Janus Kinase Inhibitors and Adverse Events of Acne

Key Points

Question

Is there an association between Janus kinase (JAK) inhibitor use and acne?

Findings

In this systematic review and meta-analysis of 25 unique phase 2 and 3 randomized clinical trials with 10 839 participants, JAK inhibitors were associated with a 3.83 times higher odds of acne compared with placebo. This association was stronger among those studies conducted for dermatologic indications as well as among JAK1-specific, combined JAK1 and JAK2, and tyrosine kinase 2–specific inhibitor treatments.

Meaning

These findings suggest that patients should be properly counseled on the potential adverse effect of acne before beginning treatment with JAK inhibitors.

This systematic review and meta-analysis analyzes the risk of acne associated with Janus kinase (JAK) inhibitors across published phase 2 and 3 placebo-controlled randomized clinical trials of JAK inhibitors.

Abstract

Importance

Janus kinase (JAK) inhibitors are increasingly used across a range of dermatologic conditions. Adverse events of acne have been noted in some studies in clinical practice, but the scope of this outcome across JAK inhibitors has not been established.

Objective

To systematically analyze all published phase 2 and 3 placebo-controlled randomized clinical trials (RCTs) of JAK inhibitors for the risk of acne as an adverse effect of these medications.

Data Sources

Comprehensive search of Ovid MEDLINE and PubMed databases through January 31, 2023.

Study Selection

Inclusion criteria were phase 2 and 3 placebo-controlled RCTs of JAK inhibitors published in English with reported adverse events of acne.

Data Extraction and Synthesis

Two reviewers independently reviewed and extracted information from all included studies.

Main Outcomes and Measures

The primary outcome of interest was the incidence of acne following JAK inhibitor use. A meta-analysis was conducted using random-effects models.

Results

A total of 25 unique studies (10 839 unique participants; 54% male and 46% female) were included in the final analysis. The pooled odds ratio (OR) was calculated to be 3.83 (95% CI, 2.76-5.32) with increased ORs for abrocitinib (13.47 [95% CI, 3.25-55.91]), baricitinib (4.96 [95% CI, 2.52-9.78]), upadacitinib (4.79 [95% CI, 3.61-6.37]), deucravacitinib (2.64 [95% CI, 1.44-4.86]), and deuruxolitinib (3.30 [95% CI, 1.22-8.93]). Estimated ORs were higher across studies investigating the use of JAK inhibitors for the management of dermatologic compared with nondermatologic conditions (4.67 [95% CI, 3.10-7.05]) as well as for JAK1-specific inhibitors (4.69 [95% CI, 3.56-6.18]), combined JAK1 and JAK2 inhibitors (3.43 [95% CI, 2.14-5.49]), and tyrosine kinase 2 inhibitors (2.64 [95% CI, 1.44-4.86]).

Conclusions and Relevance

In this systematic review and meta-analysis, JAK inhibitor use was associated with an elevated odds of acne. Patients should be properly counseled on this potential adverse effect of these medications before treatment initiation. Future studies are needed to further elucidate the pathophysiology of this association.

Introduction

The Janus kinase signal transducer and activator of transcription (JAK-STAT) is a key inflammatory pathway in a number of dermatologic, hematologic, and rheumatologic conditions,^1,2,3,4,5 and JAK inhibitors are an emerging treatment option for these disorders.^{6,7,8,9,10,11,12,13} There are 4 recognized JAK proteins—JAK1, JAK2, JAK3, and tryosine kinase 2 (TYK2)—and 8 medications approved by the US Food and Drug Administration that inhibit this pathway. Baricitinib and ruxolitinib phosphate target JAK1 and JAK2; fedratinib hydrochloride and pacritinib citrate target JAK2; tofacitinib citrate targets JAK1, JAK2, and JAK3; upadacitinib and abrocitinib target only JAK1; and deucravacitinib targets TYK2.¹⁴

While JAK inhibitors hold promise as a valuable treatment option for patients with inflammatory conditions, acne has been a commonly reported adverse effect in clinical trials of JAK inhibitors.¹⁵ However, little is known about the overall incidence of acne from JAK inhibitors and how this differs between different drug classes and the underlying conditions being treated.¹⁶ In this systematic review and meta-analysis, we analyzed the risk of acne associated with JAK inhibitors across published phase 2 and 3 placebo-controlled randomized clinical trials (RCTs) of JAK inhibitors.

