SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Objective

To explore the impact of an early treatment response on maintenance of work capacity in patients with early, active rheumatoid arthritis (RA).

Methods

In the Finnish Rheumatoid Arthritis Combination Therapy trial, 195 patients with recent-onset RA were randomized to receive either a combination of disease-modifying antirheumatic drugs (DMARDs) with prednisolone or a single DMARD with or without prednisolone for 2 years. Treatment responses were evaluated according to the American College of Rheumatology (ACR) criteria. After a 5-year followup, the cumulative number of days of sick leave and RA-related permanent work disability was calculated for each of the 162 patients who were available for the active work force at baseline.

Results

Of the 159 patients assessed at 6 months, 29 were in clinical remission, 66 achieved an ACR50 response but not remission, 29 achieved an ACR20 response but not an ACR50 response, and 35 failed to achieve an ACR20 response. In these 4 groups, the median numbers of work disability days per patient-year from 6 months through 60 months of followup were 0 (interquartile range [IQR] 0–3), 4 (IQR 0–131), 16 (IQR 0–170), and 352 (16–365), respectively (P < 0.001). Pairwise multiple comparisons showed a statistically significant difference between all groups except the ACR50 and ACR20 groups. At 12 months, 30 patients were in remission. None of the 44 patients in remission at 6 or 12 months became permanently work disabled over the 5-year followup, as compared with 15 patients in the ACR50 group (23%), 6 in the ACR20 group (21%), and 19 without an ACR20 response at 6 months (56%).

Conclusion

Prompt induction of remission translates into maintenance of work capacity. At 6 months, an ACR50 response is no better than an ACR20 response with regard to future productivity, while failure to achieve an ACR20 response carries a high risk for work disability.

Work disability is the most expensive consequence of rheumatoid arthritis (RA) because of reduced earnings of the individual and reduced productivity of society (1–3). We have reported that initial aggressive treatment with a combination of 3 disease-modifying antirheumatic drugs (DMARDs), as compared with treatment with a single DMARD, reduced the cumulative number of work disability days per patient-year and lowered the costs to society (4). The patient's and physician's global assessments of RA severity and the self-reported physical function (Health Assessment Questionnaire score) were the only statistically significant disease-related factors that predicted sick leave and permanent work disability at baseline (5). Remissions were more frequent in the combination-treatment arm, but some patients did very well on single-DMARD therapy, as well (6). In the present study, we analyzed the impact of an early treatment response in RA on maintaining the capacity to work.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Study design.

From April 1993 through May 1995, 195 DMARD-naive patients with recent-onset (disease duration <2 years), active RA who fulfilled the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) 1987 criteria (7) started a multicenter, parallel-group, randomized study comparing the efficacy and tolerability of therapy with a combination of DMARDs (simultaneous sulfasalazine, methotrexate, hydroxychloroquine; 97 patients) plus prednisolone with the efficacy and tolerability of therapy with a single DMARD (initially, sulfasalazine, which was replaced with methotrexate at 6 months if the response was <25%; 98 patients) with or without prednisolone. The study has been described in detail previously (4, 6). After 2 years, the drug treatment strategy was no longer restricted, and 51 of the patients who had been randomized to single-DMARD therapy were eventually treated with a combination of DMARDs. None of our patients received any biologic agents during the 5-year followup.

The patients were assessed clinically at the beginning of the study and at regular followup visits for 5 years (8). Treatment responses were evaluated using the ACR criteria (9). The ACR preliminary criteria for remission (10) were modified to exclude fatigue (a vaguely defined criterion) and duration definition, and a patient might be receiving any drug treatment or no drug treatment. However, a patient in remission had to fulfill all of the other 5 criteria.

At the 5-year visit, the patients were asked for permission to access data on their sick leaves and pensions as recorded in the registries of the Finnish social security system. Of the 162 patients available for the active work force at study entry, 146 gave their permission; 3 patients withheld permission, and 13 were lost to followup. For these latter 16 patients, sufficient information about sick leave and retirement was obtained from their case records, based on the patients' written informed consent that was given at study entry. The cumulative number of days of sick leave (representing work disability of <1 year in the Finnish social security system [for details, see ref.4]) or of RA-related work disability pension (awarded after work disability for >1 year, with few exceptions, equivalent to permanent work disability) plus the cumulative benefits paid by the social insurance system as compensation for the lost income were calculated for each patient until either 5 years of followup or the end of availability for work (without taking RA into account), or until the patient was lost to followup. The number of work disability days was divided by that observation time (in years), and the permanent RA-related disability pensions (equivalent to permanent work disability; n = 40) were recorded (4). Euros were converted to US dollars (€1.00 = $1.22).

Statistical analysis.

