Applied Evidence

CASE 1 › Mr. L is a 47-year-old construction worker who comes to the clinic with a 3-month history of bothersome, constant, high-pitched ringing in his ears that is worse on his right side. He also reports mild hearing loss. Mr. L notes that he keeps a busy schedule, working weekends at a local shooting range. He is a 30-pack-year smoker and takes sumatriptan and nasal fluticasone spray as needed for migraines and allergies. On inspection, his ears appear normal.

CASE 2 › Ms. B, age 68 years, seeks treatment for a constant pulsatile noise in her left ear, which has been bothering her for the past 5 months. She lives alone and says this noise is worse when the house is quiet. She takes chlorthalidone for hypertension and prophylactic aspirin. She indicates that she has no problems with her hearing. On exam, the noise synchronizes with her pulse.

If Mr. L and Ms. B were your patients, what would your next steps be?

An estimated 50 million people in the United States experience some form of tinnitus,¹ and the incidence is on the rise, which some have attributed to the increased use of personal music devices.^2,3 Patients often describe tinnitus as a ringing noise, but it also can be perceived as buzzing, chirping, hissing, whistling, humming, or other sound. It is more often bilateral than unilateral⁴ and more often intermittent than continuous.¹

Tinnitus may be present in childhood, but the prevalence increases with age. Surveys show that approximately 25% of adults experience symptoms and one-fourth of these patients report that it interferes with daily activities.^1,2 The prevalence peaks at 31% in patients between the ages of 60 and 69 years.¹

The severity of the condition ranges from causing patients to merely be aware of the noise to having substantial adverse effects on their quality of life. Because not all patients will report tinnitus symptoms, it is important to be aware of risk factors, which include advanced age, male sex, history of military service, and a work history that includes exposure to loud noise.^1,2 Smoking and hypertension also are associated with higher rates of tinnitus, as is living in the southern United States.¹

A subjective, or objective case of tinnitus?

While subjective tinnitus consists of noises only the patient can hear, objective tinnitus refers to noises, including somatosounds such as turbulent blood flow or palatal myoclonus, that a physician could at least theoretically detect by auscultation or with an amplifying device. Objective tinnitus is less common than subjective tinnitus and more often has an identifiable and correctable source,⁵ though it may herald a serious underlying condition. When tinnitus is pulsatile or rhythmic, it may be the result of an arteriovenous fistula, arteriovenous malformation, cerebral aneurysm, arterial bruit, or other vascular lesion, such as a glomus tumor.⁶ Nonvascular conditions like palatal myoclonus present with clicking or low-pitched buzzing and may be a result of multiple sclerosis.⁷

The causes of both subjective and objective tinnitus are detailed in TABLE 1.^2,6,7

Tinnitus and hearing loss: The connection

Most tinnitus is associated with hearing loss and probably results from a disruption in the normal suppression of neuronal activity in the central nervous system.^2,8

When a patient reports pulsatile tinnitus, perform auscultation over vascular structures in the neck, temple, and around the ear.Conductive hearing loss can be caused by cerumen impaction, otosclerosis, or cholesteatoma. Sensorineural hearing loss (SNHL), which is more common than conductive hearing loss, often is irreversible. The damage typically occurs in the stereocilia cells of the cochlea. These cells trigger the release of neurotransmitters that activate the eighth cranial nerve and cause abnormal excitation along the auditory pathway, giving the perception of sound in a quiet environment.²

Patients with SNHL usually have a history of prolonged exposure to loud noise (eg, heavy machinery, firearms, personal musical devices such as an iPod, or musical instruments) and often describe their tinnitus as a bilateral, high-pitched, continuous ringing. The other major category of SNHL that causes tinnitus is presbycusis—the hearing loss associated with aging—which has clinical features similar to noise-induced hearing loss.⁸

What to look for

Evaluation of tinnitus begins with a thorough history and physical exam (FIGURE 1).^2,6,7,9 Key components of the exam include inspecting the ears, nose, and throat and evaluating cranial nerve function. Weber and Rinne tuning fork testing can help to confirm a conductive hearing loss. When evaluating a patient who reports pulsatile tinnitus, perform auscultation over vascular structures in the neck, temple, and around the ear.

Obtain targeted laboratory studies if there is a suggested metabolic etiology for tinnitus (TABLE 1).^2,6,7 Handheld tympanometry that is flatlined or fluctuates with breathing can help support the diagnosis of a subtle middle ear effusion or patulous eustachian tube, respectively.

It is important to quantify how tinnitus affects a patient’s mood, including irritability and concentration. Tinnitus can be measured on several scales, including the Tinnitus Functional Index (TFI),¹⁰ which is easily completed in the office. It has been validated to quantify the severity of symptoms and can be used to monitor a patient’s progress. A copy of the TFI and its scoring instructions are available at http://www.ohsu.edu/xd/health/services/ent/services/tinnitus-clinic/tinnitus-functional-index.cfm.

Up to 25% of patients with chronic tinnitus report an increase in severity over time. Thus, prevention efforts can be beneficial for patients who already have the condition.Refer most patients to audiology. Patient’s symptoms often correlate poorly with acoustic functioning.⁶ Unless you find simple, reversible causes of tinnitus on history and physical, a comprehensive audiologic evaluation is essential. Components of these evaluations include pure-tone thresholds, tympanometry, speech thresholds, and speech discrimination testing.⁷

Image when necessary. If audiometric testing indicates cochlear damage, imaging generally is unnecessary because SNHL has been confirmed.⁷ However, if a retrocochlear hearing deficit is detected, auditory brainstem response testing is useful to help locate the lesion. Gadolinium-enhanced magnetic resonance imaging of the internal auditory canals also can be performed to evaluate for central nervous system lesions.^2,7 This will detect vestibular schwannoma, which is the most frequent cause of tinnitus apparent on imaging.¹¹

Pulsatility is the one true red flag feature of tinnitus and regardless of audiometry, patients with pulsatile tinnitus require imaging to rule out vascular lesions. The petrous carotid system is a common culprit; therefore, contrast-enhanced high-resolution computed tomography (CT) of the temporal bone is a reasonable initial study since it also will detect osseous abnormalities of the inner ear. However, angiography often is eventually necessary (conventional, magnetic resonance, or CT) to exclude a dural arteriovenous fistula or malformation-the most common cause of objective, pulsatile tinnitus.¹¹ When tinnitus is pulsatile, unilateral, atypical in nature, or associated with deafness, imaging plus referral to a neurologist or otolaryngologist is advisable.⁶

Medications, and other factors to consider

Many different types of medications and substances can have ototoxic effects, mainly on the cochlear hair cells (TABLE 2).¹² The damage may be reversible or irreversible. When doing so would be clinically prudent, consider tapering a patient off a drug that may be causing tinnitus.⁷

Other causes to consider. Pain in the jaw or neck may be due to a temporomandibular joint disorder or a cervical spine problem like whiplash; these conditions are associated with tinnitus and vertigo.^7,13 The combination of low-pitched tinnitus, vertigo, aural fullness, and hearing loss often signifies Meniere’s disease—especially if symptoms are episodic.

Address mood disorders. Although insomnia, anxiety, depression, and posttraumatic stress disorder generally are not considered causes, these conditions are associated with tinnitus and can exacerbate the condition. Tinnitus can trigger depression, and vice versa. Optimizing treatment for these common problems can significantly reduce suffering.^6,7

For most patients, you'll focus on prevention, rather than Tx

Treatment for tinnitus (which we’ll describe in a bit) is necessary only for patients for whom the condition has substantially affected the quality of their life.² Greater emphasis should be placed on prevention.

Most tinnitus originates from the auditory system and is considered irreversible, but up to 25% of patients with chronic tinnitus report an increase in severity over time.¹⁴ Therefore, prevention can be beneficial not only for patients at risk of developing tinnitus, but also for those already affected by it. Prevention efforts should focus on protecting hearing by reducing noise levels and exposure time to certain noise thresholds.

Recommend ear plugs or muffs for patients exposed to >80 dB and both forms of protection for those exposed to >95 dB.The decibel (dB) scale is logarithmic; perception of sound loudness doubles every 10 dB. The sound of a vacuum cleaner is approximately 70 dB; the average human pain threshold is roughly 110 dB, which is the loudness of live rock music. Eardrum rupture occurs at approximately 150 dB—the equivalent of hearing a jet take off at 25 meters.

Talk to patients about hearing protection devices. The US Environmental Protection Agency monitors all hearing protection devices and assigns them a Noise Reduction Rating (NRR). The adequacy of single vs double hearing protection depends on the dB exposure level, duration of exposure, and NRR for the protective device(s). In general, recommend single hearing protection (ear plugs, which are inserted in the ear canal, or ear muffs, which fit around the ears) to patients exposed to >80 dB and dual hearing protection (ear plugs and muffs) to those exposed to >95 dB. More guidance on single or dual hearing protection can be obtained from a local occupational health physician or from https://www.osha.gov/dts/osta/otm/noise/hcp/attenuation_estimation.html.

There are drawbacks to using certain forms of ear protection. Regular use can increase the likelihood of cerumen impaction or otitis externa, both of which can actually cause tinnitus. Proper training on how to use hearing protection devices and routine otologic examinations are advisable for patients who frequently use ear protection.

Techniques that can help patients to better cope

The most common therapies used to treat tinnitus are cognitive-behavioral therapy (CBT) and tinnitus retraining therapy (TRT). Both are techniques of habituation designed to change the way patients think about, and emotionally respond to, tinnitus.^15,16

CBT is administered by a skilled therapist and employs relaxation exercises, coping strategies, and deconditioning techniques.¹⁶ The goal of CBT is to reduce arousal levels and reverse negative thoughts about tinnitus.¹⁶ A recent Cochrane review found that although CBT does not subjectively reduce the loudness of tinnitus, it does significantly improve quality of life and depression caused by tinnitus.¹⁷ CBT’s benefits also extend to other common comorbidities such as SNHL, insomnia, depression, and anxiety.¹⁶ Up to 75% of patients experience improvement in their score on the standardized Tinnitus Handicap Questionnaire one year after completing therapy.¹⁶

TRT combines counseling, education, and acoustic therapy—using soft music or a sound machine—to minimize the bothersome nature of the condition.¹⁵ TRT is delivered by a team of physicians, audiologists, and psychologists and requires commitment from patients because most therapies are performed at a specialized tinnitus center over the course of up to 2 years.¹⁵ Retrospective trials of TRT have generally been positive, finding that this approach minimizes the annoyance patients experience.¹⁵

Even in the absence of a formal TRT protocol, patients can take advantage of acoustic therapy. Patients should be advised to add pleasant noise to quiet environments with soft music or sound machines. “Masking” devices are also an option. These commercially available sound generators fit in the ear and may lessen patients’ perception of tinnitus.^2,18

Evidence supporting medications is weak

Though many medications have been investigated for treating tinnitus, most have been studied in small clinical trials and none is FDA-approved for tinnitus.

Acamprosate, which is FDA-approved for maintaining alcohol abstinence in alcohol-dependent patients, is a relatively new tinnitus treatment option. In small randomized, double-blinded, placebo controlled trials, approximately 90% of patients treated with acamprosate experienced improvement in tinnitus severity and quality of life.¹⁹ Larger studies will be necessary to determine if frequent adverse effects (including depression, anxiety, diarrhea, and drowsiness) will hamper its usefulness.

Benzodiazepines (mainly alprazolam) tend to reduce tinnitus-associated anxiety and also may decrease tinnitus intensity via central suppression of the auditory pathway. However, because evidence is limited to small trials with methodological flaws, and because benzodiazepines have the potential for dependence, the risks and benefits of these agents must be weighed carefully.^7,16

Lidocaine has a long history of use for tinnitus, by both intravenous and intratympanic routes. Its benefits are unclear. In some trials, lidocaine was moderately effective in the short term, whereas in others, it appeared to make tinnitus worse.^7,20

Oral misoprostol also may be an option, according to a series of placebo-controlled trials.²⁰ But the benefit of this medication may be limited to the perception of loudness, and not other tinnitus measures, such as improved sleep and concentration.²⁰

Antidepressants can have a profound positive effect on tinnitus in patients with severe depression but do not have the same effect on patients who do not suffer from depression.²⁰ Anticonvulsants such as gabapentin have not been found to be effective for tinnitus.²¹

Researchers are investigating centrally acting agents, such as the N-methyl-D-aspartate antagonist neramexane, for the treatment of tinnitus. With safety and tolerability now established, neramexane is in European phase III trials.²²

Acupuncture and electromagnetic stimulation have not proven efficacious in the treatment of tinnitus.The jury is out on alternative therapies

With data lacking for prescription medications, patients may look into complementary and alternative therapies. However, consistent evidence is lacking for these therapies, as well. Ginkgo biloba has been found to reduce tinnitus severity and loudness in limited studies,²³ although some preparations may be superior to others.¹⁴ In a double-blind, randomized controlled trial, melatonin decreased tinnitus intensity significantly, particularly for men and for patients with severe symptoms or a history of noise exposure.²⁴ Zinc supplements may improve tinnitus in patients with zinc deficiency.^20,25 Acupuncture and electromagnetic stimulation have not proven efficacious in the treatment of tinnitus.²¹

Additional steps that your patient can take

A trial of a hearing aid is often worthwhile as a noninvasive, first-line intervention for patients with tinnitus and SNHL. Hearing aids reduce the perception of tinnitus by amplifying ambient sounds.⁸ Some hearing aids also incorporate masking devices and are used to treat tinnitus in patients with hearing loss. Cochlear implants also are an option for certain patients with confirmed severe SNHL. One study found that tinnitus intensity and awareness were reduced in up to 86% of patients who received cochlear implants.²⁶

The American Tinnitus Association also advises patients with tinnitus to eliminate potential aggravating factors, including salt, artificial sweeteners, sugar, alcohol, tobacco, and caffeine.²⁷

CASE 1 › Mr. L was referred for audiometric testing, which revealed severe high-frequency SNHL that was worse in the right ear. His symptoms improved slightly following a trial of a combination hearing aid/masking device and participation in TRT. He was counseled to quit smoking and use dual hearing protection for future high-noise exposure.

CASE 2 › Ms. B had normal audiometric testing and was referred for angiography. This revealed a dural arteriovenous fistula that was categorized as type III (draining directly into subarachnoid veins) with a small adjacent aneurysm. She underwent a successful clipping of the draining vein to prevent future hemorrhage. Her tinnitus subsequently resolved. Her aspirin use was not modified because it was low dose.

CORRESPONDENCE
Ethan Zimmerman, MD, Mike O’Callaghan Federal Hospital,
4700 Las Vegas Boulevard North, Nellis Air Force Base, NV 89191; [email protected]

CASE 1 › Mr. L is a 47-year-old construction worker who comes to the clinic with a 3-month history of bothersome, constant, high-pitched ringing in his ears that is worse on his right side. He also reports mild hearing loss. Mr. L notes that he keeps a busy schedule, working weekends at a local shooting range. He is a 30-pack-year smoker and takes sumatriptan and nasal fluticasone spray as needed for migraines and allergies. On inspection, his ears appear normal.

CASE 2 › Ms. B, age 68 years, seeks treatment for a constant pulsatile noise in her left ear, which has been bothering her for the past 5 months. She lives alone and says this noise is worse when the house is quiet. She takes chlorthalidone for hypertension and prophylactic aspirin. She indicates that she has no problems with her hearing. On exam, the noise synchronizes with her pulse.

If Mr. L and Ms. B were your patients, what would your next steps be?

An estimated 50 million people in the United States experience some form of tinnitus,¹ and the incidence is on the rise, which some have attributed to the increased use of personal music devices.^2,3 Patients often describe tinnitus as a ringing noise, but it also can be perceived as buzzing, chirping, hissing, whistling, humming, or other sound. It is more often bilateral than unilateral⁴ and more often intermittent than continuous.¹

Tinnitus may be present in childhood, but the prevalence increases with age. Surveys show that approximately 25% of adults experience symptoms and one-fourth of these patients report that it interferes with daily activities.^1,2 The prevalence peaks at 31% in patients between the ages of 60 and 69 years.¹

The severity of the condition ranges from causing patients to merely be aware of the noise to having substantial adverse effects on their quality of life. Because not all patients will report tinnitus symptoms, it is important to be aware of risk factors, which include advanced age, male sex, history of military service, and a work history that includes exposure to loud noise.^1,2 Smoking and hypertension also are associated with higher rates of tinnitus, as is living in the southern United States.¹

A subjective, or objective case of tinnitus?

While subjective tinnitus consists of noises only the patient can hear, objective tinnitus refers to noises, including somatosounds such as turbulent blood flow or palatal myoclonus, that a physician could at least theoretically detect by auscultation or with an amplifying device. Objective tinnitus is less common than subjective tinnitus and more often has an identifiable and correctable source,⁵ though it may herald a serious underlying condition. When tinnitus is pulsatile or rhythmic, it may be the result of an arteriovenous fistula, arteriovenous malformation, cerebral aneurysm, arterial bruit, or other vascular lesion, such as a glomus tumor.⁶ Nonvascular conditions like palatal myoclonus present with clicking or low-pitched buzzing and may be a result of multiple sclerosis.⁷

The causes of both subjective and objective tinnitus are detailed in TABLE 1.^2,6,7

Tinnitus and hearing loss: The connection

Most tinnitus is associated with hearing loss and probably results from a disruption in the normal suppression of neuronal activity in the central nervous system.^2,8

When a patient reports pulsatile tinnitus, perform auscultation over vascular structures in the neck, temple, and around the ear.Conductive hearing loss can be caused by cerumen impaction, otosclerosis, or cholesteatoma. Sensorineural hearing loss (SNHL), which is more common than conductive hearing loss, often is irreversible. The damage typically occurs in the stereocilia cells of the cochlea. These cells trigger the release of neurotransmitters that activate the eighth cranial nerve and cause abnormal excitation along the auditory pathway, giving the perception of sound in a quiet environment.²

Patients with SNHL usually have a history of prolonged exposure to loud noise (eg, heavy machinery, firearms, personal musical devices such as an iPod, or musical instruments) and often describe their tinnitus as a bilateral, high-pitched, continuous ringing. The other major category of SNHL that causes tinnitus is presbycusis—the hearing loss associated with aging—which has clinical features similar to noise-induced hearing loss.⁸

What to look for

Evaluation of tinnitus begins with a thorough history and physical exam (FIGURE 1).^2,6,7,9 Key components of the exam include inspecting the ears, nose, and throat and evaluating cranial nerve function. Weber and Rinne tuning fork testing can help to confirm a conductive hearing loss. When evaluating a patient who reports pulsatile tinnitus, perform auscultation over vascular structures in the neck, temple, and around the ear.

Obtain targeted laboratory studies if there is a suggested metabolic etiology for tinnitus (TABLE 1).^2,6,7 Handheld tympanometry that is flatlined or fluctuates with breathing can help support the diagnosis of a subtle middle ear effusion or patulous eustachian tube, respectively.

It is important to quantify how tinnitus affects a patient’s mood, including irritability and concentration. Tinnitus can be measured on several scales, including the Tinnitus Functional Index (TFI),¹⁰ which is easily completed in the office. It has been validated to quantify the severity of symptoms and can be used to monitor a patient’s progress. A copy of the TFI and its scoring instructions are available at http://www.ohsu.edu/xd/health/services/ent/services/tinnitus-clinic/tinnitus-functional-index.cfm.

Up to 25% of patients with chronic tinnitus report an increase in severity over time. Thus, prevention efforts can be beneficial for patients who already have the condition.Refer most patients to audiology. Patient’s symptoms often correlate poorly with acoustic functioning.⁶ Unless you find simple, reversible causes of tinnitus on history and physical, a comprehensive audiologic evaluation is essential. Components of these evaluations include pure-tone thresholds, tympanometry, speech thresholds, and speech discrimination testing.⁷

Image when necessary. If audiometric testing indicates cochlear damage, imaging generally is unnecessary because SNHL has been confirmed.⁷ However, if a retrocochlear hearing deficit is detected, auditory brainstem response testing is useful to help locate the lesion. Gadolinium-enhanced magnetic resonance imaging of the internal auditory canals also can be performed to evaluate for central nervous system lesions.^2,7 This will detect vestibular schwannoma, which is the most frequent cause of tinnitus apparent on imaging.¹¹

Pulsatility is the one true red flag feature of tinnitus and regardless of audiometry, patients with pulsatile tinnitus require imaging to rule out vascular lesions. The petrous carotid system is a common culprit; therefore, contrast-enhanced high-resolution computed tomography (CT) of the temporal bone is a reasonable initial study since it also will detect osseous abnormalities of the inner ear. However, angiography often is eventually necessary (conventional, magnetic resonance, or CT) to exclude a dural arteriovenous fistula or malformation-the most common cause of objective, pulsatile tinnitus.¹¹ When tinnitus is pulsatile, unilateral, atypical in nature, or associated with deafness, imaging plus referral to a neurologist or otolaryngologist is advisable.⁶

Medications, and other factors to consider

Many different types of medications and substances can have ototoxic effects, mainly on the cochlear hair cells (TABLE 2).¹² The damage may be reversible or irreversible. When doing so would be clinically prudent, consider tapering a patient off a drug that may be causing tinnitus.⁷

Other causes to consider. Pain in the jaw or neck may be due to a temporomandibular joint disorder or a cervical spine problem like whiplash; these conditions are associated with tinnitus and vertigo.^7,13 The combination of low-pitched tinnitus, vertigo, aural fullness, and hearing loss often signifies Meniere’s disease—especially if symptoms are episodic.

Address mood disorders. Although insomnia, anxiety, depression, and posttraumatic stress disorder generally are not considered causes, these conditions are associated with tinnitus and can exacerbate the condition. Tinnitus can trigger depression, and vice versa. Optimizing treatment for these common problems can significantly reduce suffering.^6,7

For most patients, you'll focus on prevention, rather than Tx

Treatment for tinnitus (which we’ll describe in a bit) is necessary only for patients for whom the condition has substantially affected the quality of their life.² Greater emphasis should be placed on prevention.

Most tinnitus originates from the auditory system and is considered irreversible, but up to 25% of patients with chronic tinnitus report an increase in severity over time.¹⁴ Therefore, prevention can be beneficial not only for patients at risk of developing tinnitus, but also for those already affected by it. Prevention efforts should focus on protecting hearing by reducing noise levels and exposure time to certain noise thresholds.

Recommend ear plugs or muffs for patients exposed to >80 dB and both forms of protection for those exposed to >95 dB.The decibel (dB) scale is logarithmic; perception of sound loudness doubles every 10 dB. The sound of a vacuum cleaner is approximately 70 dB; the average human pain threshold is roughly 110 dB, which is the loudness of live rock music. Eardrum rupture occurs at approximately 150 dB—the equivalent of hearing a jet take off at 25 meters.

Talk to patients about hearing protection devices. The US Environmental Protection Agency monitors all hearing protection devices and assigns them a Noise Reduction Rating (NRR). The adequacy of single vs double hearing protection depends on the dB exposure level, duration of exposure, and NRR for the protective device(s). In general, recommend single hearing protection (ear plugs, which are inserted in the ear canal, or ear muffs, which fit around the ears) to patients exposed to >80 dB and dual hearing protection (ear plugs and muffs) to those exposed to >95 dB. More guidance on single or dual hearing protection can be obtained from a local occupational health physician or from https://www.osha.gov/dts/osta/otm/noise/hcp/attenuation_estimation.html.

There are drawbacks to using certain forms of ear protection. Regular use can increase the likelihood of cerumen impaction or otitis externa, both of which can actually cause tinnitus. Proper training on how to use hearing protection devices and routine otologic examinations are advisable for patients who frequently use ear protection.

Techniques that can help patients to better cope

The most common therapies used to treat tinnitus are cognitive-behavioral therapy (CBT) and tinnitus retraining therapy (TRT). Both are techniques of habituation designed to change the way patients think about, and emotionally respond to, tinnitus.^15,16

CBT is administered by a skilled therapist and employs relaxation exercises, coping strategies, and deconditioning techniques.¹⁶ The goal of CBT is to reduce arousal levels and reverse negative thoughts about tinnitus.¹⁶ A recent Cochrane review found that although CBT does not subjectively reduce the loudness of tinnitus, it does significantly improve quality of life and depression caused by tinnitus.¹⁷ CBT’s benefits also extend to other common comorbidities such as SNHL, insomnia, depression, and anxiety.¹⁶ Up to 75% of patients experience improvement in their score on the standardized Tinnitus Handicap Questionnaire one year after completing therapy.¹⁶

TRT combines counseling, education, and acoustic therapy—using soft music or a sound machine—to minimize the bothersome nature of the condition.¹⁵ TRT is delivered by a team of physicians, audiologists, and psychologists and requires commitment from patients because most therapies are performed at a specialized tinnitus center over the course of up to 2 years.¹⁵ Retrospective trials of TRT have generally been positive, finding that this approach minimizes the annoyance patients experience.¹⁵

Even in the absence of a formal TRT protocol, patients can take advantage of acoustic therapy. Patients should be advised to add pleasant noise to quiet environments with soft music or sound machines. “Masking” devices are also an option. These commercially available sound generators fit in the ear and may lessen patients’ perception of tinnitus.^2,18

Evidence supporting medications is weak

Though many medications have been investigated for treating tinnitus, most have been studied in small clinical trials and none is FDA-approved for tinnitus.

Acamprosate, which is FDA-approved for maintaining alcohol abstinence in alcohol-dependent patients, is a relatively new tinnitus treatment option. In small randomized, double-blinded, placebo controlled trials, approximately 90% of patients treated with acamprosate experienced improvement in tinnitus severity and quality of life.¹⁹ Larger studies will be necessary to determine if frequent adverse effects (including depression, anxiety, diarrhea, and drowsiness) will hamper its usefulness.

Benzodiazepines (mainly alprazolam) tend to reduce tinnitus-associated anxiety and also may decrease tinnitus intensity via central suppression of the auditory pathway. However, because evidence is limited to small trials with methodological flaws, and because benzodiazepines have the potential for dependence, the risks and benefits of these agents must be weighed carefully.^7,16

Lidocaine has a long history of use for tinnitus, by both intravenous and intratympanic routes. Its benefits are unclear. In some trials, lidocaine was moderately effective in the short term, whereas in others, it appeared to make tinnitus worse.^7,20

Oral misoprostol also may be an option, according to a series of placebo-controlled trials.²⁰ But the benefit of this medication may be limited to the perception of loudness, and not other tinnitus measures, such as improved sleep and concentration.²⁰

Antidepressants can have a profound positive effect on tinnitus in patients with severe depression but do not have the same effect on patients who do not suffer from depression.²⁰ Anticonvulsants such as gabapentin have not been found to be effective for tinnitus.²¹

Researchers are investigating centrally acting agents, such as the N-methyl-D-aspartate antagonist neramexane, for the treatment of tinnitus. With safety and tolerability now established, neramexane is in European phase III trials.²²

Acupuncture and electromagnetic stimulation have not proven efficacious in the treatment of tinnitus.The jury is out on alternative therapies

With data lacking for prescription medications, patients may look into complementary and alternative therapies. However, consistent evidence is lacking for these therapies, as well. Ginkgo biloba has been found to reduce tinnitus severity and loudness in limited studies,²³ although some preparations may be superior to others.¹⁴ In a double-blind, randomized controlled trial, melatonin decreased tinnitus intensity significantly, particularly for men and for patients with severe symptoms or a history of noise exposure.²⁴ Zinc supplements may improve tinnitus in patients with zinc deficiency.^20,25 Acupuncture and electromagnetic stimulation have not proven efficacious in the treatment of tinnitus.²¹

Additional steps that your patient can take

A trial of a hearing aid is often worthwhile as a noninvasive, first-line intervention for patients with tinnitus and SNHL. Hearing aids reduce the perception of tinnitus by amplifying ambient sounds.⁸ Some hearing aids also incorporate masking devices and are used to treat tinnitus in patients with hearing loss. Cochlear implants also are an option for certain patients with confirmed severe SNHL. One study found that tinnitus intensity and awareness were reduced in up to 86% of patients who received cochlear implants.²⁶

The American Tinnitus Association also advises patients with tinnitus to eliminate potential aggravating factors, including salt, artificial sweeteners, sugar, alcohol, tobacco, and caffeine.²⁷

CASE 1 › Mr. L was referred for audiometric testing, which revealed severe high-frequency SNHL that was worse in the right ear. His symptoms improved slightly following a trial of a combination hearing aid/masking device and participation in TRT. He was counseled to quit smoking and use dual hearing protection for future high-noise exposure.