Methods

This systematic review and meta-analysis was reported in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) reporting guideline for RCTs. The methodological quality of this systematic review was assessed using the Measurement Tool to Assess systematic Reviews (AMSTAR 2) (eTable 4 in Supplement 1). The population targeted consisted of patients in clinical trials receiving JAK inhibitors. The comparator group was the placebo population in those same trials, and the outcome of interest was new-onset acne.

A literature search of Ovid MEDLINE and PubMed for placebo-controlled RCTs involving JAK inhibitors was conducted through June 22, 2022, and then expanded to include all dates up to and including January 31, 2023 (eTable 1 in Supplement 1). Search terms were developed with the assistance of a librarian to include all RCTs involving any combination of Janus kinase, JAK, antagonist, inhibitor, block, ruxolitinib, tofacitinib, oclacitinib, baricitinib, peficitinib, fedratinib, upadacitinib, filgotinib, delgocitinib, abrocitinib, cerdulatinib, gandotinib, lestaurtinib, momelotinib, pacritinib, deucravacitinib, or cucurbitacin. Only studies published in English were included. Duplicates were removed using the online service Covidence. All methods, including the question, search strategy, inclusion and exclusion criteria, and risk of bias assessment were established prior to the conduct of the review.

Three reviewers (J.M., C.M., and P.M.) independently screened each study title and abstract for compliance with predefined exclusion and inclusion criteria. To be considered for full text review, studies must have been phase 2 or 3 placebo-controlled RCTs of JAK inhibitors. Studies published in a language other than English or without data on adverse events of acne were excluded. Full texts were then analyzed by 2 reviewers (J.M. and C.M.), and duplicate studies that had passed initial screening were manually removed. Any discrepancies were resolved by discussion with the principal investigator (A.M.).

Studies with multiple published interim reports were manually identified, and only data from the most recent report were used. In the case of studies that transitioned to an open-label phase, only data from the blinded phases were used. Data on JAK inhibitor type, disease indication, patient characteristics, and incidence of acne were extracted from included studies. Heterogeneity was assessed by I² values and was accounted for through use of random-effects model estimators. A Cochrane bias risk assessment was performed according to the most updated guidelines and is included in eTable 2 in Supplement 1.¹⁷ In addition, we performed planned subgroup analysis stratified by JAK inhibitor type and indication (dermatologic vs nondermatologic). When possible, data were quantitatively synthesized using random-effects meta-analysis. Analyses were performed using R, version 4 (R Project for Statistical Computing). Statistical significance was indicated by a 2-sided P < .05.

Results

Of the 4427 identified articles, 25 were ultimately suitable for data extraction (Figure 1),^{18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41} including a publication describing 2 studies. Characteristics of the included trials are summarized in the Table. All trials were funded by the pharmaceutical industry with the exception of 2 studies by Nakagawa et al,^24,25 which were funded by Japan Tobacco Inc. The cumulative sample size was 10 839 participants (46% female and 54% male), of whom 7572 received treatment with a JAK inhibitor. The Egger test for publication bias produced P = .20, suggesting no evidence of publication bias. A funnel plot can be found in the eFigure in Supplement 1.

Figure 1. Study Flow Diagram.

Table. Study Characteristics^a.