The patients were divided into 4 groups according to the clinical treatment response at 6 months, using the ACR response criteria (9, 10) as cut points: remission; ACR50 response with some symptoms remaining; ACR20 response but less than ACR50 improvement; and less than ACR20 response. Descriptive values of the variables are expressed as means with SDs or 95% confidence intervals (95% CIs) and as medians with interquartile ranges (IQRs). CIs for the means of work disability benefits were obtained by bias-corrected and accelerated bootstrapping (5,000 replications) (11). The Cox proportional hazards regression model with robust estimate of variance was used to estimate adjusted risks for RA-related permanent work disability. Statistical comparisons between the response groups were made by the Kruskal-Wallis test with Monte Carlo P values and by the permutation one-way analysis of variance with general scores. The global test for differences was followed by pairwise multiple comparisons with the Dwass-Steel-Chritchlow-Flinger method (12).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Of the 159 patients assessed at 6 months, 29 (18%) were in clinical remission (group I), 66 (42%) achieved an ACR50 response but not remission (group II), 29 (18%) achieved an ACR20 response but not an ACR50 response (group III), and 35 patients (22%) achieved a less than ACR20 response (group IV). Table 1 shows the baseline demographic and clinical variables for these patient groups. Patients in group III had the most-active RA at study entry, while those in group IV had the longest sick leave during the initial 6 months.

Table 1. Baseline demographic and clinical variables and duration of sick leave during the first 6 months of followup of the rheumatoid arthritis patients, by 6-month response group*
VariableRemission (n = 29)ACR50 response, with some symptoms remaining (n = 66)ACR20 response, but less than ACR50 response (n = 29)Less than ACR20 response (n = 35)P
  • *

    Clinical responses were defined as 20% or 50% improvement in disease activity according to the American College of Rheumatology criteria (ACR20 or ACR50). ESR = erythrocyte sedimentation rate; HAQ = Health Assessment Questionnaire; IQR = interquartile range.

Females, %665486510.015
Age, mean ± SD years45 ± 945 ± 944 ± 949 ± 90.040
Duration of education, mean ± SD years12 ± 410 ± 312 ± 310 ± 3NS
Physically demanding job, %48553160NS
Disease duration, median (range) months6 (3–19)6 (3–24)7 (2–23)7 (3–24)NS
Rheumatoid factor present, %83686971NS
ESR, mean ± SD mm/hour34 ± 2039 ± 2241 ± 2437 ± 23NS
No. of swollen joints, mean ± SD12 ± 414 ± 716 ± 713 ± 50.051
No. of tender joints, mean ± SD15 ± 620 ± 1122 ± 1017 ± 60.019
Pain score, mean ± SD39 ± 2052 ± 2650 ± 2344 ± 210.045
Patient's global assessment, mean ± SD37 ± 2052 ± 2551 ± 2244 ± 190.022
Physician's global assessment, mean ± SD39 ± 1648 ± 1849 ± 2046 ± 20NS
Physical function, mean ± SD HAQ score0.6 ± 0.50.9 ± 0.61.1 ± 0.70.8 ± 0.6NS
Larsen score, median (IQR)2 (0–6)2 (0–6)0 (0–4)2 (0–10)NS
Combination treatment, %695031460.036
Sick leave during first 6 months, median (IQR) days20 (0–72)48 (0–142)24 (0–140)182 (16–182)0.001

The improvement at 6 months, as defined by the response category, also reflected the remaining disease activity at that time (Table 2). There was a monotonic trend across the response groups in all activity variables except the ESR. There was no statistically significant difference in the radiologic joint damage at 6 months, but the Larsen score was lowest in the group that was in remission.

Table 2. Clinical variables at 6 months in the 159 rheumatoid arthritis patients, by 6-month response group*
VariableRemission (n = 29)ACR50 response, with some symptoms remaining (n = 66)ACR20 response, but less than ACR50 response (n = 29)Less than ACR20 response (n = 35)P
  • *

    Clinical responses were defined as 20% or 50% improvement in disease activity according to the American College of Rheumatology criteria (ACR20 or ACR50). ESR = erythrocyte sedimentation rate; HAQ = Health Assessment Questionnaire; IQR = interquartile range.

ESR, mean ± SD mm/hour7 ± 59 ± 623 ± 1820 ± 13<0.001
No. of swollen joints, mean ± SD02 ± 25 ± 58 ± 7<0.001
No. of tender joints, mean ± SD04 ± 410 ± 415 ± 10<0.001
Pain score, mean ± SD3 ± 515 ± 1527 ± 1740 ± 23<0.001
Patient's global assessment, mean ± SD4 ± 516 ± 1431 ± 1947 ± 43<0.001
Physician's global assessment, mean ± SD1 ± 411 ± 828 ± 1438 ± 19<0.001
Physical function, mean ± SD HAQ score0.0 ± 0.20.2 ± 0.30.4 ± 0.40.6 ± 0.5<0.001
Larsen score, median (IQR)2 (0–8)7 (2–11)6 (0–10)9 (0–15)0.23

A trend was noted between the response category and the cumulative duration of work disability per patient-observation year from 6 months through 60 months of followup (P < 0.001) (Table 3). Pairwise multiple comparisons showed statistically significant differences between all groups except ACR50 and ACR20 (P = 0.71).

