Charlotte Lasenna, MD

Medication-induced telogen effluvium (TE) is a nonscarring alopecia that typically is reversible. Appropriate management requires identification of the underlying trigger and cessation of potential culprit medications. In part 2 of this series, we review anticoagulant and antihypertensive medications as potential contributors to TE.

Anticoagulants

Anticoagulants target various parts of the coagulation cascade to prevent clot formation in patients with conditions that increase their risk for thromboembolic events. Common indications for initiating anticoagulant therapy include atrial fibrillation,¹ venous thromboembolism,² acute myocardial infarction,³ malignancy,⁴ and hypercoagulable states.⁵ Traditional anticoagulants include heparin and warfarin. Heparin is a glycosaminoglycan that exerts its anticoagulant effects through binding with antithrombin, greatly increasing its inactivation of thrombin and factor Xa of the coagulation cascade.⁶ Warfarin is a coumarin derivative that inhibits activation of vitamin K, subsequently limiting the function of vitamin K–dependent factors II, VII, IX, and X.^7,8 Watras et al⁹ noted that heparin and warfarin were implicated in alopecia as their clinical use became widespread throughout the mid-20th century. Onset of alopecia following the use of heparin or warfarin was reported at 3 weeks to 3 months following medication initiation, with most cases clinically consistent with TE.⁹ Heparin and warfarin both have alopecia reported as a potential adverse effect in their structured product labeling documents.^10,11

Heparin is further classified into unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH); the latter is a heterogeneous group of medications derived from chemical or enzymatic depolymerization of UFH.¹² In contrast to UFH, LMWH exerts its anticoagulant effects through inactivation of factor Xa without the ability to bind thrombin.¹² An animal study using anagen-induced mice demonstrated that intraperitoneal administration of heparin inhibited the development of anagen follicles, while in vitro studies showed that the addition of heparin inhibited mouse dermal papilla cell proliferation.¹³ Other animal and in vitro studies have examined the inhibitory effects of heparin on signaling pathways in tumor lymphangiogenesis, including the vascular endothelial growth factor C/vascular endothelial growth factor receptor 3 axis.^14,15 Clinically, it has been demonstrated that heparin, especially LMWHs, may be associated with a survival benefit among certain cancer patients,^16,17 with the impact of LMWHs attributed to antimitotic and antimetastatic effects of heparin on tumor growth.¹⁴ It is hypothesized that such antiangiogenic and antimitotic effects also are involved in the pathomechanisms of heparin-induced alopecia.¹⁸

More recently, the use of direct oral anticoagulants (DOACs) such as dabigatran, rivaroxaban, and apixaban has increased due to their more favorable adverse-effect profile and minimal monitoring requirements. Bonaldo et al¹⁹ conducted an analysis of reports submitted to the World Health Organization’s VigiBase database of alopecia associated with DOACs until May 2, 2018. They found 1316 nonduplicate DOAC-induced cases of alopecia, with rivaroxaban as the most reported drug associated with alopecia development (58.8% [774/1316]). Only 4 cases demonstrated alopecia with DOAC rechallenge, suggesting onset of alopecia may have been unrelated to DOAC use or caused by a different trigger. Among 243 cases with a documented time to onset of alopecia, the median was 28 days, with an interquartile range of 63 days. Because TE most commonly occurs 3 to 4 months after the inciting event or medication trigger, there is little evidence to suggest DOACs as the cause of TE, and the observed cases of alopecia may be attributable to another preceding medical event and/or medication exposure.¹⁹ More studies are needed to examine the impact of anticoagulant medications on the hair cycle.

Antihypertensives

Hypertension is a modifiable risk factor for several ­cardiovascular diseases.²⁰ According to the 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease,²¹ first-line medications include thiazide diuretics, calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin receptor ­blockers (ARBs).

Angiotensin-converting enzyme inhibitors exert their antihypertensive effects by reducing conversion of angiotensin I to angiotensin II, thereby limiting the downstream effects of vasoconstriction as well as sodium and water retention. Given the proven mortality benefit of ACE inhibition in patients with congestive heart failure, ACE inhibitors are used as first-line therapy in these patients.^22,23 Alopecia associated with ACE inhibitors is rare and limited to case reports following their introduction and approval in 1981.^24-28 In one case, a woman in her 60s with congestive heart failure initiated captopril with development of an erythematous pruritic rash on the extremities and diffuse scalp hair loss 2 months later; spontaneous hair growth resumed 1 month following captopril discontinuation.²⁵ In this case, the hair loss may be secondary to the drug eruption rather than true medication-induced TE. Initiation of enalapril in a woman in her 30s with hypertension was associated with diffuse scalp alopecia 4 weeks later that resolved with cessation of the suspected culprit, enalapril; rechallenge with enalapril several months later reproduced the hair loss.²⁷ Given limited reports of ACE inhibitor–associated hair loss relative to their pervasive use, a direct causal role between ACE inhibition and TE is unlikely, or it has not been rigorously identified. The structured product labeling for captopril includes alopecia in its list of adverse effects reported in approximately 0.5% to 2% of patients but did not appear at increased frequency compared to placebo or other treatments used in controlled trials.²⁹ Alternative inciting causes of alopecia in patients prescribed ACE inhibitors may include use of other medications, hospitalization, or metabolic derangements related to their underlying cardiac disease.

Although not indicated as a primary treatment for hypertension, β-blockers have US Food and Drug Administration approval for the treatment of certain arrhythmias, hypertension, heart failure, myocardial infarction, hyperthyroidism, and other conditions.³⁰β-Blockers are competitive antagonists of β-adrenergic receptors that limit the production of intracellular cyclic adenosine monophosphate, but the mechanism of β-blockers as antihypertensives is unclear.³¹ Evidence supporting the role of β-adrenergic antagonists in TE is limited to case reports. Widespread alopecia across the scalp and arms was noted in a man in his 30s several months after starting propranolol.³² Biopsy of an affected area of the scalp demonstrated an increased number of telogen follicles with no other abnormalities. Near-complete resolution of alopecia was seen 4 months following cessation of propranolol, which recurred within 4 weeks of rechallenge.³² Although the histopathologic features are compatible with TE, the loss of body hair and rapid recurrence within 4 weeks of rechallenge are atypical for TE. As such, the use of propranolol and the reported alopecia may be coincidental or evidence of an atypical drug reaction distinct from medication-induced TE. Only a handful of other case reports have been published describing TE in patients treated with β-blockers, including metoprolol and propranolol.^33,34 Alopecia has been reported with the use of carvedilol in up to 0.1% of participants.³⁵ Although cases have been reported, TE appears to be an uncommon occurrence following β-blocker therapy.

Minoxidil—Oral minoxidil originally was approved for use in patients with resistant hypertension, defined as blood pressure elevated above goal despite concurrent use of the maximum dose of 3 classes of antihypertensives.³⁶ Unlike other antihypertensive medications, minoxidil appears to cause reversible hypertrichosis that affects nearly all patients using oral minoxidil for longer than 1 month.³⁷ This common adverse effect was a desired outcome in patients affected by hair loss, and a topical formulation of minoxidil was approved for androgenetic alopecia in men and women in 1988 and 1991, respectively.³⁸ Since its approval, topical minoxidil has been commonly prescribed in the treatment of several types of alopecia, though evidence of its efficacy in the treatment of TE is limited.^39,40 Low-dose oral minoxidil also has been reported to aid hair growth in androgenetic alopecia and TE.⁴¹ Taken orally, minoxidil is converted by sulfotransferases in the liver to minoxidil sulfate, which causes opening of plasma membrane adenosine ­triphosphate–sensitive potassium channels.^42-44 The subsequent membrane hyperpolarization reduces calcium ion influx, which also reduces cell excitability, and inhibits contraction in vascular smooth muscle cells, which results in the arteriolar vasodilatory and antihypertensive effects of minoxidil.^43,45 The potassium channel–opening effects of minoxidil may underly its hair growth stimulatory action. Unrelated potassium channel openers such as diazoxide and pinacidil also cause hypertrichosis.^46-48 An animal study showed that topical minoxidil, cromakalim (potassium channel opener), and P1075 (pinacidil analog) applied daily to the scalps of balding stump-tailed macaques led to significant increases in hair weight over a 20-week treatment period compared with the vehicle control group (P<.05 for minoxidil 100 mM and 250 mM, cromakalim 100 mM, and P1075 100 mM and 250 mM).⁵⁰ For minoxidil, this effect on hair growth appears to be dose dependent, as cumulative hair weights for the study period were significantly greater in the 250-mM concentration compared with 100-mM minoxidil (P<.05).⁴⁹ The potassium channel–opening activity of minoxidil may induce stimulation of microcirculation around hair follicles conducive to hair growth.⁵⁰ Other proposed mechanisms for hair growth with minoxidil include effects on keratinocyte and fibroblast cell proliferation,^51-53 collagen synthesis,^52,54 and prostaglandin activity.^44,55

Final Thoughts

Medication-induced TE is an undesired adverse effect of many commonly used medications, including retinoids, azole antifungals, mood stabilizers, anticoagulants, and antihypertensives. In part 1⁵⁶ of this 2-part series, we reviewed the existing literature on hair loss from retinoids, antifungals, and psychotropic medications. Herein, we focused on anticoagulant and antihypertensive medications as potential culprits of TE. Heparin and its derivatives have been associated with development of diffuse alopecia weeks to months after the start of treatment. Alopecia associated with ACE inhibitors and β-blockers has been described only in case reports, suggesting that they may be unlikely causes of TE. In contrast, minoxidil is an antihypertensive that can result in hypertrichosis and is used in the treatment of androgenetic alopecia. It should not be assumed that medications that share an indication or are part of the same medication class would similarly induce TE. The development of diffuse nonscarring alopecia should prompt suspicion for TE and thorough investigation of medications initiated 1 to 6 months prior to onset of clinically apparent alopecia. Suspected culprit medications should be carefully assessed for their likelihood of inducing TE.

Medication-induced telogen effluvium (TE) is a nonscarring alopecia that typically is reversible. Appropriate management requires identification of the underlying trigger and cessation of potential culprit medications. In part 2 of this series, we review anticoagulant and antihypertensive medications as potential contributors to TE.

Anticoagulants

Anticoagulants target various parts of the coagulation cascade to prevent clot formation in patients with conditions that increase their risk for thromboembolic events. Common indications for initiating anticoagulant therapy include atrial fibrillation,¹ venous thromboembolism,² acute myocardial infarction,³ malignancy,⁴ and hypercoagulable states.⁵ Traditional anticoagulants include heparin and warfarin. Heparin is a glycosaminoglycan that exerts its anticoagulant effects through binding with antithrombin, greatly increasing its inactivation of thrombin and factor Xa of the coagulation cascade.⁶ Warfarin is a coumarin derivative that inhibits activation of vitamin K, subsequently limiting the function of vitamin K–dependent factors II, VII, IX, and X.^7,8 Watras et al⁹ noted that heparin and warfarin were implicated in alopecia as their clinical use became widespread throughout the mid-20th century. Onset of alopecia following the use of heparin or warfarin was reported at 3 weeks to 3 months following medication initiation, with most cases clinically consistent with TE.⁹ Heparin and warfarin both have alopecia reported as a potential adverse effect in their structured product labeling documents.^10,11

Heparin is further classified into unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH); the latter is a heterogeneous group of medications derived from chemical or enzymatic depolymerization of UFH.¹² In contrast to UFH, LMWH exerts its anticoagulant effects through inactivation of factor Xa without the ability to bind thrombin.¹² An animal study using anagen-induced mice demonstrated that intraperitoneal administration of heparin inhibited the development of anagen follicles, while in vitro studies showed that the addition of heparin inhibited mouse dermal papilla cell proliferation.¹³ Other animal and in vitro studies have examined the inhibitory effects of heparin on signaling pathways in tumor lymphangiogenesis, including the vascular endothelial growth factor C/vascular endothelial growth factor receptor 3 axis.^14,15 Clinically, it has been demonstrated that heparin, especially LMWHs, may be associated with a survival benefit among certain cancer patients,^16,17 with the impact of LMWHs attributed to antimitotic and antimetastatic effects of heparin on tumor growth.¹⁴ It is hypothesized that such antiangiogenic and antimitotic effects also are involved in the pathomechanisms of heparin-induced alopecia.¹⁸

More recently, the use of direct oral anticoagulants (DOACs) such as dabigatran, rivaroxaban, and apixaban has increased due to their more favorable adverse-effect profile and minimal monitoring requirements. Bonaldo et al¹⁹ conducted an analysis of reports submitted to the World Health Organization’s VigiBase database of alopecia associated with DOACs until May 2, 2018. They found 1316 nonduplicate DOAC-induced cases of alopecia, with rivaroxaban as the most reported drug associated with alopecia development (58.8% [774/1316]). Only 4 cases demonstrated alopecia with DOAC rechallenge, suggesting onset of alopecia may have been unrelated to DOAC use or caused by a different trigger. Among 243 cases with a documented time to onset of alopecia, the median was 28 days, with an interquartile range of 63 days. Because TE most commonly occurs 3 to 4 months after the inciting event or medication trigger, there is little evidence to suggest DOACs as the cause of TE, and the observed cases of alopecia may be attributable to another preceding medical event and/or medication exposure.¹⁹ More studies are needed to examine the impact of anticoagulant medications on the hair cycle.

Antihypertensives

Hypertension is a modifiable risk factor for several ­cardiovascular diseases.²⁰ According to the 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease,²¹ first-line medications include thiazide diuretics, calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin receptor ­blockers (ARBs).

Angiotensin-converting enzyme inhibitors exert their antihypertensive effects by reducing conversion of angiotensin I to angiotensin II, thereby limiting the downstream effects of vasoconstriction as well as sodium and water retention. Given the proven mortality benefit of ACE inhibition in patients with congestive heart failure, ACE inhibitors are used as first-line therapy in these patients.^22,23 Alopecia associated with ACE inhibitors is rare and limited to case reports following their introduction and approval in 1981.^24-28 In one case, a woman in her 60s with congestive heart failure initiated captopril with development of an erythematous pruritic rash on the extremities and diffuse scalp hair loss 2 months later; spontaneous hair growth resumed 1 month following captopril discontinuation.²⁵ In this case, the hair loss may be secondary to the drug eruption rather than true medication-induced TE. Initiation of enalapril in a woman in her 30s with hypertension was associated with diffuse scalp alopecia 4 weeks later that resolved with cessation of the suspected culprit, enalapril; rechallenge with enalapril several months later reproduced the hair loss.²⁷ Given limited reports of ACE inhibitor–associated hair loss relative to their pervasive use, a direct causal role between ACE inhibition and TE is unlikely, or it has not been rigorously identified. The structured product labeling for captopril includes alopecia in its list of adverse effects reported in approximately 0.5% to 2% of patients but did not appear at increased frequency compared to placebo or other treatments used in controlled trials.²⁹ Alternative inciting causes of alopecia in patients prescribed ACE inhibitors may include use of other medications, hospitalization, or metabolic derangements related to their underlying cardiac disease.