Source	Drug	Dosage	Indication	Treatment group		Control group		Age, y	Female, %
				No. of participants	% With acne	No. of participants	% With acne
Bieber et al,18 2021	Abrocitinib	100 mg/d; 200 mg/d	Atopic dermatitis	464	4.7	131	0	37.9 (14.8)b	49.4
Eichenfield et al,19 2021	Abrocitinib	100 mg/d; 200 mg/d	Atopic dermatitis	189	3.7	96	1.0	15 (13-17)c	49.1
Silverberg et al,20 2020	Abrocitinib	100 mg/d; 200 mg/d	Atopic dermatitis	313	3.5	78	0	35.1 (15.1)b	41.4
Pooled proportion of treatment group with acne, 4.2%
King et al,21 2021	Baricitinib	1 mg/d; 2 mg/d; 4 mg/d	Alopecia areata	54	9.3	28	0	41 (13.6)b	74.5
King et al,22 2022	Baricitinib	2 mg/d; 4 mg/d	Alopecia areata	852	5.4	343	2.3	37.5 (12.9)b	60.6
Reich et al,23 2020	Baricitinib	2 mg/d; 4 mg/d	Atopic dermatitis	220	2.3	108	0.9	33.8(12.4)b	34
Pooled proportion of treatment group with acne, 5.0%
Nakagawa et al,24 2018	Delgocitinib	0.25%, 0.5%, 1%, 3% Ointment twice daily	Atopic dermatitis	265	2.6	32	0	30.6 (9.8)b	33.6
Nakagawa et al,25 2021	Delgocitinib	0.25% Ointment twice daily	Atopic dermatitis	69	1.4	68	1.5	8.3 (3.8)b	48.9
Pooled proportion of treatment group with acne, 2.3%
Morand et al,26 2023	Deucravacitinib	3 mg twice daily; 6 mg twice daily; 12 mg/d	Systemic lupus erythematosus	273	6.6	90	4.4	40.1 (12)b	92
Armstrong et al,27 2023	Deucravacitinib	6 mg/d	Psoriasis	531	2.8	165	0	46.1 (13.4)b	31.1
Papp et al,28 2018	Deucravacitinib	3 mg/d; 3 mg/d; 3 mg/d; 6 mg twice daily; 12 mg twice daily	Psoriasis	222	3.6	45	0	45 (13)b	27
Strober et al,29 2023	Deucravacitinib	6 mg/d	Psoriasis	833	1.6	501	0.2	40.4 (13.47)b	31.6
Pooled proportion of treatment group with acne, 2.9%
King et al,30 2022	Deuruxolitinib	4 mg twice daily; 8 mg twice daily; 12 mg twice daily	Alopecia areata	103	13.6	44	4.5	36.8 (12.85)b	70.5
Pooled proportion of treatment group with acne, 13.6%
Papp et al,31 2015	Peficitinib	10 mg twice daily; 25 mg twice daily; 50 mg/d; 60 mg twice daily; 100 mg twice daily	Psoriasis	95	3.2	29	0	48.1 (12.7)b	21.8
Pooled proportion of treatment group with acne, 3.2%
Robinson et al,32 2020	Ritlecitinib	200 mg/d	Rheumatoid arthritis	42	2.4	28	0	54.9 (11.7)b	81.4
Pooled proportion of treatment group with acne, 2.4%
Rosmarin et al,33 2020	Ruxolitinib	0.15% Cream daily; 0.5% cream daily; 1.5% cream twice daily; 1.5% cream daily	Vitiligo	125	14.4	32	3.1	48.3 (12.9)b	53.5
Rosmarin et al,34 2022	Ruxolitinib	1.5% Cream daily	Vitiligo	449	0.7	224	0.9	39.5 (15.1)b	53.1
Pooled proportion of treatment group with acne, 3.7%
Bissonnette et al,35 2016	Tofacitinib	2% Ointment twice daily	Atopic dermatitis	35	0	34	8.8	31.4 (9.9)b	53.6
Papp et al,36 2012	Tofacitinib	2 mg/d; 5 mg/d; 15 mg/d	Psoriasis	147	1.4	50	4.0	44.3 (13.7)b	36.5
Pooled proportion of treatment group with acne, 1.1%
Danese et al,37 2022	Upadacitinib	45 mg/d	Ulcerative colitis	302	3.3	149	4.0	42.0 (23.5)b	37.7
Guttman-Yassky et al,38 2020	Upadacitinib	7.5 mg/d; 15 mg/d; 30 mg/d	Atopic dermatitis	86	14.0	40	2.5	40.0 (15.7)b	37.7
Guttman-Yassky et al,39 2021	Upadacitinib	15 mg/d; 30 mg/d	Atopic dermatitis	1124	12.8	559	1.1	33.8 (15.6)b	44.9
Katoh et al,40 2023	Upadacitinib	15 mg/d; 30 mg/d	Atopic dermatitis	182	16.5	90	5.6	35.8 (12.8)b	77.6
Reich et al,41 2021	Upadacitinib	15 mg/d; 30 mg/d	Atopic dermatitis	597	11.9	303	2.0	34.1(NR)	39.3

