Article Text
Abstract
Objective The Swedish Early Psoriatic Arthritis Register describes the course of early psoriatic arthritis (PsA) in a real life clinical setting in Sweden. The aim of this study was to obtain information on predictors of clinical outcomes over a 5-year period with special focus on effects of gender, joint patterns, diagnostic delay and initial disease activity.
Methods In six centres, patients with signs suggestive of PsA were included in the Swedish Early Psoriatic Arthritis Register within 2 years of symptom onset. CASPAR (classification for psoriatic arthritis) criteria were fulfilled by 197 patients who had passed the 5-year follow-up. Disease activity was measured by the Disease Activity Score including 28 joints (DAS28) and the Disease Activity Index for Psoriatic Arthritis (DAPSA). Remission and minimal disease activity (MDA) were used as outcome measures.
Results Mean age at inclusion was 46 years, younger in male than female patients (43 vs 48 years).
Mean DAS28 was 3.7 and 3.0 at inclusion and 2.8 and 2.1 at follow-up for women and men, respectively—significantly higher in women at both visits. Likewise, DAPSA scores were significantly higher in women. The degree of improvement (change in DAS28 and DAPSA) was similar. Men achieved MDA or remission (50% vs 33%, 25% vs 13%, respectively) more often, and women had significantly more polyarthritis at inclusion (49% vs 27%) and after 5 years (25% vs 15%). Axial or mono/oligoarticular disease was predominant in men. Independent predictors of MDA at the 5-year follow-up were: shorter symptom duration; greater general well-being (global visual analogue scale); and low Health Assessment Questionnaire at inclusion.
Conclusions In early PsA, short delay between onset of symptoms and diagnosis, preserved function, and male gender are the most important predictors of favourable clinical outcome at the 5-year follow-up. Early recognition of PsA and active treatment may be important, particularly in women with polyarticular disease.
- Psoriatic Arthritis
- Outcomes research
- DAS28
- Disease Activity
Statistics from Altmetric.com
Psoriatic arthritis (PsA) is part of the psoriatic disease entity consisting of skin psoriasis, PsA, psoriatic nail disease, dactylitis and enthesitis.1 PsA presents with highly variable patterns of joint and skin involvement and degree of severity. In Southern Sweden, 18% of those with psoriasis have joint disease, and the prevalence of PsA is 0.34% (own unpublished data). Manifestations include the whole spectrum of mild mono/oligoarthritis to very severe, erosive and destructive polyarthritis, as well as spondyloarthritis, enthesitis and dactylitis. Several disease subsets have been defined according to joint involvement. The clinical picture may, however, change over time, either spontaneously or due to the effects of disease-modifying anti-rheumatic drugs (DMARDs).2 Awareness of PsA-associated comorbidity, such as cardiovascular disease, the metabolic syndrome and premature death, is increasing.3
Diagnosis and treatment as early as possible has become the standard of care in rheumatoid arthritis (RA) in order to prevent disease progression, irreversible functional impairment and premature death.4 Only a few studies have focused on early PsA, and thus it is unknown if early detection and treatment of inflammatory peripheral or axial joint involvement is equally important for prevention of destruction, deformity and functional disability. Kane et al2 reported a 2-year follow-up in a cohort of patients with early PsA demonstrating high rates of erosive disease. Gladman et al5 compared patients with up to 2 years disease duration with those of more longstanding disease. They concluded that early disease should be detected and treated actively, as those presenting with longer symptom duration had a higher degree of damage. Queiro-Silva et al6 studied 71 patients with early PsA over a 10-year period and found that 45% had developed erosive or deforming disease at the end of the study. Polyarticular disease at onset, defined as five or more swollen joints, was strongly associated with erosive disease. In general, long-term outcome with regard to premature mortality has improved over calendar periods possibly because of identification of patients with milder or earlier disease and more effective antirheumatic treatment in severe cases.7
We here present a prospective 5-year follow-up of 197 patients with early PsA from the Swedish Early Psoriatic Arthritis Register (SwePsA).8 The aim of this register is to: enable studies on early PsA in the Swedish population; evaluate the course of the disease, socioeconomic issues and workforce participation; identify markers of disease progression; and describe treatment patterns and effects in real life care. The aim of the present study was to obtain information on predictors of clinical outcomes over a 5-year period with special focus on effects of gender, joint patterns, diagnostic delay and initial disease activity. Radiographic outcome and predictors of radiographic progression, as well as impact on health-related quality of life and social functioning will be reported separately.
