Abstract
Sum of Pain Intensity Difference (SPID) is an outcome measure that summarizes treatment response over a clinically relevant period. SPID is widely reported in clinical trials of analgesics but has been little used in migraine trials. We compared SPID over 2 h with the standard migraine outcome measures of pain-free at 2 h and headache relief at 2 h using data from four published clinical trials of rizatriptan in migraine patients. In assessing treatment response (rizatriptan and sumatriptan versus placebo, rizatriptan versus sumatriptan, within-treatment dose effects), SPID usually yielded similar results to the more easily understood pain-free measure.
Pain intensity differences and summed pain intensity differences have been widely used in randomized controlled clinical trials (RCTs) of analgesics (1–3). In the recent second edition of the International Headache Society (IHS) guidelines for controlled trials in migraine it is stated that ‘Sum of Pain Intensity Difference (SPID), based on the verbal/numerical scale (0=no headache; 1=mild headache; 2=moderate headache; 3=severe headache) could theoretically be useful since it has the advantage of summarizing the benefits of treatment over a clinically relevant period, e.g. 2 hours’ (4). There is, however, very limited experience with SPID in RCTs in headache and migraine (5–7).
During the development program for rizatriptan, SPID was calculated as a secondary efficacy measure in four published RCTs (7–10), but reported in only one (7). In order to increase the database with SPID in migraine RCTs, we now present these data and compare them with results obtained for the new IHS-recommended primary efficacy measure, percentage of patients pain-free at 2 h, as well as the most often used efficacy measure, headache relief at 2 h (a decrease in headache from severe or moderate to none or mild). In order to compare the results across RCTs, only results for the first treated attack are shown for the two crossover RCTs which involved treatment of more than one attack (7, 9). It should be emphasized that the primary prespecified analyses in the studies comparing rizatriptan with sumatriptan (7, 10) were time-to-event analyses rather than those presented here, and used two-attack data where appropriate. Rizatriptan was superior to sumatriptan in these analyses.
A summary of results are shown in Table 1. All active treatments were superior to placebo on all three measures at at least the <0.01 level of significance, with levels of significance (<0.01 or <0.001) being identical on the three measures in each case. The results of the statistical comparisons between the active treatments for the three measures are shown in Table 2. Assuming a significance level of 0.05, the SPID and pain-free measures always agreed with each other in showing significance or non-significance. The headache relief measure generally agreed with the other measures, but there were four instances of non-significant findings for headache relief but significant differences on the other two measures for the same treatment comparisons.
Results for SPID, pain-free at 2 h, and headache relief at 2 h
n given is for SPID. The ns for pain-free and headache relief were slightly higher in some cases. SPID is defined as the sum of pain intensity differences weighted by the length of the interval since the previous observation. Since observations in the rizatriptan studies were 0.5 h apart up to 2 h, 0.5 was used as the weight. Headache severity was reported on a 4-grade scale (0=no headache, 1=mild headache, 2=moderate headache, 3=severe headache). The pain intensity difference from baseline (2 or 3) was calculated at each time point and multiplied by 0.5. The four weighted differences were then summed.
Statistical comparison of active treatment results for SPID, pain-free at 2 h, and headache relief at 2 h
Placebo comparisons are not shown. All active treatments were superior to placebo on all three measures at at least the<0.01 level. The Kramer et al. study (9) is not included because it did not include any active treatment comparisons. SPID was analysed using Student's t-test. Headache relief and pain-free were analysed using Fisher's exact test.
∗ P<0.05;
∗∗ P<0.01;
∗∗∗ P<0.001.
†When the two attacks treated were analysed together using
Comments
SPID can be regarded from a clinical or a methodological point of view. For clinicians it is difficult to understand the relevance, for example, of a mean SPID of 1.66 vs. 0.72 (P<0.001), whereas the relevance of 42% pain-free vs. 10% (P<0.001) (results for rizatriptan 10 mg vs. placebo from Teall et al. (8)), is obvious. From a methodological point of view SPID could, as mentioned above, be theoretically useful since it has the advantage of summarizing the benefits of treatment over a clinically relevant period of 2 h. However, to be useful, SPID should be demonstrated to add something to the statistical analysis of RCTs in migraine; i.e. it should be helpful in demonstrating differences between drugs or doses of drugs that are not found with analyses of pain-free or headache relief. Our review of single attack data suggests that analysis of SPID does not appear to add anything. Pain-free and headache relief, especially pain-free, tell the same story with regard to detecting differences between treatments. On the other hand, when two-attack data were analysed in the Goldstein et al. study (7), SPID did show a significant difference between rizatriptan 10 mg and sumatriptan 50 mg (P=0.026), whereas pain-free and headache relief differences narrowly failed to achieve the significance criteria (P-values of 0.066 and 0.068, respectively). In most cases, though, SPID usually gives similar results to the more easily understood pain-free measure.