CASE 2 › Ms. B had normal audiometric testing and was referred for angiography. This revealed a dural arteriovenous fistula that was categorized as type III (draining directly into subarachnoid veins) with a small adjacent aneurysm. She underwent a successful clipping of the draining vein to prevent future hemorrhage. Her tinnitus subsequently resolved. Her aspirin use was not modified because it was low dose.

CORRESPONDENCE
Ethan Zimmerman, MD, Mike O’Callaghan Federal Hospital,
4700 Las Vegas Boulevard North, Nellis Air Force Base, NV 89191; [email protected]

CASE 1 › Mr. L is a 47-year-old construction worker who comes to the clinic with a 3-month history of bothersome, constant, high-pitched ringing in his ears that is worse on his right side. He also reports mild hearing loss. Mr. L notes that he keeps a busy schedule, working weekends at a local shooting range. He is a 30-pack-year smoker and takes sumatriptan and nasal fluticasone spray as needed for migraines and allergies. On inspection, his ears appear normal.

CASE 2 › Ms. B, age 68 years, seeks treatment for a constant pulsatile noise in her left ear, which has been bothering her for the past 5 months. She lives alone and says this noise is worse when the house is quiet. She takes chlorthalidone for hypertension and prophylactic aspirin. She indicates that she has no problems with her hearing. On exam, the noise synchronizes with her pulse.

If Mr. L and Ms. B were your patients, what would your next steps be?

An estimated 50 million people in the United States experience some form of tinnitus,¹ and the incidence is on the rise, which some have attributed to the increased use of personal music devices.^2,3 Patients often describe tinnitus as a ringing noise, but it also can be perceived as buzzing, chirping, hissing, whistling, humming, or other sound. It is more often bilateral than unilateral⁴ and more often intermittent than continuous.¹

Tinnitus may be present in childhood, but the prevalence increases with age. Surveys show that approximately 25% of adults experience symptoms and one-fourth of these patients report that it interferes with daily activities.^1,2 The prevalence peaks at 31% in patients between the ages of 60 and 69 years.¹

The severity of the condition ranges from causing patients to merely be aware of the noise to having substantial adverse effects on their quality of life. Because not all patients will report tinnitus symptoms, it is important to be aware of risk factors, which include advanced age, male sex, history of military service, and a work history that includes exposure to loud noise.^1,2 Smoking and hypertension also are associated with higher rates of tinnitus, as is living in the southern United States.¹

A subjective, or objective case of tinnitus?

While subjective tinnitus consists of noises only the patient can hear, objective tinnitus refers to noises, including somatosounds such as turbulent blood flow or palatal myoclonus, that a physician could at least theoretically detect by auscultation or with an amplifying device. Objective tinnitus is less common than subjective tinnitus and more often has an identifiable and correctable source,⁵ though it may herald a serious underlying condition. When tinnitus is pulsatile or rhythmic, it may be the result of an arteriovenous fistula, arteriovenous malformation, cerebral aneurysm, arterial bruit, or other vascular lesion, such as a glomus tumor.⁶ Nonvascular conditions like palatal myoclonus present with clicking or low-pitched buzzing and may be a result of multiple sclerosis.⁷

The causes of both subjective and objective tinnitus are detailed in TABLE 1.^2,6,7

Tinnitus and hearing loss: The connection

Most tinnitus is associated with hearing loss and probably results from a disruption in the normal suppression of neuronal activity in the central nervous system.^2,8

When a patient reports pulsatile tinnitus, perform auscultation over vascular structures in the neck, temple, and around the ear.Conductive hearing loss can be caused by cerumen impaction, otosclerosis, or cholesteatoma. Sensorineural hearing loss (SNHL), which is more common than conductive hearing loss, often is irreversible. The damage typically occurs in the stereocilia cells of the cochlea. These cells trigger the release of neurotransmitters that activate the eighth cranial nerve and cause abnormal excitation along the auditory pathway, giving the perception of sound in a quiet environment.²

Patients with SNHL usually have a history of prolonged exposure to loud noise (eg, heavy machinery, firearms, personal musical devices such as an iPod, or musical instruments) and often describe their tinnitus as a bilateral, high-pitched, continuous ringing. The other major category of SNHL that causes tinnitus is presbycusis—the hearing loss associated with aging—which has clinical features similar to noise-induced hearing loss.⁸

What to look for

Evaluation of tinnitus begins with a thorough history and physical exam (FIGURE 1).^2,6,7,9 Key components of the exam include inspecting the ears, nose, and throat and evaluating cranial nerve function. Weber and Rinne tuning fork testing can help to confirm a conductive hearing loss. When evaluating a patient who reports pulsatile tinnitus, perform auscultation over vascular structures in the neck, temple, and around the ear.

Obtain targeted laboratory studies if there is a suggested metabolic etiology for tinnitus (TABLE 1).^2,6,7 Handheld tympanometry that is flatlined or fluctuates with breathing can help support the diagnosis of a subtle middle ear effusion or patulous eustachian tube, respectively.

It is important to quantify how tinnitus affects a patient’s mood, including irritability and concentration. Tinnitus can be measured on several scales, including the Tinnitus Functional Index (TFI),¹⁰ which is easily completed in the office. It has been validated to quantify the severity of symptoms and can be used to monitor a patient’s progress. A copy of the TFI and its scoring instructions are available at http://www.ohsu.edu/xd/health/services/ent/services/tinnitus-clinic/tinnitus-functional-index.cfm.

Up to 25% of patients with chronic tinnitus report an increase in severity over time. Thus, prevention efforts can be beneficial for patients who already have the condition.Refer most patients to audiology. Patient’s symptoms often correlate poorly with acoustic functioning.⁶ Unless you find simple, reversible causes of tinnitus on history and physical, a comprehensive audiologic evaluation is essential. Components of these evaluations include pure-tone thresholds, tympanometry, speech thresholds, and speech discrimination testing.⁷

Image when necessary. If audiometric testing indicates cochlear damage, imaging generally is unnecessary because SNHL has been confirmed.⁷ However, if a retrocochlear hearing deficit is detected, auditory brainstem response testing is useful to help locate the lesion. Gadolinium-enhanced magnetic resonance imaging of the internal auditory canals also can be performed to evaluate for central nervous system lesions.^2,7 This will detect vestibular schwannoma, which is the most frequent cause of tinnitus apparent on imaging.¹¹

Pulsatility is the one true red flag feature of tinnitus and regardless of audiometry, patients with pulsatile tinnitus require imaging to rule out vascular lesions. The petrous carotid system is a common culprit; therefore, contrast-enhanced high-resolution computed tomography (CT) of the temporal bone is a reasonable initial study since it also will detect osseous abnormalities of the inner ear. However, angiography often is eventually necessary (conventional, magnetic resonance, or CT) to exclude a dural arteriovenous fistula or malformation-the most common cause of objective, pulsatile tinnitus.¹¹ When tinnitus is pulsatile, unilateral, atypical in nature, or associated with deafness, imaging plus referral to a neurologist or otolaryngologist is advisable.⁶

Medications, and other factors to consider

Many different types of medications and substances can have ototoxic effects, mainly on the cochlear hair cells (TABLE 2).¹² The damage may be reversible or irreversible. When doing so would be clinically prudent, consider tapering a patient off a drug that may be causing tinnitus.⁷

Other causes to consider. Pain in the jaw or neck may be due to a temporomandibular joint disorder or a cervical spine problem like whiplash; these conditions are associated with tinnitus and vertigo.^7,13 The combination of low-pitched tinnitus, vertigo, aural fullness, and hearing loss often signifies Meniere’s disease—especially if symptoms are episodic.

Address mood disorders. Although insomnia, anxiety, depression, and posttraumatic stress disorder generally are not considered causes, these conditions are associated with tinnitus and can exacerbate the condition. Tinnitus can trigger depression, and vice versa. Optimizing treatment for these common problems can significantly reduce suffering.^6,7

For most patients, you'll focus on prevention, rather than Tx

Treatment for tinnitus (which we’ll describe in a bit) is necessary only for patients for whom the condition has substantially affected the quality of their life.² Greater emphasis should be placed on prevention.

Most tinnitus originates from the auditory system and is considered irreversible, but up to 25% of patients with chronic tinnitus report an increase in severity over time.¹⁴ Therefore, prevention can be beneficial not only for patients at risk of developing tinnitus, but also for those already affected by it. Prevention efforts should focus on protecting hearing by reducing noise levels and exposure time to certain noise thresholds.

Recommend ear plugs or muffs for patients exposed to >80 dB and both forms of protection for those exposed to >95 dB.The decibel (dB) scale is logarithmic; perception of sound loudness doubles every 10 dB. The sound of a vacuum cleaner is approximately 70 dB; the average human pain threshold is roughly 110 dB, which is the loudness of live rock music. Eardrum rupture occurs at approximately 150 dB—the equivalent of hearing a jet take off at 25 meters.

Talk to patients about hearing protection devices. The US Environmental Protection Agency monitors all hearing protection devices and assigns them a Noise Reduction Rating (NRR). The adequacy of single vs double hearing protection depends on the dB exposure level, duration of exposure, and NRR for the protective device(s). In general, recommend single hearing protection (ear plugs, which are inserted in the ear canal, or ear muffs, which fit around the ears) to patients exposed to >80 dB and dual hearing protection (ear plugs and muffs) to those exposed to >95 dB. More guidance on single or dual hearing protection can be obtained from a local occupational health physician or from https://www.osha.gov/dts/osta/otm/noise/hcp/attenuation_estimation.html.

There are drawbacks to using certain forms of ear protection. Regular use can increase the likelihood of cerumen impaction or otitis externa, both of which can actually cause tinnitus. Proper training on how to use hearing protection devices and routine otologic examinations are advisable for patients who frequently use ear protection.

Techniques that can help patients to better cope

The most common therapies used to treat tinnitus are cognitive-behavioral therapy (CBT) and tinnitus retraining therapy (TRT). Both are techniques of habituation designed to change the way patients think about, and emotionally respond to, tinnitus.^15,16

CBT is administered by a skilled therapist and employs relaxation exercises, coping strategies, and deconditioning techniques.¹⁶ The goal of CBT is to reduce arousal levels and reverse negative thoughts about tinnitus.¹⁶ A recent Cochrane review found that although CBT does not subjectively reduce the loudness of tinnitus, it does significantly improve quality of life and depression caused by tinnitus.¹⁷ CBT’s benefits also extend to other common comorbidities such as SNHL, insomnia, depression, and anxiety.¹⁶ Up to 75% of patients experience improvement in their score on the standardized Tinnitus Handicap Questionnaire one year after completing therapy.¹⁶

TRT combines counseling, education, and acoustic therapy—using soft music or a sound machine—to minimize the bothersome nature of the condition.¹⁵ TRT is delivered by a team of physicians, audiologists, and psychologists and requires commitment from patients because most therapies are performed at a specialized tinnitus center over the course of up to 2 years.¹⁵ Retrospective trials of TRT have generally been positive, finding that this approach minimizes the annoyance patients experience.¹⁵

Even in the absence of a formal TRT protocol, patients can take advantage of acoustic therapy. Patients should be advised to add pleasant noise to quiet environments with soft music or sound machines. “Masking” devices are also an option. These commercially available sound generators fit in the ear and may lessen patients’ perception of tinnitus.^2,18

Evidence supporting medications is weak

Though many medications have been investigated for treating tinnitus, most have been studied in small clinical trials and none is FDA-approved for tinnitus.

Acamprosate, which is FDA-approved for maintaining alcohol abstinence in alcohol-dependent patients, is a relatively new tinnitus treatment option. In small randomized, double-blinded, placebo controlled trials, approximately 90% of patients treated with acamprosate experienced improvement in tinnitus severity and quality of life.¹⁹ Larger studies will be necessary to determine if frequent adverse effects (including depression, anxiety, diarrhea, and drowsiness) will hamper its usefulness.

Benzodiazepines (mainly alprazolam) tend to reduce tinnitus-associated anxiety and also may decrease tinnitus intensity via central suppression of the auditory pathway. However, because evidence is limited to small trials with methodological flaws, and because benzodiazepines have the potential for dependence, the risks and benefits of these agents must be weighed carefully.^7,16

Lidocaine has a long history of use for tinnitus, by both intravenous and intratympanic routes. Its benefits are unclear. In some trials, lidocaine was moderately effective in the short term, whereas in others, it appeared to make tinnitus worse.^7,20

Oral misoprostol also may be an option, according to a series of placebo-controlled trials.²⁰ But the benefit of this medication may be limited to the perception of loudness, and not other tinnitus measures, such as improved sleep and concentration.²⁰

Antidepressants can have a profound positive effect on tinnitus in patients with severe depression but do not have the same effect on patients who do not suffer from depression.²⁰ Anticonvulsants such as gabapentin have not been found to be effective for tinnitus.²¹

Researchers are investigating centrally acting agents, such as the N-methyl-D-aspartate antagonist neramexane, for the treatment of tinnitus. With safety and tolerability now established, neramexane is in European phase III trials.²²

Acupuncture and electromagnetic stimulation have not proven efficacious in the treatment of tinnitus.The jury is out on alternative therapies

With data lacking for prescription medications, patients may look into complementary and alternative therapies. However, consistent evidence is lacking for these therapies, as well. Ginkgo biloba has been found to reduce tinnitus severity and loudness in limited studies,²³ although some preparations may be superior to others.¹⁴ In a double-blind, randomized controlled trial, melatonin decreased tinnitus intensity significantly, particularly for men and for patients with severe symptoms or a history of noise exposure.²⁴ Zinc supplements may improve tinnitus in patients with zinc deficiency.^20,25 Acupuncture and electromagnetic stimulation have not proven efficacious in the treatment of tinnitus.²¹

Additional steps that your patient can take

A trial of a hearing aid is often worthwhile as a noninvasive, first-line intervention for patients with tinnitus and SNHL. Hearing aids reduce the perception of tinnitus by amplifying ambient sounds.⁸ Some hearing aids also incorporate masking devices and are used to treat tinnitus in patients with hearing loss. Cochlear implants also are an option for certain patients with confirmed severe SNHL. One study found that tinnitus intensity and awareness were reduced in up to 86% of patients who received cochlear implants.²⁶

The American Tinnitus Association also advises patients with tinnitus to eliminate potential aggravating factors, including salt, artificial sweeteners, sugar, alcohol, tobacco, and caffeine.²⁷

CASE 1 › Mr. L was referred for audiometric testing, which revealed severe high-frequency SNHL that was worse in the right ear. His symptoms improved slightly following a trial of a combination hearing aid/masking device and participation in TRT. He was counseled to quit smoking and use dual hearing protection for future high-noise exposure.

CASE 2 › Ms. B had normal audiometric testing and was referred for angiography. This revealed a dural arteriovenous fistula that was categorized as type III (draining directly into subarachnoid veins) with a small adjacent aneurysm. She underwent a successful clipping of the draining vein to prevent future hemorrhage. Her tinnitus subsequently resolved. Her aspirin use was not modified because it was low dose.

CORRESPONDENCE
Ethan Zimmerman, MD, Mike O’Callaghan Federal Hospital,
4700 Las Vegas Boulevard North, Nellis Air Force Base, NV 89191; [email protected]

With type 2 diabetes now affecting 8.3% of the US population, most primary care physicians see patients with this disorder every day.¹ Based on the concurrent obesity epidemic, aging population, and emergence of type 2 diabetes in children and adolescents, it is estimated that by 2050, the prevalence will have risen from one in 12 Americans to one in 3.¹

Type 2 diabetes is a progressive disorder, with a relentless decline in beta cells. By the time of diagnosis, patients typically have lost at least 50% of insulin secretion; within 6 years of diagnosis, insulin secretion decreases to less than 25%.²

The American Association of Clinical Endocrinologists (AACE)³ and the American Diabetes Association/European Association for the Study of Diabetes (ADA/EASD)⁴ have recently published guidelines for the management of type 2 diabetes. While the AACE’s guidelines (available at https://www.aace.com/files/aace_algorithm.pdf) focus on different treatments at different stages of disease and both glycemic and nonglycemic benefits of treatment,³ the ADA/EASD’s guidelines (see http://care.diabetesjournals.org/content/early/2012/04/17/dc12-0413.full.pdf+html) emphasize a patient-centered approach, shared decision making, and individualization of treatment goals based on both patient preference and comorbid disease states.⁴

One thing both sets of guidelines have in common is a purposeful intensification of therapy every 2 to 3 months, as needed, and the introduction of insulin one year after diagnosis if the patient is still not at goal.^3,4 But all too often, this does not occur, particularly in primary care settings.

This article will review the “when” and “how” of insulin initiation. But first, a look at barriers to insulin therapy and evidence in support of earlier use.

Clinical inertia and patient fear are associated with delays

Both the AACE and the ADA/EASD guidelines agree that metformin is best used as early as possible.^5,6 With typical use, however, metformin fails to prevent the progression of diabetes, as measured by the climb of hemoglobin A1c (HbA1c), at a failure rate of about 17% of patients per year.⁵ Physicians have been slow to intensify treatment for type 2 diabetes⁶—a phenomenon referred to as clinical inertia. In a recent study, patients were out of control—with an HbA1c >8%—for an average of 4.6 years before insulin was initiated.

Typically, physicians adopt a stepwise approach, which often results in patients spending more than 10 years with an HbA1c >7% and 5 years with an HbA1c >8% before insulin is started.⁵ In a recent Veterans Administration study, patients were out of control, with an HbA1c >8%, for an average of 4.6 years before insulin was initiated.⁷

Both patient and physician factors contribute to the delay. Patient factors include the fear of injection, the belief that insulin will interfere with their lifestyle, and the idea that the use of insulin signifies impending complications or even death.⁸ But such beliefs are starting to change. In a recent multinational study of patients with type 2 diabetes, less than 20% stated they were unwilling to start insulin.⁹

For their part, primary care physicians are much less likely to prescribe insulin than clinicians specializing in diabetes.⁶ Physician-reported barriers to insulin initiation include the time required to train patients to use it correctly; the lack of support, including access to diabetes educators; and the absence of clear guidelines on the use of insulin.¹⁰

A case for earlier insulin

There has been recent momentum in favor of earlier initiation of insulin. In fact, some researchers regard intensive insulin as an excellent first treatment for type 2 diabetes,¹¹ based on the belief that early insulin (used for a brief time) can provide not only immediate improvement in glucose control, but also a lasting “legacy” effect. The ADA/EASD guidelines support the use of insulin as a first-line treatment for patients with symptoms of insulin deficiency,⁴ but do not recommend it for everyone with newly diagnosed type 2 diabetes. 

Primary care physicians are much less likely to prescribe insulin than clinicians specializing in diabetes.There have also been a number of advances in insulin therapy over the past 2 decades. These include insulin analogs with physiologic profiles that better match daily schedules, as well as improvements in the way insulin is delivered. Insulin pens, smaller needles, disposable devices, and insulin pumps have made it easier to administer and fine-tune insulin delivery. Despite these improvements and recommendations for earlier implementation, the use of insulin in type 2 diabetes is significantly lower today than in the 1990s.¹²

When to introduce insulin

Insulin is indicated for patients with type 2 diabetes whose disease is not easily controlled. That includes individuals with decompensated type 2 diabetes, those whose HbA1c remains high even with 2 or more oral agents, and individuals who have not reached goal after a year of treatment.

Glucose toxicity. It is generally agreed that insulin is the most effective treatment for patients who present with decompensated type 2 diabetes⁴—ie, with significant hyperglycemia and catabolic symptoms such as polydipsia, polyuria, and weight loss. Initiation of insulin promotes reversal of glucose toxicity and stabilization of metabolic status. In such cases, insulin can be started at a low dose to expose the patient to the complexities of injection therapy (more about this in a bit), then titrated as needed for stabilization.

HbA1c ≥8% even with 2 or more drugs. In my experience, an oral diabetes drug will lead to a drop in HbA1c of about one percentage point. Generally, the further from goal the patient is, the greater the effect the medication will have. As HbA1c inches closer to 7%, the effect diminishes. And when 2 oral agents fail to lower a patient’s HbA1c adequately, the incremental change expected from the addition of a third, fourth, or fifth agent is small.

Thus, in a patient with an HbA1c ≥8%, there is still a significant fasting hyperglycemic component. In such a case, a basal insulin is likely the best treatment option.

Not at goal at one year. Both the AACE and the ADA/EASD guidelines agree that treatment titration should be considered every 2 to 3 months to achieve metabolic control and that if a patient is not at goal after a year, insulin should be started.^3,4 However, traditionally this is not done. The delayed implementation of this recommendation is an example of clinical inertia, which can contribute to further misunderstandings about the role and effect of insulin therapy.

Getting started with basal insulin

Most patients who are started on insulin have global hyperglycemia. But because fasting hyperglycemia can affect pancreatic insulin secretion, it is important to get control of the fasting glucose first. This can often be done with insulin sensitizers (metformin, thiazolidinediones, and incretin-based agents).

Suppression of excessive hepatic glucose production, which is very common in type 2 diabetes, is one of the biggest challenges in normalizing fasting glucose. This is well managed with a basal insulin. When starting basal insulin, however, it is critical that current treatments not be stopped. Oral agents such as sulfonylureas and meglitinides can be reduced to lower the risk of hypoglycemia, but stopping them altogether will only prolong the time it takes to get to goal.

There are 3 insulin formulations that can serve as basal insulin (TABLE 1).¹³ Neutral protamine Hagedorn (NPH) is a human insulin that can be used 2 to 3 times daily to provide basal insulin coverage. But long-acting basal analog insulins glargine and detemir, typically administered once a day when used by patients with type 2 diabetes, are a better option.¹⁴

In my experience, adherence to an insulin regimen is much greater if the first injection is administered in an office setting.While all 3 formulations have similar efficacy for lowering HbA1c, the analog basal insulins have numerous advantages: less weight gain, less hypoglycemia for the same level of glucose control, and less frequent dosing. In addition, glargine and detemir are available in a pen or vial, while generic NPH is available only in a vial. The primary disadvantage of the analogs is cost: A month’s supply—one vial—of NPH sells for approximately $25 (generic) or $94 (brand name); in comparison, a month’s supply (one box of 5 3-mL pens) of detemir and glargine costs about $300 and $320, respectively.¹⁵ (Humulin N, a brand-name NPH, is available in a pen, at a cost of approximately $315 per box.)

Use a weight-based initial dose

The recommended starting dose is 0.1 to 0.2 U/kg daily for patients with an HbA1c <8%. If HbA1c is ≥8%, the ADA/EASD guidelines recommend a starting dose of 0.3 to 0.4 U/kg daily⁴(TABLE 2).^3,4,16 While basal insulin is most commonly dosed at bedtime, in fact, basal analog insulins can be given at any time that’s convenient for the patient. Morning dosing may be preferable for individuals with a significant fear of hypoglycemia—a phobia that sometimes causes patients to skip insulin doses and engage in “defensive eating” (ie, eating in an attempt to prevent hypoglycemia rather than because of hunger or the need for nutrition).

Teach injection technique

It is critically important that patients get the first shot in the office, guided by a clinician who can teach proper injection technique. This also helps to dispel the apprehension of self-injection.

In addition to being surprised at how easy and painless injection can be, patients have the opportunity to observe the results and gain confidence in insulin’s efficacy. And, in my experience, adherence to an insulin regimen is much greater if the first injection is administered in an office setting.

(Tech-savvy patients may find it helpful to use a smartphone app, such as Glucose Buddy or Dbees.com, to help manage their diabetes. See “The 13 best diabetes iPhone & Android apps of 2013” at http://www.healthline.com/health-slideshow/top-iphone-android-apps-diabetes.)

Establish a titration schedule

It is important, too, to teach the patient how to titrate the insulin dose from the start, rather than waiting until the next visit to address this. Patient titration—facilitated by a clinician-provided titration schedule (available from the AACE and the ADA/EASD^3,4)—has been shown to achieve target glucose levels faster than physician titration.¹⁷

Teach the patient how to titrate the insulin dose from the start, rather than waiting until the next visit to address this.I usually suggest that patients increase the basal insulin dose by 3 units every 3 days, with an upper limit of 0.5 U/kg/d, until fasting glucose is consistently between 100 and 150 mg/dL. I advise every patient who starts taking insulin to track morning readings and titrate the dose until one of 3 things occurs:
1) the 0.5 U/kg/d limit is reached;
2) the patient has a glucose reading <100 mg/dL; or
3) the patient achieves his or her HbA1c target (<7% for most patients).

In every case, I recommend that the patient call my office for further instruction.

If the patient has any low glucose readings, I reduce the basal insulin by 5 U/kg/d. If he or she is still above goal, I advise the patient to continue titration, but more slowly. If the patient is at goal, I advise continuing at the current dose.

Basal titration vs mealtime coverage. Most people with type 2 diabetes require between 0.2 and 1 U/kg of basal insulin daily. It is currently recommended that when a patient has titrated to a dose of 0.5 U/kg/d, it is time to look at the glucose pattern to determine whether further titrating basal insulin or addressing prandial hyperglycemia should be the next step.^4,18 This requires a change in fingerstick pattern.

The patient can stop the first morning glucose check and start checking before meals and 90 to 120 minutes postmeal. This allows for exploration of the mealtime excursion. Generally, a difference of <50 mg/dL is preferred. If the morning glucose level is at target but HbA1c is high, it is likely that postprandial glucose is contributing to this difference. This is particularly true when the HbA1c is between 7% and 8%. If the glucose pattern shows high postmeal glucose readings, it is much safer to address mealtime insulin (not discussed in this article) than to continue to titrate the basal insulin.^4,18

Avoid “overbasalization”—ie, titrating basal insulin beyond its normal role to suppress hepatic glucose production and get the fasting glucose to goal. Doing so puts the patient at risk for unexpected hypoglycemia, as the insulin will now try to overcome hyperglycemia with meals, as well. Basal insulins are not designed to meet insulin requirements at meals. If a patient misses a meal yet continues the same dose of basal insulin, the risk of a hypoglycemic episode increases substantially.