Although not indicated as a primary treatment for hypertension, β-blockers have US Food and Drug Administration approval for the treatment of certain arrhythmias, hypertension, heart failure, myocardial infarction, hyperthyroidism, and other conditions.³⁰β-Blockers are competitive antagonists of β-adrenergic receptors that limit the production of intracellular cyclic adenosine monophosphate, but the mechanism of β-blockers as antihypertensives is unclear.³¹ Evidence supporting the role of β-adrenergic antagonists in TE is limited to case reports. Widespread alopecia across the scalp and arms was noted in a man in his 30s several months after starting propranolol.³² Biopsy of an affected area of the scalp demonstrated an increased number of telogen follicles with no other abnormalities. Near-complete resolution of alopecia was seen 4 months following cessation of propranolol, which recurred within 4 weeks of rechallenge.³² Although the histopathologic features are compatible with TE, the loss of body hair and rapid recurrence within 4 weeks of rechallenge are atypical for TE. As such, the use of propranolol and the reported alopecia may be coincidental or evidence of an atypical drug reaction distinct from medication-induced TE. Only a handful of other case reports have been published describing TE in patients treated with β-blockers, including metoprolol and propranolol.^33,34 Alopecia has been reported with the use of carvedilol in up to 0.1% of participants.³⁵ Although cases have been reported, TE appears to be an uncommon occurrence following β-blocker therapy.

Minoxidil—Oral minoxidil originally was approved for use in patients with resistant hypertension, defined as blood pressure elevated above goal despite concurrent use of the maximum dose of 3 classes of antihypertensives.³⁶ Unlike other antihypertensive medications, minoxidil appears to cause reversible hypertrichosis that affects nearly all patients using oral minoxidil for longer than 1 month.³⁷ This common adverse effect was a desired outcome in patients affected by hair loss, and a topical formulation of minoxidil was approved for androgenetic alopecia in men and women in 1988 and 1991, respectively.³⁸ Since its approval, topical minoxidil has been commonly prescribed in the treatment of several types of alopecia, though evidence of its efficacy in the treatment of TE is limited.^39,40 Low-dose oral minoxidil also has been reported to aid hair growth in androgenetic alopecia and TE.⁴¹ Taken orally, minoxidil is converted by sulfotransferases in the liver to minoxidil sulfate, which causes opening of plasma membrane adenosine ­triphosphate–sensitive potassium channels.^42-44 The subsequent membrane hyperpolarization reduces calcium ion influx, which also reduces cell excitability, and inhibits contraction in vascular smooth muscle cells, which results in the arteriolar vasodilatory and antihypertensive effects of minoxidil.^43,45 The potassium channel–opening effects of minoxidil may underly its hair growth stimulatory action. Unrelated potassium channel openers such as diazoxide and pinacidil also cause hypertrichosis.^46-48 An animal study showed that topical minoxidil, cromakalim (potassium channel opener), and P1075 (pinacidil analog) applied daily to the scalps of balding stump-tailed macaques led to significant increases in hair weight over a 20-week treatment period compared with the vehicle control group (P<.05 for minoxidil 100 mM and 250 mM, cromakalim 100 mM, and P1075 100 mM and 250 mM).⁵⁰ For minoxidil, this effect on hair growth appears to be dose dependent, as cumulative hair weights for the study period were significantly greater in the 250-mM concentration compared with 100-mM minoxidil (P<.05).⁴⁹ The potassium channel–opening activity of minoxidil may induce stimulation of microcirculation around hair follicles conducive to hair growth.⁵⁰ Other proposed mechanisms for hair growth with minoxidil include effects on keratinocyte and fibroblast cell proliferation,^51-53 collagen synthesis,^52,54 and prostaglandin activity.^44,55

Final Thoughts

Medication-induced TE is an undesired adverse effect of many commonly used medications, including retinoids, azole antifungals, mood stabilizers, anticoagulants, and antihypertensives. In part 1⁵⁶ of this 2-part series, we reviewed the existing literature on hair loss from retinoids, antifungals, and psychotropic medications. Herein, we focused on anticoagulant and antihypertensive medications as potential culprits of TE. Heparin and its derivatives have been associated with development of diffuse alopecia weeks to months after the start of treatment. Alopecia associated with ACE inhibitors and β-blockers has been described only in case reports, suggesting that they may be unlikely causes of TE. In contrast, minoxidil is an antihypertensive that can result in hypertrichosis and is used in the treatment of androgenetic alopecia. It should not be assumed that medications that share an indication or are part of the same medication class would similarly induce TE. The development of diffuse nonscarring alopecia should prompt suspicion for TE and thorough investigation of medications initiated 1 to 6 months prior to onset of clinically apparent alopecia. Suspected culprit medications should be carefully assessed for their likelihood of inducing TE.

Medication-induced telogen effluvium (TE) is a nonscarring alopecia that typically is reversible. Appropriate management requires identification of the underlying trigger and cessation of potential culprit medications. In part 2 of this series, we review anticoagulant and antihypertensive medications as potential contributors to TE.

Anticoagulants

Anticoagulants target various parts of the coagulation cascade to prevent clot formation in patients with conditions that increase their risk for thromboembolic events. Common indications for initiating anticoagulant therapy include atrial fibrillation,¹ venous thromboembolism,² acute myocardial infarction,³ malignancy,⁴ and hypercoagulable states.⁵ Traditional anticoagulants include heparin and warfarin. Heparin is a glycosaminoglycan that exerts its anticoagulant effects through binding with antithrombin, greatly increasing its inactivation of thrombin and factor Xa of the coagulation cascade.⁶ Warfarin is a coumarin derivative that inhibits activation of vitamin K, subsequently limiting the function of vitamin K–dependent factors II, VII, IX, and X.^7,8 Watras et al⁹ noted that heparin and warfarin were implicated in alopecia as their clinical use became widespread throughout the mid-20th century. Onset of alopecia following the use of heparin or warfarin was reported at 3 weeks to 3 months following medication initiation, with most cases clinically consistent with TE.⁹ Heparin and warfarin both have alopecia reported as a potential adverse effect in their structured product labeling documents.^10,11

Heparin is further classified into unfractionated heparin (UFH) and low-molecular-weight heparin (LMWH); the latter is a heterogeneous group of medications derived from chemical or enzymatic depolymerization of UFH.¹² In contrast to UFH, LMWH exerts its anticoagulant effects through inactivation of factor Xa without the ability to bind thrombin.¹² An animal study using anagen-induced mice demonstrated that intraperitoneal administration of heparin inhibited the development of anagen follicles, while in vitro studies showed that the addition of heparin inhibited mouse dermal papilla cell proliferation.¹³ Other animal and in vitro studies have examined the inhibitory effects of heparin on signaling pathways in tumor lymphangiogenesis, including the vascular endothelial growth factor C/vascular endothelial growth factor receptor 3 axis.^14,15 Clinically, it has been demonstrated that heparin, especially LMWHs, may be associated with a survival benefit among certain cancer patients,^16,17 with the impact of LMWHs attributed to antimitotic and antimetastatic effects of heparin on tumor growth.¹⁴ It is hypothesized that such antiangiogenic and antimitotic effects also are involved in the pathomechanisms of heparin-induced alopecia.¹⁸

More recently, the use of direct oral anticoagulants (DOACs) such as dabigatran, rivaroxaban, and apixaban has increased due to their more favorable adverse-effect profile and minimal monitoring requirements. Bonaldo et al¹⁹ conducted an analysis of reports submitted to the World Health Organization’s VigiBase database of alopecia associated with DOACs until May 2, 2018. They found 1316 nonduplicate DOAC-induced cases of alopecia, with rivaroxaban as the most reported drug associated with alopecia development (58.8% [774/1316]). Only 4 cases demonstrated alopecia with DOAC rechallenge, suggesting onset of alopecia may have been unrelated to DOAC use or caused by a different trigger. Among 243 cases with a documented time to onset of alopecia, the median was 28 days, with an interquartile range of 63 days. Because TE most commonly occurs 3 to 4 months after the inciting event or medication trigger, there is little evidence to suggest DOACs as the cause of TE, and the observed cases of alopecia may be attributable to another preceding medical event and/or medication exposure.¹⁹ More studies are needed to examine the impact of anticoagulant medications on the hair cycle.

Antihypertensives

Hypertension is a modifiable risk factor for several ­cardiovascular diseases.²⁰ According to the 2019 American College of Cardiology/American Heart Association Guideline on the Primary Prevention of Cardiovascular Disease,²¹ first-line medications include thiazide diuretics, calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin receptor ­blockers (ARBs).

Angiotensin-converting enzyme inhibitors exert their antihypertensive effects by reducing conversion of angiotensin I to angiotensin II, thereby limiting the downstream effects of vasoconstriction as well as sodium and water retention. Given the proven mortality benefit of ACE inhibition in patients with congestive heart failure, ACE inhibitors are used as first-line therapy in these patients.^22,23 Alopecia associated with ACE inhibitors is rare and limited to case reports following their introduction and approval in 1981.^24-28 In one case, a woman in her 60s with congestive heart failure initiated captopril with development of an erythematous pruritic rash on the extremities and diffuse scalp hair loss 2 months later; spontaneous hair growth resumed 1 month following captopril discontinuation.²⁵ In this case, the hair loss may be secondary to the drug eruption rather than true medication-induced TE. Initiation of enalapril in a woman in her 30s with hypertension was associated with diffuse scalp alopecia 4 weeks later that resolved with cessation of the suspected culprit, enalapril; rechallenge with enalapril several months later reproduced the hair loss.²⁷ Given limited reports of ACE inhibitor–associated hair loss relative to their pervasive use, a direct causal role between ACE inhibition and TE is unlikely, or it has not been rigorously identified. The structured product labeling for captopril includes alopecia in its list of adverse effects reported in approximately 0.5% to 2% of patients but did not appear at increased frequency compared to placebo or other treatments used in controlled trials.²⁹ Alternative inciting causes of alopecia in patients prescribed ACE inhibitors may include use of other medications, hospitalization, or metabolic derangements related to their underlying cardiac disease.

Although not indicated as a primary treatment for hypertension, β-blockers have US Food and Drug Administration approval for the treatment of certain arrhythmias, hypertension, heart failure, myocardial infarction, hyperthyroidism, and other conditions.³⁰β-Blockers are competitive antagonists of β-adrenergic receptors that limit the production of intracellular cyclic adenosine monophosphate, but the mechanism of β-blockers as antihypertensives is unclear.³¹ Evidence supporting the role of β-adrenergic antagonists in TE is limited to case reports. Widespread alopecia across the scalp and arms was noted in a man in his 30s several months after starting propranolol.³² Biopsy of an affected area of the scalp demonstrated an increased number of telogen follicles with no other abnormalities. Near-complete resolution of alopecia was seen 4 months following cessation of propranolol, which recurred within 4 weeks of rechallenge.³² Although the histopathologic features are compatible with TE, the loss of body hair and rapid recurrence within 4 weeks of rechallenge are atypical for TE. As such, the use of propranolol and the reported alopecia may be coincidental or evidence of an atypical drug reaction distinct from medication-induced TE. Only a handful of other case reports have been published describing TE in patients treated with β-blockers, including metoprolol and propranolol.^33,34 Alopecia has been reported with the use of carvedilol in up to 0.1% of participants.³⁵ Although cases have been reported, TE appears to be an uncommon occurrence following β-blocker therapy.

Minoxidil—Oral minoxidil originally was approved for use in patients with resistant hypertension, defined as blood pressure elevated above goal despite concurrent use of the maximum dose of 3 classes of antihypertensives.³⁶ Unlike other antihypertensive medications, minoxidil appears to cause reversible hypertrichosis that affects nearly all patients using oral minoxidil for longer than 1 month.³⁷ This common adverse effect was a desired outcome in patients affected by hair loss, and a topical formulation of minoxidil was approved for androgenetic alopecia in men and women in 1988 and 1991, respectively.³⁸ Since its approval, topical minoxidil has been commonly prescribed in the treatment of several types of alopecia, though evidence of its efficacy in the treatment of TE is limited.^39,40 Low-dose oral minoxidil also has been reported to aid hair growth in androgenetic alopecia and TE.⁴¹ Taken orally, minoxidil is converted by sulfotransferases in the liver to minoxidil sulfate, which causes opening of plasma membrane adenosine ­triphosphate–sensitive potassium channels.^42-44 The subsequent membrane hyperpolarization reduces calcium ion influx, which also reduces cell excitability, and inhibits contraction in vascular smooth muscle cells, which results in the arteriolar vasodilatory and antihypertensive effects of minoxidil.^43,45 The potassium channel–opening effects of minoxidil may underly its hair growth stimulatory action. Unrelated potassium channel openers such as diazoxide and pinacidil also cause hypertrichosis.^46-48 An animal study showed that topical minoxidil, cromakalim (potassium channel opener), and P1075 (pinacidil analog) applied daily to the scalps of balding stump-tailed macaques led to significant increases in hair weight over a 20-week treatment period compared with the vehicle control group (P<.05 for minoxidil 100 mM and 250 mM, cromakalim 100 mM, and P1075 100 mM and 250 mM).⁵⁰ For minoxidil, this effect on hair growth appears to be dose dependent, as cumulative hair weights for the study period were significantly greater in the 250-mM concentration compared with 100-mM minoxidil (P<.05).⁴⁹ The potassium channel–opening activity of minoxidil may induce stimulation of microcirculation around hair follicles conducive to hair growth.⁵⁰ Other proposed mechanisms for hair growth with minoxidil include effects on keratinocyte and fibroblast cell proliferation,^51-53 collagen synthesis,^52,54 and prostaglandin activity.^44,55

Final Thoughts

Medication-induced TE is an undesired adverse effect of many commonly used medications, including retinoids, azole antifungals, mood stabilizers, anticoagulants, and antihypertensives. In part 1⁵⁶ of this 2-part series, we reviewed the existing literature on hair loss from retinoids, antifungals, and psychotropic medications. Herein, we focused on anticoagulant and antihypertensive medications as potential culprits of TE. Heparin and its derivatives have been associated with development of diffuse alopecia weeks to months after the start of treatment. Alopecia associated with ACE inhibitors and β-blockers has been described only in case reports, suggesting that they may be unlikely causes of TE. In contrast, minoxidil is an antihypertensive that can result in hypertrichosis and is used in the treatment of androgenetic alopecia. It should not be assumed that medications that share an indication or are part of the same medication class would similarly induce TE. The development of diffuse nonscarring alopecia should prompt suspicion for TE and thorough investigation of medications initiated 1 to 6 months prior to onset of clinically apparent alopecia. Suspected culprit medications should be carefully assessed for their likelihood of inducing TE.

