Direct‐acting oral anticoagulants and alopecia: The valuable support of postmarketing data

Abstract

Little is known about the administration of direct‐acting oral anticoagulants (DOACs) and the occurrence of alopecia. Our aim was to analyse the reports of alopecia following DOAC administration received until 2 May 2018 from VigiBase, the World Health Organization database. A descriptive analysis of age, sex, seriousness and dechallenge/rechallenge outcome was carried out. For each report, the time‐to‐onset was evaluated and the causality was assessed by using Naranjo algorithm. Overall, 1316 reports were retrieved, most concerning rivaroxaban (58.8%); 80% of the reports were related to females, in particular to those aged ≥65 years (23.1%). The median value of the time‐to‐onset was 28 days, with an interquartile range of 63 days. In 54.3% of the reports the causality was assessed as possible. In conclusion, a possible association could exist between DOACs administration and alopecia, but further observational studies are needed to confirm these findings.

What is already known about this subject

• Alopecia is a disorder that can occur during drug therapy.

• Anticoagulants, especially warfarin and enoxaparin, have been shown to induce alopecia.

• Little is known about the use of direct‐acting oral anticoagulants (DOACs: dabigatran, rivaroxaban, apixaban, edoxaban) and alopecia.

What this study adds

• Several reports of DOAC‐induced alopecia have been retrieved in VigiBase.

• Our data suggest that a possible association could exist between DOACs and alopecia that should be investigated thoroughly.

• If this association is confirmed, an update of the summaries of products characteristics would be necessary.

1. BACKGROUND

Alopecia is a disorder characterized by hair loss that can occur in a variety of clinical conditions, including drug therapy.1 Several drugs have been associated with alopecia and this adverse drug reaction (ADR) is listed in the summary of product characteristics (SPC) for some of them. Drugs known to cause alopecia include antineoplastic agents, for which hair loss is 1 of the most common side effects especially in patients administered with combined therapy.1 Chemotherapy kills cells that grow and spread quickly as cancer cells but can also cause the degeneration of the hair‐producing cells.2 Others drugs causing hair loss are antiepileptics (e.g. valproic acid, which induces a dose‐dependent and reversible alopecia),3 oral contraceptives4 and protease inhibitors in HIV‐therapy (e.g. ritonavir plus lopinavir therapy).5, 6

Anticoagulants, especially warfarin and enoxaparin, have also been shown to induce alopecia, and this ADR is reported both in literature7, 8 and in the SPCs.9, 10 For both drugs, alopecia was detected thanks to the postmarketing evaluations.

In recent years, direct acting oral anticoagulants (DOACs) have entered the market and little is available in literature about the possible association with alopecia/hair loss. Considering the several noninferiority studies of DOACs compared to warfarin in the treatment of nonvalvular atrial fibrillation11, 12, 13 and different guidelines on anticoagulant therapies management,14, 15 DOACs could be identified as an elective therapy in patients who experienced alopecia during or after the treatment with warfarin. Indeed, the European Medicine Agency focused on the possible association between dabigatran and alopecia asking for an assessment in the periodic safety update report16 in order to collect supplementary information.

The purpose of this study was to analyse all the reports of suspected alopecia following DOACs administration using the World Health Organization (WHO) database VigiBase.

VigiBase is the global database of individual case safety reports that gathers all the reports of suspected ADRs from the member countries of the WHO Program for International Drug Monitoring since 1968. The ADRs are coded using the MedDRA dictionary that enable structured data entry at different levels of accuracy. This dictionary has a hierarchic terminology: the high‐level MedDRA terms (HLTs) comprise several preferred terms (PTs), that include lowest‐level terms in turn. This approach allowed us to collect up‐to‐date information about this ADR associated with the use of direct acting oral anticoagulants.