Methods
Patients
The SwePsA cohort
Patients with early PsA—defined as having arthritis, enthesitis or dactylitis suggestive of PsA in the investigator's opinion, independently of the presence of skin psoriasis—were included in the prospectively followed register between November 1999 and December 2010.9 The patients were referred to one of six rheumatological outpatient clinics, covering all parts of Sweden from Malmö in the south to Umeå in the north, and were assessed on inclusion and after 2, 5 and 10 years according to the programme for the SwePsA.9 Symptom duration at inclusion was at most 2 years. By April 2011, 223 5-year visits had been performed. Finally, 197 patients fulfilling the CASPAR (classification for psoriatic arthritis) criteria were included in the study (figure 1 and online supplementary table S1).
Classification criteria
The most widely accepted classification criteria for PsA are those of the CASPAR criteria set,10 which also seem suitable for diagnosis of early disease.11 Although not available yet when the SwePsa was established, it was applied retrospectively.
Measures of disease activity
Measuring disease activity in PsA can be performed with Disease Activity Score including 28 joints (DAS28), which is widely used in RA, but criticised for being too restricted because of not including feet, distal interphalangeal joints, enthesitis, dactylitis, axial disease or skin disease.12 This is improved by the Disease Activity Index for Reactive Arthritis (DAREA)/Disease Activity Index for Psoriatic Arthritis (DAPSA) score proposed by Schoels et al,13 which is calculated as follows: DAREA/DAPSA=swollen joint count of 66 joints+tender joint count 68+patient global assessment visual analogue scale (VAS) (in cm)+pain VAS (in cm)+C-reactive protein (CRP) (in mg/dl). Mumtaz et al proposed the more comprehensive Composite Psoriatic Disease Activity Index (CPDAI)14 including all relevant items from the psoriatic disease spectrum. In the present investigation, DAS28 and DAPSA were applied; CPDAI could not be calculated because of several missing variables. We have no separate activity measure for patients with axial disease. The DAPSA covers patient global, patient pain and CRP from the Ankylosing Spondylitis Disease Activity Score,15 but includes no direct information on back pain and function.
Outcome measures
Treat to target requires definition of the optimal treatment goal. Complete remission or Minimal Disease Activity (MDA) are plausible outcomes. We defined our criteria for remission in our previously published 2-year follow-up8 as absence of any swollen or tender joints as well as erythrocyte sedimentation rate (ESR) <20 mm during first hour and CRP <0.5 mg/dl. In this paper, in axial disease, remission also required absence of signs of axial enthesitis and low-level (none or minimal) joint/back pain. Coates et al16 proposed the following definition for MDA in PsA: ‘A patient is classified as achieving MDA when meeting 5 of the 7 following criteria: tender joint count ≤1; swollen joint count ≤1; Psoriasis Activity and Severity Index (PASI) ≤1 or body surface area ≤3; patient pain VAS ≤15; patient global disease activity VAS ≤20; Health Assessment Questionnaire (HAQ) ≤0.5; tender entheseal points ≤1’.
Assessment of clinical disease, disease activity and outcome measures
At each visit, counts for tender and swollen joints, dactylitis, deformities (ankylosis, subluxation and/or loss of function), enthesitis (clinical diagnosis supported by imaging when necessary, at shoulders, elbows, hips and heels), tenosynovitis, presence of active axial disease, PASI score and presence of nail psoriasis were reported. Patients were classified according to proposed joint pattern subgroups.2 Polyarthritis was defined as five or more inflamed (swollen or tender) joints. Axial disease required the presence of typical inflammatory back pain in combination with clinical signs of axial enthesitis or sacroiliitis. Radiographic confirmation was not required. Participants also filled in HAQ and Medical Outcome Survey Short-Form (SF)-36 Quality of Life instrument, questionnaires on work participation and comorbidities, as well as VAS scales for global health, pain and skin disease. DAS28 and DAPSA were calculated. Both remission and MDA were assessed at inclusion and 5-year follow-up. Antirheumatic medication was reported.
