of motor disease was finally redressed, mood often improved. A multicenter randomized controlled trial of pramipexole in PD patients without motor fluctuations but with mild to moderate depressive symptoms showed the drug improved scores on the Beck Depression Inventory over a 12-week period. The improvement in mood was 6 points overall, but by 2 points over placebo, illustrating the importance of the size of the placebo effect [97]. Given the potential side effect profile of dopamine agonists, it may be useful to weigh the antidepressant effects only when their motor benefits are already being employed.

Controlled trials have demonstrated efficacy of both selective serotonin reuptake inhibitors (SSRI) and selective norepinephrine reuptake inhibitors (SNRI) antidepressants in PD. Clinical trials have demonstrated efficacy against placebo or with other antidepressant comparators. Examples of drugs with demonstrated efficacy include citalopram, paroxetine, venlafaxine, and nortriptyline. Results have attempted to illuminate the small unique differences between classes of antidepressants or dynamic properties between drugs within a class. For example, desipramine may nudge scores on a mood scale a few weeks sooner than a purer SSRI. Paroxetine (SSRI) versus venlafaxine (SNRI) improved mood scores comparably in a multicenter trial with a placebo comparator. In general, all have all been demonstrated to be effective and with a relatively low side-effect profile, comparable to the general population[98–102]. While case reports exist in the literature, the interaction of monoamine oxidase B inhibitors and SSRIs has not caused significant hypertensive crises or risk of serotonin syndrome [103,104]. Electroconvulsive therapy (ECT) can be used for severe refractory depression in PD as for non-PD patients, with case reports of very effective results. Due to the rarity of use, systematic evidence for its use is lacking [105,106].

Other novel agents and techniques such as omega-3 fatty acids [107] and repetitive transcranial magnetic stimulation [108] have been reported with promising early results. Cognitive behavioral therapy (CBT), which may involve stress management techniques, sleep hygiene, and caregiver support, additionally almost always provided improvement in measured outcomes, whether the trial was controlled or open label in design. In one RCT of CBT in PD of 14 weeks’ duration, there were significantly more treatment responders in the CBT group, with a number needed to treat of only 2 [109].

Anxiety

Anxiety is also common in PD, at least as common as depression considering that prevalence estimates suggest up to 50% of patients experience it [110–112]. Manifestations of anxiety may include panic attacks, generalized anxiety disorder, social anxiety, or other phobias [113]. Anxiety has an important negative impact on health-related quality of life and is often underrecognized by clinicians [114]. While reliable and valid scales to measure anxiety have been lacking in PD, a new effort has yielded the “Parkinson Anxiety Scale” though full clinimetric properties of the scale remain to be demonstrated (sensitivity to change) [115].

Anxiety that parallels the timing of motor OFF-ON cycling is important to recognize. This “subsyndromic” anxiety or anxiety disorder not otherwise specified (ie, the anxiety does not meet DSM-IV criteria) can respond to improvement in antiparkinsonian medication dosing patterns that reduce fluctuations [116,117]. Indeed, the presence of motor fluctuations is the principle marker of anxiety in many studies [118–120]. In an analogous manner, anxiety can predate PD by years and be part of the nonmotor amalgam of features heralding the disease [6,121].

Treatment

Systematic controlled trials of anxiolytic treatment for PD are lacking; therefore, SSRIs are prescribed for this purpose as in non-PD patients. Until SSRIs are demonstrated to be of benefit in anxiety, they are likely safer than use of benzodiazepines, which are associated with risk for falling, cognitive dysfunction, or autonomic dysregulation in PD patients when used during waking hours. Psychotherapy and other nonpharmacologic approaches are likely to be of benefit. A small study of neuromuscular (massage) therapy demonstrated improvement on the Beck Anxiety Inventory in PD [122]. A case report of ECT for severe anxiety has been published [123].

Conclusion

Neuropsychiatric symptoms are common in PD and new knowledge about clinical features, epidemiology, and treatment options has been gained in the last decade, though much remains to be discovered. The development of valid instruments to measure neuropsychiatric symptoms has been vital in these research efforts to bridge the gaps in our understanding. Further elucidation of the pathophysiologies of neuropsychiatric symptoms will help to define treatment targets and likely fuel drug development and the discovery of drugs with more potent benefit and fewer side effects.

 

Corresponding author: Kathryn A. Chung, MD, Department of Neurology, Oregon Health & Science University, Portland, OR, [email protected].

Financial disclosures: None.

From the Department of Neurology, Oregon Health & Science University, Portland, OR.

 

Abstract

• Objective: To review the clinical characteristics, epidemiology, and management of the most common neuropsychiatric symptoms (NPS) in Parkinson’s disease (PD).
• Methods: Literature review.
• Results: PD has traditionally been considered a disease of impaired motor function. However, neuropsychiatric complications, such as fatigue, depression, anxiety, psychosis, impulse control disorders, and apathy, frequently complicate the course of the illness. Although the development of new medication options in recent years has had a positive benefit on the management of these troublesome symptoms, responses are frequently suboptimal. The development of valid instruments to measure neuropsychiatric symptoms has been vital in research efforts to bridge the gaps in our understanding. Further elucidation of neuropsychiatric pathophysiologies will help to define treatment targets and has the potential to expand our therapeutic armamentarium.
• Conclusion: While NPS affect patients with established disease, recent investigations have demonstrated risk of symptoms in those with early untreated stages of PD; therefore, better understanding of NPS should be the goal of practitioners treating the entire continuum of PD.

Parkinson’s disease (PD) has traditionally been considered a disease of impaired motor function, but increased recognition of nonmotor symptoms and in particular neuropsychiatric symptoms, such as fatigue, depression, anxiety, psychosis, impulse control disorders, and apathy, offer new opportunities for better care of patients. While neuropsychiatric symptoms affect patients with established disease, recent investigations have clearly demonstrated risk of symptoms in those with early untreated stages of PD; therefore, better understanding of neuropsychiatric symptoms should be the goal of practitioners treating the entire continuum of PD. This review will focus on the clinical characteristics, epidemiology, and management of the most common neuropsychiatric symptoms in PD.

Impulse Control Disorders

The recognition that dopaminergic drugs were successful at treating many symptoms of PD was followed by the disturbing realization that impulse control disorders could be an unfortunate side effect in a substantial minority. Impulse control disorders as defined by DSM-IV [1] are disinhibited behaviors that are maladaptive and recurrent, causing personal and relationship consequences. The impulse control disorders that became associated with PD and medication intake, particularly dopamine agonist use, included gambling, hobbyism, punding (stereotyped, seemingly purposeless behaviors), excessive sexual behavior, shopping, hoarding, and less commonly, compulsive eating. The prevalence estimates of these behavioral disturbances range from 6% to 15.5%, compared with < 2% in the general population [2,3]. The addiction-like dopamine dysregulation syndrome, whereby patients self-medicate with high doses of levodopa and short-acting dopamine agonists beyond what is needed for motor control, can lead to significant impairment of the therapeutic alliance in addition to other patient personal relations. With the advent of surgical options to treat PD and its medication complications, it was observed that stimulation of the subthalamic nucleus could be associated with the spectrum of impulse control disorders [4].

Epidemiology/Risk Factors

In a recent systematic review of the literature of impulse control disorders in PD [5], the authors determined that dopaminergic therapy caused compulsive or impulsive behaviors in approximately 10% of PD patients in the course of their treatment, with pathologic gambling and hypersexuality most frequently experienced. Multiple impulse control disorders are not uncommon and may coexist in one-quarter of patients with compulsivity. There appeared to be more disordered behavior with higher comparable doses of agonists. The authors concluded that impulse control disorder symptoms tended to occur with initiation or dose increases of direct D₂/D₃ agonists, such as pramipexole and ropinirole. Importantly, impulse control disorder behavior improved if not resolved with discontinuation or reduction of dosage of the agonist, even if a compensatory levodopa dosage is added or increased. Perhaps not surprisingly, it was observed that if patients had a preexisting impulse control disorder prior to PD or the initiation of treatment, there was a high likelihood of worsening of symptoms. This small subgroup is estimated at about 1% of PD subjects, which corresponds to the prevalence of impulse control disorders in the general population. Other identified potential risk factors for impulse control disorder development include male gender, young age at onset, a personal or family history of addiction, novelty or risk seeking personality, and a concurrent diagnosis of depression [3]. In a recent study of early PD patients, the risk of developing an impulse control disorder became important once treatment with dopaminergic drugs began and continued for a year or more [6].

Pathogenesis

The pathogenesis is not fully understood, however, mesolimbic dopamine alterations are strongly suspected. It has been long speculated that the high doses of dopamine needed to replete the relatively depleted dorsal striatum overdose the “intact” ventral striatum and cause this neuropsychiatric disorder [7–9]. The additional cognitive impairments in PD, which can include problems with attention, working memory, planning, forethought and decision-making, are faculties that can markedly increase susceptibility to impulse control disorder [8].

The role of serotonin deficiency in the PD brain and its part in inhibiting the patient’s ability to delay rewards adds to the complexity of impulse control disorder pathogenesis. Dorsal raphe nuclei disease in PD results in loss of serotonin innervation to substantial portions of the prefrontal and motor cortices in addition to basal ganglia substructures like the striatum, pallidum and subthalamic nucleus [10]. Together with dopamine, serotonin may work to regulate risk-sensitive decision making, response inhibition, waiting for future rewards, and overall impulse control. Its relative loss therefore also likely contributes to tipping the balance towards impulse dyscontrol [11,12]. The role of other neurotransmitters such as opiate systems involved in the process of acquisition and maintenance of addictive behaviors like dopamine dysregulation syndrome remains to be fully understood.

Treatment

The most successful strategy to address this problem is to reduce or eliminate the offending medication, usually the dopamine agonist. This may be associated with worsening apathy, anxiety or depression; however, substituting levodopa can be a successful strategy in many cases [13]. Zonisamide was described to be possibly effective in a trial of 15 subjects; however, the open label nature of this evidence must be considered as with other case reports using valproate, donepezil, and selective serotonin reuptake inhibitors (SSRIs) [14–16].

Fatigue

An easy to understand operational definition of fatigue is that it is a state of extreme tiredness, weakness, or exhaustion, either physical or mental or both. Fatigue is not uncommon in the general population [17] but is increasingly recognized to occur in numerous disease conditions and is frequently encountered in PD and multiple sclerosis. The latter is of special significance in the consideration of the neurotransmission of fatigue, as it is not thought to be a disease of dopamine deficiency. The pathophysiology remains unclear, and it may differ depending on whether the fatigue is experienced as more physical or mental, or rather motor versus nonmotor as some authors propose.

Fatigue has been conceptualized as central or peripheral in character. Peripheral fatigue is best understood as muscular fatigue caused by repetitive muscular contraction or reduced force generation [18]. Central fatigue however, is divided into mental or physical fatigue. Mental fatigue can occur after sustained attentive or emotional activity. It may alternatively be provoked after boring repetitive tasks or lack of intellectually stimulating activity. Physical fatigue is the sense of body exhaustion or energy to perform physical tasks even though the ability to carry them out exists.

Epidemiology

As recognition of the problem of fatigue increased in the last 2 decades, the realization that one-third to one-half of patients experience it at some point has improved opportunities for recognition and treatment [19]. Fatigue may be the presenting symptom in one-third of patients prior to actual motor symptom onset [20]. Half of untreated PD patients in a biomarker cohort study reported fatigue [6]. Unfortunately, it is also described by patients as one of the most disabling symptoms, causing significant impact on quality of life [19]. Fatigue in PD is associated with higher rates of depressive symptoms, but occurs with higher prevalence in nondepressed patients [21]. Poor ability to initiate and sustain activity due to fatigue is different from depression, excessive sleepiness, or impaired motor function [22,23].

Pathophysiology

The pathophysiology of fatigue remains somewhat unclear, though physical fatigue is likely a significant part of the problem and related to dopamine deficiency based on studies of time and force generation of keyboard strikes in PD subjects before and after L-dopa administration. These subjects had declines in force and increased physical fatigue which improved after L-dopa [24]. In other studies using transcranial magnetic stimulation to study changes in cortical excitability, the degree of physical fatigue correlated with abnormalities in motor evoked potentials during fatiguing exercising. These studies also support the hypothesis that fatigue is a motor symptom [25,26]. In the ELLDOPA study, fatigue worsened more in PD subjects treated with placebo [27]. Other imaging studies have suggested suggested nondopaminergic mechanisms including serotonergic pathway abnormalities [28], thus the question behind the etiology and solution for all cases of fatigue remains to be settled.

Diagnosis

The diagnosis is fatigue may be challenging as it may mask as depression or apathy. There are a number of fatigue rating scales available; however, the validated Parkinson’s Fatigue Scale (PFS) supersedes many of the problems of using a generic scale which could overlap motor questions and potentially be confounding [29,30].

Treatment

Most important is awareness and vigilance for the symptoms of fatigue, depression, and apathy and effort to distinguish between them. It may require structured interviews or assessment tools to properly diagnose the problem. Treatment is less clear in that few studies have clearly indicated the best treatment options. In placebo-controlled trials, methylphenidate did improve fatigue as did levodopa [31]. Modafinil, a hypocretin modulator and a drug first approved by the FDA for treatment of narcolepsy, has demonstrated mixed results in recent years. It may reduce physical fatigue and reduce excessive daytime sleepiness but likely does not reduce subjective symptoms of fatigue [32]. L-dopa can significantly reduce fatigue in many patients, which would argue that it often is a motor symptom [33,24]. In a post-hoc analysis of the ADAGIO delayed start study, patients taking rasagiline 1 mg/day and 2 mg/day (the latter dose exceeds the usual clinical dosing) showed significantly less worsening of symptoms on the PFS compared to placebo over time [34]. It is important to realize that once motor symptoms are optimally treated with dopaminergic medications, while many patients will feel significant relief from fatigue some patients will continue to feel symptomatic.

Apathy

The definition of apathy has become more complicated and refined, incorporating findings from the study of brain disease and behavioral analysis. Marin’s classic elaboration of apathy as lack of motivation not attributable to diminished level of consciousness, cognitive impairment, or emotional distress has been built upon by Levy and Dubois [35–37]. They suggest apathy may be better thought of as an observable behavioral syndrome characterized by a quantitative reduction of self-generated voluntary and purposeful behaviors. They suggest 3 apathetic subtypes: emotional, cognitive, and auto-activational, which reflect different disease states accounting for failure of normal goal-directed behavior.

Epidemiology

Prevalence estimates for apathy in PD vary. This is likely due to the varying recruitment criteria among studies, with some including patients with comorbid depression and dementia and others containing only “pure apathy.” Other reports may have had referral bias issues, as community-based studies report lower prevalences in general. In a group of newly diagnosed PD patients, using more restrictive criteria (apathy subscale of the neuropsychiatric inventory and the diagnostic consensus criteria for apathy validated in PD), Pedersen reported a prevalence of apathy of 14.3% [38]. In a 4-year prospective longitudinal cohort study, an annual incidence rate of 12.3% was reported, with apathy developing in 60% of the cohort by the study’s conclusion [39].

Apathy has been associated with longer disease duration, male gender [40], higher daily levodopa doses [41], more severe parkinsonism [38], and lower education status, though the latter feature remains under debate. Early cognitive deficits appear to be a risk factor for development of apathy [42]. The patterns of cognitive dysfunction and apathy remain unsettled in the literature.

Pathology

The pathology of atrophy remains unexplained and is unlikely to be reduced to a simple atrophy of one nucleus or the tone of one circuit. However, in a small neuroimaging study, severity of apathy correlated with atrophy of the bilateral nucleus accumbens [43], and it is notable that one major input to the nucleus accumbens is the amygdala. According to Braak staging, by stage 4 significant involvement of the amygdala by Lewy bodies has occurred. Others have found changes in grey matter density that could correlate with deficits of the prefrontal-basal ganglia circuitry to produce dysfunction of segregated frontal-subcortical loops. These may correlate with the “autoactivation” deficit pattern of apathy in which patients have a lack of self-initiated actions, even thoughts, though appear more normal when giving externally prompted responses [37,44].

Assessment

Clinically, the relationship between apathy and depression can be hard to disentangle, especially since many studies have found an association between them, especially with regards to apathy and anhedonia. Depression may feature negative self thoughts and sadness while apathy is notable for lack of initiation and effort. Viewed over a longer period of time, apathy tended to worsen in a linear fashion, where depression tended to fluctuate with improvements and exacerbations.

The Movement Disorders Society task force has recommended the Lille Apathy Rating Scale (LARS) for assessment of apathy; English and French versions have been validated in PD patients. It uses a semi-structured interview format assessing 4 dimensions of apathy: self awareness, intellectual curiosity, emotion, and action initiation [45–47].

The impact of apathy cannot be underestimated as this poor show of motivation or effort leads to lack of engagement in old activities or interest in new ones. Spouses may misinterpret this change in behavior as laziness or deliberate social withdrawal, or perhaps entitlement. It is not surprising that apathy routinely shows up on quality of life (QoL) questionnaires as highly impacting patients and families. In one study, apathy was the nonmotor symptom most likely to cause caregiver distress in PD [40,48–50].

Treatment

No approved drugs exist for treatment of apathy. However, clinical experience often confirms that dopaminergic modulation can be helpful in the treatment of apathy as indirect evidence suggests. A meta-analysis of controlled trials using pramipexole and Part I of the Unified Parkinson's disease rating scale (UPDRS) (secondary measure) showed the medication improved scores on this measure of motivation and mood in non-depressed subjects [51] with PD. Rare patients undergoing subthalamic deep brain stimulation have been reported to experience new and sometimes severe apathy after surgery [52]. This was posited at least in part to be the result of reduction of dopaminergic medication due to surgery.

Nondopaminergic pharmacotherapy of apathy is in its infancy. A recent controlled trial of rivastigmine in 31 French subjects with moderate to severe apathy based on LARS showed that 6 months of treatment at 9.5 mg/day improved average scores from –11.5 to –20 compared with placebo. While quality of life did not improve, caregiver burden did. The investigators found in this group of subjects that apathy was a possible herald for early dementia in PD [53].

A post-hoc analysis of the ADAGIO study (rasagiline or placebo in PD patients taking antidepressants) found that rasagiline use was associated with a nonsignificant slowing of apathy development during the trial [54].

Psychosis

Psychotic symptoms are a common occurrence in drug-treated patients, with visual hallucinations occurring in up to 30%, though over a 20-year period up to three-quarters of patients may develop visual hallucinations.After visual, the most common type of hallucination is auditory, followed by the other affected senses such as tactile, olfactory, or even taste [57]. Delusions, which tend to be paranoid in nature, occur in about 5% of patients [55–57]. The presence of psychotic symptoms is associated with poorer quality of life [58].

Symptomatology

The visual hallucinations of PD are usually quite stereotyped, and have been described as “minor” and “non-minor”[59]. Minor hallucinations refer to transient peripheral field stimuli that disappear when brought into central focus, “something flashed by,” a sense of a living being nearby, “a presence in the room,” or illusions whereby objects are transformed, eg, a bush in the yard is a deer.

Auditory hallucinations tend to be vague or indistinct sounds, like music in another room as opposed to voices speaking directly to the patient as might be experienced in a primary psychotic disorder. Tactile forms often involve insects or other animals crawling on the skin. Olfactory hallucinations may take the form of smelling perfume, toxic odors from room vents, etc.

Early in the experience, the visual hallucinations may be amusing in that they consistently remain nonthreatening, similar day to day, and sometimes oddly provide an aspect of comfort or companionship to the patient. More commonly, the hallucinations are bothersome to the patient because the experience indicates to the patient that there is something wrong with their mind. Visual hallucinations often begin in low-stimulus environments, often in the evening or other low-light conditions, but as the problem advances they can occur at any time of day. While visual hallucinations may initially occur for only seconds at a time many days apart, the frequency and duration can increase until they occur hours at a time every day and are accompanied by multiple other visual hallucinations, delusions, and confusion [60].

Delusions tend to be more distressing to patients and caregivers because they are often paranoid in nature. The patient is more likely to act out due to the anxiety the paranoia creates. For example, she may change passwords to online accounts due to a belief that unknown assailants are after her finances. He may go to great lengths trying to prove his wife is cheating.

Risk Factors

While the primary risk factor for psychotic symptom development is dementia [57], it occurs in nondemented patients. Other associations include reduced visual acuity [56], visual processing impairment [61–65], use of dopamine agonists, REM behavior disorder, duration of PD, axial rigidity subtype of disease [61,66–68]. The pathophysiology of psychosis in PD is likely complex and remains currently unexplained. The role of excess dopamine has been described above, but there is also data suggesting cholinergic deficits in the cortex may also contribute. Excess serotonin (increased 5HT2A receptor subtypes) in the temporal lobe within the visual processing pathway has been postulated to be of significance [69,70]. Hypometabolism in visual association areas of the brain in subjects with visual hallucinations has been demonstrated in PET and functional MRI studies [64,71]. This is similar to findings in patients with dementia with Lewy bodies [72].

This review focuses on the primary forms of PD-related psychosis, which occur with a clear sensorium and generally longer exposure to dopaminergic medication. It is important to distinguish 2 other common scenarios in which hallucinations or delusions may occur. In the common toxic-metabolic delirium, a clouded sensorium with attention deficits may be the only clue to the etiology of new onset confusion with visual hallucinations. It is highly likely that resolution of the underlying medical problem will lead to resolution of the new onset psychosis and encephalopathy. In a second scenario, hallucinations precede or occur very shortly after the onset of initiation of dopaminergic medication. This differs from the classic syndrome described earlier, in particular when visual hallucinations precede any initiation of medication, and likely represents the distinction between a diagnosis of Lewy body disease and PD [60].

Treatment

Management of psychosis is approachable, but often the outcome is unsatisfactory and associated with trade-offs in motor control. It is unfortunately true that psychotic symptoms are often associated with increased caregiver burden and are a cause of increased nursing home placements [73]. When considering the workup of psychotic symptoms, the differential diagnosis includes delirium, dream enactment (REM behavior disorder), or less commonly, Bonnet syndrome.

A delirium may be precipitated by a difficult to diagnose infection; new-onset confusion and psychotic symptoms may be the heralding presentation. Urinary tract or upper respiratory tract are common vulnerable sources of infection. Once infection is ruled out, the next practical step is to review the patient’s medication list and manage centrally acting drugs that could be contributing to the altered sensorium. A recent prescription of opioids for a dental treatment or a new muscle relaxant may be a culprit, though it is not that usual. A bladder anticholinergic could be suspect and is worth eliminating especially if its addition coincided with the appearance of the psychotic symptoms. Once the non-dopaminergic medications have been reduced/eliminated, then the PD medications should be considered. The general approach is to eliminate the medications that provide the least benefit while being more likely to contribute to psychotic symptoms. Anticholinergic medications, dopamine agonists, selegiline should all be uppermost in that consideration until one is left with L-dopa and COMT inhibitors (the latter function to increase levodopa availability). Then COMT inhibitors and levodopa can be reduced; however, at any point motor control can suffer with the loss of symptomatic therapies [74].

Clozapine is effective against psychotic symptoms in PD, at doses much lower than used in schizophrenia (300-600mg/day). The average dose in the US randomized controlled clinical trial was 25 mg/day, with no associated motor worsening. Patients in the United States are required by the FDA to be placed in a computer-based registry and monitored for agranulocytosis for the duration of clozapine therapy. This rare adverse event is not dose related. Orthostasis can occur at these low doses however. Fortunately the metabolic syndrome is not associated with this range of administration [75,76].

Quetiapine was not found to be effective in 3 blinded randomized controlled trials despite its rather common use for this purpose. It was not associated with motor worsening, however.

Other neuroleptic medications have not resulted in widespread use, because trials have been open label, or outcomes demonstrated motor worsening. Cholinesterase inhibitors have been the subject of a few positive case series, however results appear to be sporadic, the effect size is relatively small, and side effects of this medication class are common [77–79]. It is clear that there is an unmet need for a medication for psychotic symptoms. Clozapine is effective but onerous in its monitoring requirements. Practically speaking, there are relatively few PD patients who take advantage of it because of its feasibility challenges. Yet the problem of psychotic symptoms is a significant one that imposes important challenges to the patient and caregiver, and may limit the number of medications that the patient needs in order to optimize quality of life.

Pimavanserin, a novel medication which acts as a selective serotonin inverse agonist, is in the early application stages for FDA approval for treatment of psychotic symptoms in PD [80]. In its pivotal phase III controlled trial, the drug reduced not only positive symptoms (hallucinations/delusions) without causing motoric worsening, but also reduced caregiver burden. Pimavanserin improved certain sleep features without causing daytime sedation. If this drug meets final approval, it may present an exciting option for many patients for whom treatment was previously limited.

Depression

A study of early PD suggested that depression is often unrecognized and frequently untreated [1]. Indeed it is not unusual for depression to predate the diagnosis of PD by an average of 4 to 6 years [81]. Expanding to the larger PD population, it is generally accepted that about 30% to 50% of PD patients experience clinically significant depression, and once diagnosed may have a long term course, or may recur [82,83]. This is important as untreated depression is an important cause of poor quality of life in early PD. In addition depression can exacerbate motor disability, lead to earlier motor treatment with medication, and increase caregiver stress [83–85].

Diagnosis

Diagnosing depression in PD is more challenging as somatic, cognitive, and vegetative symptoms of altered mood can be imitated features in PD, such as facial masking, fatigue, sleep changes, weight loss, and working memory dysfunction (Table). Therefore, nonsomatic features tend to be more valuable to the clinician to query, such as the hedonic state (does the patient seek pleasurable activity?), mood, pessimism, state of hope, sense of capability[86,87]. Diagnosis of major and minor depression as well as dysthymia in PD using DSM-IV criteria has been validated [88]. To avoid underdiagnosis of this important commonly coexisting comorbidity, being inclusive with somatic overlapping features is still more important than trying to distinguish which disease state it belongs to.

A number of clinimetric rating scales for depression have been used and their advantages have been largely related to their objective nature (quantifiable); thus, they tend to be most useful in epidemiologic research studies or for larger scale screening purposes. Examples include the the Beck Depression Inventory, the Geriatric Depression Scale, and the Hamilton Depression Scale, all of which have been shown to be valid tools in PD (with the exception of the UPDRS Depression). It is important to note that they do not substitute for a diagnostic clinical interview [89].

Suicide is not common in PD, however suicidal ideation is estimated at about 11% in PD patients [90], and while there was concern initially after deep brain stimulation procedures began that suicide incidence was increased, evidence does not support this [91].

Pathophysiology

The pathophysiology of depression in PD is largely unknown however is thought to be less causally due to psychosocial factors and more etiologically driven by brainstem monoamine and serotonergic dysfunction [92]. Nonetheless, similar to other chronic conditions, PD patients can certainly develop fear of disability, guilt about impact on others, or other reactive mood changes. Overall, rates of depression are higher in PD compared with patients with similar conditions matched for disability [93].

Treatment

First, the clinician must determine if depression is a result of short-term fluctuations, chronic undertreatment of motor disease, or longer-term mood phenomenon. One important pattern to recognize are mood fluctuations, which can parallel motor OFF-ON cycling. It can be valuable to distinguish this as “subsyndromic” depression or anxiety (sometimes referred to as “OFF dysphoria”), as it can respond to improvement in antiparkinsonian medication dosing patterns that reduce fluctuations[94–96]. Similarly, elevating chronic motor undertreatment to goal therapy can result in mood normalization.

If symptoms persist despite optimization of motor/nonmotor fluctuations or chronic undertreatment and are severe enough to warrant treatment, then therapies used can range from nonpharmacologic education, support, and mental health referrals, as well as pharmacologic support in the form of medications.

A frequent but uncontrolled observation was that when undertreatment of motor disease was finally redressed, mood often improved. A multicenter randomized controlled trial of pramipexole in PD patients without motor fluctuations but with mild to moderate depressive symptoms showed the drug improved scores on the Beck Depression Inventory over a 12-week period. The improvement in mood was 6 points overall, but by 2 points over placebo, illustrating the importance of the size of the placebo effect [97]. Given the potential side effect profile of dopamine agonists, it may be useful to weigh the antidepressant effects only when their motor benefits are already being employed.

Controlled trials have demonstrated efficacy of both selective serotonin reuptake inhibitors (SSRI) and selective norepinephrine reuptake inhibitors (SNRI) antidepressants in PD. Clinical trials have demonstrated efficacy against placebo or with other antidepressant comparators. Examples of drugs with demonstrated efficacy include citalopram, paroxetine, venlafaxine, and nortriptyline. Results have attempted to illuminate the small unique differences between classes of antidepressants or dynamic properties between drugs within a class. For example, desipramine may nudge scores on a mood scale a few weeks sooner than a purer SSRI. Paroxetine (SSRI) versus venlafaxine (SNRI) improved mood scores comparably in a multicenter trial with a placebo comparator. In general, all have all been demonstrated to be effective and with a relatively low side-effect profile, comparable to the general population[98–102]. While case reports exist in the literature, the interaction of monoamine oxidase B inhibitors and SSRIs has not caused significant hypertensive crises or risk of serotonin syndrome [103,104]. Electroconvulsive therapy (ECT) can be used for severe refractory depression in PD as for non-PD patients, with case reports of very effective results. Due to the rarity of use, systematic evidence for its use is lacking [105,106].

Other novel agents and techniques such as omega-3 fatty acids [107] and repetitive transcranial magnetic stimulation [108] have been reported with promising early results. Cognitive behavioral therapy (CBT), which may involve stress management techniques, sleep hygiene, and caregiver support, additionally almost always provided improvement in measured outcomes, whether the trial was controlled or open label in design. In one RCT of CBT in PD of 14 weeks’ duration, there were significantly more treatment responders in the CBT group, with a number needed to treat of only 2 [109].

Anxiety

Anxiety is also common in PD, at least as common as depression considering that prevalence estimates suggest up to 50% of patients experience it [110–112]. Manifestations of anxiety may include panic attacks, generalized anxiety disorder, social anxiety, or other phobias [113]. Anxiety has an important negative impact on health-related quality of life and is often underrecognized by clinicians [114]. While reliable and valid scales to measure anxiety have been lacking in PD, a new effort has yielded the “Parkinson Anxiety Scale” though full clinimetric properties of the scale remain to be demonstrated (sensitivity to change) [115].

Anxiety that parallels the timing of motor OFF-ON cycling is important to recognize. This “subsyndromic” anxiety or anxiety disorder not otherwise specified (ie, the anxiety does not meet DSM-IV criteria) can respond to improvement in antiparkinsonian medication dosing patterns that reduce fluctuations [116,117]. Indeed, the presence of motor fluctuations is the principle marker of anxiety in many studies [118–120]. In an analogous manner, anxiety can predate PD by years and be part of the nonmotor amalgam of features heralding the disease [6,121].