Alopecia is a commonly reported side effect of various medications. Anagen effluvium and telogen effluvium (TE) are considered the most common mechanisms underlying medication-related hair loss. Anagen effluvium is associated with chemotherapeutic agents and radiation therapy, with anagen shedding typically occurring within 2 weeks of medication administration.^1,2 Medication-induced TE is a diffuse nonscarring alopecia that is a reversible reactive process.^3-5 Telogen effluvium is clinically apparent as a generalized shedding of scalp hair 1 to 6 months after an inciting cause.⁶ The underlying cause of TE may be multifactorial and difficult to identify given the delay between the trigger and the onset of clinically apparent hair loss. Other known triggers of TE include acute illness,^7,8 nutritional deficiencies,^4,9 and/or major surgery.¹⁰

Each hair follicle independently and sequentially progresses through anagen growth, catagen transition, and telogen resting phases. In the human scalp, the telogen phase typically lasts 3 months, at the end of which the telogen hair is extruded from the scalp. Anagen and telogen follicles typically account for an average of 90% and 10% of follicles on the human scalp, respectively.¹¹ Immediate anagen release is hypothesized to be the mechanism underlying medication-induced TE.¹² This theory suggests that an increased percentage of anagen follicles prematurely enter the telogen phase, with a notable increase in hair shedding at the conclusion of the telogen phase approximately 1 to 6 months later.¹² First-line management of medication-induced TE is identification and cessation of the causative agent, if possible. Notable regrowth of hair is expected several months after removal of the inciting medication. In part 1 of this 2-part series, we review the existing literature to identify common culprits of medication-induced TE, including retinoids, antifungals, and psychotropic medications.

Retinoids

Retinoids are vitamin A derivatives used in the treatment of a myriad of dermatologic and nondermatologic conditions.^13,14 Retinoids modulate sebum production,¹⁵ keratinocyte proliferation,¹⁶ and epithelial differentiation through signal transduction downstream of the ligand-activated nuclear retinoic acid receptors and retinoid X receptors.^13,14,17 The recommended daily dosage of retinol is 900 µg retinol activity equivalent (3000 IU) for men and 700 µg retinol activity equivalent (2333 IU) for women. Retinoids are used in the treatment of acne vulgaris,¹⁸ psoriasis,¹⁹ and ichthyosis.²⁰ The most commonly reported adverse effects of systemic retinoid therapy include cheilitis, alopecia, and xerosis.²¹ Retinoid-associated alopecia is dose and duration dependent.^19,21-24 A prospective study of acitretin therapy in plaque psoriasis reported that more than 63% (42/66) of patients on 50 mg or more of acitretin daily for 6 months or longer experienced alopecia that reversed with discontinuation.²³ A systematic review of isotretinoin use in acne showed alopecia was seen in 3.2% (18/565) of patients on less than 0.5 mg/kg/d of isotretinoin and in 5.7% (192/3375) of patients on 0.5 mg/kg/d or less of isotretinoin.²⁴ In a phase 2 clinical trial of orally administered 9-cis-retinoic acid (alitretinoin) in the treatment of Kaposi sarcoma related to AIDS, 42% (24/57) of adult male patients receiving 60, 100, or 140 mg/m² alitretinoin daily (median treatment duration, 15.1 weeks) reported alopecia as an adverse effect of treatment.²⁵ In one case report, a patient who ingested 500,000 IU of vitamin A daily for 4 months and then 100,000 IU monthly for 6 months experienced diffusely increased shedding of scalp hair along with muscle soreness, nail dystrophy, diffuse skin rash, and refractory ascites; he was found to have severe liver damage secondary to hypervitaminosis A that required liver transplantation.²⁶ Regarding the pathomechanism of retinoid-induced alopecia, animal and in vitro studies similarly have demonstrated that all-trans-retinoic acid appears to exert its inhibitory effects on hair follicle growth via the influence of the transforming growth factor β2 and SMAD2/3 pathway influence on dermal papillae cells.^14,27 Development of hair loss secondary to systemic retinoid therapy may be managed with dose reduction or cessation.

Antifungals

Azole medications have broad-spectrum fungistatic activity against a wide range of yeast and filamentous fungi. Azoles inhibit sterol 14α-demethylase activity, impairing ergosterol synthesis and thereby disrupting plasma membrane synthesis and activity of membrane-bound enzymes.²⁸ Fluconazole is a systemic oral agent in this class that was first approved by the US Food and Drug Administration (FDA) for use in the 1990s.²⁹ A retrospective study by the National Institute of Allergy and Infectious Disease Mycoses Study Group followed the clinical course of 33 patients who developed alopecia while receiving fluconazole therapy for various mycoses.³⁰ The majority (88% [29/33]) of patients received 400 mg or more of fluconazole daily. The median time to hair loss after starting fluconazole was 3 months, and the scalp was involved in all cases. In 97% (32/33) of patients, resolution of alopecia was noted following discontinuation of fluconazole or a dose reduction of 50% or more. In 85% (28/33) of patients, complete resolution of alopecia occurred within 6 months of fluconazole cessation or dose reduction.³⁰ Fluconazole-induced TE was reproducible in an animal model using Wistar rats³¹; however, further studies are required to clarify the molecular pathways of its effect on hair growth.

Voriconazole is an azole approved for the treatment of invasive aspergillosis, candidemia, and fungal infections caused by Scedosporium apiospermum and Fusarium species. A retrospective survey study of patients who received voriconazole for 1 month or longer found a considerable proportion of patients developed diffuse reversible hair loss.³² Scalp alopecia was noted in 79% (120/152) of patients who completed the survey, with a mean (SD) time to alopecia of 75 (54) days after initiation of voriconazole. Notable regrowth was reported in 69% (79/114) of patients who discontinued voriconazole for at least 3 months. A subgroup of 32 patients were changed to itraconazole or posaconazole, and hair loss stopped in 84% (27/32) with regrowth noted in 69% (22/32) of patients.³² Voriconazole and fluconazole share structural similarity not present with other triazoles.^33,34 Because voriconazole-associated alopecia was reversed in the majority of patients who switched to itraconazole or posaconazole, the authors hypothesized that structural similarity of fluconazole and voriconazole may underly the greater risk for TE that is not a class effect of azole medications.³¹

Psychotropic Medications

Various psychotropic medications have been associated with hair loss. Valproic acid (or sodium valproate) is an anticonvulsant and mood-stabilizing agent used for the treatment of seizures, bipolar disorder (BD), migraines, and neuropathic pain.^35,36 Divalproex sodium (or divalproex) is an enteric-coated formulation of sodium valproate and valproic acid with similar indications. Valproate is a notorious culprit of medication-induced hair loss, with alopecia listed among the most common adverse reactions (reported >5%) on its structure product labeling document.³⁷ A systemic review and meta-analysis by Wang et al³⁸ estimated the overall incidence of valproate-related alopecia to be 11% (95% CI, 0.08-0.13). Although this meta-analysis did not find an association between incidence of alopecia and dose or duration of valproate therapy,³⁸ a separate review suggested that valproate-induced alopecia is dose dependent and can be managed with dose reduction.³⁹ A 12-month, randomized, double-blind study of treatment of BD with divalproex (valproate derivative), lithium, or placebo (2:1:1 ratio) showed a significantly higher frequency of alopecia in the divalproex group compared with placebo (16% [30/187] vs 6% [6/94]; P=.03).⁴⁰ Valproate-related hair loss is characteristically diffuse and nonscarring, often noted 3 to 6 months following initiation of valproate.^41,42 The proposed mechanism of valproate-induced alopecia includes chelation of zinc and selenium,⁴³ and a reduction in serum biotinidase activity, thereby decreasing the availability of these essential micronutrients required for hair growth.⁴¹ Studies examining the effects of valproate administration and serum biotinidase activity in patients have yielded conflicting results.^44-46 In a study of children with seizures including 57 patients treated with valproic acid, 17 treated with carbamazepine, and 75 age- and sex-matched healthy controls, the authors found no significant differences in serum biotinidase enzyme activity across the 3 groups.⁴⁴ In contrast, a study of 75 children with seizures on valproic acid therapy stratified by dose (mean [SD])—group A: 28.7 [8.5] mg/kg/d; group B: 41.6 [4.9] mg/kg/d; group C: 64.5 [5.8] mg/kg/d—found that patients receiving higher doses (groups B and C) had significantly reduced serum biotinidase activity (1.22 [1.11] and 0.97 [0.07] mmol/min/L, respectively) compared with 50 healthy pediatric controls (5.20 [0.90] mmol/min/L; P<.001). The same study found biotin supplementation at 10 mg/d for 20 days led to resolution of alopecia in 22% (2/9) of patients with alopecia on valproic acid therapy.⁴⁵ Despite hypothesized effects of valproate on micronutrients, the role of mineral supplementation in treating valproate-associated hair loss remains unclear. There is evidence to suggest that valproic acid–associated alterations in serum biotinidase activity may be transient. In a study of 32 pediatric patients receiving valproic acid for the treatment of epilepsy, serum biotinidase activity was significantly lower after 3 months of valproic acid therapy compared with pretreatment levels (P<.05); at 6 months, the serum biotinidase activity was increased compared with 3 months (P<.05) and not significantly different from pretreatment levels (P>.05).⁴⁶ Hair regrowth has been observed following discontinuation or dose reduction of valproate therapy in some cases.^39,47

Lithium carbonate (lithium) is used in the treatment of BD. Despite its efficacy and low cost, its potential for adverse effects, narrow therapeutic index, and subsequent need for routine monitoring are factors that limit its use.⁴⁸ Some reported dermatologic adverse reactions on its structure product labeling include xerosis, thinning of hair, alopecia, xerosis cutis, psoriasis onset/exacerbation, and generalized pruritus.⁴⁹ A systematic review and meta-analysis of 385 studies identified 24 publications reporting adverse effects of lithium on hair with no significantly increased risk of alopecia overall.⁵⁰ The analysis included 2 randomized controlled trials comparing the effects of lithium and placebo on hair loss in patients with BD. Hair loss was reported in 7% (7/94) of patients taking lithium and 6% (6/94) of the placebo group in the 12-month study⁴⁰ and in 3% (1/32) of the lithium group and 0% (0/28) of the divalproex group in the 20-month study.⁵¹ Despite anecdotal reports of alopecia associated with lithium, there is a lack of high-quality evidence to support this claim. Of note, hypothyroidism is a known complication of lithium use, and serum testing of thyroid function at 6-month intervals is recommended for patients on lithium treatment.⁵² Because thyroid abnormalities can cause alopecia distinct from TE, new-onset alopecia during lithium use should prompt serum testing of thyroid function. The development of hypothyroidism secondary to lithium is not a direct contraindication to its use⁵³; rather, treatment should be focused on correction with thyroid replacement therapy (eg, supplementation with thyroxine).⁵⁴

Commonly prescribed antidepressant medications include selective serotonin reuptake inhibitors (SSRIs) and bupropion. Selective serotonin reuptake inhibitors affect the neuronal serotonin transporter, increasing the concentration of serotonin in the synaptic cleft available for stimulation of postsynaptic serotonin receptors^55,56; bupropion is an antidepressant medication that inhibits norepinephrine and dopamine reuptake at the synaptic cleft.⁵⁷ Alopecia is an infrequent (1 in 100 to 1 in 1000 patients) adverse effect for several SSRIs.^58-62 A recent systematic review identified a total of 71 cases of alopecia associated with SSRI use including citalopram (n=11), escitalopram (n=7), fluoxetine (n=27), fluoxvamine (n=5), paroxetine (n=4), and sertraline (n=20), with a median time to onset of hair shedding of 8.6 weeks (range, 3 days to 5 years). Discontinuation of the suspected culprit SSRI led to improvement and/or resolution in 63% (51/81) episodes of alopecia, with a median time to improvement and/or resolution of 4 weeks.⁶³ A comparative retrospective cohort study using a large US health claims database from 2006 to 2014 included more than 1 million new and mutually exclusive patients taking fluoxetine, fluvoxamine, sertraline, citalopram, escitalopram, paroxetine, duloxetine, venlafaxine, desvenlafaxine, and bupropion.⁶⁴ Overall, 1% (1569/150,404) of patients treated with bupropion received 1 or more physician visits for alopecia. Patients on SSRIs generally had a lower risk for hair loss compared with patients using bupropion (citalopram: hazard ratio [HR], 0.80 [95% CI, 0.74-0.86]; escitalopram: HR, 0.79 [95% CI, 0.74-0.86]; fluoxetine: HR, 0.68 [95% CI, 0.63-0.74]; paroxetine: HR, 0.68 [95% CI, 0.62-0.74]; sertraline: HR, 0.74 [95% CI, 0.69-0.79]), with the exception of fluvoxamine (HR, 0.93 [95% CI, 0.64-1.37]). However, the type of alopecia, time to onset, and time to resolution were not reported, making it difficult to assess whether the reported hair loss was consistent with medication-induced TE. Additionally, the authors acknowledged that bupropion may have been prescribed for smoking cessation, which may carry a different risk profile for the development of alopecia.⁶⁴ Several other case reports have described alopecia following treatment with SSRIs, including sertraline,⁶⁵ fluvoxamine,⁶⁶ paroxetine,⁶⁷ fluoxetine,⁶⁸ and escitalopram.⁶⁹

Overall, it appears that the use of SSRIs portends relatively low risk for alopecia and medication-induced TE. Little is known regarding the molecular effects of SSRIs on hair growth and the pathomechanism of SSRI-induced TE. The potential benefits of discontinuing a suspected culprit medication should be carefully weighed against the risks of medication cessation, and consideration should be given to alternative medications in the same class that also may be associated with TE. In patients requiring antidepressant therapy with suspected medication-induced TE, consider transitioning to a different class of medication with lower risk of medication-induced alopecia; for example, discontinuing bupropion in favor of an SSRI.

Final Thoughts

Medication-induced alopecia is an undesired side effect of many commonly used drugs and drug classes, including retinoids, azole antifungals, and mood stabilizers. Although the precise pathomechanisms of medication-induced TE remain unclear, the recommended management often requires identification of the likely causative agent and its discontinuation, if possible. Suspicion for medication-induced TE should prompt a thorough history of recent changes to medications, risk factors for nutritional deficiencies, underlying illnesses, and recent surgical procedures. Underlying nutritional, electrolyte, and/or metabolic disturbances should be corrected. In part 2 of this series, we will discuss medication-induced alopecia associated with anticoagulant and antihypertensive medications.

Alopecia is a commonly reported side effect of various medications. Anagen effluvium and telogen effluvium (TE) are considered the most common mechanisms underlying medication-related hair loss. Anagen effluvium is associated with chemotherapeutic agents and radiation therapy, with anagen shedding typically occurring within 2 weeks of medication administration.^1,2 Medication-induced TE is a diffuse nonscarring alopecia that is a reversible reactive process.^3-5 Telogen effluvium is clinically apparent as a generalized shedding of scalp hair 1 to 6 months after an inciting cause.⁶ The underlying cause of TE may be multifactorial and difficult to identify given the delay between the trigger and the onset of clinically apparent hair loss. Other known triggers of TE include acute illness,^7,8 nutritional deficiencies,^4,9 and/or major surgery.¹⁰

Each hair follicle independently and sequentially progresses through anagen growth, catagen transition, and telogen resting phases. In the human scalp, the telogen phase typically lasts 3 months, at the end of which the telogen hair is extruded from the scalp. Anagen and telogen follicles typically account for an average of 90% and 10% of follicles on the human scalp, respectively.¹¹ Immediate anagen release is hypothesized to be the mechanism underlying medication-induced TE.¹² This theory suggests that an increased percentage of anagen follicles prematurely enter the telogen phase, with a notable increase in hair shedding at the conclusion of the telogen phase approximately 1 to 6 months later.¹² First-line management of medication-induced TE is identification and cessation of the causative agent, if possible. Notable regrowth of hair is expected several months after removal of the inciting medication. In part 1 of this 2-part series, we review the existing literature to identify common culprits of medication-induced TE, including retinoids, antifungals, and psychotropic medications.