In the pipeline. There are a number of new basal insulins in development, including one that has a prolonged duration of action and the potential for every-other-day injections¹⁹ and another that uses an attached polyethylene glycol moiety to slow absorption and prolong its effect.²⁰

The nuts and bolts of insulin prescribing

When you prescribe insulin, there are a number of components to consider.

Pen or vial? In addition to deciding whether to order pen or vial, it is essential to consider the volume of insulin needed. Glargine, detemir, and Humulin N are available in 10 mL vials (100 U/mL) and in 3 mL pens (100 U/mL). (Generic NPH is available in vials only.) Most patients prefer insulin pens, which are more convenient and easier to use than a vial and syringe.

The choice also depends on the dosage, however. A patient on a daily dose of 45 units would need one box of 5 pens (each prefilled pen has a 3 mL, or 300 unit, capacity) to have sufficient insulin for a month. Vials would be preferable for an individual who requires a larger single dose than a pen can dispense at one time (80 units of glargine, 60 units of detemir).

Syringe and needle size. If you are ordering insulin vials, you will also need to specify the correct syringe—available in 0.3 mL (which holds 30 units), 0.5 mL (50 units), and 1 mL (100 units) sizes. If the patient requires <50 units, order a small syringe to ensure that the unit markings are clear; a 1 mL syringe is preferable for those using a larger volume of insulin. Order the smallest syringe, which also has half-unit markings, if the patient is a child.

All needles are fine, with a 29 to 31 gauge, and available in regular (12.7 mm), short (8 mm), mini (5 mm), and nano (4 mm). Recent studies have shown that absorption, safety, and adverse events are similar for all needle lengths across a variety of patient factors,²¹ but patients generally prefer shorter needles.

Remember, too, to specify the maximum daily dose of insulin—a consideration that will be more important when prescribing mealtime insulin but is worth mentioning here.

Recent studies have shown that absorption, safety, and adverse events are similar for all needle lengths across a wide variety of patient factors.Finally, tell patients who are getting started on insulin about www.accurateinsulin.org. Hosted by The Endocrine Society in partnership with the American Association of Diabetes Educators, ADA, American Pharmacists Association, and American College of Osteopathic Family Physicians, among other clinical groups, the Web site is designed to help patients (as well as providers) navigate the initiation and adjustment of insulin.¹⁸

CORRESPONDENCE
Jay Shubrook, DO, FAAFP, FACOFP, The Diabetes Institute at Ohio University, Athens, OH 45701; [email protected]

With type 2 diabetes now affecting 8.3% of the US population, most primary care physicians see patients with this disorder every day.¹ Based on the concurrent obesity epidemic, aging population, and emergence of type 2 diabetes in children and adolescents, it is estimated that by 2050, the prevalence will have risen from one in 12 Americans to one in 3.¹

Type 2 diabetes is a progressive disorder, with a relentless decline in beta cells. By the time of diagnosis, patients typically have lost at least 50% of insulin secretion; within 6 years of diagnosis, insulin secretion decreases to less than 25%.²

The American Association of Clinical Endocrinologists (AACE)³ and the American Diabetes Association/European Association for the Study of Diabetes (ADA/EASD)⁴ have recently published guidelines for the management of type 2 diabetes. While the AACE’s guidelines (available at https://www.aace.com/files/aace_algorithm.pdf) focus on different treatments at different stages of disease and both glycemic and nonglycemic benefits of treatment,³ the ADA/EASD’s guidelines (see http://care.diabetesjournals.org/content/early/2012/04/17/dc12-0413.full.pdf+html) emphasize a patient-centered approach, shared decision making, and individualization of treatment goals based on both patient preference and comorbid disease states.⁴

One thing both sets of guidelines have in common is a purposeful intensification of therapy every 2 to 3 months, as needed, and the introduction of insulin one year after diagnosis if the patient is still not at goal.^3,4 But all too often, this does not occur, particularly in primary care settings.

This article will review the “when” and “how” of insulin initiation. But first, a look at barriers to insulin therapy and evidence in support of earlier use.

Clinical inertia and patient fear are associated with delays

Both the AACE and the ADA/EASD guidelines agree that metformin is best used as early as possible.^5,6 With typical use, however, metformin fails to prevent the progression of diabetes, as measured by the climb of hemoglobin A1c (HbA1c), at a failure rate of about 17% of patients per year.⁵ Physicians have been slow to intensify treatment for type 2 diabetes⁶—a phenomenon referred to as clinical inertia. In a recent study, patients were out of control—with an HbA1c >8%—for an average of 4.6 years before insulin was initiated.

Typically, physicians adopt a stepwise approach, which often results in patients spending more than 10 years with an HbA1c >7% and 5 years with an HbA1c >8% before insulin is started.⁵ In a recent Veterans Administration study, patients were out of control, with an HbA1c >8%, for an average of 4.6 years before insulin was initiated.⁷

Both patient and physician factors contribute to the delay. Patient factors include the fear of injection, the belief that insulin will interfere with their lifestyle, and the idea that the use of insulin signifies impending complications or even death.⁸ But such beliefs are starting to change. In a recent multinational study of patients with type 2 diabetes, less than 20% stated they were unwilling to start insulin.⁹

For their part, primary care physicians are much less likely to prescribe insulin than clinicians specializing in diabetes.⁶ Physician-reported barriers to insulin initiation include the time required to train patients to use it correctly; the lack of support, including access to diabetes educators; and the absence of clear guidelines on the use of insulin.¹⁰

A case for earlier insulin

There has been recent momentum in favor of earlier initiation of insulin. In fact, some researchers regard intensive insulin as an excellent first treatment for type 2 diabetes,¹¹ based on the belief that early insulin (used for a brief time) can provide not only immediate improvement in glucose control, but also a lasting “legacy” effect. The ADA/EASD guidelines support the use of insulin as a first-line treatment for patients with symptoms of insulin deficiency,⁴ but do not recommend it for everyone with newly diagnosed type 2 diabetes. 

Primary care physicians are much less likely to prescribe insulin than clinicians specializing in diabetes.There have also been a number of advances in insulin therapy over the past 2 decades. These include insulin analogs with physiologic profiles that better match daily schedules, as well as improvements in the way insulin is delivered. Insulin pens, smaller needles, disposable devices, and insulin pumps have made it easier to administer and fine-tune insulin delivery. Despite these improvements and recommendations for earlier implementation, the use of insulin in type 2 diabetes is significantly lower today than in the 1990s.¹²

When to introduce insulin

Insulin is indicated for patients with type 2 diabetes whose disease is not easily controlled. That includes individuals with decompensated type 2 diabetes, those whose HbA1c remains high even with 2 or more oral agents, and individuals who have not reached goal after a year of treatment.

Glucose toxicity. It is generally agreed that insulin is the most effective treatment for patients who present with decompensated type 2 diabetes⁴—ie, with significant hyperglycemia and catabolic symptoms such as polydipsia, polyuria, and weight loss. Initiation of insulin promotes reversal of glucose toxicity and stabilization of metabolic status. In such cases, insulin can be started at a low dose to expose the patient to the complexities of injection therapy (more about this in a bit), then titrated as needed for stabilization.

HbA1c ≥8% even with 2 or more drugs. In my experience, an oral diabetes drug will lead to a drop in HbA1c of about one percentage point. Generally, the further from goal the patient is, the greater the effect the medication will have. As HbA1c inches closer to 7%, the effect diminishes. And when 2 oral agents fail to lower a patient’s HbA1c adequately, the incremental change expected from the addition of a third, fourth, or fifth agent is small.

Thus, in a patient with an HbA1c ≥8%, there is still a significant fasting hyperglycemic component. In such a case, a basal insulin is likely the best treatment option.

Not at goal at one year. Both the AACE and the ADA/EASD guidelines agree that treatment titration should be considered every 2 to 3 months to achieve metabolic control and that if a patient is not at goal after a year, insulin should be started.^3,4 However, traditionally this is not done. The delayed implementation of this recommendation is an example of clinical inertia, which can contribute to further misunderstandings about the role and effect of insulin therapy.

Getting started with basal insulin

Most patients who are started on insulin have global hyperglycemia. But because fasting hyperglycemia can affect pancreatic insulin secretion, it is important to get control of the fasting glucose first. This can often be done with insulin sensitizers (metformin, thiazolidinediones, and incretin-based agents).

Suppression of excessive hepatic glucose production, which is very common in type 2 diabetes, is one of the biggest challenges in normalizing fasting glucose. This is well managed with a basal insulin. When starting basal insulin, however, it is critical that current treatments not be stopped. Oral agents such as sulfonylureas and meglitinides can be reduced to lower the risk of hypoglycemia, but stopping them altogether will only prolong the time it takes to get to goal.

There are 3 insulin formulations that can serve as basal insulin (TABLE 1).¹³ Neutral protamine Hagedorn (NPH) is a human insulin that can be used 2 to 3 times daily to provide basal insulin coverage. But long-acting basal analog insulins glargine and detemir, typically administered once a day when used by patients with type 2 diabetes, are a better option.¹⁴

In my experience, adherence to an insulin regimen is much greater if the first injection is administered in an office setting.While all 3 formulations have similar efficacy for lowering HbA1c, the analog basal insulins have numerous advantages: less weight gain, less hypoglycemia for the same level of glucose control, and less frequent dosing. In addition, glargine and detemir are available in a pen or vial, while generic NPH is available only in a vial. The primary disadvantage of the analogs is cost: A month’s supply—one vial—of NPH sells for approximately $25 (generic) or $94 (brand name); in comparison, a month’s supply (one box of 5 3-mL pens) of detemir and glargine costs about $300 and $320, respectively.¹⁵ (Humulin N, a brand-name NPH, is available in a pen, at a cost of approximately $315 per box.)

Use a weight-based initial dose

The recommended starting dose is 0.1 to 0.2 U/kg daily for patients with an HbA1c <8%. If HbA1c is ≥8%, the ADA/EASD guidelines recommend a starting dose of 0.3 to 0.4 U/kg daily⁴(TABLE 2).^3,4,16 While basal insulin is most commonly dosed at bedtime, in fact, basal analog insulins can be given at any time that’s convenient for the patient. Morning dosing may be preferable for individuals with a significant fear of hypoglycemia—a phobia that sometimes causes patients to skip insulin doses and engage in “defensive eating” (ie, eating in an attempt to prevent hypoglycemia rather than because of hunger or the need for nutrition).

Teach injection technique

It is critically important that patients get the first shot in the office, guided by a clinician who can teach proper injection technique. This also helps to dispel the apprehension of self-injection.

In addition to being surprised at how easy and painless injection can be, patients have the opportunity to observe the results and gain confidence in insulin’s efficacy. And, in my experience, adherence to an insulin regimen is much greater if the first injection is administered in an office setting.

(Tech-savvy patients may find it helpful to use a smartphone app, such as Glucose Buddy or Dbees.com, to help manage their diabetes. See “The 13 best diabetes iPhone & Android apps of 2013” at http://www.healthline.com/health-slideshow/top-iphone-android-apps-diabetes.)

Establish a titration schedule

It is important, too, to teach the patient how to titrate the insulin dose from the start, rather than waiting until the next visit to address this. Patient titration—facilitated by a clinician-provided titration schedule (available from the AACE and the ADA/EASD^3,4)—has been shown to achieve target glucose levels faster than physician titration.¹⁷

Teach the patient how to titrate the insulin dose from the start, rather than waiting until the next visit to address this.I usually suggest that patients increase the basal insulin dose by 3 units every 3 days, with an upper limit of 0.5 U/kg/d, until fasting glucose is consistently between 100 and 150 mg/dL. I advise every patient who starts taking insulin to track morning readings and titrate the dose until one of 3 things occurs:
1) the 0.5 U/kg/d limit is reached;
2) the patient has a glucose reading <100 mg/dL; or
3) the patient achieves his or her HbA1c target (<7% for most patients).

In every case, I recommend that the patient call my office for further instruction.

If the patient has any low glucose readings, I reduce the basal insulin by 5 U/kg/d. If he or she is still above goal, I advise the patient to continue titration, but more slowly. If the patient is at goal, I advise continuing at the current dose.

Basal titration vs mealtime coverage. Most people with type 2 diabetes require between 0.2 and 1 U/kg of basal insulin daily. It is currently recommended that when a patient has titrated to a dose of 0.5 U/kg/d, it is time to look at the glucose pattern to determine whether further titrating basal insulin or addressing prandial hyperglycemia should be the next step.^4,18 This requires a change in fingerstick pattern.

The patient can stop the first morning glucose check and start checking before meals and 90 to 120 minutes postmeal. This allows for exploration of the mealtime excursion. Generally, a difference of <50 mg/dL is preferred. If the morning glucose level is at target but HbA1c is high, it is likely that postprandial glucose is contributing to this difference. This is particularly true when the HbA1c is between 7% and 8%. If the glucose pattern shows high postmeal glucose readings, it is much safer to address mealtime insulin (not discussed in this article) than to continue to titrate the basal insulin.^4,18

Avoid “overbasalization”—ie, titrating basal insulin beyond its normal role to suppress hepatic glucose production and get the fasting glucose to goal. Doing so puts the patient at risk for unexpected hypoglycemia, as the insulin will now try to overcome hyperglycemia with meals, as well. Basal insulins are not designed to meet insulin requirements at meals. If a patient misses a meal yet continues the same dose of basal insulin, the risk of a hypoglycemic episode increases substantially.

In the pipeline. There are a number of new basal insulins in development, including one that has a prolonged duration of action and the potential for every-other-day injections¹⁹ and another that uses an attached polyethylene glycol moiety to slow absorption and prolong its effect.²⁰

The nuts and bolts of insulin prescribing

When you prescribe insulin, there are a number of components to consider.

Pen or vial? In addition to deciding whether to order pen or vial, it is essential to consider the volume of insulin needed. Glargine, detemir, and Humulin N are available in 10 mL vials (100 U/mL) and in 3 mL pens (100 U/mL). (Generic NPH is available in vials only.) Most patients prefer insulin pens, which are more convenient and easier to use than a vial and syringe.

The choice also depends on the dosage, however. A patient on a daily dose of 45 units would need one box of 5 pens (each prefilled pen has a 3 mL, or 300 unit, capacity) to have sufficient insulin for a month. Vials would be preferable for an individual who requires a larger single dose than a pen can dispense at one time (80 units of glargine, 60 units of detemir).

Syringe and needle size. If you are ordering insulin vials, you will also need to specify the correct syringe—available in 0.3 mL (which holds 30 units), 0.5 mL (50 units), and 1 mL (100 units) sizes. If the patient requires <50 units, order a small syringe to ensure that the unit markings are clear; a 1 mL syringe is preferable for those using a larger volume of insulin. Order the smallest syringe, which also has half-unit markings, if the patient is a child.

All needles are fine, with a 29 to 31 gauge, and available in regular (12.7 mm), short (8 mm), mini (5 mm), and nano (4 mm). Recent studies have shown that absorption, safety, and adverse events are similar for all needle lengths across a variety of patient factors,²¹ but patients generally prefer shorter needles.

Remember, too, to specify the maximum daily dose of insulin—a consideration that will be more important when prescribing mealtime insulin but is worth mentioning here.

Recent studies have shown that absorption, safety, and adverse events are similar for all needle lengths across a wide variety of patient factors.Finally, tell patients who are getting started on insulin about www.accurateinsulin.org. Hosted by The Endocrine Society in partnership with the American Association of Diabetes Educators, ADA, American Pharmacists Association, and American College of Osteopathic Family Physicians, among other clinical groups, the Web site is designed to help patients (as well as providers) navigate the initiation and adjustment of insulin.¹⁸

CORRESPONDENCE
Jay Shubrook, DO, FAAFP, FACOFP, The Diabetes Institute at Ohio University, Athens, OH 45701; [email protected]

With type 2 diabetes now affecting 8.3% of the US population, most primary care physicians see patients with this disorder every day.¹ Based on the concurrent obesity epidemic, aging population, and emergence of type 2 diabetes in children and adolescents, it is estimated that by 2050, the prevalence will have risen from one in 12 Americans to one in 3.¹

Type 2 diabetes is a progressive disorder, with a relentless decline in beta cells. By the time of diagnosis, patients typically have lost at least 50% of insulin secretion; within 6 years of diagnosis, insulin secretion decreases to less than 25%.²

The American Association of Clinical Endocrinologists (AACE)³ and the American Diabetes Association/European Association for the Study of Diabetes (ADA/EASD)⁴ have recently published guidelines for the management of type 2 diabetes. While the AACE’s guidelines (available at https://www.aace.com/files/aace_algorithm.pdf) focus on different treatments at different stages of disease and both glycemic and nonglycemic benefits of treatment,³ the ADA/EASD’s guidelines (see http://care.diabetesjournals.org/content/early/2012/04/17/dc12-0413.full.pdf+html) emphasize a patient-centered approach, shared decision making, and individualization of treatment goals based on both patient preference and comorbid disease states.⁴

One thing both sets of guidelines have in common is a purposeful intensification of therapy every 2 to 3 months, as needed, and the introduction of insulin one year after diagnosis if the patient is still not at goal.^3,4 But all too often, this does not occur, particularly in primary care settings.

This article will review the “when” and “how” of insulin initiation. But first, a look at barriers to insulin therapy and evidence in support of earlier use.

Clinical inertia and patient fear are associated with delays

Both the AACE and the ADA/EASD guidelines agree that metformin is best used as early as possible.^5,6 With typical use, however, metformin fails to prevent the progression of diabetes, as measured by the climb of hemoglobin A1c (HbA1c), at a failure rate of about 17% of patients per year.⁵ Physicians have been slow to intensify treatment for type 2 diabetes⁶—a phenomenon referred to as clinical inertia. In a recent study, patients were out of control—with an HbA1c >8%—for an average of 4.6 years before insulin was initiated.

Typically, physicians adopt a stepwise approach, which often results in patients spending more than 10 years with an HbA1c >7% and 5 years with an HbA1c >8% before insulin is started.⁵ In a recent Veterans Administration study, patients were out of control, with an HbA1c >8%, for an average of 4.6 years before insulin was initiated.⁷

Both patient and physician factors contribute to the delay. Patient factors include the fear of injection, the belief that insulin will interfere with their lifestyle, and the idea that the use of insulin signifies impending complications or even death.⁸ But such beliefs are starting to change. In a recent multinational study of patients with type 2 diabetes, less than 20% stated they were unwilling to start insulin.⁹

For their part, primary care physicians are much less likely to prescribe insulin than clinicians specializing in diabetes.⁶ Physician-reported barriers to insulin initiation include the time required to train patients to use it correctly; the lack of support, including access to diabetes educators; and the absence of clear guidelines on the use of insulin.¹⁰

A case for earlier insulin

There has been recent momentum in favor of earlier initiation of insulin. In fact, some researchers regard intensive insulin as an excellent first treatment for type 2 diabetes,¹¹ based on the belief that early insulin (used for a brief time) can provide not only immediate improvement in glucose control, but also a lasting “legacy” effect. The ADA/EASD guidelines support the use of insulin as a first-line treatment for patients with symptoms of insulin deficiency,⁴ but do not recommend it for everyone with newly diagnosed type 2 diabetes. 

Primary care physicians are much less likely to prescribe insulin than clinicians specializing in diabetes.There have also been a number of advances in insulin therapy over the past 2 decades. These include insulin analogs with physiologic profiles that better match daily schedules, as well as improvements in the way insulin is delivered. Insulin pens, smaller needles, disposable devices, and insulin pumps have made it easier to administer and fine-tune insulin delivery. Despite these improvements and recommendations for earlier implementation, the use of insulin in type 2 diabetes is significantly lower today than in the 1990s.¹²

When to introduce insulin

Insulin is indicated for patients with type 2 diabetes whose disease is not easily controlled. That includes individuals with decompensated type 2 diabetes, those whose HbA1c remains high even with 2 or more oral agents, and individuals who have not reached goal after a year of treatment.

Glucose toxicity. It is generally agreed that insulin is the most effective treatment for patients who present with decompensated type 2 diabetes⁴—ie, with significant hyperglycemia and catabolic symptoms such as polydipsia, polyuria, and weight loss. Initiation of insulin promotes reversal of glucose toxicity and stabilization of metabolic status. In such cases, insulin can be started at a low dose to expose the patient to the complexities of injection therapy (more about this in a bit), then titrated as needed for stabilization.

HbA1c ≥8% even with 2 or more drugs. In my experience, an oral diabetes drug will lead to a drop in HbA1c of about one percentage point. Generally, the further from goal the patient is, the greater the effect the medication will have. As HbA1c inches closer to 7%, the effect diminishes. And when 2 oral agents fail to lower a patient’s HbA1c adequately, the incremental change expected from the addition of a third, fourth, or fifth agent is small.

Thus, in a patient with an HbA1c ≥8%, there is still a significant fasting hyperglycemic component. In such a case, a basal insulin is likely the best treatment option.

Not at goal at one year. Both the AACE and the ADA/EASD guidelines agree that treatment titration should be considered every 2 to 3 months to achieve metabolic control and that if a patient is not at goal after a year, insulin should be started.^3,4 However, traditionally this is not done. The delayed implementation of this recommendation is an example of clinical inertia, which can contribute to further misunderstandings about the role and effect of insulin therapy.

Getting started with basal insulin

Most patients who are started on insulin have global hyperglycemia. But because fasting hyperglycemia can affect pancreatic insulin secretion, it is important to get control of the fasting glucose first. This can often be done with insulin sensitizers (metformin, thiazolidinediones, and incretin-based agents).

Suppression of excessive hepatic glucose production, which is very common in type 2 diabetes, is one of the biggest challenges in normalizing fasting glucose. This is well managed with a basal insulin. When starting basal insulin, however, it is critical that current treatments not be stopped. Oral agents such as sulfonylureas and meglitinides can be reduced to lower the risk of hypoglycemia, but stopping them altogether will only prolong the time it takes to get to goal.

There are 3 insulin formulations that can serve as basal insulin (TABLE 1).¹³ Neutral protamine Hagedorn (NPH) is a human insulin that can be used 2 to 3 times daily to provide basal insulin coverage. But long-acting basal analog insulins glargine and detemir, typically administered once a day when used by patients with type 2 diabetes, are a better option.¹⁴

In my experience, adherence to an insulin regimen is much greater if the first injection is administered in an office setting.While all 3 formulations have similar efficacy for lowering HbA1c, the analog basal insulins have numerous advantages: less weight gain, less hypoglycemia for the same level of glucose control, and less frequent dosing. In addition, glargine and detemir are available in a pen or vial, while generic NPH is available only in a vial. The primary disadvantage of the analogs is cost: A month’s supply—one vial—of NPH sells for approximately $25 (generic) or $94 (brand name); in comparison, a month’s supply (one box of 5 3-mL pens) of detemir and glargine costs about $300 and $320, respectively.¹⁵ (Humulin N, a brand-name NPH, is available in a pen, at a cost of approximately $315 per box.)

Use a weight-based initial dose

The recommended starting dose is 0.1 to 0.2 U/kg daily for patients with an HbA1c <8%. If HbA1c is ≥8%, the ADA/EASD guidelines recommend a starting dose of 0.3 to 0.4 U/kg daily⁴(TABLE 2).^3,4,16 While basal insulin is most commonly dosed at bedtime, in fact, basal analog insulins can be given at any time that’s convenient for the patient. Morning dosing may be preferable for individuals with a significant fear of hypoglycemia—a phobia that sometimes causes patients to skip insulin doses and engage in “defensive eating” (ie, eating in an attempt to prevent hypoglycemia rather than because of hunger or the need for nutrition).

Teach injection technique

It is critically important that patients get the first shot in the office, guided by a clinician who can teach proper injection technique. This also helps to dispel the apprehension of self-injection.

In addition to being surprised at how easy and painless injection can be, patients have the opportunity to observe the results and gain confidence in insulin’s efficacy. And, in my experience, adherence to an insulin regimen is much greater if the first injection is administered in an office setting.

(Tech-savvy patients may find it helpful to use a smartphone app, such as Glucose Buddy or Dbees.com, to help manage their diabetes. See “The 13 best diabetes iPhone & Android apps of 2013” at http://www.healthline.com/health-slideshow/top-iphone-android-apps-diabetes.)

Establish a titration schedule

It is important, too, to teach the patient how to titrate the insulin dose from the start, rather than waiting until the next visit to address this. Patient titration—facilitated by a clinician-provided titration schedule (available from the AACE and the ADA/EASD^3,4)—has been shown to achieve target glucose levels faster than physician titration.¹⁷

Teach the patient how to titrate the insulin dose from the start, rather than waiting until the next visit to address this.I usually suggest that patients increase the basal insulin dose by 3 units every 3 days, with an upper limit of 0.5 U/kg/d, until fasting glucose is consistently between 100 and 150 mg/dL. I advise every patient who starts taking insulin to track morning readings and titrate the dose until one of 3 things occurs:
1) the 0.5 U/kg/d limit is reached;
2) the patient has a glucose reading <100 mg/dL; or
3) the patient achieves his or her HbA1c target (<7% for most patients).

In every case, I recommend that the patient call my office for further instruction.

If the patient has any low glucose readings, I reduce the basal insulin by 5 U/kg/d. If he or she is still above goal, I advise the patient to continue titration, but more slowly. If the patient is at goal, I advise continuing at the current dose.

Basal titration vs mealtime coverage. Most people with type 2 diabetes require between 0.2 and 1 U/kg of basal insulin daily. It is currently recommended that when a patient has titrated to a dose of 0.5 U/kg/d, it is time to look at the glucose pattern to determine whether further titrating basal insulin or addressing prandial hyperglycemia should be the next step.^4,18 This requires a change in fingerstick pattern.

The patient can stop the first morning glucose check and start checking before meals and 90 to 120 minutes postmeal. This allows for exploration of the mealtime excursion. Generally, a difference of <50 mg/dL is preferred. If the morning glucose level is at target but HbA1c is high, it is likely that postprandial glucose is contributing to this difference. This is particularly true when the HbA1c is between 7% and 8%. If the glucose pattern shows high postmeal glucose readings, it is much safer to address mealtime insulin (not discussed in this article) than to continue to titrate the basal insulin.^4,18

Avoid “overbasalization”—ie, titrating basal insulin beyond its normal role to suppress hepatic glucose production and get the fasting glucose to goal. Doing so puts the patient at risk for unexpected hypoglycemia, as the insulin will now try to overcome hyperglycemia with meals, as well. Basal insulins are not designed to meet insulin requirements at meals. If a patient misses a meal yet continues the same dose of basal insulin, the risk of a hypoglycemic episode increases substantially.

In the pipeline. There are a number of new basal insulins in development, including one that has a prolonged duration of action and the potential for every-other-day injections¹⁹ and another that uses an attached polyethylene glycol moiety to slow absorption and prolong its effect.²⁰

The nuts and bolts of insulin prescribing

When you prescribe insulin, there are a number of components to consider.

Pen or vial? In addition to deciding whether to order pen or vial, it is essential to consider the volume of insulin needed. Glargine, detemir, and Humulin N are available in 10 mL vials (100 U/mL) and in 3 mL pens (100 U/mL). (Generic NPH is available in vials only.) Most patients prefer insulin pens, which are more convenient and easier to use than a vial and syringe.