Treatment

Systematic controlled trials of anxiolytic treatment for PD are lacking; therefore, SSRIs are prescribed for this purpose as in non-PD patients. Until SSRIs are demonstrated to be of benefit in anxiety, they are likely safer than use of benzodiazepines, which are associated with risk for falling, cognitive dysfunction, or autonomic dysregulation in PD patients when used during waking hours. Psychotherapy and other nonpharmacologic approaches are likely to be of benefit. A small study of neuromuscular (massage) therapy demonstrated improvement on the Beck Anxiety Inventory in PD [122]. A case report of ECT for severe anxiety has been published [123].

Conclusion

Neuropsychiatric symptoms are common in PD and new knowledge about clinical features, epidemiology, and treatment options has been gained in the last decade, though much remains to be discovered. The development of valid instruments to measure neuropsychiatric symptoms has been vital in these research efforts to bridge the gaps in our understanding. Further elucidation of the pathophysiologies of neuropsychiatric symptoms will help to define treatment targets and likely fuel drug development and the discovery of drugs with more potent benefit and fewer side effects.

 

Corresponding author: Kathryn A. Chung, MD, Department of Neurology, Oregon Health & Science University, Portland, OR, [email protected].

Financial disclosures: None.

From the Department of Neurology, Oregon Health & Science University, Portland, OR.

 

Abstract

• Objective: To review the clinical characteristics, epidemiology, and management of the most common neuropsychiatric symptoms (NPS) in Parkinson’s disease (PD).
• Methods: Literature review.
• Results: PD has traditionally been considered a disease of impaired motor function. However, neuropsychiatric complications, such as fatigue, depression, anxiety, psychosis, impulse control disorders, and apathy, frequently complicate the course of the illness. Although the development of new medication options in recent years has had a positive benefit on the management of these troublesome symptoms, responses are frequently suboptimal. The development of valid instruments to measure neuropsychiatric symptoms has been vital in research efforts to bridge the gaps in our understanding. Further elucidation of neuropsychiatric pathophysiologies will help to define treatment targets and has the potential to expand our therapeutic armamentarium.
• Conclusion: While NPS affect patients with established disease, recent investigations have demonstrated risk of symptoms in those with early untreated stages of PD; therefore, better understanding of NPS should be the goal of practitioners treating the entire continuum of PD.

Parkinson’s disease (PD) has traditionally been considered a disease of impaired motor function, but increased recognition of nonmotor symptoms and in particular neuropsychiatric symptoms, such as fatigue, depression, anxiety, psychosis, impulse control disorders, and apathy, offer new opportunities for better care of patients. While neuropsychiatric symptoms affect patients with established disease, recent investigations have clearly demonstrated risk of symptoms in those with early untreated stages of PD; therefore, better understanding of neuropsychiatric symptoms should be the goal of practitioners treating the entire continuum of PD. This review will focus on the clinical characteristics, epidemiology, and management of the most common neuropsychiatric symptoms in PD.

Impulse Control Disorders

The recognition that dopaminergic drugs were successful at treating many symptoms of PD was followed by the disturbing realization that impulse control disorders could be an unfortunate side effect in a substantial minority. Impulse control disorders as defined by DSM-IV [1] are disinhibited behaviors that are maladaptive and recurrent, causing personal and relationship consequences. The impulse control disorders that became associated with PD and medication intake, particularly dopamine agonist use, included gambling, hobbyism, punding (stereotyped, seemingly purposeless behaviors), excessive sexual behavior, shopping, hoarding, and less commonly, compulsive eating. The prevalence estimates of these behavioral disturbances range from 6% to 15.5%, compared with < 2% in the general population [2,3]. The addiction-like dopamine dysregulation syndrome, whereby patients self-medicate with high doses of levodopa and short-acting dopamine agonists beyond what is needed for motor control, can lead to significant impairment of the therapeutic alliance in addition to other patient personal relations. With the advent of surgical options to treat PD and its medication complications, it was observed that stimulation of the subthalamic nucleus could be associated with the spectrum of impulse control disorders [4].

Epidemiology/Risk Factors

In a recent systematic review of the literature of impulse control disorders in PD [5], the authors determined that dopaminergic therapy caused compulsive or impulsive behaviors in approximately 10% of PD patients in the course of their treatment, with pathologic gambling and hypersexuality most frequently experienced. Multiple impulse control disorders are not uncommon and may coexist in one-quarter of patients with compulsivity. There appeared to be more disordered behavior with higher comparable doses of agonists. The authors concluded that impulse control disorder symptoms tended to occur with initiation or dose increases of direct D₂/D₃ agonists, such as pramipexole and ropinirole. Importantly, impulse control disorder behavior improved if not resolved with discontinuation or reduction of dosage of the agonist, even if a compensatory levodopa dosage is added or increased. Perhaps not surprisingly, it was observed that if patients had a preexisting impulse control disorder prior to PD or the initiation of treatment, there was a high likelihood of worsening of symptoms. This small subgroup is estimated at about 1% of PD subjects, which corresponds to the prevalence of impulse control disorders in the general population. Other identified potential risk factors for impulse control disorder development include male gender, young age at onset, a personal or family history of addiction, novelty or risk seeking personality, and a concurrent diagnosis of depression [3]. In a recent study of early PD patients, the risk of developing an impulse control disorder became important once treatment with dopaminergic drugs began and continued for a year or more [6].

Pathogenesis

The pathogenesis is not fully understood, however, mesolimbic dopamine alterations are strongly suspected. It has been long speculated that the high doses of dopamine needed to replete the relatively depleted dorsal striatum overdose the “intact” ventral striatum and cause this neuropsychiatric disorder [7–9]. The additional cognitive impairments in PD, which can include problems with attention, working memory, planning, forethought and decision-making, are faculties that can markedly increase susceptibility to impulse control disorder [8].

The role of serotonin deficiency in the PD brain and its part in inhibiting the patient’s ability to delay rewards adds to the complexity of impulse control disorder pathogenesis. Dorsal raphe nuclei disease in PD results in loss of serotonin innervation to substantial portions of the prefrontal and motor cortices in addition to basal ganglia substructures like the striatum, pallidum and subthalamic nucleus [10]. Together with dopamine, serotonin may work to regulate risk-sensitive decision making, response inhibition, waiting for future rewards, and overall impulse control. Its relative loss therefore also likely contributes to tipping the balance towards impulse dyscontrol [11,12]. The role of other neurotransmitters such as opiate systems involved in the process of acquisition and maintenance of addictive behaviors like dopamine dysregulation syndrome remains to be fully understood.

Treatment

The most successful strategy to address this problem is to reduce or eliminate the offending medication, usually the dopamine agonist. This may be associated with worsening apathy, anxiety or depression; however, substituting levodopa can be a successful strategy in many cases [13]. Zonisamide was described to be possibly effective in a trial of 15 subjects; however, the open label nature of this evidence must be considered as with other case reports using valproate, donepezil, and selective serotonin reuptake inhibitors (SSRIs) [14–16].

Fatigue

An easy to understand operational definition of fatigue is that it is a state of extreme tiredness, weakness, or exhaustion, either physical or mental or both. Fatigue is not uncommon in the general population [17] but is increasingly recognized to occur in numerous disease conditions and is frequently encountered in PD and multiple sclerosis. The latter is of special significance in the consideration of the neurotransmission of fatigue, as it is not thought to be a disease of dopamine deficiency. The pathophysiology remains unclear, and it may differ depending on whether the fatigue is experienced as more physical or mental, or rather motor versus nonmotor as some authors propose.

Fatigue has been conceptualized as central or peripheral in character. Peripheral fatigue is best understood as muscular fatigue caused by repetitive muscular contraction or reduced force generation [18]. Central fatigue however, is divided into mental or physical fatigue. Mental fatigue can occur after sustained attentive or emotional activity. It may alternatively be provoked after boring repetitive tasks or lack of intellectually stimulating activity. Physical fatigue is the sense of body exhaustion or energy to perform physical tasks even though the ability to carry them out exists.

Epidemiology

As recognition of the problem of fatigue increased in the last 2 decades, the realization that one-third to one-half of patients experience it at some point has improved opportunities for recognition and treatment [19]. Fatigue may be the presenting symptom in one-third of patients prior to actual motor symptom onset [20]. Half of untreated PD patients in a biomarker cohort study reported fatigue [6]. Unfortunately, it is also described by patients as one of the most disabling symptoms, causing significant impact on quality of life [19]. Fatigue in PD is associated with higher rates of depressive symptoms, but occurs with higher prevalence in nondepressed patients [21]. Poor ability to initiate and sustain activity due to fatigue is different from depression, excessive sleepiness, or impaired motor function [22,23].

Pathophysiology

The pathophysiology of fatigue remains somewhat unclear, though physical fatigue is likely a significant part of the problem and related to dopamine deficiency based on studies of time and force generation of keyboard strikes in PD subjects before and after L-dopa administration. These subjects had declines in force and increased physical fatigue which improved after L-dopa [24]. In other studies using transcranial magnetic stimulation to study changes in cortical excitability, the degree of physical fatigue correlated with abnormalities in motor evoked potentials during fatiguing exercising. These studies also support the hypothesis that fatigue is a motor symptom [25,26]. In the ELLDOPA study, fatigue worsened more in PD subjects treated with placebo [27]. Other imaging studies have suggested suggested nondopaminergic mechanisms including serotonergic pathway abnormalities [28], thus the question behind the etiology and solution for all cases of fatigue remains to be settled.

Diagnosis

The diagnosis is fatigue may be challenging as it may mask as depression or apathy. There are a number of fatigue rating scales available; however, the validated Parkinson’s Fatigue Scale (PFS) supersedes many of the problems of using a generic scale which could overlap motor questions and potentially be confounding [29,30].

Treatment

Most important is awareness and vigilance for the symptoms of fatigue, depression, and apathy and effort to distinguish between them. It may require structured interviews or assessment tools to properly diagnose the problem. Treatment is less clear in that few studies have clearly indicated the best treatment options. In placebo-controlled trials, methylphenidate did improve fatigue as did levodopa [31]. Modafinil, a hypocretin modulator and a drug first approved by the FDA for treatment of narcolepsy, has demonstrated mixed results in recent years. It may reduce physical fatigue and reduce excessive daytime sleepiness but likely does not reduce subjective symptoms of fatigue [32]. L-dopa can significantly reduce fatigue in many patients, which would argue that it often is a motor symptom [33,24]. In a post-hoc analysis of the ADAGIO delayed start study, patients taking rasagiline 1 mg/day and 2 mg/day (the latter dose exceeds the usual clinical dosing) showed significantly less worsening of symptoms on the PFS compared to placebo over time [34]. It is important to realize that once motor symptoms are optimally treated with dopaminergic medications, while many patients will feel significant relief from fatigue some patients will continue to feel symptomatic.

Apathy

The definition of apathy has become more complicated and refined, incorporating findings from the study of brain disease and behavioral analysis. Marin’s classic elaboration of apathy as lack of motivation not attributable to diminished level of consciousness, cognitive impairment, or emotional distress has been built upon by Levy and Dubois [35–37]. They suggest apathy may be better thought of as an observable behavioral syndrome characterized by a quantitative reduction of self-generated voluntary and purposeful behaviors. They suggest 3 apathetic subtypes: emotional, cognitive, and auto-activational, which reflect different disease states accounting for failure of normal goal-directed behavior.

Epidemiology

Prevalence estimates for apathy in PD vary. This is likely due to the varying recruitment criteria among studies, with some including patients with comorbid depression and dementia and others containing only “pure apathy.” Other reports may have had referral bias issues, as community-based studies report lower prevalences in general. In a group of newly diagnosed PD patients, using more restrictive criteria (apathy subscale of the neuropsychiatric inventory and the diagnostic consensus criteria for apathy validated in PD), Pedersen reported a prevalence of apathy of 14.3% [38]. In a 4-year prospective longitudinal cohort study, an annual incidence rate of 12.3% was reported, with apathy developing in 60% of the cohort by the study’s conclusion [39].

Apathy has been associated with longer disease duration, male gender [40], higher daily levodopa doses [41], more severe parkinsonism [38], and lower education status, though the latter feature remains under debate. Early cognitive deficits appear to be a risk factor for development of apathy [42]. The patterns of cognitive dysfunction and apathy remain unsettled in the literature.

Pathology

The pathology of atrophy remains unexplained and is unlikely to be reduced to a simple atrophy of one nucleus or the tone of one circuit. However, in a small neuroimaging study, severity of apathy correlated with atrophy of the bilateral nucleus accumbens [43], and it is notable that one major input to the nucleus accumbens is the amygdala. According to Braak staging, by stage 4 significant involvement of the amygdala by Lewy bodies has occurred. Others have found changes in grey matter density that could correlate with deficits of the prefrontal-basal ganglia circuitry to produce dysfunction of segregated frontal-subcortical loops. These may correlate with the “autoactivation” deficit pattern of apathy in which patients have a lack of self-initiated actions, even thoughts, though appear more normal when giving externally prompted responses [37,44].

Assessment

Clinically, the relationship between apathy and depression can be hard to disentangle, especially since many studies have found an association between them, especially with regards to apathy and anhedonia. Depression may feature negative self thoughts and sadness while apathy is notable for lack of initiation and effort. Viewed over a longer period of time, apathy tended to worsen in a linear fashion, where depression tended to fluctuate with improvements and exacerbations.

The Movement Disorders Society task force has recommended the Lille Apathy Rating Scale (LARS) for assessment of apathy; English and French versions have been validated in PD patients. It uses a semi-structured interview format assessing 4 dimensions of apathy: self awareness, intellectual curiosity, emotion, and action initiation [45–47].

The impact of apathy cannot be underestimated as this poor show of motivation or effort leads to lack of engagement in old activities or interest in new ones. Spouses may misinterpret this change in behavior as laziness or deliberate social withdrawal, or perhaps entitlement. It is not surprising that apathy routinely shows up on quality of life (QoL) questionnaires as highly impacting patients and families. In one study, apathy was the nonmotor symptom most likely to cause caregiver distress in PD [40,48–50].

Treatment

No approved drugs exist for treatment of apathy. However, clinical experience often confirms that dopaminergic modulation can be helpful in the treatment of apathy as indirect evidence suggests. A meta-analysis of controlled trials using pramipexole and Part I of the Unified Parkinson's disease rating scale (UPDRS) (secondary measure) showed the medication improved scores on this measure of motivation and mood in non-depressed subjects [51] with PD. Rare patients undergoing subthalamic deep brain stimulation have been reported to experience new and sometimes severe apathy after surgery [52]. This was posited at least in part to be the result of reduction of dopaminergic medication due to surgery.

Nondopaminergic pharmacotherapy of apathy is in its infancy. A recent controlled trial of rivastigmine in 31 French subjects with moderate to severe apathy based on LARS showed that 6 months of treatment at 9.5 mg/day improved average scores from –11.5 to –20 compared with placebo. While quality of life did not improve, caregiver burden did. The investigators found in this group of subjects that apathy was a possible herald for early dementia in PD [53].

A post-hoc analysis of the ADAGIO study (rasagiline or placebo in PD patients taking antidepressants) found that rasagiline use was associated with a nonsignificant slowing of apathy development during the trial [54].

Psychosis

Psychotic symptoms are a common occurrence in drug-treated patients, with visual hallucinations occurring in up to 30%, though over a 20-year period up to three-quarters of patients may develop visual hallucinations.After visual, the most common type of hallucination is auditory, followed by the other affected senses such as tactile, olfactory, or even taste [57]. Delusions, which tend to be paranoid in nature, occur in about 5% of patients [55–57]. The presence of psychotic symptoms is associated with poorer quality of life [58].

Symptomatology

The visual hallucinations of PD are usually quite stereotyped, and have been described as “minor” and “non-minor”[59]. Minor hallucinations refer to transient peripheral field stimuli that disappear when brought into central focus, “something flashed by,” a sense of a living being nearby, “a presence in the room,” or illusions whereby objects are transformed, eg, a bush in the yard is a deer.

Auditory hallucinations tend to be vague or indistinct sounds, like music in another room as opposed to voices speaking directly to the patient as might be experienced in a primary psychotic disorder. Tactile forms often involve insects or other animals crawling on the skin. Olfactory hallucinations may take the form of smelling perfume, toxic odors from room vents, etc.

Early in the experience, the visual hallucinations may be amusing in that they consistently remain nonthreatening, similar day to day, and sometimes oddly provide an aspect of comfort or companionship to the patient. More commonly, the hallucinations are bothersome to the patient because the experience indicates to the patient that there is something wrong with their mind. Visual hallucinations often begin in low-stimulus environments, often in the evening or other low-light conditions, but as the problem advances they can occur at any time of day. While visual hallucinations may initially occur for only seconds at a time many days apart, the frequency and duration can increase until they occur hours at a time every day and are accompanied by multiple other visual hallucinations, delusions, and confusion [60].

Delusions tend to be more distressing to patients and caregivers because they are often paranoid in nature. The patient is more likely to act out due to the anxiety the paranoia creates. For example, she may change passwords to online accounts due to a belief that unknown assailants are after her finances. He may go to great lengths trying to prove his wife is cheating.

Risk Factors

While the primary risk factor for psychotic symptom development is dementia [57], it occurs in nondemented patients. Other associations include reduced visual acuity [56], visual processing impairment [61–65], use of dopamine agonists, REM behavior disorder, duration of PD, axial rigidity subtype of disease [61,66–68]. The pathophysiology of psychosis in PD is likely complex and remains currently unexplained. The role of excess dopamine has been described above, but there is also data suggesting cholinergic deficits in the cortex may also contribute. Excess serotonin (increased 5HT2A receptor subtypes) in the temporal lobe within the visual processing pathway has been postulated to be of significance [69,70]. Hypometabolism in visual association areas of the brain in subjects with visual hallucinations has been demonstrated in PET and functional MRI studies [64,71]. This is similar to findings in patients with dementia with Lewy bodies [72].

This review focuses on the primary forms of PD-related psychosis, which occur with a clear sensorium and generally longer exposure to dopaminergic medication. It is important to distinguish 2 other common scenarios in which hallucinations or delusions may occur. In the common toxic-metabolic delirium, a clouded sensorium with attention deficits may be the only clue to the etiology of new onset confusion with visual hallucinations. It is highly likely that resolution of the underlying medical problem will lead to resolution of the new onset psychosis and encephalopathy. In a second scenario, hallucinations precede or occur very shortly after the onset of initiation of dopaminergic medication. This differs from the classic syndrome described earlier, in particular when visual hallucinations precede any initiation of medication, and likely represents the distinction between a diagnosis of Lewy body disease and PD [60].

Treatment

Management of psychosis is approachable, but often the outcome is unsatisfactory and associated with trade-offs in motor control. It is unfortunately true that psychotic symptoms are often associated with increased caregiver burden and are a cause of increased nursing home placements [73]. When considering the workup of psychotic symptoms, the differential diagnosis includes delirium, dream enactment (REM behavior disorder), or less commonly, Bonnet syndrome.

A delirium may be precipitated by a difficult to diagnose infection; new-onset confusion and psychotic symptoms may be the heralding presentation. Urinary tract or upper respiratory tract are common vulnerable sources of infection. Once infection is ruled out, the next practical step is to review the patient’s medication list and manage centrally acting drugs that could be contributing to the altered sensorium. A recent prescription of opioids for a dental treatment or a new muscle relaxant may be a culprit, though it is not that usual. A bladder anticholinergic could be suspect and is worth eliminating especially if its addition coincided with the appearance of the psychotic symptoms. Once the non-dopaminergic medications have been reduced/eliminated, then the PD medications should be considered. The general approach is to eliminate the medications that provide the least benefit while being more likely to contribute to psychotic symptoms. Anticholinergic medications, dopamine agonists, selegiline should all be uppermost in that consideration until one is left with L-dopa and COMT inhibitors (the latter function to increase levodopa availability). Then COMT inhibitors and levodopa can be reduced; however, at any point motor control can suffer with the loss of symptomatic therapies [74].

Clozapine is effective against psychotic symptoms in PD, at doses much lower than used in schizophrenia (300-600mg/day). The average dose in the US randomized controlled clinical trial was 25 mg/day, with no associated motor worsening. Patients in the United States are required by the FDA to be placed in a computer-based registry and monitored for agranulocytosis for the duration of clozapine therapy. This rare adverse event is not dose related. Orthostasis can occur at these low doses however. Fortunately the metabolic syndrome is not associated with this range of administration [75,76].

Quetiapine was not found to be effective in 3 blinded randomized controlled trials despite its rather common use for this purpose. It was not associated with motor worsening, however.

Other neuroleptic medications have not resulted in widespread use, because trials have been open label, or outcomes demonstrated motor worsening. Cholinesterase inhibitors have been the subject of a few positive case series, however results appear to be sporadic, the effect size is relatively small, and side effects of this medication class are common [77–79]. It is clear that there is an unmet need for a medication for psychotic symptoms. Clozapine is effective but onerous in its monitoring requirements. Practically speaking, there are relatively few PD patients who take advantage of it because of its feasibility challenges. Yet the problem of psychotic symptoms is a significant one that imposes important challenges to the patient and caregiver, and may limit the number of medications that the patient needs in order to optimize quality of life.

Pimavanserin, a novel medication which acts as a selective serotonin inverse agonist, is in the early application stages for FDA approval for treatment of psychotic symptoms in PD [80]. In its pivotal phase III controlled trial, the drug reduced not only positive symptoms (hallucinations/delusions) without causing motoric worsening, but also reduced caregiver burden. Pimavanserin improved certain sleep features without causing daytime sedation. If this drug meets final approval, it may present an exciting option for many patients for whom treatment was previously limited.

Depression

A study of early PD suggested that depression is often unrecognized and frequently untreated [1]. Indeed it is not unusual for depression to predate the diagnosis of PD by an average of 4 to 6 years [81]. Expanding to the larger PD population, it is generally accepted that about 30% to 50% of PD patients experience clinically significant depression, and once diagnosed may have a long term course, or may recur [82,83]. This is important as untreated depression is an important cause of poor quality of life in early PD. In addition depression can exacerbate motor disability, lead to earlier motor treatment with medication, and increase caregiver stress [83–85].

Diagnosis

Diagnosing depression in PD is more challenging as somatic, cognitive, and vegetative symptoms of altered mood can be imitated features in PD, such as facial masking, fatigue, sleep changes, weight loss, and working memory dysfunction (Table). Therefore, nonsomatic features tend to be more valuable to the clinician to query, such as the hedonic state (does the patient seek pleasurable activity?), mood, pessimism, state of hope, sense of capability[86,87]. Diagnosis of major and minor depression as well as dysthymia in PD using DSM-IV criteria has been validated [88]. To avoid underdiagnosis of this important commonly coexisting comorbidity, being inclusive with somatic overlapping features is still more important than trying to distinguish which disease state it belongs to.

A number of clinimetric rating scales for depression have been used and their advantages have been largely related to their objective nature (quantifiable); thus, they tend to be most useful in epidemiologic research studies or for larger scale screening purposes. Examples include the the Beck Depression Inventory, the Geriatric Depression Scale, and the Hamilton Depression Scale, all of which have been shown to be valid tools in PD (with the exception of the UPDRS Depression). It is important to note that they do not substitute for a diagnostic clinical interview [89].

Suicide is not common in PD, however suicidal ideation is estimated at about 11% in PD patients [90], and while there was concern initially after deep brain stimulation procedures began that suicide incidence was increased, evidence does not support this [91].

Pathophysiology

The pathophysiology of depression in PD is largely unknown however is thought to be less causally due to psychosocial factors and more etiologically driven by brainstem monoamine and serotonergic dysfunction [92]. Nonetheless, similar to other chronic conditions, PD patients can certainly develop fear of disability, guilt about impact on others, or other reactive mood changes. Overall, rates of depression are higher in PD compared with patients with similar conditions matched for disability [93].

Treatment

First, the clinician must determine if depression is a result of short-term fluctuations, chronic undertreatment of motor disease, or longer-term mood phenomenon. One important pattern to recognize are mood fluctuations, which can parallel motor OFF-ON cycling. It can be valuable to distinguish this as “subsyndromic” depression or anxiety (sometimes referred to as “OFF dysphoria”), as it can respond to improvement in antiparkinsonian medication dosing patterns that reduce fluctuations[94–96]. Similarly, elevating chronic motor undertreatment to goal therapy can result in mood normalization.

If symptoms persist despite optimization of motor/nonmotor fluctuations or chronic undertreatment and are severe enough to warrant treatment, then therapies used can range from nonpharmacologic education, support, and mental health referrals, as well as pharmacologic support in the form of medications.

A frequent but uncontrolled observation was that when undertreatment of motor disease was finally redressed, mood often improved. A multicenter randomized controlled trial of pramipexole in PD patients without motor fluctuations but with mild to moderate depressive symptoms showed the drug improved scores on the Beck Depression Inventory over a 12-week period. The improvement in mood was 6 points overall, but by 2 points over placebo, illustrating the importance of the size of the placebo effect [97]. Given the potential side effect profile of dopamine agonists, it may be useful to weigh the antidepressant effects only when their motor benefits are already being employed.

Controlled trials have demonstrated efficacy of both selective serotonin reuptake inhibitors (SSRI) and selective norepinephrine reuptake inhibitors (SNRI) antidepressants in PD. Clinical trials have demonstrated efficacy against placebo or with other antidepressant comparators. Examples of drugs with demonstrated efficacy include citalopram, paroxetine, venlafaxine, and nortriptyline. Results have attempted to illuminate the small unique differences between classes of antidepressants or dynamic properties between drugs within a class. For example, desipramine may nudge scores on a mood scale a few weeks sooner than a purer SSRI. Paroxetine (SSRI) versus venlafaxine (SNRI) improved mood scores comparably in a multicenter trial with a placebo comparator. In general, all have all been demonstrated to be effective and with a relatively low side-effect profile, comparable to the general population[98–102]. While case reports exist in the literature, the interaction of monoamine oxidase B inhibitors and SSRIs has not caused significant hypertensive crises or risk of serotonin syndrome [103,104]. Electroconvulsive therapy (ECT) can be used for severe refractory depression in PD as for non-PD patients, with case reports of very effective results. Due to the rarity of use, systematic evidence for its use is lacking [105,106].

Other novel agents and techniques such as omega-3 fatty acids [107] and repetitive transcranial magnetic stimulation [108] have been reported with promising early results. Cognitive behavioral therapy (CBT), which may involve stress management techniques, sleep hygiene, and caregiver support, additionally almost always provided improvement in measured outcomes, whether the trial was controlled or open label in design. In one RCT of CBT in PD of 14 weeks’ duration, there were significantly more treatment responders in the CBT group, with a number needed to treat of only 2 [109].

Anxiety

Anxiety is also common in PD, at least as common as depression considering that prevalence estimates suggest up to 50% of patients experience it [110–112]. Manifestations of anxiety may include panic attacks, generalized anxiety disorder, social anxiety, or other phobias [113]. Anxiety has an important negative impact on health-related quality of life and is often underrecognized by clinicians [114]. While reliable and valid scales to measure anxiety have been lacking in PD, a new effort has yielded the “Parkinson Anxiety Scale” though full clinimetric properties of the scale remain to be demonstrated (sensitivity to change) [115].

Anxiety that parallels the timing of motor OFF-ON cycling is important to recognize. This “subsyndromic” anxiety or anxiety disorder not otherwise specified (ie, the anxiety does not meet DSM-IV criteria) can respond to improvement in antiparkinsonian medication dosing patterns that reduce fluctuations [116,117]. Indeed, the presence of motor fluctuations is the principle marker of anxiety in many studies [118–120]. In an analogous manner, anxiety can predate PD by years and be part of the nonmotor amalgam of features heralding the disease [6,121].

Treatment

Systematic controlled trials of anxiolytic treatment for PD are lacking; therefore, SSRIs are prescribed for this purpose as in non-PD patients. Until SSRIs are demonstrated to be of benefit in anxiety, they are likely safer than use of benzodiazepines, which are associated with risk for falling, cognitive dysfunction, or autonomic dysregulation in PD patients when used during waking hours. Psychotherapy and other nonpharmacologic approaches are likely to be of benefit. A small study of neuromuscular (massage) therapy demonstrated improvement on the Beck Anxiety Inventory in PD [122]. A case report of ECT for severe anxiety has been published [123].

Conclusion

Neuropsychiatric symptoms are common in PD and new knowledge about clinical features, epidemiology, and treatment options has been gained in the last decade, though much remains to be discovered. The development of valid instruments to measure neuropsychiatric symptoms has been vital in these research efforts to bridge the gaps in our understanding. Further elucidation of the pathophysiologies of neuropsychiatric symptoms will help to define treatment targets and likely fuel drug development and the discovery of drugs with more potent benefit and fewer side effects.

 

Corresponding author: Kathryn A. Chung, MD, Department of Neurology, Oregon Health & Science University, Portland, OR, [email protected].

Financial disclosures: None.

From the Center for Narcolepsy, Sleep and Health Research, Department of Biobehavioral Health Science, University of Illinois at Chicago, Chicago, IL.

 

Abstract

• Objective: To review the efficacy of current treatment options for adults with obstructive sleep apnea (OSA).
• Methods: Review of the literature.
• Results: OSA, characterized by repetitive ≥ 10-second interruptions (apnea) or reductions (hypopnea) in airflow, is initiated by partial or complete collapse in the upper airway despite respiratory effort. When left untreated, OSA is associated with comorbid conditions, such as cardiovascular and metabolic diseases. The current “gold standard” treatment for OSA is continuous positive air pressure (CPAP), which pneumatically stabilizes the upper airways. CPAP has proven efficacy and potential cost savings via decreases in health comorbidities and/or motor-vehicle crashes. However, CPAP treatment is not well-tolerated due to various side effects, and adherence among OSA subjects can be as low as 50% in certain populations. Other treatment options for OSA include improving CPAP tolerability, increasing CPAP adherence through patient interventions, weight loss/exercise, positional therapy, nasal expiratory positive airway pressure, oral pressure therapy, oral appliances, surgery, hypoglossal nerve stimulation, drug treatment, and combining 2 or more of the aforementioned treatments. Despite the many options available to treat OSA, none of them are as efficacious as CPAP. However, many of these treatments are tolerable, and adherence rates are higher than those of the CPAP, making them a more viable treatment option for long-term use.
• Conclusion: Patients need to weigh the benefits and risks of available treatments for OSA. More large randomized controlled studies on treatments or combination of treatments for OSA are needed that measure parameters such as treatment adherence, apnea-hypopnea index, oxygen desaturation, subjective sleepiness, quality of life, and adverse events.