Retinoids

Retinoids are vitamin A derivatives used in the treatment of a myriad of dermatologic and nondermatologic conditions.^13,14 Retinoids modulate sebum production,¹⁵ keratinocyte proliferation,¹⁶ and epithelial differentiation through signal transduction downstream of the ligand-activated nuclear retinoic acid receptors and retinoid X receptors.^13,14,17 The recommended daily dosage of retinol is 900 µg retinol activity equivalent (3000 IU) for men and 700 µg retinol activity equivalent (2333 IU) for women. Retinoids are used in the treatment of acne vulgaris,¹⁸ psoriasis,¹⁹ and ichthyosis.²⁰ The most commonly reported adverse effects of systemic retinoid therapy include cheilitis, alopecia, and xerosis.²¹ Retinoid-associated alopecia is dose and duration dependent.^19,21-24 A prospective study of acitretin therapy in plaque psoriasis reported that more than 63% (42/66) of patients on 50 mg or more of acitretin daily for 6 months or longer experienced alopecia that reversed with discontinuation.²³ A systematic review of isotretinoin use in acne showed alopecia was seen in 3.2% (18/565) of patients on less than 0.5 mg/kg/d of isotretinoin and in 5.7% (192/3375) of patients on 0.5 mg/kg/d or less of isotretinoin.²⁴ In a phase 2 clinical trial of orally administered 9-cis-retinoic acid (alitretinoin) in the treatment of Kaposi sarcoma related to AIDS, 42% (24/57) of adult male patients receiving 60, 100, or 140 mg/m² alitretinoin daily (median treatment duration, 15.1 weeks) reported alopecia as an adverse effect of treatment.²⁵ In one case report, a patient who ingested 500,000 IU of vitamin A daily for 4 months and then 100,000 IU monthly for 6 months experienced diffusely increased shedding of scalp hair along with muscle soreness, nail dystrophy, diffuse skin rash, and refractory ascites; he was found to have severe liver damage secondary to hypervitaminosis A that required liver transplantation.²⁶ Regarding the pathomechanism of retinoid-induced alopecia, animal and in vitro studies similarly have demonstrated that all-trans-retinoic acid appears to exert its inhibitory effects on hair follicle growth via the influence of the transforming growth factor β2 and SMAD2/3 pathway influence on dermal papillae cells.^14,27 Development of hair loss secondary to systemic retinoid therapy may be managed with dose reduction or cessation.

Antifungals

Azole medications have broad-spectrum fungistatic activity against a wide range of yeast and filamentous fungi. Azoles inhibit sterol 14α-demethylase activity, impairing ergosterol synthesis and thereby disrupting plasma membrane synthesis and activity of membrane-bound enzymes.²⁸ Fluconazole is a systemic oral agent in this class that was first approved by the US Food and Drug Administration (FDA) for use in the 1990s.²⁹ A retrospective study by the National Institute of Allergy and Infectious Disease Mycoses Study Group followed the clinical course of 33 patients who developed alopecia while receiving fluconazole therapy for various mycoses.³⁰ The majority (88% [29/33]) of patients received 400 mg or more of fluconazole daily. The median time to hair loss after starting fluconazole was 3 months, and the scalp was involved in all cases. In 97% (32/33) of patients, resolution of alopecia was noted following discontinuation of fluconazole or a dose reduction of 50% or more. In 85% (28/33) of patients, complete resolution of alopecia occurred within 6 months of fluconazole cessation or dose reduction.³⁰ Fluconazole-induced TE was reproducible in an animal model using Wistar rats³¹; however, further studies are required to clarify the molecular pathways of its effect on hair growth.

Voriconazole is an azole approved for the treatment of invasive aspergillosis, candidemia, and fungal infections caused by Scedosporium apiospermum and Fusarium species. A retrospective survey study of patients who received voriconazole for 1 month or longer found a considerable proportion of patients developed diffuse reversible hair loss.³² Scalp alopecia was noted in 79% (120/152) of patients who completed the survey, with a mean (SD) time to alopecia of 75 (54) days after initiation of voriconazole. Notable regrowth was reported in 69% (79/114) of patients who discontinued voriconazole for at least 3 months. A subgroup of 32 patients were changed to itraconazole or posaconazole, and hair loss stopped in 84% (27/32) with regrowth noted in 69% (22/32) of patients.³² Voriconazole and fluconazole share structural similarity not present with other triazoles.^33,34 Because voriconazole-associated alopecia was reversed in the majority of patients who switched to itraconazole or posaconazole, the authors hypothesized that structural similarity of fluconazole and voriconazole may underly the greater risk for TE that is not a class effect of azole medications.³¹

Psychotropic Medications

Various psychotropic medications have been associated with hair loss. Valproic acid (or sodium valproate) is an anticonvulsant and mood-stabilizing agent used for the treatment of seizures, bipolar disorder (BD), migraines, and neuropathic pain.^35,36 Divalproex sodium (or divalproex) is an enteric-coated formulation of sodium valproate and valproic acid with similar indications. Valproate is a notorious culprit of medication-induced hair loss, with alopecia listed among the most common adverse reactions (reported >5%) on its structure product labeling document.³⁷ A systemic review and meta-analysis by Wang et al³⁸ estimated the overall incidence of valproate-related alopecia to be 11% (95% CI, 0.08-0.13). Although this meta-analysis did not find an association between incidence of alopecia and dose or duration of valproate therapy,³⁸ a separate review suggested that valproate-induced alopecia is dose dependent and can be managed with dose reduction.³⁹ A 12-month, randomized, double-blind study of treatment of BD with divalproex (valproate derivative), lithium, or placebo (2:1:1 ratio) showed a significantly higher frequency of alopecia in the divalproex group compared with placebo (16% [30/187] vs 6% [6/94]; P=.03).⁴⁰ Valproate-related hair loss is characteristically diffuse and nonscarring, often noted 3 to 6 months following initiation of valproate.^41,42 The proposed mechanism of valproate-induced alopecia includes chelation of zinc and selenium,⁴³ and a reduction in serum biotinidase activity, thereby decreasing the availability of these essential micronutrients required for hair growth.⁴¹ Studies examining the effects of valproate administration and serum biotinidase activity in patients have yielded conflicting results.^44-46 In a study of children with seizures including 57 patients treated with valproic acid, 17 treated with carbamazepine, and 75 age- and sex-matched healthy controls, the authors found no significant differences in serum biotinidase enzyme activity across the 3 groups.⁴⁴ In contrast, a study of 75 children with seizures on valproic acid therapy stratified by dose (mean [SD])—group A: 28.7 [8.5] mg/kg/d; group B: 41.6 [4.9] mg/kg/d; group C: 64.5 [5.8] mg/kg/d—found that patients receiving higher doses (groups B and C) had significantly reduced serum biotinidase activity (1.22 [1.11] and 0.97 [0.07] mmol/min/L, respectively) compared with 50 healthy pediatric controls (5.20 [0.90] mmol/min/L; P<.001). The same study found biotin supplementation at 10 mg/d for 20 days led to resolution of alopecia in 22% (2/9) of patients with alopecia on valproic acid therapy.⁴⁵ Despite hypothesized effects of valproate on micronutrients, the role of mineral supplementation in treating valproate-associated hair loss remains unclear. There is evidence to suggest that valproic acid–associated alterations in serum biotinidase activity may be transient. In a study of 32 pediatric patients receiving valproic acid for the treatment of epilepsy, serum biotinidase activity was significantly lower after 3 months of valproic acid therapy compared with pretreatment levels (P<.05); at 6 months, the serum biotinidase activity was increased compared with 3 months (P<.05) and not significantly different from pretreatment levels (P>.05).⁴⁶ Hair regrowth has been observed following discontinuation or dose reduction of valproate therapy in some cases.^39,47

Lithium carbonate (lithium) is used in the treatment of BD. Despite its efficacy and low cost, its potential for adverse effects, narrow therapeutic index, and subsequent need for routine monitoring are factors that limit its use.⁴⁸ Some reported dermatologic adverse reactions on its structure product labeling include xerosis, thinning of hair, alopecia, xerosis cutis, psoriasis onset/exacerbation, and generalized pruritus.⁴⁹ A systematic review and meta-analysis of 385 studies identified 24 publications reporting adverse effects of lithium on hair with no significantly increased risk of alopecia overall.⁵⁰ The analysis included 2 randomized controlled trials comparing the effects of lithium and placebo on hair loss in patients with BD. Hair loss was reported in 7% (7/94) of patients taking lithium and 6% (6/94) of the placebo group in the 12-month study⁴⁰ and in 3% (1/32) of the lithium group and 0% (0/28) of the divalproex group in the 20-month study.⁵¹ Despite anecdotal reports of alopecia associated with lithium, there is a lack of high-quality evidence to support this claim. Of note, hypothyroidism is a known complication of lithium use, and serum testing of thyroid function at 6-month intervals is recommended for patients on lithium treatment.⁵² Because thyroid abnormalities can cause alopecia distinct from TE, new-onset alopecia during lithium use should prompt serum testing of thyroid function. The development of hypothyroidism secondary to lithium is not a direct contraindication to its use⁵³; rather, treatment should be focused on correction with thyroid replacement therapy (eg, supplementation with thyroxine).⁵⁴

Commonly prescribed antidepressant medications include selective serotonin reuptake inhibitors (SSRIs) and bupropion. Selective serotonin reuptake inhibitors affect the neuronal serotonin transporter, increasing the concentration of serotonin in the synaptic cleft available for stimulation of postsynaptic serotonin receptors^55,56; bupropion is an antidepressant medication that inhibits norepinephrine and dopamine reuptake at the synaptic cleft.⁵⁷ Alopecia is an infrequent (1 in 100 to 1 in 1000 patients) adverse effect for several SSRIs.^58-62 A recent systematic review identified a total of 71 cases of alopecia associated with SSRI use including citalopram (n=11), escitalopram (n=7), fluoxetine (n=27), fluoxvamine (n=5), paroxetine (n=4), and sertraline (n=20), with a median time to onset of hair shedding of 8.6 weeks (range, 3 days to 5 years). Discontinuation of the suspected culprit SSRI led to improvement and/or resolution in 63% (51/81) episodes of alopecia, with a median time to improvement and/or resolution of 4 weeks.⁶³ A comparative retrospective cohort study using a large US health claims database from 2006 to 2014 included more than 1 million new and mutually exclusive patients taking fluoxetine, fluvoxamine, sertraline, citalopram, escitalopram, paroxetine, duloxetine, venlafaxine, desvenlafaxine, and bupropion.⁶⁴ Overall, 1% (1569/150,404) of patients treated with bupropion received 1 or more physician visits for alopecia. Patients on SSRIs generally had a lower risk for hair loss compared with patients using bupropion (citalopram: hazard ratio [HR], 0.80 [95% CI, 0.74-0.86]; escitalopram: HR, 0.79 [95% CI, 0.74-0.86]; fluoxetine: HR, 0.68 [95% CI, 0.63-0.74]; paroxetine: HR, 0.68 [95% CI, 0.62-0.74]; sertraline: HR, 0.74 [95% CI, 0.69-0.79]), with the exception of fluvoxamine (HR, 0.93 [95% CI, 0.64-1.37]). However, the type of alopecia, time to onset, and time to resolution were not reported, making it difficult to assess whether the reported hair loss was consistent with medication-induced TE. Additionally, the authors acknowledged that bupropion may have been prescribed for smoking cessation, which may carry a different risk profile for the development of alopecia.⁶⁴ Several other case reports have described alopecia following treatment with SSRIs, including sertraline,⁶⁵ fluvoxamine,⁶⁶ paroxetine,⁶⁷ fluoxetine,⁶⁸ and escitalopram.⁶⁹

Overall, it appears that the use of SSRIs portends relatively low risk for alopecia and medication-induced TE. Little is known regarding the molecular effects of SSRIs on hair growth and the pathomechanism of SSRI-induced TE. The potential benefits of discontinuing a suspected culprit medication should be carefully weighed against the risks of medication cessation, and consideration should be given to alternative medications in the same class that also may be associated with TE. In patients requiring antidepressant therapy with suspected medication-induced TE, consider transitioning to a different class of medication with lower risk of medication-induced alopecia; for example, discontinuing bupropion in favor of an SSRI.

Final Thoughts

Medication-induced alopecia is an undesired side effect of many commonly used drugs and drug classes, including retinoids, azole antifungals, and mood stabilizers. Although the precise pathomechanisms of medication-induced TE remain unclear, the recommended management often requires identification of the likely causative agent and its discontinuation, if possible. Suspicion for medication-induced TE should prompt a thorough history of recent changes to medications, risk factors for nutritional deficiencies, underlying illnesses, and recent surgical procedures. Underlying nutritional, electrolyte, and/or metabolic disturbances should be corrected. In part 2 of this series, we will discuss medication-induced alopecia associated with anticoagulant and antihypertensive medications.

Alopecia is a commonly reported side effect of various medications. Anagen effluvium and telogen effluvium (TE) are considered the most common mechanisms underlying medication-related hair loss. Anagen effluvium is associated with chemotherapeutic agents and radiation therapy, with anagen shedding typically occurring within 2 weeks of medication administration.^1,2 Medication-induced TE is a diffuse nonscarring alopecia that is a reversible reactive process.^3-5 Telogen effluvium is clinically apparent as a generalized shedding of scalp hair 1 to 6 months after an inciting cause.⁶ The underlying cause of TE may be multifactorial and difficult to identify given the delay between the trigger and the onset of clinically apparent hair loss. Other known triggers of TE include acute illness,^7,8 nutritional deficiencies,^4,9 and/or major surgery.¹⁰

Each hair follicle independently and sequentially progresses through anagen growth, catagen transition, and telogen resting phases. In the human scalp, the telogen phase typically lasts 3 months, at the end of which the telogen hair is extruded from the scalp. Anagen and telogen follicles typically account for an average of 90% and 10% of follicles on the human scalp, respectively.¹¹ Immediate anagen release is hypothesized to be the mechanism underlying medication-induced TE.¹² This theory suggests that an increased percentage of anagen follicles prematurely enter the telogen phase, with a notable increase in hair shedding at the conclusion of the telogen phase approximately 1 to 6 months later.¹² First-line management of medication-induced TE is identification and cessation of the causative agent, if possible. Notable regrowth of hair is expected several months after removal of the inciting medication. In part 1 of this 2-part series, we review the existing literature to identify common culprits of medication-induced TE, including retinoids, antifungals, and psychotropic medications.

