Current methods of dividing scleroderma patients into groups, based on characteristics of their disease in order to predict its likely progression, gives highly variable results, researchers reported, noting that no single method is able to correctly predict outcomes across groups.
The study, “Subsets in systemic sclerosis: one size does not fit all,” published in the Journal of Scleroderma and Related Disorders, suggested that better ways of studying scleroderma are needed.
Researchers at the Jewish General Hospital in Montreal, Canada, recruited a total of 560 scleroderma patients. Since the prediction of disease course is most valuable in initial disease stages, they focused on patients who had been ill for less than four years.
The team analyzed disease outcomes separately for patients with a disease duration of two years or less, and those with a duration between two and four years. Patients were followed for a median of 5.1 and 7.1 years in the two groups, respectively.
To assess disease progression, researchers examined health-related quality of life, forced vital capacity (FVC, a measure of lung function), the presence of pulmonary hypertension or interstitial lung disease, and death.
Three different methods for grouping patients were scrutinized. The most common is based on skin involvement, grouping patients into limited and diffuse cutaneous scleroderma.
Patients are also often grouped based on the presence of specific autoantibodies. In the analysis, researchers looked at antibodies against centromeres (ACA), topoisomerase (ATA), RNA polymerase III (RNAP), and a group consisting of other antibodies.
Finally, the team performed a so-called cluster analysis. This type of statistical approach clusters patients based on a host of disease features. The team constructed three groups, and noted that patients in the two disease duration groups clustered differently.
Importantly, none of the methods could predict all five disease outcomes. For instance, skin involvement could not predict quality of life changes in either those with short or long disease duration, but results suggested it worked better in patients with limited scleroderma when considering FVC reductions and interstitial lung disease — if they had been sick for less than two years.
Limited scleroderma group patients with longer disease duration had a smaller chance of FVC decline and death, and marginally lower chances of lung hypertension and interstitial lung disease.
In contrast, the presence of antibodies could not predict pulmonary hypertension in patients with longer disease durations, but researchers suggested that the presence of ACA antibodies was linked to a lower chance of lung hypertension than RNAP antibodies. The presence of ACA was also linked to better FVC than other antibodies in those with longer disease duration, but was only better than the presence of ATA antibodies in the shorter duration group.
The cluster analysis, in turn, offered other predictions, and it could not predict the development of interstitial lung disease at all.
The analysis demonstrated that different groups had different predictive power depending on the outcome analyzed and disease duration, suggesting that none of the methods can be used for general assessments of likely scleroderma progression across patients.