Look Beyond the Disease Activity Score of 28 Joints (DAS28): Tender Points Influence the DAS28 in Patients with Rheumatoid Arthritis

MATERIALS AND METHODS

For this explorative study, 200 consecutive outpatients with RA seen in one university hospital by 3 rheumatologists and in 4 general hospitals by 6 rheumatologists, collaborating in the Utrecht Rheumatoid Arthritis Cohort study group, were included. All patients had established RA according to the 1987 classification criteria¹³ and all patients had given informed consent. They received antirheumatic medication according to treatment protocols used in all participating hospitals; 65% currently used methotrexate as single disease-modifying antirheumatic drug (DMARD), 7% used hydroxychloroquine, 3% intramuscular gold, 2% azathioprine, 1% sulfasalazine, 0.5% cyclosporine, 0.5% leflunomide, and 0.5% oral gold; 10% used a combination of these DMARD and 10% currently used no DMARD. Two patients used anti-tumor necrosis factor-α drugs. During an outpatient clinic visit, ESR, TJC, SJC, and VAS-GH (0–100; 100 = worst score) were assessed and the DAS28 was calculated³. In addition, tender point assessment according to the 1990 American College of Rheumatology (ACR) criteria for FM was performed¹⁴. Data for duration of early morning stiffness (0–180 minutes), VAS pain (0–100 scale; 0 = no pain), and the Health Assessment Questionnaire (HAQ; 0–3 scale; 0 = no disability) were obtained.

RESULTS

Of the 200 patients studied, 4 were excluded from analyses because of missing data, leaving 196 eligible for evaluation. Patient characteristics are summarized in Table 1. According to published criteria³, 49% of the patients had active RA defined as high (DAS28 > 5.1, 8% of patients) or moderate (DAS28 > 3.2 and ≤ 5.1, 41% of patients) disease activity; 14% had low disease activity (DAS28 ≥ 2.6 and ≤ 3.2); and 37% were in remission (DAS28 < 2.6). Among all patients, 43% had no TP, 30% had a TPC of 1–5, 12% (15% of women vs 3% of men) had TPC of 6–10, and 15% (17% of women vs 8% of men) had TPC ≥ 11 (Table 1). Overall, women had significantly more TP than men: median 2 versus 0, respectively (p < 0.005). The distribution of TP and joint counts according to 4 regions of the body — upper and lower and left and right extremities — is illustrated in Figure 2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Distribution of tender points and joint counts. Mean prevalences of individual tender points shown as percentages of mean total tender point count in italic. Mean prevalences of tender joints for the upper and lower and right and left body regions separately are shown in the boxes as percentages of mean total tender joint count (TJC); similarly, prevalences of swollen joints are shown as percentages of mean total swollen joint count (SJC). TJC and SJC are part of DAS28 (assessing 28 joints), of which the only joint that is assessed at the lower extremity is the knee.

Figure 3 shows that in the TPC classes, with increasing TPC, the median DAS28, TJC, and VAS-GH increased significantly; in contrast, the ESR and SJC were not statistically significantly different among the 4 groups (Table 1). Similarly, the DAS28, TJC, and VAS-GH correlated significantly with TPC, in contrast to SJC and ESR; the TPC also correlated significantly with early morning stiffness, VAS pain, and the HAQ results (Table 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Medians of DAS28 and its individual components in the 4 tender point classes. TPC 0, n = 84; TPC 1–5, n = 60; TPC 6–11, n = 23; TPC 11–18, n = 29 patients. TPC: tender point count (0–18); TJC: tender joint count (0–28); SJC: swollen joint count (0–28); ESR: erythrocyte sedimentation rate (0–140 mm/h); VAS-GH: visual analog scale for general health (0–100; 100 = worst score).

Among patients with active RA according to DAS28 > 3.2, 13% had no swollen joint of the 28 joints assessed, indicating misclassification.