Retinoids

Retinoids are vitamin A derivatives used in the treatment of a myriad of dermatologic and nondermatologic conditions.^13,14 Retinoids modulate sebum production,¹⁵ keratinocyte proliferation,¹⁶ and epithelial differentiation through signal transduction downstream of the ligand-activated nuclear retinoic acid receptors and retinoid X receptors.^13,14,17 The recommended daily dosage of retinol is 900 µg retinol activity equivalent (3000 IU) for men and 700 µg retinol activity equivalent (2333 IU) for women. Retinoids are used in the treatment of acne vulgaris,¹⁸ psoriasis,¹⁹ and ichthyosis.²⁰ The most commonly reported adverse effects of systemic retinoid therapy include cheilitis, alopecia, and xerosis.²¹ Retinoid-associated alopecia is dose and duration dependent.^19,21-24 A prospective study of acitretin therapy in plaque psoriasis reported that more than 63% (42/66) of patients on 50 mg or more of acitretin daily for 6 months or longer experienced alopecia that reversed with discontinuation.²³ A systematic review of isotretinoin use in acne showed alopecia was seen in 3.2% (18/565) of patients on less than 0.5 mg/kg/d of isotretinoin and in 5.7% (192/3375) of patients on 0.5 mg/kg/d or less of isotretinoin.²⁴ In a phase 2 clinical trial of orally administered 9-cis-retinoic acid (alitretinoin) in the treatment of Kaposi sarcoma related to AIDS, 42% (24/57) of adult male patients receiving 60, 100, or 140 mg/m² alitretinoin daily (median treatment duration, 15.1 weeks) reported alopecia as an adverse effect of treatment.²⁵ In one case report, a patient who ingested 500,000 IU of vitamin A daily for 4 months and then 100,000 IU monthly for 6 months experienced diffusely increased shedding of scalp hair along with muscle soreness, nail dystrophy, diffuse skin rash, and refractory ascites; he was found to have severe liver damage secondary to hypervitaminosis A that required liver transplantation.²⁶ Regarding the pathomechanism of retinoid-induced alopecia, animal and in vitro studies similarly have demonstrated that all-trans-retinoic acid appears to exert its inhibitory effects on hair follicle growth via the influence of the transforming growth factor β2 and SMAD2/3 pathway influence on dermal papillae cells.^14,27 Development of hair loss secondary to systemic retinoid therapy may be managed with dose reduction or cessation.

Antifungals

Azole medications have broad-spectrum fungistatic activity against a wide range of yeast and filamentous fungi. Azoles inhibit sterol 14α-demethylase activity, impairing ergosterol synthesis and thereby disrupting plasma membrane synthesis and activity of membrane-bound enzymes.²⁸ Fluconazole is a systemic oral agent in this class that was first approved by the US Food and Drug Administration (FDA) for use in the 1990s.²⁹ A retrospective study by the National Institute of Allergy and Infectious Disease Mycoses Study Group followed the clinical course of 33 patients who developed alopecia while receiving fluconazole therapy for various mycoses.³⁰ The majority (88% [29/33]) of patients received 400 mg or more of fluconazole daily. The median time to hair loss after starting fluconazole was 3 months, and the scalp was involved in all cases. In 97% (32/33) of patients, resolution of alopecia was noted following discontinuation of fluconazole or a dose reduction of 50% or more. In 85% (28/33) of patients, complete resolution of alopecia occurred within 6 months of fluconazole cessation or dose reduction.³⁰ Fluconazole-induced TE was reproducible in an animal model using Wistar rats³¹; however, further studies are required to clarify the molecular pathways of its effect on hair growth.

Voriconazole is an azole approved for the treatment of invasive aspergillosis, candidemia, and fungal infections caused by Scedosporium apiospermum and Fusarium species. A retrospective survey study of patients who received voriconazole for 1 month or longer found a considerable proportion of patients developed diffuse reversible hair loss.³² Scalp alopecia was noted in 79% (120/152) of patients who completed the survey, with a mean (SD) time to alopecia of 75 (54) days after initiation of voriconazole. Notable regrowth was reported in 69% (79/114) of patients who discontinued voriconazole for at least 3 months. A subgroup of 32 patients were changed to itraconazole or posaconazole, and hair loss stopped in 84% (27/32) with regrowth noted in 69% (22/32) of patients.³² Voriconazole and fluconazole share structural similarity not present with other triazoles.^33,34 Because voriconazole-associated alopecia was reversed in the majority of patients who switched to itraconazole or posaconazole, the authors hypothesized that structural similarity of fluconazole and voriconazole may underly the greater risk for TE that is not a class effect of azole medications.³¹

Psychotropic Medications

Various psychotropic medications have been associated with hair loss. Valproic acid (or sodium valproate) is an anticonvulsant and mood-stabilizing agent used for the treatment of seizures, bipolar disorder (BD), migraines, and neuropathic pain.^35,36 Divalproex sodium (or divalproex) is an enteric-coated formulation of sodium valproate and valproic acid with similar indications. Valproate is a notorious culprit of medication-induced hair loss, with alopecia listed among the most common adverse reactions (reported >5%) on its structure product labeling document.³⁷ A systemic review and meta-analysis by Wang et al³⁸ estimated the overall incidence of valproate-related alopecia to be 11% (95% CI, 0.08-0.13). Although this meta-analysis did not find an association between incidence of alopecia and dose or duration of valproate therapy,³⁸ a separate review suggested that valproate-induced alopecia is dose dependent and can be managed with dose reduction.³⁹ A 12-month, randomized, double-blind study of treatment of BD with divalproex (valproate derivative), lithium, or placebo (2:1:1 ratio) showed a significantly higher frequency of alopecia in the divalproex group compared with placebo (16% [30/187] vs 6% [6/94]; P=.03).⁴⁰ Valproate-related hair loss is characteristically diffuse and nonscarring, often noted 3 to 6 months following initiation of valproate.^41,42 The proposed mechanism of valproate-induced alopecia includes chelation of zinc and selenium,⁴³ and a reduction in serum biotinidase activity, thereby decreasing the availability of these essential micronutrients required for hair growth.⁴¹ Studies examining the effects of valproate administration and serum biotinidase activity in patients have yielded conflicting results.^44-46 In a study of children with seizures including 57 patients treated with valproic acid, 17 treated with carbamazepine, and 75 age- and sex-matched healthy controls, the authors found no significant differences in serum biotinidase enzyme activity across the 3 groups.⁴⁴ In contrast, a study of 75 children with seizures on valproic acid therapy stratified by dose (mean [SD])—group A: 28.7 [8.5] mg/kg/d; group B: 41.6 [4.9] mg/kg/d; group C: 64.5 [5.8] mg/kg/d—found that patients receiving higher doses (groups B and C) had significantly reduced serum biotinidase activity (1.22 [1.11] and 0.97 [0.07] mmol/min/L, respectively) compared with 50 healthy pediatric controls (5.20 [0.90] mmol/min/L; P<.001). The same study found biotin supplementation at 10 mg/d for 20 days led to resolution of alopecia in 22% (2/9) of patients with alopecia on valproic acid therapy.⁴⁵ Despite hypothesized effects of valproate on micronutrients, the role of mineral supplementation in treating valproate-associated hair loss remains unclear. There is evidence to suggest that valproic acid–associated alterations in serum biotinidase activity may be transient. In a study of 32 pediatric patients receiving valproic acid for the treatment of epilepsy, serum biotinidase activity was significantly lower after 3 months of valproic acid therapy compared with pretreatment levels (P<.05); at 6 months, the serum biotinidase activity was increased compared with 3 months (P<.05) and not significantly different from pretreatment levels (P>.05).⁴⁶ Hair regrowth has been observed following discontinuation or dose reduction of valproate therapy in some cases.^39,47

Lithium carbonate (lithium) is used in the treatment of BD. Despite its efficacy and low cost, its potential for adverse effects, narrow therapeutic index, and subsequent need for routine monitoring are factors that limit its use.⁴⁸ Some reported dermatologic adverse reactions on its structure product labeling include xerosis, thinning of hair, alopecia, xerosis cutis, psoriasis onset/exacerbation, and generalized pruritus.⁴⁹ A systematic review and meta-analysis of 385 studies identified 24 publications reporting adverse effects of lithium on hair with no significantly increased risk of alopecia overall.⁵⁰ The analysis included 2 randomized controlled trials comparing the effects of lithium and placebo on hair loss in patients with BD. Hair loss was reported in 7% (7/94) of patients taking lithium and 6% (6/94) of the placebo group in the 12-month study⁴⁰ and in 3% (1/32) of the lithium group and 0% (0/28) of the divalproex group in the 20-month study.⁵¹ Despite anecdotal reports of alopecia associated with lithium, there is a lack of high-quality evidence to support this claim. Of note, hypothyroidism is a known complication of lithium use, and serum testing of thyroid function at 6-month intervals is recommended for patients on lithium treatment.⁵² Because thyroid abnormalities can cause alopecia distinct from TE, new-onset alopecia during lithium use should prompt serum testing of thyroid function. The development of hypothyroidism secondary to lithium is not a direct contraindication to its use⁵³; rather, treatment should be focused on correction with thyroid replacement therapy (eg, supplementation with thyroxine).⁵⁴

Commonly prescribed antidepressant medications include selective serotonin reuptake inhibitors (SSRIs) and bupropion. Selective serotonin reuptake inhibitors affect the neuronal serotonin transporter, increasing the concentration of serotonin in the synaptic cleft available for stimulation of postsynaptic serotonin receptors^55,56; bupropion is an antidepressant medication that inhibits norepinephrine and dopamine reuptake at the synaptic cleft.⁵⁷ Alopecia is an infrequent (1 in 100 to 1 in 1000 patients) adverse effect for several SSRIs.^58-62 A recent systematic review identified a total of 71 cases of alopecia associated with SSRI use including citalopram (n=11), escitalopram (n=7), fluoxetine (n=27), fluoxvamine (n=5), paroxetine (n=4), and sertraline (n=20), with a median time to onset of hair shedding of 8.6 weeks (range, 3 days to 5 years). Discontinuation of the suspected culprit SSRI led to improvement and/or resolution in 63% (51/81) episodes of alopecia, with a median time to improvement and/or resolution of 4 weeks.⁶³ A comparative retrospective cohort study using a large US health claims database from 2006 to 2014 included more than 1 million new and mutually exclusive patients taking fluoxetine, fluvoxamine, sertraline, citalopram, escitalopram, paroxetine, duloxetine, venlafaxine, desvenlafaxine, and bupropion.⁶⁴ Overall, 1% (1569/150,404) of patients treated with bupropion received 1 or more physician visits for alopecia. Patients on SSRIs generally had a lower risk for hair loss compared with patients using bupropion (citalopram: hazard ratio [HR], 0.80 [95% CI, 0.74-0.86]; escitalopram: HR, 0.79 [95% CI, 0.74-0.86]; fluoxetine: HR, 0.68 [95% CI, 0.63-0.74]; paroxetine: HR, 0.68 [95% CI, 0.62-0.74]; sertraline: HR, 0.74 [95% CI, 0.69-0.79]), with the exception of fluvoxamine (HR, 0.93 [95% CI, 0.64-1.37]). However, the type of alopecia, time to onset, and time to resolution were not reported, making it difficult to assess whether the reported hair loss was consistent with medication-induced TE. Additionally, the authors acknowledged that bupropion may have been prescribed for smoking cessation, which may carry a different risk profile for the development of alopecia.⁶⁴ Several other case reports have described alopecia following treatment with SSRIs, including sertraline,⁶⁵ fluvoxamine,⁶⁶ paroxetine,⁶⁷ fluoxetine,⁶⁸ and escitalopram.⁶⁹

Overall, it appears that the use of SSRIs portends relatively low risk for alopecia and medication-induced TE. Little is known regarding the molecular effects of SSRIs on hair growth and the pathomechanism of SSRI-induced TE. The potential benefits of discontinuing a suspected culprit medication should be carefully weighed against the risks of medication cessation, and consideration should be given to alternative medications in the same class that also may be associated with TE. In patients requiring antidepressant therapy with suspected medication-induced TE, consider transitioning to a different class of medication with lower risk of medication-induced alopecia; for example, discontinuing bupropion in favor of an SSRI.

Final Thoughts

Medication-induced alopecia is an undesired side effect of many commonly used drugs and drug classes, including retinoids, azole antifungals, and mood stabilizers. Although the precise pathomechanisms of medication-induced TE remain unclear, the recommended management often requires identification of the likely causative agent and its discontinuation, if possible. Suspicion for medication-induced TE should prompt a thorough history of recent changes to medications, risk factors for nutritional deficiencies, underlying illnesses, and recent surgical procedures. Underlying nutritional, electrolyte, and/or metabolic disturbances should be corrected. In part 2 of this series, we will discuss medication-induced alopecia associated with anticoagulant and antihypertensive medications.

Ferritin is an iron storage protein crucial to human iron homeostasis. Because serum ferritin levels are in dynamic equilibrium with the body’s iron stores, ferritin often is measured as a reflection of iron status; however, ferritin also is an acute-phase reactant whose levels may be nonspecifically elevated in a wide range of inflammatory conditions. The various processes that alter serum ferritin levels complicate the clinical interpretation of this laboratory value. In this article, we review the structure and function of ferritin and provide a guide for clinical use.

Overview of Iron

Iron is an essential element of key biologic functions including DNA synthesis and repair, oxygen transport, and oxidative phosphorylation. The body’s iron stores are mainly derived from internal iron recycling following red blood cell breakdown, while 5% to 10% is supplied by dietary intake.^1-3 Iron metabolism is of particular importance in cells of the reticuloendothelial system (eg, spleen, liver, bone marrow), where excess iron must be appropriately sequestered and from which iron can be mobilized.⁴ Sufficient iron stores are necessary for proper cellular function and survival, as iron is a necessary component of hemoglobin for oxygen delivery, iron-sulfur clusters in electron transport, and enzyme cofactors in other cellular processes.

Although labile pools of biologically active free iron exist in limited amounts within cells, excess free iron can generate free radicals that damage cellular proteins, lipids, and nucleic acids.^5-7 As such, most intracellular iron is captured within ferritin molecules. The excretion of iron is unregulated and occurs through loss in sweat, menstruation, hair shedding, skin desquamation, and enterocyte turnover.⁸ The lack of regulated excretion highlights the need for a tightly regulated system of uptake, transportation, storage, and sequestration to maintain iron homeostasis.

