Article Text

Download PDFPDF

Extended report
Long-term work disability in patients with psoriatic arthritis treated with anti-tumour necrosis factor: a population-based regional Swedish cohort study
  1. Lars Erik Kristensen1,2,3,
  2. Martin Englund1,4,
  3. Martin Neovius5,
  4. Johan Askling5,
  5. Lennart T H Jacobsson6,
  6. Ingemar F Petersson1,2,3
  1. 1Department of Orthopedics, Clinical Sciences Lund, Lund University, Lund, Sweden
  2. 2Department of Clinical Sciences, Lund, Section of Rheumatology, Lund University, Lund, Sweden
  3. 3Department of Rheumatology, Skåne University Hospital, Lund, Sweden
  4. 4Clinical Epidemiology Research & Training Unit, Boston University School of Medicine, Massachusetts, USA
  5. 5Department of Medicine, Clinical Epidemiology Unit, Karolinska Institutet, Stockholm, Sweden
  6. 6Department of Inflammation Research and Rheumatology, The Sahlgrenska Academy at Gothenburg University, Gothenburg, Sweden
  1. Correspondence to Dr Lars Erik Kristensen, Department of Rheumatology, Skåne University Hospital, Lund SE-221 85, Sweden; larserik_kristensen{at}yahoo.com

Abstract

Objective To study long-term work disability before and after tumour necrosis factor (TNF)-antagonist therapy in patients with psoriatic arthritis (PsA).

Methods Using the population-based South Swedish Arthritis Treatment Group Register, we identified 191 patients with PsA (median age 43 years, range 18–58 years, 54% men), who between January 2003 and December 2007 started treatment with adalimumab, etanercept or infliximab. We linked data to the Swedish Social Insurance Agency and calculated the proportion of work disability in 30-day intervals from 12 months before the start of treatment until 3 years after. For each patient with PsA we randomly selected four matched reference subjects from the general population.

Results At treatment initiation 67% of the patients with PsA were work disabled—that is, either on sick leave (41.5%) or receiving a disability pension (25.3%). Patients sustaining treatment were, on average, work disabled 12.5 days a month at treatment initiation declining to 10.6 days a month after 3 years of treatment. Patients for whom the first treatment course failed were work disabled 16.5 days at treatment start decreasing to 15.6 days after 3 years. The background population were 2.5 days and 3.0 days off work each month, respectively. Regression modelling identified prior work disability status, anti-TNF treatment failure, higher age, female gender and longer disease duration as significant predictors of working disability.

Conclusions There was a decline in net work disability after initiation of anti-TNF treatment in patients with PsA. Patients withdrawing from treatment had a 50% increased risk of being work disabled. Prior work disability, higher age, female gender and longer disease duration were also associated with long-term work disability.

  • Psoriatic Arthritis
  • TNF-alpha
  • Anti-TNF
  • Economic Evaluations
  • Epidemiology
View Full Text

Statistics from Altmetric.com

Introduction

Psoriatic arthritis (PsA) is a multifaceted disease associated with psoriasis in the skin or nails, chronic peripheral and/or axial arthritis, enthesopathy and dactylitis.1 The disease affects around 0.25% of the adult population2 and often causes substantial functional impairment and reduced health-related quality of life.3 Reduced work ability leading to sick leave (SL) is a common consequence of PsA3–5 and may contribute to psychosocial and adverse economic consequences for the individual and for society.4–6

During the past decade the development of anti-tumour necrosis factor (TNF) therapy has been a major advance in the treatment of PsA, having a significant effect on many aspects of this disease, including skin and nail manifestations, spondyloarthritis, peripheral arthritis and enthesitis.7–12 However, anti-TNF treatment is costly, as is the burden of the disease, and thus awareness of health economic calculations has increased. The proportion of work-disabled patients with PsA has been reported to be around 40%.6 Only a few studies to date have focused on the impact of anti-TNF treatment on work disability among patients with PsA.4 ,5 These studies indicate that there might be an increase in productivity among treated patients, but no substantial effect on the degree of total work disability.4 ,5 The studies mentioned have limitations, relying on self-reported work status and selected randomised controlled trial cohorts, and it is hard to have a clear picture of the effect of anti-TNF treatment on work disability in PsA.