The choice also depends on the dosage, however. A patient on a daily dose of 45 units would need one box of 5 pens (each prefilled pen has a 3 mL, or 300 unit, capacity) to have sufficient insulin for a month. Vials would be preferable for an individual who requires a larger single dose than a pen can dispense at one time (80 units of glargine, 60 units of detemir).

Syringe and needle size. If you are ordering insulin vials, you will also need to specify the correct syringe—available in 0.3 mL (which holds 30 units), 0.5 mL (50 units), and 1 mL (100 units) sizes. If the patient requires <50 units, order a small syringe to ensure that the unit markings are clear; a 1 mL syringe is preferable for those using a larger volume of insulin. Order the smallest syringe, which also has half-unit markings, if the patient is a child.

All needles are fine, with a 29 to 31 gauge, and available in regular (12.7 mm), short (8 mm), mini (5 mm), and nano (4 mm). Recent studies have shown that absorption, safety, and adverse events are similar for all needle lengths across a variety of patient factors,²¹ but patients generally prefer shorter needles.

Remember, too, to specify the maximum daily dose of insulin—a consideration that will be more important when prescribing mealtime insulin but is worth mentioning here.

Recent studies have shown that absorption, safety, and adverse events are similar for all needle lengths across a wide variety of patient factors.Finally, tell patients who are getting started on insulin about www.accurateinsulin.org. Hosted by The Endocrine Society in partnership with the American Association of Diabetes Educators, ADA, American Pharmacists Association, and American College of Osteopathic Family Physicians, among other clinical groups, the Web site is designed to help patients (as well as providers) navigate the initiation and adjustment of insulin.¹⁸

CORRESPONDENCE
Jay Shubrook, DO, FAAFP, FACOFP, The Diabetes Institute at Ohio University, Athens, OH 45701; [email protected]

CASE 1 Sally J is a 72-year-old Caucasian woman who comes to your clinic after being diagnosed with atrial fibrillation (AF). The patient has a 10-year history of type 2 diabetes; she also has a history of hypertension and chronic kidney disease (CKD), with a baseline creatinine clearance of approximately 40 mL/min. Ms. J tells you she knows people who take warfarin and really dislike it. She asks for your opinion of the new anticoagulants she’s seen advertised on TV, and wonders whether one of them would be right for her.

CASE 2 Bobby W, a 35-year-old African American man, was recently diagnosed with deep vein thrombosis (DVT). This was his second clot in 5 years, and occurred after a long flight home from Europe. The patient explains that he leads a very active lifestyle and doesn’t have the time to come in for the monthly international normalized ratio (INR) checks that warfarin requires. What would you recommend for these patients?

Troubled by warfarin’s narrow therapeutic index, numerous medication and dietary interactions, and need for frequent monitoring, patients requiring long-term oral anticoagulation therapy have been seeking alternatives for years. Finally, they have a choice. The US Food and Drug Administration (FDA) approved 3 oral anticoagulants—dabigatran (Pradaxa), rivaroxaban (Xarelto), and apixaban (Eliquis)—in less than 4 years. Known as novel oral anticoagulants (NOACs), they are the first such drugs to enter the market in more than 50 years.^1,2

While warfarin inhibits a wide range of clotting factors (including II, VII, IX, and X), NOACs work further down the clotting cascade (TABLE 1).^1,3-7 Dabigatran, a direct thrombin inhibitor, only inhibits factor IIa.^3,5 Rivaroxaban and apixaban directly inhibit factor Xa and indirectly inhibit factor IIa.^3,6,7

There are notable advantages to these newer agents, but some disadvantages that must be considered, as well. Appropriate patient selection, guided by a thorough understanding of the benefits and risks of NOACs, is key.

Stroke prevention in atrial fibrillation

All 3 NOACs are approved for stroke prevention in patients with nonvalvular atrial fibrillation (AF). The approvals are based on a small number of well-designed trials: RE-LY (dabigatran), ROCKET-AF (rivaroxaban), and ARISTOTLE (apixaban).^8-10 Compared with warfarin, dabigatran is the only oral anticoagulant with a lower rate of both hemorrhagic and ischemic stroke.⁸ Both rivaroxaban and apixaban were found to decrease overall stroke risk relative to warfarin, but the difference was driven by a lower risk for hemorrhagic, not ischemic, stroke.^9,10

In these trials, overall rates of major bleeding were similar to that of warfarin.^8-10 Patients taking warfarin generally experienced higher rates of intracranial hemorrhage but lower rates of gastrointestinal (GI) bleeding than those on NOACs. Relative to warfarin, apixaban was the only NOAC that did not have a higher rate of GI bleeding and the only one with a lower rate of major bleeding.^8-10 In addition, apixaban remains the only NOAC found to have a statistically significant decrease in all-cause mortality compared with warfarin.¹⁰ Although dabigatran and rivaroxaban were associated with a strong trend towards decreased mortality, both studies were underpowered for this secondary outcome.^8,9

Adding NOACs to stroke guidelines. The role of NOACs in the prevention of stroke in patients with nonvalvular AF is beginning to be reflected in newer guidelines. The American College of Chest Physicians (ACCP)’s 2012 guidelines recommend dabigatran over warfarin (grade 2B—weak recommendation; moderate quality evidence) unless the patient is well controlled on warfarin.¹¹ The European Society of Cardiology (ESC)’s 2012 guidelines recommend dabigatran, apixaban, and rivaroxaban as broadly preferable to warfarin, while noting that experience with these agents is limited and appropriate patient selection is important.¹²

Anticoagulation to treat—and prevent—VTE

The standard of care for acute venous thromboembolism (VTE) is to initiate warfarin along with a parenteral anticoagulant, such as unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.¹³ Due to warfarin’s slow onset to peak effect, a parenteral anticoagulant is overlapped for ≥5 days—until warfarin reaches a therapeutic level and can be continued as monotherapy.¹³ But many patients find subcutaneous delivery of LMWH disagreeable and costly and frequent INR monitoring inconvenient, so the new agents offer notable advantages.

In well-designed studies, dabigatran, rivaroxaban, and apixaban have all been shown to be noninferior to warfarin in the initial treatment of acute DVT and pulmonary embolism (PE).^14-17 All 3 agents were also shown to have lower rates of major bleeding than warfarin. Rivaroxaban and apixaban were also superior to warfarin with regard to bleeding events, and dabigatran was noninferior to warfarin for this outcome.^14-17

NOACs help prevent recurrence

All 3 NOACs have been studied for long-term prevention of recurrent VTE after 3 to 18 months of anticoagulation, as well. Dabigatran was found in the RE-MEDY trial to be noninferior to warfarin for the risk of recurrent VTE, and to have lower rates of bleeding.¹⁸ In separate trials, all 3 agents were superior to placebo in preventing recurrent VTE. Rates of long-term major bleeding were significantly higher than placebo with rivaroxaban and dabigatran, but not with apixaban.^15,18,19

Rivaroxaban is the only NOAC to be FDA approved for the treatment of acute DVT and PE, and the ACCP’s 2012 guidelines list it as a viable alternative to parenteral anticoagulation when initiating treatment for acute VTE.^6,13 When treating VTE long term, the guidelines continue to recommend warfarin or LMWH rather than dabigatran or rivaroxaban.¹³ Recommendations may change in coming years, as physicians gain more experience with NOACs and more clinical trials are published.^16-19

Starting or converting to NOAC therapy

In patients who have not been on anticoagulant therapy, any NOAC can be initiated immediately, with no need for parenteral, or “bridge” therapy. This is because of the rapid onset of action of the NOACs.¹²

To transition a patient from warfarin to a NOAC, it is necessary to discontinue warfarin therapy completely and closely monitor INR, then initiate NOAC therapy when INR≤2. No parenteral anticoagulation is necessary (TABLE 2).^5-7

If it is necessary to transition a patient from a NOAC to warfarin, the protocol depends on the agent. Because dabigatran has no significant impact on prolongation of prothrombin time (PT), it can be overlapped with warfarin. Rivaroxaban and apixaban have a significant impact on PT prolongation, however, and overlapping either agent with warfarin is not recommended.⁴ Keep in mind that the recommended dosages for the NOACs are not standardized, and can differ drastically depending on the indication for use as well as on patient-specific factors, including renal function, body weight, and age.

Laboratory monitoring is not required

While warfarin has a great deal of interpatient variability and requires frequent lab monitoring, an oft-cited advantage of the NOACs is that they do not require regular monitoring. However, that also has a downside (TABLE 3).^{1,3-10,12,14-17,20-22} Monitoring INR in patients on warfarin allows physicians to assess patient compliance. And, if a patient on warfarin requires an invasive procedure, coagulation status and bleeding risk can easily be determined. That is not the case with the NOACs.

While some routine laboratory tests may be elevated in a patient taking a NOAC, the degree of elevation does not correlate well with anticoagulant concentration. And, because each NOAC has a different mechanism of action, different measures will be elevated in a patient taking dabigatran vs apixaban or rivaroxaban.⁴

Activated partial thromboplastin time (aPTT) is the most readily available lab test to assess the presence or absence of dabigatran.⁴ A normal aPTT indicates that there is little to no dabigatran present.⁴ But, while an elevated aPTT suggests the presence of dabigatran, it provides little information about how much.⁴

PT is a useful test to assess coagulation status in patients on either rivaroxaban or apixaban. A normal PT suggests that minimal amounts (or none) of the NOAC are present in the plasma.⁴ (A direct thrombin inhibitor assay, calibrated to more accurately assess dabigatran concentration, is being developed for clinical use, but is currently available only for research purposes in the United States; a chromogenic antifactor Xa test specific to apixaban and rivaroxaban is also being developed, but is not yet commercially available.⁴)

What to do when NOAC reversal is required

Patients often need to stop taking an oral anticoagulation in the days leading up to a planned invasive procedure. In an individual with normal renal function who will undergo a procedure with a standard bleeding risk, a NOAC would generally need to be withheld for one to 2 days prior to surgery, given the relatively short half-life. If a patient has acute renal failure or CKD, however, dabigatran may need to be withheld for a prolonged period (3-6 days) in order to safely proceed to surgery.^4,23 NOACs may also need to be withheld for 2 to 6 days prior to any surgery with a high risk for bleeding.⁴

When speed is of the essenceIn patients who have not been on anticoagulant therapy, any NOAC can be initiated immediately, with no need for "bridge" therapy.

There is no known antidote to aid in the reversal of dabigatran, rivaroxaban, or apixaban.⁴ Because of their relatively short half-lives, withholding the medication and providing supportive care is generally sufficient to ensure adequate hemostasis in cases of mild to moderate bleeding.⁴ If a patient presents with acute ingestion or an overdose, activated charcoal should be administered if the ingestion has occurred within the past 3 hours.^4,24 The lack of a clear-cut reversal strategy can be extremely problematic in cases of trauma or life-threatening bleeding, however. (Fresh frozen plasma has not been shown to be effective at reversing NOACs’ effects.⁴)

In instances of severe bleeding or the need for urgent surgery, a more aggressive approach may be needed. Hemodialysis can be used to assist in the removal of dabigatran, but not rivaroxaban or apixaban.⁴ However, evidence suggests that the most effective therapy for patients who need rapid reversal of any NOAC is to administer 75 to 80 units/kg of activated prothrombin complex concentrate (aPCC).^4,25-27 Recombinant factor VIIa has shown some promise in reversing the anticoagulant effects of these novel agents, but evidence is insufficient to recommend it as first-line therapy at this time.^{26, 27}

Patients are more satisfied

A direct thrombin inhibitor assay, calibrated to more accurately assess dabigatran concentration, is being developed, as is a chromogenic antifactor Xa test specific to apixaban and rivaroxaban.The most obvious advantage of the NOACs as a group compared with warfarin is the lack of need for laboratory monitoring or continuous dose titration. Reliably stable pharmacokinetics make once or twice daily dosing possible. A rapid onset of action negates the need for bridging therapy with parenteral anticoagulants in patients at high risk of thrombosis. This may improve compliance, as many patients are averse to the use of subcutaneous injections or need extensive education before they can safely self-inject. The incidence of heparin-induced thrombocytopenia may also be decreased if unfractionated heparin and LMWH are used less frequently.

NOACs also appear to improve patient satisfaction.^20-22 In one study that included patients with AF on dabigatran or warfarin, satisfaction was higher in those taking dabigatran, particularly among those who did not experience significant GI adverse effects.²⁰ Another study showed improved patient satisfaction with rivaroxaban compared with LMWH following lower extremity joint replacement, which led to significantly higher rates of compliance.²²

… but problems and pitfalls remain

In addition to the lack of a readily available and clinically validated reversal agent, the absence of a lab test that reliably measures the concentration of NOACs makes it difficult to determine whether patients are following their prescribed regimen.^3,4

Medication compliance must be assessed when considering a transition from warfarin to a NOAC. Switching patients with poor INR control on warfarin to a NOAC should be done only after determining that the poor control is not the result of nonadherence. Because of the NOACs’ shorter half-life, patients who don’t take them regularly may be at higher risk for thromboembolic events.^1,12

Cost is a serious consideration. While there are some costs associated with the monitoring warfarin requires, the medication itself has been generic for several decades and can be found on many “$4 lists” at pharmacies nationwide. In contrast, all 3 NOACs are available only as branded drugs, and can cost a patient with limited drug coverage anywhere from $250 to $350 per month²⁸—a serious concern, given that the likelihood of noncompliance increases as out-of-pocket costs rise. This was highlighted in a recent study that found patients were twice as likely to discontinue their cholesterol-lowering medication if 100% of the cost was out of pocket, compared with patients who had no prescription copay.²⁹ From the perspective of the US health care system, however, NOACs have been found to be cost effective compared with warfarin, mostly due to the lack of laboratory monitoring.^30-32

Adverse effects. The risk of GI bleeds has been shown to be higher in patients taking rivaroxaban and dabigatran vs warfarin.^8,9 Dabigatran has also been associated with a significant risk for dyspepsia.^5,8,14 In clinical trials, the reported rate for dyspepsia in patients taking dabigatran was 3% to 11%; subsequent investigations have found the incidence to be far higher (33%).^8,14,33

Drug interactions. Warfarin has a large number of drug interactions, of course, but because of its long history, these interactions are well established. NOACs also have a number of drug interactions, but the true clinical impact has not yet been established. All 3 agents are substrates for the P-glycoprotein (P-gp) transport system, so any known inhibitors or inducers of the P-gp system should be used cautiously in patients on NOACs.^1,3-7,12 Rivaroxaban and apixaban are also substrates for the CYP3A4 hepatic enzyme system, so any drugs known to inhibit or induce this system require caution, as well.^1,3-7,12

Who should not take a NOAC?The most effective therapy for patients who need rapid reversal of any NOAC is to administer 75 to 80 unites/kg of activated prothrombin complex concentrate.

NOACs should not be prescribed for patients with mechanical heart valves.³⁴ Dabigatran is the only NOAC to have been studied in this patient population, and the phase II trial was stopped prematurely due to increased risk for both bleeding and stroke in patients on dabigatran compared with warfarin.³⁴

Renal impairment must be considered, as well. Do a baseline assessment of renal function in all patients before transitioning them to a NOAC, and periodic reassessment during therapy. While this is important for patients on rivaroxaban and apixaban, it is essential for those on dabigatran, as 80% of the drug is excreted by the kidneys.^1,5,12 NOACs have not been adequately assessed in patients with severe renal dysfunction and should be avoided in this patient population. Caution should be exercised in patients with moderate renal dysfunction, as well.^5-10,14-19 Apixaban appears to be the safest NOAC for patients with moderate renal dysfunction, as it has the least renal clearance.^1,12

Who should take a NOAC?

No well-established criteria for patient selection for NOACs exist, yet appropriate patient selection is crucial. Evidence suggests that NOAC therapy is best suited to those who:
• are relatively young (<65 years)
• have normal renal function
• have poorly controlled INR with warfarin that is unrelated to noncompliance
• are unable to have regular INR monitoring.

NOACs have not been adequately assessed in severe renal dysfunction and should be avoided in this patient population.Patients best suited for continued use of warfarin would be those whose INR is well controlled, those who have higher goal INR ranges (eg, because of the presence of mechanical heart valves), patients with significant renal dysfunction, and individuals with a history of peptic ulcer disease or GI bleeding. Warfarin may also be the best option for patients with a history of noncompliance and for uninsured or underinsured patients.

CASE 1 Warfarin and any of the NOACs were all feasible options for Ms. J, but apixaban was deemed to be the safest because of her moderate renal dysfunction. However, after she was told that apixaban has little “real world” clinical data, no effective antidote if bleeding were to occur, and a much higher cost than warfarin, she opted for warfarin therapy, despite the laboratory monitoring required.

CASE 2 Mr. W was excited to learn that there were new alternatives to warfarin; he had taken warfarin for 6 months after his last DVT and had a hard time coming in for INR checks. The patient reported that he had no history of bleeding and was compliant with medications. Rivaroxaban was the best option for Mr. W, as it is the only NOAC with FDA approval for the treatment of acute VTE.

CORRESPONDENCE
Jeremy Vandiver, PharmD, BCPS, Swedish Medical Center, Room 3260, 501 East Hampden Avenue, Englewood, CO 80013; [email protected]

CASE 1 Sally J is a 72-year-old Caucasian woman who comes to your clinic after being diagnosed with atrial fibrillation (AF). The patient has a 10-year history of type 2 diabetes; she also has a history of hypertension and chronic kidney disease (CKD), with a baseline creatinine clearance of approximately 40 mL/min. Ms. J tells you she knows people who take warfarin and really dislike it. She asks for your opinion of the new anticoagulants she’s seen advertised on TV, and wonders whether one of them would be right for her.

CASE 2 Bobby W, a 35-year-old African American man, was recently diagnosed with deep vein thrombosis (DVT). This was his second clot in 5 years, and occurred after a long flight home from Europe. The patient explains that he leads a very active lifestyle and doesn’t have the time to come in for the monthly international normalized ratio (INR) checks that warfarin requires. What would you recommend for these patients?

Troubled by warfarin’s narrow therapeutic index, numerous medication and dietary interactions, and need for frequent monitoring, patients requiring long-term oral anticoagulation therapy have been seeking alternatives for years. Finally, they have a choice. The US Food and Drug Administration (FDA) approved 3 oral anticoagulants—dabigatran (Pradaxa), rivaroxaban (Xarelto), and apixaban (Eliquis)—in less than 4 years. Known as novel oral anticoagulants (NOACs), they are the first such drugs to enter the market in more than 50 years.^1,2

While warfarin inhibits a wide range of clotting factors (including II, VII, IX, and X), NOACs work further down the clotting cascade (TABLE 1).^1,3-7 Dabigatran, a direct thrombin inhibitor, only inhibits factor IIa.^3,5 Rivaroxaban and apixaban directly inhibit factor Xa and indirectly inhibit factor IIa.^3,6,7

There are notable advantages to these newer agents, but some disadvantages that must be considered, as well. Appropriate patient selection, guided by a thorough understanding of the benefits and risks of NOACs, is key.

Stroke prevention in atrial fibrillation

All 3 NOACs are approved for stroke prevention in patients with nonvalvular atrial fibrillation (AF). The approvals are based on a small number of well-designed trials: RE-LY (dabigatran), ROCKET-AF (rivaroxaban), and ARISTOTLE (apixaban).^8-10 Compared with warfarin, dabigatran is the only oral anticoagulant with a lower rate of both hemorrhagic and ischemic stroke.⁸ Both rivaroxaban and apixaban were found to decrease overall stroke risk relative to warfarin, but the difference was driven by a lower risk for hemorrhagic, not ischemic, stroke.^9,10

In these trials, overall rates of major bleeding were similar to that of warfarin.^8-10 Patients taking warfarin generally experienced higher rates of intracranial hemorrhage but lower rates of gastrointestinal (GI) bleeding than those on NOACs. Relative to warfarin, apixaban was the only NOAC that did not have a higher rate of GI bleeding and the only one with a lower rate of major bleeding.^8-10 In addition, apixaban remains the only NOAC found to have a statistically significant decrease in all-cause mortality compared with warfarin.¹⁰ Although dabigatran and rivaroxaban were associated with a strong trend towards decreased mortality, both studies were underpowered for this secondary outcome.^8,9

Adding NOACs to stroke guidelines. The role of NOACs in the prevention of stroke in patients with nonvalvular AF is beginning to be reflected in newer guidelines. The American College of Chest Physicians (ACCP)’s 2012 guidelines recommend dabigatran over warfarin (grade 2B—weak recommendation; moderate quality evidence) unless the patient is well controlled on warfarin.¹¹ The European Society of Cardiology (ESC)’s 2012 guidelines recommend dabigatran, apixaban, and rivaroxaban as broadly preferable to warfarin, while noting that experience with these agents is limited and appropriate patient selection is important.¹²

Anticoagulation to treat—and prevent—VTE

The standard of care for acute venous thromboembolism (VTE) is to initiate warfarin along with a parenteral anticoagulant, such as unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.¹³ Due to warfarin’s slow onset to peak effect, a parenteral anticoagulant is overlapped for ≥5 days—until warfarin reaches a therapeutic level and can be continued as monotherapy.¹³ But many patients find subcutaneous delivery of LMWH disagreeable and costly and frequent INR monitoring inconvenient, so the new agents offer notable advantages.

In well-designed studies, dabigatran, rivaroxaban, and apixaban have all been shown to be noninferior to warfarin in the initial treatment of acute DVT and pulmonary embolism (PE).^14-17 All 3 agents were also shown to have lower rates of major bleeding than warfarin. Rivaroxaban and apixaban were also superior to warfarin with regard to bleeding events, and dabigatran was noninferior to warfarin for this outcome.^14-17

NOACs help prevent recurrence

All 3 NOACs have been studied for long-term prevention of recurrent VTE after 3 to 18 months of anticoagulation, as well. Dabigatran was found in the RE-MEDY trial to be noninferior to warfarin for the risk of recurrent VTE, and to have lower rates of bleeding.¹⁸ In separate trials, all 3 agents were superior to placebo in preventing recurrent VTE. Rates of long-term major bleeding were significantly higher than placebo with rivaroxaban and dabigatran, but not with apixaban.^15,18,19

Rivaroxaban is the only NOAC to be FDA approved for the treatment of acute DVT and PE, and the ACCP’s 2012 guidelines list it as a viable alternative to parenteral anticoagulation when initiating treatment for acute VTE.^6,13 When treating VTE long term, the guidelines continue to recommend warfarin or LMWH rather than dabigatran or rivaroxaban.¹³ Recommendations may change in coming years, as physicians gain more experience with NOACs and more clinical trials are published.^16-19

Starting or converting to NOAC therapy

In patients who have not been on anticoagulant therapy, any NOAC can be initiated immediately, with no need for parenteral, or “bridge” therapy. This is because of the rapid onset of action of the NOACs.¹²

To transition a patient from warfarin to a NOAC, it is necessary to discontinue warfarin therapy completely and closely monitor INR, then initiate NOAC therapy when INR≤2. No parenteral anticoagulation is necessary (TABLE 2).^5-7

If it is necessary to transition a patient from a NOAC to warfarin, the protocol depends on the agent. Because dabigatran has no significant impact on prolongation of prothrombin time (PT), it can be overlapped with warfarin. Rivaroxaban and apixaban have a significant impact on PT prolongation, however, and overlapping either agent with warfarin is not recommended.⁴ Keep in mind that the recommended dosages for the NOACs are not standardized, and can differ drastically depending on the indication for use as well as on patient-specific factors, including renal function, body weight, and age.

Laboratory monitoring is not required

While warfarin has a great deal of interpatient variability and requires frequent lab monitoring, an oft-cited advantage of the NOACs is that they do not require regular monitoring. However, that also has a downside (TABLE 3).^{1,3-10,12,14-17,20-22} Monitoring INR in patients on warfarin allows physicians to assess patient compliance. And, if a patient on warfarin requires an invasive procedure, coagulation status and bleeding risk can easily be determined. That is not the case with the NOACs.

While some routine laboratory tests may be elevated in a patient taking a NOAC, the degree of elevation does not correlate well with anticoagulant concentration. And, because each NOAC has a different mechanism of action, different measures will be elevated in a patient taking dabigatran vs apixaban or rivaroxaban.⁴

Activated partial thromboplastin time (aPTT) is the most readily available lab test to assess the presence or absence of dabigatran.⁴ A normal aPTT indicates that there is little to no dabigatran present.⁴ But, while an elevated aPTT suggests the presence of dabigatran, it provides little information about how much.⁴

PT is a useful test to assess coagulation status in patients on either rivaroxaban or apixaban. A normal PT suggests that minimal amounts (or none) of the NOAC are present in the plasma.⁴ (A direct thrombin inhibitor assay, calibrated to more accurately assess dabigatran concentration, is being developed for clinical use, but is currently available only for research purposes in the United States; a chromogenic antifactor Xa test specific to apixaban and rivaroxaban is also being developed, but is not yet commercially available.⁴)

What to do when NOAC reversal is required

Patients often need to stop taking an oral anticoagulation in the days leading up to a planned invasive procedure. In an individual with normal renal function who will undergo a procedure with a standard bleeding risk, a NOAC would generally need to be withheld for one to 2 days prior to surgery, given the relatively short half-life. If a patient has acute renal failure or CKD, however, dabigatran may need to be withheld for a prolonged period (3-6 days) in order to safely proceed to surgery.^4,23 NOACs may also need to be withheld for 2 to 6 days prior to any surgery with a high risk for bleeding.⁴

When speed is of the essenceIn patients who have not been on anticoagulant therapy, any NOAC can be initiated immediately, with no need for "bridge" therapy.

There is no known antidote to aid in the reversal of dabigatran, rivaroxaban, or apixaban.⁴ Because of their relatively short half-lives, withholding the medication and providing supportive care is generally sufficient to ensure adequate hemostasis in cases of mild to moderate bleeding.⁴ If a patient presents with acute ingestion or an overdose, activated charcoal should be administered if the ingestion has occurred within the past 3 hours.^4,24 The lack of a clear-cut reversal strategy can be extremely problematic in cases of trauma or life-threatening bleeding, however. (Fresh frozen plasma has not been shown to be effective at reversing NOACs’ effects.⁴)

In instances of severe bleeding or the need for urgent surgery, a more aggressive approach may be needed. Hemodialysis can be used to assist in the removal of dabigatran, but not rivaroxaban or apixaban.⁴ However, evidence suggests that the most effective therapy for patients who need rapid reversal of any NOAC is to administer 75 to 80 units/kg of activated prothrombin complex concentrate (aPCC).^4,25-27 Recombinant factor VIIa has shown some promise in reversing the anticoagulant effects of these novel agents, but evidence is insufficient to recommend it as first-line therapy at this time.^{26, 27}

Patients are more satisfied

A direct thrombin inhibitor assay, calibrated to more accurately assess dabigatran concentration, is being developed, as is a chromogenic antifactor Xa test specific to apixaban and rivaroxaban.The most obvious advantage of the NOACs as a group compared with warfarin is the lack of need for laboratory monitoring or continuous dose titration. Reliably stable pharmacokinetics make once or twice daily dosing possible. A rapid onset of action negates the need for bridging therapy with parenteral anticoagulants in patients at high risk of thrombosis. This may improve compliance, as many patients are averse to the use of subcutaneous injections or need extensive education before they can safely self-inject. The incidence of heparin-induced thrombocytopenia may also be decreased if unfractionated heparin and LMWH are used less frequently.