Overview of Ferritin Structure and Function

Ferritin is a key regulator of iron homeostasis that also serves as an important clinical indicator of body iron status. Ferritin mainly is found as an intracellular cytosolic iron storage and detoxification protein structured as a hollow 24-subunit polymer shell that can sequester up to 4500 atoms of iron within its core.^9,10 The 24-mer is composed of both ferritin L (FTL) and ferritin H (FTH) subunits, with dynamic regulation of the H:L ratio dependent on the context and tissue in which ferritin is found.⁶

Ferritin H possesses ferroxidase, which facilitates oxidation of ferrous (Fe²⁺) iron into ferric (Fe³⁺) iron, which can then be incorporated into the mineral core of the ferritin heteropolymer.¹¹ Ferritin L is more abundant in the spleen and liver, while FTH is found predominantly in the heart and kidneys where the increased ferroxidase activity may confer an increased ability to oxidize Fe²⁺ and limit oxidative stress.⁶

Regulation of Ferritin Synthesis and Secretion

Ferritin synthesis is regulated by intracellular nonheme iron levels, governed mainly by an iron response element (IRE) and iron response protein (IRP) translational repression system. Both FTH and FTL messenger RNA (mRNA) contain an IRE that is a regulatory stem-loop structure in the 5´ untranslated region. When the IRE is bound by IRP1 or IRP2, mRNA translation of ferritin subunits is suppressed.⁶ In low iron conditions, IRPs have greater affinity for IRE, and binding suppresses ferritin translation.¹² In high iron conditions, IRPs have a decreased affinity for IRE, and ferritin mRNA synthesis is increased.¹³ Additionally, inflammatory cytokines such as tumor necrosis factor α and IL-1α transcriptionally induce FTH synthesis, resulting in an increased population of H-rich ferritins.^11,14-16 A study using cultured human primary skin fibroblasts demonstrated UV radiation–induced increases in free intracellular iron content.^17,18 Pourzand et al¹⁸ suggested that UV-mediated damage of lysosomal membranes results in leakage of lysosomal proteases into the cytosol, contributing to degradation of intracellular ferritin and subsequent release of iron within skin fibroblasts. The increased intracellular iron downregulates IRPs and increases ferritin mRNA synthesis,¹⁸ consistent with prior findings of increased ferritin synthesis in skin that is induced by UV radiation.¹⁹

Molecular analysis of serum ferritin in iron-overloaded mice revealed that extracellular ferritin found in the serum is composed of a greater fraction of FTL and has lower iron content than intracellular ferritin. The low iron content of serum ferritin compared with intracellular ferritin and transferrin suggests that serum ferritin is not a major pathway of systemic iron transport.¹⁰ However, locally secreted ferritins may play a greater role in iron transport and release in selected tissues. Additionally, in vitro studies of cell cultures from humans and mice have demonstrated the ability of macrophages to secrete ferritin, suggesting that macrophages are an important cellular source of serum ferritin.^10,20 As such, serum ferritin generally may reflect body iron status but more specifically reflects macrophage iron status.¹⁰ Although the exact pathways of ferritin release are unknown, it is hypothesized that ferritin secretion occurs through cytosolic autophagy followed by secretion of proteins from the lysosomal compartment.^10,18,21

Low Ferritin and Iron Deficiency—Although bone marrow biopsy with iron staining remains the gold standard for diagnosis of iron deficiency, serum ferritin is a much more accessible and less invasive tool for evaluation of iron status. A serum ferritin level below 12 μg/L is highly specific for iron depletion,²² with a higher cutoff recommended in clinical practice to improve diagnostic sensitivity.^23,24 Conditions independent of iron deficiency that may reduce serum ferritin include hypothyroidism and ascorbate deficiency, though neither condition has been reported to interfere with appropriate diagnosis of iron deficiency.²⁵ Guyatt et al²⁶ conducted a systematic review of laboratory tests used in the diagnosis of iron deficiency anemia and identified 55 studies suitable for inclusion. Based on an area under the receiver operating characteristic curve (AUROC) of 0.95, serum ferritin values were superior to transferrin saturation (AUROC, 0.74), red cell protoporphyrin (AUROC, 0.77), red cell volume distribution width (AUROC, 0.62), and mean cell volume (AUROC, 0.76) for diagnosis of IDA, verified by histologic examination of aspirated bone marrow.²⁶ The likelihood ratio of iron deficiency begins to decrease for serum ferritin values of 40 μg/L or greater. For patients with inflammatory conditions—patients with concomitant chronic renal failure, inflammatory disease, infection, rheumatoid arthritis, liver disease, inflammatory bowel disease, and malignancy—the likelihood of iron deficiency begins to decrease at serum ferritin levels of 70 μg/L or greater.²⁶ Similarly, the World Health Organization recommends that in adults with infection or inflammation, serum ferritin levels lower than 70 μg/L may be used to indicate iron deficiency.²⁴ A serum ferritin level of 41 μg/L or lower was found to have a sensitivity and specificity of 98% for discriminating between iron-deficiency anemia and anemia of chronic disease (diagnosed based on bone marrow biopsy with iron staining), with an AUROC of 0.98.²⁷ As such, we recommend using a serum ferritin level of 40 μg/L or lower in patients who are otherwise healthy as an indicator of iron deficiency.

The threshold for iron supplementation may vary based on age, sex, and race. In women, ferritin levels increase during menopause and peak after menopause; ferritin levels are higher in men than in women.^28-30 A multisite longitudinal cohort study of 70 women in the United States found that the mean (SD) ferritin valuewas 69.5 (81.7) μg/L premenopause and 128.8 (125.7) μg/L postmenopause (P<.01).³¹ A separate longitudinal survey study of 8564 patients in China found that the mean (SE) ferritin value was 201.55 (3.60) μg/L for men and 80.46 (1.64) μg/L for women (P<.0001).³² Analysis of serum ferritin levels of 3554 male patients from the third National Health and Nutrition Examination Survey demonstrated that patients who self-reported as non-Hispanic Black (n=1616) had significantly higher serum ferritin levels than non-Hispanic White patients (n=1938)(serum ferritin difference of 37.1 μg/L)(P<.0001).³³ The British Society for Haematology guidelines recommend that the threshold of serum ferritin for diagnosing iron deficiency should take into account age-, sex-, and race-based differences.³⁴ Ferritin and Hair—Cutaneous manifestations of iron deficiency include koilonychia, glossitis, pruritus, angular cheilitis, and telogen effluvium (TE).¹ A case-control study of 30 females aged 15 to 45 years demonstrated that the mean (SD) ferritin level was significantly lower in patients with TE than those with no hair loss (16.3 [12.6] ng/mL vs 60.3 [50.1] ng/mL; P<.0001). Using a threshold of 30 μg/L or lower, the investigators found that the odds ratio for TE was 21.0 (95% CI, 4.2-105.0) in patients with low serum ferritin.³⁵

Another retrospective review of 54 patients with diffuse hair loss and 55 controls compared serum vitamin B₁₂, folate, thyroid-stimulating hormone, zinc, ferritin, and 25-hydroxy vitamin D levels between the 2 groups.³⁶ Exclusion criteria were clinical diagnoses of female pattern hair loss (androgenetic alopecia), pregnancy, menopause, metabolic and endocrine disorders, hormone replacement therapy, chemotherapy, immunosuppressive therapy, vitamin and mineral supplementation, scarring alopecia, eating disorders, and restrictive diets. Compared with controls, patients with diffuse nonscarring hair loss were found to have significantly lower ferritin (mean [SD], 14.72 [10.70] ng/mL vs 25.30 [14.41] ng/mL; P<.001) and 25-hydroxy vitamin D levels (mean [SD], 14.03 [8.09] ng/mL vs 17.01 [8.59] ng/mL; P=.01).³⁶

In contrast, a separate case-control study of 381 cases and 76 controls found no increase in the rate of iron deficiency—defined as ferritin ≤15 μg/L or ≤40 μg/L—among women with female pattern hair loss or chronic TE vs controls.³⁷ Taken together, these studies suggest that iron status may play a role in TE, a process that may result from nutritional deficiency, trauma, or physical or psychological stress³⁸; however, there is insufficient evidence to suggest that low iron status impacts androgenetic alopecia, in which its multifactorial pathogenesis implicates genetic and hormonal factors.³⁹ More research is needed to clarify the potential associations between iron deficiency and types of hair loss. Additionally, it is unclear whether iron supplementation improves hair growth parameters such as density and caliber.⁴⁰

Low serum ferritin (<40 μg/L) with concurrent symptoms of iron deficiency, including fatigue, pallor, dyspnea on exertion, or hair loss, should prompt treatment with supplemental iron.^41-43 Generally, ferrous (Fe2⁺) salts are preferred to ferric (Fe3⁺) salts, as the former is more readily absorbed through the duodenal mucosa⁴⁴ and is the more common formulation in commercially available supplements in the United States.⁴⁵ Oral supplementation with ferrous sulfate 325 mg (65 mg elemental iron) tablets is the first-line therapy for iron deficiency anemia.¹ Alternatively, ferrous gluconate 324 mg (38 mg elemental iron) over-the-counter and its liquid form has demonstrated superior absorption compared to ferrous sulfate tablets in a clinical study with peritoneal dialysis patients.^1,46 One study suggested that oral iron 40 to 80 mg should be taken every other day to increase absorption.⁴⁷ Due to improved bioavailability, intravenous iron may be utilized in patients with malabsorption, renal failure, or intolerance to oral iron (including those with gastric ulcers or active inflammatory bowel disease), with the formulation chosen based on underlying comorbidities and potential risks.^1,48 The theoretical risk for potentiating bacterial growth by increasing the amount of unbound iron in the blood raises concerns of iron supplementation in patients with infection or sepsis. Although far from definitive, existing data suggest that risk for infection is greater with intravenous iron supplementation and should be carefully considered prior to use.^49,50Elevated Ferritin—Elevated ferritin may be difficult to interpret given the multitude of conditions that can cause it.^23,51,52 Elevated serum ferritin can be broadly characterized by increased synthesis due to iron overload, increased synthesis due to inflammation, or increased ferritin release from cellular damage.³⁴ Further complicating interpretation is the potential diurnal fluctuations in serum iron levels dependent on dietary intake and timing of laboratory evaluation, choice of assay, differences in reference standards, and variations in calibration procedures that can lead to analytic variability in the measurement of ferritin.^23,53,54

Among healthy patients, serum ferritin is directly proportional to iron status.^9,51 A study utilizing weekly phlebotomy of 22 healthy participants to measure serum ferritin and calculate mobilizable storage iron found a strong positive correlation between the 2 variables (r=0.83, P<.001), with each 1-μg/L increase of serum ferritin corresponding to approximately an 8-mg increase of storage iron; the initial serum ferritin values ranged from 2 to 83 μg/L in females and 36 to 224 μg/L in males.⁵⁵ The correlation of ferritin with iron status also was supported by the significant correlation between the number of transfusions received in patients with transfusion-related iron overload and serum ferritin levels (r=0.89, P<.001), with an average increase of 60 μg/L per transfusion.⁵¹

Clinical guidelines on the interpretation of serum ferritin levels by Cullis et al³⁴ recommend a normal upper limit of 200 μg/L for healthy females and 300 μg/L for healthy males. Outside of clinical syndromes associated with iron overload, Lee and Means⁵⁶ found that serum ferritin of 1000 μg/L or higher was a nonspecific marker of disease, including infection and/or neoplastic disorders. We have adapted these guidelines to propose a workflow for evaluation of serum ferritin levels (Figure). In patients with inflammatory conditions or those affected by metabolic syndrome, elevated serum ferritin does not correlate with body iron status.^57,58 It is believed that inflammatory cytokines, including tumor necrosis factor α and IL-1α, can upregulate ferritin synthesis independent of cellular iron stores.^15,16 Several studies have examined the relationship between insulin resistance and/or metabolic syndrome with serum ferritin levels.^31,32 Han et al³² found that elevated serum ferritin was significantly associated with higher risk for metabolic syndrome in men (P<.0001) but not in women.

Although cutaneous manifestations of iron overload can be seen as skin hyperpigmentation due to increased iron deposits and increased melanin production,²² the effects of elevated ferritin on the skin and hair are not well known. Iron overload is a known trigger of porphyria cutanea tarda (PCT),⁵⁹ a condition in which reduced or absent enzymatic activity of uroporphyrinogen decarboxylase (UROD) leads to build up of toxic porphyrins in various organs.⁶⁰ In the skin, PCT manifests as a blistering photosensitive eruption that may resolve as dyspigmentation, scarring, and milia.⁶¹ Phlebotomy is first-line therapy in PCT to reduce serum iron and subsequent formation of UROD inhibitors, with guidelines suggesting discontinuation of phlebotomy when serum ferritin levels reach 20 ng/mL or lower.⁶⁰ Hyperferritinemia (serum ferritin >500 μg/L) is a common finding in several inflammatory disorders often accompanied by clinically apparent cutaneous symptoms such as adult-onset Still disease,⁶² hemophagocytic lymphohistiocytosis,^63,64 and anti-melanoma differentiation-associated gene 5 dermatomyositis.⁶⁵ Among these conditions, serum ferritin levels have been reported to correlate with disease activity, raising the question of whether ferritin is a bystander or a driver of the underlying pathology.^62,66,67 However, rapid decline of serum ferritin levels with treatment and control of inflammatory cytokines suggest that ferritin is unlikely to contribute to pathology.^62,67

Final Thoughts

Many clinical studies have examined the association between hair health and body iron status, the collective findings of which suggest that iron deficiency may be associated with TE. Among commonly measured serum iron parameters, low ferritin is a highly specific and sensitive marker for diagnosing iron deficiency. Serum ferritin may be a clinically useful tool for ruling out underlying iron deficiency in patients presenting with hair loss. Despite advances in our understanding of the molecular mechanisms of ferritin synthesis and regulation, whether ferritin itself contributes to cutaneous pathology is poorly understood.^35,36,68-74 For patients who are otherwise healthy with low suspicion for inflammatory disorders, chronic systemic illnesses, or malignancy, serum ferritin can be used as an indicator of body iron status. The workup for slightly elevated serum ferritin should be interpreted in the context of other laboratory findings and should be reassessed over time. Serum ferritin levels above 1000 μg/L warrant further investigation into causes such as iron overload conditions and underlying inflammatory conditions or malignancy.

Ferritin is an iron storage protein crucial to human iron homeostasis. Because serum ferritin levels are in dynamic equilibrium with the body’s iron stores, ferritin often is measured as a reflection of iron status; however, ferritin also is an acute-phase reactant whose levels may be nonspecifically elevated in a wide range of inflammatory conditions. The various processes that alter serum ferritin levels complicate the clinical interpretation of this laboratory value. In this article, we review the structure and function of ferritin and provide a guide for clinical use.

Overview of Iron

Iron is an essential element of key biologic functions including DNA synthesis and repair, oxygen transport, and oxidative phosphorylation. The body’s iron stores are mainly derived from internal iron recycling following red blood cell breakdown, while 5% to 10% is supplied by dietary intake.^1-3 Iron metabolism is of particular importance in cells of the reticuloendothelial system (eg, spleen, liver, bone marrow), where excess iron must be appropriately sequestered and from which iron can be mobilized.⁴ Sufficient iron stores are necessary for proper cellular function and survival, as iron is a necessary component of hemoglobin for oxygen delivery, iron-sulfur clusters in electron transport, and enzyme cofactors in other cellular processes.

Although labile pools of biologically active free iron exist in limited amounts within cells, excess free iron can generate free radicals that damage cellular proteins, lipids, and nucleic acids.^5-7 As such, most intracellular iron is captured within ferritin molecules. The excretion of iron is unregulated and occurs through loss in sweat, menstruation, hair shedding, skin desquamation, and enterocyte turnover.⁸ The lack of regulated excretion highlights the need for a tightly regulated system of uptake, transportation, storage, and sequestration to maintain iron homeostasis.