The aim of this study was to assess the extent of SL and receipt of a disability pension (DP) in patients with PsA before and after the start of adalimumab, etanercept or infliximab therapy (referred to as anti-TNF treatment). Moreover, treatment failure and other potential factors with a possible effect on working status were studied. The study is based on patient data from a regional anti-TNF treatment register13 and information on SL and DP retrieved from an independent source, the Swedish Social Insurance Agency (SSIA) via the Epi-centrum Skåne. The results were matched with those for subjects randomly drawn from the general population.

Methods

The PsA cohort was identified as patients with active PsA, starting anti-TNF therapy for the first time between January 2003 and December 2007, thus allowing for 4 years of follow-up, ranging from 1 year before anti-TNF treatment to 3 years after initiation of treatment. The patients were prospectively monitored as part of the South Swedish Arthritis Treatment Group (SSATG) Register cohort, as previously described.13 The coverage of the Register has been reported to be >90% of all treatments for patients with rheumatic arthritis starting treatment with biological agents.13 ,14 Patients eligible for inclusion had a clinical diagnosis of PsA requiring a first course of anti-TNF treatment. A review of 50 patients treated at University hospital of Skane showed that 92% (n=46) fulfilled the CASPAR (ClASsification criteria for Psoriatic ARthritis) criteria at study inclusion.15 Previously, the diagnosis of the patients had been validated according to European Spondyloarthropathy Study Group (ESSG) criteria for spondyloarthropathy.16 Ninety-four per cent of the studied patients in this cohort fulfilled the ESSG criteria.17 Also the pattern of joint distribution was validated and found to be consistent with that of other cohorts.18 ,19 Thus around 6% of the patients had spondylitis only, 55–60% had peripheral arthritis only and about 35–40% of the patients had both spondylitis and peripheral arthritis.17

Anti-TNF therapy in the patients with PsA included in the SSATG Register was initiated according to national guidelines. In accordance with international recommendation a conventional disease-modifying antirheumatic drug (DMARD) was given before the start of anti-TNF therapy unless the patient had predominantly axial disease.7 Glucocorticoids were administered systemically or by local injections before or during the study period and concomitant use of non-steroidal anti-inflammatory drugs or non-biological DMARDs was allowed.

At the start of anti-TNF therapy, baseline characteristics were reported by the treating doctors using a standardised protocol. These included information on demographics, disease duration, symptoms, data on past and present antirheumatic treatment, results of laboratory examinations and clinical variables allowing the calculation of disease activity using the validated Swedish version of the Health Assessment Questionnaire.20 Furthermore, results of self-scored visual analogue scales for pain (VAS pain) and general health (VAS global) were also reported, together with the evaluator's global assessment of disease activity on a five-grade Likert scale. Anti-TNF therapies were administered according to standard clinical practice. The average dosage of infliximab after 6 months was about 4.5 mg/kg every 8 weeks.

Only patients of working age (18–59 years) and with a potentially complete follow-up period of 1 year before treatment and 3 years after the first treatment course were included. A total of 121 patients of the original cohort were excluded: 61 patients were excluded owing to age, nine patients missed valid follow-up recordings and 51 patients were enrolled after December 2007. None of the patients included died during the study. Using the Swedish population register, which includes information on date of birth, sex and residential address of all Swedish residents, we randomly selected four subjects for each patient with PsA matched for age, sex, area of residence and months of study inclusion. This reference cohort represents the background population in this study. The cohort of patients and the background population were subsequently cross-linked to data from the SSIA and we calculated the proportion of patients with ongoing SL or receiving a DP in 30-day intervals from 12 months before treatment start until 12 months after.21

The SSIA, which provides financial protection for individuals of working age affected by illness or injury, covers everyone who legally lives or works in Sweden regardless of employment status. We defined SL as days with sick pay or sickness benefit of any degree paid by the SSIA (all SL of ≥15 days is continuously and prospectively registered). Sick pay for shorter periods of inability to work (≤14 days) is paid by the employer and therefore not included in this study. A DP is defined as a permanent social benefit paid by the SSIA. All kinds of work disability for both the patients with PsA and the background population, irrespective of the cause, were included. In Sweden people can work part time (25%, 50% or 75%) and receive sick pay or DP for the other part. Therefore, the net amount of time (number of days out of 30-day periods) the patients with PsA and the matched background population were work disabled during the observed period was also calculated.