2. METHODS

We collected the reports of alopecia in VigiBase using the HLT alopecias, which includes the following PTs: alopecia, alopecia areata, alopecia scarring, alopecia syphilitic, alopecia totalis, alopecia universalis, hypotrichosis, progeria, madarosis, follicular mucinosis, application site alopecia, androgenetic alopecia, Satoyoshi syndrome, radiation alopecia, diffuse alopecia. From these PTs, we deleted those related to specific conditions that could have other causes, such as progeria, alopecia syphilitic, follicular mucinosis, androgenetic alopecia, Satoyoshi syndrome and radiation alopecia. We selected all the reports received until 2 May 2018 concerning any of the DOACs as suspected or interacting drug and any of the above mentioned reactions. To detect and exclude possible duplicates, we performed an analysis using a record‐linkage strategy by grouping the overlapping records in 4 key fields: report_id (i.e. the number identifying each case report), preferred‐base‐name, MedDRA PtCode and OnsetDate. After exclusion of the duplicates, we performed a descriptive analysis evaluating age and sex of the patients, reports' seriousness, reporting country, MedDRA PTs and concomitants medications. We also calculated the time elapsed between the date of administration and the appearance of the adverse reaction. Lastly, dechallenge and rechallenge evaluation with a focus on cases that had a positive rechallenge was performed. Different algorithms for causality assessment are used around the world and this evaluation is often not made for all the reports. A case‐by‐case evaluation of each report using the Naranjo algorithm17 was performed. Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY.

3. RESULTS

By using the above described MedDRA PTs, we retrieved a total of 1397 reports as related to dabigatran, rivaroxaban, apixaban or edoxaban. Out of the total, 810 concerned rivaroxaban as suspected/interacting drug, 305 apixaban, 264 dabigatran and 18 edoxaban. After the exclusion of duplicates, 1316 cases were analysed. The most reported drug was rivaroxaban (774 cases, 58.8%), followed by apixaban (290, 22.0%), dabigatran (235, 17.9%) and edoxaban (17, 1.3%). Table 1 shows the distribution of the cases by age and sex of the patients for each drug. Overall, about the 80% of the reports were related to females, in particular, those aged ≥65 years (23.1%). For rivaroxaban, 1 case was in the age group 0–23 months, whereas rivaroxaban and apixaban had 1 report each in 2–17‐year‐old subjects.

Table 1.

Population characteristics related to age groups and sex for each drug

AgeClass	Dabigatran							Rivaroxaban							Apixaban							Edoxaban
	M	%*	F	%*	NA	%*	TOT	M	%*	F	%*	NA	%*	TOT	M	%*	F	%*	NA	%*	TOT	M	%*	F	%*	NA	%*	TOT
0–23 mo (infants/newborns)	0	0	0	0	0	0	0	1	0.1	0	0.0	0	0.0	1	0	0.0	0	0.0	0	0.0	0	0	0.0	0	0.0	0	0.0	0
2–17 y (children/teens)	0	0	0	0	0	0	0	0	0.0	1	0.1	0	0.0	1	0	0.0	0	0.0	1	0.3	1	0	0.0	0	0.0	0	0.0	0
18–64 y (adults)	7	3.0	18	7.7	0	0	25	25	3.2	148	19.1	0	0.0	173	5	1.7	29	10.0	1	0.3	35	0	0.0	3	17.6	0	0.0	3
≥65 y (elderly)	12	5.1	81	34.5	0	0	93	30	3.9	207	26.7	0	0.0	237	15	5.2	78	26.9	1	0.3	94	2	11.8	4	23.5	0	0.0	6
Not available	15	6.4	92	39.1	10	4.3	117	31	4.0	253	32.7	78	10.1	362	19	6.6	130	44.8	11	3.8	160	0	0.0	7	41.2	1	5.9	8
Total	34	14.5	191	81.3	10	4.3	235	87	11.2	609	78.7	78	10.1	774	39	13.4	237	81.7	14	4.8	290	2	11.8	14	82.4	1	5.9	17

^* Percentages calculated for each drug separately depending on age and sex.