Overview of Ferritin Structure and Function

Ferritin is a key regulator of iron homeostasis that also serves as an important clinical indicator of body iron status. Ferritin mainly is found as an intracellular cytosolic iron storage and detoxification protein structured as a hollow 24-subunit polymer shell that can sequester up to 4500 atoms of iron within its core.^9,10 The 24-mer is composed of both ferritin L (FTL) and ferritin H (FTH) subunits, with dynamic regulation of the H:L ratio dependent on the context and tissue in which ferritin is found.⁶

Ferritin H possesses ferroxidase, which facilitates oxidation of ferrous (Fe²⁺) iron into ferric (Fe³⁺) iron, which can then be incorporated into the mineral core of the ferritin heteropolymer.¹¹ Ferritin L is more abundant in the spleen and liver, while FTH is found predominantly in the heart and kidneys where the increased ferroxidase activity may confer an increased ability to oxidize Fe²⁺ and limit oxidative stress.⁶

Regulation of Ferritin Synthesis and Secretion

Ferritin synthesis is regulated by intracellular nonheme iron levels, governed mainly by an iron response element (IRE) and iron response protein (IRP) translational repression system. Both FTH and FTL messenger RNA (mRNA) contain an IRE that is a regulatory stem-loop structure in the 5´ untranslated region. When the IRE is bound by IRP1 or IRP2, mRNA translation of ferritin subunits is suppressed.⁶ In low iron conditions, IRPs have greater affinity for IRE, and binding suppresses ferritin translation.¹² In high iron conditions, IRPs have a decreased affinity for IRE, and ferritin mRNA synthesis is increased.¹³ Additionally, inflammatory cytokines such as tumor necrosis factor α and IL-1α transcriptionally induce FTH synthesis, resulting in an increased population of H-rich ferritins.^11,14-16 A study using cultured human primary skin fibroblasts demonstrated UV radiation–induced increases in free intracellular iron content.^17,18 Pourzand et al¹⁸ suggested that UV-mediated damage of lysosomal membranes results in leakage of lysosomal proteases into the cytosol, contributing to degradation of intracellular ferritin and subsequent release of iron within skin fibroblasts. The increased intracellular iron downregulates IRPs and increases ferritin mRNA synthesis,¹⁸ consistent with prior findings of increased ferritin synthesis in skin that is induced by UV radiation.¹⁹

Molecular analysis of serum ferritin in iron-overloaded mice revealed that extracellular ferritin found in the serum is composed of a greater fraction of FTL and has lower iron content than intracellular ferritin. The low iron content of serum ferritin compared with intracellular ferritin and transferrin suggests that serum ferritin is not a major pathway of systemic iron transport.¹⁰ However, locally secreted ferritins may play a greater role in iron transport and release in selected tissues. Additionally, in vitro studies of cell cultures from humans and mice have demonstrated the ability of macrophages to secrete ferritin, suggesting that macrophages are an important cellular source of serum ferritin.^10,20 As such, serum ferritin generally may reflect body iron status but more specifically reflects macrophage iron status.¹⁰ Although the exact pathways of ferritin release are unknown, it is hypothesized that ferritin secretion occurs through cytosolic autophagy followed by secretion of proteins from the lysosomal compartment.^10,18,21

Low Ferritin and Iron Deficiency—Although bone marrow biopsy with iron staining remains the gold standard for diagnosis of iron deficiency, serum ferritin is a much more accessible and less invasive tool for evaluation of iron status. A serum ferritin level below 12 μg/L is highly specific for iron depletion,²² with a higher cutoff recommended in clinical practice to improve diagnostic sensitivity.^23,24 Conditions independent of iron deficiency that may reduce serum ferritin include hypothyroidism and ascorbate deficiency, though neither condition has been reported to interfere with appropriate diagnosis of iron deficiency.²⁵ Guyatt et al²⁶ conducted a systematic review of laboratory tests used in the diagnosis of iron deficiency anemia and identified 55 studies suitable for inclusion. Based on an area under the receiver operating characteristic curve (AUROC) of 0.95, serum ferritin values were superior to transferrin saturation (AUROC, 0.74), red cell protoporphyrin (AUROC, 0.77), red cell volume distribution width (AUROC, 0.62), and mean cell volume (AUROC, 0.76) for diagnosis of IDA, verified by histologic examination of aspirated bone marrow.²⁶ The likelihood ratio of iron deficiency begins to decrease for serum ferritin values of 40 μg/L or greater. For patients with inflammatory conditions—patients with concomitant chronic renal failure, inflammatory disease, infection, rheumatoid arthritis, liver disease, inflammatory bowel disease, and malignancy—the likelihood of iron deficiency begins to decrease at serum ferritin levels of 70 μg/L or greater.²⁶ Similarly, the World Health Organization recommends that in adults with infection or inflammation, serum ferritin levels lower than 70 μg/L may be used to indicate iron deficiency.²⁴ A serum ferritin level of 41 μg/L or lower was found to have a sensitivity and specificity of 98% for discriminating between iron-deficiency anemia and anemia of chronic disease (diagnosed based on bone marrow biopsy with iron staining), with an AUROC of 0.98.²⁷ As such, we recommend using a serum ferritin level of 40 μg/L or lower in patients who are otherwise healthy as an indicator of iron deficiency.

The threshold for iron supplementation may vary based on age, sex, and race. In women, ferritin levels increase during menopause and peak after menopause; ferritin levels are higher in men than in women.^28-30 A multisite longitudinal cohort study of 70 women in the United States found that the mean (SD) ferritin valuewas 69.5 (81.7) μg/L premenopause and 128.8 (125.7) μg/L postmenopause (P<.01).³¹ A separate longitudinal survey study of 8564 patients in China found that the mean (SE) ferritin value was 201.55 (3.60) μg/L for men and 80.46 (1.64) μg/L for women (P<.0001).³² Analysis of serum ferritin levels of 3554 male patients from the third National Health and Nutrition Examination Survey demonstrated that patients who self-reported as non-Hispanic Black (n=1616) had significantly higher serum ferritin levels than non-Hispanic White patients (n=1938)(serum ferritin difference of 37.1 μg/L)(P<.0001).³³ The British Society for Haematology guidelines recommend that the threshold of serum ferritin for diagnosing iron deficiency should take into account age-, sex-, and race-based differences.³⁴ Ferritin and Hair—Cutaneous manifestations of iron deficiency include koilonychia, glossitis, pruritus, angular cheilitis, and telogen effluvium (TE).¹ A case-control study of 30 females aged 15 to 45 years demonstrated that the mean (SD) ferritin level was significantly lower in patients with TE than those with no hair loss (16.3 [12.6] ng/mL vs 60.3 [50.1] ng/mL; P<.0001). Using a threshold of 30 μg/L or lower, the investigators found that the odds ratio for TE was 21.0 (95% CI, 4.2-105.0) in patients with low serum ferritin.³⁵

Another retrospective review of 54 patients with diffuse hair loss and 55 controls compared serum vitamin B₁₂, folate, thyroid-stimulating hormone, zinc, ferritin, and 25-hydroxy vitamin D levels between the 2 groups.³⁶ Exclusion criteria were clinical diagnoses of female pattern hair loss (androgenetic alopecia), pregnancy, menopause, metabolic and endocrine disorders, hormone replacement therapy, chemotherapy, immunosuppressive therapy, vitamin and mineral supplementation, scarring alopecia, eating disorders, and restrictive diets. Compared with controls, patients with diffuse nonscarring hair loss were found to have significantly lower ferritin (mean [SD], 14.72 [10.70] ng/mL vs 25.30 [14.41] ng/mL; P<.001) and 25-hydroxy vitamin D levels (mean [SD], 14.03 [8.09] ng/mL vs 17.01 [8.59] ng/mL; P=.01).³⁶

In contrast, a separate case-control study of 381 cases and 76 controls found no increase in the rate of iron deficiency—defined as ferritin ≤15 μg/L or ≤40 μg/L—among women with female pattern hair loss or chronic TE vs controls.³⁷ Taken together, these studies suggest that iron status may play a role in TE, a process that may result from nutritional deficiency, trauma, or physical or psychological stress³⁸; however, there is insufficient evidence to suggest that low iron status impacts androgenetic alopecia, in which its multifactorial pathogenesis implicates genetic and hormonal factors.³⁹ More research is needed to clarify the potential associations between iron deficiency and types of hair loss. Additionally, it is unclear whether iron supplementation improves hair growth parameters such as density and caliber.⁴⁰

Low serum ferritin (<40 μg/L) with concurrent symptoms of iron deficiency, including fatigue, pallor, dyspnea on exertion, or hair loss, should prompt treatment with supplemental iron.^41-43 Generally, ferrous (Fe2⁺) salts are preferred to ferric (Fe3⁺) salts, as the former is more readily absorbed through the duodenal mucosa⁴⁴ and is the more common formulation in commercially available supplements in the United States.⁴⁵ Oral supplementation with ferrous sulfate 325 mg (65 mg elemental iron) tablets is the first-line therapy for iron deficiency anemia.¹ Alternatively, ferrous gluconate 324 mg (38 mg elemental iron) over-the-counter and its liquid form has demonstrated superior absorption compared to ferrous sulfate tablets in a clinical study with peritoneal dialysis patients.^1,46 One study suggested that oral iron 40 to 80 mg should be taken every other day to increase absorption.⁴⁷ Due to improved bioavailability, intravenous iron may be utilized in patients with malabsorption, renal failure, or intolerance to oral iron (including those with gastric ulcers or active inflammatory bowel disease), with the formulation chosen based on underlying comorbidities and potential risks.^1,48 The theoretical risk for potentiating bacterial growth by increasing the amount of unbound iron in the blood raises concerns of iron supplementation in patients with infection or sepsis. Although far from definitive, existing data suggest that risk for infection is greater with intravenous iron supplementation and should be carefully considered prior to use.^49,50Elevated Ferritin—Elevated ferritin may be difficult to interpret given the multitude of conditions that can cause it.^23,51,52 Elevated serum ferritin can be broadly characterized by increased synthesis due to iron overload, increased synthesis due to inflammation, or increased ferritin release from cellular damage.³⁴ Further complicating interpretation is the potential diurnal fluctuations in serum iron levels dependent on dietary intake and timing of laboratory evaluation, choice of assay, differences in reference standards, and variations in calibration procedures that can lead to analytic variability in the measurement of ferritin.^23,53,54

Among healthy patients, serum ferritin is directly proportional to iron status.^9,51 A study utilizing weekly phlebotomy of 22 healthy participants to measure serum ferritin and calculate mobilizable storage iron found a strong positive correlation between the 2 variables (r=0.83, P<.001), with each 1-μg/L increase of serum ferritin corresponding to approximately an 8-mg increase of storage iron; the initial serum ferritin values ranged from 2 to 83 μg/L in females and 36 to 224 μg/L in males.⁵⁵ The correlation of ferritin with iron status also was supported by the significant correlation between the number of transfusions received in patients with transfusion-related iron overload and serum ferritin levels (r=0.89, P<.001), with an average increase of 60 μg/L per transfusion.⁵¹

Clinical guidelines on the interpretation of serum ferritin levels by Cullis et al³⁴ recommend a normal upper limit of 200 μg/L for healthy females and 300 μg/L for healthy males. Outside of clinical syndromes associated with iron overload, Lee and Means⁵⁶ found that serum ferritin of 1000 μg/L or higher was a nonspecific marker of disease, including infection and/or neoplastic disorders. We have adapted these guidelines to propose a workflow for evaluation of serum ferritin levels (Figure). In patients with inflammatory conditions or those affected by metabolic syndrome, elevated serum ferritin does not correlate with body iron status.^57,58 It is believed that inflammatory cytokines, including tumor necrosis factor α and IL-1α, can upregulate ferritin synthesis independent of cellular iron stores.^15,16 Several studies have examined the relationship between insulin resistance and/or metabolic syndrome with serum ferritin levels.^31,32 Han et al³² found that elevated serum ferritin was significantly associated with higher risk for metabolic syndrome in men (P<.0001) but not in women.

Although cutaneous manifestations of iron overload can be seen as skin hyperpigmentation due to increased iron deposits and increased melanin production,²² the effects of elevated ferritin on the skin and hair are not well known. Iron overload is a known trigger of porphyria cutanea tarda (PCT),⁵⁹ a condition in which reduced or absent enzymatic activity of uroporphyrinogen decarboxylase (UROD) leads to build up of toxic porphyrins in various organs.⁶⁰ In the skin, PCT manifests as a blistering photosensitive eruption that may resolve as dyspigmentation, scarring, and milia.⁶¹ Phlebotomy is first-line therapy in PCT to reduce serum iron and subsequent formation of UROD inhibitors, with guidelines suggesting discontinuation of phlebotomy when serum ferritin levels reach 20 ng/mL or lower.⁶⁰ Hyperferritinemia (serum ferritin >500 μg/L) is a common finding in several inflammatory disorders often accompanied by clinically apparent cutaneous symptoms such as adult-onset Still disease,⁶² hemophagocytic lymphohistiocytosis,^63,64 and anti-melanoma differentiation-associated gene 5 dermatomyositis.⁶⁵ Among these conditions, serum ferritin levels have been reported to correlate with disease activity, raising the question of whether ferritin is a bystander or a driver of the underlying pathology.^62,66,67 However, rapid decline of serum ferritin levels with treatment and control of inflammatory cytokines suggest that ferritin is unlikely to contribute to pathology.^62,67

Final Thoughts

Many clinical studies have examined the association between hair health and body iron status, the collective findings of which suggest that iron deficiency may be associated with TE. Among commonly measured serum iron parameters, low ferritin is a highly specific and sensitive marker for diagnosing iron deficiency. Serum ferritin may be a clinically useful tool for ruling out underlying iron deficiency in patients presenting with hair loss. Despite advances in our understanding of the molecular mechanisms of ferritin synthesis and regulation, whether ferritin itself contributes to cutaneous pathology is poorly understood.^35,36,68-74 For patients who are otherwise healthy with low suspicion for inflammatory disorders, chronic systemic illnesses, or malignancy, serum ferritin can be used as an indicator of body iron status. The workup for slightly elevated serum ferritin should be interpreted in the context of other laboratory findings and should be reassessed over time. Serum ferritin levels above 1000 μg/L warrant further investigation into causes such as iron overload conditions and underlying inflammatory conditions or malignancy.

Ferritin is an iron storage protein crucial to human iron homeostasis. Because serum ferritin levels are in dynamic equilibrium with the body’s iron stores, ferritin often is measured as a reflection of iron status; however, ferritin also is an acute-phase reactant whose levels may be nonspecifically elevated in a wide range of inflammatory conditions. The various processes that alter serum ferritin levels complicate the clinical interpretation of this laboratory value. In this article, we review the structure and function of ferritin and provide a guide for clinical use.

Overview of Iron

Iron is an essential element of key biologic functions including DNA synthesis and repair, oxygen transport, and oxidative phosphorylation. The body’s iron stores are mainly derived from internal iron recycling following red blood cell breakdown, while 5% to 10% is supplied by dietary intake.^1-3 Iron metabolism is of particular importance in cells of the reticuloendothelial system (eg, spleen, liver, bone marrow), where excess iron must be appropriately sequestered and from which iron can be mobilized.⁴ Sufficient iron stores are necessary for proper cellular function and survival, as iron is a necessary component of hemoglobin for oxygen delivery, iron-sulfur clusters in electron transport, and enzyme cofactors in other cellular processes.

