This post hoc analysis compared the efficacy and safety of tofacitinib in patients with active AS by baseline BMI. To the best of our knowledge, this is the first analysis to evaluate the effect of BMI on the efficacy and safety of a Janus kinase inhibitor in patients with AS. In general, regardless of BMI category, tofacitinib demonstrated greater efficacy than placebo at week 12, and no differences in treatment effects were observed between patients with baseline BMI < 25 kg/m2 (underweight/normal) versus ≥ 25 to < 30 kg/m2 (overweight) or ≥ 30 kg/m2 (obese), although there were a few exceptions in the obese BMI category, and sample sizes and some baseline characteristics varied between categories.

In patients with baseline BMI ≥ 30 kg/m2, ASAS40 response, ASAS partial remission, BASDAI50 response, and ASDAS-CRP inactive disease rates were similar with tofacitinib and placebo, and differences in BASDAI50 response rates with tofacitinib versus placebo were smaller than in patients with baseline BMI < 25 kg/m2. In this analysis, patients with BMI ≥ 30 kg/m2 were older, had a higher mean waist circumference and more active disease at baseline since they had a higher mean swollen joint count and were more likely to have enthesitis, a baseline hsCRP level of > 5 mg/L, and an inadequate response to a TNFi or to have prior biological DMARD use (non-IR), which may have contributed to the limited response observed in the higher BMI category. Of note, obesity can be associated with elevated CRP levels [14]; as such, elevations in CRP levels may be due to AS disease activity as well as obesity. In addition, it is important to acknowledge that the overall number of patients was small in the BMI ≥ 30 kg/m2 category, and smaller compared with the other BMI categories. Finally, it has been hypothesized that enthesitis in AS may arise from mechanical stress [15]. While it is possible that enthesitis-like symptoms in obese patients with spondyloarthritis are due to an overlap of both inflammation and mechanical stress, a real-world study of 477 patients with AS from the CorEvitas (previously Corrona) Psoriatic Arthritis/Spondyloarthritis Registry did not find any correlation between BMI and the presence of enthesitis [16].

Measures of efficacy used in this analysis, such as ASAS, BASDAI, and ASDAS-CRP, comprise several patient-reported outcomes [17,18,19] and may have been affected by other related conditions, such as depression and fibromyalgia. Obesity is known to increase the risk of depression, and depression can be a predictor for obesity [20]. Additionally, fibromyalgia, a condition mainly related to central sensitization, is more prevalent in patients with AS than in the general population [21]; an association between obesity and greater pain sensitivity, poorer sleep quality, and reduced physical strength and flexibility has been reported in patients with fibromyalgia syndrome [22]. Therefore, it is plausible that patients with a higher BMI in the current analysis may have been more apt to feeling unwell and, thus, scored lower in subjective measurements of the ASAS, BASDAI, and ASDAS-CRP assessments, which could then have led to lower overall responses. Furthermore, when treating patients with AS, addressing the origin of pain and other related conditions, such as fibromyalgia and depression, should be considered in order to improve overall treatment response and ensure optimal disease management [23, 24], especially in those with a high BMI.

Improvements in SF-36 PCS and MCS scores by baseline BMI were assessed separately by trial, given that these outcomes only had a single post-baseline visit measure. Differences in improvements from baseline in SF-36 PCS scores between tofacitinib and placebo were not consistent between trials, and no differences in improvement from baseline in SF-36 MCS scores were observed between tofacitinib and placebo in either trial, regardless of BMI category. It should be noted that these results should be interpreted with caution given the low number of patients.

