Search and Selection

Our search across Medline (142), Embase (393) and the Cochrane Library [65] identified 604 studies. After duplicate removal, 366 articles underwent screening, resulting in the inclusion of 36 articles (Fig. 1). The exclusion of studies was necessitated by challenges of aggregating results due to heterogeneous data reporting. The subsequent quantitative analysis, involving 19 selected articles, was conducted.

Fig. 1

PRISMA 2020 flow diagram for the selection process

Basic Characteristics of Studies Included

The basic characteristics of the studies, including randomised controlled trials (RCTs), cohort studies and case reports, are reported in Tables 1, 2, and 3, respectively. The meta-analysis of RCTs included two separate articles, collectively featuring three RCTs [27, 28], all employing topical ruxolitinib cream. A meta-analysis comprising four single-arm trials was conducted, with all but one employing topical JAKi. In addition, a cohort was curated from case reports and case series documenting improvements in vitiligo presented as a percentage of the affected body surface area. All case reports included in the quantitative analysis utilised oral JAKi.

Table 1 Characteristics of included RCTsTable 2 Basic characteristics of single-arm trials includedTable 3 Characteristics of the included case reports and case series were used to construct the cohortMeta-analysis of RCTs

A meta-analysis of three RCTs, including 726 patients, was conducted [27, 28]. The primary efficacy outcome was a 75% improvement in the F-VASI score, denoted as FVASI-75, relative to baseline measurements. In the work of Rosmarin et al. [28], the primary endpoint was defined as a 50% improvement in the F-VASI score. However, we chose to include a 75% improvement in F-VASI, as it was more methodologically conducive to conducting the meta-analysis. Secondary endpoints in the work of Rosmarin et al. [28] encompassed both a 50% and a 75% improvement in the total VASI score. Nevertheless, due to incomplete reporting across all included RCTs, these secondary endpoints could not be systematically pooled and were consequently excluded from the meta-analysis.

All RCTs investigated a topical JAK inhibitor, ruxolitinib, at a concentration of 1.5% in the intervention arm, with a vehicle cream serving as the control over a period of 24 weeks. Varied concentrations of topical JAK inhibitors were explored in the study by Rosmarin et al. [28]; however, the available data were insufficient to facilitate a meta-analysis.

In the intervention group, comprising 476 patients, 144 reached the FVASI-75 endpoint (30.3%), whereas in the control group of 250 patients, 20 reached the FVASI-75 endpoint (8.0%), giving a RR to reach FVASI-75 of 3.47 (95% CI 0.98–12.22; p = 0.051) for the two groups (Fig. 2).

Fig. 2

Forest plot of the number of patients who achieved 75% improvement in their F-VASI score

The analysis of adverse events involved 482 patients in the intervention arm and 256 patients in the control arm. In the intervention arm, 238 patients experienced any adverse event (49.4%), compared with 42.4% (111 patients) in the control group (RR 1.27; 95% CI 0.88–1.82; p = 0.107) (Fig. 3). This encompasses both newly occurring adverse events and the exacerbation of pre-existing ones following the initial drug dose. In addition, serious adverse events occurred in nine patients in the intervention group and one patient in the control group (1.9% versus 0.4%), with a RR of 2.83 (95% CI 2.31–3.47; p = 0.002) (Figs. S1 and S2). In total, there were four cases of adverse events leading to discontinuation: two in the intervention arm and two in the control arm. In the intervention arm, discontinuations were due to fatigue and application site rash, whereas in the control arm, they were due to nausea, headache and seizure.

Fig. 3

Forest plot of the number of patients who experienced any adverse events throughout the trial

Meta-analysis of Single-Arm Trials

A meta-analysis consisting of four single-arm trials was conducted to evaluate the efficacy of JAK inhibitors, both alone and in combination with conventional treatments, in the repigmentation of vitiligo lesions, with the minimum follow-up time ranging from 9 to 20 weeks [29,30,31,32]. The primary outcome measure assessed was the single mean percentage change in VASI. Among these trials, all but one employed topical JAK inhibitors, with two studies utilising tofacitinib 2% and one using ruxolitinib 1.5%. The sole oral JAK inhibitor utilised was upadacitinib (15 mg/qd).

The pooled analysis of single-arm trials utilising either JAKi alone or in combination with other treatments revealed a statistically significant improvement, with a single mean percentage change in VASI of 43.79% (95% CI 0.71–0.93; p < 0.001) (Fig. 4).

Fig. 4

Forest plot of the mean percentage repigmentation in vitiligo lesions of patients included in single-arm trials employing JAK inhibitors with or without conventional treatment

Upon subgroup analysis of trials utilising solely JAK inhibitors (comprising three studies, with two employing topical JAK inhibitors – tofacitinib and ruxolitinib, respectively – and one using an oral JAK inhibitor), the observed single mean percentage change in VASI was 30.83 (95% CI 5.14–56.5) (Fig. S3). This shows a clinically relevant and statistically significant improvement.

Furthermore, our analysis of single-arm trials exclusively investigating topical JAK inhibitors revealed a single mean percentage change in VASI of 44.7% (95% CI 19.36–70.1), further supporting the robust efficacy of topical JAK inhibitors (Fig. S4).

Analysis of the Novel Cohort Constructed from Case Reports and Case Series

Due to the lack of RCTs investigating oral JAKi, we constructed a cohort of 28 patients compiled from case reports and series evaluating the efficacy of JAK inhibitors, either alone or in combination with other treatments, with follow-up times ranging from 9 to 12 months. In all cases, different compounds of oral JAK inhibitors were employed. Patients were divided into three groups: group A received JAK inhibitors alone, group B received JAK inhibitors with UV-B phototherapy, and group C received JAK inhibitors in combination with other treatments, except UV-B phototherapy. The primary outcome measure was the percentage repigmentation of vitiligo lesions relative to baseline. Basic characteristics of patients are detailed in Supplementary Table 7.

Group A [33,34,35,36,37,38] showed a significant improvement in vitiligo lesions, with an average percentage increase of 48.7% (95% CI 28% to ∞; p = 0.0018). Group B [33,34,35, 39,40,41,42] showed similar significant improvement (average increase of 63.7%; 95% CI 49% to ∞; p < 0.0001). However, group C [43, 44] did not reach statistical significance (average increase of 56.3%; 95% CI 18% to ∞; p = 0.061), although a trend favouring JAK inhibitors was suggested by the confidence interval excluding zero. The limited sample size precludes definitive conclusions on the efficacy of group C. The analysis of the cohort, without grouping it into the three groups, revealed a significant repigmentation compared with baseline of 55.68% (95% CI 44.51% to ∞, p < 0.0001).

Two Fisher’s exact tests were used to compare differences in the proportion of patients below and above specified repigmentation thresholds within the groups. Both tests, conducted at thresholds of 0% and 50%, respectively, revealed non-significant differences between the ratios of patients above and below the threshold (p = 0.32 and p = 0.35, respectively), likely attributable to the limited statistical power resulting from the small sample size. Moreover, an ANOVA test was conducted to compare the three groups against each other, revealing no significant differences between them (p = 0.456) (Fig. 5), possibly due to the small sample size. However, nearly all patients receiving combination therapies (14/15) demonstrated clinical improvement in their vitiligo lesions in contrast to patients receiving JAK inhibitors alone (9/13).

Fig. 5

Analysis of variance of the three groups assembled from the case reports. N = 28, p = 0.456, group A – JAK inhibitors alone, group B – JAK inhibitors and UVB, group C– JAK inhibitor and other treatments except for UV

Systematic Review

Several studies were excluded from the statistical analysis. NCT03715829 examined the efficacy and safety of various doses of ritlecitinib over 48 weeks in patients with non-segmental vitiligo and found notable efficacy in the 50 mg group (p < 0.001) and 30 mg group (p = 0.01), with further improvement using 200/50 mg in the extension period (n = 187). Treatment-related adverse events were well tolerated and not dose dependent [45].

Fang et al. conducted a pilot study on tofacitinib with phototherapy, reporting effective repigmentation and no adverse events [46]. Joshipura et al. extended the study by Rothstein et al. [28] on ruxolitinib with optional phototherapy, showing significant VASI score improvement at week 52, especially in facial vitiligo [47]. These results align with findings from Gianfaldoni et al. and Song et al., who also combined tofacitinib and phototherapy [48, 49].

Furthermore, numerous case reports detailed successful outcomes with minimal side effects using various JAKi in vitiligo management. Murakami et al. showed that delgocitinib with phototherapy was effective for vitiligo and alopecia areata, without side effects [50]. Camacho et al. reported a positive response to ruxolitinib in an 8-year-old child [51]. Aickara et al. achieved near-complete facial and scalp repigmentation with oral tofacitinib, without adverse events [52]. Yagi et al. documented significant repigmentation with delgocitinib and phototherapy in two cases, without adverse events [53]. Moore et al. reported repigmentation with oral tofacitinib, but noted gastrointestinal side effects requiring dosage reduction [41]. Ferreira et al. reported near-complete repigmentation with minor adverse events using topical tofacitinib in an adolescent [54]. Olamiju et al. reported a rapid and complete response to topical tofacitinib in a child with segmental vitiligo [55]. Joshipura et al. documented successful repigmentation in sun-exposed areas using ruxolitinib and tofacitinib in separate cases [56]. Hu et al. showed successful repigmentation of halo nevi in two paediatric patients using topical tofacitinib [57]. Vu et al. observed marginal improvement with mild side effects using oral and topical steroids, phototherapy and topical tofacitinib [58].

Risk-of-Bias Assessment and Certainty of Evidence

The risk-of-bias assessment of the RCTs included showed low risk for all outcomes. However, for the single-arm trials, the majority showed a moderate risk of bias. The detailed evaluation is presented in Supplementary Tables 2, 3 and 5.

Funnel plots for our outcomes showed no evidence of publication bias (Fig. S6). Detailed results of the GRADE assessment are presented in Supplementary Table 9.