Table 3. Cumulative work disability days, work disability benefits paid by the social insurance system, and percentage of permanent rheumatoid arthritis-related work disability during months 6–60 of followup, by 6-month response group*
Response at 6 monthsNo. of patients (n = 159)Work disability days per year, median (IQR)Work disability benefits per year, mean (95% CI) dollarsNo. (%) of patients with permanent work disability
  • *

    Clinical responses were defined as 20% or 50% improvement in disease activity according to the American College of Rheumatology criteria (ACR20 or ACR50).

  • Per patient-observation year from 6 months through 60 months of followup.

  • Per patient-observation year from 6 through 60 months of followup. The 95% confidence intervals (95% CIs) were obtained by bias-corrected and accelerated bootstrapping (5,000 replications).

Remission290 (0–3)117 (33–353)0 (0)
ACR50 response, with some symptoms remaining664 (0–131)2,877 (1,818–4,315)15 (23)
ACR20 response, but less than ACR50 response2915 (0–170)2,873 (1,556–4,741)6 (21)
Less than ACR20 response35337 (27–365)7,344 (5,300–9,442)19 (54)

From 6 months through 60 months of followup, the social insurance system paid a total of $533,600 in work disability benefits. The mean cumulative benefit per patient-observation year over the same period was $3,355 (95% CI 2,514–4,198). The annual cost for a patient with a poor treatment response at 6 months (group IV) was $7,226 (95% CI 5,124–9,272) more than the cost for a patient in remission. The difference across the groups was statistically significant (P < 0.001) (Table 3).

None of the 29 patients who were in remission at 6 months became permanently work disabled because of RA during the 5-year followup, whereas 19 patients in the group without an ACR20 response became permanently work disabled (54%) (Table 3). If a patient was not in remission at 6 months, the risk for permanent work disability was no better in the ACR50 group than in the ACR20 group. Figure 1 illustrates the age-, sex-, job type–, and education level–adjusted rates of RA-related permanent work disability in the 4 response groups.

thumbnail image

Figure 1. Age-, sex-, job type–, and education level–adjusted risk for rheumatoid arthritis (RA)–related permanent disability pension, by response group at 6 months: group I = clinical remission; group II = American College of Rheumatology 50% response (ACR50) but no remission; group III = ACR20, but not ACR50, response; and group IV = less than ACR20 response. P < 0.001 for difference across groups.

Download figure to PowerPoint

At 12 months, 30 patients were in remission; each of these patients had achieved at least an ACR50 response at 6 months. Over the 5-year followup, none of the 44 patients who were in remission at either 6 months or 12 months experienced RA-related permanent work disability.

In most cases, continuous work disability for >1 year translated into permanent dropout from the work force, but 2 patients receiving single-DMARD therapy and 1 patient receiving combination therapy regained their capacity to work at about 2 years of followup. After 3 years, this was possible only for 2 of the patients who were originally randomized to single-DMARD treatment.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

In this 5-year prospective study of patients with early, clinically active RA (6), we found that the 6-month clinical response predicted future ability to work and, conversely, future loss of productivity. Remission during the first year was an indicator of maintenance of work capacity. Lower risk for work disability associated with early control of RA also has been reported from studies in the US (13). In the present study, failure to achieve an ACR20 response at 6 months strongly predicted permanent work disability and high costs to society, but unexpectedly, achievement of an ACR50 response was, in this respect, no better than achievement of an ACR20 response. Notably, the patients without an ACR20 response dropped out of the work force early and at a fast rate (Figure 1). Differences between response groups in work disability benefits paid were considerable: induction of remission translated into annual savings of about $7,000 per patient for the social insurance system, compared with costs from failure to reach an ACR20 response. This amount of money, however, represents only a fraction of the total value of the productivity saved.

When the Finnish Rheumatoid Arthritis Combination Therapy (FIN-RACo) trial started in 1993, single-DMARD treatment with sulfasalazine was standard therapy. The single-DMARD treatment applied the sawtooth strategy (14), with remission as the target, and if the response with sulfasalazine was <25% at 6 months, a switch to methotrexate was mandatory. Eventually, 62 patients in the single-treatment arm received methotrexate during the 5-year followup. Furthermore, the “aggressive” combination treatment was not actually very vigorous during the initial months. The highest methotrexate dosage of 15 mg/week was not allowed until 9 months and onward. Consequently, the highest number of remissions (n = 54) was achieved at 2 years (6, 8), and 3 patients were able to return to the work force after long-term work disability at that time. After 2 years, an increase in the methotrexate dosage to as much as 25 mg/week was allowed, as was a switch from single-DMARD therapy to treatment with a combination of DMARDs. In most cases, however, it was too late to institute a more intensive treatment: only 2 patients regained work capacity after 3 years.