Laboratory measurements and radiographs
Markers of inflammation (ESR, CRP) were measured at each visit, and rheumatoid factor (RF) at inclusion. Radiography was performed at baseline (and reinvestigated at follow-up visits) of hands and feet for patients with polyarticular disease, of spine and pelvis in axial disease, and of affected joints in those with mono/oligoarticular disease or mixed patterns.
Statistical analysis
Descriptive statistics were used to report demographic and clinical features. Comparisons between inclusion and 5-year follow-up were performed with paired t tests. Subgroups of patients were compared by χ2 tests and t tests. Correlation between measures and variables was assessed by Spearman rank test. To analyse predictors of remission or MDA, age-adjusted logistic regression analysis for relevant explanatory variables was performed. For the multivariate analysis, among the statistically significant variables from model one, those with the highest explanatory potential were depicted while avoiding highly correlated variables (so-called multicollinearity) in the same model. Therefore composite indices were excluded as predictors from the multivariate model. The low number of MDA events further limited the number of variables in the multivariate subanalyses, separately performed for women and men. Results were expressed as OR with 95% CI. Gender was strongly associated with most other variables, but was kept in the multivariate model as a basic biological factor similar to age.
Results
Patient characteristics
Only patients fulfilling the CASPAR criteria (114 women, 83 men) are included in the statistical analysis. Age at inclusion was significantly lower in men (43.1±13.4 vs 48.3±15.2 years, p=0.004). Diagnosis before the age of 45 occurred in 55% of men compared with 41% of women (p=0.049). Symptom duration before inclusion was similar in men and women (11.6 vs 10.6 months (ns)). RF was positive in 3.7% of women and 8.8% of men (ns). Table 1 gives detailed information on core outcome measures at baseline and 5-year follow-up for women and men.
Patterns of joint involvement at inclusion and follow-up
Owing to the mean symptom duration of 11 months before inclusion, three patients (1.5%) were in spontaneous remission at inclusion. Another seven (3.6%) had only slightly elevated biochemical inflammatory activity.
Women and men presented with different distribution of inflamed joints. The predominant female pattern was polyarthritis versus mono/oligoarthritis in men. As expected, axial disease, either isolated or in combination with peripheral joint disease, was more prevalent in men.17–19 Subclinical axial disease may have been missed, as only symptomatic joints were subjected to x-ray analysis. At the 5-year visit, twice as many men as women had achieved remission (25.3 vs 12.5%, p=0.022). Also, MDA was predominantly present among men (18.2% vs 7.7% at inclusion, p=0.023, and 50% vs 32.7% at follow-up, p=0.017). The persisting high prevalence of polyarticular disease in women after 5 years was caused by the persistently high tender joint count, while swollen joint counts were low in both genders at that time point (figure 2 and see online supplementary table S2). One woman had chronic dactylitis at inclusion, and, after 5 years, nine women and six men had developed chronic dactylitis or ankylosis. There was no gender difference in the frequency of joint deformities.
Measures of inflammation and functional impairment at inclusion and follow-up
Table 1 reports in detail results of measurements from single items and composite indices related to inflammation and physical function. Most signs and symptoms as well as composite indices improved over the 5 years of observation. The HAQ was only slightly abnormal at inclusion, resulting in limited potential for improvement. Beside the HAQ in both genders, only tender joint count in women did not show significant improvement.