Although labile pools of biologically active free iron exist in limited amounts within cells, excess free iron can generate free radicals that damage cellular proteins, lipids, and nucleic acids.^5-7 As such, most intracellular iron is captured within ferritin molecules. The excretion of iron is unregulated and occurs through loss in sweat, menstruation, hair shedding, skin desquamation, and enterocyte turnover.⁸ The lack of regulated excretion highlights the need for a tightly regulated system of uptake, transportation, storage, and sequestration to maintain iron homeostasis.

Overview of Ferritin Structure and Function

Ferritin is a key regulator of iron homeostasis that also serves as an important clinical indicator of body iron status. Ferritin mainly is found as an intracellular cytosolic iron storage and detoxification protein structured as a hollow 24-subunit polymer shell that can sequester up to 4500 atoms of iron within its core.^9,10 The 24-mer is composed of both ferritin L (FTL) and ferritin H (FTH) subunits, with dynamic regulation of the H:L ratio dependent on the context and tissue in which ferritin is found.⁶

Ferritin H possesses ferroxidase, which facilitates oxidation of ferrous (Fe²⁺) iron into ferric (Fe³⁺) iron, which can then be incorporated into the mineral core of the ferritin heteropolymer.¹¹ Ferritin L is more abundant in the spleen and liver, while FTH is found predominantly in the heart and kidneys where the increased ferroxidase activity may confer an increased ability to oxidize Fe²⁺ and limit oxidative stress.⁶

Regulation of Ferritin Synthesis and Secretion

Ferritin synthesis is regulated by intracellular nonheme iron levels, governed mainly by an iron response element (IRE) and iron response protein (IRP) translational repression system. Both FTH and FTL messenger RNA (mRNA) contain an IRE that is a regulatory stem-loop structure in the 5´ untranslated region. When the IRE is bound by IRP1 or IRP2, mRNA translation of ferritin subunits is suppressed.⁶ In low iron conditions, IRPs have greater affinity for IRE, and binding suppresses ferritin translation.¹² In high iron conditions, IRPs have a decreased affinity for IRE, and ferritin mRNA synthesis is increased.¹³ Additionally, inflammatory cytokines such as tumor necrosis factor α and IL-1α transcriptionally induce FTH synthesis, resulting in an increased population of H-rich ferritins.^11,14-16 A study using cultured human primary skin fibroblasts demonstrated UV radiation–induced increases in free intracellular iron content.^17,18 Pourzand et al¹⁸ suggested that UV-mediated damage of lysosomal membranes results in leakage of lysosomal proteases into the cytosol, contributing to degradation of intracellular ferritin and subsequent release of iron within skin fibroblasts. The increased intracellular iron downregulates IRPs and increases ferritin mRNA synthesis,¹⁸ consistent with prior findings of increased ferritin synthesis in skin that is induced by UV radiation.¹⁹

Molecular analysis of serum ferritin in iron-overloaded mice revealed that extracellular ferritin found in the serum is composed of a greater fraction of FTL and has lower iron content than intracellular ferritin. The low iron content of serum ferritin compared with intracellular ferritin and transferrin suggests that serum ferritin is not a major pathway of systemic iron transport.¹⁰ However, locally secreted ferritins may play a greater role in iron transport and release in selected tissues. Additionally, in vitro studies of cell cultures from humans and mice have demonstrated the ability of macrophages to secrete ferritin, suggesting that macrophages are an important cellular source of serum ferritin.^10,20 As such, serum ferritin generally may reflect body iron status but more specifically reflects macrophage iron status.¹⁰ Although the exact pathways of ferritin release are unknown, it is hypothesized that ferritin secretion occurs through cytosolic autophagy followed by secretion of proteins from the lysosomal compartment.^10,18,21

Low Ferritin and Iron Deficiency—Although bone marrow biopsy with iron staining remains the gold standard for diagnosis of iron deficiency, serum ferritin is a much more accessible and less invasive tool for evaluation of iron status. A serum ferritin level below 12 μg/L is highly specific for iron depletion,²² with a higher cutoff recommended in clinical practice to improve diagnostic sensitivity.^23,24 Conditions independent of iron deficiency that may reduce serum ferritin include hypothyroidism and ascorbate deficiency, though neither condition has been reported to interfere with appropriate diagnosis of iron deficiency.²⁵ Guyatt et al²⁶ conducted a systematic review of laboratory tests used in the diagnosis of iron deficiency anemia and identified 55 studies suitable for inclusion. Based on an area under the receiver operating characteristic curve (AUROC) of 0.95, serum ferritin values were superior to transferrin saturation (AUROC, 0.74), red cell protoporphyrin (AUROC, 0.77), red cell volume distribution width (AUROC, 0.62), and mean cell volume (AUROC, 0.76) for diagnosis of IDA, verified by histologic examination of aspirated bone marrow.²⁶ The likelihood ratio of iron deficiency begins to decrease for serum ferritin values of 40 μg/L or greater. For patients with inflammatory conditions—patients with concomitant chronic renal failure, inflammatory disease, infection, rheumatoid arthritis, liver disease, inflammatory bowel disease, and malignancy—the likelihood of iron deficiency begins to decrease at serum ferritin levels of 70 μg/L or greater.²⁶ Similarly, the World Health Organization recommends that in adults with infection or inflammation, serum ferritin levels lower than 70 μg/L may be used to indicate iron deficiency.²⁴ A serum ferritin level of 41 μg/L or lower was found to have a sensitivity and specificity of 98% for discriminating between iron-deficiency anemia and anemia of chronic disease (diagnosed based on bone marrow biopsy with iron staining), with an AUROC of 0.98.²⁷ As such, we recommend using a serum ferritin level of 40 μg/L or lower in patients who are otherwise healthy as an indicator of iron deficiency.

The threshold for iron supplementation may vary based on age, sex, and race. In women, ferritin levels increase during menopause and peak after menopause; ferritin levels are higher in men than in women.^28-30 A multisite longitudinal cohort study of 70 women in the United States found that the mean (SD) ferritin valuewas 69.5 (81.7) μg/L premenopause and 128.8 (125.7) μg/L postmenopause (P<.01).³¹ A separate longitudinal survey study of 8564 patients in China found that the mean (SE) ferritin value was 201.55 (3.60) μg/L for men and 80.46 (1.64) μg/L for women (P<.0001).³² Analysis of serum ferritin levels of 3554 male patients from the third National Health and Nutrition Examination Survey demonstrated that patients who self-reported as non-Hispanic Black (n=1616) had significantly higher serum ferritin levels than non-Hispanic White patients (n=1938)(serum ferritin difference of 37.1 μg/L)(P<.0001).³³ The British Society for Haematology guidelines recommend that the threshold of serum ferritin for diagnosing iron deficiency should take into account age-, sex-, and race-based differences.³⁴ Ferritin and Hair—Cutaneous manifestations of iron deficiency include koilonychia, glossitis, pruritus, angular cheilitis, and telogen effluvium (TE).¹ A case-control study of 30 females aged 15 to 45 years demonstrated that the mean (SD) ferritin level was significantly lower in patients with TE than those with no hair loss (16.3 [12.6] ng/mL vs 60.3 [50.1] ng/mL; P<.0001). Using a threshold of 30 μg/L or lower, the investigators found that the odds ratio for TE was 21.0 (95% CI, 4.2-105.0) in patients with low serum ferritin.³⁵

Another retrospective review of 54 patients with diffuse hair loss and 55 controls compared serum vitamin B₁₂, folate, thyroid-stimulating hormone, zinc, ferritin, and 25-hydroxy vitamin D levels between the 2 groups.³⁶ Exclusion criteria were clinical diagnoses of female pattern hair loss (androgenetic alopecia), pregnancy, menopause, metabolic and endocrine disorders, hormone replacement therapy, chemotherapy, immunosuppressive therapy, vitamin and mineral supplementation, scarring alopecia, eating disorders, and restrictive diets. Compared with controls, patients with diffuse nonscarring hair loss were found to have significantly lower ferritin (mean [SD], 14.72 [10.70] ng/mL vs 25.30 [14.41] ng/mL; P<.001) and 25-hydroxy vitamin D levels (mean [SD], 14.03 [8.09] ng/mL vs 17.01 [8.59] ng/mL; P=.01).³⁶

In contrast, a separate case-control study of 381 cases and 76 controls found no increase in the rate of iron deficiency—defined as ferritin ≤15 μg/L or ≤40 μg/L—among women with female pattern hair loss or chronic TE vs controls.³⁷ Taken together, these studies suggest that iron status may play a role in TE, a process that may result from nutritional deficiency, trauma, or physical or psychological stress³⁸; however, there is insufficient evidence to suggest that low iron status impacts androgenetic alopecia, in which its multifactorial pathogenesis implicates genetic and hormonal factors.³⁹ More research is needed to clarify the potential associations between iron deficiency and types of hair loss. Additionally, it is unclear whether iron supplementation improves hair growth parameters such as density and caliber.⁴⁰

Low serum ferritin (<40 μg/L) with concurrent symptoms of iron deficiency, including fatigue, pallor, dyspnea on exertion, or hair loss, should prompt treatment with supplemental iron.^41-43 Generally, ferrous (Fe2⁺) salts are preferred to ferric (Fe3⁺) salts, as the former is more readily absorbed through the duodenal mucosa⁴⁴ and is the more common formulation in commercially available supplements in the United States.⁴⁵ Oral supplementation with ferrous sulfate 325 mg (65 mg elemental iron) tablets is the first-line therapy for iron deficiency anemia.¹ Alternatively, ferrous gluconate 324 mg (38 mg elemental iron) over-the-counter and its liquid form has demonstrated superior absorption compared to ferrous sulfate tablets in a clinical study with peritoneal dialysis patients.^1,46 One study suggested that oral iron 40 to 80 mg should be taken every other day to increase absorption.⁴⁷ Due to improved bioavailability, intravenous iron may be utilized in patients with malabsorption, renal failure, or intolerance to oral iron (including those with gastric ulcers or active inflammatory bowel disease), with the formulation chosen based on underlying comorbidities and potential risks.^1,48 The theoretical risk for potentiating bacterial growth by increasing the amount of unbound iron in the blood raises concerns of iron supplementation in patients with infection or sepsis. Although far from definitive, existing data suggest that risk for infection is greater with intravenous iron supplementation and should be carefully considered prior to use.^49,50Elevated Ferritin—Elevated ferritin may be difficult to interpret given the multitude of conditions that can cause it.^23,51,52 Elevated serum ferritin can be broadly characterized by increased synthesis due to iron overload, increased synthesis due to inflammation, or increased ferritin release from cellular damage.³⁴ Further complicating interpretation is the potential diurnal fluctuations in serum iron levels dependent on dietary intake and timing of laboratory evaluation, choice of assay, differences in reference standards, and variations in calibration procedures that can lead to analytic variability in the measurement of ferritin.^23,53,54

Among healthy patients, serum ferritin is directly proportional to iron status.^9,51 A study utilizing weekly phlebotomy of 22 healthy participants to measure serum ferritin and calculate mobilizable storage iron found a strong positive correlation between the 2 variables (r=0.83, P<.001), with each 1-μg/L increase of serum ferritin corresponding to approximately an 8-mg increase of storage iron; the initial serum ferritin values ranged from 2 to 83 μg/L in females and 36 to 224 μg/L in males.⁵⁵ The correlation of ferritin with iron status also was supported by the significant correlation between the number of transfusions received in patients with transfusion-related iron overload and serum ferritin levels (r=0.89, P<.001), with an average increase of 60 μg/L per transfusion.⁵¹

Clinical guidelines on the interpretation of serum ferritin levels by Cullis et al³⁴ recommend a normal upper limit of 200 μg/L for healthy females and 300 μg/L for healthy males. Outside of clinical syndromes associated with iron overload, Lee and Means⁵⁶ found that serum ferritin of 1000 μg/L or higher was a nonspecific marker of disease, including infection and/or neoplastic disorders. We have adapted these guidelines to propose a workflow for evaluation of serum ferritin levels (Figure). In patients with inflammatory conditions or those affected by metabolic syndrome, elevated serum ferritin does not correlate with body iron status.^57,58 It is believed that inflammatory cytokines, including tumor necrosis factor α and IL-1α, can upregulate ferritin synthesis independent of cellular iron stores.^15,16 Several studies have examined the relationship between insulin resistance and/or metabolic syndrome with serum ferritin levels.^31,32 Han et al³² found that elevated serum ferritin was significantly associated with higher risk for metabolic syndrome in men (P<.0001) but not in women.

Although cutaneous manifestations of iron overload can be seen as skin hyperpigmentation due to increased iron deposits and increased melanin production,²² the effects of elevated ferritin on the skin and hair are not well known. Iron overload is a known trigger of porphyria cutanea tarda (PCT),⁵⁹ a condition in which reduced or absent enzymatic activity of uroporphyrinogen decarboxylase (UROD) leads to build up of toxic porphyrins in various organs.⁶⁰ In the skin, PCT manifests as a blistering photosensitive eruption that may resolve as dyspigmentation, scarring, and milia.⁶¹ Phlebotomy is first-line therapy in PCT to reduce serum iron and subsequent formation of UROD inhibitors, with guidelines suggesting discontinuation of phlebotomy when serum ferritin levels reach 20 ng/mL or lower.⁶⁰ Hyperferritinemia (serum ferritin >500 μg/L) is a common finding in several inflammatory disorders often accompanied by clinically apparent cutaneous symptoms such as adult-onset Still disease,⁶² hemophagocytic lymphohistiocytosis,^63,64 and anti-melanoma differentiation-associated gene 5 dermatomyositis.⁶⁵ Among these conditions, serum ferritin levels have been reported to correlate with disease activity, raising the question of whether ferritin is a bystander or a driver of the underlying pathology.^62,66,67 However, rapid decline of serum ferritin levels with treatment and control of inflammatory cytokines suggest that ferritin is unlikely to contribute to pathology.^62,67

Final Thoughts

Many clinical studies have examined the association between hair health and body iron status, the collective findings of which suggest that iron deficiency may be associated with TE. Among commonly measured serum iron parameters, low ferritin is a highly specific and sensitive marker for diagnosing iron deficiency. Serum ferritin may be a clinically useful tool for ruling out underlying iron deficiency in patients presenting with hair loss. Despite advances in our understanding of the molecular mechanisms of ferritin synthesis and regulation, whether ferritin itself contributes to cutaneous pathology is poorly understood.^35,36,68-74 For patients who are otherwise healthy with low suspicion for inflammatory disorders, chronic systemic illnesses, or malignancy, serum ferritin can be used as an indicator of body iron status. The workup for slightly elevated serum ferritin should be interpreted in the context of other laboratory findings and should be reassessed over time. Serum ferritin levels above 1000 μg/L warrant further investigation into causes such as iron overload conditions and underlying inflammatory conditions or malignancy.