We prospectively evaluated whether obesity impacts achievement of minimal disease activity (MDA) in subjects with psoriatic arthritis (PsA).
We prospectively evaluated whether obesity impacts achievement of minimal disease activity (MDA) in subjects with psoriatic arthritis (PsA).
Among PsA subjects with an active disease and who were starting a treatment with tumor necrosis factor α blockers, 135 obese (body mass index [BMI] >30 kg/m2) patients and 135 patients of normal weight (controls) were followed up for 24 months. At baseline and at the 12- and 24-month followup, all subjects underwent a clinical, rheumatologic, and laboratory assessment.
With the exception of the prevalence of hypercholesterolemia and hypertriglyceridemia, case and control subjects were similar for all the clinical and demographic characteristics analyzed. At the 12-month followup, in both cases and controls, no significant changes in body weight were found (P > 0.05 for all). MDA was achieved by 98 (36.3%) of the 270 PsA individuals. The prevalence of obesity was higher in those that did not achieve MDA than in those that did (64.0% versus 25.5%; P < 0.001). After adjusting for all the other variables, obesity was associated with a higher risk of not achieving MDA (hazard ratio [HR] 4.90, 95% confidence interval [95% CI] 3.04–7.87; P < 0.001). The HR of not achieving MDA was 3.98 (95% CI 1.96–8.06, P < 0.001) and 5.40 (95% CI 3.09–9.43, P < 0.001) in subjects with first-degree (BMI <30 kg/m2) and second-degree (BMI 30–35 kg/m2) obesity, respectively. Among the 98 subjects who had achieved MDA at the 12-month followup, the presence of obesity was associated with a poor probability of sustained MDA at the 24-month followup (HR 2.04, 95% CI 1.015–3.61; P = 0.014).
Obesity is a negative predictor of achieving and maintaining MDA.
Subjects with rheumatic diseases, such as psoriatic arthritis (PsA), exhibit an enhanced prevalence of the metabolic syndrome and of some of its major features (obesity, hypertension, hypercholesterolemia, hypertriglyceridemia, and impaired fasting glucose) (1). As for other patients with rheumatic diseases, clinical studies (2) show a high prevalence of obesity in PsA patients as compared with the general population. In view of the fact that chronic inflammation in PsA acts synergistically with traditional vascular risk factors (VRFs) (3, 4), a tight interrelation between obesity and PsA can be suspected. By leading to an abnormal expression of “adipokines” (e.g., tumor necrosis factor α [TNFα], interleukin-6 [IL-6], leptin, adiponectin), obesity leads to a proinflammatory status (5–7). In view of this, obesity has been defined as a low-grade, chronic systemic inflammatory disease (8, 9). Further support for a link between obesity and inflammation emerges from nutritional intervention trials. In parallel with reducing obesity, caloric restriction affects a variety of circulating inflammatory markers, thus affecting the inflammatory status in humans (10). In this prospective study, we evaluated whether obesity impacts the clinical response in PsA patients starting a treatment with TNFα blockers.
Obesity has been recognized as a low-grade chronic inflammation.
Obesity-related inflammatory status can interact with the immune-related inflammation in obese subjects with psoriatic arthritis (PsA).
Obesity is a negative predictor of achieving minimal disease activity in PsA patients starting treatment with tumor necrosis factor α blockers.
During a 3-year period (January 2007–January 2010) we enrolled, as the case group, 135 consecutive PsA subjects with obesity (see below for obesity definition) who had a PsA diagnosis according to Classification of Psoriatic Arthritis (Study Group) criteria (11), who were nonresponders to traditional disease-modifying antirheumatic drugs (DMARDs), and who were referred to the Psoriatic Arthritis Clinic of the Federico II University of Naples to start a treatment with TNFα blockers. A total of 135 consecutive PsA subjects of normal weight, referred during the same time period to start a treatment with TNFα blockers, served as the control group.