M = male; F = female; NA = not available. TOT = total

Most of the reports were from the USA (n = 944, 71.7%) and Germany (112, 8.5%). Other reporting countries were UK (n = 60, 4.6%), Sweden (28, 2.1%), Netherlands (25, 1.9%), France (23, 1.7%), Switzerland (18, 1.4%), Italy (16, 1.2%), Austria (15, 1.1%), Norway (11, 0.8%) and Canada (10, 0.8%), overall accounting for >95% of the reports. Within the MedDRA HLT alopecia, the top 2 reported lowest‐level terms were alopecia (n = 1046, 79.5%) and hair loss (229, 17.4%), corresponding to about the 97% of all the ADRs reported. Most reported concomitants drugs were omeprazole, furosemide, acetylsalicylic acid, levothyroxine sodium and digoxin. Alopecia is listed in the corresponding SPC only for omeprazole. Overall 79.2% of the cases were reported as not serious, 19.8% were serious and seriousness was not available for only 1% of the reports. Based on the WHO seriousness criteria, 49 out of the 260 serious cases caused or prolonged hospitalization. Of the 260 serious cases, 170 (68.8%) reported additional ADRs besides alopecia, such as: contusion, dyspnoea, arthralgia, asthenia, fatigue, dizziness and epistaxis. Only 77 cases reported alopecia as ADR exclusively. As for the case‐by‐case causality assessment (Table 2), 54.3% of the reports were assessed as possible according to Naranjo algorithm, 14.7% as probable and 3.3% as doubtful.

Table 2.

Assessment of the causality case‐by‐case using Naranjo algorithm

Causality Naranjo algorithm	Dabigatran		Rivaroxaban		Apixaban		Edoxaban		Total
Doubtful	18	7.7%	21	2.7%	4	1.4%	0	0.0%	43	3.3%
Possible	76	32.3%	476	61.5%	153	52.7%	10	58.8%	715	54.3%
Probable	45	19.2%	88	11.4%	57	19.6%	4	23.5%	194	14.7%
Unclassifiable	96	40.8%	189	24.4%	76	26.3%	3	17.7%	364	27.7%
Total	235	100.0%	774	100.0%	290	100.0%	17	100.0%	1316	100.0%

It was possible in 243 reports to calculate the time interval between the date of administration and the appearance of the adverse reaction. The time‐to‐onset had a median of 28 days, with an interquartile range of 63 days.

An analysis of the dechallenge and rechallenge outcome was also performed (Table 3). The most important information comes from the analysis of the reports where the dechallenge action was drug withdrawn. For rivaroxaban and apixaban, in 59 reports (7.6%) and 23 (7.9%) respectively, withdrawal of the drugs was followed by the abatement of the ADR.

Table 3.

Outcomes of the dechallenge (i.e. action taken after the ADR) and the rechallenge (i.e. re‐administration of the drug after the ADR)

Dechallenge	Outcome	Dabigatran	%	Rivaroxaban	%	Apixaban	%	Edoxaban	%
Dose not changed	Reaction abated	3	1.3	13	1.7	3	1.0	0	0.0
	Effect unknown	14	6.0	21	2.7	16	5.5	1	5.9
	No effect observed	109	46.4	283	36.6	9	3.1	4	23.5
Dose reduced	Reaction abated	0	0.0	1	0.1	1	0.3	0	0.0
	No effect observed	6	2.6	9	1.2	1	0.3	0	0.0
	Effect unknown	1	0.4	4	0.5	4	1.4	0	0.0
Drug withdrawn	Reaction abated	13	5.5	59	7.6	23	7.9	2	11.8
	No effect observed	18	7.7	47	6.1	13	4.5	2	11.8
	Effect unknown	13	5.5	46	5.9	23	7.9	3	17.6
Not available	Not available	58	24.7	291	37.6	197	67.9	5	29.4
All	Total	235	100.0	774	100.0	290	100.0	17	100.0