Gender differences
With the exception of tender (at inclusion only) and swollen joint counts, all variables and composite indices showed worse results in women than men. Persistent high tender joint count in women resulted in persistence of polyarticular pattern, and significantly higher disease activity measured as DAS28 and DAPSA. HAQ was significantly higher in women and, in contrast with men, above the cut-off for normal function of 0.5 at both visits. Also MDA and remission were achieved less often in women, partly because of the persistence of the high tender joint count. The pain level measured as VAS improved significantly in both genders. The swollen joint count improved more in women than in men. ΔDAS28 and ΔDAPSA were comparable between men and women (table 1). No difference in the presence of dactylitis and peripheral enthesitis was documented between women and men.
Correlations between measures of disease activity and function
DAS28 and DAPSA correlated well at both inclusion and follow-up (Spearman ρ 0.783 and 0.721, respectively). Both DAS28 and DAPSA correlated well with HAQ (Spearman ρ 0.586 and 0.675 at inclusion and 0.556 and 0.636 at 5 years). DAS remission was achieved by 54% at the 5-year follow-up (corresponding DAPSA: mean 6.32±5.23), low disease activity by 21% (corresponding DAPSA: mean 12.84±6.14), moderate disease activity by 20% (corresponding DAPSA: mean 21.03±9.12), and 5% had high disease activity (DAPSA: mean 42.15±14.47). DAS–remission correlated well (p<0.001) with our stricter definition of remission (101 vs 34 patients, 32 overlapping, κ 0.28), but is more similar to the concept of MDA16 (101 vs 72 patients, 60 in common, κ 0.46).
Medication
For SwePsA patients, treatment decisions are made in accordance with standard of care, usually following national Swedish guidelines. Sixty-one per cent had been treated with DMARDs at some time. Thirty-three (16.8%) were receiving methotrexate continuously from the first to last visit. Biological agents (etanercept, adalimumab, infliximab, rarely tocilizumab) were used equally in women and men (table 2). After exclusion of patients with pure axial disease, women and men who received DMARD treatment had a mean DAS28 baseline of 4.09 (DAPSA 27.01) and 3.28 (DAPSA 18.28), respectively (p=0.001). The use of DMARD or biological agent was predicted by higher baseline levels of DAS28, DAPSA and polyarticular disease (data not shown). Thus patients with presumably good prognosis were treated less aggressively, which may explain the lack of association between medication and MDA or remission at 5 years.
Predictors of MDA or remission after 5 years
MDA16 was present in 22 (11.3%) patients at baseline and in 75 (40.1%) at follow-up; 15 (8.1%) had sustained low disease activity at both visits. Only 35 (17.9%) achieved remission (no swollen or tender joints and normal ESR and CRP). In univariate analysis, factors at inclusion associated with a state of MDA after 5 years were male gender, short delay between symptom onset and inclusion, the presence of axial disease, a low swollen and tender joint count, lower disease activity measured by DAS28 or DAPSA, lower patient global VAS and pain VAS, low HAQ and presence of MDA at inclusion. Presenting with polyarthritis reduced the chance of MDA.
In multivariate analysis, shorter delay between symptom onset and inclusion and low HAQ at inclusion were independent predictors of MDA. Age at inclusion, ESR or CRP, PASI score, presence of dactylitis, enthesitis or tenosynovitis at presentation, and treatment with non-steroidal anti-inflammatory drugs, DMARDs or biological agents during the observation period did not influence outcome (data except for age and medication not presented). In women, only symptom duration before inclusion and low HAQ predicted MDA. In men, low pain VAS was also predictive of MDA (table 3).
Remission according to our strict definition was only predicted by male gender in univariate analysis. However, only 35 patients achieved remission, giving very limited power for the statistical evaluation (table 4).
Discussion
We present a longitudinal 5-year follow-up of 197 prospectively followed patients with early PsA from the SwePsA.