Both case and control subjects had a clinically active disease at the time of enrollment; both groups were classified into different clinical subsets according to the Moll and Wright criteria (12). For all enrolled subjects, exclusion criteria were the lack of informed consent signature, age <18 years, previous treatment with TNFα blockers, malignancy, hematologic diseases, autoimmune diseases other than PsA, unstable medical conditions, and ongoing pregnancy. After informed consent signature, data about age, sex, height, weight, disease duration, disease activity, previous and/or current treatments, and VRFs were collected from all patients as previously described (13). According to current guidelines (14), obesity was defined as a body mass index (BMI) >30 kg/m2. In view of this definition, both case and control subjects were stratified into 3 BMI groups as follows: 1) normal weight: BMI <30 kg/m2; 2) first-degree obesity: BMI 30–35 kg/m2; and 3) second-degree obesity: BMI 35–40 kg/m2. The National Cholesterol Education Program criteria (15) define VRFs as follows: 1) hypertriglyceridemia as triglycerides levels ≥150 mg/dl, 2) hypercholesterolemia with low high-density lipoprotein (HDL) cholesterol as a total cholesterol of ≥200 mg/dl, with HDL cholesterol <40 mg/dl for men and <50 mg/dl for women, 3) hypertension as a blood pressure ≥130 and/or 85 mm Hg, and 4) impaired fasting glucose as a fasting glucose ≥100 mg/dl. All these criteria were employed to stratify cases and controls.
After baseline evaluation (T0), all subjects started treatment with TNFα blockers and were followed up (every 3 months) for 12 months (T1). At baseline and at each followup visit, all PsA subjects underwent a complete clinical rheumatologic and laboratory evaluation that included tender joint count (TJC), swollen joint count (SJC), tender entheseal count, Psoriasis Area Severity Index (PASI), Health Assessment Questionnaire (HAQ), visual analog scale (VAS) for pain, patient global disease activity VAS score, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) level. During the first 12 months of followup (from T0 to T1), subjects were classified as having achieved minimal disease activity (MDA) when fulfilling 5 of the following 7 outcome measures at T1: TJC ≤1, SJC ≤1, PASI score ≤1 or body surface area ≤3, VAS for pain ≤15, patient global disease activity VAS score of ≤20, HAQ score ≤0.5, and tender entheseal points ≤1 (16). Otherwise, the subjects were considered as not having achieved MDA (non-MDA). Those that had achieved MDA were reevaluated every 3 months thereafter during the additional 12 months (T2) to test the achieving of sustained MDA (s-MDA). S-MDA was defined as the fulfilling of the same criteria employed for MDA in all the visits carried out between T1 and T2. In addition to body weight and BMI, cardiometabolic risk factors were evaluated at each followup visit. To evaluate changes in nutritional habits, a periodic nutritional interview was performed by trained staff.
Statistical analysis was performed using SPSS, version 16.0. Continuous data were expressed as the mean ± SD; categorical variables were expressed as the percentage. The t-test was performed to compare continuous variables; the chi-square test was employed to analyze categorical data. When the minimum expected value was <5, Fisher's exact test was used.
A Kaplan-Meier survival model (with the log rank test) was adopted to evaluate the cumulative hazard of achieving MDA according to the presence of obesity. To adjust for all the other variables and to evaluate the achieving of MDA (or of s-MDA), Cox regression analyses (stepwise method) were adopted, with achieving MDA (or s-MDA) as the dependent variable, and obesity, TJC, SJC, tender entheseal count, PASI, HAQ, VAS for pain, patient global VAS, ESR, CRP level, concomitant treatment with methotrexate, sex, age, disease duration, smoking habit, diabetes mellitus, hypercholesterolemia, hypertriglyceridemia, and hypertension as independent variables. Multicollinearity effect in multivariable regression models was excluded by a stepwise approach with each variable included for P values less than 0.05 and excluded for P values greater than 0.1. A tolerance test was employed to exclude models in which the sum of the values exceeded the sum of the variances for all variables. All the results are presented as 2-tailed values with statistical significance if P values were less than 0.05.
As for the sample size calculation, when planning a study with a minimal predefined hazard ratio (HR) >1.5, a 1:1 ratio between case and control subjects, an accrual interval of 3 years, and a followup of 24 months, at least 130 subjects for each group are needed to achieve a >80% power with a 5% alpha error.