 

Obstructive sleep apnea (OSA), characterized by repetitive ≥ 10-second interruptions (apnea) or reductions (hypopnea) in airflow (measured as events/hour, called the apnea-hypopnea index [AHI]), is initiated by partial or complete collapse in the upper airway despite respiratory effort [1]. Current estimates of the prevalence of OSA (AHI ≥ 5 and Epworth Sleepiness Scale > 10) in American men and women (aged 30–70 years) are 14% and 5%, respectively, with prevalence rates increasing due to increasing rates of obesity, a risk factor for developing OSA [2]. Hypoxemia/hypercapnia, fragmented sleep, as well as exaggerated fluctuations in heart rhythm, blood pressure, and intrathoracic pressure are some of the acute physiological effects of untreated OSA [1]. These acute effects can develop into long-term sequelae, such as hypertension and other cardiovascular comorbidities [2,3], decrements in cognitive function [4,5], poor mood, reduced quality of life [6,7], and premature death [8,9]. In economic terms, health care cost estimates of OSA and its associated comorbidities rival that of diabetes [10]. Additionally, in the year 2000, more than 800,000 drivers were involved OSA-related motor-vehicle collisions, of which more than 1400 fatalities occurred [11].

Front-line treatment of OSA relies on mechanically stabilizing the upper airway with a column of air via continuous positive airway pressure (CPAP) treatment. Though CPAP is the “gold standard” treatment for OSA with proven efficacy and potential cost savings via decreases in health comorbidities and/or motor-vehicle crashes [10–12], CPAP treatment is not well-tolerated due to various side effects [13–15]. Adherence among OSA subjects can be as low as 50% in certain populations [16–18]. Improved strategies for current and innovative treatments have emerged in the last few years and are the subject of this review.

 

 

Improved CPAP Treatment

As stated previously, CPAP pneumatically splints the upper airway, thus preventing it from collapsing during sleep. However, CPAP is not well-tolerated. Modifications to standard CPAP to increase adherence have been met with disappointing results. Humidification with heated tubing delivering heated moistened air did not increase compliance compared to standard CPAP [19]. CPAP was also compared with auto-adjusting CPAP (APAP), where respiration is monitored and the minimum pressure of air is applied to splint the upper airway open. In a meta-analysis, APAP only had very small effect on compliance [20]. Lastly, reduction in pressure during expiration was investigated, and a meta-analysis showed no effect [21,22]. However, recent advances in CPAP delivery give hope to increasing compliance. The S9CPAP machine (Resmed, San Diego, CA), which combines a humidification system and an APAP, showed increased compliance compared to standard CPAP. Compliance increased by an average of 30 minutes per night, and variance of daily usage decreased (eg, patients used it more day-to-day) [23]. However, a randomized blinded study needs to be conducted to corroborate these results.

Promoting CPAP Adherence Through Patient Interventions

Educational, supportive, and behavioral interventions have been used to increase CPAP adherence and have been thoroughly reviewed via meta-analysis [24]. Briefly, 30 studies of various interventions were included and demonstrated that educational, supportive, or behavioral interventions increased CPAP usage in OSA-naive patients. Behavioral interventions increased CPAP usage by over an hour, but the evidence was of “low-quality.” Educational and supportive interventions also increased CPAP usage, with the former having “moderate-quality” evidence [24]. However, whether increased CPAP usage had an effect on symptoms and quality of life was statistically unclear, and the authors recommended further assessment [24]. Three more studies on interventions to increase CPAP usage have been conducted since the aforementioned review. In a randomized controlled study, investigators had OSA patients participate in a 30-minute group social cognitive therapy session (eg, increasing perceived self-efficacy, outcome expectations, and social support) to increase CPAP adherence. Compared to a social interaction control group, there was no increase in adherence rates [25]. In another smaller randomized controlled study that used a social cognition model of behavioral therapy, there were small increases of CPAP usage. At 3 months, the social cognitive intervention increased CPAP usage by an average of 23 minutes per night, increased the number of individuals using their CPAP machine for more than 4 hours compared to standard care group, and decreased symptom of sleepiness [26]. And lastly, a preliminary study looked at increasing adherence rates by utilizing easily accessible alternative care providers, such as nurses and respiratory therapists, for the management of OSA [27]. Though this study had no control group, it did show that good adherence and a decrease in symptoms of sleepiness could be achieved with non-physician management of OSA [27]. A randomized controlled study will be needed to validate the use of alternative care providers.

Interventions have shown some success in increasing adherence rates, but the question remains on who should receive those interventions. Predicting which OSA patients are in most need of an intervention has been studied. A recent study used a 19-question assessment tool called the Index of Nonadherence to PAP to screen for nonadherers (OSA patients who used CPAP for less than 4 hours a night, after 1 month of OSA diagnosis). The assessment tool was 87% sensitive and 63% specific at determining those OSA patients who would not adhere to CPAP treatment [28]. Another study investigated the reliability and validity of a self-rating scale measuring the side effects of CPAP and their consequences on adherence [15]. The investigators showed that the scale was able to reliably discriminate between those who adhered to CPAP treatment and those that did not [15]. Both of these scales can be used to screen OSA patients that need interventions to increase CPAP adherence. Lastly, a recent systematic review showed that a user’s CPAP experience was not defined by the user but by the user’s health care provider, who framed CPAP as “problematic” [29]. The authors argue that users of CPAP are “primed” to reflect negatively on their CPAP experience [29]. Interventions can be used to change the way OSA patients think or feel about their CPAP machines.

When OSA Patients Do Not Adhere to CPAP Treatment

With adherence rates as low as 50% [16–18], those who fail to tolerate CPAP are unlikely to be referred for additional treatment [30]. Those who do tolerate treatment dislike the side effects of CPAP and show an interest in other treatment options [14]. Other treatment options have been shown to decrease the severity of OSA.

Weight Loss and Exercise

OSA prevalence is correlated with body mass index (BMI), and the increasing rates of OSA has been attributed to the increasing rates of obesity in the United States [2]. A meta-analysis of 3 randomized controlled studies of weight loss induced by dieting or lifestyle change showed that weight loss decreased OSA severity. The effect was the greatest for OSA patients who lost more than 10 kg or had severe OSA at baseline [31]. A recent randomized controlled study involving OSA patients with type 2 diabetes investigated if either a weight loss intervention or a diabetes support and education intervention would be able to decrease OSA severity [32]. The weight loss intervention significantly decreased OSA severity, which was largely but not entirely attributed to weight loss. The participants regained 50% of their weight 4 years after the intervention and still had significantly less severe OSA compared to control intervention group. The downside to this intervention is the intensity of the regimen to which the subjects had to adapt: portion-controlled diets with liquid meals and snack bars for the first 4 months and moderate-intensity physical activity for a minimum of 3 hours a week for the first year. After that, patients were still required to follow through with the intervention for 3 years, which included one on-site visit per month and a second contact by phone, mail, or email [32]. One study looked at weight loss and sleep position (supine vs. lateral). The study showed a decrease in AHI in OSA patients that lost weight, and the biggest decrease was in AHI in the lateral sleeping position [33]. Another study looked at the more invasive procedure of bariatric surgery to decrease weight and OSA. At the 1-year follow-up, patients had significantly decreased their BMI and AHI [34]. Two more randomized controlled studies investigated if exercise or fitness level might be beneficial to OSA patients independent of weight loss. Exercise improved AHI even though there was not a significant decrease in weight between the exercise and stretching control group [35]. However, an increase in fitness level did not have any additive effect on the decrease of AHI when weight change was taken into account [36]. The difference in results might be attributed to the latter study using older type 2 diabetic patients and moderate physical activity, while the former studied incorporated moderate-intensity aerobic activity and resistance training for younger patients [35,36]. There is evidence that a sedentary lifestyle increases diurnal leg fluid volume that can shift to the neck during sleep and might play a role in pathogenesis of OSA [37]. Decreasing a sedentary lifestyle by exercising might therefore be beneficial to OSA patients. Given the increasing rates of obesity [2], implementing weight loss as a solution to OSA is viable, especially considering that OSA is not the only comorbid disease of obesity [38].

 

 

Positional Therapy

It has been known for some time that sleeping in a supine position doubles a patient’s AHI compared to sleeping in the lateral position [39]. A more recent analysis showed that 60% of patients were “supine predominant OSA;” these patients had supine AHI that was twice that of non-supine AHI [40]. Moreover, a drug-induced sleep endoscopy study showed that the upper airway collapses at multiple levels sleeping in the supine position as opposed to at a single level sleeping in the lateral position [41]. Another study showed that lateral sleeping position improved passive airway anatomy and decreased collapsibility [42]. Many studies have shown that patients who wear a device that alerts the sleeper that he or she is in a supine position (referred to as positional therapy) significantly decreases AHI, but long-term compliance is still an issue, and new and improved devices are needed [43]. Three new studies bolster the effectiveness of positional therapy [44–46]. In all 3 studies, sleeping in the supine position went down to 0% (no change in sleep efficiency [the ratio of total time spent sleeping to the total time spent in bed]), AHI decreased to less than 6, and sleep quality and daytime sleepiness increased and decreased, respectively [44–46]. Compliance was as low as 76% [44] and as high as 93% [46]. For those who cannot tolerate CPAP, positional therapy could be a substitute for decreasing severity of OSA. However, “phenotyping” OSA patients as “supine predominant OSA” would need to be implemented to guarantee efficacy of positional therapy.

Nasal Expiratory Positive Airway Pressure

Nasal expiratory positive airway pressure (nEPAP), sold under the commercial name Provent (Provent Sleep Therapy, Manchester, NH), consists of a single-use device that attaches to the nostrils using an adhesive to create an airtight seal (Figure 1). The device contains a mechanical valve that creates high resistance during expiration but not during inspiration [12]. The greatest risk of upper airway collapse occurs at the end of the expiratory phase of the respiratory cycle because of a lack of positive pressure or phasic activation of the upper airways [47]. nEPAP increases positive pressure at the expiratory phase, thus preventing upper airway collapse [12]. A recent review detailed the positive benefits (eg, decreased AHI, improved oxygen saturation, increased quality of life, decreased snoring) and very few negative effects (eg, dry mouth, nasal discomfort) of nEPAP; adherence was still a problem, but better compared to CPAP [12]. A recent meta-analysis of 18 studies was published about nEPAP [48]. The authors concluded that there was a 53.2% decrease in AHI and an improvement in the quality of life; however, nEPAP did not entirely eliminate OSA (residual AHI was still elevated), and similar to CPAP, adherence was still an issue. But for those patients who are intolerant of CPAP or are traveling, as well as the ease of application and low cost, makes nEPAP a decent alternative [48].

Oral Pressure Therapy

Retro-palatal collapse occurs in OSA and can be prevented by applying negative pressure to the upper airway [49]. The oral pressure therapy (OPT) device applies gentle suction anteriorly and superiorly to displace the tongue and soft palate and breathing occurs via nasopharyngeal airway [12]. A recent systematic review [49] of OPT revealed that successful OPT treatment rate was 25% to 37% if using standard and stringent definitions of treatment success. Although OPT decreased AHI, residual AHI still remained high due to positional apneas and collapse of upper airway at other levels besides retro-palatal. The authors of this systematic review recommend more rigorous and controlled studies with defined “treatment success” [49]. The advantage of OPT is that adherence was good; patients used the device on average 6 hours a night. There were no severe or serious adverse events with OPT, however oral tissue discomfort or irritation, dental discomfort, and dry mouth were reported [50].

Oral Appliances

Similar to OPT, oral appliances (OAs) attempt to prevent upper airway collapse. OAs either stabilize the tongue, advance the mandible, or lift the soft palate to increase the volumes of the upper airways to avert OSA [16, 51]. The OAs, like the mandibular advancement device, for example, have the added benefit of being fitted specifically for the OSA patient. The mandible for a patient can be advanced to alleviate obstructive apneas, but can also be pulled back if the OA is too uncomfortable or painful. However, there is still dispute on how exactly to titrate these OAs [52]. A meta-analysis recently published looked at all clinical trials of OAs through September 2015. After meeting strict exclusion/inclusion criteria, 17 studies looking at OAs were included in the meta-analysis. There were robust decreases in AHI and in symptoms of sleepiness in OSA patients that used OAs compared to control groups. However, due to the strict inclusion/exclusion criteria of the meta-analysis, all the studies except one used mandibular advancement appliances; one study used a tongue-retaining appliance. The authors concluded that there is sufficient evidence for OAs to be effective in patients with mild-to-moderate OSA [51]. Since the meta-analysis, 6 new studies have been published about OAs. In 4 of the studies (all using mandibular advancement), OAs significantly decreased AHI by 50% or more in the majority of OSA patients [53–56]. The other 2 studies looked at long-term efficacy and compliance. In both studies, there were drastic decreases in AHI when OAs were applied [57, 58]. In one study, about 40% of OSA patients stopped using the OAs. When the change in AHI was stratified between users and non-users, the users group was significantly higher that the non-user group, suggesting that the non-user group were not compliant due to less of an effect of the OA on AHI [57]. In the second study, OSA patients using OAs for a median of 16.5 years were evaluated for long-term efficacy of the OAs. At the short-term follow-up, AHI decreased by more than 50% with use of an OA. However, at the long-term follow-up, the OA lost any effect on AHI. One reason for this is that the OSA patients’ AHI without the OA at the long-term follow-up nearly doubled compared to AHI without OA at the short-term follow-up. The authors conclude that OSA patients using OAs for the long-term might undergo deteriorations in treatment efficacy of OAs, and that regular follow-up appointments with sleep apnea recordings should be implemented [58].

A similarity in all these studies is that adherence was higher for OAs compared to CPAP [51]. The caveat is that most studies have relied on self-reports for adherence rates [12]. However, there were 3 studies that implemented a sensor that measured adherence and compared those results to self-reported OA adherence. All 3 studies showed a strong relationship between self-reports and sensor adherence; patients were honest about adherence to OAs [59–61]. Studies have also been conducted to predict compliance with OAs: higher therapeutic CPAP pressure, age, OSA severity [62], decreased snoring [63], and lower BMI [64, 65] predicted compliance, while dry mouth [63], oropharyngeal crowding [65], and sleeping in a supine position [66] predicted noncompliance. Though adherence rates are high, OAs do not decrease AHI as much as CPAP [67], and a recent study showed that long-term adherence rates might not be different to CPAP adherence rates [68]. OAs, due to their higher adherence rates, are a potential second choice over CPAP. However, they are less efficacious than CPAP at decreasing AHI. That may not be as important since a recent meta-analysis comparing the effects of CPAP or OAs on blood pressure showed that both treatments significantly decreased blood pressure [69]. More studies need to be conducted over long-term efficacy of OAs compared with CPAP.

Surgeries to Treat OSA

Surgery as a treatment option has been extensively reviewed and meta-analyzed [70–78]. Surgery for the treatment of OSA includes tongue suspension [70,74], maxillomandibular advancement (MMA) [72,73,78], pharyngeal surgeries (eg, uvulopharyngopalatoplasty [UPPP]) [73], laser-assisted uvulopalatoplasty (LAUP) [73], radiofrequency ablation (RFA) [73], tracheostomy [71], nasal surgery [75], and glossectomy [77], as well as multi-level and multi-phased procedures [70,74,76,77]. Most studies done on surgeries were case studies, with a minority of investigations that were randomized and controlled. Glossectomy, as part of a multi-level surgical approach, decreased AHI and symptoms of sleepiness, but glossectomy as a stand-alone surgical procedure did not improve AHI [77]. Significant improvements in AHI and sleepiness symptoms were seen in a majority of OSA patients who underwent MMA [72,73,78] and tracheostomy, although tracheostomy was performed for the morbidly obese or those who have failed other traditional surgical treatments [71]. Stand-alone tongue suspension and nasal surgery did not decrease AHI in the majority of patients, though nasal surgery did decrease subjective sleepiness [70,72,74,75]. However, tongue suspension combined with UPPP had better outcomes [70]. LAUP showed inconsistent results with the majority of studies showing no change in AHI, while UPPP and RFA seemed to improved AHI, although some studies showed no change [73]. Multi-level or multi-phase surgeries also showed improvements on OSA severity, but whether these surgeries are better than stand-alone remains to be investigated [73,76]. Morbidity and adverse events, like infection or pain, are common in all of these surgical events [70–78], but there are significant differences between the procedures. For example, MMA had fewer adverse events reported compared to UPPP [73]. More recently, glossectomy via transoral robotic surgery with UPPP [79] or epiglottoplasty [80] has been investigated; there were decreases in AHI, but response rates were between 64% to 73%. Although it seems surgical procedures to treat OSA are plausible, most studies were not rigorous enough to say this with any certainty.

Hypoglossal Nerve Stimulation

OSA subjects experience upper airway obstruction due to loss of genioglossus muscle activity during sleep. Without tongue activation, the negative pressure of breathing causes the upper airways to collapse [81]. Transcutaneous, intraoral, and intramuscular devices used to electrically activate the tongue have been developed and tested; however, although these devices decreased AHI they also induced arousals and sleep fragmentation caused by the electrical stimulus [82–86]. A new method had to be developed that would not be felt by the OSA patient.

That new method, hypoglossal nerve stimulation (HNS) [87–93], was to electrically stimulate the hypoglossal nerve, a motor nerve innervating the protrusor and retractor muscle of the tongue. During a surgical procedure, a silicone cuff with stimulating electrodes is placed around a unilateral hypoglossal nerve and appropriate placement of the cuff is tested by stimulating the nerve and observing protrusion and electromyographic signals of the tongue. The leads of the stimulating electrodes are tunneled subplatysmally via the neck to a subcutaneous neurostimulator located on the chest. Sensory leads from the neurostimulator are then subcutaneously tunneled to the intercostal muscles to monitor respiration (Figure 2). Via inspiration detected through the sensory leads, the neurostimulator uses an algorithm to predict the onset of inspiration. The neurostimulator delivers electric pulses to the hypoglossal nerve between the end of expiration and the beginning of the next expiratory phase, thus activating and protruding the genioglossus muscle and counteracting the negative pressures and collapsing forces on the upper airways during inspiration [87–93]. After implantation, device titration occurs by gradually adjusting stimulus pulse intensity, frequency, and width to levels that are tolerable to the OSA patient and that lead to significant decreases in AHI [87–93].

In all trials to date, there were significant decreases in AHI as long as 3 years post implantation [87–93]. There were significant improvements in symptoms of sleepiness, mood, quality of life, and sleep quality [87,88,90–94]. When OSA patients had their neurostimulators turned off for 5 days to a week, AHI returned back to baseline levels [89,92]. However, all the trials excluded morbidly obese individuals [87–93] because investigations showed that HNS had no therapeutic effect with elevated BMI [88,90]. The drawbacks of HNS are that it is surgically invasive and minor adverse events have been reported: procedural-related events (eg, numbness/pain/swelling/infection at incision site, temporary tongue weakness) that resolved with time, pain medication, and/or antibiotic treatment, or therapy-related events (eg, tongue abrasions cause by tongue movement over teeth, discomfort associated with stimulation) that resolved after acclimation. Serious adverse events occurred infrequently, such as infection at incision site requiring device removal or subsequent surgery to reposition or replace electrode cuff or malfunctioning neurostimulator [87,88,90]. HNS durability at 18 and 36 months was still very effective, with decreased AHI and increase quality of life and sleep being sustained; adverse events were uncommon this long after implantation [91,93]. Although surgery is required and adverse events are reported, the long-term significant improvement of OSA makes this a very viable treatment option over CPAP. However, increasing prevalence rates of OSA are correlated to increasing obesity rates [2], which may limit the usefulness of HNS since a BMI of more than 40 might preclude individuals to this treatment.

Pharmacologic Treatment

There are no approved pharmacologic treatments for OSA. A recent Cochrane review and meta-analysis assessed clinical trials of various drugs treating OSA. These drugs targeted 5 strategies at alleviating OSA: increasing ventilatory drive (progestogens, theophylline, and acetazolamide), increasing upper airway tone (serotonergics and cholinergics), decreasing rapid eye movement sleep (antidepressants and clonidine), increasing arousal threshold (eszopiclone), and/or increasing the cross-sectional area or reducing the surface tension of the upper airway through topical therapy (fluticasone and lubricant). The review concluded that “some of the drugs may be helpful; however, their tolerability needs to be considered in long-term trials.” Some of these drugs had little or no effect on AHI, and if they did have an effect on AHI, side effects outweighed the benefit [95]. Since then, more investigations of other drugs targeted at the previously aforementioned strategies or via new strategies have been published.

Dronabinol (synthetic Δ⁹-THC), a nonselective cannabinoid type 1 and type 2 receptor agonist, significantly reduced AHI and improved subjective sleepiness and alertness in a single-blind dose-escalation (2.5, 5, or 10 mg) proof-of-concept study [96,97]. Dronabinol most likely increases upper airway tone though inhibition of vagal afferents [98,99]. There were no serious adverse events associated with dronabinol. Minor adverse events included somnolence and increased appetite. Increased appetite might lead to increased weight and contradict any beneficial effects of dronabinol; however, in the 3-week treatment period no weight gain was observed [97]. This might have been due to drug administration occurring before going to sleep with no opportunity to eat. A larger randomized controlled study will be needed to establish the safety and efficacy of dronabinol.

The sedative zopiclone was used to increase arousal threshold without effecting genioglossus activity [100]. Eszopiclone, a drug in the same class, has been used in the past with unfavorable results [95]. Zopiclone was used in a small double-blind randomized controlled cross-over study. Zopiclone significantly increased respiratory arousal threshold without effecting genioglossus activity or the upper airway’s response to negative pressure. Thus, there was a trend in the zopiclone treatment to increase sleep efficiency. However, zopiclone had no effect on AHI, and increased oxygen desaturation [100]. Similar to eszopiclone, the results for zopiclone are not promising.

A new strategy to treat OSA is to modify pharmacologically “loop gain,” a dimensionless value quantifying the stability of the ventilatory control system. A high loop gain signifies instability in the ventilatory control system and predisposes an OSA person to recurrent apneas [101–103]. Three studies used drugs that inhibit carbonic anhydrase to stabilize the ventilatory control system [104–106]. Two studies used acetazolamide, which decreased loop gain in OSA patients [104,105]. Acetazolamide only decreased AHI in non–rapid eye movement (NREM) sleep, and there was a slight correlation between decrease in loop gain and total AHI [105]. Acetazolamide also decreased ventilatory response to spontaneous arousal, thus promoting ventilatory stability [104]. In the last study, zonisamide, a carbonic anhydrase inhibitor that also causes weight loss, was investigated in OSA patients. Sleep apnea alleviation, measured in terms of absolute elimination of sleep apnea by mechanical or pharmacologic treatment, was 61% and 13% for CPAP and zonisamide, respectively, compared with placebo. In other words, zonisamide decreased AHI but not to the extent of CPAP [106]. Zonisamide also decreased arousals and marginally, but significantly, decreased weight compared to the CPAP group. Although carbonic anhydrase inhibitors have promise as an alternative treatment, long-term use is poorly tolerated [101] and further studies need to be completed.

OSA has been linked with gastroesophageal reflux disease (GERD), with studies suggesting OSA precipitates GERD [107] or GERD precipitates OSA [108]. A meta-analysis was recently published looking at studies that used proton pump inhibitors (PPI) to treat GERD and the effects it would have on OSA [109]. The meta-analysis only included 2 randomized trials and 4 prospective cohort studies. Two of the cohort studies showed a significant decrease, and one cohort showed no difference in apnea indices; and all 4 of the cohort studies showed no difference in AHI. In one trial, the frequency of apnea attacks as recorded by diaries significantly decreased. In 3 cohort studies and one trial, symptoms of sleepiness significantly decreased [109]. A study that was not included in the meta-analysis showed that 3 months of PPI treatment decreased AHI but did not alter sleep efficiency [110]. Larger randomized controlled studies need to be conducted on the effects of PPIs on OSA, especially since PPIs are well tolerated with only weak observational associations between PPI therapy and fractures, pneumonia, mortality, and nutritional deficiencies [111].

The drugs mentioned above have potential for treating OSA in patients intolerant to CPAP. The efficacy and side effects of the drugs will need to be studied for long-term use. However, development of pharmacologic treatments has been hampered by incomplete knowledge of the relevant sleep-dependent peripheral and central neural mechanisms controlling ventilatory drive and upper airway muscles. More importantly, additional basic science research needs to focus on the neurobiological and neurophysiological mechanisms underlying OSA to develop new pharmacotherapies or treatment strategies, or to modify previous treatment strategies.

Treatment Combinations and Phenotyping

It has been recently suggested that combining 2 or more of the above treatments might lead to greater decreases in AHI and greater improvements in subjective sleepiness [112,113]. In fact, one such treatment combination has occurred [114]. Both OA or positional therapy decrease AHI. However, the combination of an OA and positional therapy led to further significant decreases in AHI compared to when those treatments were used alone [114]. To correctly combine treatments, the patient will have to be “phenotyped” via polysomnography to discern the specific pathophysiology of the patient’s OSA. There are published reports of methods to phenotype patients according to their sleep positon, ventilation parameters, loop gain, arousal threshold, and upper airway gain, and if apneic events occur in REM or NREM sleep [40,115]. Defining these traits for individual OSA patients can lead to better efficacy and compliance of combination treatments for OSA. Combination treatment coupled with phenotyping are needed to try to reduce AHI to levels achieved with CPAP.

Conclusion

CPAP is the gold standard treatment because it substantially decreases the severity of OSA just by placing a mask over one’s face before going to sleep. However, it is not tolerable to continually have air forced into your upper airways, and new CPAP devices that heat and humidify the air, and auto titrate the pressure, have been developed to increase adherence rates, but with limited success. For all the treatments listed, a majority do not decrease the severity of OSA to levels achieved with CPAP. However, adherence rates are higher and therefore might, in the long-term, be a better option than CPAP. Some treatments involve invasive surgery to open or stabilize the upper airways, or to implant a stimulator, some treatments involve oral drugs with side effects, and some treatments involve placing appliances on your nose or in your mouth. And some treatments can be combined and individually tailored to the OSA patient via “phenotyping.” For all treatments, the benefits and risks need to be weighed by each patient. More importantly, more large randomized controlled studies on treatments or combination of treatments for OSA are needed using parameters such as treatment adherence, AHI, oxygen desaturation, subjective sleepiness, quality of life, and adverse events (both minor and major) to gauge treatment success in the short-term and long-term. Only then can OSA patients in partnership with their health care provider choose the best treatment option.

 

Corresponding author: Michael W. Calik, PhD, 845 S. Damen Ave (M/C 802), College of Nursing, Room 740, Chicago, IL 60612, [email protected].

Financial disclosures: None.

From the Center for Narcolepsy, Sleep and Health Research, Department of Biobehavioral Health Science, University of Illinois at Chicago, Chicago, IL.

 

Abstract

• Objective: To review the efficacy of current treatment options for adults with obstructive sleep apnea (OSA).
• Methods: Review of the literature.
• Results: OSA, characterized by repetitive ≥ 10-second interruptions (apnea) or reductions (hypopnea) in airflow, is initiated by partial or complete collapse in the upper airway despite respiratory effort. When left untreated, OSA is associated with comorbid conditions, such as cardiovascular and metabolic diseases. The current “gold standard” treatment for OSA is continuous positive air pressure (CPAP), which pneumatically stabilizes the upper airways. CPAP has proven efficacy and potential cost savings via decreases in health comorbidities and/or motor-vehicle crashes. However, CPAP treatment is not well-tolerated due to various side effects, and adherence among OSA subjects can be as low as 50% in certain populations. Other treatment options for OSA include improving CPAP tolerability, increasing CPAP adherence through patient interventions, weight loss/exercise, positional therapy, nasal expiratory positive airway pressure, oral pressure therapy, oral appliances, surgery, hypoglossal nerve stimulation, drug treatment, and combining 2 or more of the aforementioned treatments. Despite the many options available to treat OSA, none of them are as efficacious as CPAP. However, many of these treatments are tolerable, and adherence rates are higher than those of the CPAP, making them a more viable treatment option for long-term use.
• Conclusion: Patients need to weigh the benefits and risks of available treatments for OSA. More large randomized controlled studies on treatments or combination of treatments for OSA are needed that measure parameters such as treatment adherence, apnea-hypopnea index, oxygen desaturation, subjective sleepiness, quality of life, and adverse events.

 

Obstructive sleep apnea (OSA), characterized by repetitive ≥ 10-second interruptions (apnea) or reductions (hypopnea) in airflow (measured as events/hour, called the apnea-hypopnea index [AHI]), is initiated by partial or complete collapse in the upper airway despite respiratory effort [1]. Current estimates of the prevalence of OSA (AHI ≥ 5 and Epworth Sleepiness Scale > 10) in American men and women (aged 30–70 years) are 14% and 5%, respectively, with prevalence rates increasing due to increasing rates of obesity, a risk factor for developing OSA [2]. Hypoxemia/hypercapnia, fragmented sleep, as well as exaggerated fluctuations in heart rhythm, blood pressure, and intrathoracic pressure are some of the acute physiological effects of untreated OSA [1]. These acute effects can develop into long-term sequelae, such as hypertension and other cardiovascular comorbidities [2,3], decrements in cognitive function [4,5], poor mood, reduced quality of life [6,7], and premature death [8,9]. In economic terms, health care cost estimates of OSA and its associated comorbidities rival that of diabetes [10]. Additionally, in the year 2000, more than 800,000 drivers were involved OSA-related motor-vehicle collisions, of which more than 1400 fatalities occurred [11].

Front-line treatment of OSA relies on mechanically stabilizing the upper airway with a column of air via continuous positive airway pressure (CPAP) treatment. Though CPAP is the “gold standard” treatment for OSA with proven efficacy and potential cost savings via decreases in health comorbidities and/or motor-vehicle crashes [10–12], CPAP treatment is not well-tolerated due to various side effects [13–15]. Adherence among OSA subjects can be as low as 50% in certain populations [16–18]. Improved strategies for current and innovative treatments have emerged in the last few years and are the subject of this review.

 

 

Improved CPAP Treatment

As stated previously, CPAP pneumatically splints the upper airway, thus preventing it from collapsing during sleep. However, CPAP is not well-tolerated. Modifications to standard CPAP to increase adherence have been met with disappointing results. Humidification with heated tubing delivering heated moistened air did not increase compliance compared to standard CPAP [19]. CPAP was also compared with auto-adjusting CPAP (APAP), where respiration is monitored and the minimum pressure of air is applied to splint the upper airway open. In a meta-analysis, APAP only had very small effect on compliance [20]. Lastly, reduction in pressure during expiration was investigated, and a meta-analysis showed no effect [21,22]. However, recent advances in CPAP delivery give hope to increasing compliance. The S9CPAP machine (Resmed, San Diego, CA), which combines a humidification system and an APAP, showed increased compliance compared to standard CPAP. Compliance increased by an average of 30 minutes per night, and variance of daily usage decreased (eg, patients used it more day-to-day) [23]. However, a randomized blinded study needs to be conducted to corroborate these results.

Promoting CPAP Adherence Through Patient Interventions

Educational, supportive, and behavioral interventions have been used to increase CPAP adherence and have been thoroughly reviewed via meta-analysis [24]. Briefly, 30 studies of various interventions were included and demonstrated that educational, supportive, or behavioral interventions increased CPAP usage in OSA-naive patients. Behavioral interventions increased CPAP usage by over an hour, but the evidence was of “low-quality.” Educational and supportive interventions also increased CPAP usage, with the former having “moderate-quality” evidence [24]. However, whether increased CPAP usage had an effect on symptoms and quality of life was statistically unclear, and the authors recommended further assessment [24]. Three more studies on interventions to increase CPAP usage have been conducted since the aforementioned review. In a randomized controlled study, investigators had OSA patients participate in a 30-minute group social cognitive therapy session (eg, increasing perceived self-efficacy, outcome expectations, and social support) to increase CPAP adherence. Compared to a social interaction control group, there was no increase in adherence rates [25]. In another smaller randomized controlled study that used a social cognition model of behavioral therapy, there were small increases of CPAP usage. At 3 months, the social cognitive intervention increased CPAP usage by an average of 23 minutes per night, increased the number of individuals using their CPAP machine for more than 4 hours compared to standard care group, and decreased symptom of sleepiness [26]. And lastly, a preliminary study looked at increasing adherence rates by utilizing easily accessible alternative care providers, such as nurses and respiratory therapists, for the management of OSA [27]. Though this study had no control group, it did show that good adherence and a decrease in symptoms of sleepiness could be achieved with non-physician management of OSA [27]. A randomized controlled study will be needed to validate the use of alternative care providers.

Interventions have shown some success in increasing adherence rates, but the question remains on who should receive those interventions. Predicting which OSA patients are in most need of an intervention has been studied. A recent study used a 19-question assessment tool called the Index of Nonadherence to PAP to screen for nonadherers (OSA patients who used CPAP for less than 4 hours a night, after 1 month of OSA diagnosis). The assessment tool was 87% sensitive and 63% specific at determining those OSA patients who would not adhere to CPAP treatment [28]. Another study investigated the reliability and validity of a self-rating scale measuring the side effects of CPAP and their consequences on adherence [15]. The investigators showed that the scale was able to reliably discriminate between those who adhered to CPAP treatment and those that did not [15]. Both of these scales can be used to screen OSA patients that need interventions to increase CPAP adherence. Lastly, a recent systematic review showed that a user’s CPAP experience was not defined by the user but by the user’s health care provider, who framed CPAP as “problematic” [29]. The authors argue that users of CPAP are “primed” to reflect negatively on their CPAP experience [29]. Interventions can be used to change the way OSA patients think or feel about their CPAP machines.

When OSA Patients Do Not Adhere to CPAP Treatment

With adherence rates as low as 50% [16–18], those who fail to tolerate CPAP are unlikely to be referred for additional treatment [30]. Those who do tolerate treatment dislike the side effects of CPAP and show an interest in other treatment options [14]. Other treatment options have been shown to decrease the severity of OSA.

Weight Loss and Exercise

OSA prevalence is correlated with body mass index (BMI), and the increasing rates of OSA has been attributed to the increasing rates of obesity in the United States [2]. A meta-analysis of 3 randomized controlled studies of weight loss induced by dieting or lifestyle change showed that weight loss decreased OSA severity. The effect was the greatest for OSA patients who lost more than 10 kg or had severe OSA at baseline [31]. A recent randomized controlled study involving OSA patients with type 2 diabetes investigated if either a weight loss intervention or a diabetes support and education intervention would be able to decrease OSA severity [32]. The weight loss intervention significantly decreased OSA severity, which was largely but not entirely attributed to weight loss. The participants regained 50% of their weight 4 years after the intervention and still had significantly less severe OSA compared to control intervention group. The downside to this intervention is the intensity of the regimen to which the subjects had to adapt: portion-controlled diets with liquid meals and snack bars for the first 4 months and moderate-intensity physical activity for a minimum of 3 hours a week for the first year. After that, patients were still required to follow through with the intervention for 3 years, which included one on-site visit per month and a second contact by phone, mail, or email [32]. One study looked at weight loss and sleep position (supine vs. lateral). The study showed a decrease in AHI in OSA patients that lost weight, and the biggest decrease was in AHI in the lateral sleeping position [33]. Another study looked at the more invasive procedure of bariatric surgery to decrease weight and OSA. At the 1-year follow-up, patients had significantly decreased their BMI and AHI [34]. Two more randomized controlled studies investigated if exercise or fitness level might be beneficial to OSA patients independent of weight loss. Exercise improved AHI even though there was not a significant decrease in weight between the exercise and stretching control group [35]. However, an increase in fitness level did not have any additive effect on the decrease of AHI when weight change was taken into account [36]. The difference in results might be attributed to the latter study using older type 2 diabetic patients and moderate physical activity, while the former studied incorporated moderate-intensity aerobic activity and resistance training for younger patients [35,36]. There is evidence that a sedentary lifestyle increases diurnal leg fluid volume that can shift to the neck during sleep and might play a role in pathogenesis of OSA [37]. Decreasing a sedentary lifestyle by exercising might therefore be beneficial to OSA patients. Given the increasing rates of obesity [2], implementing weight loss as a solution to OSA is viable, especially considering that OSA is not the only comorbid disease of obesity [38].

 

 

Positional Therapy

It has been known for some time that sleeping in a supine position doubles a patient’s AHI compared to sleeping in the lateral position [39]. A more recent analysis showed that 60% of patients were “supine predominant OSA;” these patients had supine AHI that was twice that of non-supine AHI [40]. Moreover, a drug-induced sleep endoscopy study showed that the upper airway collapses at multiple levels sleeping in the supine position as opposed to at a single level sleeping in the lateral position [41]. Another study showed that lateral sleeping position improved passive airway anatomy and decreased collapsibility [42]. Many studies have shown that patients who wear a device that alerts the sleeper that he or she is in a supine position (referred to as positional therapy) significantly decreases AHI, but long-term compliance is still an issue, and new and improved devices are needed [43]. Three new studies bolster the effectiveness of positional therapy [44–46]. In all 3 studies, sleeping in the supine position went down to 0% (no change in sleep efficiency [the ratio of total time spent sleeping to the total time spent in bed]), AHI decreased to less than 6, and sleep quality and daytime sleepiness increased and decreased, respectively [44–46]. Compliance was as low as 76% [44] and as high as 93% [46]. For those who cannot tolerate CPAP, positional therapy could be a substitute for decreasing severity of OSA. However, “phenotyping” OSA patients as “supine predominant OSA” would need to be implemented to guarantee efficacy of positional therapy.

Nasal Expiratory Positive Airway Pressure

Nasal expiratory positive airway pressure (nEPAP), sold under the commercial name Provent (Provent Sleep Therapy, Manchester, NH), consists of a single-use device that attaches to the nostrils using an adhesive to create an airtight seal (Figure 1). The device contains a mechanical valve that creates high resistance during expiration but not during inspiration [12]. The greatest risk of upper airway collapse occurs at the end of the expiratory phase of the respiratory cycle because of a lack of positive pressure or phasic activation of the upper airways [47]. nEPAP increases positive pressure at the expiratory phase, thus preventing upper airway collapse [12]. A recent review detailed the positive benefits (eg, decreased AHI, improved oxygen saturation, increased quality of life, decreased snoring) and very few negative effects (eg, dry mouth, nasal discomfort) of nEPAP; adherence was still a problem, but better compared to CPAP [12]. A recent meta-analysis of 18 studies was published about nEPAP [48]. The authors concluded that there was a 53.2% decrease in AHI and an improvement in the quality of life; however, nEPAP did not entirely eliminate OSA (residual AHI was still elevated), and similar to CPAP, adherence was still an issue. But for those patients who are intolerant of CPAP or are traveling, as well as the ease of application and low cost, makes nEPAP a decent alternative [48].

Oral Pressure Therapy

Retro-palatal collapse occurs in OSA and can be prevented by applying negative pressure to the upper airway [49]. The oral pressure therapy (OPT) device applies gentle suction anteriorly and superiorly to displace the tongue and soft palate and breathing occurs via nasopharyngeal airway [12]. A recent systematic review [49] of OPT revealed that successful OPT treatment rate was 25% to 37% if using standard and stringent definitions of treatment success. Although OPT decreased AHI, residual AHI still remained high due to positional apneas and collapse of upper airway at other levels besides retro-palatal. The authors of this systematic review recommend more rigorous and controlled studies with defined “treatment success” [49]. The advantage of OPT is that adherence was good; patients used the device on average 6 hours a night. There were no severe or serious adverse events with OPT, however oral tissue discomfort or irritation, dental discomfort, and dry mouth were reported [50].

Oral Appliances

Similar to OPT, oral appliances (OAs) attempt to prevent upper airway collapse. OAs either stabilize the tongue, advance the mandible, or lift the soft palate to increase the volumes of the upper airways to avert OSA [16, 51]. The OAs, like the mandibular advancement device, for example, have the added benefit of being fitted specifically for the OSA patient. The mandible for a patient can be advanced to alleviate obstructive apneas, but can also be pulled back if the OA is too uncomfortable or painful. However, there is still dispute on how exactly to titrate these OAs [52]. A meta-analysis recently published looked at all clinical trials of OAs through September 2015. After meeting strict exclusion/inclusion criteria, 17 studies looking at OAs were included in the meta-analysis. There were robust decreases in AHI and in symptoms of sleepiness in OSA patients that used OAs compared to control groups. However, due to the strict inclusion/exclusion criteria of the meta-analysis, all the studies except one used mandibular advancement appliances; one study used a tongue-retaining appliance. The authors concluded that there is sufficient evidence for OAs to be effective in patients with mild-to-moderate OSA [51]. Since the meta-analysis, 6 new studies have been published about OAs. In 4 of the studies (all using mandibular advancement), OAs significantly decreased AHI by 50% or more in the majority of OSA patients [53–56]. The other 2 studies looked at long-term efficacy and compliance. In both studies, there were drastic decreases in AHI when OAs were applied [57, 58]. In one study, about 40% of OSA patients stopped using the OAs. When the change in AHI was stratified between users and non-users, the users group was significantly higher that the non-user group, suggesting that the non-user group were not compliant due to less of an effect of the OA on AHI [57]. In the second study, OSA patients using OAs for a median of 16.5 years were evaluated for long-term efficacy of the OAs. At the short-term follow-up, AHI decreased by more than 50% with use of an OA. However, at the long-term follow-up, the OA lost any effect on AHI. One reason for this is that the OSA patients’ AHI without the OA at the long-term follow-up nearly doubled compared to AHI without OA at the short-term follow-up. The authors conclude that OSA patients using OAs for the long-term might undergo deteriorations in treatment efficacy of OAs, and that regular follow-up appointments with sleep apnea recordings should be implemented [58].

A similarity in all these studies is that adherence was higher for OAs compared to CPAP [51]. The caveat is that most studies have relied on self-reports for adherence rates [12]. However, there were 3 studies that implemented a sensor that measured adherence and compared those results to self-reported OA adherence. All 3 studies showed a strong relationship between self-reports and sensor adherence; patients were honest about adherence to OAs [59–61]. Studies have also been conducted to predict compliance with OAs: higher therapeutic CPAP pressure, age, OSA severity [62], decreased snoring [63], and lower BMI [64, 65] predicted compliance, while dry mouth [63], oropharyngeal crowding [65], and sleeping in a supine position [66] predicted noncompliance. Though adherence rates are high, OAs do not decrease AHI as much as CPAP [67], and a recent study showed that long-term adherence rates might not be different to CPAP adherence rates [68]. OAs, due to their higher adherence rates, are a potential second choice over CPAP. However, they are less efficacious than CPAP at decreasing AHI. That may not be as important since a recent meta-analysis comparing the effects of CPAP or OAs on blood pressure showed that both treatments significantly decreased blood pressure [69]. More studies need to be conducted over long-term efficacy of OAs compared with CPAP.

Surgeries to Treat OSA

Surgery as a treatment option has been extensively reviewed and meta-analyzed [70–78]. Surgery for the treatment of OSA includes tongue suspension [70,74], maxillomandibular advancement (MMA) [72,73,78], pharyngeal surgeries (eg, uvulopharyngopalatoplasty [UPPP]) [73], laser-assisted uvulopalatoplasty (LAUP) [73], radiofrequency ablation (RFA) [73], tracheostomy [71], nasal surgery [75], and glossectomy [77], as well as multi-level and multi-phased procedures [70,74,76,77]. Most studies done on surgeries were case studies, with a minority of investigations that were randomized and controlled. Glossectomy, as part of a multi-level surgical approach, decreased AHI and symptoms of sleepiness, but glossectomy as a stand-alone surgical procedure did not improve AHI [77]. Significant improvements in AHI and sleepiness symptoms were seen in a majority of OSA patients who underwent MMA [72,73,78] and tracheostomy, although tracheostomy was performed for the morbidly obese or those who have failed other traditional surgical treatments [71]. Stand-alone tongue suspension and nasal surgery did not decrease AHI in the majority of patients, though nasal surgery did decrease subjective sleepiness [70,72,74,75]. However, tongue suspension combined with UPPP had better outcomes [70]. LAUP showed inconsistent results with the majority of studies showing no change in AHI, while UPPP and RFA seemed to improved AHI, although some studies showed no change [73]. Multi-level or multi-phase surgeries also showed improvements on OSA severity, but whether these surgeries are better than stand-alone remains to be investigated [73,76]. Morbidity and adverse events, like infection or pain, are common in all of these surgical events [70–78], but there are significant differences between the procedures. For example, MMA had fewer adverse events reported compared to UPPP [73]. More recently, glossectomy via transoral robotic surgery with UPPP [79] or epiglottoplasty [80] has been investigated; there were decreases in AHI, but response rates were between 64% to 73%. Although it seems surgical procedures to treat OSA are plausible, most studies were not rigorous enough to say this with any certainty.

Hypoglossal Nerve Stimulation

OSA subjects experience upper airway obstruction due to loss of genioglossus muscle activity during sleep. Without tongue activation, the negative pressure of breathing causes the upper airways to collapse [81]. Transcutaneous, intraoral, and intramuscular devices used to electrically activate the tongue have been developed and tested; however, although these devices decreased AHI they also induced arousals and sleep fragmentation caused by the electrical stimulus [82–86]. A new method had to be developed that would not be felt by the OSA patient.

That new method, hypoglossal nerve stimulation (HNS) [87–93], was to electrically stimulate the hypoglossal nerve, a motor nerve innervating the protrusor and retractor muscle of the tongue. During a surgical procedure, a silicone cuff with stimulating electrodes is placed around a unilateral hypoglossal nerve and appropriate placement of the cuff is tested by stimulating the nerve and observing protrusion and electromyographic signals of the tongue. The leads of the stimulating electrodes are tunneled subplatysmally via the neck to a subcutaneous neurostimulator located on the chest. Sensory leads from the neurostimulator are then subcutaneously tunneled to the intercostal muscles to monitor respiration (Figure 2). Via inspiration detected through the sensory leads, the neurostimulator uses an algorithm to predict the onset of inspiration. The neurostimulator delivers electric pulses to the hypoglossal nerve between the end of expiration and the beginning of the next expiratory phase, thus activating and protruding the genioglossus muscle and counteracting the negative pressures and collapsing forces on the upper airways during inspiration [87–93]. After implantation, device titration occurs by gradually adjusting stimulus pulse intensity, frequency, and width to levels that are tolerable to the OSA patient and that lead to significant decreases in AHI [87–93].

In all trials to date, there were significant decreases in AHI as long as 3 years post implantation [87–93]. There were significant improvements in symptoms of sleepiness, mood, quality of life, and sleep quality [87,88,90–94]. When OSA patients had their neurostimulators turned off for 5 days to a week, AHI returned back to baseline levels [89,92]. However, all the trials excluded morbidly obese individuals [87–93] because investigations showed that HNS had no therapeutic effect with elevated BMI [88,90]. The drawbacks of HNS are that it is surgically invasive and minor adverse events have been reported: procedural-related events (eg, numbness/pain/swelling/infection at incision site, temporary tongue weakness) that resolved with time, pain medication, and/or antibiotic treatment, or therapy-related events (eg, tongue abrasions cause by tongue movement over teeth, discomfort associated with stimulation) that resolved after acclimation. Serious adverse events occurred infrequently, such as infection at incision site requiring device removal or subsequent surgery to reposition or replace electrode cuff or malfunctioning neurostimulator [87,88,90]. HNS durability at 18 and 36 months was still very effective, with decreased AHI and increase quality of life and sleep being sustained; adverse events were uncommon this long after implantation [91,93]. Although surgery is required and adverse events are reported, the long-term significant improvement of OSA makes this a very viable treatment option over CPAP. However, increasing prevalence rates of OSA are correlated to increasing obesity rates [2], which may limit the usefulness of HNS since a BMI of more than 40 might preclude individuals to this treatment.

Pharmacologic Treatment

There are no approved pharmacologic treatments for OSA. A recent Cochrane review and meta-analysis assessed clinical trials of various drugs treating OSA. These drugs targeted 5 strategies at alleviating OSA: increasing ventilatory drive (progestogens, theophylline, and acetazolamide), increasing upper airway tone (serotonergics and cholinergics), decreasing rapid eye movement sleep (antidepressants and clonidine), increasing arousal threshold (eszopiclone), and/or increasing the cross-sectional area or reducing the surface tension of the upper airway through topical therapy (fluticasone and lubricant). The review concluded that “some of the drugs may be helpful; however, their tolerability needs to be considered in long-term trials.” Some of these drugs had little or no effect on AHI, and if they did have an effect on AHI, side effects outweighed the benefit [95]. Since then, more investigations of other drugs targeted at the previously aforementioned strategies or via new strategies have been published.

Dronabinol (synthetic Δ⁹-THC), a nonselective cannabinoid type 1 and type 2 receptor agonist, significantly reduced AHI and improved subjective sleepiness and alertness in a single-blind dose-escalation (2.5, 5, or 10 mg) proof-of-concept study [96,97]. Dronabinol most likely increases upper airway tone though inhibition of vagal afferents [98,99]. There were no serious adverse events associated with dronabinol. Minor adverse events included somnolence and increased appetite. Increased appetite might lead to increased weight and contradict any beneficial effects of dronabinol; however, in the 3-week treatment period no weight gain was observed [97]. This might have been due to drug administration occurring before going to sleep with no opportunity to eat. A larger randomized controlled study will be needed to establish the safety and efficacy of dronabinol.

The sedative zopiclone was used to increase arousal threshold without effecting genioglossus activity [100]. Eszopiclone, a drug in the same class, has been used in the past with unfavorable results [95]. Zopiclone was used in a small double-blind randomized controlled cross-over study. Zopiclone significantly increased respiratory arousal threshold without effecting genioglossus activity or the upper airway’s response to negative pressure. Thus, there was a trend in the zopiclone treatment to increase sleep efficiency. However, zopiclone had no effect on AHI, and increased oxygen desaturation [100]. Similar to eszopiclone, the results for zopiclone are not promising.

A new strategy to treat OSA is to modify pharmacologically “loop gain,” a dimensionless value quantifying the stability of the ventilatory control system. A high loop gain signifies instability in the ventilatory control system and predisposes an OSA person to recurrent apneas [101–103]. Three studies used drugs that inhibit carbonic anhydrase to stabilize the ventilatory control system [104–106]. Two studies used acetazolamide, which decreased loop gain in OSA patients [104,105]. Acetazolamide only decreased AHI in non–rapid eye movement (NREM) sleep, and there was a slight correlation between decrease in loop gain and total AHI [105]. Acetazolamide also decreased ventilatory response to spontaneous arousal, thus promoting ventilatory stability [104]. In the last study, zonisamide, a carbonic anhydrase inhibitor that also causes weight loss, was investigated in OSA patients. Sleep apnea alleviation, measured in terms of absolute elimination of sleep apnea by mechanical or pharmacologic treatment, was 61% and 13% for CPAP and zonisamide, respectively, compared with placebo. In other words, zonisamide decreased AHI but not to the extent of CPAP [106]. Zonisamide also decreased arousals and marginally, but significantly, decreased weight compared to the CPAP group. Although carbonic anhydrase inhibitors have promise as an alternative treatment, long-term use is poorly tolerated [101] and further studies need to be completed.

OSA has been linked with gastroesophageal reflux disease (GERD), with studies suggesting OSA precipitates GERD [107] or GERD precipitates OSA [108]. A meta-analysis was recently published looking at studies that used proton pump inhibitors (PPI) to treat GERD and the effects it would have on OSA [109]. The meta-analysis only included 2 randomized trials and 4 prospective cohort studies. Two of the cohort studies showed a significant decrease, and one cohort showed no difference in apnea indices; and all 4 of the cohort studies showed no difference in AHI. In one trial, the frequency of apnea attacks as recorded by diaries significantly decreased. In 3 cohort studies and one trial, symptoms of sleepiness significantly decreased [109]. A study that was not included in the meta-analysis showed that 3 months of PPI treatment decreased AHI but did not alter sleep efficiency [110]. Larger randomized controlled studies need to be conducted on the effects of PPIs on OSA, especially since PPIs are well tolerated with only weak observational associations between PPI therapy and fractures, pneumonia, mortality, and nutritional deficiencies [111].

The drugs mentioned above have potential for treating OSA in patients intolerant to CPAP. The efficacy and side effects of the drugs will need to be studied for long-term use. However, development of pharmacologic treatments has been hampered by incomplete knowledge of the relevant sleep-dependent peripheral and central neural mechanisms controlling ventilatory drive and upper airway muscles. More importantly, additional basic science research needs to focus on the neurobiological and neurophysiological mechanisms underlying OSA to develop new pharmacotherapies or treatment strategies, or to modify previous treatment strategies.

Treatment Combinations and Phenotyping

It has been recently suggested that combining 2 or more of the above treatments might lead to greater decreases in AHI and greater improvements in subjective sleepiness [112,113]. In fact, one such treatment combination has occurred [114]. Both OA or positional therapy decrease AHI. However, the combination of an OA and positional therapy led to further significant decreases in AHI compared to when those treatments were used alone [114]. To correctly combine treatments, the patient will have to be “phenotyped” via polysomnography to discern the specific pathophysiology of the patient’s OSA. There are published reports of methods to phenotype patients according to their sleep positon, ventilation parameters, loop gain, arousal threshold, and upper airway gain, and if apneic events occur in REM or NREM sleep [40,115]. Defining these traits for individual OSA patients can lead to better efficacy and compliance of combination treatments for OSA. Combination treatment coupled with phenotyping are needed to try to reduce AHI to levels achieved with CPAP.

Conclusion

CPAP is the gold standard treatment because it substantially decreases the severity of OSA just by placing a mask over one’s face before going to sleep. However, it is not tolerable to continually have air forced into your upper airways, and new CPAP devices that heat and humidify the air, and auto titrate the pressure, have been developed to increase adherence rates, but with limited success. For all the treatments listed, a majority do not decrease the severity of OSA to levels achieved with CPAP. However, adherence rates are higher and therefore might, in the long-term, be a better option than CPAP. Some treatments involve invasive surgery to open or stabilize the upper airways, or to implant a stimulator, some treatments involve oral drugs with side effects, and some treatments involve placing appliances on your nose or in your mouth. And some treatments can be combined and individually tailored to the OSA patient via “phenotyping.” For all treatments, the benefits and risks need to be weighed by each patient. More importantly, more large randomized controlled studies on treatments or combination of treatments for OSA are needed using parameters such as treatment adherence, AHI, oxygen desaturation, subjective sleepiness, quality of life, and adverse events (both minor and major) to gauge treatment success in the short-term and long-term. Only then can OSA patients in partnership with their health care provider choose the best treatment option.

 

Corresponding author: Michael W. Calik, PhD, 845 S. Damen Ave (M/C 802), College of Nursing, Room 740, Chicago, IL 60612, [email protected].

Financial disclosures: None.

From the Center for Narcolepsy, Sleep and Health Research, Department of Biobehavioral Health Science, University of Illinois at Chicago, Chicago, IL.

 

Abstract

• Objective: To review the efficacy of current treatment options for adults with obstructive sleep apnea (OSA).
• Methods: Review of the literature.
• Results: OSA, characterized by repetitive ≥ 10-second interruptions (apnea) or reductions (hypopnea) in airflow, is initiated by partial or complete collapse in the upper airway despite respiratory effort. When left untreated, OSA is associated with comorbid conditions, such as cardiovascular and metabolic diseases. The current “gold standard” treatment for OSA is continuous positive air pressure (CPAP), which pneumatically stabilizes the upper airways. CPAP has proven efficacy and potential cost savings via decreases in health comorbidities and/or motor-vehicle crashes. However, CPAP treatment is not well-tolerated due to various side effects, and adherence among OSA subjects can be as low as 50% in certain populations. Other treatment options for OSA include improving CPAP tolerability, increasing CPAP adherence through patient interventions, weight loss/exercise, positional therapy, nasal expiratory positive airway pressure, oral pressure therapy, oral appliances, surgery, hypoglossal nerve stimulation, drug treatment, and combining 2 or more of the aforementioned treatments. Despite the many options available to treat OSA, none of them are as efficacious as CPAP. However, many of these treatments are tolerable, and adherence rates are higher than those of the CPAP, making them a more viable treatment option for long-term use.
• Conclusion: Patients need to weigh the benefits and risks of available treatments for OSA. More large randomized controlled studies on treatments or combination of treatments for OSA are needed that measure parameters such as treatment adherence, apnea-hypopnea index, oxygen desaturation, subjective sleepiness, quality of life, and adverse events.

 

Obstructive sleep apnea (OSA), characterized by repetitive ≥ 10-second interruptions (apnea) or reductions (hypopnea) in airflow (measured as events/hour, called the apnea-hypopnea index [AHI]), is initiated by partial or complete collapse in the upper airway despite respiratory effort [1]. Current estimates of the prevalence of OSA (AHI ≥ 5 and Epworth Sleepiness Scale > 10) in American men and women (aged 30–70 years) are 14% and 5%, respectively, with prevalence rates increasing due to increasing rates of obesity, a risk factor for developing OSA [2]. Hypoxemia/hypercapnia, fragmented sleep, as well as exaggerated fluctuations in heart rhythm, blood pressure, and intrathoracic pressure are some of the acute physiological effects of untreated OSA [1]. These acute effects can develop into long-term sequelae, such as hypertension and other cardiovascular comorbidities [2,3], decrements in cognitive function [4,5], poor mood, reduced quality of life [6,7], and premature death [8,9]. In economic terms, health care cost estimates of OSA and its associated comorbidities rival that of diabetes [10]. Additionally, in the year 2000, more than 800,000 drivers were involved OSA-related motor-vehicle collisions, of which more than 1400 fatalities occurred [11].

Front-line treatment of OSA relies on mechanically stabilizing the upper airway with a column of air via continuous positive airway pressure (CPAP) treatment. Though CPAP is the “gold standard” treatment for OSA with proven efficacy and potential cost savings via decreases in health comorbidities and/or motor-vehicle crashes [10–12], CPAP treatment is not well-tolerated due to various side effects [13–15]. Adherence among OSA subjects can be as low as 50% in certain populations [16–18]. Improved strategies for current and innovative treatments have emerged in the last few years and are the subject of this review.

 

 

Improved CPAP Treatment

As stated previously, CPAP pneumatically splints the upper airway, thus preventing it from collapsing during sleep. However, CPAP is not well-tolerated. Modifications to standard CPAP to increase adherence have been met with disappointing results. Humidification with heated tubing delivering heated moistened air did not increase compliance compared to standard CPAP [19]. CPAP was also compared with auto-adjusting CPAP (APAP), where respiration is monitored and the minimum pressure of air is applied to splint the upper airway open. In a meta-analysis, APAP only had very small effect on compliance [20]. Lastly, reduction in pressure during expiration was investigated, and a meta-analysis showed no effect [21,22]. However, recent advances in CPAP delivery give hope to increasing compliance. The S9CPAP machine (Resmed, San Diego, CA), which combines a humidification system and an APAP, showed increased compliance compared to standard CPAP. Compliance increased by an average of 30 minutes per night, and variance of daily usage decreased (eg, patients used it more day-to-day) [23]. However, a randomized blinded study needs to be conducted to corroborate these results.

Promoting CPAP Adherence Through Patient Interventions

Educational, supportive, and behavioral interventions have been used to increase CPAP adherence and have been thoroughly reviewed via meta-analysis [24]. Briefly, 30 studies of various interventions were included and demonstrated that educational, supportive, or behavioral interventions increased CPAP usage in OSA-naive patients. Behavioral interventions increased CPAP usage by over an hour, but the evidence was of “low-quality.” Educational and supportive interventions also increased CPAP usage, with the former having “moderate-quality” evidence [24]. However, whether increased CPAP usage had an effect on symptoms and quality of life was statistically unclear, and the authors recommended further assessment [24]. Three more studies on interventions to increase CPAP usage have been conducted since the aforementioned review. In a randomized controlled study, investigators had OSA patients participate in a 30-minute group social cognitive therapy session (eg, increasing perceived self-efficacy, outcome expectations, and social support) to increase CPAP adherence. Compared to a social interaction control group, there was no increase in adherence rates [25]. In another smaller randomized controlled study that used a social cognition model of behavioral therapy, there were small increases of CPAP usage. At 3 months, the social cognitive intervention increased CPAP usage by an average of 23 minutes per night, increased the number of individuals using their CPAP machine for more than 4 hours compared to standard care group, and decreased symptom of sleepiness [26]. And lastly, a preliminary study looked at increasing adherence rates by utilizing easily accessible alternative care providers, such as nurses and respiratory therapists, for the management of OSA [27]. Though this study had no control group, it did show that good adherence and a decrease in symptoms of sleepiness could be achieved with non-physician management of OSA [27]. A randomized controlled study will be needed to validate the use of alternative care providers.

Interventions have shown some success in increasing adherence rates, but the question remains on who should receive those interventions. Predicting which OSA patients are in most need of an intervention has been studied. A recent study used a 19-question assessment tool called the Index of Nonadherence to PAP to screen for nonadherers (OSA patients who used CPAP for less than 4 hours a night, after 1 month of OSA diagnosis). The assessment tool was 87% sensitive and 63% specific at determining those OSA patients who would not adhere to CPAP treatment [28]. Another study investigated the reliability and validity of a self-rating scale measuring the side effects of CPAP and their consequences on adherence [15]. The investigators showed that the scale was able to reliably discriminate between those who adhered to CPAP treatment and those that did not [15]. Both of these scales can be used to screen OSA patients that need interventions to increase CPAP adherence. Lastly, a recent systematic review showed that a user’s CPAP experience was not defined by the user but by the user’s health care provider, who framed CPAP as “problematic” [29]. The authors argue that users of CPAP are “primed” to reflect negatively on their CPAP experience [29]. Interventions can be used to change the way OSA patients think or feel about their CPAP machines.

When OSA Patients Do Not Adhere to CPAP Treatment

With adherence rates as low as 50% [16–18], those who fail to tolerate CPAP are unlikely to be referred for additional treatment [30]. Those who do tolerate treatment dislike the side effects of CPAP and show an interest in other treatment options [14]. Other treatment options have been shown to decrease the severity of OSA.

Weight Loss and Exercise

OSA prevalence is correlated with body mass index (BMI), and the increasing rates of OSA has been attributed to the increasing rates of obesity in the United States [2]. A meta-analysis of 3 randomized controlled studies of weight loss induced by dieting or lifestyle change showed that weight loss decreased OSA severity. The effect was the greatest for OSA patients who lost more than 10 kg or had severe OSA at baseline [31]. A recent randomized controlled study involving OSA patients with type 2 diabetes investigated if either a weight loss intervention or a diabetes support and education intervention would be able to decrease OSA severity [32]. The weight loss intervention significantly decreased OSA severity, which was largely but not entirely attributed to weight loss. The participants regained 50% of their weight 4 years after the intervention and still had significantly less severe OSA compared to control intervention group. The downside to this intervention is the intensity of the regimen to which the subjects had to adapt: portion-controlled diets with liquid meals and snack bars for the first 4 months and moderate-intensity physical activity for a minimum of 3 hours a week for the first year. After that, patients were still required to follow through with the intervention for 3 years, which included one on-site visit per month and a second contact by phone, mail, or email [32]. One study looked at weight loss and sleep position (supine vs. lateral). The study showed a decrease in AHI in OSA patients that lost weight, and the biggest decrease was in AHI in the lateral sleeping position [33]. Another study looked at the more invasive procedure of bariatric surgery to decrease weight and OSA. At the 1-year follow-up, patients had significantly decreased their BMI and AHI [34]. Two more randomized controlled studies investigated if exercise or fitness level might be beneficial to OSA patients independent of weight loss. Exercise improved AHI even though there was not a significant decrease in weight between the exercise and stretching control group [35]. However, an increase in fitness level did not have any additive effect on the decrease of AHI when weight change was taken into account [36]. The difference in results might be attributed to the latter study using older type 2 diabetic patients and moderate physical activity, while the former studied incorporated moderate-intensity aerobic activity and resistance training for younger patients [35,36]. There is evidence that a sedentary lifestyle increases diurnal leg fluid volume that can shift to the neck during sleep and might play a role in pathogenesis of OSA [37]. Decreasing a sedentary lifestyle by exercising might therefore be beneficial to OSA patients. Given the increasing rates of obesity [2], implementing weight loss as a solution to OSA is viable, especially considering that OSA is not the only comorbid disease of obesity [38].

 

 

Positional Therapy

It has been known for some time that sleeping in a supine position doubles a patient’s AHI compared to sleeping in the lateral position [39]. A more recent analysis showed that 60% of patients were “supine predominant OSA;” these patients had supine AHI that was twice that of non-supine AHI [40]. Moreover, a drug-induced sleep endoscopy study showed that the upper airway collapses at multiple levels sleeping in the supine position as opposed to at a single level sleeping in the lateral position [41]. Another study showed that lateral sleeping position improved passive airway anatomy and decreased collapsibility [42]. Many studies have shown that patients who wear a device that alerts the sleeper that he or she is in a supine position (referred to as positional therapy) significantly decreases AHI, but long-term compliance is still an issue, and new and improved devices are needed [43]. Three new studies bolster the effectiveness of positional therapy [44–46]. In all 3 studies, sleeping in the supine position went down to 0% (no change in sleep efficiency [the ratio of total time spent sleeping to the total time spent in bed]), AHI decreased to less than 6, and sleep quality and daytime sleepiness increased and decreased, respectively [44–46]. Compliance was as low as 76% [44] and as high as 93% [46]. For those who cannot tolerate CPAP, positional therapy could be a substitute for decreasing severity of OSA. However, “phenotyping” OSA patients as “supine predominant OSA” would need to be implemented to guarantee efficacy of positional therapy.

Nasal Expiratory Positive Airway Pressure

Nasal expiratory positive airway pressure (nEPAP), sold under the commercial name Provent (Provent Sleep Therapy, Manchester, NH), consists of a single-use device that attaches to the nostrils using an adhesive to create an airtight seal (Figure 1). The device contains a mechanical valve that creates high resistance during expiration but not during inspiration [12]. The greatest risk of upper airway collapse occurs at the end of the expiratory phase of the respiratory cycle because of a lack of positive pressure or phasic activation of the upper airways [47]. nEPAP increases positive pressure at the expiratory phase, thus preventing upper airway collapse [12]. A recent review detailed the positive benefits (eg, decreased AHI, improved oxygen saturation, increased quality of life, decreased snoring) and very few negative effects (eg, dry mouth, nasal discomfort) of nEPAP; adherence was still a problem, but better compared to CPAP [12]. A recent meta-analysis of 18 studies was published about nEPAP [48]. The authors concluded that there was a 53.2% decrease in AHI and an improvement in the quality of life; however, nEPAP did not entirely eliminate OSA (residual AHI was still elevated), and similar to CPAP, adherence was still an issue. But for those patients who are intolerant of CPAP or are traveling, as well as the ease of application and low cost, makes nEPAP a decent alternative [48].

Oral Pressure Therapy

Retro-palatal collapse occurs in OSA and can be prevented by applying negative pressure to the upper airway [49]. The oral pressure therapy (OPT) device applies gentle suction anteriorly and superiorly to displace the tongue and soft palate and breathing occurs via nasopharyngeal airway [12]. A recent systematic review [49] of OPT revealed that successful OPT treatment rate was 25% to 37% if using standard and stringent definitions of treatment success. Although OPT decreased AHI, residual AHI still remained high due to positional apneas and collapse of upper airway at other levels besides retro-palatal. The authors of this systematic review recommend more rigorous and controlled studies with defined “treatment success” [49]. The advantage of OPT is that adherence was good; patients used the device on average 6 hours a night. There were no severe or serious adverse events with OPT, however oral tissue discomfort or irritation, dental discomfort, and dry mouth were reported [50].

Oral Appliances

Similar to OPT, oral appliances (OAs) attempt to prevent upper airway collapse. OAs either stabilize the tongue, advance the mandible, or lift the soft palate to increase the volumes of the upper airways to avert OSA [16, 51]. The OAs, like the mandibular advancement device, for example, have the added benefit of being fitted specifically for the OSA patient. The mandible for a patient can be advanced to alleviate obstructive apneas, but can also be pulled back if the OA is too uncomfortable or painful. However, there is still dispute on how exactly to titrate these OAs [52]. A meta-analysis recently published looked at all clinical trials of OAs through September 2015. After meeting strict exclusion/inclusion criteria, 17 studies looking at OAs were included in the meta-analysis. There were robust decreases in AHI and in symptoms of sleepiness in OSA patients that used OAs compared to control groups. However, due to the strict inclusion/exclusion criteria of the meta-analysis, all the studies except one used mandibular advancement appliances; one study used a tongue-retaining appliance. The authors concluded that there is sufficient evidence for OAs to be effective in patients with mild-to-moderate OSA [51]. Since the meta-analysis, 6 new studies have been published about OAs. In 4 of the studies (all using mandibular advancement), OAs significantly decreased AHI by 50% or more in the majority of OSA patients [53–56]. The other 2 studies looked at long-term efficacy and compliance. In both studies, there were drastic decreases in AHI when OAs were applied [57, 58]. In one study, about 40% of OSA patients stopped using the OAs. When the change in AHI was stratified between users and non-users, the users group was significantly higher that the non-user group, suggesting that the non-user group were not compliant due to less of an effect of the OA on AHI [57]. In the second study, OSA patients using OAs for a median of 16.5 years were evaluated for long-term efficacy of the OAs. At the short-term follow-up, AHI decreased by more than 50% with use of an OA. However, at the long-term follow-up, the OA lost any effect on AHI. One reason for this is that the OSA patients’ AHI without the OA at the long-term follow-up nearly doubled compared to AHI without OA at the short-term follow-up. The authors conclude that OSA patients using OAs for the long-term might undergo deteriorations in treatment efficacy of OAs, and that regular follow-up appointments with sleep apnea recordings should be implemented [58].

A similarity in all these studies is that adherence was higher for OAs compared to CPAP [51]. The caveat is that most studies have relied on self-reports for adherence rates [12]. However, there were 3 studies that implemented a sensor that measured adherence and compared those results to self-reported OA adherence. All 3 studies showed a strong relationship between self-reports and sensor adherence; patients were honest about adherence to OAs [59–61]. Studies have also been conducted to predict compliance with OAs: higher therapeutic CPAP pressure, age, OSA severity [62], decreased snoring [63], and lower BMI [64, 65] predicted compliance, while dry mouth [63], oropharyngeal crowding [65], and sleeping in a supine position [66] predicted noncompliance. Though adherence rates are high, OAs do not decrease AHI as much as CPAP [67], and a recent study showed that long-term adherence rates might not be different to CPAP adherence rates [68]. OAs, due to their higher adherence rates, are a potential second choice over CPAP. However, they are less efficacious than CPAP at decreasing AHI. That may not be as important since a recent meta-analysis comparing the effects of CPAP or OAs on blood pressure showed that both treatments significantly decreased blood pressure [69]. More studies need to be conducted over long-term efficacy of OAs compared with CPAP.

Surgeries to Treat OSA

Surgery as a treatment option has been extensively reviewed and meta-analyzed [70–78]. Surgery for the treatment of OSA includes tongue suspension [70,74], maxillomandibular advancement (MMA) [72,73,78], pharyngeal surgeries (eg, uvulopharyngopalatoplasty [UPPP]) [73], laser-assisted uvulopalatoplasty (LAUP) [73], radiofrequency ablation (RFA) [73], tracheostomy [71], nasal surgery [75], and glossectomy [77], as well as multi-level and multi-phased procedures [70,74,76,77]. Most studies done on surgeries were case studies, with a minority of investigations that were randomized and controlled. Glossectomy, as part of a multi-level surgical approach, decreased AHI and symptoms of sleepiness, but glossectomy as a stand-alone surgical procedure did not improve AHI [77]. Significant improvements in AHI and sleepiness symptoms were seen in a majority of OSA patients who underwent MMA [72,73,78] and tracheostomy, although tracheostomy was performed for the morbidly obese or those who have failed other traditional surgical treatments [71]. Stand-alone tongue suspension and nasal surgery did not decrease AHI in the majority of patients, though nasal surgery did decrease subjective sleepiness [70,72,74,75]. However, tongue suspension combined with UPPP had better outcomes [70]. LAUP showed inconsistent results with the majority of studies showing no change in AHI, while UPPP and RFA seemed to improved AHI, although some studies showed no change [73]. Multi-level or multi-phase surgeries also showed improvements on OSA severity, but whether these surgeries are better than stand-alone remains to be investigated [73,76]. Morbidity and adverse events, like infection or pain, are common in all of these surgical events [70–78], but there are significant differences between the procedures. For example, MMA had fewer adverse events reported compared to UPPP [73]. More recently, glossectomy via transoral robotic surgery with UPPP [79] or epiglottoplasty [80] has been investigated; there were decreases in AHI, but response rates were between 64% to 73%. Although it seems surgical procedures to treat OSA are plausible, most studies were not rigorous enough to say this with any certainty.

Hypoglossal Nerve Stimulation

OSA subjects experience upper airway obstruction due to loss of genioglossus muscle activity during sleep. Without tongue activation, the negative pressure of breathing causes the upper airways to collapse [81]. Transcutaneous, intraoral, and intramuscular devices used to electrically activate the tongue have been developed and tested; however, although these devices decreased AHI they also induced arousals and sleep fragmentation caused by the electrical stimulus [82–86]. A new method had to be developed that would not be felt by the OSA patient.

That new method, hypoglossal nerve stimulation (HNS) [87–93], was to electrically stimulate the hypoglossal nerve, a motor nerve innervating the protrusor and retractor muscle of the tongue. During a surgical procedure, a silicone cuff with stimulating electrodes is placed around a unilateral hypoglossal nerve and appropriate placement of the cuff is tested by stimulating the nerve and observing protrusion and electromyographic signals of the tongue. The leads of the stimulating electrodes are tunneled subplatysmally via the neck to a subcutaneous neurostimulator located on the chest. Sensory leads from the neurostimulator are then subcutaneously tunneled to the intercostal muscles to monitor respiration (Figure 2). Via inspiration detected through the sensory leads, the neurostimulator uses an algorithm to predict the onset of inspiration. The neurostimulator delivers electric pulses to the hypoglossal nerve between the end of expiration and the beginning of the next expiratory phase, thus activating and protruding the genioglossus muscle and counteracting the negative pressures and collapsing forces on the upper airways during inspiration [87–93]. After implantation, device titration occurs by gradually adjusting stimulus pulse intensity, frequency, and width to levels that are tolerable to the OSA patient and that lead to significant decreases in AHI [87–93].

In all trials to date, there were significant decreases in AHI as long as 3 years post implantation [87–93]. There were significant improvements in symptoms of sleepiness, mood, quality of life, and sleep quality [87,88,90–94]. When OSA patients had their neurostimulators turned off for 5 days to a week, AHI returned back to baseline levels [89,92]. However, all the trials excluded morbidly obese individuals [87–93] because investigations showed that HNS had no therapeutic effect with elevated BMI [88,90]. The drawbacks of HNS are that it is surgically invasive and minor adverse events have been reported: procedural-related events (eg, numbness/pain/swelling/infection at incision site, temporary tongue weakness) that resolved with time, pain medication, and/or antibiotic treatment, or therapy-related events (eg, tongue abrasions cause by tongue movement over teeth, discomfort associated with stimulation) that resolved after acclimation. Serious adverse events occurred infrequently, such as infection at incision site requiring device removal or subsequent surgery to reposition or replace electrode cuff or malfunctioning neurostimulator [87,88,90]. HNS durability at 18 and 36 months was still very effective, with decreased AHI and increase quality of life and sleep being sustained; adverse events were uncommon this long after implantation [91,93]. Although surgery is required and adverse events are reported, the long-term significant improvement of OSA makes this a very viable treatment option over CPAP. However, increasing prevalence rates of OSA are correlated to increasing obesity rates [2], which may limit the usefulness of HNS since a BMI of more than 40 might preclude individuals to this treatment.

Pharmacologic Treatment

There are no approved pharmacologic treatments for OSA. A recent Cochrane review and meta-analysis assessed clinical trials of various drugs treating OSA. These drugs targeted 5 strategies at alleviating OSA: increasing ventilatory drive (progestogens, theophylline, and acetazolamide), increasing upper airway tone (serotonergics and cholinergics), decreasing rapid eye movement sleep (antidepressants and clonidine), increasing arousal threshold (eszopiclone), and/or increasing the cross-sectional area or reducing the surface tension of the upper airway through topical therapy (fluticasone and lubricant). The review concluded that “some of the drugs may be helpful; however, their tolerability needs to be considered in long-term trials.” Some of these drugs had little or no effect on AHI, and if they did have an effect on AHI, side effects outweighed the benefit [95]. Since then, more investigations of other drugs targeted at the previously aforementioned strategies or via new strategies have been published.

Dronabinol (synthetic Δ⁹-THC), a nonselective cannabinoid type 1 and type 2 receptor agonist, significantly reduced AHI and improved subjective sleepiness and alertness in a single-blind dose-escalation (2.5, 5, or 10 mg) proof-of-concept study [96,97]. Dronabinol most likely increases upper airway tone though inhibition of vagal afferents [98,99]. There were no serious adverse events associated with dronabinol. Minor adverse events included somnolence and increased appetite. Increased appetite might lead to increased weight and contradict any beneficial effects of dronabinol; however, in the 3-week treatment period no weight gain was observed [97]. This might have been due to drug administration occurring before going to sleep with no opportunity to eat. A larger randomized controlled study will be needed to establish the safety and efficacy of dronabinol.

The sedative zopiclone was used to increase arousal threshold without effecting genioglossus activity [100]. Eszopiclone, a drug in the same class, has been used in the past with unfavorable results [95]. Zopiclone was used in a small double-blind randomized controlled cross-over study. Zopiclone significantly increased respiratory arousal threshold without effecting genioglossus activity or the upper airway’s response to negative pressure. Thus, there was a trend in the zopiclone treatment to increase sleep efficiency. However, zopiclone had no effect on AHI, and increased oxygen desaturation [100]. Similar to eszopiclone, the results for zopiclone are not promising.

A new strategy to treat OSA is to modify pharmacologically “loop gain,” a dimensionless value quantifying the stability of the ventilatory control system. A high loop gain signifies instability in the ventilatory control system and predisposes an OSA person to recurrent apneas [101–103]. Three studies used drugs that inhibit carbonic anhydrase to stabilize the ventilatory control system [104–106]. Two studies used acetazolamide, which decreased loop gain in OSA patients [104,105]. Acetazolamide only decreased AHI in non–rapid eye movement (NREM) sleep, and there was a slight correlation between decrease in loop gain and total AHI [105]. Acetazolamide also decreased ventilatory response to spontaneous arousal, thus promoting ventilatory stability [104]. In the last study, zonisamide, a carbonic anhydrase inhibitor that also causes weight loss, was investigated in OSA patients. Sleep apnea alleviation, measured in terms of absolute elimination of sleep apnea by mechanical or pharmacologic treatment, was 61% and 13% for CPAP and zonisamide, respectively, compared with placebo. In other words, zonisamide decreased AHI but not to the extent of CPAP [106]. Zonisamide also decreased arousals and marginally, but significantly, decreased weight compared to the CPAP group. Although carbonic anhydrase inhibitors have promise as an alternative treatment, long-term use is poorly tolerated [101] and further studies need to be completed.

OSA has been linked with gastroesophageal reflux disease (GERD), with studies suggesting OSA precipitates GERD [107] or GERD precipitates OSA [108]. A meta-analysis was recently published looking at studies that used proton pump inhibitors (PPI) to treat GERD and the effects it would have on OSA [109]. The meta-analysis only included 2 randomized trials and 4 prospective cohort studies. Two of the cohort studies showed a significant decrease, and one cohort showed no difference in apnea indices; and all 4 of the cohort studies showed no difference in AHI. In one trial, the frequency of apnea attacks as recorded by diaries significantly decreased. In 3 cohort studies and one trial, symptoms of sleepiness significantly decreased [109]. A study that was not included in the meta-analysis showed that 3 months of PPI treatment decreased AHI but did not alter sleep efficiency [110]. Larger randomized controlled studies need to be conducted on the effects of PPIs on OSA, especially since PPIs are well tolerated with only weak observational associations between PPI therapy and fractures, pneumonia, mortality, and nutritional deficiencies [111].

The drugs mentioned above have potential for treating OSA in patients intolerant to CPAP. The efficacy and side effects of the drugs will need to be studied for long-term use. However, development of pharmacologic treatments has been hampered by incomplete knowledge of the relevant sleep-dependent peripheral and central neural mechanisms controlling ventilatory drive and upper airway muscles. More importantly, additional basic science research needs to focus on the neurobiological and neurophysiological mechanisms underlying OSA to develop new pharmacotherapies or treatment strategies, or to modify previous treatment strategies.

Treatment Combinations and Phenotyping

It has been recently suggested that combining 2 or more of the above treatments might lead to greater decreases in AHI and greater improvements in subjective sleepiness [112,113]. In fact, one such treatment combination has occurred [114]. Both OA or positional therapy decrease AHI. However, the combination of an OA and positional therapy led to further significant decreases in AHI compared to when those treatments were used alone [114]. To correctly combine treatments, the patient will have to be “phenotyped” via polysomnography to discern the specific pathophysiology of the patient’s OSA. There are published reports of methods to phenotype patients according to their sleep positon, ventilation parameters, loop gain, arousal threshold, and upper airway gain, and if apneic events occur in REM or NREM sleep [40,115]. Defining these traits for individual OSA patients can lead to better efficacy and compliance of combination treatments for OSA. Combination treatment coupled with phenotyping are needed to try to reduce AHI to levels achieved with CPAP.

Conclusion

CPAP is the gold standard treatment because it substantially decreases the severity of OSA just by placing a mask over one’s face before going to sleep. However, it is not tolerable to continually have air forced into your upper airways, and new CPAP devices that heat and humidify the air, and auto titrate the pressure, have been developed to increase adherence rates, but with limited success. For all the treatments listed, a majority do not decrease the severity of OSA to levels achieved with CPAP. However, adherence rates are higher and therefore might, in the long-term, be a better option than CPAP. Some treatments involve invasive surgery to open or stabilize the upper airways, or to implant a stimulator, some treatments involve oral drugs with side effects, and some treatments involve placing appliances on your nose or in your mouth. And some treatments can be combined and individually tailored to the OSA patient via “phenotyping.” For all treatments, the benefits and risks need to be weighed by each patient. More importantly, more large randomized controlled studies on treatments or combination of treatments for OSA are needed using parameters such as treatment adherence, AHI, oxygen desaturation, subjective sleepiness, quality of life, and adverse events (both minor and major) to gauge treatment success in the short-term and long-term. Only then can OSA patients in partnership with their health care provider choose the best treatment option.

 

Corresponding author: Michael W. Calik, PhD, 845 S. Damen Ave (M/C 802), College of Nursing, Room 740, Chicago, IL 60612, [email protected].

Financial disclosures: None.

Most medical malpractice cases are still resolved in a courtroom—typically after years of preparation and personal torment. Yet, overall rates of paid medical malpractice claims among all physicians have been steadily decreasing over the past two decades, with reports showing decreases of 30% to 50% in paid claims since 2000.^1-3 At the same time, while median payments and insurance premiums continued to increase until the mid-2000s, they now appear to have plateaued.¹

None of these changes occurred in isolation. More than 30 states now have caps on noneconomic or total damages.² As noted in part 1, since 2000, some states have enacted comprehensive tort reform.⁴ However, whether these changes in malpractice patterns can be attributed directly to specific policy changes remains a hotly contested issue.

Malpractice Risk in Emergency Medicine

To what extent do the trends in medical malpractice apply to emergency medicine (EM)? While emergency physicians’ (EPs’) perception of malpractice risk ranks higher than any other medical specialty,5 in a review of a large sample of malpractice claims from 1991 through 2005, EPs ranked in the middle among specialties with respect to annual risk of a malpractice claim.6 Moreover, the annual risk of a claim for EPs is just under 8%, compared to 7.4% for all physicians. Yet, for neurosurgery and cardiothoracic surgery—the specialties with the highest overall risk of malpractice claims—the annual risk approaches 20%.6 Regarding payout statistics, less than one-fifth of the claims against EPs resulted in payment.6  In a review of a separate insurance database of closed claims, EPs were named as the primary defendant in only 19% of cases.7

Despite the discrepancies between perceived risk and absolute risk of malpractice claims among EPs, malpractice lawsuits continue to affect the practice of EM. This is evidenced in several surveys, in which the majority of EP participants admitted to practicing “defensive medicine” by ordering tests that were felt to be unnecessary and did so in response to perceived malpractice risk.^8-10 Perceived risk also accounts for the significant variation in decision-making in the ED with respect to diagnostic testing and hospitalization of patients.¹¹ One would expect that lowering malpractice risk would result in less so-called unnecessary testing, but whether or not this is truly the case remains to be seen.

Effects of Malpractice Reform

A study by Waxman et al¹² on the effects of significant malpractice tort reform in ED care in Texas, Georgia, and South Carolina found no difference in rates of imaging studies, charges, or patient admissions. Furthermore, legislation reform did not increase plaintiff onus to prove proximate “gross negligence” rather than simply a breach from “reasonably skillful and careful” medicine.¹² These findings suggest that perception of malpractice risk might simply be serving as a proxy for physicians’ underlying risk tolerance, and be less subject to influence by external forces.

Areas Associated With Malpractice Risk

A number of closed-claim databases attempted to identify the characteristics of patient encounters that can lead to malpractice claims, including patient conditions and sources of error. Diagnostic errors have consistently been found to be the leading cause of malpractice claims, accounting for 28% to 65% of claims, followed by inappropriate management of medical treatment and improper performance of a procedure.^7,13-16 A January 2016 benchmarking system report by CRICO Strategies found that 30% of 23,658 medical malpractice claims filed between 2009 through 2013 cited failures in communication as a factor.¹⁷ The report also revealed that among these failed communications, those that occurred between health care providers are more likely to result in payout compared to miscommunications between providers and patients.¹⁷ This report further noted 70% to 80% of claims closed without payment.^7,16 Closed claims were significantly more likely to involve serious injuries or death.^7,18 Leading conditions that resulted in claims include myocardial infarction, nonspecific chest pain, symptoms involving the abdomen or pelvis, appendicitis, and orthopedic injuries.^7,13,16

Diagnostic Errors

Errors in diagnosis have been attributed to multiple factors in the ED. The two most common factors were failure to order tests and failure to perform an adequate history and physical examination, both of which contribute to rationalization of the practice of defensive medicine under the current tort system.¹³ Other significant factors associated with errors in diagnosis include misinterpretation of test results or imaging studies and failure to obtain an appropriate consultation. Processes contributing to each of these potential errors include mistakes in judgment, lack of knowledge, miscommunication, and insufficient documentation (Table).¹⁵

Strategies for Reducing Malpractice Risk

In part 1, we listed several strategies EPs could adopt to help reduce malpractice risk. In this section, we will discuss in further detail how these strategies help mitigate malpractice claims.

 

Patient Communication

Open communication with patients is paramount in reducing the risk of a malpractice allegation. Patients are more likely to become angry or frustrated if they sense a physician is not listening to or addressing their concerns. These patients are in turn more likely to file a complaint if they are harmed or experience a bad outcome during their stay in the ED.

Situations in which patients are unable to provide pertinent information also place the EP at significant risk, as the provider must make decisions without full knowledge of the case. Communication with potential resources such as nursing home staff, the patient’s family, and emergency medical service providers to obtain additional information can help reduce risk. 

Of course, when evaluating and treating patients, the EP should always take the time to listen to the patient’s concerns during the encounter to ensure his or her needs have been addressed. In the event of a patient allegation or complaint, the EP should make the effort to explore and de-escalate the situation before the patient is discharged.

Discharge Care and Instructions

According to CRICO, premature discharge as a factor in medical malpractice liability results from inadequate assessment and missed opportunities in 41% of diagnosis-related ED cases.¹⁶ The following situation illustrates a brief example of such a missed opportunity: A provider makes a diagnosis of urinary tract infection (UTI) in a patient presenting with fever and abdominal pain but whose urinalysis is suspect for contamination and in whom no pelvic examination was performed to rule out other etiologies. When the same patient later returns to the ED with worse abdominal pain, a sterile urine culture invalidates the diagnosis of UTI, and further evaluation leads to a final diagnosis of ruptured appendix.

Prior to discharging any patient, the EP should provide clear and concise at-home care instructions in a manner in which the patient can understand. Clear instructions on how the patient is to manage his or her care after discharge are vital, and failure to do so in terms the patient can understand can create problems if a harmful result occurs. This is especially important in patients with whom there is a communication barrier—eg, language barrier, hearing impairment, cognitive deficit, intoxication, or violent or irrational behavior. In these situations, the EP should always take advantage of available resources and tools such as language lines, interpreters, discharge planners, psychiatric staff, and supportive family members to help reconcile any communication barriers. These measures will in turn optimize patient outcome and reduce the risk of a later malpractice allegation.

Board Certification

All physicians should maintain their respective board certification and specialty training requirements. Efforts in this area help providers to stay up to date in current practice standards and new developments, thus reducing one’s risk of incurring a malpractice claim.

Patient Safety

All members of the care team should engender an environment that is focused on patient safety, including open communication between providers and with nursing staff and technical support teams. Although interruptions can be detrimental to patient care, simply having an understanding of this phenomenon among all staff members can alleviate some of the working stressors in the ED. Effort must be made to create an environment that allows for clarification between nursing staff and physicians without causing undue antagonism. Fostering supportive communication, having a questioning attitude, and seeking clarification can only enhance patient safety.

The importance of the supervisory role of attending physicians to trainees, physician extenders, and nursing staff must be emphasized, and appropriate guidance from the ED attending is germane in keeping patients safe in teaching environments. Additionally, in departments that suffer the burden of high numbers of admitted patient boarders in the ED, attention must be given to the transitional period between decision to admit and termination of ED care and the acquisition of care of the admitting physician. A clear plan of responsibility must be in place for these high-risk situations.

Policies and Procedures

Departmental policies and procedures should be designed to identify and address all late laboratory results data, radiological discrepancies, and culture results in a timely and uniform manner. Since unaddressed results and discrepancies can result in patient harm, patient-callback processes should be designed to reduce risk by addressing these hazards regularly, thoroughly, and in a timely fashion.

Cognitive Biases

An awareness of inherent biases in the medical decision-making process is also helpful to maintain mindfulness in the routine practice of EM and avoid medical errors. The EP should take care not to be influenced by recent events and diagnostic information that is easy to recall or common, and to ensure the differential addresses possibilities beyond the readily available diagnoses. Further, reliance on an existing opinion may be misleading if subsequent judgments are based on this “anchor,” whether it is true or false.

 

If the data points of the case do not line up as expected, or if there are unexplained outliers, the EP should expand the frame of reference to seek more appropriate possibilities, and avoid attempts to make the data fit a preferred or favored conclusion.

When one fails to recognize that data do not fit the diagnostic presumption, the true diagnosis can be undermined. Such confirmation bias in turn challenges diagnostic success. Hasty judgment without considering and seeking out relevant information can set up diagnostic failure and premature closure.

Remembering the Basics

Finally, providers should follow the basic principles for every patient. Vital signs are vital for a reason, and all abnormal data must be accounted for prior to patient hand off or discharge. Patient turnover is a high-risk occasion, and demands careful attention to case details between the off-going physician, the accepting physician, and the patient.

All patients presenting to the ED for care should leave the ED at their baseline functional level (ie, if they walk independently, they should still walk independently at discharge). If not, the reason should be sought out and clarified with appropriate recommendations for treatment and follow-up.

Patients and staff should always be treated with respect, which in turn will encourage effective communication. Providers should be honest with patients, document truthfully, respect privacy and confidentiality, practice within one’s competence, confirm information, and avoid assumptions. Compassion goes hand in hand with respectful and open communication. Physicians perceived as compassionate and trustworthy are less likely to be the target of a malpractice suit, even when harm has occurred.

Conclusion

Even though the number of paid medical malpractice claims has continued to decrease over the past 20 years, a discrepancy between perceived and absolute risk persists among EPs—one that perpetuates the practice of defensive medicine and continues to affect EM. Despite the current perceptions and climate, EPs can allay their risk of incurring a malpractice claim by employing the strategies outlined above.