Abbreviation: NR, not reported.

^a Proportion in pooled treatment group with acne was 11.7%.

^b Reported as mean (SD).

^c Reported as median (IQR).

Among individuals exposed to JAK inhibitors, adverse events of acne following treatment were identified in 467 individuals (6.2%) compared with 44 controls (1.3%). The pooled odds ratio (OR) was calculated to be 3.83 (95% CI, 2.76-5.32), with increased ORs for abrocitinib (13.47 [95% CI, 3.25-55.91]), baricitinib (4.96 [95% CI, 2.52-9.78]), upadacitinib (4.79 [95% CI, 3.61-6.37]), deucravacitinib (2.64 [95% CI, 1.44-4.86), and deuruxolitinib (3.30 [95% CI, 1.22-8.93]) (Figure 2). There was moderate heterogeneity in reporting (I² = 24.0%).

Figure 2. Summary of Janus Kinase Inhibitors and Associations With Acne.

Arrows indicate upper bound of 95% CI greater than 15.00; size of markers, number of participants in each group; and width of diamond, upper and lower bound of 95% CI for pooled estimate of the odds ratio (OR).

Subgroup analyses by JAK inhibitor class were also performed and revealed an increased OR for JAK1-specific inhibitors (4.69 [95% CI, 3.56-6.18]), combined JAK1 and JAK2 inhibitors (3.43 [95% CI, 2.14-5.49]), and TYK2 inhibitors (2.64 [95% CI, 1.44-4.86]). There were no differences in acne incidence in pan-JAK inhibitors (2.83 [95% CI, 0.92-8.74]) or JAK3-specific inhibitors (2.06 [95% CI, 0.08-52.39]) (Figure 3). There was moderate heterogeneity in reporting (I² = 27.9%).

Figure 3. Subgroup Analysis by Janus Kinase (JAK) Inhibitor Specificity.

Arrows indicate upper bound of 95% CI greater than 8.00; size of markers, number of participants in each group; and width of diamond, upper and lower bound of 95% CI for pooled estimate of the odds ratio (OR). TYK2 indicates tyrosine kinase 2.

In addition, a subgroup analysis was performed comparing those studies investigating the role of JAK inhibitors in the treatment of dermatologic conditions compared with those studies looking at JAK inhibitors for other applications such as ulcerative colitis or rheumatoid arthritis. In this analysis, the pooled OR for dermatologic studies was 4.67 (95% CI, 3.10-7.05) compared with 1.18 (95% CI, 0.67-2.07) for nondermatologic studies (Figure 4). The heterogeneity of these outcomes was small, with I² = 31.1% and I² = 2.11%, respectively.

Figure 4. Subgroup Analyses by Indication.

Arrows indicate upper bound of 95% CI greater than 15.00; size of markers, number of participants in each group; and width of diamond, upper and lower bound of 95% CI for pooled estimate of the odds ratio (OR).

Two final subgroup analyses based on age and sex were conducted by meta-regression. These analyses failed to show that either sex (estimate, −0.006; P = .30) or age (estimate, −0.0175; P = .41) were associated with the effect size of JAK inhibitor use on the development of acne (eTable 3 in Supplement 1).