Clinical and demographic characteristics of case and control populations are reported in Table 1. As to the TNFα blockers used, adalimumab (40 mg/2 weeks) was employed in 80 (29.6%) subjects, etanercept (50 mg/week) in 111 (41.1%) subjects, and infliximab (5 mg/kg every 8 weeks) in 79 (29.3%) subjects.
|Variable||Normal-weight patients (n = 135)||Obese patients (n = 135)||P|
|Age, years||51.12 ± 12.95||52.33 ± 9.83||0.386|
|Male sex, no. (%)||70 (51.9)||54 (40.0)||0.067|
|Clinical subset, no. (%)|
|Axial and peripheral||46 (34.1)||49 (36.3)||0.799|
|Peripheral||52 (38.5)||43 (31.9)||0.308|
|Axial||26 (19.3)||38 (28.1)||0.115|
|Mutilans||11 (8.1)||5 (3.7)||0.196|
|Disease duration, months||114.71 ± 59.66||107.02 ± 78.76||0.367|
|ESR, mm/hour||17.82 ± 12.27||18.28 ± 14.09||0.779|
|CRP level, mg/dl||2.93 ± 3.17||2.35 ± 2.65||0.110|
|SJC||3.50 ± 3.77||4.03 ± 4.36||0.291|
|TJC||11.44 ± 5.05||12.41 ± 5.86||0.147|
|PASI score||1.45 ± 0.70||1.51 ± 0.72||0.497|
|HAQ score||2.65 ± 1.45||2.49 ± 1.44||0.147|
|VAS||76.96 ± 22.23||76.44 ± 17.84||0.833|
|Patient global VAS||75.11 ± 23.74||73.40 ± 20.98||0.533|
|Tender entheseal count||10.77 ± 5.23||11.16 ± 5.57||0.559|
|Hypercholesterolemia, no. (%)||60 (44.4)||79 (58.5)||0.028|
|Hypertriglyceridemia, no. (%)||36 (26.7)||56 (41.5)||0.014|
|Impaired fasting glucose, no. (%)||9 (6.7)||11 (8.1)||0.817|
|Hypertension, no. (%)||29 (21.5)||28 (20.7)||1.000|
|Smoking habit, no. (%)||27 (20.0)||38 (28.1)||0.154|
|Obesity, no. (%)||0 (0)||135 (100)||< 0.001|
|First-degree||0 (0)||100 (74.1)||< 0.001†|
|Second-degree||0 (0)||35 (25.9)||< 0.001†|
With the exception of the prevalence of obesity, hypercholesterolemia, and hypertriglyceridemia, the 2 study groups did not significantly differ for other clinical and demographic variables considered. Nor did they differ as to the rheumatologic variables whose outcome measurements have been shown to account for the achieving of MDA (17, 18). No difference in the prevalence of a concomitant treatment with methotrexate was found between normal weight and obese PsA subjects (43.7% versus 38.5%; P = 0.458).
During the followup, no changes in treatment schedules and/or in nutritional habits of both case and control subjects were reported. When comparing T0 and T1 data, no significant differences in body weight, as well as in the metabolic variables analyzed, were found (P > 0.05 for all by paired-sample t-test). As a whole, MDA was achieved by 98 (36.3%) of 270 PsA individuals.
As to the type of TNFα blocker used, MDA was achieved in 33 (41.8%) subjects receiving infliximab, in 41 (36.9%) of those receiving etanercept, and in 24 (30%) of those receiving adalimumab (P = 0.299 for trend). By evaluating baseline measurements (Table 2), those that had achieved MDA were more often males (P = 0.001), of a younger age (P = 0.042), had a longer disease duration (P = 0.004), a higher ESR (P < 0.001) and CRP level (P < 0.001), and a lower TJC (P < 0.001) and tender entheseal counts (P = 0.003) than those that did not. Among cardiometabolic risk factors, hypercholesterolemia (P < 0.001) and obesity (P < 0.001) were more frequent in those that did not achieve MDA compared with those that did.