The impact of BMI on treatment response has been explored in patients with AS receiving TNFi. In an analysis of 57 patients with AS, those with BMI > 30 kg/m2 who were receiving adalimumab had a lower probability of achieving BASDAI ≤ 4 or ASDAS ≤ 2.1 responses compared with those with BMI < 25 kg/m2 [25]. A retrospective study of infliximab in 155 patients with AS also showed that a higher BMI was associated with a reduced treatment response; significantly lower proportions of patients with a higher BMI versus those with a lower BMI achieved a decrease of 50% in BASDAI, the visual analog scale of pain, and CRP level [6]. Furthermore, a study of 41 patients with AS receiving etanercept or adalimumab demonstrated that a higher body fat percentage, fat mass index, or fat mass index percentile was independently associated with a lower chance of achieving at least a clinically important improvement based on the ASDAS-CRP criteria and a lower chance of ASDAS-CRP improvement [8]. Additionally, a higher fat mass index percentile was associated with a lower chance of achieving the minimum clinically important improvement for BASDAI, and a higher body fat percentage percentile and fat mass index percentile was associated with a negative change in BASDAI scores [8]. However, in contrast with other TNFi studies, no relationship between a higher BMI and an impaired clinical response was reported, indicating that the BMI may not fully reflect the effect of body composition in patients with AS [8].

Irrespective of baseline BMI category, the overall safety of tofacitinib presented herein was generally consistent with the prior phase 2 [9] and 3 [10] trials. Up to week 16, AEs and SAEs occurred more frequently in tofacitinib-treated patients with baseline BMI < 25 kg/m2 versus those in higher baseline BMI categories. However, it should be noted that patients with a baseline BMI < 25 kg/m2 were more likely to be current smokers versus those in other BMI categories. Notably, an association between smoking and higher disease activity, increased inflammation and structural damage has been reported in patients with early axial spondyloarthritis [26]. Similar to a study in RA, where patients with a low or normal BMI (i.e., BMI < 20 kg/m2 or ≥ 20–25 kg/m2, respectively) were generally more likely to discontinue conventional synthetic DMARD or TNFi therapy compared with those who were overweight (i.e., BMI ≥ 25–30 kg/m2) [27], it is possible that patients with AS who have a low BMI may have a worse health status that leads to more treatment side effects.

Across BMI categories, elevated lipid parameters at weeks 12 and 16 in the phase 2 and 3 RCTs, respectively, were generally reported in patients receiving tofacitinib versus placebo, while liver enzyme levels were generally similar between treatments. Differences in laboratory parameters were observed across BMI categories in tofacitinib-treated patients but were considered not clinically meaningful. Notably, in a previous report of patients with psoriatic arthritis, elevated aspartate aminotransferase and alanine aminotransferase levels were observed in patients with baseline BMI ≥ 35 kg/m2 who were receiving tofacitinib 5 mg BID but not in those receiving tofacitinib 10 mg BID; however, the small sample size within each tofacitinib group in the baseline BMI ≥ 35 kg/m2 category may have led to this inconclusive finding [13].

Several limitations of this analysis should be considered. This analysis was post hoc in nature and, therefore, the trials were not powered to compare differences across BMI categories. The overall sample size was low, and there were differences in sample sizes between BMI categories. In addition, the phase 2 trial did not include efficacy data at week 16, which led to smaller pooled sample sizes for some analyses, and the trial designs limited the evaluation of the long-term impact of BMI on treatment response. Furthermore, waist circumference data were only available for a very small number of patients enrolled in the phase 3 trial across treatments and baseline BMI groups, further limiting the application of these data in this analysis. There was also some imbalance in baseline characteristics, which was expected to impact efficacy and safety outcomes across BMI categories. Lastly, BMI is an anthropometric measure and does not assess functional health and fitness [28]. Therefore, future studies could benefit from a combined approach of objective measures of body fat (e.g., waist circumference and dual-energy X-ray absorptiometry data) and comprehensive blood variables (e.g., laboratory measures of inflammation [28] and the active level of the drug on board over time).

In summary, regardless of baseline BMI category, efficacy was greater with tofacitinib versus placebo in patients with active AS, and no differences in treatment effects were observed across categories, though there were a few exceptions in patients with BMI ≥ 30 kg/m2, who had more active and treatment-refractory disease and a smaller sample size versus other BMI categories. The overall safety of tofacitinib was generally comparable across BMI categories; however, AEs and SAEs occurred more frequently in tofacitinib-treated patients with BMI < 25 kg/m2, who were more likely to be current smokers, compared with other categories.