Discussion

This systematic review explores the incidence of acne as an adverse event following treatment with JAK inhibitors. These data demonstrate the significant increases in incidence of acne associated with combined JAK1 and JAK2 inhibitors, JAK1 inhibitors, as well as the TYK2 inhibitor deucravacitinib. Pan-JAK inhibitors and the JAK3-specific inhibitor ritlecitinib do not appear to be associated with incident acne; however, analysis of the effects of TYK2 inhibitors and ritlecitinib should be interpreted cautiously given the imprecision of the effect estimates and small sample sizes.

Acne vulgaris is well known to affect individuals’ self-esteem, mental health, and psychosocial functioning.^42,43,44 Thus, the occurrence of acne following treatment with certain classes of JAK inhibitors is of potential concern, as this adverse effect may jeopardize treatment adherence among some patients. While the role of JAK signaling in the pathogenesis of acne vulgaris is unclear, a previous study demonstrated that certain JAK proteins, namely JAK1 and JAK3, are overexpressed in acne vulgaris lesions.⁴⁵ Furthermore, the JAK-STAT pathway has been implicated in other chronic cutaneous inflammatory conditions, including rosacea.^46,47 Although this framework suggests the possibility that JAK1 inhibitors could be beneficial in cases of acne vulgaris, our findings suggest a more complicated association. Parallels could be drawn to the association between acne and oral corticosteroid use, where a short course of oral corticosteroids can help bring severe acne under control; however, prolonged corticosteroid use can lead to a paradoxical increase in acne.^48,49 Further studies are needed to elucidate the potential association between JAK1 inhibitors and incident acne.

Our subgroup analysis comparing the results of studies conducted by dermatologists compared with those conducted by others revealed a stark difference in the incidence of acne with JAK inhibitor use. The reasons underlying this difference are not clear but may be a consequence of increased sensitivity to the diagnosis of acne among dermatologists compared with nondermatologists. Although age and sex differences may exist between trials, our data suggest that differences between drugs are not due to these variations. Future research is required to determine whether these observed differences are due to clinician evaluation or a function of the underlying physiology of specific diseases.

Limitations

These results must be interpreted in the context of the limitations of the study design. As a systematic review and meta-analysis, our study relied on data reported from several phase 2 and 3 RCTs. While we attempted to include all relevant studies, we cannot definitively rule out the possibility that some studies were inappropriately excluded. Furthermore, there is moderate heterogeneity in the studies that were included and, while our use of a random-effects estimator maintains the validity of our pooled effect estimates, our results are more conservative than they would have been had we used a fixed-effects model. Additionally, as phase 2 and 3 clinical trials often implement strict inclusion criteria, the samples included in our study may not be representative of the more diverse patient population seen in clinical practice. Therefore, caution is warranted when attempting to generalize our results to a broader or otherwise dissimilar patient group. It is also worth noting that, for some drugs, very few studies were included. This reduces the power of our analysis substantially, limiting our ability to detect any association of acne with certain medications.

Another limitation is the varied sensitivity of studies to the outcome of interest (ie, acne). It is likely that some instances of acne may have been misclassified, for example as a rash or skin irritation. Furthermore, given the ubiquitous nature of acne, it is also likely that this adverse effect was occasionally left unreported in certain studies. This would lead to potentially significant underestimates of incident acne following JAK inhibitor use. Last, while some studies may have accurately listed the occurrence of new-onset acne as an adverse event, improvements in preexisting acne likely would not have been recorded. Therefore, our study would not be able to quantify possible improvements in acne with certain JAK inhibitors, even if this were the case in certain instances.

Conclusions

In this systematic review and meta-analysis, we identified higher rates of acne following treatment with a JAK inhibitor. Further studies are needed to characterize the underlying mechanism of acne with JAK inhibitor use and to identify best practices for treatment.

Supplement 1.

eTable 1. Search Strategy

eTable 2. Cochrane Risk of Bias Assessment

eTable 3. Summary of Subgroup Analyses

eFigure. Funnel Plot for Publication Bias

eTable 4. A Measurement Tool to Assess Systematic Reviews 2 (AMSTAR 2)

Supplement 2.

Data Sharing Statement