Cross-linking data from the SSIA with healthcare data was approved by the local ethical committee in Lund (No 514/2007).

Statistics

We present baseline characteristics of patients starting anti-TNF treatment as medians with IQR unless otherwise stated. Baseline clinical characteristics were analysed by Mann–Whitney U test for comparison of groups for continuous variables, whereas a χ2 test was used for categorical variables. McNemar's test was used to compare the average proportion of patients with PsA on SL at the start of anti-TNF therapy with the proportions 3 years after treatment initiation. The relative risk of being on SL compared with the background population including the 95% CI was calculated using crude proportions.

Predictors of work disability status after 3 years of observation were modelled using multivariable binary logistic regression modelling. After backward deletion of covariates with p>0.10, the final regression model consisted of baseline work disability, treatment continuation, disease duration, age and number of previous DMARDs. Two-tailed p values <0.05 were considered statistically significant.

Results

During the study period, 191 biologically naïve patients aged 18–58 years with active PsA started their first anti-TNF inhibitor treatment and were studied for 1 year before treatment and 3 years after the initial treatment. Men constituted 54% of the cohort and median age at inclusion was 43 years, with median disease duration of 7 years. Three years after the start of treatment, 110 (58%) patients continued to receive the initial treatment. Table 1 displays demographic and clinical characteristics of patients continuing treatment and those who discontinued the first treatment course. At baseline patients sustaining treatment over the 3-year observational period had significantly lower median Health Assessment Questionnaire scores (0.75, IQR 0.38–1.13 vs 1.00, IQR 0.63–1.25; p=0.02), higher C-reactive protein levels (10.0, IQR 3.3–25.5 vs 5.0 IQR 1.25–13.0; p=0.01) and longer disease duration (8.6, IQR 4.1–15.1 vs 5.8, IQR 2.7–10.6; p=0.02). Other baseline characteristics were similar between the two groups. Of the 81 patients for whom the first treatment failed, 13 stopped biological treatment altogether, while 68 patients went on to receive a second or third (15 patients) treatment course during the 3-year follow-up period. There were no significant baseline differences between patients stopping treatment and patients switching to subsequent anti-TNF therapies (data not shown).

Table 1

Demographic and clinical characteristics at start of anti-TNF therapy split according to patients who sustain treatment throughout the 3-year observational period and patients who withdraw from first treatment course

Any degree of work disability

Figure 1 shows the prevalence of any degree (binary) of SL and DP, for periods of 30 days for patients with PsA and the background population during the 1-year period before and 3 years after initiation of anti-TNF treatment. At the start of treatment, 67% of the patients with PsA were either on SL (41.5%) or receiving a DP (25.3%). During the first 12 months after the start of treatment this fraction dropped to 58% (SL 29.5%; DP 28.6%) and finally, 62.5% (SL 20.2%; DP 42.3%) after 3 years. Thus there was a significant decline in SL (p<0.001) from baseline to the end of the study period, whereas receipt of a DP increased significantly (p<0.001). Comparing patients with PsA with the general population, the relative risk of being work disabled at the start of treatment, at 12 months and 3 years after the start of treatment was 4.2 (95% CI 3.4 to 5.1), 3.7 (3.0 to 4.5) and 4.1 (3.4 to 5.1), respectively. The proportion of patients with PsA receiving a DP steadily increased throughout the study period, initially constituting 38% of the total work disability increasing to 68% after 3 years of treatment (p<0.001). Only three patients with any level of DP returned to work after 3 years of follow-up.

Figure 1

Proportion (with 95% CI) of patients with psoriatic arthritis (PsA) and matched background population with any degree (yes/no) of sick leave (SL) or disability pension (DP), in the period of 1 year before anti-tumour necrosis factor (TNF) treatment until 3 years after initiation of treatment (n=191).