DISCUSSION

The DAS28 is a widely used instrument for assessing disease activity in patients with RA. Reliability of DAS28 in the individual patient can be questioned, especially if there is concomitant FM, as shown in other studies^7,8, or if there are tender points, even in the non-FM range, as shown in our study. Although the increase in the median DAS28 of 2.6 in the group without TP to 3.3 in the group with 1–5 TP (Table 2) might not seem impressive on first consideration, it was statistically significant, and one should keep in mind that DAS28 2.6 is the cutoff for remission and DAS 3.2 that for moderate disease activity. In the group with 1–5 TP compared to the group with 6–10 TP, the median DAS28 scores were 3.3 and 3.4, but mean DAS28 3.3 and 3.6, respectively (Table 1). With increasing TPC, the less objective DAS28 components (i.e., based on patient report: the VAS for general well-being and TJC) showed increasingly higher scores, but the more objective DAS28 components (i.e., observer- and laboratory-based SJC and ESR) did not. This is in accord with a previous study in which patients with FM — without RA and with normal ESR values and no swollen joints — had high scores on the DAS28¹⁶. Similarly to our study, in another study in patients with both RA and FM, statistically significant associations were found between TPC and DAS28 but not between TPC and ESR; in contrast to our findings, there was also a significant correlation between TPC and SJC¹².

These findings reduce the sensitivity of DAS28 to assess low disease activity or remission in individuals (more false-negative cases). We did not investigate this, but other causes of a raised ESR than disease activity (low serum albumin, anemia, or a paraprotein) would have the same effect. In the DAS28, the absence of joints of the feet, which are frequently involved in RA¹⁷, reduces the specificity of the DAS28 for this aim (more false-positive cases)⁶.

To improve the specificity of assessing remission in patients using individual DAS28-guided “tight control” and “treat-to-target” strategies, one could add to the DAS28 criterion of remission the criterion of absence of any swollen joint. This would mean assessment of all joints frequently involved in RA. If the presence or absence of arthritis cannot be assessed reliably, ultrasonography could be applied¹⁸. To improve the sensitivity of the DAS28 to assess low disease activity or remission in individuals, other influences falsely increasing the DAS28 should be taken into account, such as TP and elevations of the ESR not specific for RA. This would mean looking not only at the DAS28, but also at its individual components.

In contrast to the prevalence of FM, the prevalence of TP in the non-FM range in patients with RA using antirheumatic medication (including nonsteroidal antiinflammatory drugs and analgesics) is not known. A substantial part of our population had no TP. Our study comprised a population-based RA cohort visiting academic and general rheumatology outpatient clinics on a regular basis. In The Netherlands, virtually all patients with RA are treated by rheumatologists, not by general practitioners. Thus our sample reflects a common RA population.

One could speculate on the pathophysiological mechanisms of the origin of TP in RA. Our findings that ESR and SJC are not associated with TPC refute a direct causative relation of joint inflammation and TP. It is possible that physical deconditioning induced by less physical activity in the past because of signs and symptoms of RA plays an indirect role. Also, other mechanisms could influence pain and tenderness in a patient with RA, e.g., joint destruction, central amplification mechanisms, and sleep disorders.

The prevalence of concomitant FM in our study population was not known, but the number of patients meeting the cutoff of 11 TP according to the 1990 criteria for FM was 15% (Table 2), and this is consistent with the published prevalence of secondary FM of 12%–17% among patients with RA^8,9,10,11,12. We chose to assess TP but not to apply the ACR 1990 criteria for FM, as the FM 1990 criterion of chronic generalized pain is difficult to interpret in patients with RA. As the 1990 ACR criteria for FM are intended for classifying groups, especially for research, their relevance in clinical practice for individuals has been questioned. It has been suggested that in clinical practice for individuals, a TPC ≥ 6 might discriminate better between patients with FM and those without FM¹⁵. In our study, this TPC would also influence the reliability of the DAS28. New criteria sets have been developed not only for FM and RA, but also for RA remission^19,20,21; the new Boolean criteria of remission permit only 1 swollen joint to be present. In our view this is a real improvement compared to DAS28 remission, because in RA patients with DAS28 remission it is not infrequently the case that 5–10 swollen joints are present⁶.

The DAS28 disease activity index is influenced by coexistence of tender points, even in the non-FM range, due to the clear association of the TPC with the less objective DAS28 components VAS-GH and TJC. When applying DAS28 for “treat-to-target” treatment strategies, evaluation of not only the DAS28 but also its individual components along with a full physical evaluation, according to good clinical practice, including assessment of all joints frequently involved in RA (and also ankles and feet) and of TP is required for adequate estimation of the individual’s disease activity, and for making appropriate therapeutic decisions.