|Variable||Non-MDA patients (n = 172)||MDA patients (n = 98)||P|
|Age, years||52.66 ± 10.60||50.08 ± 8.81||0.042|
|Male sex, no. (%)||65 (37.8)||59 (60.2)||0.001|
|Clinical subset, no. (%)|
|Axial and peripheral||63 (36.6)||32 (32.7)||0.596|
|Peripheral||59 (34.3)||36 (36.7)||0.693|
|Axial||41 (23.8)||23 (23.5)||1.000|
|Mutilans||9 (5.2)||7 (7.1)||0.595|
|Disease duration, months||101.72 ± 74.01||126.93 ± 58.83||0.004|
|ESR, mm/hour||14.35 ± 11.06||20.16 ± 13.86||< 0.001|
|CRP level, mg/dl||1.66 ± 1.83||3.20 ± 3.28||< 0.001|
|SJC||3.75 ± 4.30||3.80 ± 3.68||0.929|
|TJC||12.99 ± 5.62||10.06 ± 4.71||< 0.001|
|PASI score||1.49 ± 0.68||1.45 ± 0.76||0.700|
|HAQ score||2.39 ± 1.41||2.61 ± 1.52||0.230|
|VAS||77.50 ± 17.30||75.30 ± 24.33||0.390|
|Patient global VAS||74.24 ± 21.08||74.28 ± 24.62||0.988|
|Tender entheseal count||11.71 ± 5.45||9.66 ± 5.06||0.003|
|Hypercholesterolemia, no. (%)||103 (59.9)||36 (36.7)||< 0.001|
|Hypertriglyceridemia, no. (%)||60 (34.9)||32 (32.7)||0.790|
|Impaired fasting glucose, no. (%)||11 (6.4)||9 (9.2)||0.470|
|Hypertension, no. (%)||39 (22.7)||18 (18.4)||0.441|
|Smoking habit, no. (%)||36 (20.9)||29 (29.6)||0.138|
|Obesity, no. (%)||110 (64.0)||25 (25.5)||< 0.001|
|First-degree||84 (48.8)||16 (16.3)||< 0.001†|
|Second-degree||26 (15.1)||9 (9.2)||< 0.001†|
With respect to the time elapsed from the beginning of the TNFα blockers treatment, a Kaplan-Meier survival model (Figure 1) showed a significant difference in achieving MDA according to the presence/absence of obesity (log rank 41.77, P < 0.001). After adjusting for all the other demographic characteristics and for rheumatologic and cardiovascular variables, obesity was the strongest predictor of the risk of not achieving MDA (HR 4.90, 95% confidence interval [95% CI] 3.04–7.87; P < 0.001). A progressively increasing HR of not achieving MDA was found for increasing degrees of obesity (as evaluated by BMI) (Figure 2).
Together with obesity, low TJC (HR 1.15, 95% CI 1.08–1.23; P < 0.001), male sex (HR 2.25, 95% CI 1.47–3.43; P < 0.001), and a high CRP level (HR 1.18, 95% CI 1.07–1.30; P < 0.001) were found to independently predict achieving MDA in the regression model.
Among the 98 subjects that had achieved MDA during the first 12 months of followup, 17 (17.3%) relapsed and 81 (82.7%) maintained their MDA during the additional 12-month followup. In a separate regression model that evaluated only those 98 subjects achieving MDA during the first 12-month followup, obesity was associated with a high risk of disease relapsing during the additional 12 months of followup (HR 2.04, 95% CI 1.015–3.61; P = 0.014).
In PsA subjects starting treatment with TNFα blockers, this prospective study shows that obesity is associated with a high risk of not achieving MDA. Analysis for increasing BMI levels confirmed and extended this finding.
In the setting of PsA subjects, a variety of 6-month and 3-month clinical and laboratory predictors of response to TNFα blockers have been identified (17, 18). So far, however, no ad hoc study has tested obesity as a predictor of success of the TNFα blockers treatment in PsA subjects.
A variety of markers (CRP level, C3, IL-6, and TNFα) implicated in the etiology and the development of insulin resistance, type 2 diabetes mellitus, and atherosclerosis (19–21) are also important in the regulation of the inflammatory process (22) and, in turn, in the pathophysiology of arthritides (23). Obesity is a central event in insulin resistance and type 2 diabetes mellitus. Thus, an interaction between obesity-related and immunity-related inflammatory status may be postulated in obese PsA subjects. Intervention studies (10) support such formulation. On the other hand, major markers of atherosclerosis (carotid intima-media thickness and hepatic steatosis) that assess the global damage secondary to metabolic and inflammatory determinants (24) are severely affected in PsA subjects (2, 13, 25). These data provide the rationale for testing whether, among PsA subjects starting a treatment with TNFα blockers, obesity would impact achieving MDA and s-MDA.