Absolute amount of work disability

Sixty (47%) of the work-disabled patients were part time work disabled at baseline. The pooled fractions of net SL and DP (days out of 30-day periods) for the entire group of anti-TNF treated patients were 47% at treatment initiation, declining to 42% after 3 years of treatment (p<0.001). The corresponding numbers for the background population were 8.2% with work disability increasing to 10.0% after 3 years. When patients with PsA were compared with the general population the relative risk of being work disabled at the start of treatment and 3 years after the start of treatment was 5.7 (95% CI 4.3 to 7.5) and 4.1 (3.1 to 5.4). Figure 2 displays the number of net working disability days (either DP or SL) per 30-day periods for patients with PsA split according to treatment continuation and discontinuation together with matched background population, in the period of 1 year before anti-TNF treatment until 3 years after initiation of treatment. Patients sustaining treatment were on average work disabled 12.5 days out of 30 days at treatment initiation declining to 10.6 days out of 30 days after 3 years of treatment. On the other hand, patients for whom the first treatment course failed were work disabled 16.5 days out of 30 days at treatment start decreasing to 15.6 days out of 30 days after 3 years. The corresponding numbers for the background population were 2.5 days and 3.0 days out of 30 days, respectively.

Figure 2

Number of net working disability days (either disability pension or sick leave) per 30-day periods (with 95% CI) for patients with psoriatic arthritis (PsA) split according to treatment continuation and discontinuation, together with the matched background population, in the period of 1 year before anti-tumour necrosis factor (TNF) treatment until 3 years after initiation of treatment (n=191).

Based on crude proportions, a comparison of patients withdrawing from treatment with patients sustaining treatment showed that the relative risk of being work disabled tended to be already raised at the start of treatment 1.3 (0.99 to 1.78), and after the 3-year observational period there was a significant relative risk of 1.5 (95% CI 1.05 to 2.03).

Clinical response

The difference in decreased level of work disability during anti-TNF treatment between patients sustaining treatment and those who stopped treatment is reflected by the clinical American College of Rheumatology (ACR) response. At 3 months of follow up 74% of patients (n=85) sustaining treatment achieved an ACR20 compared with 55% of patients (n=53) stopping treatment (p=0.02). An ACR50 response was achieved in 57% and 38%, respectively (p=0.03), and an ACR70 response was achieved in 32% and 11%, respectively (p=0.01).

Regression analysis

In order to search for predictors of long-term work disability a multivariable binary logistic regression model was computed based on any degree of work disability (SL or DP yes/no) 3 years after the start of the first anti-TNF therapy. Table 2 shows OR and 95% CI for the predictors studied. Prior work disability status (yes/no) at treatment start appeared to be a strong predictor of long-term work disability with an OR of 25.3 (95% CI 9.8 to 65.5; p<0.001). Likewise withdrawal from treatment was also associated with work disability after 3 years with an OR of 4.3 (95% CI 1.7 to 11.2; p=0.003). Each additional year of age (OR=1.1, 95% CI 1.0 to 1.1; p=0.02) and each additional year of disease duration (OR=1.1, 95% CI 1.0 to 1.2; p=0.02) were also significantly associated with being work disabled. On the other hand, male sex was associated with a decreased risk of being work disabled after 3 years of anti-TNF treatment (OR=0.4, 95% CI 0.2 to 0.9; p=0.04). A subgroup analysis of patients not previously receiving a DP consistently showed treatment failure and prior SL to be strong predictors of 3-year work disability, while the other baseline predictors showed the same trends, although non-significantly (data not shown).

Table 2

OR, 95% CI and level of significance for the risk of being work disabled 3 years after the start of anti-tumour necrosis factor treatment based on multivariable logistic regression modelling

Furthermore, a univariate regression analysis using ACR50 response at 3 months showed a highly significant reduced risk of being long-term work disabled in patients who had a positive treatment response (OR=0.2; 95% CI 0.1 to 0.5; p<0.01).

Discussion

This population-based cohort study provides evidence of a significant and sustained decrease in SL among patients with PsA after starting anti-TNF therapy. In contrast, the fraction receiving a permanent DP increased significantly during the 3-year observational period with no significant effect of anti-TNF therapy. Total work disability—that is, the sum of DP and SL, decreased slightly during the study period; when only crude work disability (any degree; yes/no) was taken into account this decrease was not evident. Total work disability decreased in patients tolerating treatment compared with patients who stopped treatment. The different effect on work disability status was also associated with significant differences in clinical response according to ACR20, ACR50 and ACR70. Moreover, favourable ACR50 response at 3 months significantly reduced the risk of being long-term work disabled. Comparison with the background population showed a fivefold increased risk of being work disabled in patients with PsA at the start of anti-TNF treatment, declining to fourfold after 3 years of treatment. Prior working status, treatment withdrawal, higher age, female gender and longer disease duration were identified as significant predictors of work disability after 3 years.