NOACs also appear to improve patient satisfaction.^20-22 In one study that included patients with AF on dabigatran or warfarin, satisfaction was higher in those taking dabigatran, particularly among those who did not experience significant GI adverse effects.²⁰ Another study showed improved patient satisfaction with rivaroxaban compared with LMWH following lower extremity joint replacement, which led to significantly higher rates of compliance.²²

… but problems and pitfalls remain

In addition to the lack of a readily available and clinically validated reversal agent, the absence of a lab test that reliably measures the concentration of NOACs makes it difficult to determine whether patients are following their prescribed regimen.^3,4

Medication compliance must be assessed when considering a transition from warfarin to a NOAC. Switching patients with poor INR control on warfarin to a NOAC should be done only after determining that the poor control is not the result of nonadherence. Because of the NOACs’ shorter half-life, patients who don’t take them regularly may be at higher risk for thromboembolic events.^1,12

Cost is a serious consideration. While there are some costs associated with the monitoring warfarin requires, the medication itself has been generic for several decades and can be found on many “$4 lists” at pharmacies nationwide. In contrast, all 3 NOACs are available only as branded drugs, and can cost a patient with limited drug coverage anywhere from $250 to $350 per month²⁸—a serious concern, given that the likelihood of noncompliance increases as out-of-pocket costs rise. This was highlighted in a recent study that found patients were twice as likely to discontinue their cholesterol-lowering medication if 100% of the cost was out of pocket, compared with patients who had no prescription copay.²⁹ From the perspective of the US health care system, however, NOACs have been found to be cost effective compared with warfarin, mostly due to the lack of laboratory monitoring.^30-32

Adverse effects. The risk of GI bleeds has been shown to be higher in patients taking rivaroxaban and dabigatran vs warfarin.^8,9 Dabigatran has also been associated with a significant risk for dyspepsia.^5,8,14 In clinical trials, the reported rate for dyspepsia in patients taking dabigatran was 3% to 11%; subsequent investigations have found the incidence to be far higher (33%).^8,14,33

Drug interactions. Warfarin has a large number of drug interactions, of course, but because of its long history, these interactions are well established. NOACs also have a number of drug interactions, but the true clinical impact has not yet been established. All 3 agents are substrates for the P-glycoprotein (P-gp) transport system, so any known inhibitors or inducers of the P-gp system should be used cautiously in patients on NOACs.^1,3-7,12 Rivaroxaban and apixaban are also substrates for the CYP3A4 hepatic enzyme system, so any drugs known to inhibit or induce this system require caution, as well.^1,3-7,12

Who should not take a NOAC?The most effective therapy for patients who need rapid reversal of any NOAC is to administer 75 to 80 unites/kg of activated prothrombin complex concentrate.

NOACs should not be prescribed for patients with mechanical heart valves.³⁴ Dabigatran is the only NOAC to have been studied in this patient population, and the phase II trial was stopped prematurely due to increased risk for both bleeding and stroke in patients on dabigatran compared with warfarin.³⁴

Renal impairment must be considered, as well. Do a baseline assessment of renal function in all patients before transitioning them to a NOAC, and periodic reassessment during therapy. While this is important for patients on rivaroxaban and apixaban, it is essential for those on dabigatran, as 80% of the drug is excreted by the kidneys.^1,5,12 NOACs have not been adequately assessed in patients with severe renal dysfunction and should be avoided in this patient population. Caution should be exercised in patients with moderate renal dysfunction, as well.^5-10,14-19 Apixaban appears to be the safest NOAC for patients with moderate renal dysfunction, as it has the least renal clearance.^1,12

Who should take a NOAC?

No well-established criteria for patient selection for NOACs exist, yet appropriate patient selection is crucial. Evidence suggests that NOAC therapy is best suited to those who:
• are relatively young (<65 years)
• have normal renal function
• have poorly controlled INR with warfarin that is unrelated to noncompliance
• are unable to have regular INR monitoring.

NOACs have not been adequately assessed in severe renal dysfunction and should be avoided in this patient population.Patients best suited for continued use of warfarin would be those whose INR is well controlled, those who have higher goal INR ranges (eg, because of the presence of mechanical heart valves), patients with significant renal dysfunction, and individuals with a history of peptic ulcer disease or GI bleeding. Warfarin may also be the best option for patients with a history of noncompliance and for uninsured or underinsured patients.

CASE 1 Warfarin and any of the NOACs were all feasible options for Ms. J, but apixaban was deemed to be the safest because of her moderate renal dysfunction. However, after she was told that apixaban has little “real world” clinical data, no effective antidote if bleeding were to occur, and a much higher cost than warfarin, she opted for warfarin therapy, despite the laboratory monitoring required.

CASE 2 Mr. W was excited to learn that there were new alternatives to warfarin; he had taken warfarin for 6 months after his last DVT and had a hard time coming in for INR checks. The patient reported that he had no history of bleeding and was compliant with medications. Rivaroxaban was the best option for Mr. W, as it is the only NOAC with FDA approval for the treatment of acute VTE.

CORRESPONDENCE
Jeremy Vandiver, PharmD, BCPS, Swedish Medical Center, Room 3260, 501 East Hampden Avenue, Englewood, CO 80013; [email protected]

CASE 1 Sally J is a 72-year-old Caucasian woman who comes to your clinic after being diagnosed with atrial fibrillation (AF). The patient has a 10-year history of type 2 diabetes; she also has a history of hypertension and chronic kidney disease (CKD), with a baseline creatinine clearance of approximately 40 mL/min. Ms. J tells you she knows people who take warfarin and really dislike it. She asks for your opinion of the new anticoagulants she’s seen advertised on TV, and wonders whether one of them would be right for her.

CASE 2 Bobby W, a 35-year-old African American man, was recently diagnosed with deep vein thrombosis (DVT). This was his second clot in 5 years, and occurred after a long flight home from Europe. The patient explains that he leads a very active lifestyle and doesn’t have the time to come in for the monthly international normalized ratio (INR) checks that warfarin requires. What would you recommend for these patients?

Troubled by warfarin’s narrow therapeutic index, numerous medication and dietary interactions, and need for frequent monitoring, patients requiring long-term oral anticoagulation therapy have been seeking alternatives for years. Finally, they have a choice. The US Food and Drug Administration (FDA) approved 3 oral anticoagulants—dabigatran (Pradaxa), rivaroxaban (Xarelto), and apixaban (Eliquis)—in less than 4 years. Known as novel oral anticoagulants (NOACs), they are the first such drugs to enter the market in more than 50 years.^1,2

While warfarin inhibits a wide range of clotting factors (including II, VII, IX, and X), NOACs work further down the clotting cascade (TABLE 1).^1,3-7 Dabigatran, a direct thrombin inhibitor, only inhibits factor IIa.^3,5 Rivaroxaban and apixaban directly inhibit factor Xa and indirectly inhibit factor IIa.^3,6,7

There are notable advantages to these newer agents, but some disadvantages that must be considered, as well. Appropriate patient selection, guided by a thorough understanding of the benefits and risks of NOACs, is key.

Stroke prevention in atrial fibrillation

All 3 NOACs are approved for stroke prevention in patients with nonvalvular atrial fibrillation (AF). The approvals are based on a small number of well-designed trials: RE-LY (dabigatran), ROCKET-AF (rivaroxaban), and ARISTOTLE (apixaban).^8-10 Compared with warfarin, dabigatran is the only oral anticoagulant with a lower rate of both hemorrhagic and ischemic stroke.⁸ Both rivaroxaban and apixaban were found to decrease overall stroke risk relative to warfarin, but the difference was driven by a lower risk for hemorrhagic, not ischemic, stroke.^9,10

In these trials, overall rates of major bleeding were similar to that of warfarin.^8-10 Patients taking warfarin generally experienced higher rates of intracranial hemorrhage but lower rates of gastrointestinal (GI) bleeding than those on NOACs. Relative to warfarin, apixaban was the only NOAC that did not have a higher rate of GI bleeding and the only one with a lower rate of major bleeding.^8-10 In addition, apixaban remains the only NOAC found to have a statistically significant decrease in all-cause mortality compared with warfarin.¹⁰ Although dabigatran and rivaroxaban were associated with a strong trend towards decreased mortality, both studies were underpowered for this secondary outcome.^8,9

Adding NOACs to stroke guidelines. The role of NOACs in the prevention of stroke in patients with nonvalvular AF is beginning to be reflected in newer guidelines. The American College of Chest Physicians (ACCP)’s 2012 guidelines recommend dabigatran over warfarin (grade 2B—weak recommendation; moderate quality evidence) unless the patient is well controlled on warfarin.¹¹ The European Society of Cardiology (ESC)’s 2012 guidelines recommend dabigatran, apixaban, and rivaroxaban as broadly preferable to warfarin, while noting that experience with these agents is limited and appropriate patient selection is important.¹²

Anticoagulation to treat—and prevent—VTE

The standard of care for acute venous thromboembolism (VTE) is to initiate warfarin along with a parenteral anticoagulant, such as unfractionated heparin, low-molecular-weight heparin (LMWH), or fondaparinux.¹³ Due to warfarin’s slow onset to peak effect, a parenteral anticoagulant is overlapped for ≥5 days—until warfarin reaches a therapeutic level and can be continued as monotherapy.¹³ But many patients find subcutaneous delivery of LMWH disagreeable and costly and frequent INR monitoring inconvenient, so the new agents offer notable advantages.

In well-designed studies, dabigatran, rivaroxaban, and apixaban have all been shown to be noninferior to warfarin in the initial treatment of acute DVT and pulmonary embolism (PE).^14-17 All 3 agents were also shown to have lower rates of major bleeding than warfarin. Rivaroxaban and apixaban were also superior to warfarin with regard to bleeding events, and dabigatran was noninferior to warfarin for this outcome.^14-17

NOACs help prevent recurrence

All 3 NOACs have been studied for long-term prevention of recurrent VTE after 3 to 18 months of anticoagulation, as well. Dabigatran was found in the RE-MEDY trial to be noninferior to warfarin for the risk of recurrent VTE, and to have lower rates of bleeding.¹⁸ In separate trials, all 3 agents were superior to placebo in preventing recurrent VTE. Rates of long-term major bleeding were significantly higher than placebo with rivaroxaban and dabigatran, but not with apixaban.^15,18,19

Rivaroxaban is the only NOAC to be FDA approved for the treatment of acute DVT and PE, and the ACCP’s 2012 guidelines list it as a viable alternative to parenteral anticoagulation when initiating treatment for acute VTE.^6,13 When treating VTE long term, the guidelines continue to recommend warfarin or LMWH rather than dabigatran or rivaroxaban.¹³ Recommendations may change in coming years, as physicians gain more experience with NOACs and more clinical trials are published.^16-19

Starting or converting to NOAC therapy

In patients who have not been on anticoagulant therapy, any NOAC can be initiated immediately, with no need for parenteral, or “bridge” therapy. This is because of the rapid onset of action of the NOACs.¹²

To transition a patient from warfarin to a NOAC, it is necessary to discontinue warfarin therapy completely and closely monitor INR, then initiate NOAC therapy when INR≤2. No parenteral anticoagulation is necessary (TABLE 2).^5-7

If it is necessary to transition a patient from a NOAC to warfarin, the protocol depends on the agent. Because dabigatran has no significant impact on prolongation of prothrombin time (PT), it can be overlapped with warfarin. Rivaroxaban and apixaban have a significant impact on PT prolongation, however, and overlapping either agent with warfarin is not recommended.⁴ Keep in mind that the recommended dosages for the NOACs are not standardized, and can differ drastically depending on the indication for use as well as on patient-specific factors, including renal function, body weight, and age.

Laboratory monitoring is not required

While warfarin has a great deal of interpatient variability and requires frequent lab monitoring, an oft-cited advantage of the NOACs is that they do not require regular monitoring. However, that also has a downside (TABLE 3).^{1,3-10,12,14-17,20-22} Monitoring INR in patients on warfarin allows physicians to assess patient compliance. And, if a patient on warfarin requires an invasive procedure, coagulation status and bleeding risk can easily be determined. That is not the case with the NOACs.

While some routine laboratory tests may be elevated in a patient taking a NOAC, the degree of elevation does not correlate well with anticoagulant concentration. And, because each NOAC has a different mechanism of action, different measures will be elevated in a patient taking dabigatran vs apixaban or rivaroxaban.⁴

Activated partial thromboplastin time (aPTT) is the most readily available lab test to assess the presence or absence of dabigatran.⁴ A normal aPTT indicates that there is little to no dabigatran present.⁴ But, while an elevated aPTT suggests the presence of dabigatran, it provides little information about how much.⁴

PT is a useful test to assess coagulation status in patients on either rivaroxaban or apixaban. A normal PT suggests that minimal amounts (or none) of the NOAC are present in the plasma.⁴ (A direct thrombin inhibitor assay, calibrated to more accurately assess dabigatran concentration, is being developed for clinical use, but is currently available only for research purposes in the United States; a chromogenic antifactor Xa test specific to apixaban and rivaroxaban is also being developed, but is not yet commercially available.⁴)

What to do when NOAC reversal is required

Patients often need to stop taking an oral anticoagulation in the days leading up to a planned invasive procedure. In an individual with normal renal function who will undergo a procedure with a standard bleeding risk, a NOAC would generally need to be withheld for one to 2 days prior to surgery, given the relatively short half-life. If a patient has acute renal failure or CKD, however, dabigatran may need to be withheld for a prolonged period (3-6 days) in order to safely proceed to surgery.^4,23 NOACs may also need to be withheld for 2 to 6 days prior to any surgery with a high risk for bleeding.⁴

When speed is of the essenceIn patients who have not been on anticoagulant therapy, any NOAC can be initiated immediately, with no need for "bridge" therapy.

There is no known antidote to aid in the reversal of dabigatran, rivaroxaban, or apixaban.⁴ Because of their relatively short half-lives, withholding the medication and providing supportive care is generally sufficient to ensure adequate hemostasis in cases of mild to moderate bleeding.⁴ If a patient presents with acute ingestion or an overdose, activated charcoal should be administered if the ingestion has occurred within the past 3 hours.^4,24 The lack of a clear-cut reversal strategy can be extremely problematic in cases of trauma or life-threatening bleeding, however. (Fresh frozen plasma has not been shown to be effective at reversing NOACs’ effects.⁴)

In instances of severe bleeding or the need for urgent surgery, a more aggressive approach may be needed. Hemodialysis can be used to assist in the removal of dabigatran, but not rivaroxaban or apixaban.⁴ However, evidence suggests that the most effective therapy for patients who need rapid reversal of any NOAC is to administer 75 to 80 units/kg of activated prothrombin complex concentrate (aPCC).^4,25-27 Recombinant factor VIIa has shown some promise in reversing the anticoagulant effects of these novel agents, but evidence is insufficient to recommend it as first-line therapy at this time.^{26, 27}

Patients are more satisfied

A direct thrombin inhibitor assay, calibrated to more accurately assess dabigatran concentration, is being developed, as is a chromogenic antifactor Xa test specific to apixaban and rivaroxaban.The most obvious advantage of the NOACs as a group compared with warfarin is the lack of need for laboratory monitoring or continuous dose titration. Reliably stable pharmacokinetics make once or twice daily dosing possible. A rapid onset of action negates the need for bridging therapy with parenteral anticoagulants in patients at high risk of thrombosis. This may improve compliance, as many patients are averse to the use of subcutaneous injections or need extensive education before they can safely self-inject. The incidence of heparin-induced thrombocytopenia may also be decreased if unfractionated heparin and LMWH are used less frequently.

NOACs also appear to improve patient satisfaction.^20-22 In one study that included patients with AF on dabigatran or warfarin, satisfaction was higher in those taking dabigatran, particularly among those who did not experience significant GI adverse effects.²⁰ Another study showed improved patient satisfaction with rivaroxaban compared with LMWH following lower extremity joint replacement, which led to significantly higher rates of compliance.²²

… but problems and pitfalls remain

In addition to the lack of a readily available and clinically validated reversal agent, the absence of a lab test that reliably measures the concentration of NOACs makes it difficult to determine whether patients are following their prescribed regimen.^3,4

Medication compliance must be assessed when considering a transition from warfarin to a NOAC. Switching patients with poor INR control on warfarin to a NOAC should be done only after determining that the poor control is not the result of nonadherence. Because of the NOACs’ shorter half-life, patients who don’t take them regularly may be at higher risk for thromboembolic events.^1,12

Cost is a serious consideration. While there are some costs associated with the monitoring warfarin requires, the medication itself has been generic for several decades and can be found on many “$4 lists” at pharmacies nationwide. In contrast, all 3 NOACs are available only as branded drugs, and can cost a patient with limited drug coverage anywhere from $250 to $350 per month²⁸—a serious concern, given that the likelihood of noncompliance increases as out-of-pocket costs rise. This was highlighted in a recent study that found patients were twice as likely to discontinue their cholesterol-lowering medication if 100% of the cost was out of pocket, compared with patients who had no prescription copay.²⁹ From the perspective of the US health care system, however, NOACs have been found to be cost effective compared with warfarin, mostly due to the lack of laboratory monitoring.^30-32

Adverse effects. The risk of GI bleeds has been shown to be higher in patients taking rivaroxaban and dabigatran vs warfarin.^8,9 Dabigatran has also been associated with a significant risk for dyspepsia.^5,8,14 In clinical trials, the reported rate for dyspepsia in patients taking dabigatran was 3% to 11%; subsequent investigations have found the incidence to be far higher (33%).^8,14,33

Drug interactions. Warfarin has a large number of drug interactions, of course, but because of its long history, these interactions are well established. NOACs also have a number of drug interactions, but the true clinical impact has not yet been established. All 3 agents are substrates for the P-glycoprotein (P-gp) transport system, so any known inhibitors or inducers of the P-gp system should be used cautiously in patients on NOACs.^1,3-7,12 Rivaroxaban and apixaban are also substrates for the CYP3A4 hepatic enzyme system, so any drugs known to inhibit or induce this system require caution, as well.^1,3-7,12

Who should not take a NOAC?The most effective therapy for patients who need rapid reversal of any NOAC is to administer 75 to 80 unites/kg of activated prothrombin complex concentrate.

NOACs should not be prescribed for patients with mechanical heart valves.³⁴ Dabigatran is the only NOAC to have been studied in this patient population, and the phase II trial was stopped prematurely due to increased risk for both bleeding and stroke in patients on dabigatran compared with warfarin.³⁴

Renal impairment must be considered, as well. Do a baseline assessment of renal function in all patients before transitioning them to a NOAC, and periodic reassessment during therapy. While this is important for patients on rivaroxaban and apixaban, it is essential for those on dabigatran, as 80% of the drug is excreted by the kidneys.^1,5,12 NOACs have not been adequately assessed in patients with severe renal dysfunction and should be avoided in this patient population. Caution should be exercised in patients with moderate renal dysfunction, as well.^5-10,14-19 Apixaban appears to be the safest NOAC for patients with moderate renal dysfunction, as it has the least renal clearance.^1,12

Who should take a NOAC?

No well-established criteria for patient selection for NOACs exist, yet appropriate patient selection is crucial. Evidence suggests that NOAC therapy is best suited to those who:
• are relatively young (<65 years)
• have normal renal function
• have poorly controlled INR with warfarin that is unrelated to noncompliance
• are unable to have regular INR monitoring.

NOACs have not been adequately assessed in severe renal dysfunction and should be avoided in this patient population.Patients best suited for continued use of warfarin would be those whose INR is well controlled, those who have higher goal INR ranges (eg, because of the presence of mechanical heart valves), patients with significant renal dysfunction, and individuals with a history of peptic ulcer disease or GI bleeding. Warfarin may also be the best option for patients with a history of noncompliance and for uninsured or underinsured patients.

CASE 1 Warfarin and any of the NOACs were all feasible options for Ms. J, but apixaban was deemed to be the safest because of her moderate renal dysfunction. However, after she was told that apixaban has little “real world” clinical data, no effective antidote if bleeding were to occur, and a much higher cost than warfarin, she opted for warfarin therapy, despite the laboratory monitoring required.

CASE 2 Mr. W was excited to learn that there were new alternatives to warfarin; he had taken warfarin for 6 months after his last DVT and had a hard time coming in for INR checks. The patient reported that he had no history of bleeding and was compliant with medications. Rivaroxaban was the best option for Mr. W, as it is the only NOAC with FDA approval for the treatment of acute VTE.

CORRESPONDENCE
Jeremy Vandiver, PharmD, BCPS, Swedish Medical Center, Room 3260, 501 East Hampden Avenue, Englewood, CO 80013; [email protected]

CASE Jane K, a 40-year-old multiparous woman, is seeking treatment for heavy menstrual bleeding and cramping, both of which have troubled her for 4 years. Another physician had given her oral contraceptive pills (OCPs) to decrease the pain and bleeding, she reports, but she has difficulty remembering to take a pill every day.

On physical exam, you note an enlarged uterus (approximately 16-week size). A pregnancy test is negative, her thyroidstimulating hormone level is normal, and her hemoglobin is 9 g/dL. Transvaginal ultrasound reveals multiple fibroids.

What can you offer her?

At some point in their reproductive years, 10% to 15% of women experience heavy menstrual bleeding (HMB), or menorrhagia.^1,2 In fact, HMB and dysmenorrhea are among the most common reasons for office visits and missed work among women in this age group.^3,4 In addition to having a negative impact on quality of life, HMB can cause severe anemia.⁵

All too often, the suggested solution is a hysterectomy. In fact, 90% of the more than 600,000 hysterectomies performed annually in the United States are for benign disease.^6,7 Yet many women with HMB want nonsurgical treatments, and some seek to preserve their fertility. A progestin IUD can often fulfill both of these desires.

An IUD containing 52 mg levonorgestrel (Mirena) has US Food and Drug Administration (FDA) approval both for use as a contraceptive and for the treatment of HMB in women who want intrauterine contraception.⁸ Recent studies have confirmed its efficacy in treating a wide variety of conditions associated with menorrhagia and dysmenorrhea. In 2013, a smaller, lower dose (13.5 mg) levonorgestrel-releasing IUD (Skyla) received FDA approval as a contraceptive.⁹ But because this device has been neither tested nor approved for other applications, the following review applies only to Mirena—referred to throughout this article as the LNG-IUD.

A proven (and often superior) treatment for menorrhagia

Leiomyomas, or fibroids—the most common benign tumor of the female genital tract—are a frequent cause of menorrhagia.¹⁰ OCPs are often used for the treatment of symptomatic fibroids for women who wish to avoid surgery. When compared with low-dose OCPs for the treatment of menorrhagia secondary to fibroids, however, the LNG-IUD resulted in a significantly greater reduction in blood loss.¹¹

While fibroid size does not appear to decrease significantly after insertion of the device, a systematic review found that menstrual bleeding lessens and hemoglobin levels improve in women with symptomatic fibroids.^12,13 The LNG-IUD has also been shown to improve symptoms in women with dysmenorrhea secondary to fibroids.¹⁴

The 52 mg levonorgestrol-releasing IUD has been shown to be effective in treating multiple conditions associated with menorrhagia and dysmenorrhea.Hemostatic disorders. Anticoagulants are vital for women with hemostatic disorders such as Von Willebrand disease, immune thrombocytopenia, or a clotting factor deficiency (See “Is a novel anticoagulant right for your patient?”), but their use may cause or worsen menorrhagia. In such cases, the LNG-IUD appears to be an effective treatment.

In a retrospective case review of 28 women with menorrhagia secondary to various hemostatic disorders, 68% experienced improvement after insertion of the IUD.¹⁵ Another study compared women on anticoagulant therapy for cardiac valve disease (N=40) with and without the LNG-IUD. Compared with the control group, those with the IUD were found to have significant increases in hemoglobin levels 3 months after insertion.¹⁶

Obesity-related uterine bleeding. Obese women are at higher risk for excessive uterine bleeding, the result of increased conversion of plasma androstenedione to estrogen in adipose tissue. In one study evaluating the use of the LNG-IUD in this patient population, 75% of participants experienced a reduction in bleeding.¹⁷

Idiopathic HMB. The IUD is as good as, or better than, other treatments for idiopathic menorrhagia. It results in a significantly higher reduction in both blood loss and days out of work than OCPs.¹ The device also reduces blood loss more effectively than oral medroxyprogesterone,¹⁸ another common approach to idiopathic HMB; and, compared with hysterectomy, it results in similar patient satisfaction—but lower costs and complication rates.¹⁹

In a randomized controlled trial that compared the LNG-IUD with tranexamic acid, mefenamic acid, combined estrogenprogesterone, or progesterone alone over a 2-year period, scores on a menorrhagia symptom scale were significantly higher (indicating greater improvement) in the LNG-IUD group.²⁰

Medical management of endometrial proliferation

Endometrial hyperplasia, often found in women with abnormal uterine bleeding patterns and recurrent anovulatory cycles,²¹ is sometimes treated with supplemental progesterone. According to the Centers for Disease Control and Prevention (CDC)’s US Medical Eligibility Criteria for Contraceptive Use, the LNG-IUD can be used without restriction in this patient population, as well.²²

A systematic review of 9 studies of women with endometrial hyperplasia without atypia found the LNG-IUD to be both safe and effective. In 7 of the 9 studies, 100% of participants experienced disease regression; regression rates for the other 2 studies were 90% and 67%.²³ The only caveat: Both endometrial atypia and endometrial cancer should be excluded prior to IUD insertion.

Adenomyosis is caused by the presence of ectopic endometrial glands and stroma within the myometrium and frequently results in pelvic pain, menorrhagia, and dysmenorrhea.¹⁰ Hysterectomy is often regarded as the mainstay of treatment. But medical management with the LNG-IUD is also an option, as it has demonstrated similar improvements to hysterectomy both in hemoglobin levels and quality of life.²⁴ Three years after insertion of the LNG-IUD to treat moderate or severe adenomyosis-associated dysmenorrhea, one study found, women reported significant improvement in their symptoms—and 73% were satisfied with their treatment.²⁵ The LNG-IUD also appears to decrease uterine volume, although this effect may begin to decrease 2 years after insertion.²⁶

Endometriosis, another common cause of dysmenorrhea and chronic pelvic pain,²⁷ can also be treated with the LNG-IUD. The local progestin administration to pelvic structures that the device provides has been found to significantly decrease both endometrial proliferation and monthly blood flow.²⁸ Additional studies of the LNG-IUD as a treatment for endometriosis and pelvic pain are ongoing and encouraging. After surgery for endometriosis, a Cochrane review found, women who had the LNG-IUD inserted had a lower rate of recurrence of dysmenorrhea than those without it.²⁹

Helping women through perimenopause

Compared with oral or intramuscular progesterone therapy, the LNG-IUD has been found to be superior for the treatment of perimenopausal symptoms.³⁰ Two years after insertion, one study found, perimenopausal women had a 95% reduction in blood loss and a 63% decrease in dysmenorrhea.³¹ The LNG-IUD also provides reliable endometrial protection for women receiving estrogen therapy³² and for those who are taking adjuvant tamoxifen because of a history of breast cancer.^33,34

IUD insertion is a safe office procedure

The Society of Teachers of Family Medicine cites IUD insertion in its core list of routine procedures to be included in family medicine residency programs.³⁵ The risk associated with insertion is small—uterine perforation occurs in about 2.6/1000 insertions³⁶—and there is a small and transient increase in the risk of IUD-related infection in the first few weeks to months after insertion. IUD insertion does not increase the overall risk of pelvic inflammatory disease in women at low risk for sexually transmitted infections.^37,38

Who is not a candidate?