Among the 98 (of 270) PsA patients who achieved MDA at 12 months, the prevalence of obesity was significantly lower than in those that did not. Accordingly, obesity was associated with a high risk of not achieving MDA (P < 0.001). This concept was strengthened by stratifying the population according to increasing BMI degrees. Moreover, obesity was associated with a poor probability of s-MDA at the 24-month followup (P = 0.014). The combined data lend credence to the possibility that obesity is a negative predictor of achieving and maintaining MDA.
The relevance of obesity as an MDA predictor has been analyzed in a regression model in which the large majority of demographic, clinical, and laboratory variables known to predict the achieving of MDA were present (18). The HRs reported here have been computed after adjusting for all these variables, and argue for obesity as being a major determinant of achieving MDA.
At variance with the other TNFα blockers used, infliximab needs a weight adjustment of dosages. The lack of a meticulous dose adjustment might have hampered an optimal response to TNFα blocker treatment in obese PsA subjects. However, in spite of its potential inherent limitations (the drug presentation is in 100-mg vials), special attention has been paid to this issue during the study period. As reported in the Results, no difference in achieving MDA was found when the data were stratified according to type of TNFα blocker used (P = 0.299 for trend). Moreover, by evaluating only obese subjects, MDA was achieved in 7 (15.6%) subjects receiving infliximab, in 10 (21.3%) of those receiving etanercept, and in 8 (18.6%) of those receiving adalimumab (P = 0.779 for trend).
By translating the results of the present study to clinical practice, one has to argue that, in addition to being a predictor of a poor response, obesity should be primarily considered as an indication to adjust the dosing of TNFα blockers in PsA.
Potential limitations of this study need to be addressed. In the present setting, hypercholesterolemia and hypertriglyceridemia were more prevalent in the obese PsA group. Based on the tight relationship between obesity and other VRFs, this prevalence was expected. To overcome a potential bias, all the variables examined were included in the regression model to obtain adjusted values of the HR.
As obesity is correlated with significantly higher levels of major inflammatory markers (9), in addition to the ESR and CRP level, differences in TNFα and IL-6 levels between obese and PsA subjects of normal weight could have been of interest. However, these measurements were available only in few cases.
Among PsA subjects enrolled in the present setting, the prevalence of an axial involvement is high (approximately 23%). However, at variance with most data in the literature, we have enrolled subjects experiencing a failure of treatment with traditional DMARDs and with the indication to start a treatment with TNFα blockers. Because of its high rate of refractoriness to DMARDs, the presence of an axial subset has been recognized as a major criterion to start treatment with TNFα blockers (26). The selection criteria used could have determined such a high prevalence of axial subset in our study population.
In addition to obesity, other demographic (male sex) and rheumatologic (CRP level and TJC) variables predicted achieving MDA. The multicollinearity effect was excluded in the present setting by a stepwise approach in which obesity was the strongest negative predictor of the achieving of MDA. Most previous studies in PsA subjects employed different criteria to define good clinical response. We have chosen the criteria of achieving MDA proven to be appropriate in providing an outcome measure for clinical trials (17). As for variables known to predict MDA other than obesity, the HRs found in the present study are in line with those of previous reports (18).
In addition to their obvious pathophysiologic significance, these data may have pharmacoeconomic implications (27). Because of their high cost and their side effects (i.e., higher than normal risk of infections and malignancies) (28), the identification of predictors of treatment success with TNFα blockers is mandatory. By stratifying the achieving of a good clinical response, a risk/benefit evaluation may be performed in each case and for each patient. This, in turn, would hamper the use of TNFα blockers in subjects with a low probability of success. Further studies with a higher number of subjects are needed to address this important issue. In this respect, determining whether, in addition to its effect on weight loss, calorie restriction lowers the levels of inflammatory markers in PsA is a major direction to be pursued to strengthen the interrelationship between obesity and inflammation.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. M. N. D. Di Minno had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. M. N. D. Di Minno, Peluso, Scarpa, G. Di Minno.
Acquisition of data. Iervolino, Lupoli, Russolillo.
Analysis and interpretation of data. M. N. D. Di Minno, Peluso, G. Di Minno.