|Home | About | Journals | Submit | Contact Us | Français|
Objective. The modified Rodnan skin score (MRSS) is a standard outcome measure for skin disease in SSc and calculated by summation of skin thickness in 17 different body sites (total score = 51). Our objective was to evaluate the sensitivity of change over time of individual body sites used in the calculation of total MRSS.
Methods. We analysed two randomized placebo controlled clinical trials investigating the effect of either recombinant human relaxin or type I oral collagen in dcSSc. Both trials used MRSS as the primary outcome measure. We used a change of >10 mm (on a 0–100 VAS) in the patient global assessment (PGA) as the clinically important improvement. We calculated the mean change and the effect size (ES) for each individual body site used in the total MRSS for each study. Magnitude of ES was assessed using Cohen's rule of thumb for ES.
Results. In the relaxin and collagen studies, 71 of 199 patients (36%) and 54 of 129 (42%) of patients had an improvement >10 mm on the PGA at 6 and 12 months, respectively. Total MRSS had large ES in both studies (0.85–0.98); the chest, forearms and hands had moderate ES (0.50–0.74); and the lower extremities, face, abdomen and fingers had small ES (0.16–0.49).
Conclusion. Certain body sites (hands, forearms and chest) are more sensitive to change compared with other body sites in two randomized clinical trials. It remains to be seen whether decreasing the number of body sites by exclusion of relatively static areas would further increase the sensitivity to change over time of the total MRSS.
The modified Rodnan skin score (MRSS), a measure of skin thickness, has been used as the primary outcome measure in clinical trials of SSc (scleroderma). It is feasible, reliable, valid and responsive to change in multicentre clinical trials . Measurement of skin thickness is also used as a surrogate measure of disease severity and mortality in patients with dcSSc—an increase in skin thickening is associated with increased risk of involvement of internal organs and increased mortality . In addition, spontaneous improvement in skin score is associated with more favourable outcomes . Our current objective was to evaluate the sensitivity to change of each body site used in calculating the total MRSS. The sensitivity to change is defined as the ability of a measure (or instrument) to reflect underlying change over time.
We used data from two different double-blind randomized clinical trials (RCTs) in patients with dcSSc.
Subjects with dcSSc were enrolled in this clinical trial evaluating the safety, efficacy and dose–response effect of continuous subcutaneously infused recombinant human relaxin . All participants had SSc disease duration of 5 years since the onset of the first sign or symptom of SSc other than RP. Other inclusion criteria included moderately severe skin disease with a skin score of 20 or 16 if truncal skin was involved. Patients were randomized to receive either relaxin (25 µg/kg/day or 10 µg/kg/day) or placebo in a 2:1:2 ratio for 24 weeks. Two hundred and thirty-one patients were randomized in the 24-week study. Analysis of the trial results revealed no statistically significant difference in the change in the skin score and functional disability between groups who received recombinant human relaxin and those who received placebo. We combined the two treatment groups in our study.
This trial randomized 168 dcSSc patients with baseline disease duration of up to 10 years to receive 12 months of oral bovine type I collagen (500 μg/day) or placebo . One hundred and twenty-four patients (74%) completed the 12-month trial. MRSS was obtained at 4-, 8- and 12-month follow-up visits. The results showed no statistically significant difference in the change in the skin score at month 12 in the two groups. We combined the two treatment groups in our study.
In both clinical trials, skin thickness was evaluated using the MRSS . The skin thickening was assessed by palpation of the skin in 17 areas of the body (fingers, hands, forearms, arms, feet, legs and thighs, face, chest and abdomen) using a 0–3 scale, where 0 = normal, 1 = mild thickness, 2 = moderate thickness and 3 = severe thickness. Total skin score can range from 0 (no thickening) to 51 (severe thickening in all 17 areas). All investigators underwent formal skin score training prior to both clinical trials.
We used the patient global assessment (PGA; 0–100 mm) as our anchor. A change of >10 mm (0–100 mm) in patient-reported outcome was considered clinically meaningful (7–10). For the collagen study, we assessed change in PGA between baseline and 12-month visit and for the relaxin study, we assessed change in PGA between baseline and 6-month visit.
The changed PGA (mean scoretime2 − mean scoretime1) score was dichotomized with 10 mm as ‘not improved’ and a change of >10 mm considered ‘improved’. To assess the usefulness of an anchor, experts recommend reporting correlation between the anchor (PGA) and changed score; for example, a correlation of zero will make the anchor useless so a correlation coefficient of 0.30–0.35 has been suggested (for details, see references  and ). The association between anchor and total MRSS change score was assessed using the Spearman correlation coefficient.
The sensitivity to change was assessed using effect size (ES) of each body site: (mean scoretime2 − mean scoretime1)/s.d. at baseline. We used Cohen's rule of thumb for determining ES for each site of the body for the 17-site MRSS. Cohen's rule of thumb for interpreting ES is that a value of 0.20–0.49 represents a small change, 0.50–0.79 a medium change and 0.80 a large change [13–15]. All analyses were performed using STATA 9.2 and P < 0.05 was indicative of statistical significance. Multiple comparisons were not corrected for since this was an exploratory analysis.
Detailed baseline characteristics have previously been published for two trials. In these studies, the majority of the participants were females and Caucasian. Mean (s.d.) age in the relaxin and collagen studies was 47.3 (10.3) and 50.8 (12.2) years, respectively. The mean (s.d.) MRSS scores in the relaxin and collagen studies were 27.3 (6.9) and 26.1 (7.8) U, respectively. In the relaxin study, 71 of 199 evaluable patients (36%) had an improvement >10 mm on the PGA at 6 months (Table 1). Of the 129 patients evaluated in the collagen study, 54 (42%) reported improvement of >10 mm on the PGA (Table 2).