Most medical malpractice cases are still resolved in a courtroom—typically after years of preparation and personal torment. Yet, overall rates of paid medical malpractice claims among all physicians have been steadily decreasing over the past two decades, with reports showing decreases of 30% to 50% in paid claims since 2000.^1-3 At the same time, while median payments and insurance premiums continued to increase until the mid-2000s, they now appear to have plateaued.¹

None of these changes occurred in isolation. More than 30 states now have caps on noneconomic or total damages.² As noted in part 1, since 2000, some states have enacted comprehensive tort reform.⁴ However, whether these changes in malpractice patterns can be attributed directly to specific policy changes remains a hotly contested issue.

Malpractice Risk in Emergency Medicine

To what extent do the trends in medical malpractice apply to emergency medicine (EM)? While emergency physicians’ (EPs’) perception of malpractice risk ranks higher than any other medical specialty,5 in a review of a large sample of malpractice claims from 1991 through 2005, EPs ranked in the middle among specialties with respect to annual risk of a malpractice claim.6 Moreover, the annual risk of a claim for EPs is just under 8%, compared to 7.4% for all physicians. Yet, for neurosurgery and cardiothoracic surgery—the specialties with the highest overall risk of malpractice claims—the annual risk approaches 20%.6 Regarding payout statistics, less than one-fifth of the claims against EPs resulted in payment.6  In a review of a separate insurance database of closed claims, EPs were named as the primary defendant in only 19% of cases.7

Despite the discrepancies between perceived risk and absolute risk of malpractice claims among EPs, malpractice lawsuits continue to affect the practice of EM. This is evidenced in several surveys, in which the majority of EP participants admitted to practicing “defensive medicine” by ordering tests that were felt to be unnecessary and did so in response to perceived malpractice risk.^8-10 Perceived risk also accounts for the significant variation in decision-making in the ED with respect to diagnostic testing and hospitalization of patients.¹¹ One would expect that lowering malpractice risk would result in less so-called unnecessary testing, but whether or not this is truly the case remains to be seen.

Effects of Malpractice Reform

A study by Waxman et al¹² on the effects of significant malpractice tort reform in ED care in Texas, Georgia, and South Carolina found no difference in rates of imaging studies, charges, or patient admissions. Furthermore, legislation reform did not increase plaintiff onus to prove proximate “gross negligence” rather than simply a breach from “reasonably skillful and careful” medicine.¹² These findings suggest that perception of malpractice risk might simply be serving as a proxy for physicians’ underlying risk tolerance, and be less subject to influence by external forces.

Areas Associated With Malpractice Risk

A number of closed-claim databases attempted to identify the characteristics of patient encounters that can lead to malpractice claims, including patient conditions and sources of error. Diagnostic errors have consistently been found to be the leading cause of malpractice claims, accounting for 28% to 65% of claims, followed by inappropriate management of medical treatment and improper performance of a procedure.^7,13-16 A January 2016 benchmarking system report by CRICO Strategies found that 30% of 23,658 medical malpractice claims filed between 2009 through 2013 cited failures in communication as a factor.¹⁷ The report also revealed that among these failed communications, those that occurred between health care providers are more likely to result in payout compared to miscommunications between providers and patients.¹⁷ This report further noted 70% to 80% of claims closed without payment.^7,16 Closed claims were significantly more likely to involve serious injuries or death.^7,18 Leading conditions that resulted in claims include myocardial infarction, nonspecific chest pain, symptoms involving the abdomen or pelvis, appendicitis, and orthopedic injuries.^7,13,16

Diagnostic Errors

Errors in diagnosis have been attributed to multiple factors in the ED. The two most common factors were failure to order tests and failure to perform an adequate history and physical examination, both of which contribute to rationalization of the practice of defensive medicine under the current tort system.¹³ Other significant factors associated with errors in diagnosis include misinterpretation of test results or imaging studies and failure to obtain an appropriate consultation. Processes contributing to each of these potential errors include mistakes in judgment, lack of knowledge, miscommunication, and insufficient documentation (Table).¹⁵

Strategies for Reducing Malpractice Risk

In part 1, we listed several strategies EPs could adopt to help reduce malpractice risk. In this section, we will discuss in further detail how these strategies help mitigate malpractice claims.

 

Patient Communication

Open communication with patients is paramount in reducing the risk of a malpractice allegation. Patients are more likely to become angry or frustrated if they sense a physician is not listening to or addressing their concerns. These patients are in turn more likely to file a complaint if they are harmed or experience a bad outcome during their stay in the ED.

Situations in which patients are unable to provide pertinent information also place the EP at significant risk, as the provider must make decisions without full knowledge of the case. Communication with potential resources such as nursing home staff, the patient’s family, and emergency medical service providers to obtain additional information can help reduce risk. 

Of course, when evaluating and treating patients, the EP should always take the time to listen to the patient’s concerns during the encounter to ensure his or her needs have been addressed. In the event of a patient allegation or complaint, the EP should make the effort to explore and de-escalate the situation before the patient is discharged.

Discharge Care and Instructions

According to CRICO, premature discharge as a factor in medical malpractice liability results from inadequate assessment and missed opportunities in 41% of diagnosis-related ED cases.¹⁶ The following situation illustrates a brief example of such a missed opportunity: A provider makes a diagnosis of urinary tract infection (UTI) in a patient presenting with fever and abdominal pain but whose urinalysis is suspect for contamination and in whom no pelvic examination was performed to rule out other etiologies. When the same patient later returns to the ED with worse abdominal pain, a sterile urine culture invalidates the diagnosis of UTI, and further evaluation leads to a final diagnosis of ruptured appendix.

Prior to discharging any patient, the EP should provide clear and concise at-home care instructions in a manner in which the patient can understand. Clear instructions on how the patient is to manage his or her care after discharge are vital, and failure to do so in terms the patient can understand can create problems if a harmful result occurs. This is especially important in patients with whom there is a communication barrier—eg, language barrier, hearing impairment, cognitive deficit, intoxication, or violent or irrational behavior. In these situations, the EP should always take advantage of available resources and tools such as language lines, interpreters, discharge planners, psychiatric staff, and supportive family members to help reconcile any communication barriers. These measures will in turn optimize patient outcome and reduce the risk of a later malpractice allegation.

Board Certification

All physicians should maintain their respective board certification and specialty training requirements. Efforts in this area help providers to stay up to date in current practice standards and new developments, thus reducing one’s risk of incurring a malpractice claim.

Patient Safety

All members of the care team should engender an environment that is focused on patient safety, including open communication between providers and with nursing staff and technical support teams. Although interruptions can be detrimental to patient care, simply having an understanding of this phenomenon among all staff members can alleviate some of the working stressors in the ED. Effort must be made to create an environment that allows for clarification between nursing staff and physicians without causing undue antagonism. Fostering supportive communication, having a questioning attitude, and seeking clarification can only enhance patient safety.

The importance of the supervisory role of attending physicians to trainees, physician extenders, and nursing staff must be emphasized, and appropriate guidance from the ED attending is germane in keeping patients safe in teaching environments. Additionally, in departments that suffer the burden of high numbers of admitted patient boarders in the ED, attention must be given to the transitional period between decision to admit and termination of ED care and the acquisition of care of the admitting physician. A clear plan of responsibility must be in place for these high-risk situations.

Policies and Procedures

Departmental policies and procedures should be designed to identify and address all late laboratory results data, radiological discrepancies, and culture results in a timely and uniform manner. Since unaddressed results and discrepancies can result in patient harm, patient-callback processes should be designed to reduce risk by addressing these hazards regularly, thoroughly, and in a timely fashion.

Cognitive Biases

An awareness of inherent biases in the medical decision-making process is also helpful to maintain mindfulness in the routine practice of EM and avoid medical errors. The EP should take care not to be influenced by recent events and diagnostic information that is easy to recall or common, and to ensure the differential addresses possibilities beyond the readily available diagnoses. Further, reliance on an existing opinion may be misleading if subsequent judgments are based on this “anchor,” whether it is true or false.

 

If the data points of the case do not line up as expected, or if there are unexplained outliers, the EP should expand the frame of reference to seek more appropriate possibilities, and avoid attempts to make the data fit a preferred or favored conclusion.

When one fails to recognize that data do not fit the diagnostic presumption, the true diagnosis can be undermined. Such confirmation bias in turn challenges diagnostic success. Hasty judgment without considering and seeking out relevant information can set up diagnostic failure and premature closure.

Remembering the Basics

Finally, providers should follow the basic principles for every patient. Vital signs are vital for a reason, and all abnormal data must be accounted for prior to patient hand off or discharge. Patient turnover is a high-risk occasion, and demands careful attention to case details between the off-going physician, the accepting physician, and the patient.

All patients presenting to the ED for care should leave the ED at their baseline functional level (ie, if they walk independently, they should still walk independently at discharge). If not, the reason should be sought out and clarified with appropriate recommendations for treatment and follow-up.

Patients and staff should always be treated with respect, which in turn will encourage effective communication. Providers should be honest with patients, document truthfully, respect privacy and confidentiality, practice within one’s competence, confirm information, and avoid assumptions. Compassion goes hand in hand with respectful and open communication. Physicians perceived as compassionate and trustworthy are less likely to be the target of a malpractice suit, even when harm has occurred.

Conclusion

Even though the number of paid medical malpractice claims has continued to decrease over the past 20 years, a discrepancy between perceived and absolute risk persists among EPs—one that perpetuates the practice of defensive medicine and continues to affect EM. Despite the current perceptions and climate, EPs can allay their risk of incurring a malpractice claim by employing the strategies outlined above.

Most medical malpractice cases are still resolved in a courtroom—typically after years of preparation and personal torment. Yet, overall rates of paid medical malpractice claims among all physicians have been steadily decreasing over the past two decades, with reports showing decreases of 30% to 50% in paid claims since 2000.^1-3 At the same time, while median payments and insurance premiums continued to increase until the mid-2000s, they now appear to have plateaued.¹

None of these changes occurred in isolation. More than 30 states now have caps on noneconomic or total damages.² As noted in part 1, since 2000, some states have enacted comprehensive tort reform.⁴ However, whether these changes in malpractice patterns can be attributed directly to specific policy changes remains a hotly contested issue.

Malpractice Risk in Emergency Medicine

To what extent do the trends in medical malpractice apply to emergency medicine (EM)? While emergency physicians’ (EPs’) perception of malpractice risk ranks higher than any other medical specialty,5 in a review of a large sample of malpractice claims from 1991 through 2005, EPs ranked in the middle among specialties with respect to annual risk of a malpractice claim.6 Moreover, the annual risk of a claim for EPs is just under 8%, compared to 7.4% for all physicians. Yet, for neurosurgery and cardiothoracic surgery—the specialties with the highest overall risk of malpractice claims—the annual risk approaches 20%.6 Regarding payout statistics, less than one-fifth of the claims against EPs resulted in payment.6  In a review of a separate insurance database of closed claims, EPs were named as the primary defendant in only 19% of cases.7

Despite the discrepancies between perceived risk and absolute risk of malpractice claims among EPs, malpractice lawsuits continue to affect the practice of EM. This is evidenced in several surveys, in which the majority of EP participants admitted to practicing “defensive medicine” by ordering tests that were felt to be unnecessary and did so in response to perceived malpractice risk.^8-10 Perceived risk also accounts for the significant variation in decision-making in the ED with respect to diagnostic testing and hospitalization of patients.¹¹ One would expect that lowering malpractice risk would result in less so-called unnecessary testing, but whether or not this is truly the case remains to be seen.

Effects of Malpractice Reform

A study by Waxman et al¹² on the effects of significant malpractice tort reform in ED care in Texas, Georgia, and South Carolina found no difference in rates of imaging studies, charges, or patient admissions. Furthermore, legislation reform did not increase plaintiff onus to prove proximate “gross negligence” rather than simply a breach from “reasonably skillful and careful” medicine.¹² These findings suggest that perception of malpractice risk might simply be serving as a proxy for physicians’ underlying risk tolerance, and be less subject to influence by external forces.

Areas Associated With Malpractice Risk

A number of closed-claim databases attempted to identify the characteristics of patient encounters that can lead to malpractice claims, including patient conditions and sources of error. Diagnostic errors have consistently been found to be the leading cause of malpractice claims, accounting for 28% to 65% of claims, followed by inappropriate management of medical treatment and improper performance of a procedure.^7,13-16 A January 2016 benchmarking system report by CRICO Strategies found that 30% of 23,658 medical malpractice claims filed between 2009 through 2013 cited failures in communication as a factor.¹⁷ The report also revealed that among these failed communications, those that occurred between health care providers are more likely to result in payout compared to miscommunications between providers and patients.¹⁷ This report further noted 70% to 80% of claims closed without payment.^7,16 Closed claims were significantly more likely to involve serious injuries or death.^7,18 Leading conditions that resulted in claims include myocardial infarction, nonspecific chest pain, symptoms involving the abdomen or pelvis, appendicitis, and orthopedic injuries.^7,13,16

Diagnostic Errors

Errors in diagnosis have been attributed to multiple factors in the ED. The two most common factors were failure to order tests and failure to perform an adequate history and physical examination, both of which contribute to rationalization of the practice of defensive medicine under the current tort system.¹³ Other significant factors associated with errors in diagnosis include misinterpretation of test results or imaging studies and failure to obtain an appropriate consultation. Processes contributing to each of these potential errors include mistakes in judgment, lack of knowledge, miscommunication, and insufficient documentation (Table).¹⁵

Strategies for Reducing Malpractice Risk

In part 1, we listed several strategies EPs could adopt to help reduce malpractice risk. In this section, we will discuss in further detail how these strategies help mitigate malpractice claims.

 

Patient Communication

Open communication with patients is paramount in reducing the risk of a malpractice allegation. Patients are more likely to become angry or frustrated if they sense a physician is not listening to or addressing their concerns. These patients are in turn more likely to file a complaint if they are harmed or experience a bad outcome during their stay in the ED.

Situations in which patients are unable to provide pertinent information also place the EP at significant risk, as the provider must make decisions without full knowledge of the case. Communication with potential resources such as nursing home staff, the patient’s family, and emergency medical service providers to obtain additional information can help reduce risk. 

Of course, when evaluating and treating patients, the EP should always take the time to listen to the patient’s concerns during the encounter to ensure his or her needs have been addressed. In the event of a patient allegation or complaint, the EP should make the effort to explore and de-escalate the situation before the patient is discharged.

Discharge Care and Instructions

According to CRICO, premature discharge as a factor in medical malpractice liability results from inadequate assessment and missed opportunities in 41% of diagnosis-related ED cases.¹⁶ The following situation illustrates a brief example of such a missed opportunity: A provider makes a diagnosis of urinary tract infection (UTI) in a patient presenting with fever and abdominal pain but whose urinalysis is suspect for contamination and in whom no pelvic examination was performed to rule out other etiologies. When the same patient later returns to the ED with worse abdominal pain, a sterile urine culture invalidates the diagnosis of UTI, and further evaluation leads to a final diagnosis of ruptured appendix.

Prior to discharging any patient, the EP should provide clear and concise at-home care instructions in a manner in which the patient can understand. Clear instructions on how the patient is to manage his or her care after discharge are vital, and failure to do so in terms the patient can understand can create problems if a harmful result occurs. This is especially important in patients with whom there is a communication barrier—eg, language barrier, hearing impairment, cognitive deficit, intoxication, or violent or irrational behavior. In these situations, the EP should always take advantage of available resources and tools such as language lines, interpreters, discharge planners, psychiatric staff, and supportive family members to help reconcile any communication barriers. These measures will in turn optimize patient outcome and reduce the risk of a later malpractice allegation.

Board Certification

All physicians should maintain their respective board certification and specialty training requirements. Efforts in this area help providers to stay up to date in current practice standards and new developments, thus reducing one’s risk of incurring a malpractice claim.

Patient Safety

All members of the care team should engender an environment that is focused on patient safety, including open communication between providers and with nursing staff and technical support teams. Although interruptions can be detrimental to patient care, simply having an understanding of this phenomenon among all staff members can alleviate some of the working stressors in the ED. Effort must be made to create an environment that allows for clarification between nursing staff and physicians without causing undue antagonism. Fostering supportive communication, having a questioning attitude, and seeking clarification can only enhance patient safety.

The importance of the supervisory role of attending physicians to trainees, physician extenders, and nursing staff must be emphasized, and appropriate guidance from the ED attending is germane in keeping patients safe in teaching environments. Additionally, in departments that suffer the burden of high numbers of admitted patient boarders in the ED, attention must be given to the transitional period between decision to admit and termination of ED care and the acquisition of care of the admitting physician. A clear plan of responsibility must be in place for these high-risk situations.

Policies and Procedures

Departmental policies and procedures should be designed to identify and address all late laboratory results data, radiological discrepancies, and culture results in a timely and uniform manner. Since unaddressed results and discrepancies can result in patient harm, patient-callback processes should be designed to reduce risk by addressing these hazards regularly, thoroughly, and in a timely fashion.

Cognitive Biases

An awareness of inherent biases in the medical decision-making process is also helpful to maintain mindfulness in the routine practice of EM and avoid medical errors. The EP should take care not to be influenced by recent events and diagnostic information that is easy to recall or common, and to ensure the differential addresses possibilities beyond the readily available diagnoses. Further, reliance on an existing opinion may be misleading if subsequent judgments are based on this “anchor,” whether it is true or false.

 

If the data points of the case do not line up as expected, or if there are unexplained outliers, the EP should expand the frame of reference to seek more appropriate possibilities, and avoid attempts to make the data fit a preferred or favored conclusion.

When one fails to recognize that data do not fit the diagnostic presumption, the true diagnosis can be undermined. Such confirmation bias in turn challenges diagnostic success. Hasty judgment without considering and seeking out relevant information can set up diagnostic failure and premature closure.

Remembering the Basics

Finally, providers should follow the basic principles for every patient. Vital signs are vital for a reason, and all abnormal data must be accounted for prior to patient hand off or discharge. Patient turnover is a high-risk occasion, and demands careful attention to case details between the off-going physician, the accepting physician, and the patient.

All patients presenting to the ED for care should leave the ED at their baseline functional level (ie, if they walk independently, they should still walk independently at discharge). If not, the reason should be sought out and clarified with appropriate recommendations for treatment and follow-up.

Patients and staff should always be treated with respect, which in turn will encourage effective communication. Providers should be honest with patients, document truthfully, respect privacy and confidentiality, practice within one’s competence, confirm information, and avoid assumptions. Compassion goes hand in hand with respectful and open communication. Physicians perceived as compassionate and trustworthy are less likely to be the target of a malpractice suit, even when harm has occurred.

Conclusion

Even though the number of paid medical malpractice claims has continued to decrease over the past 20 years, a discrepancy between perceived and absolute risk persists among EPs—one that perpetuates the practice of defensive medicine and continues to affect EM. Despite the current perceptions and climate, EPs can allay their risk of incurring a malpractice claim by employing the strategies outlined above.

Prehospital spinal immobilization has long been the standard of care (SOC) to prevent spinal cord injury in trauma patients, but utilizing the best data currently available, some professional societies recently released new recommendations that question this practice. Guidelines released in 2014 from the National Association of EMS Physicians (NAEMSP) and the American College of Surgeons Committee on Trauma (ACS-COT) support limited application of spinal immobilization.¹ These guidelines note, “Given the rarity of unstable spinal injuries in EMS trauma patients, the number that might benefit from immobilization to prevent secondary injury is likely extremely small. For each patient who has potential benefit, hundreds to thousands of patients must undergo immobilization with no potential benefit.” Further, they advise “utilization of backboards for spinal immobilization during transport should be judicious, so that potential benefits outweigh risks.”¹ Spinal immobilization should not be used at all in patients with penetrating trauma who do not present with obvious neurological injury and should be selective, based on objective findings of injury or the high potential for same.¹

Questioning a Long-standing Practice

Fear of the consequences of spinal cord injury from significant vertebral fractures has dictated prehospital spinal immobilization to manage injured trauma patients for decades. For almost 50 years, it has been the SOC. However, increasing evidence that spinal immobilization is not only unnecessary, but may even cause harm has resulted in questioning this paradigm, which has lead to promoting a change in the SOC.

Spinal immobilization dates back to the mid-1960s, when Geisler et al² reported on a cohort of patients who suffered long-term paralysis from what was believed to be improper handling and failure to discover spinal injuries. Soon after, Farrington^3,4 developed and published a systematic approach to spinal immobilization during extrication following blunt force trauma, supporting the widespread acceptance of backboards and cervical collars to immobilize the spine in injured trauma patients. Logic dictated that an unstable spine fracture could be worsened, or a cord injury could result, by unnecessary movement during extrication, transport, and initial evaluation in the ED, resulting in avoidable injury. This fear of potential secondary injury grew as more papers were published examining the link between prehospital handling of blunt force trauma patients and delayed paralysis. This resulted in the use of spinal immobilization on the majority of trauma patients, regardless of mechanism of injury or presenting symptoms.^5,6

One review estimated that over 50% of trauma patients with no complaint of neck or back pain were transported with full spinal immobilization.⁷ This immobilization on uncomfortable long backboards typically continued in the ED for prolonged periods, until the spine could be cleared by physical examination and/or imaging studies. Yet a 2001 Cochrane review found that despite increasing use of spinal immobilization, no prospective, randomized controlled trial of the appropriate use of spinal immobilization or patient outcomes had ever been conducted.⁸

What the Evidence Says

How much evidence exists that supports the benefits of spinal immobilization? Not much. Studies on healthy volunteers and cadavers evaluating spinal motion with immobilization have been contradictory.⁹ One study found there was less motion with a cervical collar in place than without,¹⁰ whereas others found that the use of a cervical collar did not effectively reduce motion in an unstable spine.^11,12 Perry et al¹³ studied the effectiveness of different head immobilization techniques and found that none could eliminate head and neck motion during emergency medical services (EMS) transport. Still other reports, including two biomechanical studies, demonstrated increased neck motion when using conventional extrication techniques (cervical collar with backboard) versus controlled self-extrication with cervical collar only.^14,15

An Abundance of Literature on the Risks

Whereas data regarding the actual benefits of spinal immobilization is lacking, an abundance of literature details the risks. One of the most frequently cited studies is also one of the most controversial. Hauswald et al¹⁶ compared the outcomes of two groups of patients with blunt force trauma who were either immobilized during transport (in New Mexico) or non-immobilized (in Malaysia) and found that the risk of disability was higher in the immobilized group (odds ratio, 2.03). Although these environments are very different, the authors noted that mechanism of injury, resources, and the size of the hospitals were similar.¹⁶

Studies of spinal immobilization in patients with penetrating trauma report even worse outcomes. In separate studies, Haut et al¹⁷ and Vanderlan et al¹⁸ demonstrated increased mortality when immobilization led to increased transport times and interference with other resuscitative measures. These and other studies have led the American College of Emergency Physicians, NAEMSP, ACS-COT, the Prehospital Trauma Life Support Executive Committee, and other national organizations to recommend no spinal immobilization in patients with penetrating neck trauma.^1,19,20

 

Many trauma patients arrive with complaints of pain at one or more sites. Some of these complaints, particularly back pain, may be secondary to the use of the backboard itself, especially in cases of prolonged transport.^21,22 In a study of healthy volunteers who were immobilized on a backboard for 30 minutes, all of them reported pain, along with headaches, most often involving the occipital and sacral regions.²³ A 1996 study compared spinal immobilization utilizing a backboard versus a vacuum mattress in 37 healthy volunteers with no history of back pain or spinal disease.²⁴ Compared to those immobilized with the vacuum mattress, patients immobilized with a backboard for 30 minutes were 3.1 times more likely to have symptoms, 7.9 times more likely to complain of occipital pain, and 4.3 times more likely to have lumbosacral pain.²⁴

Increased pain complaints in the setting of trauma can result in increased imaging, leading to increased costs and unnecessary radiation exposure.²⁵ Prolonged backboard times can also result in sacral pressure ulcers.²⁶ A recent study has shown that patients who undergo computed tomography (CT) scans with automatic tube current modulation (as most modern multidetector row CT systems utilize) while on a backboard  may be exposed to a significant increase in radiation dose.²⁷

Spinal immobilization has also been linked to respiratory compromise, particularly with the use of straps across the chest, even when not applied tightly. One study found worse lung function test results in healthy immobilized volunteers.²⁸ Other studies have shown that older patients (even when healthy) and those with lung or chest injury have an even larger degree of restriction and respiratory compromise.^29,30

Risks from immobilization are not isolated to backboards. The use of cervical collars alone also carries potential risks. (See “What About Cervical Collars?”^8,31-39)

Risk of Secondary Neurological Deterioration Is Low

Many EMS systems have already adopted the new standards calling for less use of spinal immobilization. Though the evidence is compelling, not all EMS systems have adopted these standards due to strongly rooted beliefs and fears of long-term patient disability and subsequent litigation. However, these fears do not appear justified.

A recent review by Oto et al⁴⁰ found only 42 cases of early secondary neurological deterioration after blunt trauma in all of the indexed medical literature. They noted, “In twelve cases the authors did attribute deterioration to temporally associated precipitants, seven of which were possibly iatrogenic; these included removal of a cervical collar, placement of a halo device, patient agitation, performance of flexion/extension films, ‘unintentional manipulation,’ falling in or near the ED, and forced collar application in patients with ankylosing spondylitis.” Thirteen of these cases occurred during prehospital care, none of them sudden and movement-provoked, and all reported by a single study.” This review highlights the rarity of secondary deterioration.

When Should Immobilization Be Used?

So what’s the next step for spinal immobilization in the field? How do we appropriately protect trauma patients during transport? As always seems to be the case in medicine, more evidence is needed. Oteir et al⁴¹ recently published a review of new literature on the epidemiology and current practice of prehospital spine management. They reported that early (8-24 hours) transfer of patients with spinal injury to spinal care units, along with effective resuscitation, was the most important determinant of better neurological outcomes.⁴¹ This review reaffirms the need for more data evaluating the relationship between spinal immobilization and neurological outcomes.

Currently, recommendations call for selective spinal immobilization to decrease unnecessary application and potential harm. Use of backboards for spinal immobilization should be limited to the following types of patients:^1,20

• Blunt trauma and altered level of consciousness;
• Spinal pain or tenderness;
• Neurological complaint (eg, numbness or motor weakness);
• Anatomic deformity of the spine;
• High-energy mechanism of injury and:
  
  • Drug or alcohol intoxication;
  • Inability to communicate; and/or
  • Distracting injury.

Patients for whom immobilization on a backboard is not necessary include those with all of the following:

• Normal level of consciousness (GCS 15);
• No spine tenderness or anatomic abnormality;
• No neurological findings or complaints;
• No distracting injury;
• No intoxication.

Cervical collars alone are still recommended for use in patients who do not meet validated clinical rules, such as the NEXUS or Canadian C spine rules.^1,20,42,43 As these rules are well validated, they can be safely used to determine who should have a cervical collar placed, with or without a backboard. In a retrospective review, selective spinal immobilization was found to be 99% sensitive in identifying patients with cervical injuries.⁴⁴

Clearly, there is still work to be done. Due to the relative rarity of actual spinal cord injury with the conse­quences of neurological injury, prospective trials in this area are rare and very difficult to safely design. However, there is growing confidence that selective spinal protocols, together with the inclusion of validated clinical rules, can effectively limit exposure to unnecessary spinal immobilization. As the current evidence continues to mount for the potential harm in indiscriminate backboard and cervical collar use, it seems clear we should strive to decrease the overuse of prehospital and early spinal immobilization consistent with current position statements and validated clinical rules.

Prehospital spinal immobilization has long been the standard of care (SOC) to prevent spinal cord injury in trauma patients, but utilizing the best data currently available, some professional societies recently released new recommendations that question this practice. Guidelines released in 2014 from the National Association of EMS Physicians (NAEMSP) and the American College of Surgeons Committee on Trauma (ACS-COT) support limited application of spinal immobilization.¹ These guidelines note, “Given the rarity of unstable spinal injuries in EMS trauma patients, the number that might benefit from immobilization to prevent secondary injury is likely extremely small. For each patient who has potential benefit, hundreds to thousands of patients must undergo immobilization with no potential benefit.” Further, they advise “utilization of backboards for spinal immobilization during transport should be judicious, so that potential benefits outweigh risks.”¹ Spinal immobilization should not be used at all in patients with penetrating trauma who do not present with obvious neurological injury and should be selective, based on objective findings of injury or the high potential for same.¹

Questioning a Long-standing Practice

Fear of the consequences of spinal cord injury from significant vertebral fractures has dictated prehospital spinal immobilization to manage injured trauma patients for decades. For almost 50 years, it has been the SOC. However, increasing evidence that spinal immobilization is not only unnecessary, but may even cause harm has resulted in questioning this paradigm, which has lead to promoting a change in the SOC.