Our findings are in agreement with earlier reports from the British Society for Rheumatology Biologics Register.6 In that study, which was based on repeat cross-sectional questionnaire reporting, around 40% of the patients with PsA were work disabled, corresponding well with our data which showed a net work disability of 47% at treatment initiation declining to 42% after 3 years. In contrast to our results the study by Kavanaugh et al5 indicates that there might be an increase in productivity among anti-TNF-treated patients, but no substantial effect on the degree of total work disability. However, owing to the differences in study design, the heterogeneity of populations studied, and the diverse social security systems, comparisons between countries should be interpreted with care. In agreement with our findings for patients with PsA, a recent publication of an anti-TNF-treated cohort with ankylosing spondylitis from southern Sweden showed a fourfold relative risk of being work disabled compared with matched reference subjects after 1 year of treatment.22

Notably, anti-TNF therapy did not have any significant effect on receipt of a permanent DP, which increased during the observation period. This finding, in combination with the increased level of long-term work disability in patients with longer disease duration and higher age, suggests that anti-TNF agents should preferably be used in an earlier phase of the disease, as longer disease duration and higher age seems to result in an irreversible stage of work disability.

Periods of sickness of employees shorter than 15 days are not registered by the SSIA and hence not included. However, it is reasonable to assume that short periods of SL in most cases might be attributable to intermittent infections or injuries, thus being somewhat more evenly distributed between patients with PsA and the general population. In addition, the study does not yield data on potential productivity of patients available at the working market, and thus presenteeism, although hard to measure objectively, as a possible confounder of the results cannot be estimated.

The open, observational setting of this study naturally has limitations such as possible bias in patients’ assignment to treatment and collection of clinical data, possibly inflating the effect size of clinical responses.23 ,24 In addition, information about potential confounders such as education, employment and socioeconomic status of the studied patients was not available and only indirectly controlled for by matching by area of residence. However, the work disability data, which is the primary study outcome, were collected from an independent and external source and are of high quality as they are linked to the payment system. It should be noted that the results obtained are not valid for patients with mild PsA or severe comorbidities preventing the use of anti-TNF therapies, thus reducing the external validity of the study. Also, having no control cohort with similar disease severity, but receiving conventional treatment only, limits the possibility of evaluating effects of confounding by indication and regression towards the mean. Nonetheless, the continuous decline in SL and changed trajectory of total work disability seen throughout the 3-year period after initiation of anti-TNF therapy remains firm.

In conclusion, we provide evidence that anti-TNF treatment of patients with PsA is associated with a decline in net work disability sustained for 3 years. When any degree of work disability was crudely calculated and partial work ability was not taken into account, no such decrease was seen. Patients withdrawing from treatment have a 50% increased risk of being work disabled after 3 years of follow up compared with patients continuing to receive treatment. Prior work disability, higher age, female gender and longer disease duration were also associated with long-term work disability.

Acknowledgments

We are indebted to all colleagues and staff in the South Swedish Arthritis Treatment Group for cooperation and data supply.

References

View Abstract

Footnotes

  • Handling editor Tore K Kvien

  • Contributors LEK is guarantor and participated in study design, acquisition of data, draft and revision of the manuscript, analysis and interpretation of data. ME, JA, MN, IFP and LTHJ participated in study design, revision of the manuscript and interpretation of data. All authors read and approved the final manuscript.

  • Funding This study was funded by unrestricted grants from Region Skåne, Skåne University Hospital, the Swedish Social Insurance Agency, the Swedish Research Council, the Faculty of Medicine at Lund University, Österlund and Kock Foundations, the Swedish Rheumatism Association and King Gustav V 80-year fund.

  • Competing interests LEK, IFP and LTHJ have received fees for speaking and consultancy by Pfizer, Wyeth, Abbott, BMS, MSD and Schering-Plough. JA has received speaker's honoraria from Merck. ME and MN have no competing interests.

  • Ethics approval Local ethical committee of Lund Sweden: No. 514/2007.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.