The Spearman correlation coefficient between anchor (change in PGA) and total MRSS changed score ranged from 0.20 (P = 0.006) for the relaxin study to 0.40 (P < 0.001) for the collagen study.
The total MRSS change score showed a large ES in both studies (0.85–0.98). The regions of the chest, hands and forearms showed moderate ESs in both studies (0.50–0.74). In addition, the upper arms had a moderate ES in the relaxin study (ES 0.55–0.63). The lower extremities (including thighs, legs and feet), face and fingers had small ESs in both studies (0.16–0.49). Individual body sites (except fingers) with low ES had greater proportion of patients with skin score of 0 at baseline (no skin thickness) compared with sites that had a moderate ES (Table 3).
Patients in the ‘not improved’ group consistently had lower ESs in both studies (ES ranging from 0.07 to 0.53) compared with patients who were deemed to have ‘improved’ (ES ranging from 0.16 to 0.98; Tables 1 and and22).
The 17-site MRSS is currently the most commonly used primary outcome measure in the evaluation of skin thickening in dcSSc clinical trials . A previous study by Silman et al.  assessed the impact of reducing the number of body sites used in the MRSS. The researchers found that using only five bilateral body sites (forearm, lower legs, chest, upper arms and fingers) did not affect the precision of the total MRSS suggesting that using fewer body sites in the MRSS may be valid in clinical studies. In our study, we show that the lower extremities, abdomen, fingers and the face are not sensitive to change over time. Although Silman and colleagues did not evaluate sensitivity to change, three of the five body sites they used were found by our analysis to have moderate ESs.
We found a correlation coefficient of 0.20 (relaxin) to 0.40 (collagen) between anchor and total MRSS change score. Although 0.20 is lower than the recommended coefficient of 0.30, similar results were found in both the studies providing confidence in our anchor. In addition, other studies have shown the validity of using PGA as an anchor [10, 17] and in this study, a cut-off of 10 mm was able to differentiate ‘improved’ vs ‘not improved’ groups.
Our study shows that certain body sites had moderate ESs, whereas others showed negligible to small ESs. These differences may relate to the anatomy of the sites studied, higher intra-observer variability for those sites, or that these sites are not sensitive to change. Another possibility would be that individual body sites that did not respond had a large proportion of patients with no skin thickness (i.e. skin score of 0). Generally, individual body sites (except fingers) with low ES had greater proportion of patients with skin score of 0 compared with sites that had a moderate ES. Floor effect (proportion of patients having a baseline skin score of 0) can affect the senstivity to change .
Our analysis validates the total MRSS score as an outcome measure. The total MRSS score, which can be viewed as a combined index of separate body sites, showed large ES in both studies. This is consonant with the data that well-validated combined response indices are more likely to be responsive to change than individual measures  facilitating drug development and improving assessment of efficacy of therapeutic agents in other rheumatic diseases. Among individual body sites, forearms, hands and chest were consistently found to have moderate ES across all patients in both studies. This fact suggests that regions which are least affected by confounding variables are most likely to be sensitive to change across time. The upper arms were found to have moderate ES in the relaxin study and a small ES in the collagen group; this may be due to the different time periods over which the studies were performed. Another explanation may be due to higher floor effect seen in the collagen study (18–21% had a baseline skin score = 0 vs 14–15% in the relaxin study).
Our study has strengths and weaknesses. The strength of our study lies in the fact that we used large RCTs performed at experienced scleroderma centres in USA and that researchers evaluating the MRSS underwent training sessions prior to each of these trials. The main weakness of our study is that it is a post hoc analysis lending itself to the possible biases of such analyses and that some of the patients in the studies had relatively low skin scores, resulting in floor effects.
Our study has implications for clinical care and future trial design, similar to the use of 28- vs 66-joint count in RA . Seventeen-site MRSS is useful in clinical care and ‘gold standard’ for RCTs. We show that certain body sites (hands, forearms and chest) are more sensitive to change compared with other body sites in RCTs. It remains to be seen whether decreasing the number of body sites with the exclusion of those which are not sensitive to change over time would further increase the sensitivity to change over time of the total MRSS.
Disclosure statement: D.K. was supported by a National Institutes of Health Award (NIAMS K23 AR053858-01A1) and the Scleroderma Foundation (New Investigator Award). A.E.P. is a consultant to Argentis Pharmaceuticals and holds minor stock in Argentis and Fibrogen Corporation. Their university (the University of Tennessee) research foundation owns a patent on type I collagen for use in scleroderma. A.E.P. is named on this patent. Argentis has licensed type I collagen for use in sclerderma from the University of Tennessee Research Foundation. P.P.K is supported by a Ruth L. Kirschstein National Research Service Award (NRSA) Institutional Research Training Grant NIAMS 1 T32 AR053463. D.E.F. is on the Consultancy and Advisory Board (study, design and advice on future directions) for Abbott, Actelion, Amgen, BMS, Biog Idec, Centocor, Genentech, Gilead, GSK, Merck, Nitec, Novartis, UCB, Wyeth and Xoma and has received research grants from Abbott, Actelion, Amgen, BMS, Genentech, Gilead, GSK, Nitec, Novartis, Roche, UCB, Wyeth and Xoma. D.E.F. has received honoraria from Abbott, Actelion, Amgen, BMS, Biog Idec, Centocor, Genentech, Gilead, Merck and Nitec and has done CME only talks for Abbott, Actelion and UCB. All other authors have declared no conflicts of interest.