Spinal immobilization dates back to the mid-1960s, when Geisler et al² reported on a cohort of patients who suffered long-term paralysis from what was believed to be improper handling and failure to discover spinal injuries. Soon after, Farrington^3,4 developed and published a systematic approach to spinal immobilization during extrication following blunt force trauma, supporting the widespread acceptance of backboards and cervical collars to immobilize the spine in injured trauma patients. Logic dictated that an unstable spine fracture could be worsened, or a cord injury could result, by unnecessary movement during extrication, transport, and initial evaluation in the ED, resulting in avoidable injury. This fear of potential secondary injury grew as more papers were published examining the link between prehospital handling of blunt force trauma patients and delayed paralysis. This resulted in the use of spinal immobilization on the majority of trauma patients, regardless of mechanism of injury or presenting symptoms.^5,6

One review estimated that over 50% of trauma patients with no complaint of neck or back pain were transported with full spinal immobilization.⁷ This immobilization on uncomfortable long backboards typically continued in the ED for prolonged periods, until the spine could be cleared by physical examination and/or imaging studies. Yet a 2001 Cochrane review found that despite increasing use of spinal immobilization, no prospective, randomized controlled trial of the appropriate use of spinal immobilization or patient outcomes had ever been conducted.⁸

What the Evidence Says

How much evidence exists that supports the benefits of spinal immobilization? Not much. Studies on healthy volunteers and cadavers evaluating spinal motion with immobilization have been contradictory.⁹ One study found there was less motion with a cervical collar in place than without,¹⁰ whereas others found that the use of a cervical collar did not effectively reduce motion in an unstable spine.^11,12 Perry et al¹³ studied the effectiveness of different head immobilization techniques and found that none could eliminate head and neck motion during emergency medical services (EMS) transport. Still other reports, including two biomechanical studies, demonstrated increased neck motion when using conventional extrication techniques (cervical collar with backboard) versus controlled self-extrication with cervical collar only.^14,15

An Abundance of Literature on the Risks

Whereas data regarding the actual benefits of spinal immobilization is lacking, an abundance of literature details the risks. One of the most frequently cited studies is also one of the most controversial. Hauswald et al¹⁶ compared the outcomes of two groups of patients with blunt force trauma who were either immobilized during transport (in New Mexico) or non-immobilized (in Malaysia) and found that the risk of disability was higher in the immobilized group (odds ratio, 2.03). Although these environments are very different, the authors noted that mechanism of injury, resources, and the size of the hospitals were similar.¹⁶

Studies of spinal immobilization in patients with penetrating trauma report even worse outcomes. In separate studies, Haut et al¹⁷ and Vanderlan et al¹⁸ demonstrated increased mortality when immobilization led to increased transport times and interference with other resuscitative measures. These and other studies have led the American College of Emergency Physicians, NAEMSP, ACS-COT, the Prehospital Trauma Life Support Executive Committee, and other national organizations to recommend no spinal immobilization in patients with penetrating neck trauma.^1,19,20

 

Many trauma patients arrive with complaints of pain at one or more sites. Some of these complaints, particularly back pain, may be secondary to the use of the backboard itself, especially in cases of prolonged transport.^21,22 In a study of healthy volunteers who were immobilized on a backboard for 30 minutes, all of them reported pain, along with headaches, most often involving the occipital and sacral regions.²³ A 1996 study compared spinal immobilization utilizing a backboard versus a vacuum mattress in 37 healthy volunteers with no history of back pain or spinal disease.²⁴ Compared to those immobilized with the vacuum mattress, patients immobilized with a backboard for 30 minutes were 3.1 times more likely to have symptoms, 7.9 times more likely to complain of occipital pain, and 4.3 times more likely to have lumbosacral pain.²⁴

Increased pain complaints in the setting of trauma can result in increased imaging, leading to increased costs and unnecessary radiation exposure.²⁵ Prolonged backboard times can also result in sacral pressure ulcers.²⁶ A recent study has shown that patients who undergo computed tomography (CT) scans with automatic tube current modulation (as most modern multidetector row CT systems utilize) while on a backboard  may be exposed to a significant increase in radiation dose.²⁷

Spinal immobilization has also been linked to respiratory compromise, particularly with the use of straps across the chest, even when not applied tightly. One study found worse lung function test results in healthy immobilized volunteers.²⁸ Other studies have shown that older patients (even when healthy) and those with lung or chest injury have an even larger degree of restriction and respiratory compromise.^29,30

Risks from immobilization are not isolated to backboards. The use of cervical collars alone also carries potential risks. (See “What About Cervical Collars?”^8,31-39)

Risk of Secondary Neurological Deterioration Is Low

Many EMS systems have already adopted the new standards calling for less use of spinal immobilization. Though the evidence is compelling, not all EMS systems have adopted these standards due to strongly rooted beliefs and fears of long-term patient disability and subsequent litigation. However, these fears do not appear justified.

A recent review by Oto et al⁴⁰ found only 42 cases of early secondary neurological deterioration after blunt trauma in all of the indexed medical literature. They noted, “In twelve cases the authors did attribute deterioration to temporally associated precipitants, seven of which were possibly iatrogenic; these included removal of a cervical collar, placement of a halo device, patient agitation, performance of flexion/extension films, ‘unintentional manipulation,’ falling in or near the ED, and forced collar application in patients with ankylosing spondylitis.” Thirteen of these cases occurred during prehospital care, none of them sudden and movement-provoked, and all reported by a single study.” This review highlights the rarity of secondary deterioration.

When Should Immobilization Be Used?

So what’s the next step for spinal immobilization in the field? How do we appropriately protect trauma patients during transport? As always seems to be the case in medicine, more evidence is needed. Oteir et al⁴¹ recently published a review of new literature on the epidemiology and current practice of prehospital spine management. They reported that early (8-24 hours) transfer of patients with spinal injury to spinal care units, along with effective resuscitation, was the most important determinant of better neurological outcomes.⁴¹ This review reaffirms the need for more data evaluating the relationship between spinal immobilization and neurological outcomes.

Currently, recommendations call for selective spinal immobilization to decrease unnecessary application and potential harm. Use of backboards for spinal immobilization should be limited to the following types of patients:^1,20

• Blunt trauma and altered level of consciousness;
• Spinal pain or tenderness;
• Neurological complaint (eg, numbness or motor weakness);
• Anatomic deformity of the spine;
• High-energy mechanism of injury and:
  
  • Drug or alcohol intoxication;
  • Inability to communicate; and/or
  • Distracting injury.

Patients for whom immobilization on a backboard is not necessary include those with all of the following:

• Normal level of consciousness (GCS 15);
• No spine tenderness or anatomic abnormality;
• No neurological findings or complaints;
• No distracting injury;
• No intoxication.

Cervical collars alone are still recommended for use in patients who do not meet validated clinical rules, such as the NEXUS or Canadian C spine rules.^1,20,42,43 As these rules are well validated, they can be safely used to determine who should have a cervical collar placed, with or without a backboard. In a retrospective review, selective spinal immobilization was found to be 99% sensitive in identifying patients with cervical injuries.⁴⁴

Clearly, there is still work to be done. Due to the relative rarity of actual spinal cord injury with the conse­quences of neurological injury, prospective trials in this area are rare and very difficult to safely design. However, there is growing confidence that selective spinal protocols, together with the inclusion of validated clinical rules, can effectively limit exposure to unnecessary spinal immobilization. As the current evidence continues to mount for the potential harm in indiscriminate backboard and cervical collar use, it seems clear we should strive to decrease the overuse of prehospital and early spinal immobilization consistent with current position statements and validated clinical rules.

Prehospital spinal immobilization has long been the standard of care (SOC) to prevent spinal cord injury in trauma patients, but utilizing the best data currently available, some professional societies recently released new recommendations that question this practice. Guidelines released in 2014 from the National Association of EMS Physicians (NAEMSP) and the American College of Surgeons Committee on Trauma (ACS-COT) support limited application of spinal immobilization.¹ These guidelines note, “Given the rarity of unstable spinal injuries in EMS trauma patients, the number that might benefit from immobilization to prevent secondary injury is likely extremely small. For each patient who has potential benefit, hundreds to thousands of patients must undergo immobilization with no potential benefit.” Further, they advise “utilization of backboards for spinal immobilization during transport should be judicious, so that potential benefits outweigh risks.”¹ Spinal immobilization should not be used at all in patients with penetrating trauma who do not present with obvious neurological injury and should be selective, based on objective findings of injury or the high potential for same.¹

Questioning a Long-standing Practice

Fear of the consequences of spinal cord injury from significant vertebral fractures has dictated prehospital spinal immobilization to manage injured trauma patients for decades. For almost 50 years, it has been the SOC. However, increasing evidence that spinal immobilization is not only unnecessary, but may even cause harm has resulted in questioning this paradigm, which has lead to promoting a change in the SOC.

Spinal immobilization dates back to the mid-1960s, when Geisler et al² reported on a cohort of patients who suffered long-term paralysis from what was believed to be improper handling and failure to discover spinal injuries. Soon after, Farrington^3,4 developed and published a systematic approach to spinal immobilization during extrication following blunt force trauma, supporting the widespread acceptance of backboards and cervical collars to immobilize the spine in injured trauma patients. Logic dictated that an unstable spine fracture could be worsened, or a cord injury could result, by unnecessary movement during extrication, transport, and initial evaluation in the ED, resulting in avoidable injury. This fear of potential secondary injury grew as more papers were published examining the link between prehospital handling of blunt force trauma patients and delayed paralysis. This resulted in the use of spinal immobilization on the majority of trauma patients, regardless of mechanism of injury or presenting symptoms.^5,6

One review estimated that over 50% of trauma patients with no complaint of neck or back pain were transported with full spinal immobilization.⁷ This immobilization on uncomfortable long backboards typically continued in the ED for prolonged periods, until the spine could be cleared by physical examination and/or imaging studies. Yet a 2001 Cochrane review found that despite increasing use of spinal immobilization, no prospective, randomized controlled trial of the appropriate use of spinal immobilization or patient outcomes had ever been conducted.⁸

What the Evidence Says

How much evidence exists that supports the benefits of spinal immobilization? Not much. Studies on healthy volunteers and cadavers evaluating spinal motion with immobilization have been contradictory.⁹ One study found there was less motion with a cervical collar in place than without,¹⁰ whereas others found that the use of a cervical collar did not effectively reduce motion in an unstable spine.^11,12 Perry et al¹³ studied the effectiveness of different head immobilization techniques and found that none could eliminate head and neck motion during emergency medical services (EMS) transport. Still other reports, including two biomechanical studies, demonstrated increased neck motion when using conventional extrication techniques (cervical collar with backboard) versus controlled self-extrication with cervical collar only.^14,15

An Abundance of Literature on the Risks

Whereas data regarding the actual benefits of spinal immobilization is lacking, an abundance of literature details the risks. One of the most frequently cited studies is also one of the most controversial. Hauswald et al¹⁶ compared the outcomes of two groups of patients with blunt force trauma who were either immobilized during transport (in New Mexico) or non-immobilized (in Malaysia) and found that the risk of disability was higher in the immobilized group (odds ratio, 2.03). Although these environments are very different, the authors noted that mechanism of injury, resources, and the size of the hospitals were similar.¹⁶

Studies of spinal immobilization in patients with penetrating trauma report even worse outcomes. In separate studies, Haut et al¹⁷ and Vanderlan et al¹⁸ demonstrated increased mortality when immobilization led to increased transport times and interference with other resuscitative measures. These and other studies have led the American College of Emergency Physicians, NAEMSP, ACS-COT, the Prehospital Trauma Life Support Executive Committee, and other national organizations to recommend no spinal immobilization in patients with penetrating neck trauma.^1,19,20

 

Many trauma patients arrive with complaints of pain at one or more sites. Some of these complaints, particularly back pain, may be secondary to the use of the backboard itself, especially in cases of prolonged transport.^21,22 In a study of healthy volunteers who were immobilized on a backboard for 30 minutes, all of them reported pain, along with headaches, most often involving the occipital and sacral regions.²³ A 1996 study compared spinal immobilization utilizing a backboard versus a vacuum mattress in 37 healthy volunteers with no history of back pain or spinal disease.²⁴ Compared to those immobilized with the vacuum mattress, patients immobilized with a backboard for 30 minutes were 3.1 times more likely to have symptoms, 7.9 times more likely to complain of occipital pain, and 4.3 times more likely to have lumbosacral pain.²⁴

Increased pain complaints in the setting of trauma can result in increased imaging, leading to increased costs and unnecessary radiation exposure.²⁵ Prolonged backboard times can also result in sacral pressure ulcers.²⁶ A recent study has shown that patients who undergo computed tomography (CT) scans with automatic tube current modulation (as most modern multidetector row CT systems utilize) while on a backboard  may be exposed to a significant increase in radiation dose.²⁷

Spinal immobilization has also been linked to respiratory compromise, particularly with the use of straps across the chest, even when not applied tightly. One study found worse lung function test results in healthy immobilized volunteers.²⁸ Other studies have shown that older patients (even when healthy) and those with lung or chest injury have an even larger degree of restriction and respiratory compromise.^29,30

Risks from immobilization are not isolated to backboards. The use of cervical collars alone also carries potential risks. (See “What About Cervical Collars?”^8,31-39)

Risk of Secondary Neurological Deterioration Is Low

Many EMS systems have already adopted the new standards calling for less use of spinal immobilization. Though the evidence is compelling, not all EMS systems have adopted these standards due to strongly rooted beliefs and fears of long-term patient disability and subsequent litigation. However, these fears do not appear justified.

A recent review by Oto et al⁴⁰ found only 42 cases of early secondary neurological deterioration after blunt trauma in all of the indexed medical literature. They noted, “In twelve cases the authors did attribute deterioration to temporally associated precipitants, seven of which were possibly iatrogenic; these included removal of a cervical collar, placement of a halo device, patient agitation, performance of flexion/extension films, ‘unintentional manipulation,’ falling in or near the ED, and forced collar application in patients with ankylosing spondylitis.” Thirteen of these cases occurred during prehospital care, none of them sudden and movement-provoked, and all reported by a single study.” This review highlights the rarity of secondary deterioration.

When Should Immobilization Be Used?

So what’s the next step for spinal immobilization in the field? How do we appropriately protect trauma patients during transport? As always seems to be the case in medicine, more evidence is needed. Oteir et al⁴¹ recently published a review of new literature on the epidemiology and current practice of prehospital spine management. They reported that early (8-24 hours) transfer of patients with spinal injury to spinal care units, along with effective resuscitation, was the most important determinant of better neurological outcomes.⁴¹ This review reaffirms the need for more data evaluating the relationship between spinal immobilization and neurological outcomes.

Currently, recommendations call for selective spinal immobilization to decrease unnecessary application and potential harm. Use of backboards for spinal immobilization should be limited to the following types of patients:^1,20

• Blunt trauma and altered level of consciousness;
• Spinal pain or tenderness;
• Neurological complaint (eg, numbness or motor weakness);
• Anatomic deformity of the spine;
• High-energy mechanism of injury and:
  
  • Drug or alcohol intoxication;
  • Inability to communicate; and/or
  • Distracting injury.

Patients for whom immobilization on a backboard is not necessary include those with all of the following:

• Normal level of consciousness (GCS 15);
• No spine tenderness or anatomic abnormality;
• No neurological findings or complaints;
• No distracting injury;
• No intoxication.

Cervical collars alone are still recommended for use in patients who do not meet validated clinical rules, such as the NEXUS or Canadian C spine rules.^1,20,42,43 As these rules are well validated, they can be safely used to determine who should have a cervical collar placed, with or without a backboard. In a retrospective review, selective spinal immobilization was found to be 99% sensitive in identifying patients with cervical injuries.⁴⁴

Clearly, there is still work to be done. Due to the relative rarity of actual spinal cord injury with the conse­quences of neurological injury, prospective trials in this area are rare and very difficult to safely design. However, there is growing confidence that selective spinal protocols, together with the inclusion of validated clinical rules, can effectively limit exposure to unnecessary spinal immobilization. As the current evidence continues to mount for the potential harm in indiscriminate backboard and cervical collar use, it seems clear we should strive to decrease the overuse of prehospital and early spinal immobilization consistent with current position statements and validated clinical rules.

IN THIS ARTICLE

• Diagnosis
• Management
• Newborn hip evaluation algorithm

Developmental dysplasia of the hip (DDH), previously known as congenital dislocation of the hip, follows a spectrum of irregular anatomic hip development spanning from acetabular dysplasia to irreducible dislocation at birth. Early detection is critical to improve the overall prognosis. Prompt diagnosis requires understanding of potential risk factors, proficiency in physical examination techniques, and implementation of appropriate screening tools when indicated. Although current guidelines direct timing for physical exam screenings, imaging, and treatment, it is ultimately up to the provider to determine the best course of action on a case-by-case basis. This article provides a review of these topics and more.

CURRENT GUIDELINES

In 2000, the American Academy of Pediatrics (AAP) developed guidelines for detection of hip dysplasia, including recommendation of relevant physical exam screenings for all newborns.¹ In 2007, the Pediatric Orthopaedic Society of North America (POSNA) encouraged providers to follow the AAP guidelines with a continued recommendation to perform newborn screening for hip instability and routine follow-up evaluations until the child achieves walking.² The American Academy of Orthopaedic Surgeons (AAOS) also established clinical guidelines in 2014 that are endorsed by both AAP and POSNA.³ These guidelines support routine clinical screening; research evaluated infants up to 6 months old, however, limiting the recommendations to that age-group.

Failure to treat DDH early has been associated with serious negative sequelae that include chronic pain, degenerative arthritis, postural scoliosis, and early gait disturbances.⁴ Primary care providers are expected to perform thorough newborn hip exams with associated specialized tests (ie, Ortolani and Barlow, which are discussed in “Physical exam”) at each routine follow-up. Heightened clinical suspicion and risk factor awareness are key for primary care providers to promptly identify patients requiring orthopedic referral. With early diagnosis, a removable soft abduction brace can be applied as the initial treatment. When treatment is delayed, however, closed reduction under anesthesia or complex surgical intervention may be required.

EPIDEMIOLOGY

The etiology for DDH remains unknown. Hip dysplasia typically presents unilaterally but can also occur bilaterally. DDH is more likely to affect the left hip than the right.⁵

Reported incidence varies, ranging from 0.06 to 76.1 per 1,000 live births, and is largely affected by race and geographic location.⁵ Incidence is higher in countries where routine screening is required, by either physical examination or ultrasound (1.6 to 28.5 and 34.0 to 60.3 per 1,000, respectively), compared with countries not requiring routine screening (1.3 per 1,000). This may suggest that the majority of hip dysplasia cases are transient and resolve spontaneously without treatment.^6,7

RISK FACTORS AND PATIENT HISTORY

Known risk factors for DDH include breech presentation (see Figure 1), positive family history, and female gender.^5,8-10 Female infants are eight times more likely than males to develop DDH.¹⁰ Firstborn status is also recognized as an associated risk factor, which may be attributable to space constraints in utero. This hypothesis is further supported by the relative DDH-protective effect of prematurity and low birth weight. Other potential risk factors include advanced maternal age, birth weight that is high for gestational age, decreased hip abduction, and joint laxity. However, the majority of patients with hip dysplasia have no identifiable risk ­factors.^3,5,9,11,12

Swaddling, which often maintains the hips in an adducted and/or extended position, has also been strongly associated with hip dysplasia.^5,13 Multiple organizations, including the AAOS,AAP, POSNA, and the International Hip Dysplasia Institute, have developed or endorsed hip-healthy swaddling recommendations to minimize the risk for DDH in swaddled infants.^13-15 Such practices allow the infant’s legs to bend up and out at the hips, promoting free hip movement, flexion, and abduction.^13,15 Swaddling has demonstrated multiple benefits (including improved sleep and relief of excessive crying¹³) and continues to be recommended by many US providers; however, those caring for infants at risk for DDH should avoid traditional swaddling and/or practice hip-healthy swaddling techniques.^10,13,14 Early diagnosis starts with the clinician’s knowledge of DDH risk factors and the recommended screening protocols. The presence of multiple risk factors will increase the likelihood of this condition and should lower the clinician’s threshold for ordering additional screening, regardless of hip exam findings.

PHYSICAL EXAM

Both AAP and AAOS guidelines recommend clinical screening for DDH with physical exam in all newborns.^1,3 A head-to-toe musculoskeletal exam is warranted during the initial evaluation of every newborn in order to assess for any known DDH-associated conditions, which may include neuromuscular disorders, torticollis, and metatarsus adductus.⁵

Initial evaluation of an infant with DDH may reveal nonspecific findings, including asymmetric skin folds and limb-length inequality. The Galeazzi sign should be sought by aligning flexed knees with the child in the supine position and assessing for uneven knee heights (see Figure 2). Unilateral posterior hip dislocation or femoral shortening represents a positive Galeazzi sign.¹⁶ Joint laxity and limited hip abduction have also been associated with DDH.^1,10

Barlow and Ortolani exams are more specific to DDH and should be completed at newborn screening and each subsequent well-baby exam.¹ The Barlow maneuver is a provocative test with flexion, adduction, and posterior pressure through the infant’s hip (Figure 3). A palpable clunk during the Barlow maneuver indicates positive instability with posterior displacement. The Ortolani test is a reductive maneuver requiring abduction with posterior pressure to lift the greater trochanter (Figure 4). A clunk sensation with this test is positive for reduction of the hip.

The infant’s diaper should be removed during the hip evaluation. These exams are more reliable when each hip is evaluated separately with the pelvis stabilized.¹⁰ All physical exam findings must be carefully documented at each encounter.^1,17

It is critical for the examiner to understand the appropriate technique and potential results when conducting each of these specialized hip exams. A true positive finding is the clunking sensation that occurs with the dislocation or relocation of the affected hip; this finding is better felt than heard. In contrast, a benign hip click with these maneuvers is a more subtle sensation—typically, a soft-tissue snapping or catching—and is not diagnostic of DDH. A click is not a clunk and is not indicative of DDH.^1,3

DDH may present later in infancy or early childhood; therefore, DDH should remain within the differential diagnosis for gait asymmetry, unequal hip motion, or limb-length discrepancy. It may be beneficial to continue to evaluate for these developments during routine exams as part of a thorough pediatric musculoskeletal assessment, particularly in patients with documented risk factors for DDH.^1,3,4 Delay in diagnosis of DDH, it should be noted, is a relatively common complaint in pediatric medical malpractice lawsuits; until the early 2000s, this condition represented about 75% of claims in one medical malpractice database.The decrease in claims has been attributed to better awareness and earlier diagnosis of DDH. ¹⁷

Continue for the diagnosis >>

 

DIAGNOSIS

A positive Ortolani or Barlow sign is diagnostic and warrants prompt orthopedic referral (Figure 5). If physical examination results are equivocal or inconclusive, follow-up at two weeks is recommended, with continued routine follow-up until walking is achieved. Patients with persistent equivocal findings at the two-week follow-up warrant ultrasound at age 3 to 4 weeks or orthopedic referral. Infants with significant risk factors, particularly breech presentation at birth, should also undergo imaging.¹⁸ AAP recommends ultrasound at age 6 weeks or radiograph after 4 months of age.^1,18 AAOS recommends performing an imaging study before age 6 months when at least one of the following risk factors is present: breech presentation, positive family history of DDH, or previous clinical instability (moderate level of evidence).³

IMAGING

Ultrasound is the diagnostic test of choice for infants because radiographs have limited value until the femoral heads begin to ossify at age 4 to 6 months.¹⁸ Ultrasonography allows for visualization of the cartilaginous portion of the acetabulum and femoral head.¹ Dynamic stressing is performed during ultrasound to assess the level of hip stability. A provider trained in ultrasound will measure the depth of the acetabulum and identify any potential laxity or instability of the hip joint. Accuracy of these findings is largely dependent on the experience and skill of the examiner.

Ultrasound evaluation is not recommended until after age 3 to 4 weeks. Earlier findings may include mild laxity and immature morphology of the acetabulum, which often resolve spontaneously.^1,18 Use of ultrasound is currently recommended only to confirm diagnostic suspicion, based on clinical findings, or for infants with significant risk factors.¹⁸ Universal ultrasound screening in newborns is not recommended and would incur unnecessary costs.^1,3,9 Plain radiographs are used after age 4 months to confirm a diagnosis of DDH or to assess for residual dysplasia.^3,18

Continue for management >>

 

MANAGEMENT

Once hip dysplasia is suggested by physical exam or imaging study, the child’s subsequent care should be provided by an orthopedic specialist with experience in treating this condition. Treatment is preferably initiated before age 6 weeks.¹² The specifics of treatment are largely based on age at diagnosis and the severity of dysplasia.

The goal of treatment is to maintain the hips in a stable position with the femoral head well covered by the acetabulum. This will improve anatomic development and function. Early clinical diagnosis is often sufficient to justify initiating conservative treatment; additionally, early detection of DDH can considerably reduce the need for surgical intervention.¹² Although the potential for spontaneous resolution is high, the consequences associated with delay in care can be significant.

Preferred initial management, which can be initiated before confirmation of DDH by ultrasound, involves implementation of soft abduction support.¹⁹ The Pavlik harness is the support design of choice (Figure 6).¹² This harness maintains hip flexion and abduction, creating concentric reduction of the femoral head. The brace is highly successful when its use is initiated early. Treatment in a Pavlik harness requires nearly full-time wear and close monitoring by a clinician. Unlikely potential risks associated with this treatment include avascular necrosis and femoral nerve palsy.⁴

Ultrasonography is used to further monitor treatment and to determine length of wear. Long-term results suggest a success rate exceeding 90%.^20,21 However, this rate may be falsely elevated due to the number of hips that likely would have improved spontaneously without treatment.^6,19

The Pavlik harness becomes less effective with increasing age, and a more rigid abduction brace may be considered in infants older than 6 months.²⁰ Overall outcomes improve once the femoral head is consistently maintained in the acetabulum. Delay in treatment is associated with an increase in the long-term complications associated with residual hip dysplasia.²²

Once an infant is undergoing treatment for DDH in a Pavlik harness, there is no need for primary care providers to continue to perform provocative testing, such as the Ortolani or Barlow test, at routine well-baby checks. Unnecessary stress to the hips is not beneficial, and any new results will not change the treatment being provided by the orthopedic specialist. Adjustments to the fit of the harness should be made only by the orthopedist, unless femoral nerve palsy is noted on exam. This development warrants immediate discontinuation of harness use until symptoms resolve.²¹

Abduction bracing may not be suitable for all cases of hip dysplasia. Newborns with irreducible hips, more advanced dysplasia, or associated neuromuscular or syndromic disorder may require closed versus open reduction and casting. More invasive surgical options may also be considered in advanced dysplasia in order to reshape the joint and improve function.^20,22

Continue for patient education >>

 

PATIENT EDUCATION

Parents should be fully educated on the options for managing hip dysplasia. Once DDH is diagnosed, prompt referral to an orthopedic specialist is critical in order to weigh the treatment options and to develop the appropriate individualized plan for each child. Once treatment is initiated, parental compliance is essential; frequent meetings between parents and the specialist are important.

Parents of infants with known risk factors for and/or suspicion of hip dysplasia should also be educated on hip-healthy swaddling to allow for free motion of the hips and knees.^10,13 Advise them that some commercial baby carriers and slings may maintain the hips in an undesirable extended position. In both swaddling and with baby carriers, care should be taken to allow for hip abduction and flexion. Caution should also be taken during diaper changes to avoid lifting the legs and thereby causing unnecessary stress to the hips.

CONCLUSION

Developmental dysplasia of the hip can be a disabling pediatric condition. Early diagnosis improves the likelihood of successful treatment during infancy and can prevent serious complications. If untreated, DDH can lead to joint degeneration and premature arthritis. Recognition and treatment within the first six weeks of life is crucial to the overall outcome.

The role of a primary care provider is to identify hip dysplasia risk factors and recognize associated physical exam findings in order to refer to an orthopedic specialist in a timely manner. Guidelines from the AAP, POSNA, and AAOS help direct this process in order to effectively identify infants at risk and in need of treatment.

REFERENCES
1. American Academy of Pediatrics. Committee on Quality Improvement, Subcommittee on Developmental Dysplasia of the Hip. Clinical practice guideline: early detection of developmental dysplasia of the hip. Pediatrics. 2000;105(4 pt 1):896-905.
2. Schwend RM, Schoenecker P, Richards BS, et al. Screening the newborn for developmental dysplasia of the hip: now what do we do? J Pediatr Orthop. 2007;27(6):607-610.
3. Mulpuri K, Song KM, Goldberg MJ, Sevarino K. Detection and nonoperative management of pediatric developmental dysplasia of the hip in infants up to six months of age. J Am Acad Orthop Surg. 2015;23(3):202-205.
4. Thomas SRYW. A review of long-term outcomes for late presenting developmental hip dysplasia. Bone Joint J. 2015;97-B(6):729-733.
5. Loder RT, Skopelja EN. The epidemiology and demographics of hip dysplasia. ISRN Orthop. 2011;2011:238607.
6. US Preventive Services Task Force. Screening for developmental dysplasia of the hip: recommendation statement. Pediatrics. 2006;117(3):898-902.
7. Shorter D, Hong T, Osborn DA. Screening programmes for developmental dysplasia of the hip in newborn infants. Cochrane Database Syst Rev. 2011;(9):CD004595.
8. Loder RT, Shafer C. The demographics of developmental hip dysplasia in the Midwestern United States (Indiana). J Child Orthop. 2015;9(1):93-98.
9. Paton RW, Hinduja K, Thomas CD. The significance of at-risk factors in ultrasound surveillance of developmental dysplasia of the hip: a ten-year prospective study. J Bone Joint Surg Br. 2005;87(9):1264-1266.
10. Alsaleem M, Set KK, Saadeh L. Developmental dysplasia of hip: a review. Clin Pediatr (Phila). 2015;54(10):921-928.
11. Chan A, McCaul KA, Cundy PJ, et al. Perinatal risk factors for developmental dysplasia of the hip. Arch Dis Child. 1997;76(2):F94-F100.
12. Godley DR. Assessment, diagnosis, and treatment of developmental dysplasia of the hip. JAAPA. 2013;26(3):54-58.
13. Van Sleuwen BE, Engelberts AC, Boere-Boonekamp MM, et al. Swaddling: a systematic review. Pediatrics. 2007;120(4):e1097-e1106.
14. American Academy of Orthopaedic Surgeons, American Association of Orthopaedic Surgeons. Position statement: swaddling and developmental hip dysplasia. www.aaos.org/uploadedFiles/PreProduction/About/Opinion_Statements/position/1186%20Swaddling%20and%20Developmental%20Hip%20Dysplasia.pdf. Accessed January 22, 2016.
15. Clarke NM. Swaddling and hip dysplasia: an orthopaedic perspective. Arch Dis Child. 2014;99(1):5-6.
16. Storer SK, Skaggs DL. Developmental dysplasia of the hip. Am Fam Physician. 2006;74(8):1310-1316.
17. McAbee GN, Donn SM, Mendelson RA, et al. Medical diagnoses commonly associated with pediatric malpractice lawsuits in the United States. Pediatrics. 2008;122(6):e1282-e1286.
18. Imrie M, Scott V, Stearns P, et al. Is ultrasound screening for DDH in babies born breech sufficient? J Child Orthop. 2010;4(1):3-8.
19. Chen HW, Chang CH, Tsai ST, et al. Natural progression of hip dysplasia in newborns: a reflection of hip ultrasonographic screenings in newborn nurseries. J Pediatr Orthop B. 2010;19(5):418-423.
20. Gans I, Flynn JM, Sankar WN. Abduction bracing for residual acetabular dysplasia in infantile DDH. J Pediatr Orthop. 2013;33(7):714-718.
21. Murnaghan ML, Browne RH, Sucato DJ, Birch J. Femoral nerve palsy in Pavlik harness treatment for developmental dysplasia of the hip. J Bone Joint Surg Am. 2011;93(5):493-499.
22. Dezateux C, Rosendahl K. Developmental dysplasia of the hip. Lancet. 2007;369(9572):1541-1552.

IN THIS ARTICLE

• Diagnosis
• Management
• Newborn hip evaluation algorithm

Developmental dysplasia of the hip (DDH), previously known as congenital dislocation of the hip, follows a spectrum of irregular anatomic hip development spanning from acetabular dysplasia to irreducible dislocation at birth. Early detection is critical to improve the overall prognosis. Prompt diagnosis requires understanding of potential risk factors, proficiency in physical examination techniques, and implementation of appropriate screening tools when indicated. Although current guidelines direct timing for physical exam screenings, imaging, and treatment, it is ultimately up to the provider to determine the best course of action on a case-by-case basis. This article provides a review of these topics and more.

CURRENT GUIDELINES

In 2000, the American Academy of Pediatrics (AAP) developed guidelines for detection of hip dysplasia, including recommendation of relevant physical exam screenings for all newborns.¹ In 2007, the Pediatric Orthopaedic Society of North America (POSNA) encouraged providers to follow the AAP guidelines with a continued recommendation to perform newborn screening for hip instability and routine follow-up evaluations until the child achieves walking.² The American Academy of Orthopaedic Surgeons (AAOS) also established clinical guidelines in 2014 that are endorsed by both AAP and POSNA.³ These guidelines support routine clinical screening; research evaluated infants up to 6 months old, however, limiting the recommendations to that age-group.