In general, early PsA showed a favourable course. Most variables related to disease activity improved. However, while 54% achieved DAS28 remission, only 40% reached MDA and 18% remission according to our strict definition (see above); in all cases, significantly more men did so. DAS28 does not take into account typical features of PsA, and DAS28 remission does not seem an adequate outcome measure compared with the more comprehensive PsA-specific MDA index. The PsA-specific and simply calculated composite disease activity index, DAPSA, is applied for the first time to an early PsA cohort. While correlating well with DAS28, it seemed somewhat more useful as a prognostic marker in relation to MDA. A mean DAPSA level of 6.0 and 4.6 corresponded to MDA at inclusion and 5-year follow-up, respectively. A mean DAPSA of 9.74 and 6.31 was found in patients with DAS28 remission at inclusion and 5-year follow-up, respectively. We would propose DAPSA levels of 5 as the treatment goal indicating acceptable low-disease activity state in PsA trials. DAPSA does not take into account the impact of skin disease and does not fully cover axial disease. A more holistic disease activity instrument should be used when these aspects of the psoriatic disease are to be addressed, especially in prospective intervention trials.
An important message of our study is the more severe presentation and course in women. Not surprisingly, men more often have axial disease, a predictor of MDA in crude univariate analysis. Women have higher disease activity and more joint pain and functional impairment. They more often have persistent polyarticular disease. Independently of chosen outcome, male gender was associated with a more favourable course after 5 years. The persisting higher disease activity in women was mainly due to lack of improvement in tender joint counts and not to persisting high levels of general pain (VAS) or global lack of well-being. We interpret this finding as possible persisting specific joint-related inflammation rather than a general gender-related widespread pain problem or fibromyalgia. It might therefore be treatable by intensified therapy. A gender-related propensity for higher pain levels cannot be excluded, and, before intensified anti-inflammatory therapy is argued for, the inflammatory origin of this joint pain should be confirmed by ultrasound or MRI. An analysis of the contribution of tender joint count to the radiographic progression may clarify its importance for treatment strategies. Even if non-inflammatory, pain contributes to decreased quality of life and functioning and must not be ignored.
Another important finding of our study is the effect of delay before specialist care on long-term outcome. Shorter symptom duration before inclusion independently predicted better outcome in both men and women. This finding emphasises the need of strategies for early recognition and referral of patients with PsA to rheumatology units. Disease activity was similar in patients included within 12 months of symptom onset or between 12 and 24 months after onset (data not shown). Thus the effect of short duration does not seem to be due to milder or self-limiting disease in those with shorter delay. Furthermore, this finding is supported by results from Toronto, where disease progression was more marked in those presenting with more than 2 years of disease.5 It also mirrors findings in patients with early RA, in whom longer disease duration and lower HAQ are predictors of unfavourable outcome.20
There are no previous reports analysing longitudinal clinical outcomes, such as remission and MDA, in early PsA. In agreement with our data, Gladman et al5 have demonstrated the importance of early care for radiographic progression. In consensus with Queiro-Silva et al,6 we confirm the negative prognostic value of polyarticular onset. As in their work, in our study, polyarticular disease leading to high DAPSA, often associated with female gender, is a negative predictor. While older age at onset or dactylitis was not associated with adverse clinical outcomes, effects on radiographic progress cannot be excluded, and will be studied and presented separately.
In conclusion, our results describe an overall improvement in PsA during the first 5 years of follow-up. However, only a minority of the patients achieve MDA or remission. Shorter symptom duration, male gender and axial instead of polyarticular disease as well as preserved function and subjective well-being at entry predict favourable outcome. Awareness of PsA should be increased for earlier referral, and women with polyarticular disease particularly may need aggressive treatment.
Acknowledgments
We thank Professor Björn Svensson for initiating the SwePsA, Dr Gunilla Holmström for earlier fruitful cooperation and support for SwePsA, and biostatistician Jan-Åke Nilsson for his professional help in performing the statistical analyses.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Tore K Kvien
-
Contributors ET, TH, GMA, PTL, AT and URCL contributed to the study design and acquisition and reporting of clinical data. MG was involved in planning, performance and reading of x-rays. All authors contributed to manuscript preparation and critical revision.
-
Funding The study was supported by independent grants from the Swedish Psoriasis Foundation.
-
Competing interests None.
-
Ethics approval The ethic's committee at Uppsala University, Sweden approved the register as a research register.
-
Patient consent Obtained.
-
Provenance and peer review Not commissioned; externally peer reviewed.