Rechallenge	Outcome	Dabigatran	%	Rivaroxaban	%	Apixaban	%	Edoxaban	%
Readministration	Reaction recurred	2	0.8	2	0.3	0	0	0	0
	No recurrence	0	0.0	338	43.6	1	0.3	0	0
	Effect unknown	11	4.7	254	32.8	142	49.0	1	5.9
Not available	Not available	222	94.5	180	23.3	147	50.7	16	94.1
All	Total	235	100.0	774	100.0	290	100.0	17	100.0

Conversely, for dabigatran the largest percentage showed no effect after drug withdrawal. No significant information is available for edoxaban. Only 4 cases had a positive rechallenge, 2 with dabigatran and 2 with rivaroxaban. In 43.6% of rivaroxaban re‐administration, the reaction did not recur. For 42.9% of the reports, no information was available about the rechallenge. All 4 were female patients and for only 3 was the age reported: 38, 59 and 77 years.

4. DISCUSSION

A great number of reports concerning alopecia related to direct acting oral anticoagulants administration are available in VigiBase, although this ADR is not acknowledged in the corresponding SPCs. Rivaroxaban is the most reported drug, followed by apixaban, dabigatran and edoxaban. The numerical difference of reports between DOACs of the same class (rivaroxaban, apixaban and edoxaban), may be due to their different marketing period. Rivaroxaban received the first European marketing authorization in 2008, while apixaban was authorized in 2011.

There are still few articles describing this ADR in the literature but some have focused the attention on the possible association between DOACs and alopecia.18, 19 Gelbricht et al.19 evaluated the incidence and the risk profile of hair loss in patients receiving dabigatran or rivaroxaban. The total incidence was 4.4 per 100 patient‐years of newly reported hair loss cases using the DOAC registry of the district of Saxony, in Germany.18

To provide a general description of the drug–reaction pair DOAC–alopecia, it is important to focus on the sex difference. In our survey, the majority of the patients were female (80%) as was also the case in the study of Gelbricht et al.19 In general, women appear to be more affected by alopecia20 due to different physiological conditions. Moreover, the most reported age group appears to be superimposable with the major oral anticoagulants users21 and with the prevalence of atrial fibrillation.22 By contrast, considering the same age group in males, most of them may be affected by androgenetic alopecia. The hair loss in men before beginning of anticoagulants therapy makes it difficult to detect a possible association between the therapy and alopecia.

About 80% of the cases in our analysis were reported as not serious. Although alopecia is often referred as a nonserious ADR, its occurrence should be taken into account by the doctor when prescribing the anticoagulant therapy. In fact, although nonserious based on WHO's seriousness criteria, this ADR may represent a troubling event for the patient.

Out of the total, 49 of the serious cases caused or prolonged hospitalization but only the 5.8% of them reported alopecia exclusively as ADR.

While a report may contain multiple ADRs, seriousness can be assigned only once for that report. This explains the difference in percentage between the serious reports with more ADRs and those reporting only alopecia.

Identifying alopecia as an ADR may be difficult. Basing on the moment in which hair starts falling out, alopecia can be described as anagen or telogen effluvium.23 Anagen effluvium is a condition in which people have their hair falling out during the growing phase of the hair growth cycle. Conversely, telogen effluvium consists of hair loss during the resting phase. As suggested by Watras et al., telogen effluvium may be the cause of the ADR associated with anticoagulants,24 recognizing the telogen effluvium may be complicated because this effect generally appears from 6 weeks to 3 months after treatment. Also, in the reports analysed in our research, alopecia appears with a median value of 28 days and interquartile range of 63 days, similar to the data reported for telogen effluvium. During this time, the patient could take other medicinal products, stop the anticoagulant therapy, or suffer from psychophysical stress or other pathological conditions that will lead to an impossibility to determine the real cause of alopecia and the possible causal association with the anticoagulant.

In their conclusion, Watras et al. reported that Naranjo algorithm could help to confirm the causality. Our case‐by‐case assessment of the causality shows that 54.3% of the reports were possible according to Naranjo algorithm. For 14.7%, the association resulted as probable and only for the 3.3% as doubtful. Another important information that could be helpful in establishing the causal association in the drug–reaction pairs considered is the dechellange outcome (i.e. the outcome after the drug dose reduction, or after the discontinuation or withdrawal of the drug after the ADR) and the rechallenge (i.e. the readministration of the drug after the withdrawal). In 7.9 and 7.6% of the cases, the withdrawal of apixaban and rivaroxaban resulted in ADR abatement, respectively. Only 4 cases had a positive rechallenge after an abatement of the ADR after the drug withdrawal.