Failure to treat DDH early has been associated with serious negative sequelae that include chronic pain, degenerative arthritis, postural scoliosis, and early gait disturbances.⁴ Primary care providers are expected to perform thorough newborn hip exams with associated specialized tests (ie, Ortolani and Barlow, which are discussed in “Physical exam”) at each routine follow-up. Heightened clinical suspicion and risk factor awareness are key for primary care providers to promptly identify patients requiring orthopedic referral. With early diagnosis, a removable soft abduction brace can be applied as the initial treatment. When treatment is delayed, however, closed reduction under anesthesia or complex surgical intervention may be required.

EPIDEMIOLOGY

The etiology for DDH remains unknown. Hip dysplasia typically presents unilaterally but can also occur bilaterally. DDH is more likely to affect the left hip than the right.⁵

Reported incidence varies, ranging from 0.06 to 76.1 per 1,000 live births, and is largely affected by race and geographic location.⁵ Incidence is higher in countries where routine screening is required, by either physical examination or ultrasound (1.6 to 28.5 and 34.0 to 60.3 per 1,000, respectively), compared with countries not requiring routine screening (1.3 per 1,000). This may suggest that the majority of hip dysplasia cases are transient and resolve spontaneously without treatment.^6,7

RISK FACTORS AND PATIENT HISTORY

Known risk factors for DDH include breech presentation (see Figure 1), positive family history, and female gender.^5,8-10 Female infants are eight times more likely than males to develop DDH.¹⁰ Firstborn status is also recognized as an associated risk factor, which may be attributable to space constraints in utero. This hypothesis is further supported by the relative DDH-protective effect of prematurity and low birth weight. Other potential risk factors include advanced maternal age, birth weight that is high for gestational age, decreased hip abduction, and joint laxity. However, the majority of patients with hip dysplasia have no identifiable risk ­factors.^3,5,9,11,12

Swaddling, which often maintains the hips in an adducted and/or extended position, has also been strongly associated with hip dysplasia.^5,13 Multiple organizations, including the AAOS,AAP, POSNA, and the International Hip Dysplasia Institute, have developed or endorsed hip-healthy swaddling recommendations to minimize the risk for DDH in swaddled infants.^13-15 Such practices allow the infant’s legs to bend up and out at the hips, promoting free hip movement, flexion, and abduction.^13,15 Swaddling has demonstrated multiple benefits (including improved sleep and relief of excessive crying¹³) and continues to be recommended by many US providers; however, those caring for infants at risk for DDH should avoid traditional swaddling and/or practice hip-healthy swaddling techniques.^10,13,14 Early diagnosis starts with the clinician’s knowledge of DDH risk factors and the recommended screening protocols. The presence of multiple risk factors will increase the likelihood of this condition and should lower the clinician’s threshold for ordering additional screening, regardless of hip exam findings.

PHYSICAL EXAM

Both AAP and AAOS guidelines recommend clinical screening for DDH with physical exam in all newborns.^1,3 A head-to-toe musculoskeletal exam is warranted during the initial evaluation of every newborn in order to assess for any known DDH-associated conditions, which may include neuromuscular disorders, torticollis, and metatarsus adductus.⁵

Initial evaluation of an infant with DDH may reveal nonspecific findings, including asymmetric skin folds and limb-length inequality. The Galeazzi sign should be sought by aligning flexed knees with the child in the supine position and assessing for uneven knee heights (see Figure 2). Unilateral posterior hip dislocation or femoral shortening represents a positive Galeazzi sign.¹⁶ Joint laxity and limited hip abduction have also been associated with DDH.^1,10

Barlow and Ortolani exams are more specific to DDH and should be completed at newborn screening and each subsequent well-baby exam.¹ The Barlow maneuver is a provocative test with flexion, adduction, and posterior pressure through the infant’s hip (Figure 3). A palpable clunk during the Barlow maneuver indicates positive instability with posterior displacement. The Ortolani test is a reductive maneuver requiring abduction with posterior pressure to lift the greater trochanter (Figure 4). A clunk sensation with this test is positive for reduction of the hip.

The infant’s diaper should be removed during the hip evaluation. These exams are more reliable when each hip is evaluated separately with the pelvis stabilized.¹⁰ All physical exam findings must be carefully documented at each encounter.^1,17

It is critical for the examiner to understand the appropriate technique and potential results when conducting each of these specialized hip exams. A true positive finding is the clunking sensation that occurs with the dislocation or relocation of the affected hip; this finding is better felt than heard. In contrast, a benign hip click with these maneuvers is a more subtle sensation—typically, a soft-tissue snapping or catching—and is not diagnostic of DDH. A click is not a clunk and is not indicative of DDH.^1,3

DDH may present later in infancy or early childhood; therefore, DDH should remain within the differential diagnosis for gait asymmetry, unequal hip motion, or limb-length discrepancy. It may be beneficial to continue to evaluate for these developments during routine exams as part of a thorough pediatric musculoskeletal assessment, particularly in patients with documented risk factors for DDH.^1,3,4 Delay in diagnosis of DDH, it should be noted, is a relatively common complaint in pediatric medical malpractice lawsuits; until the early 2000s, this condition represented about 75% of claims in one medical malpractice database.The decrease in claims has been attributed to better awareness and earlier diagnosis of DDH. ¹⁷

Continue for the diagnosis >>

 

DIAGNOSIS

A positive Ortolani or Barlow sign is diagnostic and warrants prompt orthopedic referral (Figure 5). If physical examination results are equivocal or inconclusive, follow-up at two weeks is recommended, with continued routine follow-up until walking is achieved. Patients with persistent equivocal findings at the two-week follow-up warrant ultrasound at age 3 to 4 weeks or orthopedic referral. Infants with significant risk factors, particularly breech presentation at birth, should also undergo imaging.¹⁸ AAP recommends ultrasound at age 6 weeks or radiograph after 4 months of age.^1,18 AAOS recommends performing an imaging study before age 6 months when at least one of the following risk factors is present: breech presentation, positive family history of DDH, or previous clinical instability (moderate level of evidence).³

IMAGING

Ultrasound is the diagnostic test of choice for infants because radiographs have limited value until the femoral heads begin to ossify at age 4 to 6 months.¹⁸ Ultrasonography allows for visualization of the cartilaginous portion of the acetabulum and femoral head.¹ Dynamic stressing is performed during ultrasound to assess the level of hip stability. A provider trained in ultrasound will measure the depth of the acetabulum and identify any potential laxity or instability of the hip joint. Accuracy of these findings is largely dependent on the experience and skill of the examiner.

Ultrasound evaluation is not recommended until after age 3 to 4 weeks. Earlier findings may include mild laxity and immature morphology of the acetabulum, which often resolve spontaneously.^1,18 Use of ultrasound is currently recommended only to confirm diagnostic suspicion, based on clinical findings, or for infants with significant risk factors.¹⁸ Universal ultrasound screening in newborns is not recommended and would incur unnecessary costs.^1,3,9 Plain radiographs are used after age 4 months to confirm a diagnosis of DDH or to assess for residual dysplasia.^3,18

Continue for management >>

 

MANAGEMENT

Once hip dysplasia is suggested by physical exam or imaging study, the child’s subsequent care should be provided by an orthopedic specialist with experience in treating this condition. Treatment is preferably initiated before age 6 weeks.¹² The specifics of treatment are largely based on age at diagnosis and the severity of dysplasia.

The goal of treatment is to maintain the hips in a stable position with the femoral head well covered by the acetabulum. This will improve anatomic development and function. Early clinical diagnosis is often sufficient to justify initiating conservative treatment; additionally, early detection of DDH can considerably reduce the need for surgical intervention.¹² Although the potential for spontaneous resolution is high, the consequences associated with delay in care can be significant.

Preferred initial management, which can be initiated before confirmation of DDH by ultrasound, involves implementation of soft abduction support.¹⁹ The Pavlik harness is the support design of choice (Figure 6).¹² This harness maintains hip flexion and abduction, creating concentric reduction of the femoral head. The brace is highly successful when its use is initiated early. Treatment in a Pavlik harness requires nearly full-time wear and close monitoring by a clinician. Unlikely potential risks associated with this treatment include avascular necrosis and femoral nerve palsy.⁴

Ultrasonography is used to further monitor treatment and to determine length of wear. Long-term results suggest a success rate exceeding 90%.^20,21 However, this rate may be falsely elevated due to the number of hips that likely would have improved spontaneously without treatment.^6,19

The Pavlik harness becomes less effective with increasing age, and a more rigid abduction brace may be considered in infants older than 6 months.²⁰ Overall outcomes improve once the femoral head is consistently maintained in the acetabulum. Delay in treatment is associated with an increase in the long-term complications associated with residual hip dysplasia.²²

Once an infant is undergoing treatment for DDH in a Pavlik harness, there is no need for primary care providers to continue to perform provocative testing, such as the Ortolani or Barlow test, at routine well-baby checks. Unnecessary stress to the hips is not beneficial, and any new results will not change the treatment being provided by the orthopedic specialist. Adjustments to the fit of the harness should be made only by the orthopedist, unless femoral nerve palsy is noted on exam. This development warrants immediate discontinuation of harness use until symptoms resolve.²¹

Abduction bracing may not be suitable for all cases of hip dysplasia. Newborns with irreducible hips, more advanced dysplasia, or associated neuromuscular or syndromic disorder may require closed versus open reduction and casting. More invasive surgical options may also be considered in advanced dysplasia in order to reshape the joint and improve function.^20,22

Continue for patient education >>

 

PATIENT EDUCATION

Parents should be fully educated on the options for managing hip dysplasia. Once DDH is diagnosed, prompt referral to an orthopedic specialist is critical in order to weigh the treatment options and to develop the appropriate individualized plan for each child. Once treatment is initiated, parental compliance is essential; frequent meetings between parents and the specialist are important.

Parents of infants with known risk factors for and/or suspicion of hip dysplasia should also be educated on hip-healthy swaddling to allow for free motion of the hips and knees.^10,13 Advise them that some commercial baby carriers and slings may maintain the hips in an undesirable extended position. In both swaddling and with baby carriers, care should be taken to allow for hip abduction and flexion. Caution should also be taken during diaper changes to avoid lifting the legs and thereby causing unnecessary stress to the hips.

CONCLUSION

Developmental dysplasia of the hip can be a disabling pediatric condition. Early diagnosis improves the likelihood of successful treatment during infancy and can prevent serious complications. If untreated, DDH can lead to joint degeneration and premature arthritis. Recognition and treatment within the first six weeks of life is crucial to the overall outcome.

The role of a primary care provider is to identify hip dysplasia risk factors and recognize associated physical exam findings in order to refer to an orthopedic specialist in a timely manner. Guidelines from the AAP, POSNA, and AAOS help direct this process in order to effectively identify infants at risk and in need of treatment.

REFERENCES
1. American Academy of Pediatrics. Committee on Quality Improvement, Subcommittee on Developmental Dysplasia of the Hip. Clinical practice guideline: early detection of developmental dysplasia of the hip. Pediatrics. 2000;105(4 pt 1):896-905.
2. Schwend RM, Schoenecker P, Richards BS, et al. Screening the newborn for developmental dysplasia of the hip: now what do we do? J Pediatr Orthop. 2007;27(6):607-610.
3. Mulpuri K, Song KM, Goldberg MJ, Sevarino K. Detection and nonoperative management of pediatric developmental dysplasia of the hip in infants up to six months of age. J Am Acad Orthop Surg. 2015;23(3):202-205.
4. Thomas SRYW. A review of long-term outcomes for late presenting developmental hip dysplasia. Bone Joint J. 2015;97-B(6):729-733.
5. Loder RT, Skopelja EN. The epidemiology and demographics of hip dysplasia. ISRN Orthop. 2011;2011:238607.
6. US Preventive Services Task Force. Screening for developmental dysplasia of the hip: recommendation statement. Pediatrics. 2006;117(3):898-902.
7. Shorter D, Hong T, Osborn DA. Screening programmes for developmental dysplasia of the hip in newborn infants. Cochrane Database Syst Rev. 2011;(9):CD004595.
8. Loder RT, Shafer C. The demographics of developmental hip dysplasia in the Midwestern United States (Indiana). J Child Orthop. 2015;9(1):93-98.
9. Paton RW, Hinduja K, Thomas CD. The significance of at-risk factors in ultrasound surveillance of developmental dysplasia of the hip: a ten-year prospective study. J Bone Joint Surg Br. 2005;87(9):1264-1266.
10. Alsaleem M, Set KK, Saadeh L. Developmental dysplasia of hip: a review. Clin Pediatr (Phila). 2015;54(10):921-928.
11. Chan A, McCaul KA, Cundy PJ, et al. Perinatal risk factors for developmental dysplasia of the hip. Arch Dis Child. 1997;76(2):F94-F100.
12. Godley DR. Assessment, diagnosis, and treatment of developmental dysplasia of the hip. JAAPA. 2013;26(3):54-58.
13. Van Sleuwen BE, Engelberts AC, Boere-Boonekamp MM, et al. Swaddling: a systematic review. Pediatrics. 2007;120(4):e1097-e1106.
14. American Academy of Orthopaedic Surgeons, American Association of Orthopaedic Surgeons. Position statement: swaddling and developmental hip dysplasia. www.aaos.org/uploadedFiles/PreProduction/About/Opinion_Statements/position/1186%20Swaddling%20and%20Developmental%20Hip%20Dysplasia.pdf. Accessed January 22, 2016.
15. Clarke NM. Swaddling and hip dysplasia: an orthopaedic perspective. Arch Dis Child. 2014;99(1):5-6.
16. Storer SK, Skaggs DL. Developmental dysplasia of the hip. Am Fam Physician. 2006;74(8):1310-1316.
17. McAbee GN, Donn SM, Mendelson RA, et al. Medical diagnoses commonly associated with pediatric malpractice lawsuits in the United States. Pediatrics. 2008;122(6):e1282-e1286.
18. Imrie M, Scott V, Stearns P, et al. Is ultrasound screening for DDH in babies born breech sufficient? J Child Orthop. 2010;4(1):3-8.
19. Chen HW, Chang CH, Tsai ST, et al. Natural progression of hip dysplasia in newborns: a reflection of hip ultrasonographic screenings in newborn nurseries. J Pediatr Orthop B. 2010;19(5):418-423.
20. Gans I, Flynn JM, Sankar WN. Abduction bracing for residual acetabular dysplasia in infantile DDH. J Pediatr Orthop. 2013;33(7):714-718.
21. Murnaghan ML, Browne RH, Sucato DJ, Birch J. Femoral nerve palsy in Pavlik harness treatment for developmental dysplasia of the hip. J Bone Joint Surg Am. 2011;93(5):493-499.
22. Dezateux C, Rosendahl K. Developmental dysplasia of the hip. Lancet. 2007;369(9572):1541-1552.

IN THIS ARTICLE

• Diagnosis
• Management
• Newborn hip evaluation algorithm

Developmental dysplasia of the hip (DDH), previously known as congenital dislocation of the hip, follows a spectrum of irregular anatomic hip development spanning from acetabular dysplasia to irreducible dislocation at birth. Early detection is critical to improve the overall prognosis. Prompt diagnosis requires understanding of potential risk factors, proficiency in physical examination techniques, and implementation of appropriate screening tools when indicated. Although current guidelines direct timing for physical exam screenings, imaging, and treatment, it is ultimately up to the provider to determine the best course of action on a case-by-case basis. This article provides a review of these topics and more.

CURRENT GUIDELINES

In 2000, the American Academy of Pediatrics (AAP) developed guidelines for detection of hip dysplasia, including recommendation of relevant physical exam screenings for all newborns.¹ In 2007, the Pediatric Orthopaedic Society of North America (POSNA) encouraged providers to follow the AAP guidelines with a continued recommendation to perform newborn screening for hip instability and routine follow-up evaluations until the child achieves walking.² The American Academy of Orthopaedic Surgeons (AAOS) also established clinical guidelines in 2014 that are endorsed by both AAP and POSNA.³ These guidelines support routine clinical screening; research evaluated infants up to 6 months old, however, limiting the recommendations to that age-group.

Failure to treat DDH early has been associated with serious negative sequelae that include chronic pain, degenerative arthritis, postural scoliosis, and early gait disturbances.⁴ Primary care providers are expected to perform thorough newborn hip exams with associated specialized tests (ie, Ortolani and Barlow, which are discussed in “Physical exam”) at each routine follow-up. Heightened clinical suspicion and risk factor awareness are key for primary care providers to promptly identify patients requiring orthopedic referral. With early diagnosis, a removable soft abduction brace can be applied as the initial treatment. When treatment is delayed, however, closed reduction under anesthesia or complex surgical intervention may be required.

EPIDEMIOLOGY

The etiology for DDH remains unknown. Hip dysplasia typically presents unilaterally but can also occur bilaterally. DDH is more likely to affect the left hip than the right.⁵

Reported incidence varies, ranging from 0.06 to 76.1 per 1,000 live births, and is largely affected by race and geographic location.⁵ Incidence is higher in countries where routine screening is required, by either physical examination or ultrasound (1.6 to 28.5 and 34.0 to 60.3 per 1,000, respectively), compared with countries not requiring routine screening (1.3 per 1,000). This may suggest that the majority of hip dysplasia cases are transient and resolve spontaneously without treatment.^6,7

RISK FACTORS AND PATIENT HISTORY

Known risk factors for DDH include breech presentation (see Figure 1), positive family history, and female gender.^5,8-10 Female infants are eight times more likely than males to develop DDH.¹⁰ Firstborn status is also recognized as an associated risk factor, which may be attributable to space constraints in utero. This hypothesis is further supported by the relative DDH-protective effect of prematurity and low birth weight. Other potential risk factors include advanced maternal age, birth weight that is high for gestational age, decreased hip abduction, and joint laxity. However, the majority of patients with hip dysplasia have no identifiable risk ­factors.^3,5,9,11,12

Swaddling, which often maintains the hips in an adducted and/or extended position, has also been strongly associated with hip dysplasia.^5,13 Multiple organizations, including the AAOS,AAP, POSNA, and the International Hip Dysplasia Institute, have developed or endorsed hip-healthy swaddling recommendations to minimize the risk for DDH in swaddled infants.^13-15 Such practices allow the infant’s legs to bend up and out at the hips, promoting free hip movement, flexion, and abduction.^13,15 Swaddling has demonstrated multiple benefits (including improved sleep and relief of excessive crying¹³) and continues to be recommended by many US providers; however, those caring for infants at risk for DDH should avoid traditional swaddling and/or practice hip-healthy swaddling techniques.^10,13,14 Early diagnosis starts with the clinician’s knowledge of DDH risk factors and the recommended screening protocols. The presence of multiple risk factors will increase the likelihood of this condition and should lower the clinician’s threshold for ordering additional screening, regardless of hip exam findings.

PHYSICAL EXAM

Both AAP and AAOS guidelines recommend clinical screening for DDH with physical exam in all newborns.^1,3 A head-to-toe musculoskeletal exam is warranted during the initial evaluation of every newborn in order to assess for any known DDH-associated conditions, which may include neuromuscular disorders, torticollis, and metatarsus adductus.⁵

Initial evaluation of an infant with DDH may reveal nonspecific findings, including asymmetric skin folds and limb-length inequality. The Galeazzi sign should be sought by aligning flexed knees with the child in the supine position and assessing for uneven knee heights (see Figure 2). Unilateral posterior hip dislocation or femoral shortening represents a positive Galeazzi sign.¹⁶ Joint laxity and limited hip abduction have also been associated with DDH.^1,10

Barlow and Ortolani exams are more specific to DDH and should be completed at newborn screening and each subsequent well-baby exam.¹ The Barlow maneuver is a provocative test with flexion, adduction, and posterior pressure through the infant’s hip (Figure 3). A palpable clunk during the Barlow maneuver indicates positive instability with posterior displacement. The Ortolani test is a reductive maneuver requiring abduction with posterior pressure to lift the greater trochanter (Figure 4). A clunk sensation with this test is positive for reduction of the hip.

The infant’s diaper should be removed during the hip evaluation. These exams are more reliable when each hip is evaluated separately with the pelvis stabilized.¹⁰ All physical exam findings must be carefully documented at each encounter.^1,17

It is critical for the examiner to understand the appropriate technique and potential results when conducting each of these specialized hip exams. A true positive finding is the clunking sensation that occurs with the dislocation or relocation of the affected hip; this finding is better felt than heard. In contrast, a benign hip click with these maneuvers is a more subtle sensation—typically, a soft-tissue snapping or catching—and is not diagnostic of DDH. A click is not a clunk and is not indicative of DDH.^1,3

DDH may present later in infancy or early childhood; therefore, DDH should remain within the differential diagnosis for gait asymmetry, unequal hip motion, or limb-length discrepancy. It may be beneficial to continue to evaluate for these developments during routine exams as part of a thorough pediatric musculoskeletal assessment, particularly in patients with documented risk factors for DDH.^1,3,4 Delay in diagnosis of DDH, it should be noted, is a relatively common complaint in pediatric medical malpractice lawsuits; until the early 2000s, this condition represented about 75% of claims in one medical malpractice database.The decrease in claims has been attributed to better awareness and earlier diagnosis of DDH. ¹⁷

Continue for the diagnosis >>

 

DIAGNOSIS

A positive Ortolani or Barlow sign is diagnostic and warrants prompt orthopedic referral (Figure 5). If physical examination results are equivocal or inconclusive, follow-up at two weeks is recommended, with continued routine follow-up until walking is achieved. Patients with persistent equivocal findings at the two-week follow-up warrant ultrasound at age 3 to 4 weeks or orthopedic referral. Infants with significant risk factors, particularly breech presentation at birth, should also undergo imaging.¹⁸ AAP recommends ultrasound at age 6 weeks or radiograph after 4 months of age.^1,18 AAOS recommends performing an imaging study before age 6 months when at least one of the following risk factors is present: breech presentation, positive family history of DDH, or previous clinical instability (moderate level of evidence).³

IMAGING

Ultrasound is the diagnostic test of choice for infants because radiographs have limited value until the femoral heads begin to ossify at age 4 to 6 months.¹⁸ Ultrasonography allows for visualization of the cartilaginous portion of the acetabulum and femoral head.¹ Dynamic stressing is performed during ultrasound to assess the level of hip stability. A provider trained in ultrasound will measure the depth of the acetabulum and identify any potential laxity or instability of the hip joint. Accuracy of these findings is largely dependent on the experience and skill of the examiner.

Ultrasound evaluation is not recommended until after age 3 to 4 weeks. Earlier findings may include mild laxity and immature morphology of the acetabulum, which often resolve spontaneously.^1,18 Use of ultrasound is currently recommended only to confirm diagnostic suspicion, based on clinical findings, or for infants with significant risk factors.¹⁸ Universal ultrasound screening in newborns is not recommended and would incur unnecessary costs.^1,3,9 Plain radiographs are used after age 4 months to confirm a diagnosis of DDH or to assess for residual dysplasia.^3,18

Continue for management >>

 

MANAGEMENT

Once hip dysplasia is suggested by physical exam or imaging study, the child’s subsequent care should be provided by an orthopedic specialist with experience in treating this condition. Treatment is preferably initiated before age 6 weeks.¹² The specifics of treatment are largely based on age at diagnosis and the severity of dysplasia.

The goal of treatment is to maintain the hips in a stable position with the femoral head well covered by the acetabulum. This will improve anatomic development and function. Early clinical diagnosis is often sufficient to justify initiating conservative treatment; additionally, early detection of DDH can considerably reduce the need for surgical intervention.¹² Although the potential for spontaneous resolution is high, the consequences associated with delay in care can be significant.

Preferred initial management, which can be initiated before confirmation of DDH by ultrasound, involves implementation of soft abduction support.¹⁹ The Pavlik harness is the support design of choice (Figure 6).¹² This harness maintains hip flexion and abduction, creating concentric reduction of the femoral head. The brace is highly successful when its use is initiated early. Treatment in a Pavlik harness requires nearly full-time wear and close monitoring by a clinician. Unlikely potential risks associated with this treatment include avascular necrosis and femoral nerve palsy.⁴

Ultrasonography is used to further monitor treatment and to determine length of wear. Long-term results suggest a success rate exceeding 90%.^20,21 However, this rate may be falsely elevated due to the number of hips that likely would have improved spontaneously without treatment.^6,19

The Pavlik harness becomes less effective with increasing age, and a more rigid abduction brace may be considered in infants older than 6 months.²⁰ Overall outcomes improve once the femoral head is consistently maintained in the acetabulum. Delay in treatment is associated with an increase in the long-term complications associated with residual hip dysplasia.²²

Once an infant is undergoing treatment for DDH in a Pavlik harness, there is no need for primary care providers to continue to perform provocative testing, such as the Ortolani or Barlow test, at routine well-baby checks. Unnecessary stress to the hips is not beneficial, and any new results will not change the treatment being provided by the orthopedic specialist. Adjustments to the fit of the harness should be made only by the orthopedist, unless femoral nerve palsy is noted on exam. This development warrants immediate discontinuation of harness use until symptoms resolve.²¹

Abduction bracing may not be suitable for all cases of hip dysplasia. Newborns with irreducible hips, more advanced dysplasia, or associated neuromuscular or syndromic disorder may require closed versus open reduction and casting. More invasive surgical options may also be considered in advanced dysplasia in order to reshape the joint and improve function.^20,22

Continue for patient education >>

 

PATIENT EDUCATION

Parents should be fully educated on the options for managing hip dysplasia. Once DDH is diagnosed, prompt referral to an orthopedic specialist is critical in order to weigh the treatment options and to develop the appropriate individualized plan for each child. Once treatment is initiated, parental compliance is essential; frequent meetings between parents and the specialist are important.

Parents of infants with known risk factors for and/or suspicion of hip dysplasia should also be educated on hip-healthy swaddling to allow for free motion of the hips and knees.^10,13 Advise them that some commercial baby carriers and slings may maintain the hips in an undesirable extended position. In both swaddling and with baby carriers, care should be taken to allow for hip abduction and flexion. Caution should also be taken during diaper changes to avoid lifting the legs and thereby causing unnecessary stress to the hips.

CONCLUSION

Developmental dysplasia of the hip can be a disabling pediatric condition. Early diagnosis improves the likelihood of successful treatment during infancy and can prevent serious complications. If untreated, DDH can lead to joint degeneration and premature arthritis. Recognition and treatment within the first six weeks of life is crucial to the overall outcome.

The role of a primary care provider is to identify hip dysplasia risk factors and recognize associated physical exam findings in order to refer to an orthopedic specialist in a timely manner. Guidelines from the AAP, POSNA, and AAOS help direct this process in order to effectively identify infants at risk and in need of treatment.

REFERENCES
1. American Academy of Pediatrics. Committee on Quality Improvement, Subcommittee on Developmental Dysplasia of the Hip. Clinical practice guideline: early detection of developmental dysplasia of the hip. Pediatrics. 2000;105(4 pt 1):896-905.
2. Schwend RM, Schoenecker P, Richards BS, et al. Screening the newborn for developmental dysplasia of the hip: now what do we do? J Pediatr Orthop. 2007;27(6):607-610.
3. Mulpuri K, Song KM, Goldberg MJ, Sevarino K. Detection and nonoperative management of pediatric developmental dysplasia of the hip in infants up to six months of age. J Am Acad Orthop Surg. 2015;23(3):202-205.
4. Thomas SRYW. A review of long-term outcomes for late presenting developmental hip dysplasia. Bone Joint J. 2015;97-B(6):729-733.
5. Loder RT, Skopelja EN. The epidemiology and demographics of hip dysplasia. ISRN Orthop. 2011;2011:238607.
6. US Preventive Services Task Force. Screening for developmental dysplasia of the hip: recommendation statement. Pediatrics. 2006;117(3):898-902.
7. Shorter D, Hong T, Osborn DA. Screening programmes for developmental dysplasia of the hip in newborn infants. Cochrane Database Syst Rev. 2011;(9):CD004595.
8. Loder RT, Shafer C. The demographics of developmental hip dysplasia in the Midwestern United States (Indiana). J Child Orthop. 2015;9(1):93-98.
9. Paton RW, Hinduja K, Thomas CD. The significance of at-risk factors in ultrasound surveillance of developmental dysplasia of the hip: a ten-year prospective study. J Bone Joint Surg Br. 2005;87(9):1264-1266.
10. Alsaleem M, Set KK, Saadeh L. Developmental dysplasia of hip: a review. Clin Pediatr (Phila). 2015;54(10):921-928.
11. Chan A, McCaul KA, Cundy PJ, et al. Perinatal risk factors for developmental dysplasia of the hip. Arch Dis Child. 1997;76(2):F94-F100.
12. Godley DR. Assessment, diagnosis, and treatment of developmental dysplasia of the hip. JAAPA. 2013;26(3):54-58.
13. Van Sleuwen BE, Engelberts AC, Boere-Boonekamp MM, et al. Swaddling: a systematic review. Pediatrics. 2007;120(4):e1097-e1106.
14. American Academy of Orthopaedic Surgeons, American Association of Orthopaedic Surgeons. Position statement: swaddling and developmental hip dysplasia. www.aaos.org/uploadedFiles/PreProduction/About/Opinion_Statements/position/1186%20Swaddling%20and%20Developmental%20Hip%20Dysplasia.pdf. Accessed January 22, 2016.
15. Clarke NM. Swaddling and hip dysplasia: an orthopaedic perspective. Arch Dis Child. 2014;99(1):5-6.
16. Storer SK, Skaggs DL. Developmental dysplasia of the hip. Am Fam Physician. 2006;74(8):1310-1316.
17. McAbee GN, Donn SM, Mendelson RA, et al. Medical diagnoses commonly associated with pediatric malpractice lawsuits in the United States. Pediatrics. 2008;122(6):e1282-e1286.
18. Imrie M, Scott V, Stearns P, et al. Is ultrasound screening for DDH in babies born breech sufficient? J Child Orthop. 2010;4(1):3-8.
19. Chen HW, Chang CH, Tsai ST, et al. Natural progression of hip dysplasia in newborns: a reflection of hip ultrasonographic screenings in newborn nurseries. J Pediatr Orthop B. 2010;19(5):418-423.
20. Gans I, Flynn JM, Sankar WN. Abduction bracing for residual acetabular dysplasia in infantile DDH. J Pediatr Orthop. 2013;33(7):714-718.
21. Murnaghan ML, Browne RH, Sucato DJ, Birch J. Femoral nerve palsy in Pavlik harness treatment for developmental dysplasia of the hip. J Bone Joint Surg Am. 2011;93(5):493-499.
22. Dezateux C, Rosendahl K. Developmental dysplasia of the hip. Lancet. 2007;369(9572):1541-1552.