The analysis of sustained progression in our observational cohort demonstrated that sustained progression by any definition occurs infrequently in the disease modifying therapy era. In addition, sustained progression on the EDSS was only maintained for the remainder of followup for at most 50% of patients, which is consistent with previous findings from the placebo arms of clinical trials [8
]. Therefore, the use of any definition presented here as an outcome for clinical trials or observational studies has uncertain clinical meaning in terms of long-term disease progression. A possible reason for the poor performance is the requirement of an elevated EDSS for only two visits to qualify as sustained. Alternative definitions could require elevated EDSS for at least three visits, which corresponds to twelve months in CLIMB patients. All of the previous analyses were also completed requiring progression to be sustained for three consecutive visits, but the results observed were largely unchanged.
One potential explanation for the poor performance of the definitions of progression in this RRMS cohort is the limited relationship between the early phase of the disease and long-term disease progression. A recent study based on the London Ontario cohort demonstrated that only a high relapse rate very early in the disease course was associated with a worse prognosis in terms of progression [13
]. A sample of progressive patients from this cohort showed that the rate of progression was roughly the same in patients with no, one, or many relapses in the early phase of the disease [14
]. These results demonstrate the lack of association between relapse rate and long-term disease progression. To address the potential problems caused by transient increases in the EDSS due to relapses, the concept of sustained progression was developed, but the inability of any of the definitions presented in this paper to identify true disease progression provides further evidence that short-term changes in the EDSS are likely not identifying true irreversible disability.
The definitions of sustained progression require the EDSS to remain elevated for consecutive visits to ensure that the increase is not due solely to a relapse. At the same time, the presence of a relapse at the time of sustained progression might explain some of the reversions observed in our dataset. When all observed EDSS measurements taken during a relapse according the scoring provider, the high probability of reversion remained. Therefore, the reversions observed in our sample were not due to relapses and likely show true short-term disease fluctuations.
To address the problem with maintaining sustained progression on the EDSS, sustained progression on each of the FS scales was investigated. Since each FS measures only one system, it was hypothesized that definitions based on a specific FS scale would be less subject to fluctuations. Surprisingly, sustained progression on any individual FS was actually less sensitive to true change than the EDSS. Less than half of patients who experienced sustained progression on a specific FS maintained the progression. Therefore, using the FS scores for future studies does not appear to allow better measurement of irreversible worsening.
The effect of baseline EDSS on the time to sustained progression has important implications for future studies because the probability of sustained progression is a function of the starting EDSS value. Several clinical trials have already incorporated the difference in probability of progression for patients with an EDSS
0 through the use of D4 [6
], but our results demonstrate that the problem persists for other EDSS values. Therefore, any imbalance in the baseline EDSS may lead to biased estimates of group differences if EDSS is not appropriately controlled in statistical models. The very low chance of sustained progression for patients with an EDSS of 3–3.5 may also be due to regression to the mean/measurement error. Regression to the mean is a well-described phenomenon in MS clinical trials with regard to disease activity [15
], but this is rarely considered in analysis of sustained progression.
Our study has several limitations that warrant further discussion. One limitation is that different neurologists scored patients at different time points over the course of the study. The interrater reliability of the EDSS is lower than the intrarater reliability of the EDSS [16
], so some of the observed changes may be due to changes in rater. Although it is impossible to know what the EDSS would have been scored if the same rater was used throughout the study, changing rater was not significantly associated with maintaining sustained progression in this study. This result does not imply that rater had no effect on EDSS score, but it does demonstrate that the reversions observed in our study were not exclusively due to rater changes. A second limitation is that no restrictions were placed on treatment regimens followed by patients. Lastly, our results reflect those of patients enrolled in the CLIMB study. Although patients in the CLIMB study are treated in the same manner as other patients within our clinic, which, to our knowledge is in line with standard practices of tertiary MS Centers in the United States, it is possible that this population may differ from others seen within different practice settings or different geographic locations.
In conclusion, no definition of sustained progression based on the EDSS or the component FS scores was sensitive for irreversible progression of the disease. Since changing the definition failed to lead to improved performance, alternative measures of disease progression/severity must be found to better classify patients with short-term disease progression.