While IUDs are safe for most women, there are several absolute contraindications to the LNG-IUD:
• current breast, cervical, or endometrial cancer
• current pelvic inflammatory disease, cervicitis, chlamydia, or gonorrhea
• having just had a septic abortion.³⁸

Teach patients about the benefits and adverse effects

For women who are potential candidates for the LNG-IUD, education is vital. Evidence suggests that satisfaction levels are very high, provided patients receive adequate counseling about the benefits and adverse effects. Risks (of uterine perforation and infection) are small, as noted earlier.³⁹

Contraceptive efficacy, of course, is a major benefit, and has been well The risk associated with insertion is small—uterine perforation occurs in about 2.6/1000 insertions—and there is a small and transient risk of infection in the first few weeks to months after insertion.documented: The LNG-IUD has an estimated failure rate of just 0.2%.⁴⁰ Unlike user-dependent methods such as OCPs, the patch, and the ring, the IUD has a perfect-use failure rate that is the same as the typical use rate. Thus, it is an excellent choice for women who want to preserve their fertility yet avoid an unintended pregnancy. For women approaching menopause—a time when estrogen may be contraindicated—the LNG-IUD can safely protect women against unwanted pregnancy.

Lower cost, less invasive. The ability to treat HMB and dysmennorhea with an IUD inserted in a family practice setting, without referrals to specialists for additional invasive treatments, increases cost savings.¹⁹ In addition, the LNG-IUD is less invasive and generally more acceptable to women than hysterectomy, endometrial ablation, uterine artery embolization, and myomectomy.^18,41 It leads to a greater reduction in menstrual bleeding than OCPs, oral progestins, tranexamic acid, and oral mefenamic acid.⁴¹ And, unlike some progestational agents, there is no evidence that the LNG-IUD has any adverse effects on bone density, vaginal tone, or urinary continence.⁴²

The LNG-IUD is less invasive and generally more acceptable to women than hysterectomy, endometrial ablation, uterine artery embolization, and myomectomy. Adverse effects. Vaginal spotting is the most commonly reported adverse effect associated with the LNG-IUD, particularly in the first 3 to 6 months postinsertion.⁴² The other common adverse effect is increased cramping, which some women experience in the first few months after insertion. Rarely women may experience ovarian cysts from unruptured follicles, which will regress on their own. Another potential problem is expulsion, which is more common in women who are using the device to control heavy bleeding (8.9%-13.6%).⁴² After the device is in place for several years, many women experience amenorrhea—a side effect that patients who have suffered from HMB and dysmenorrhea may consider a benefit.

CASE After counseling regarding her treatment options, Ms. K decides on the LNG-IUD, which her family physician inserts. At 3-month follow-up, she reports significantly less bleeding and decreased perimenstrual discomfort. If her workup had revealed adenomyosis or a hemostatic disorder, the LNG-IUD would still have been a first-line option.

CORRESPONDENCE

Erin Hendriks, MD, Wayne State University Family Medicine, 1135 West University Drive, Suite 250, Rochester Hills, MI 48307; [email protected]

CASE Jane K, a 40-year-old multiparous woman, is seeking treatment for heavy menstrual bleeding and cramping, both of which have troubled her for 4 years. Another physician had given her oral contraceptive pills (OCPs) to decrease the pain and bleeding, she reports, but she has difficulty remembering to take a pill every day.

On physical exam, you note an enlarged uterus (approximately 16-week size). A pregnancy test is negative, her thyroidstimulating hormone level is normal, and her hemoglobin is 9 g/dL. Transvaginal ultrasound reveals multiple fibroids.

What can you offer her?

At some point in their reproductive years, 10% to 15% of women experience heavy menstrual bleeding (HMB), or menorrhagia.^1,2 In fact, HMB and dysmenorrhea are among the most common reasons for office visits and missed work among women in this age group.^3,4 In addition to having a negative impact on quality of life, HMB can cause severe anemia.⁵

All too often, the suggested solution is a hysterectomy. In fact, 90% of the more than 600,000 hysterectomies performed annually in the United States are for benign disease.^6,7 Yet many women with HMB want nonsurgical treatments, and some seek to preserve their fertility. A progestin IUD can often fulfill both of these desires.

An IUD containing 52 mg levonorgestrel (Mirena) has US Food and Drug Administration (FDA) approval both for use as a contraceptive and for the treatment of HMB in women who want intrauterine contraception.⁸ Recent studies have confirmed its efficacy in treating a wide variety of conditions associated with menorrhagia and dysmenorrhea. In 2013, a smaller, lower dose (13.5 mg) levonorgestrel-releasing IUD (Skyla) received FDA approval as a contraceptive.⁹ But because this device has been neither tested nor approved for other applications, the following review applies only to Mirena—referred to throughout this article as the LNG-IUD.

A proven (and often superior) treatment for menorrhagia

Leiomyomas, or fibroids—the most common benign tumor of the female genital tract—are a frequent cause of menorrhagia.¹⁰ OCPs are often used for the treatment of symptomatic fibroids for women who wish to avoid surgery. When compared with low-dose OCPs for the treatment of menorrhagia secondary to fibroids, however, the LNG-IUD resulted in a significantly greater reduction in blood loss.¹¹

While fibroid size does not appear to decrease significantly after insertion of the device, a systematic review found that menstrual bleeding lessens and hemoglobin levels improve in women with symptomatic fibroids.^12,13 The LNG-IUD has also been shown to improve symptoms in women with dysmenorrhea secondary to fibroids.¹⁴

The 52 mg levonorgestrol-releasing IUD has been shown to be effective in treating multiple conditions associated with menorrhagia and dysmenorrhea.Hemostatic disorders. Anticoagulants are vital for women with hemostatic disorders such as Von Willebrand disease, immune thrombocytopenia, or a clotting factor deficiency (See “Is a novel anticoagulant right for your patient?”), but their use may cause or worsen menorrhagia. In such cases, the LNG-IUD appears to be an effective treatment.

In a retrospective case review of 28 women with menorrhagia secondary to various hemostatic disorders, 68% experienced improvement after insertion of the IUD.¹⁵ Another study compared women on anticoagulant therapy for cardiac valve disease (N=40) with and without the LNG-IUD. Compared with the control group, those with the IUD were found to have significant increases in hemoglobin levels 3 months after insertion.¹⁶

Obesity-related uterine bleeding. Obese women are at higher risk for excessive uterine bleeding, the result of increased conversion of plasma androstenedione to estrogen in adipose tissue. In one study evaluating the use of the LNG-IUD in this patient population, 75% of participants experienced a reduction in bleeding.¹⁷

Idiopathic HMB. The IUD is as good as, or better than, other treatments for idiopathic menorrhagia. It results in a significantly higher reduction in both blood loss and days out of work than OCPs.¹ The device also reduces blood loss more effectively than oral medroxyprogesterone,¹⁸ another common approach to idiopathic HMB; and, compared with hysterectomy, it results in similar patient satisfaction—but lower costs and complication rates.¹⁹

In a randomized controlled trial that compared the LNG-IUD with tranexamic acid, mefenamic acid, combined estrogenprogesterone, or progesterone alone over a 2-year period, scores on a menorrhagia symptom scale were significantly higher (indicating greater improvement) in the LNG-IUD group.²⁰

Medical management of endometrial proliferation

Endometrial hyperplasia, often found in women with abnormal uterine bleeding patterns and recurrent anovulatory cycles,²¹ is sometimes treated with supplemental progesterone. According to the Centers for Disease Control and Prevention (CDC)’s US Medical Eligibility Criteria for Contraceptive Use, the LNG-IUD can be used without restriction in this patient population, as well.²²

A systematic review of 9 studies of women with endometrial hyperplasia without atypia found the LNG-IUD to be both safe and effective. In 7 of the 9 studies, 100% of participants experienced disease regression; regression rates for the other 2 studies were 90% and 67%.²³ The only caveat: Both endometrial atypia and endometrial cancer should be excluded prior to IUD insertion.

Adenomyosis is caused by the presence of ectopic endometrial glands and stroma within the myometrium and frequently results in pelvic pain, menorrhagia, and dysmenorrhea.¹⁰ Hysterectomy is often regarded as the mainstay of treatment. But medical management with the LNG-IUD is also an option, as it has demonstrated similar improvements to hysterectomy both in hemoglobin levels and quality of life.²⁴ Three years after insertion of the LNG-IUD to treat moderate or severe adenomyosis-associated dysmenorrhea, one study found, women reported significant improvement in their symptoms—and 73% were satisfied with their treatment.²⁵ The LNG-IUD also appears to decrease uterine volume, although this effect may begin to decrease 2 years after insertion.²⁶

Endometriosis, another common cause of dysmenorrhea and chronic pelvic pain,²⁷ can also be treated with the LNG-IUD. The local progestin administration to pelvic structures that the device provides has been found to significantly decrease both endometrial proliferation and monthly blood flow.²⁸ Additional studies of the LNG-IUD as a treatment for endometriosis and pelvic pain are ongoing and encouraging. After surgery for endometriosis, a Cochrane review found, women who had the LNG-IUD inserted had a lower rate of recurrence of dysmenorrhea than those without it.²⁹

Helping women through perimenopause

Compared with oral or intramuscular progesterone therapy, the LNG-IUD has been found to be superior for the treatment of perimenopausal symptoms.³⁰ Two years after insertion, one study found, perimenopausal women had a 95% reduction in blood loss and a 63% decrease in dysmenorrhea.³¹ The LNG-IUD also provides reliable endometrial protection for women receiving estrogen therapy³² and for those who are taking adjuvant tamoxifen because of a history of breast cancer.^33,34

IUD insertion is a safe office procedure

The Society of Teachers of Family Medicine cites IUD insertion in its core list of routine procedures to be included in family medicine residency programs.³⁵ The risk associated with insertion is small—uterine perforation occurs in about 2.6/1000 insertions³⁶—and there is a small and transient increase in the risk of IUD-related infection in the first few weeks to months after insertion. IUD insertion does not increase the overall risk of pelvic inflammatory disease in women at low risk for sexually transmitted infections.^37,38

Who is not a candidate?

While IUDs are safe for most women, there are several absolute contraindications to the LNG-IUD:
• current breast, cervical, or endometrial cancer
• current pelvic inflammatory disease, cervicitis, chlamydia, or gonorrhea
• having just had a septic abortion.³⁸

Teach patients about the benefits and adverse effects

For women who are potential candidates for the LNG-IUD, education is vital. Evidence suggests that satisfaction levels are very high, provided patients receive adequate counseling about the benefits and adverse effects. Risks (of uterine perforation and infection) are small, as noted earlier.³⁹

Contraceptive efficacy, of course, is a major benefit, and has been well The risk associated with insertion is small—uterine perforation occurs in about 2.6/1000 insertions—and there is a small and transient risk of infection in the first few weeks to months after insertion.documented: The LNG-IUD has an estimated failure rate of just 0.2%.⁴⁰ Unlike user-dependent methods such as OCPs, the patch, and the ring, the IUD has a perfect-use failure rate that is the same as the typical use rate. Thus, it is an excellent choice for women who want to preserve their fertility yet avoid an unintended pregnancy. For women approaching menopause—a time when estrogen may be contraindicated—the LNG-IUD can safely protect women against unwanted pregnancy.

Lower cost, less invasive. The ability to treat HMB and dysmennorhea with an IUD inserted in a family practice setting, without referrals to specialists for additional invasive treatments, increases cost savings.¹⁹ In addition, the LNG-IUD is less invasive and generally more acceptable to women than hysterectomy, endometrial ablation, uterine artery embolization, and myomectomy.^18,41 It leads to a greater reduction in menstrual bleeding than OCPs, oral progestins, tranexamic acid, and oral mefenamic acid.⁴¹ And, unlike some progestational agents, there is no evidence that the LNG-IUD has any adverse effects on bone density, vaginal tone, or urinary continence.⁴²

The LNG-IUD is less invasive and generally more acceptable to women than hysterectomy, endometrial ablation, uterine artery embolization, and myomectomy. Adverse effects. Vaginal spotting is the most commonly reported adverse effect associated with the LNG-IUD, particularly in the first 3 to 6 months postinsertion.⁴² The other common adverse effect is increased cramping, which some women experience in the first few months after insertion. Rarely women may experience ovarian cysts from unruptured follicles, which will regress on their own. Another potential problem is expulsion, which is more common in women who are using the device to control heavy bleeding (8.9%-13.6%).⁴² After the device is in place for several years, many women experience amenorrhea—a side effect that patients who have suffered from HMB and dysmenorrhea may consider a benefit.

CASE After counseling regarding her treatment options, Ms. K decides on the LNG-IUD, which her family physician inserts. At 3-month follow-up, she reports significantly less bleeding and decreased perimenstrual discomfort. If her workup had revealed adenomyosis or a hemostatic disorder, the LNG-IUD would still have been a first-line option.

CORRESPONDENCE

Erin Hendriks, MD, Wayne State University Family Medicine, 1135 West University Drive, Suite 250, Rochester Hills, MI 48307; [email protected]

CASE Jane K, a 40-year-old multiparous woman, is seeking treatment for heavy menstrual bleeding and cramping, both of which have troubled her for 4 years. Another physician had given her oral contraceptive pills (OCPs) to decrease the pain and bleeding, she reports, but she has difficulty remembering to take a pill every day.

On physical exam, you note an enlarged uterus (approximately 16-week size). A pregnancy test is negative, her thyroidstimulating hormone level is normal, and her hemoglobin is 9 g/dL. Transvaginal ultrasound reveals multiple fibroids.

What can you offer her?

At some point in their reproductive years, 10% to 15% of women experience heavy menstrual bleeding (HMB), or menorrhagia.^1,2 In fact, HMB and dysmenorrhea are among the most common reasons for office visits and missed work among women in this age group.^3,4 In addition to having a negative impact on quality of life, HMB can cause severe anemia.⁵

All too often, the suggested solution is a hysterectomy. In fact, 90% of the more than 600,000 hysterectomies performed annually in the United States are for benign disease.^6,7 Yet many women with HMB want nonsurgical treatments, and some seek to preserve their fertility. A progestin IUD can often fulfill both of these desires.

An IUD containing 52 mg levonorgestrel (Mirena) has US Food and Drug Administration (FDA) approval both for use as a contraceptive and for the treatment of HMB in women who want intrauterine contraception.⁸ Recent studies have confirmed its efficacy in treating a wide variety of conditions associated with menorrhagia and dysmenorrhea. In 2013, a smaller, lower dose (13.5 mg) levonorgestrel-releasing IUD (Skyla) received FDA approval as a contraceptive.⁹ But because this device has been neither tested nor approved for other applications, the following review applies only to Mirena—referred to throughout this article as the LNG-IUD.

A proven (and often superior) treatment for menorrhagia

Leiomyomas, or fibroids—the most common benign tumor of the female genital tract—are a frequent cause of menorrhagia.¹⁰ OCPs are often used for the treatment of symptomatic fibroids for women who wish to avoid surgery. When compared with low-dose OCPs for the treatment of menorrhagia secondary to fibroids, however, the LNG-IUD resulted in a significantly greater reduction in blood loss.¹¹

While fibroid size does not appear to decrease significantly after insertion of the device, a systematic review found that menstrual bleeding lessens and hemoglobin levels improve in women with symptomatic fibroids.^12,13 The LNG-IUD has also been shown to improve symptoms in women with dysmenorrhea secondary to fibroids.¹⁴

The 52 mg levonorgestrol-releasing IUD has been shown to be effective in treating multiple conditions associated with menorrhagia and dysmenorrhea.Hemostatic disorders. Anticoagulants are vital for women with hemostatic disorders such as Von Willebrand disease, immune thrombocytopenia, or a clotting factor deficiency (See “Is a novel anticoagulant right for your patient?”), but their use may cause or worsen menorrhagia. In such cases, the LNG-IUD appears to be an effective treatment.

In a retrospective case review of 28 women with menorrhagia secondary to various hemostatic disorders, 68% experienced improvement after insertion of the IUD.¹⁵ Another study compared women on anticoagulant therapy for cardiac valve disease (N=40) with and without the LNG-IUD. Compared with the control group, those with the IUD were found to have significant increases in hemoglobin levels 3 months after insertion.¹⁶

Obesity-related uterine bleeding. Obese women are at higher risk for excessive uterine bleeding, the result of increased conversion of plasma androstenedione to estrogen in adipose tissue. In one study evaluating the use of the LNG-IUD in this patient population, 75% of participants experienced a reduction in bleeding.¹⁷

Idiopathic HMB. The IUD is as good as, or better than, other treatments for idiopathic menorrhagia. It results in a significantly higher reduction in both blood loss and days out of work than OCPs.¹ The device also reduces blood loss more effectively than oral medroxyprogesterone,¹⁸ another common approach to idiopathic HMB; and, compared with hysterectomy, it results in similar patient satisfaction—but lower costs and complication rates.¹⁹

In a randomized controlled trial that compared the LNG-IUD with tranexamic acid, mefenamic acid, combined estrogenprogesterone, or progesterone alone over a 2-year period, scores on a menorrhagia symptom scale were significantly higher (indicating greater improvement) in the LNG-IUD group.²⁰

Medical management of endometrial proliferation

Endometrial hyperplasia, often found in women with abnormal uterine bleeding patterns and recurrent anovulatory cycles,²¹ is sometimes treated with supplemental progesterone. According to the Centers for Disease Control and Prevention (CDC)’s US Medical Eligibility Criteria for Contraceptive Use, the LNG-IUD can be used without restriction in this patient population, as well.²²

A systematic review of 9 studies of women with endometrial hyperplasia without atypia found the LNG-IUD to be both safe and effective. In 7 of the 9 studies, 100% of participants experienced disease regression; regression rates for the other 2 studies were 90% and 67%.²³ The only caveat: Both endometrial atypia and endometrial cancer should be excluded prior to IUD insertion.

Adenomyosis is caused by the presence of ectopic endometrial glands and stroma within the myometrium and frequently results in pelvic pain, menorrhagia, and dysmenorrhea.¹⁰ Hysterectomy is often regarded as the mainstay of treatment. But medical management with the LNG-IUD is also an option, as it has demonstrated similar improvements to hysterectomy both in hemoglobin levels and quality of life.²⁴ Three years after insertion of the LNG-IUD to treat moderate or severe adenomyosis-associated dysmenorrhea, one study found, women reported significant improvement in their symptoms—and 73% were satisfied with their treatment.²⁵ The LNG-IUD also appears to decrease uterine volume, although this effect may begin to decrease 2 years after insertion.²⁶

Endometriosis, another common cause of dysmenorrhea and chronic pelvic pain,²⁷ can also be treated with the LNG-IUD. The local progestin administration to pelvic structures that the device provides has been found to significantly decrease both endometrial proliferation and monthly blood flow.²⁸ Additional studies of the LNG-IUD as a treatment for endometriosis and pelvic pain are ongoing and encouraging. After surgery for endometriosis, a Cochrane review found, women who had the LNG-IUD inserted had a lower rate of recurrence of dysmenorrhea than those without it.²⁹

Helping women through perimenopause

Compared with oral or intramuscular progesterone therapy, the LNG-IUD has been found to be superior for the treatment of perimenopausal symptoms.³⁰ Two years after insertion, one study found, perimenopausal women had a 95% reduction in blood loss and a 63% decrease in dysmenorrhea.³¹ The LNG-IUD also provides reliable endometrial protection for women receiving estrogen therapy³² and for those who are taking adjuvant tamoxifen because of a history of breast cancer.^33,34

IUD insertion is a safe office procedure

The Society of Teachers of Family Medicine cites IUD insertion in its core list of routine procedures to be included in family medicine residency programs.³⁵ The risk associated with insertion is small—uterine perforation occurs in about 2.6/1000 insertions³⁶—and there is a small and transient increase in the risk of IUD-related infection in the first few weeks to months after insertion. IUD insertion does not increase the overall risk of pelvic inflammatory disease in women at low risk for sexually transmitted infections.^37,38

Who is not a candidate?

While IUDs are safe for most women, there are several absolute contraindications to the LNG-IUD:
• current breast, cervical, or endometrial cancer
• current pelvic inflammatory disease, cervicitis, chlamydia, or gonorrhea
• having just had a septic abortion.³⁸

Teach patients about the benefits and adverse effects

For women who are potential candidates for the LNG-IUD, education is vital. Evidence suggests that satisfaction levels are very high, provided patients receive adequate counseling about the benefits and adverse effects. Risks (of uterine perforation and infection) are small, as noted earlier.³⁹

Contraceptive efficacy, of course, is a major benefit, and has been well The risk associated with insertion is small—uterine perforation occurs in about 2.6/1000 insertions—and there is a small and transient risk of infection in the first few weeks to months after insertion.documented: The LNG-IUD has an estimated failure rate of just 0.2%.⁴⁰ Unlike user-dependent methods such as OCPs, the patch, and the ring, the IUD has a perfect-use failure rate that is the same as the typical use rate. Thus, it is an excellent choice for women who want to preserve their fertility yet avoid an unintended pregnancy. For women approaching menopause—a time when estrogen may be contraindicated—the LNG-IUD can safely protect women against unwanted pregnancy.

Lower cost, less invasive. The ability to treat HMB and dysmennorhea with an IUD inserted in a family practice setting, without referrals to specialists for additional invasive treatments, increases cost savings.¹⁹ In addition, the LNG-IUD is less invasive and generally more acceptable to women than hysterectomy, endometrial ablation, uterine artery embolization, and myomectomy.^18,41 It leads to a greater reduction in menstrual bleeding than OCPs, oral progestins, tranexamic acid, and oral mefenamic acid.⁴¹ And, unlike some progestational agents, there is no evidence that the LNG-IUD has any adverse effects on bone density, vaginal tone, or urinary continence.⁴²

The LNG-IUD is less invasive and generally more acceptable to women than hysterectomy, endometrial ablation, uterine artery embolization, and myomectomy. Adverse effects. Vaginal spotting is the most commonly reported adverse effect associated with the LNG-IUD, particularly in the first 3 to 6 months postinsertion.⁴² The other common adverse effect is increased cramping, which some women experience in the first few months after insertion. Rarely women may experience ovarian cysts from unruptured follicles, which will regress on their own. Another potential problem is expulsion, which is more common in women who are using the device to control heavy bleeding (8.9%-13.6%).⁴² After the device is in place for several years, many women experience amenorrhea—a side effect that patients who have suffered from HMB and dysmenorrhea may consider a benefit.

CASE After counseling regarding her treatment options, Ms. K decides on the LNG-IUD, which her family physician inserts. At 3-month follow-up, she reports significantly less bleeding and decreased perimenstrual discomfort. If her workup had revealed adenomyosis or a hemostatic disorder, the LNG-IUD would still have been a first-line option.

CORRESPONDENCE

Erin Hendriks, MD, Wayne State University Family Medicine, 1135 West University Drive, Suite 250, Rochester Hills, MI 48307; [email protected]

Morbidity and mortality from atherosclerotic disease have decreased significantly in the last several decades, in large part because of advances in therapies targeting serum lipids—including statins.¹ Since their introduction in 1987, HMG Coenzyme A inhibitors have been intensively studied, and their use has increased dramatically. A recent report from the National Center for Health Statistics reveals that in the years 1999 to 2002, 26% of men ages 65 to 74 years were taking statins; several years later (2005-2008), that number had soared to 50%. In the same time frame, statin use among women ages 65 to 74 went from 24% to 36%.²

Statins lower serum low-density lipoprotein cholesterol (LDL-C) and triglycerides, raise high-density lipoprotein cholesterol (HDL-C), and improve surrogate markers for cardiovascular events. Most importantly, statins reduce the risk for major cardiovascular events, such as myocardial infarction (MI) and death from cardiovascular disease (CVD), in select populations. Yet doubts about the benefits of statins, alone or in combination with other lipid-lowering agents, for certain patient populations remain.

Primary care physicians need to know when, or whether, to add a second lipid-lowering agent to the drug regimen of patients whose response to a statin is less than hoped for, and which patient populations and clinical indicators statins have not been found to help. You’ll find answers, the results of the latest studies, and details of the 2013 cholesterol guideline released last month by the American Heart Association/American College of Cardiology and in this evidence-based update.

A statin is not enough? Don't add these drugs

In patients with very elevated LDL-C or mixed dyslipidemias that fail to reach the desired lipid levels on statin monotherapy, other classes of lipid-modifying agents are often added in an attempt to improve clinical outcomes.³ Fibrates, extended release (niacin, and n-3 polyunsaturated fatty acids have been frequently used for this purpose. But it is only in the last several years that large, well-designed studies have looked closely at patient-oriented outcomes associated with statins in combination with other lipid-modifying drugs.^4-9

Fenofibrate + a statin yields little benefit

The ACCORD lipid placebo-controlled trial studied fenofibrate as a simvastatin add-on in patients with diabetes.⁴ Its findings? While the lipid levels of patients receiving this drug combination improved significantly, the primary endpoint (MI, stroke, or death from cardiovascular causes) was no different from that of the controls, who were taking the statin alone.

Sub-group analysis suggested that the simvastatin-fenofibrate combination benefitted only one particular group: patients with high triglyceride levels (≥204 mg/dL) and low HDL-C (≤34 mg/dL). This finding prompted the US Food and Drug Administration to call for an additional clinical trial to evaluate the effectiveness of add-on therapy with fenofibrate in patients who meet this criteria.¹⁰ The status of such a study is uncertain.

Adding niacin to a statin does more harm than good

The HATS trial, published in 2001, found the addition of niacin to a statin regimen to be beneficial.⁷ But because of the wide confidence interval associated with the clinical endpoints and the small number of subjects (N=160) in that study, larger trials were needed to confirm the positive results. In fact, they found the opposite.

Niacin increases stroke risk. In both the AIM-HIGH⁵ (N=3414) and HPS2-THRIVE⁶ (N=25,673) trials, the addition of extended release niacin not only failed to reduce the risk of major cardiovascular events, it was shown to increase the risk of stroke.

n-3 polyunsaturated fatty acids don’t help much

Studies evaluating the addition of n-3 polyunsaturated fatty acids to statin therapy have had mixed results. The JELIS⁸ trial had more than 18,000 participants, 20% of whom had known coronary artery disease. All were taking statins and randomized to either open-label eicosapentaenoic acid (EPA) 600 mg 3 times daily or placebo. The primary endpoint, a composite of sudden cardiac death, fatal or nonfatal MI, unstable angina, angioplasty, and stenting or coronary artery bypass grafting, was lower in the intervention group: (2.8% vs 3.5%; number needed to treat [NNT]: 143).

It is important to note, however, that only one of the individual components of the primary endpoint—unstable angina—was significantly reduced by EPA (2.1% vs 1.6%; P=.014).8 In the Alpha Omega trial,9 various n-3 polyunsaturated fatty acids were tested in combination with statins. None was found to be superior to placebo in reducing cardiovascular outcomes.