Further data from larger cohort study or clinical trial are needed to evaluate potential risk factors for alopecia with the novel oral anticoagulants.

Given that warfarin‐ and heparin‐induced alopecia is a known ADR, DOACs could be considered as elective therapy for patients who have experienced alopecia. However, our data suggest that more information is needed to better define the outcome of this possible therapeutic shift. In VigiBase, we retrieved a total of 999 reports (duplicates included) related to warfarin and alopecia, a number comparable to those found for all the DOACs.

Finally, it should be considered that, in the WHO Pharmaceuticals newsletter 2016 No. 4,25 the possible risk of association between DOACs and alopecia was also reported by the Medicines and Medical Devices Safety Authority (Medsafe) of New Zealand as well as focused by the Pharmacovigilance Risk Assessment Committee.16

Overall, our data suggest that there might be a possible association between DOACs and alopecia. Considering the clinician's perceptions about DOACs superior safety vs warfarin26 and given the preference of patients towards the treatment with DOACs,27 all information regarding their safety profile must be monitored and updated where necessary. Spontaneous reporting allows retrieval of real‐life data regarding the safety of medicinal products without the restricted inclusion criteria of the clinical trials.

However, it is important to remember that spontaneous reporting has several limitations to consider for the interpretation of the results. First of all, the quality of the data; the information reported might be not in‐depth enough to allow a correct and complete comprehension of the case. Moreover, for drugs of recent marketing there is usually an increase in spontaneous reporting (Weber effect)28 but, in pharmacovigilance, under‐reporting or selective reporting may also represent possible biases. Lastly, there are many possible confounders to be considered such as concomitant therapies, predisposing conditions and other risk factors. In the reports analysed, the most frequent concomitant medications were: omeprazole, furosemide, acetylsalicylic acid, levothyroxine sodium and digoxin.

Alopecia is listed in the corresponding SPC only for omeprazole. However, some disease conditions requiring such drugs could cause alopecia, such as hypothyroidism, for which levothyroxine is used. Beyond the limitations, greater attention must be given to consider alopecia as a possible ADR.

These data add information to the few already present in the literature and help to better outline the potential association for the drug‐reaction pairs considered. A possible SPC update may be required for these drugs.

5. CONCLUSION

Our data provide information regarding a topic that has not yet been clarified. Although already considered, our analysis supports the possible association between the direct acting oral anticoagulants and alopecia with additional data. Further observational studies focusing on the evaluation of alopecia as a possible DOAC adverse reactions are needed to confirm this.

COMPETING INTERESTS

All the authors declare that they have no conflict of interest. This manuscript has no funding source.

CONTRIBUTORS

Substantial contributions to conception or design of the work (G.B., A.V., D.M.),or the acquisition (G.B., D.M.), analysis (G.B., D.M.) or interpretation of data for the work (G.B., A.V., D.M.). Drafting of the work (G.B., A.V., D.M.) or revising it critically for important intellectual content (A.V., D.M.). All authors approved the submitted final version to be published. All authors agree to be accountable for all aspects of the work in ensuring thatquestions related to the accuracy or integrity of any part of the work areappropriately investigated and resolved.

ETHICAL APPROVAL

The manuscript does not contain clinical studies or patient data. For this type of study, formal consent is not required.

Bonaldo G, Vaccheri A, Motola D. Direct‐acting oral anticoagulants and alopecia: The valuable support of postmarketing data. Br J Clin Pharmacol. 2020;86:1654–1660. 10.1111/bcp.14221

DATA AVAILABILITY STATEMENT

The datasets generated during and analysed during the current study are available from the corresponding author on reasonable request.