Background TNF-α inhibitor (TNFI) treatment fails among half of patients with psoriatic arthritis (PsA) treated in routine care. Obesity is a frequent, pro-inflammatory condition in PsA that may affect pharmacokinetics, immunological processes and patient-reported outcomes.
Objectives To investigate the impact of obesity on disease activity, treatment response and drug adherence among PsA patients initiating their first TNFI treatment course in routine care.
Methods Observational cohort-study based on the Danish (DANBIO) and Icelandic (ICEBIO) registries. Kaplan-Meier plots, Cox regression and logistic regression analyses studied the impact of obesity (body mass index (BMI) ≥30kg/m2) on drug adherence and TNFI response. Six-month response was defined according to ACR20/50/70 and EULAR criteria. Sub-analyses stratified according to gender and TNFI (adalimumab/etanercept/infliximab) were performed.
Results Among 1943 patients (1750 Danish, 193 Icelandic) included in the study, 1271 (65%) had available BMI and 408 (32%) of these were obese (327 (80%) Danish, 81 (20%) Icelandic). Prescription rates were: adalimumab (39%), etanercept (26%), infliximab (24%), golimumab (7%) and certolizumab (4%). Median follow-up-time was 1.5 years, IQR: 0.5–3.9. At baseline obese patients smoked less (23% vs 30%), were older (49±12 vs 47±13 years, (mean (SD)) and had higher disease activity measured as DAS28 (4.6±1.2 vs 4.4±1.2), tender joint count (6 (3–13) vs 5 (2–11); median, (IQR)), C-reactive protein (9 (5–19) vs 7 (3–18)), HAQ score (1.1 (0.8–1.6) vs 0.9 (0.5–1.4)), Visual-Analogue-Scale (VAS) patient global (72 (54–87) vs 65 (50–80)) and VAS pain (66 (48–76) vs 60 (38–74)), compared with non-obese (all p<0.05). Obese patients had shorter treatment adherence than non-obese (Figure 1), especially among men (5.9 (4.1–7.7) vs 2.5 (1.7–3.2)); median years (95%CI), (p<0.01). Similar results were found when Icelandic and Danish patients were analyzed separately. Obesity was an independent predictor of TNFI withdrawal both overall: HR 1.6 (95%CI 1.3–2.0), in men: HR 1.8 (1.3–2.5), women: HR 1.5 (1.1–2.0) and for every TNFI: adalimumab: HR 1.6 (1.2–2.2), etanercept: HR 2.0 (1.3–3.3), infliximab: HR 1.6 (1.1–2.5). The EULAR good-or-moderate response rate was significantly lower in obese than non-obese (55% vs 65%), (p=0.02). In men, obesity reduced the chance of both EULAR good response: OR 0.5 (95%CI 0.3–0.9) and EULAR good-or-moderate, ACR20 and ACR50 responses (all OR 0.4 (0.2–0.8)). Obesity also predicted a poorer EULAR good-or-moderate response overall (OR 0.6 (0.4–0.9)), in women (OR 0.5 (0.3–0.9)) and for all TNFI types (data not shown).
Conclusions Obesity is associated with increased disease activity and attenuates response and adherence to TNFIs in PsA.
Acknowledgement Thank you to Elinborg Stefansdottir who assisted in the achievement of Icelandic data, to Niels Steen Krogh from Zitelab for extraction of data and to Peder Frederiksen and Robin Christensen for statistical advises. Thanks to all Danish and Icelandic Departments of Rheumatology for providing data to DANBIO and ICEBIO.
Disclosure of Interest None declared