Based on the evidence, the new cholesterol guideline does not support the routine use of these agents to reduce atherosclerotic CVD (See “The new cholesterol guideline: Beyond the headlines”.)¹¹

Statins and kidney disease: Factors to consider

More than half of the deaths in patients with end-stage renal disease (ESRD) are from cardiovascular causes.¹² The relationship between renal dysfunction and cardiovascular events is independent of other risk factors, including a history of CVD. Risk rises with an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73m², with a sharp increase when the rate <45 mL/min/1.73m.² Thus, strategies known to reduce major cardiovascular events in the general population, including statins, have the potential to offer substantial benefit for patients with chronic kidney disease (CKD).

Among patients with heart failure, statin therapy substantially reduced LDL-C, but the time to death or hospitalization for cardiovascular causes was not significantly reduced.Statin use in patients with CKD has been evaluated in post-hoc and subgroup analyses of large clinical trials and, more recently, in RCTs targeting patients with both moderate and end-stage disease (TABLE).^13-18

Three post-hoc analyses of large multicenter, double-blind RCTs^13-15,19 compared patients with normal renal function with those with CKD. All 3 found that moderate or high-intensity statin therapy significantly reduced the incidence of the primary outcome—a composite of major cardiovascular events—compared with either placebo or a lower-intensity statin.

For patients with CKD, drug combo lowered the risk

The SHARP trial¹⁷ (N=9270) was the first large prospective, double-blind, multicenter RCT to compare the effect of a statin plus a second lipid-lowering drug (simvastatin plus ezetimibe) vs placebo in patients with CKD. A third of the participants were on dialysis at the start of the trial (ESRD was defined as starting long-term dialysis or requiring kidney transplantation).

Patients in the intervention group were significantly (17%) less likely to experience a major atherosclerotic event compared with those on placebo. This translated into an NNT of 47 over a period of 4.9 years. (Since no group received only simvastatin, it is not known what role ezetimibe had in the reduction of cardiovascular events.) Although no difference in outcomes was found when the results were stratified based on whether participants were on dialysis, this trial was not adequately powered for this subgroup analysis.¹⁷

Little benefit from statins in patients with end-stage disease

Two major prospective randomized, double-blinded, placebo-controlled, multicenter trials evaluating the effects of statin use on cardiovascular outcomes in ESRD patients on dialysis have been published.^18,19 Both found a significant decline in LDL-C in patients receiving statin therapy. But neither found a significant difference in mortality rates in the statin vs placebo groups.

One group of researchers speculated that the lack of effect may be due to a difference in the pathogenesis of vascular events in patients with and without ESRD. Delayed use of statins until patients have ESRD will offer limited benefit, they concluded, and recommended against routine statin treatment in an attempt to reduce the incidence of CVD in this patient population.¹⁸ Based on these results, the new cholesterol guideline indicates that this group of patients may not benefit from statin therapy.

For patients with heart failure, statins offer limited benefit
More than half of the heart failure (HF) in the United States is caused by ischemic heart disease.²⁰ Improvements in post-MI survival have increased the prevalence of chronic HF.

Statins have a well-established role in the prevention and treatment of atherosclerosis because of their ability to modify the natural course of the disease and reduce major adverse cardiovascular events. Thus, it seems reasonable to assume that, in patients who have or are at high risk for coronary heart disease, statins would help to prevent the occurrence or slow the progression of HF.

Early studies of statins either excluded patients with HF or enrolled so few HF patients that no conclusions could be reached regarding the safety or efficacy of statin use in this population.^21-25 More recently, 2 large RCTs have studied the effect of statins in patients with HF. Both have found them to be ineffective.^26,27

The CORONA trial enrolled elderly patients with HF of ischemic causes and ejection fraction ≤40% (≤35% in patients with New York Heart Association [NYHA] Class II), randomized to either rosuvastatin 10 mg/d or placebo.²⁵ More than 40% of the participants had a history of MI, and more than 60% were NYHA Class III or IV. HF medications were well-managed; more than 90% of the patients were being treated with angiotensin-converting enzyme inhibitors or angiotensin receptor blockers; 75%, with beta-blockers; and 39%, with aldosterone antagonists.

The researchers found no significant difference between the rosuvastatin and placebo groups in the primary outcome of death from cardiovascular causes, nonfatal MI, or nonfatal stroke (11.4% in the rosuvastatin group vs 12.3% among those on placebo; 95% confidence interval, 0.83-1.02; P=.12). No difference was found in patients with a history of MI (13.9% placebo event rate vs 12.7% rosuvastatin arm; P=not significant [NS]). Neither death from worsening HF nor sudden death was reduced.²⁶

There were fewer hospitalizations among those taking rosuvastatin, however (NNT=17 per year). And there was a trend towards a benefit among those with more advanced HF (NYHA III/IV), with primary outcome rates of 12.7% for those in the rosuvastatin group vs 14.2% in the placebo group. (P=NS). Rosuvastatin was safe for HF patients, as most types of adverse events were more common in the placebo group. Assessments of muscle toxicity were similar in both groups.

The GISSI-HF study, another RCT of rosuvastatin vs placebo in HF patients, also showed a lack of benefit from statin treatment.²⁷ Researchers enrolled more than 4500 patients with NYHA Class II to IV HF, from both ischemic and nonischemic causes.26 Similar to the findings in the CORONA trial, LDL-C was substantially reduced by rosuvastatin 10 mg/d (-32% vs no change for those in the placebo group), but this did not translate into clinically relevant endpoints. After 3.9 years of therapy, the primary endpoints of time to death and time to death or hospitalization for cardiovascular causes were not significantly reduced, nor were any of the secondary endpoints.²⁶

"We think it is important that primary care as well as specialty organizations be included in guideline development so that special interests do not interfere with good science." --Jeff Cain, MD, Board chair, AAFPThe similar lack of benefit in these 2 trials is striking in view of the benefit of statins in patients with coronary heart disease but without HF. Given these findings, focusing on optimizing therapies known to reduce mortality in patients with HF rather than adding a statin in an attempt to alter the atherosclerotic process appears to be a better approach. Thus, the recently published cholesterol guideline does not advocate the initiation or continuation of statin therapy in patients with NYHA Class II-IV HF.

Morbidity and mortality from atherosclerotic disease have decreased significantly in the last several decades, in large part because of advances in therapies targeting serum lipids—including statins.¹ Since their introduction in 1987, HMG Coenzyme A inhibitors have been intensively studied, and their use has increased dramatically. A recent report from the National Center for Health Statistics reveals that in the years 1999 to 2002, 26% of men ages 65 to 74 years were taking statins; several years later (2005-2008), that number had soared to 50%. In the same time frame, statin use among women ages 65 to 74 went from 24% to 36%.²

Statins lower serum low-density lipoprotein cholesterol (LDL-C) and triglycerides, raise high-density lipoprotein cholesterol (HDL-C), and improve surrogate markers for cardiovascular events. Most importantly, statins reduce the risk for major cardiovascular events, such as myocardial infarction (MI) and death from cardiovascular disease (CVD), in select populations. Yet doubts about the benefits of statins, alone or in combination with other lipid-lowering agents, for certain patient populations remain.

Primary care physicians need to know when, or whether, to add a second lipid-lowering agent to the drug regimen of patients whose response to a statin is less than hoped for, and which patient populations and clinical indicators statins have not been found to help. You’ll find answers, the results of the latest studies, and details of the 2013 cholesterol guideline released last month by the American Heart Association/American College of Cardiology and in this evidence-based update.

A statin is not enough? Don't add these drugs

In patients with very elevated LDL-C or mixed dyslipidemias that fail to reach the desired lipid levels on statin monotherapy, other classes of lipid-modifying agents are often added in an attempt to improve clinical outcomes.³ Fibrates, extended release (niacin, and n-3 polyunsaturated fatty acids have been frequently used for this purpose. But it is only in the last several years that large, well-designed studies have looked closely at patient-oriented outcomes associated with statins in combination with other lipid-modifying drugs.^4-9

Fenofibrate + a statin yields little benefit

The ACCORD lipid placebo-controlled trial studied fenofibrate as a simvastatin add-on in patients with diabetes.⁴ Its findings? While the lipid levels of patients receiving this drug combination improved significantly, the primary endpoint (MI, stroke, or death from cardiovascular causes) was no different from that of the controls, who were taking the statin alone.

Sub-group analysis suggested that the simvastatin-fenofibrate combination benefitted only one particular group: patients with high triglyceride levels (≥204 mg/dL) and low HDL-C (≤34 mg/dL). This finding prompted the US Food and Drug Administration to call for an additional clinical trial to evaluate the effectiveness of add-on therapy with fenofibrate in patients who meet this criteria.¹⁰ The status of such a study is uncertain.

Adding niacin to a statin does more harm than good

The HATS trial, published in 2001, found the addition of niacin to a statin regimen to be beneficial.⁷ But because of the wide confidence interval associated with the clinical endpoints and the small number of subjects (N=160) in that study, larger trials were needed to confirm the positive results. In fact, they found the opposite.

Niacin increases stroke risk. In both the AIM-HIGH⁵ (N=3414) and HPS2-THRIVE⁶ (N=25,673) trials, the addition of extended release niacin not only failed to reduce the risk of major cardiovascular events, it was shown to increase the risk of stroke.

n-3 polyunsaturated fatty acids don’t help much

Studies evaluating the addition of n-3 polyunsaturated fatty acids to statin therapy have had mixed results. The JELIS⁸ trial had more than 18,000 participants, 20% of whom had known coronary artery disease. All were taking statins and randomized to either open-label eicosapentaenoic acid (EPA) 600 mg 3 times daily or placebo. The primary endpoint, a composite of sudden cardiac death, fatal or nonfatal MI, unstable angina, angioplasty, and stenting or coronary artery bypass grafting, was lower in the intervention group: (2.8% vs 3.5%; number needed to treat [NNT]: 143).

It is important to note, however, that only one of the individual components of the primary endpoint—unstable angina—was significantly reduced by EPA (2.1% vs 1.6%; P=.014).8 In the Alpha Omega trial,9 various n-3 polyunsaturated fatty acids were tested in combination with statins. None was found to be superior to placebo in reducing cardiovascular outcomes.

Based on the evidence, the new cholesterol guideline does not support the routine use of these agents to reduce atherosclerotic CVD (See “The new cholesterol guideline: Beyond the headlines”.)¹¹

Statins and kidney disease: Factors to consider

More than half of the deaths in patients with end-stage renal disease (ESRD) are from cardiovascular causes.¹² The relationship between renal dysfunction and cardiovascular events is independent of other risk factors, including a history of CVD. Risk rises with an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73m², with a sharp increase when the rate <45 mL/min/1.73m.² Thus, strategies known to reduce major cardiovascular events in the general population, including statins, have the potential to offer substantial benefit for patients with chronic kidney disease (CKD).

Among patients with heart failure, statin therapy substantially reduced LDL-C, but the time to death or hospitalization for cardiovascular causes was not significantly reduced.Statin use in patients with CKD has been evaluated in post-hoc and subgroup analyses of large clinical trials and, more recently, in RCTs targeting patients with both moderate and end-stage disease (TABLE).^13-18

Three post-hoc analyses of large multicenter, double-blind RCTs^13-15,19 compared patients with normal renal function with those with CKD. All 3 found that moderate or high-intensity statin therapy significantly reduced the incidence of the primary outcome—a composite of major cardiovascular events—compared with either placebo or a lower-intensity statin.

For patients with CKD, drug combo lowered the risk

The SHARP trial¹⁷ (N=9270) was the first large prospective, double-blind, multicenter RCT to compare the effect of a statin plus a second lipid-lowering drug (simvastatin plus ezetimibe) vs placebo in patients with CKD. A third of the participants were on dialysis at the start of the trial (ESRD was defined as starting long-term dialysis or requiring kidney transplantation).

Patients in the intervention group were significantly (17%) less likely to experience a major atherosclerotic event compared with those on placebo. This translated into an NNT of 47 over a period of 4.9 years. (Since no group received only simvastatin, it is not known what role ezetimibe had in the reduction of cardiovascular events.) Although no difference in outcomes was found when the results were stratified based on whether participants were on dialysis, this trial was not adequately powered for this subgroup analysis.¹⁷

Little benefit from statins in patients with end-stage disease

Two major prospective randomized, double-blinded, placebo-controlled, multicenter trials evaluating the effects of statin use on cardiovascular outcomes in ESRD patients on dialysis have been published.^18,19 Both found a significant decline in LDL-C in patients receiving statin therapy. But neither found a significant difference in mortality rates in the statin vs placebo groups.

One group of researchers speculated that the lack of effect may be due to a difference in the pathogenesis of vascular events in patients with and without ESRD. Delayed use of statins until patients have ESRD will offer limited benefit, they concluded, and recommended against routine statin treatment in an attempt to reduce the incidence of CVD in this patient population.¹⁸ Based on these results, the new cholesterol guideline indicates that this group of patients may not benefit from statin therapy.

For patients with heart failure, statins offer limited benefit
More than half of the heart failure (HF) in the United States is caused by ischemic heart disease.²⁰ Improvements in post-MI survival have increased the prevalence of chronic HF.

Statins have a well-established role in the prevention and treatment of atherosclerosis because of their ability to modify the natural course of the disease and reduce major adverse cardiovascular events. Thus, it seems reasonable to assume that, in patients who have or are at high risk for coronary heart disease, statins would help to prevent the occurrence or slow the progression of HF.

Early studies of statins either excluded patients with HF or enrolled so few HF patients that no conclusions could be reached regarding the safety or efficacy of statin use in this population.^21-25 More recently, 2 large RCTs have studied the effect of statins in patients with HF. Both have found them to be ineffective.^26,27

The CORONA trial enrolled elderly patients with HF of ischemic causes and ejection fraction ≤40% (≤35% in patients with New York Heart Association [NYHA] Class II), randomized to either rosuvastatin 10 mg/d or placebo.²⁵ More than 40% of the participants had a history of MI, and more than 60% were NYHA Class III or IV. HF medications were well-managed; more than 90% of the patients were being treated with angiotensin-converting enzyme inhibitors or angiotensin receptor blockers; 75%, with beta-blockers; and 39%, with aldosterone antagonists.

The researchers found no significant difference between the rosuvastatin and placebo groups in the primary outcome of death from cardiovascular causes, nonfatal MI, or nonfatal stroke (11.4% in the rosuvastatin group vs 12.3% among those on placebo; 95% confidence interval, 0.83-1.02; P=.12). No difference was found in patients with a history of MI (13.9% placebo event rate vs 12.7% rosuvastatin arm; P=not significant [NS]). Neither death from worsening HF nor sudden death was reduced.²⁶

There were fewer hospitalizations among those taking rosuvastatin, however (NNT=17 per year). And there was a trend towards a benefit among those with more advanced HF (NYHA III/IV), with primary outcome rates of 12.7% for those in the rosuvastatin group vs 14.2% in the placebo group. (P=NS). Rosuvastatin was safe for HF patients, as most types of adverse events were more common in the placebo group. Assessments of muscle toxicity were similar in both groups.

The GISSI-HF study, another RCT of rosuvastatin vs placebo in HF patients, also showed a lack of benefit from statin treatment.²⁷ Researchers enrolled more than 4500 patients with NYHA Class II to IV HF, from both ischemic and nonischemic causes.26 Similar to the findings in the CORONA trial, LDL-C was substantially reduced by rosuvastatin 10 mg/d (-32% vs no change for those in the placebo group), but this did not translate into clinically relevant endpoints. After 3.9 years of therapy, the primary endpoints of time to death and time to death or hospitalization for cardiovascular causes were not significantly reduced, nor were any of the secondary endpoints.²⁶

"We think it is important that primary care as well as specialty organizations be included in guideline development so that special interests do not interfere with good science." --Jeff Cain, MD, Board chair, AAFPThe similar lack of benefit in these 2 trials is striking in view of the benefit of statins in patients with coronary heart disease but without HF. Given these findings, focusing on optimizing therapies known to reduce mortality in patients with HF rather than adding a statin in an attempt to alter the atherosclerotic process appears to be a better approach. Thus, the recently published cholesterol guideline does not advocate the initiation or continuation of statin therapy in patients with NYHA Class II-IV HF.

Morbidity and mortality from atherosclerotic disease have decreased significantly in the last several decades, in large part because of advances in therapies targeting serum lipids—including statins.¹ Since their introduction in 1987, HMG Coenzyme A inhibitors have been intensively studied, and their use has increased dramatically. A recent report from the National Center for Health Statistics reveals that in the years 1999 to 2002, 26% of men ages 65 to 74 years were taking statins; several years later (2005-2008), that number had soared to 50%. In the same time frame, statin use among women ages 65 to 74 went from 24% to 36%.²

Statins lower serum low-density lipoprotein cholesterol (LDL-C) and triglycerides, raise high-density lipoprotein cholesterol (HDL-C), and improve surrogate markers for cardiovascular events. Most importantly, statins reduce the risk for major cardiovascular events, such as myocardial infarction (MI) and death from cardiovascular disease (CVD), in select populations. Yet doubts about the benefits of statins, alone or in combination with other lipid-lowering agents, for certain patient populations remain.

Primary care physicians need to know when, or whether, to add a second lipid-lowering agent to the drug regimen of patients whose response to a statin is less than hoped for, and which patient populations and clinical indicators statins have not been found to help. You’ll find answers, the results of the latest studies, and details of the 2013 cholesterol guideline released last month by the American Heart Association/American College of Cardiology and in this evidence-based update.

A statin is not enough? Don't add these drugs

In patients with very elevated LDL-C or mixed dyslipidemias that fail to reach the desired lipid levels on statin monotherapy, other classes of lipid-modifying agents are often added in an attempt to improve clinical outcomes.³ Fibrates, extended release (niacin, and n-3 polyunsaturated fatty acids have been frequently used for this purpose. But it is only in the last several years that large, well-designed studies have looked closely at patient-oriented outcomes associated with statins in combination with other lipid-modifying drugs.^4-9

Fenofibrate + a statin yields little benefit

The ACCORD lipid placebo-controlled trial studied fenofibrate as a simvastatin add-on in patients with diabetes.⁴ Its findings? While the lipid levels of patients receiving this drug combination improved significantly, the primary endpoint (MI, stroke, or death from cardiovascular causes) was no different from that of the controls, who were taking the statin alone.

Sub-group analysis suggested that the simvastatin-fenofibrate combination benefitted only one particular group: patients with high triglyceride levels (≥204 mg/dL) and low HDL-C (≤34 mg/dL). This finding prompted the US Food and Drug Administration to call for an additional clinical trial to evaluate the effectiveness of add-on therapy with fenofibrate in patients who meet this criteria.¹⁰ The status of such a study is uncertain.

Adding niacin to a statin does more harm than good

The HATS trial, published in 2001, found the addition of niacin to a statin regimen to be beneficial.⁷ But because of the wide confidence interval associated with the clinical endpoints and the small number of subjects (N=160) in that study, larger trials were needed to confirm the positive results. In fact, they found the opposite.

Niacin increases stroke risk. In both the AIM-HIGH⁵ (N=3414) and HPS2-THRIVE⁶ (N=25,673) trials, the addition of extended release niacin not only failed to reduce the risk of major cardiovascular events, it was shown to increase the risk of stroke.

n-3 polyunsaturated fatty acids don’t help much

Studies evaluating the addition of n-3 polyunsaturated fatty acids to statin therapy have had mixed results. The JELIS⁸ trial had more than 18,000 participants, 20% of whom had known coronary artery disease. All were taking statins and randomized to either open-label eicosapentaenoic acid (EPA) 600 mg 3 times daily or placebo. The primary endpoint, a composite of sudden cardiac death, fatal or nonfatal MI, unstable angina, angioplasty, and stenting or coronary artery bypass grafting, was lower in the intervention group: (2.8% vs 3.5%; number needed to treat [NNT]: 143).

It is important to note, however, that only one of the individual components of the primary endpoint—unstable angina—was significantly reduced by EPA (2.1% vs 1.6%; P=.014).8 In the Alpha Omega trial,9 various n-3 polyunsaturated fatty acids were tested in combination with statins. None was found to be superior to placebo in reducing cardiovascular outcomes.

Based on the evidence, the new cholesterol guideline does not support the routine use of these agents to reduce atherosclerotic CVD (See “The new cholesterol guideline: Beyond the headlines”.)¹¹

Statins and kidney disease: Factors to consider

More than half of the deaths in patients with end-stage renal disease (ESRD) are from cardiovascular causes.¹² The relationship between renal dysfunction and cardiovascular events is independent of other risk factors, including a history of CVD. Risk rises with an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73m², with a sharp increase when the rate <45 mL/min/1.73m.² Thus, strategies known to reduce major cardiovascular events in the general population, including statins, have the potential to offer substantial benefit for patients with chronic kidney disease (CKD).

Among patients with heart failure, statin therapy substantially reduced LDL-C, but the time to death or hospitalization for cardiovascular causes was not significantly reduced.Statin use in patients with CKD has been evaluated in post-hoc and subgroup analyses of large clinical trials and, more recently, in RCTs targeting patients with both moderate and end-stage disease (TABLE).^13-18

Three post-hoc analyses of large multicenter, double-blind RCTs^13-15,19 compared patients with normal renal function with those with CKD. All 3 found that moderate or high-intensity statin therapy significantly reduced the incidence of the primary outcome—a composite of major cardiovascular events—compared with either placebo or a lower-intensity statin.

For patients with CKD, drug combo lowered the risk

The SHARP trial¹⁷ (N=9270) was the first large prospective, double-blind, multicenter RCT to compare the effect of a statin plus a second lipid-lowering drug (simvastatin plus ezetimibe) vs placebo in patients with CKD. A third of the participants were on dialysis at the start of the trial (ESRD was defined as starting long-term dialysis or requiring kidney transplantation).

Patients in the intervention group were significantly (17%) less likely to experience a major atherosclerotic event compared with those on placebo. This translated into an NNT of 47 over a period of 4.9 years. (Since no group received only simvastatin, it is not known what role ezetimibe had in the reduction of cardiovascular events.) Although no difference in outcomes was found when the results were stratified based on whether participants were on dialysis, this trial was not adequately powered for this subgroup analysis.¹⁷

Little benefit from statins in patients with end-stage disease

Two major prospective randomized, double-blinded, placebo-controlled, multicenter trials evaluating the effects of statin use on cardiovascular outcomes in ESRD patients on dialysis have been published.^18,19 Both found a significant decline in LDL-C in patients receiving statin therapy. But neither found a significant difference in mortality rates in the statin vs placebo groups.

One group of researchers speculated that the lack of effect may be due to a difference in the pathogenesis of vascular events in patients with and without ESRD. Delayed use of statins until patients have ESRD will offer limited benefit, they concluded, and recommended against routine statin treatment in an attempt to reduce the incidence of CVD in this patient population.¹⁸ Based on these results, the new cholesterol guideline indicates that this group of patients may not benefit from statin therapy.

For patients with heart failure, statins offer limited benefit
More than half of the heart failure (HF) in the United States is caused by ischemic heart disease.²⁰ Improvements in post-MI survival have increased the prevalence of chronic HF.

Statins have a well-established role in the prevention and treatment of atherosclerosis because of their ability to modify the natural course of the disease and reduce major adverse cardiovascular events. Thus, it seems reasonable to assume that, in patients who have or are at high risk for coronary heart disease, statins would help to prevent the occurrence or slow the progression of HF.

Early studies of statins either excluded patients with HF or enrolled so few HF patients that no conclusions could be reached regarding the safety or efficacy of statin use in this population.^21-25 More recently, 2 large RCTs have studied the effect of statins in patients with HF. Both have found them to be ineffective.^26,27

The CORONA trial enrolled elderly patients with HF of ischemic causes and ejection fraction ≤40% (≤35% in patients with New York Heart Association [NYHA] Class II), randomized to either rosuvastatin 10 mg/d or placebo.²⁵ More than 40% of the participants had a history of MI, and more than 60% were NYHA Class III or IV. HF medications were well-managed; more than 90% of the patients were being treated with angiotensin-converting enzyme inhibitors or angiotensin receptor blockers; 75%, with beta-blockers; and 39%, with aldosterone antagonists.

The researchers found no significant difference between the rosuvastatin and placebo groups in the primary outcome of death from cardiovascular causes, nonfatal MI, or nonfatal stroke (11.4% in the rosuvastatin group vs 12.3% among those on placebo; 95% confidence interval, 0.83-1.02; P=.12). No difference was found in patients with a history of MI (13.9% placebo event rate vs 12.7% rosuvastatin arm; P=not significant [NS]). Neither death from worsening HF nor sudden death was reduced.²⁶

There were fewer hospitalizations among those taking rosuvastatin, however (NNT=17 per year). And there was a trend towards a benefit among those with more advanced HF (NYHA III/IV), with primary outcome rates of 12.7% for those in the rosuvastatin group vs 14.2% in the placebo group. (P=NS). Rosuvastatin was safe for HF patients, as most types of adverse events were more common in the placebo group. Assessments of muscle toxicity were similar in both groups.

The GISSI-HF study, another RCT of rosuvastatin vs placebo in HF patients, also showed a lack of benefit from statin treatment.²⁷ Researchers enrolled more than 4500 patients with NYHA Class II to IV HF, from both ischemic and nonischemic causes.26 Similar to the findings in the CORONA trial, LDL-C was substantially reduced by rosuvastatin 10 mg/d (-32% vs no change for those in the placebo group), but this did not translate into clinically relevant endpoints. After 3.9 years of therapy, the primary endpoints of time to death and time to death or hospitalization for cardiovascular causes were not significantly reduced, nor were any of the secondary endpoints.²⁶

"We think it is important that primary care as well as specialty organizations be included in guideline development so that special interests do not interfere with good science." --Jeff Cain, MD, Board chair, AAFPThe similar lack of benefit in these 2 trials is striking in view of the benefit of statins in patients with coronary heart disease but without HF. Given these findings, focusing on optimizing therapies known to reduce mortality in patients with HF rather than adding a statin in an attempt to alter the atherosclerotic process appears to be a better approach. Thus, the recently published cholesterol guideline does not advocate the initiation or continuation of statin therapy in patients with NYHA Class II-IV HF.

CASE › A mother brings in her 3-year-old son for a regular check-up. Her only concern is that for the past 2 weeks, he has not been sleeping through the night. She indicates that the sleeping problem began after he was diagnosed with and treated for an ear infection. Fortunately, this hasn’t affected his daily activity or energy, she says.

The child’s appetite is good and he speaks clearly, in 5-word sentences. He is meeting his developmental milestones, and appears well—sitting in his mother’s lap and playing with her smartphone. His head, eyes, ears, nose, and throat exam only turns up fluid behind his left tympanic membrane, which is not red or bulging. The right membrane appears normal, and he has no cervical lymphadenopathy. The rest of his exam is normal. How would you manage a patient like this?

"Glue ear" is often asymptomatic

Otitis media with effusion (OME) is defined as middle-ear effusion (MEE) in the absence of acute signs of infection. In children, OME—also referred to as “glue ear”—most often arises after acute otitis media (AOM). In adults, it often occurs in association with eustachian tube dysfunction, although OME is a separate diagnosis. (To learn more, see “What about OME in adults?”^1,2.)