The plan of the FIN-RACo trial was to compare the 2 treatment strategies, and the study design does not allow comparison of the impact of individual DMARDs on maintenance of work capacity. Further, since biologic therapy was not used, the possible effect of the biologic agents in this respect can only be speculated. A recent study has shown an association between etanercept use and higher employment rates (15). Our data suggest, however, that it is the magnitude of the early response that is important, not the drug or drugs with which the response has been achieved. In the present study, the induction of remission and avoidance of work disability was obtained with very low cost compared with the expensive biologic drugs. The impact of these new agents on the maintenance of work capacity and on societal costs, as well as the optimal schedule for their institution, is important subjects for future studies.

We have previously shown that over the entire 5-year followup period, accumulation of work disability days continued almost linearly, but at different rates, in both treatment arms; after 2 years, there was no slow-down in the accumulation of work disability days in the single-therapy arm (4). Furthermore, 65% of the permanent RA-related disability pensions (equivalent to permanent work disability) were awarded during the first year (4). Thus, both our earlier and present results lend support to the concept of “window of opportunity” during the early course of RA (8, 16, 17). We found the trend we expected between treatment response and duration of sick leave over the preceding 6 months. Prolonged sick leave may impose injurious psychosocial consequences that also partly account for the risk of future work disability (18–20). Initiation of vocational rehabilitation early in the disease course is of great importance (2, 22).

There were differences in some baseline variables between the 6-month response groups. This was not unexpected because, for example, the number of joints with active disease has been shown to be a prognostic factor in many studies of early RA (23), and the disease-related baseline variables anticipated the 6-month treatment response. We adjusted our results for age, sex, job type, and education level to rule out any potential confounding factors.

In terms of days off work and permanent work disability, the ACR50 response group was no better than the ACR20 response group despite the between-group difference in disease activity at 6 months (Table 2). The response category “remission” defines a state of no arthritis, and the ACR50 and ACR20 responses define a change in disease activity. In our cohort, however, the ACR response category was a good surrogate for the remaining disease activity. No “safe” range for RA activity appeared: without remission, the risk of temporary and permanent work disability rose considerably. There are many possible interacting causes of these phenomena: disability associated with the remaining inflammation, permanent joint damage (which develops most rapidly during the first months of RA) (Table 2) (24), the aforementioned psychosocial consequences of prolonged sick leave, and the opportunity for compensation for days off work provided by the social insurance system.

Our method of counting days on sick leave and on a work disability pension is a different, albeit more accurate and more sensitive, indicator of the impact of RA on work capacity than that of previous studies, which merely recorded the often self-reported permanent work disability. Differences in social security systems, as well as possible disparities in educational and occupational profiles, may lead to dissimilarities in disease outcomes, such as sick leave and permanent work disability, between countries. Incidence and risk factors for permanent work disability have, however, been quite parallel in studies conducted in countries with differing insurance systems (25), including our FIN-RACo study (4, 5). In addition, the indirect costs (i.e., lost productivity) have dominated the societal expenses of RA (1–3). Our results may thus be generalized to patients with early, active RA in other countries.

In summary, the present results indicate that there is a window of opportunity during the early course of RA, and the level of early treatment response determines the fate of a patient's ability to continue working. Our data suggest that a patient with recent-onset RA should be treated aggressively from the very start, with the target being early remission. Even an ACR50 response at 6 months is often inadequate to prevent permanent work disability and the high costs to society. We propose that if remission is not achieved by 6 months, the best choice is to institute more-vigorous therapy.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

The authors wish to thank the other members of the FIN-RACo Trial group, as follows: Heikki Julkunen, MD, PhD, Reijo Luukkainen, MD, PhD, Kaisa Vuori, MD, Leena Paimela, MD, PhD, Harri Blåfield, MD, Martti Nissilä, MD, PhD, Urpo Yli-Kerttula, MD, PhD, Kirsti Ilva, MD, Mikko Hakola, MD, Ilppo Pälvimäki, MD, Heikki Piirainen, MD, PhD, Kalevi Koota, MD, PhD, Claes Friman, MD, PhD, Oili Kaipiainen-Seppänen, MD, PhD, Per Franzen, MD, Tapani Helve, MD, PhD, Juhani Koski, MD, PhD, Marianne Gripenberg-Gahmberg, MD, PhD, Riitta Luosujärvi, MD, PhD, and Anna Karjalainen, MD, PhD.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES