Article Text
Abstract
Background Systemic sclerosis (SSc) has substantial mortality. We hypothesise that survival rates from prevalent cohorts may underestimate mortality early in the disease course, particularly in patients with diffuse disease who may have died before having the opportunity to enter the cohort. A mortality study in an inception cohort of patients with SSc, followed from disease onset, would enable more accurate evaluation of survival, minimising the effect of ‘survivor’ bias.
Objectives We sought to compare 10-year survival and cause of death in an inception cohort with a prevalent cohort of Australian patients with SSc.
Methods All patients enrolled in the Australian Scleroderma Cohort Study (ASCS) were included. Losses to follow-up were traced to December 2012. For each patient, date of disease onset (defined as date of onset of first non-Raynaud manifestation), date of recruitment, and date of death or date of last visit for those alive, were used in survival analysis. Kaplan Meier curves were used to compare cumulative survival according to disease subtype, diffuse cutaneous (dcSSc) v. limited cutaneous (lcSSc) in (i) the cohort as a whole and (ii) patients recruited within 5 years of disease onset who will be referred to as the ‘incident cohort’. Standardised death report forms were used to record causes of death. The single primary cause of death was classified as either SSc-related or non-SSc related.
Results There were 1001 patients in the cohort as a whole, in whom 27 deaths (15 diffuse, 12 limited) were recorded over 8 years of follow-up from disease onset. There were 333 patients in the incident cohort, in whom 25 deaths (15 diffuse, 9 limited) were recorded over 8 years follow-up from disease onset. In the prevalent cohort, the 10-year survival of patients with dcSSc was 90%, whereas it was only 70% in the incident cohort. Fifteen of 27 deaths (55.6%) in the first decade were SSc-related (7 pulmonary arterial hypertension [PAH], 4 interstitial lung disease [ILD], 4 PAH and ILD), while 12 (44.4%) were non-SSc related (8 malignancy, 1 sepsis, 1 cerebrovascular disease, 2 ‘other’). Similar survival curves were obtained from data in the Canadian Scleroderma Research Group database (N=1300; data not shown).
Conclusions There is substantial mortality early in the course (within first decade) of SSc, specifically in patients with dcSSc, which is quantifiable only in an inception cohort of patients who are recruited from disease onset and followed prospectively. This provides a compelling rationale for forming a large inception cohort of patients with SSc to evaluate prognosis and disease outcomes. Among other objectives, the newly formed INternational SYstemic Sclerosis Inception Cohort (INSYNC) study will evaluate early mortality in SSc.
Acknowledgements We thank Ms Candice Rabusa for assistance with data analysis.
Disclosure of Interest: None Declared