Survival is improving in systemic sclerosis: true or false?

It depends on the cohort studied and what you compare it with This editorial refers to ‘Trends in mortality in patients with systemic sclerosis over 40 years: a systematic review and meta-analysis of cohort studies’, by Muriel Elhai et al. , doi:10.1093/rheumatology/ker269, on pages 1017–1026 and ‘The impact of cardio-pulmonary manifestations on the mortality of SSc: a systematic review and meta-analysis of observational studies’, by András Komócsi et al. , doi:10.1093/rheumatology/ker357, on pages 1027–1036 . Survival rates in SSc can be studied from data in registries [ 1 ]. For instance, in France the 3-year survival rate is 91%, whereas in Spain the 10-year survival rate is >70% [ 2 , 3 ]. A review showed that survival in SSc is improving worldwide [ 4 ]. However, there was no comparison with the general population. National data from the USA and France have found between-country mortality differences in SSc. In fact, the standardized mortality ratio (SMR) increased by >40% between 1980 and 1998 in the USA, mostly in women born before 1940 [ 5 ]. SMR is the ratio of observed deaths to expected deaths, where expected deaths are calculated for a comparison group (age and gender matched). However, when comparing the survival rates of SSc patients diagnosed over more recent decades to older cohorts, survival is improving. Ascertainment bias and variability between cohorts could account for these findings. Less severe patients may be detected now compared with the past because of widely available SSc-specific antibodies and the changing case mix (more lcSSc). Two meta-analyses in this issue of Rheumatology show a trend for improved SSc survival for organ manifestations (lung, heart, pulmonary arteries and renal) [ 6 ] and yet a large gap between SSc and background population survival exists, where no improvement in SSc survival compared with age- and gender-matched survival has been observed over four decades [ 7 ]. The survival gap between dcSSc and lcSSc is decreasing but dcSSc has a poor SMR [ 6 ]. It is possible that early detection leads to both earlier intervention and better outcomes, but one questions whether there is lead time bias. This bias is probably not entirely true as trials of Class II pulmonary arterial hypertension (PAH) showed a better outcome with treatment in 6-month studies [ 8 ]. Cohort size seems to be important for detecting PAH by screening echocardiograms [ 9 ]. Therefore, why are the two meta-analyses of survival in SSc different? How does one reconcile these two articles where mortality in SSc is improving in one but not in the other using meta-analyses of cohort studies? Both articles found that SMR with organ involvement is high (three times more deaths than the matched population) and stable. The cardiopulmonary (organ-based) meta-analysis included twice as many studies and performed a cumulative patient analysis to determine SMR. The hazard ratio showed a negative slope (lower SMR) over time, which was significant [ 6 ]. The meta-analysis by Elhai et al. [ 7 ] contained nine studies and did not use cumulative patient numbers and found the log change of the mid cohort year not significantly different over time ( P < 0.333). The techniques used are slightly different between the two papers. Both the papers have similar SMRs throughout the study years. Variable results depend on the inclusion criteria for selecting studies more than the methods used to study the SSc cohorts. If more studies are included, the CI would narrow around the point estimate, provided that there is no significant heterogeneity between the cohorts. There are biases in cohort studies such as selection bias, drop-outs and variable times of follow-up. Other than the estimated expected mortality in the general population, there is no control group, which is different from meta-analyses of randomized controlled trials. Risk factors associated with increased mortality in SSc are not identical among all papers but some are common to many papers, e.g. male, younger age at onset, dcSSc subset, higher skin scores, antibodies (topo I, RNP) and organ involvement [ 10 ]. Angiotensin converting enzyme inhibitors markedly changed early mortality from scleroderma renal crisis (SRC), and SRC may be decreasing temporally, which should result in less overall SRC mortality. However, SRC may be associated with severe SSc. For instance, there is a 3-fold increased mortality from pulmonary hypertension if renal insufficiency is present [ 11 ]. Hence, does early detection of organ involvement affect survival of patients with SSc? SSc-associated interstitial lung disease (ILD) has a better chance of survival than other CTDs [ 12 ], perhaps due to routine screening [ 13 ]. The use of CYC shows positive data in ILD [ 14 ], but relapse occurs after cessation of therapy [ 15 ]. PAH is detected earlier with screening [ 16 ]. Survival with PAH in SSc is better than in historical data, but still has a 4-year mortality rate of 40–50% [ 17 ]. Between-country survival varies: 22% for 4-year survival in Sweden and 56% for 3-year survival in France [ 18 , 19 ]. In the EULAR Scleroderma Trials and Research database, half the deaths were caused by SSc (pulmonary fibrosis, PAH and cardiac) [ 20 ]. Defining changes in mortality depends on how the study is designed. It would not be surprising if SSc mortality was improving but not closing the gap compared with the reference population (an unaltered SMR), which is the case in RA. SMR in RA is stable despite good treatment options, earlier intervention, targeted outcomes and observations by clinicians of decreasing severity [ 21 ]. Temporal changes in SSc mortality will be dependent on comparisons that are made. For instance, a study found that older age of onset of SSc had worse survival than younger age, even though they were more likely to have the better subset (lcSSc). However, when SMR was used to adjust to the expected background population mortality rates, there was no increase in mortality in the elderly subset [ 22 ]. The case mix of SSc patients included in each cohort will influence mortality trends. In the UK database, although more organ involvement was detected in recent years, the mortality was better, but the improvement was seen in dcSSc and not lcSSc, with 5-year survival rates going from 69 to 84% in the former for patients seen in 2000–03 vs 1990–93, whereas it was 93 and 91% in lcSSc [ 12 ]. Thus, depending on how you study it, SSc survival may or may not be improving. SSc survival may be better than previous survival when compared with SSc cohorts over several decades, but it still has a large gap between SSc and the age- and gender-matched referent population. Researchers from both the meta-analyses agree that SMR for SSc patients with significant internal organ involvement is high. Perhaps we will see the trends improving with current best practice and future treatments. Disclosure statement : The author has declared no conflicts of interest.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.