Experts have found it difficult to determine the exact incidence of OME because it is often asymptomatic. In addition, many cases quickly resolve on their own, making it challenging to diagnose. A 2-year prospective study of 2- to 6-year-old preschoolers revealed that MEE, diagnosed via monthly otoscopy and tympanometry, occurred at least once in 53% of the children in the first year and in 61% of the children in the second year.³ A second study followed 7-year-olds monthly for one year and found a 31% incidence of MEE using tympanometry.⁴ In the 25% of children found to have persistent MEE, the researchers noted spontaneous recovery after an average of 2 months.

We believe that nearly all children have experienced one episode of OME by the age of 3 years, but the prevalence of OME varies with age and the time of year. It is more prevalent in the winter than the summer months.⁵ OME is more common in Caucasian children than in African American or Asian children.⁶

Etiology remains elusive

Risk factors for children include a family history of OME, bottle-feeding, day care attendance, exposure to tobacco smoke, and a personal history of allergies.^7,8 One study conducted on mice suggested that inherited structural abnormalities of the middle ear and eustachian tube may play a role as well.⁹ Some have suggested that effusions of OME in children result from chronic inflammation, for example, after AOM, and that the effusions are sterile; however, recent studies have demonstrated that a biofilm is formed by bacterial otopathogens in the effusion.^10-12 The common pathogens found include nontypeable Haemophilus influenza, Streptococcus pneumoniae, and Moraxella catarrhalis. Inflammatory exudate or neutrophil infiltration is rare in the fluid, however.

The contribution of allergies to OME in children remains somewhat controversial. A retrospective review from the United Kingdom of 209 children with OME found a history of allergic rhinitis, asthma, and eczema in 89%, 36%, and 24%, respectively.¹³ However, this study was done at an allergy clinic, and it is possible that the data from the clinic’s specialized patient population are not generalizable. Gastroesophageal reflux may also be associated with OME in children. However, studies measuring the concentration of pepsin and pepsinogen in middle-ear fluid have provided conflicting results.^14,15

Look for these signs and symptoms

OME is often asymptomatic. If a patient has clinical signs of an acute illness, including fever and an erythematous tympanic membrane, it’s important to evaluate for another cause. OME can present with hearing loss or a sense of fullness in the ear. While an infant cannot express the hearing loss, the parent may detect it when observing and interacting mwith the child. Parents are also likely to report that the child is experiencing sleep disturbances.¹⁶

Vertigo may occur with OME, although not often. It may manifest itself if the child stumbles or falls. An older child or adult with vertigo may say that it feels like the room is spinning.

Diagnosis relies on pneumatic otoscopy

On physical exam, the patient will likely appear well. Otoscopic examination reveals fluid behind a normal or retracted tympanic membrane; the fluid is often clear or yellowish in color.

A subcommittee comprised of members of the American Academy of Pediatrics, American Academy of Family Physicians, and the American Academy of Otolaryngology-Head and Neck Surgery (AAP/AAFP/AAOHNS) published a clinical practice guideline in 2004 that delineates the current diagnosis and management of children between 2 months and 12 years of age with OME.¹⁷

Pneumatic otoscopy, which can reveal decreased or absent movement of the tympanic membrane (the result of fluid behind the membrane), is the primary diagnostic method recommended by the guideline. Tympanometry and acoustic reflectometry may also be used to make the diagnosis, especially when the presence of MEE is difficult to determine using pneumatic otoscopy.

CASE › Upon further discussion with the patient’s mother, you learn that the boy goes to day care 3 days a week and stays with his grandmother 2 days a week. His grandmother smokes outside of the home when he is staying with her.

After reviewing these risk factors with the mother (including the importance of smoking cessation for the grandmother), the mother asks if he needs antibiotics or a referral to a specialist for treating his OME.

How best to approach treatment

There are several management options to choose from, including watchful waiting, medication, and/or surgery. (Another option, autoinflation, which has shown some short-term benefits, is described in “Should you recommend autoinflation?”^17-19.)

The goals of management are to resolve the effusion, restore normal hearing (if diminished secondary to the effusion), and prevent future episodes or sequelae. The most significant complication of OME is permanent conductive hearing loss, but tinnitus, cholesteatoma, or tympanosclerosis may also occur.

In most patients, OME resolves without medical intervention. If additional action is required, however, the following options may be explored.

Medication. While the AAP/AAFP/AAOHNS guideline recommends against routine antibiotics for OME,¹⁷ it does note that a short course may provide short-term benefit to some patients (eg, those for whom a specialist referral or surgery is being considered).

A separate meta-analysis found that antibiotics improve clearance of the effusion within the first month after treatment (rate difference [RD]=0.16; 95% confidence interval [CI], 0.03-0.29 in 12 studies analyzed), but effusion relapses were common, and no significant benefit was noted past the first month (RD=0.06; 95% CI, -0.03 to 0.14 in 8 studies).²⁰

If you do use antibiotics, a 10- to 14-day course is preferred.¹⁷ Amoxicillin, amoxicillin-clavulanate and ceftibuten have been evaluated in separate clinical trials, but none has been clearly shown to have significant advantage over any other.^21,22

Antihistamines, decongestants, and oral and intranasal corticosteroids have little effect on OME in children and are not recommended.¹⁷ A Cochrane review including 16 studies found that children receiving antihistamines and decongestants are unlikely to see their symptoms improve significantly, and many patients experience adverse effects from the medications²³ (number needed to harm=9).

A randomized, double-blind trial involving 144 children <9 years of age with OME for at least 2 months evaluated 4 regimens involving amoxicillin alone or in combination with prednisolone. Children in the amoxicillin+prednisolone arms were significantly more likely to clear their effusions at 2 weeks (number needed to treat=6; P=.03), but not at 4 weeks (P=.12). At 4-month follow-up, effusions had recurred in 68.4% and 69.2% of those receiving amoxicillin+prednisolone and those receiving amoxicillin alone, respectively (P= .94).²⁴

Surgery—or not? The AAP/AAFP/AAOHNS guideline recommends physicians perform hearing testing when OME is present for 3 months or longer, or at any time if language delay, learning problems, or a significant hearing loss is suspected in a child with OME. The results of the hearing test can help determine how to proceed, based on the hearing level noted for the better hearing ear.

You can manage children with hearing loss ≤20 dB and without speech, language, or developmental problems with watchful waiting. Children with hearing loss of 21 to 39 dB can be managed with watchful waiting or referred for surgery. If watchful waiting is pursued, there are interventions at home and at school that can help. These include speaking near the child, facing the child when speaking, and providing accommodations in school so the child sits closer to the teacher. Consider re-examination and repeat hearing tests every 3 to 6 months until the effusion has resolved or the child develops symptoms indicating surgical referral.

When hearing loss is ≥40 dB, the AAP/AAFP/AAOHNS guideline recommends that you make a referral for surgical evaluation (ALGORITHM).¹⁷

Other indications for referral to a surgeon for evaluation of tympanostomy tube placement include situations in which there is:
• structural damage to the tympanic membrane or middle ear (prompt referral is recommended)
• OME of ≥4 months’ duration with persistent hearing loss (≥40 dB) or other signs or symptoms related to the effusion
• bilateral OME for ≥3 months, unilateral OME ≥6 months, or total duration of any degree of OME ≥12 months.¹⁷

Any decision regarding surgery should involve an otolaryngologist, the primary care provider, and the patient and/or family. The AAP/AAFP/AAOHNS guideline recommends against adenoidectomy in children with persistent OME without an indication for the procedure other than OME (eg, chronic sinusitis or nasal obstruction).¹⁷

Keep in mind that evidence of lasting benefit (>12 months) is limited for surgery in most patients, and the surgical and anesthetic risks must be considered before moving forward.¹⁷ (For more on the evidence regarding surgery, see “Cochrane weighs in on tympanostomy tubes”.²⁵) Tonsillectomy also does not appear to affect outcomes and is not advised.¹⁷

When a referral is always needed. Regardless of hearing status, promptly refer children with recurrent or persistent OME who are at risk of speech, language, or learning problems (including those with autism spectrum disorder, developmental delay, Down’s syndrome, diagnosed speech or language delay, or craniofacial disorders such as cleft palate) to a specialist.¹⁷

CASE › You tell your young patient’s mother that watchful waiting is appropriate at this point, since his acute otitis media was only 2 weeks ago, and his OME likely started after the acute infection. Given that his speech is clear and he is otherwise meeting his milestones, you tell her that he does not need a referral at this time, but that she should bring him back in 4 weeks for reassessment. At the next visit, his effusion has resolved, and his mother reports he is sleeping well through the night again.

CASE › A mother brings in her 3-year-old son for a regular check-up. Her only concern is that for the past 2 weeks, he has not been sleeping through the night. She indicates that the sleeping problem began after he was diagnosed with and treated for an ear infection. Fortunately, this hasn’t affected his daily activity or energy, she says.

The child’s appetite is good and he speaks clearly, in 5-word sentences. He is meeting his developmental milestones, and appears well—sitting in his mother’s lap and playing with her smartphone. His head, eyes, ears, nose, and throat exam only turns up fluid behind his left tympanic membrane, which is not red or bulging. The right membrane appears normal, and he has no cervical lymphadenopathy. The rest of his exam is normal. How would you manage a patient like this?

"Glue ear" is often asymptomatic

Otitis media with effusion (OME) is defined as middle-ear effusion (MEE) in the absence of acute signs of infection. In children, OME—also referred to as “glue ear”—most often arises after acute otitis media (AOM). In adults, it often occurs in association with eustachian tube dysfunction, although OME is a separate diagnosis. (To learn more, see “What about OME in adults?”^1,2.)

Experts have found it difficult to determine the exact incidence of OME because it is often asymptomatic. In addition, many cases quickly resolve on their own, making it challenging to diagnose. A 2-year prospective study of 2- to 6-year-old preschoolers revealed that MEE, diagnosed via monthly otoscopy and tympanometry, occurred at least once in 53% of the children in the first year and in 61% of the children in the second year.³ A second study followed 7-year-olds monthly for one year and found a 31% incidence of MEE using tympanometry.⁴ In the 25% of children found to have persistent MEE, the researchers noted spontaneous recovery after an average of 2 months.

We believe that nearly all children have experienced one episode of OME by the age of 3 years, but the prevalence of OME varies with age and the time of year. It is more prevalent in the winter than the summer months.⁵ OME is more common in Caucasian children than in African American or Asian children.⁶

Etiology remains elusive

Risk factors for children include a family history of OME, bottle-feeding, day care attendance, exposure to tobacco smoke, and a personal history of allergies.^7,8 One study conducted on mice suggested that inherited structural abnormalities of the middle ear and eustachian tube may play a role as well.⁹ Some have suggested that effusions of OME in children result from chronic inflammation, for example, after AOM, and that the effusions are sterile; however, recent studies have demonstrated that a biofilm is formed by bacterial otopathogens in the effusion.^10-12 The common pathogens found include nontypeable Haemophilus influenza, Streptococcus pneumoniae, and Moraxella catarrhalis. Inflammatory exudate or neutrophil infiltration is rare in the fluid, however.

The contribution of allergies to OME in children remains somewhat controversial. A retrospective review from the United Kingdom of 209 children with OME found a history of allergic rhinitis, asthma, and eczema in 89%, 36%, and 24%, respectively.¹³ However, this study was done at an allergy clinic, and it is possible that the data from the clinic’s specialized patient population are not generalizable. Gastroesophageal reflux may also be associated with OME in children. However, studies measuring the concentration of pepsin and pepsinogen in middle-ear fluid have provided conflicting results.^14,15

Look for these signs and symptoms

OME is often asymptomatic. If a patient has clinical signs of an acute illness, including fever and an erythematous tympanic membrane, it’s important to evaluate for another cause. OME can present with hearing loss or a sense of fullness in the ear. While an infant cannot express the hearing loss, the parent may detect it when observing and interacting mwith the child. Parents are also likely to report that the child is experiencing sleep disturbances.¹⁶

Vertigo may occur with OME, although not often. It may manifest itself if the child stumbles or falls. An older child or adult with vertigo may say that it feels like the room is spinning.

Diagnosis relies on pneumatic otoscopy

On physical exam, the patient will likely appear well. Otoscopic examination reveals fluid behind a normal or retracted tympanic membrane; the fluid is often clear or yellowish in color.

A subcommittee comprised of members of the American Academy of Pediatrics, American Academy of Family Physicians, and the American Academy of Otolaryngology-Head and Neck Surgery (AAP/AAFP/AAOHNS) published a clinical practice guideline in 2004 that delineates the current diagnosis and management of children between 2 months and 12 years of age with OME.¹⁷

Pneumatic otoscopy, which can reveal decreased or absent movement of the tympanic membrane (the result of fluid behind the membrane), is the primary diagnostic method recommended by the guideline. Tympanometry and acoustic reflectometry may also be used to make the diagnosis, especially when the presence of MEE is difficult to determine using pneumatic otoscopy.

CASE › Upon further discussion with the patient’s mother, you learn that the boy goes to day care 3 days a week and stays with his grandmother 2 days a week. His grandmother smokes outside of the home when he is staying with her.

After reviewing these risk factors with the mother (including the importance of smoking cessation for the grandmother), the mother asks if he needs antibiotics or a referral to a specialist for treating his OME.

How best to approach treatment

There are several management options to choose from, including watchful waiting, medication, and/or surgery. (Another option, autoinflation, which has shown some short-term benefits, is described in “Should you recommend autoinflation?”^17-19.)

The goals of management are to resolve the effusion, restore normal hearing (if diminished secondary to the effusion), and prevent future episodes or sequelae. The most significant complication of OME is permanent conductive hearing loss, but tinnitus, cholesteatoma, or tympanosclerosis may also occur.

In most patients, OME resolves without medical intervention. If additional action is required, however, the following options may be explored.

Medication. While the AAP/AAFP/AAOHNS guideline recommends against routine antibiotics for OME,¹⁷ it does note that a short course may provide short-term benefit to some patients (eg, those for whom a specialist referral or surgery is being considered).

A separate meta-analysis found that antibiotics improve clearance of the effusion within the first month after treatment (rate difference [RD]=0.16; 95% confidence interval [CI], 0.03-0.29 in 12 studies analyzed), but effusion relapses were common, and no significant benefit was noted past the first month (RD=0.06; 95% CI, -0.03 to 0.14 in 8 studies).²⁰

If you do use antibiotics, a 10- to 14-day course is preferred.¹⁷ Amoxicillin, amoxicillin-clavulanate and ceftibuten have been evaluated in separate clinical trials, but none has been clearly shown to have significant advantage over any other.^21,22

Antihistamines, decongestants, and oral and intranasal corticosteroids have little effect on OME in children and are not recommended.¹⁷ A Cochrane review including 16 studies found that children receiving antihistamines and decongestants are unlikely to see their symptoms improve significantly, and many patients experience adverse effects from the medications²³ (number needed to harm=9).

A randomized, double-blind trial involving 144 children <9 years of age with OME for at least 2 months evaluated 4 regimens involving amoxicillin alone or in combination with prednisolone. Children in the amoxicillin+prednisolone arms were significantly more likely to clear their effusions at 2 weeks (number needed to treat=6; P=.03), but not at 4 weeks (P=.12). At 4-month follow-up, effusions had recurred in 68.4% and 69.2% of those receiving amoxicillin+prednisolone and those receiving amoxicillin alone, respectively (P= .94).²⁴

Surgery—or not? The AAP/AAFP/AAOHNS guideline recommends physicians perform hearing testing when OME is present for 3 months or longer, or at any time if language delay, learning problems, or a significant hearing loss is suspected in a child with OME. The results of the hearing test can help determine how to proceed, based on the hearing level noted for the better hearing ear.

You can manage children with hearing loss ≤20 dB and without speech, language, or developmental problems with watchful waiting. Children with hearing loss of 21 to 39 dB can be managed with watchful waiting or referred for surgery. If watchful waiting is pursued, there are interventions at home and at school that can help. These include speaking near the child, facing the child when speaking, and providing accommodations in school so the child sits closer to the teacher. Consider re-examination and repeat hearing tests every 3 to 6 months until the effusion has resolved or the child develops symptoms indicating surgical referral.

When hearing loss is ≥40 dB, the AAP/AAFP/AAOHNS guideline recommends that you make a referral for surgical evaluation (ALGORITHM).¹⁷

Other indications for referral to a surgeon for evaluation of tympanostomy tube placement include situations in which there is:
• structural damage to the tympanic membrane or middle ear (prompt referral is recommended)
• OME of ≥4 months’ duration with persistent hearing loss (≥40 dB) or other signs or symptoms related to the effusion
• bilateral OME for ≥3 months, unilateral OME ≥6 months, or total duration of any degree of OME ≥12 months.¹⁷

Any decision regarding surgery should involve an otolaryngologist, the primary care provider, and the patient and/or family. The AAP/AAFP/AAOHNS guideline recommends against adenoidectomy in children with persistent OME without an indication for the procedure other than OME (eg, chronic sinusitis or nasal obstruction).¹⁷

Keep in mind that evidence of lasting benefit (>12 months) is limited for surgery in most patients, and the surgical and anesthetic risks must be considered before moving forward.¹⁷ (For more on the evidence regarding surgery, see “Cochrane weighs in on tympanostomy tubes”.²⁵) Tonsillectomy also does not appear to affect outcomes and is not advised.¹⁷

When a referral is always needed. Regardless of hearing status, promptly refer children with recurrent or persistent OME who are at risk of speech, language, or learning problems (including those with autism spectrum disorder, developmental delay, Down’s syndrome, diagnosed speech or language delay, or craniofacial disorders such as cleft palate) to a specialist.¹⁷

CASE › You tell your young patient’s mother that watchful waiting is appropriate at this point, since his acute otitis media was only 2 weeks ago, and his OME likely started after the acute infection. Given that his speech is clear and he is otherwise meeting his milestones, you tell her that he does not need a referral at this time, but that she should bring him back in 4 weeks for reassessment. At the next visit, his effusion has resolved, and his mother reports he is sleeping well through the night again.

CASE › A mother brings in her 3-year-old son for a regular check-up. Her only concern is that for the past 2 weeks, he has not been sleeping through the night. She indicates that the sleeping problem began after he was diagnosed with and treated for an ear infection. Fortunately, this hasn’t affected his daily activity or energy, she says.

The child’s appetite is good and he speaks clearly, in 5-word sentences. He is meeting his developmental milestones, and appears well—sitting in his mother’s lap and playing with her smartphone. His head, eyes, ears, nose, and throat exam only turns up fluid behind his left tympanic membrane, which is not red or bulging. The right membrane appears normal, and he has no cervical lymphadenopathy. The rest of his exam is normal. How would you manage a patient like this?

"Glue ear" is often asymptomatic

Otitis media with effusion (OME) is defined as middle-ear effusion (MEE) in the absence of acute signs of infection. In children, OME—also referred to as “glue ear”—most often arises after acute otitis media (AOM). In adults, it often occurs in association with eustachian tube dysfunction, although OME is a separate diagnosis. (To learn more, see “What about OME in adults?”^1,2.)

Experts have found it difficult to determine the exact incidence of OME because it is often asymptomatic. In addition, many cases quickly resolve on their own, making it challenging to diagnose. A 2-year prospective study of 2- to 6-year-old preschoolers revealed that MEE, diagnosed via monthly otoscopy and tympanometry, occurred at least once in 53% of the children in the first year and in 61% of the children in the second year.³ A second study followed 7-year-olds monthly for one year and found a 31% incidence of MEE using tympanometry.⁴ In the 25% of children found to have persistent MEE, the researchers noted spontaneous recovery after an average of 2 months.

We believe that nearly all children have experienced one episode of OME by the age of 3 years, but the prevalence of OME varies with age and the time of year. It is more prevalent in the winter than the summer months.⁵ OME is more common in Caucasian children than in African American or Asian children.⁶

Etiology remains elusive

Risk factors for children include a family history of OME, bottle-feeding, day care attendance, exposure to tobacco smoke, and a personal history of allergies.^7,8 One study conducted on mice suggested that inherited structural abnormalities of the middle ear and eustachian tube may play a role as well.⁹ Some have suggested that effusions of OME in children result from chronic inflammation, for example, after AOM, and that the effusions are sterile; however, recent studies have demonstrated that a biofilm is formed by bacterial otopathogens in the effusion.^10-12 The common pathogens found include nontypeable Haemophilus influenza, Streptococcus pneumoniae, and Moraxella catarrhalis. Inflammatory exudate or neutrophil infiltration is rare in the fluid, however.

The contribution of allergies to OME in children remains somewhat controversial. A retrospective review from the United Kingdom of 209 children with OME found a history of allergic rhinitis, asthma, and eczema in 89%, 36%, and 24%, respectively.¹³ However, this study was done at an allergy clinic, and it is possible that the data from the clinic’s specialized patient population are not generalizable. Gastroesophageal reflux may also be associated with OME in children. However, studies measuring the concentration of pepsin and pepsinogen in middle-ear fluid have provided conflicting results.^14,15

Look for these signs and symptoms

OME is often asymptomatic. If a patient has clinical signs of an acute illness, including fever and an erythematous tympanic membrane, it’s important to evaluate for another cause. OME can present with hearing loss or a sense of fullness in the ear. While an infant cannot express the hearing loss, the parent may detect it when observing and interacting mwith the child. Parents are also likely to report that the child is experiencing sleep disturbances.¹⁶

Vertigo may occur with OME, although not often. It may manifest itself if the child stumbles or falls. An older child or adult with vertigo may say that it feels like the room is spinning.

Diagnosis relies on pneumatic otoscopy

On physical exam, the patient will likely appear well. Otoscopic examination reveals fluid behind a normal or retracted tympanic membrane; the fluid is often clear or yellowish in color.

A subcommittee comprised of members of the American Academy of Pediatrics, American Academy of Family Physicians, and the American Academy of Otolaryngology-Head and Neck Surgery (AAP/AAFP/AAOHNS) published a clinical practice guideline in 2004 that delineates the current diagnosis and management of children between 2 months and 12 years of age with OME.¹⁷

Pneumatic otoscopy, which can reveal decreased or absent movement of the tympanic membrane (the result of fluid behind the membrane), is the primary diagnostic method recommended by the guideline. Tympanometry and acoustic reflectometry may also be used to make the diagnosis, especially when the presence of MEE is difficult to determine using pneumatic otoscopy.

CASE › Upon further discussion with the patient’s mother, you learn that the boy goes to day care 3 days a week and stays with his grandmother 2 days a week. His grandmother smokes outside of the home when he is staying with her.

After reviewing these risk factors with the mother (including the importance of smoking cessation for the grandmother), the mother asks if he needs antibiotics or a referral to a specialist for treating his OME.

How best to approach treatment

There are several management options to choose from, including watchful waiting, medication, and/or surgery. (Another option, autoinflation, which has shown some short-term benefits, is described in “Should you recommend autoinflation?”^17-19.)

The goals of management are to resolve the effusion, restore normal hearing (if diminished secondary to the effusion), and prevent future episodes or sequelae. The most significant complication of OME is permanent conductive hearing loss, but tinnitus, cholesteatoma, or tympanosclerosis may also occur.

In most patients, OME resolves without medical intervention. If additional action is required, however, the following options may be explored.

Medication. While the AAP/AAFP/AAOHNS guideline recommends against routine antibiotics for OME,¹⁷ it does note that a short course may provide short-term benefit to some patients (eg, those for whom a specialist referral or surgery is being considered).

A separate meta-analysis found that antibiotics improve clearance of the effusion within the first month after treatment (rate difference [RD]=0.16; 95% confidence interval [CI], 0.03-0.29 in 12 studies analyzed), but effusion relapses were common, and no significant benefit was noted past the first month (RD=0.06; 95% CI, -0.03 to 0.14 in 8 studies).²⁰

If you do use antibiotics, a 10- to 14-day course is preferred.¹⁷ Amoxicillin, amoxicillin-clavulanate and ceftibuten have been evaluated in separate clinical trials, but none has been clearly shown to have significant advantage over any other.^21,22

Antihistamines, decongestants, and oral and intranasal corticosteroids have little effect on OME in children and are not recommended.¹⁷ A Cochrane review including 16 studies found that children receiving antihistamines and decongestants are unlikely to see their symptoms improve significantly, and many patients experience adverse effects from the medications²³ (number needed to harm=9).

A randomized, double-blind trial involving 144 children <9 years of age with OME for at least 2 months evaluated 4 regimens involving amoxicillin alone or in combination with prednisolone. Children in the amoxicillin+prednisolone arms were significantly more likely to clear their effusions at 2 weeks (number needed to treat=6; P=.03), but not at 4 weeks (P=.12). At 4-month follow-up, effusions had recurred in 68.4% and 69.2% of those receiving amoxicillin+prednisolone and those receiving amoxicillin alone, respectively (P= .94).²⁴

Surgery—or not? The AAP/AAFP/AAOHNS guideline recommends physicians perform hearing testing when OME is present for 3 months or longer, or at any time if language delay, learning problems, or a significant hearing loss is suspected in a child with OME. The results of the hearing test can help determine how to proceed, based on the hearing level noted for the better hearing ear.

You can manage children with hearing loss ≤20 dB and without speech, language, or developmental problems with watchful waiting. Children with hearing loss of 21 to 39 dB can be managed with watchful waiting or referred for surgery. If watchful waiting is pursued, there are interventions at home and at school that can help. These include speaking near the child, facing the child when speaking, and providing accommodations in school so the child sits closer to the teacher. Consider re-examination and repeat hearing tests every 3 to 6 months until the effusion has resolved or the child develops symptoms indicating surgical referral.

When hearing loss is ≥40 dB, the AAP/AAFP/AAOHNS guideline recommends that you make a referral for surgical evaluation (ALGORITHM).¹⁷

Other indications for referral to a surgeon for evaluation of tympanostomy tube placement include situations in which there is:
• structural damage to the tympanic membrane or middle ear (prompt referral is recommended)
• OME of ≥4 months’ duration with persistent hearing loss (≥40 dB) or other signs or symptoms related to the effusion
• bilateral OME for ≥3 months, unilateral OME ≥6 months, or total duration of any degree of OME ≥12 months.¹⁷

Any decision regarding surgery should involve an otolaryngologist, the primary care provider, and the patient and/or family. The AAP/AAFP/AAOHNS guideline recommends against adenoidectomy in children with persistent OME without an indication for the procedure other than OME (eg, chronic sinusitis or nasal obstruction).¹⁷

Keep in mind that evidence of lasting benefit (>12 months) is limited for surgery in most patients, and the surgical and anesthetic risks must be considered before moving forward.¹⁷ (For more on the evidence regarding surgery, see “Cochrane weighs in on tympanostomy tubes”.²⁵) Tonsillectomy also does not appear to affect outcomes and is not advised.¹⁷

When a referral is always needed. Regardless of hearing status, promptly refer children with recurrent or persistent OME who are at risk of speech, language, or learning problems (including those with autism spectrum disorder, developmental delay, Down’s syndrome, diagnosed speech or language delay, or craniofacial disorders such as cleft palate) to a specialist.¹⁷

CASE › You tell your young patient’s mother that watchful waiting is appropriate at this point, since his acute otitis media was only 2 weeks ago, and his OME likely started after the acute infection. Given that his speech is clear and he is otherwise meeting his milestones, you tell her that he does not need a referral at this time, but that she should bring him back in 4 weeks for reassessment. At the next visit, his effusion has resolved, and his mother reports he is sleeping well through the night again.