Survival, Mortality, Causes of Death and Risk Factors of Poor Outcome Survival, Mortality, Causes of Death and Risk Factors of Poor Outcome

Systemic sclerosis is a rare autoimmune disorder with a historically bad prognosis. Survival has been improving over time and we can currently estimate a 1-year survival, 94.9; a 5-year survival, 84.4; a 10-year survival, 70.9 and a 20-year survival, 44.9%, from the time of diagnosis. Accordingly, mortality has been decreasing over time, being the overall standardized mortality ratio (SMR) 2.72 (1.90–3.83), SMR 2.4 after 1990. Among the SSc-related causes of death, the lung death is the most important cause and its relative percentage is increasing over time since the introduction of ACE inhibitors for the treat- ment of scleroderma renal crisis (SRC) in early 1990s. Among the SSc-non-related causes of death, cancer, infection and cardiovascular disorders are the leading causes of death. Risk factor predictors of poor outcomes are an elder age at diagnosis, the male gender, diffuse subset, visceral involvement and non-Raynaud’s phenomenon onset.


Introduction
Systemic sclerosis (SSc) represents one of the autoimmune systemic diseases with worse prognosis. It was actually a devastating scenario until it became well advanced in the twentieth century in terms of survival and mortality, but since late 1980s, the knowledge and course of scleroderma have been progressively improving. Nowadays, more risk factors are recognized and allow physicians to focus on patients with worse prognosis. As traditional SSc-related involvements improved, secondary involvements or SSc-non-related diseases have gained prominence.

Survival
Scleroderma was a devastating disease for ages. Physicians were short of proper tools to change significantly the prognosis of the disease since late 1980s-early 1990s due to the introduction of new therapies, firstly angiotensin converting enzyme (ACE) inhibitors for the treatment of the scleroderma renal crisis (SRC) and, in late 1990s-early 2000s, due to the implementation of pulmonary arterial hypertension (PAH) treatment with new drugs such as phosphodiesterase five (PDE5) inhibitors and antagonists of the receptor of endothelin (AREs). Survival has improved over time, measured at any time of the follow-up, from onset (in most studies described in the form of Raynaud's phenomenon) as well as from diagnosis and what is more important is that it keeps improving. This is true that survival in all stages has improved but especially 1-year and 5-year survival. Since late death is related the most to SSc-non-related causes, it might be reflecting the fact that physicians are improving significantly the prognosis of SSc-related involvements but not that much in other SSc-non-related diseases. Thus, just like in other autoimmune diseases, scleroderma is becoming step-by-step a chronic disease and not a terminal and deleterious diagnosis.
Prior to the assessment of survival of any cohort, we must pay attention to the methodology of the study because there is a huge variability among them, sometimes assessing survival from diagnosis and sometimes from the onset of disease. These last data are obviously a more imprecise data but certainly more real. Several survival and mortality studies from single cohorts and reviews have been published from the last mid-century, reporting data about cumulative survival at different times of follow-up and measured sometimes from the onset of disease and sometimes from the time of diagnosis ( Table 1) .
We show data previously released in Seminars in Arthritis and Rheumatism in the last and more precise meta-analysis, so that we can currently predict at the time of diagnosis: a 1-year survival, 94.9; a 5-year survival, 84.4; a 10-year survival, 70.9 and a 20-year survival, 44.9% ( Table 2 and Figure 1) [43]. T-test for independent groups among studies before and after 1990 (mid-cohort year). NA: non-available. Reprinted from Rubio-Rivas et al. [43], with permission from Elsevier. Table 2. Survival studies on scleroderma [43].

Mortality
SSc is an autoimmune disease with a broad spectrum of severity, ranging from a mild disease to a devastating one. The most valuable parameter in order to compare mortality (instead of a crude mortality rate) is the assessment of the SMR, a fundamental tool in the only five mortality metaanalyses reported so far in SSc ( Table 3). The SMR is the ratio between observed mortality and expected mortality in sex-and age-matched general population. These five meta-analyses are based on the assessment of the SMR: Elhai [43]. In base of this last study based on 17 studies, we could state that mortality is over 2.7-fold compared to the general population ( Table 4 and Figure 2) [5,17,24,26,29,30,31,33,37,40,42,[48][49][50][51][52][53]. Mortality has been decreasing over time and notoriously after 1990, being more reasonable to accept nowadays an SMR over 2.4-fold in a patient diagnosed today ( Figure 2). Obviously, prognosis should be individualized since different risk factors present at diagnosis or during the follow-up can modify this predicted SMR. For instance, SMR in males and dcSSc subset is expected to be worse compared to SMR in females or lcSSc subset (Figure 3).

Causes of death
As in other autoimmune diseases, the pattern of mortality has been changing over time since the autoimmune disease by itself has been the main cause of death in these patients for ages but the more physicians control the disease, the more likely it is to die due to other causes not directly related to SSc [43,[54][55][56].

SSc-related causes of death
Among the SSc-related causes, there are four major organs potentially involved: the lungs, heart, kidneys and gastrointestinal tract. Among them, lung and renal involvement were the most important as a cause of death during the twentieth century (

Lung involvement
In the case of lung involvement, death can be due to the progression towards respiratory failure due to PAH or interstitial lung disease (ILD).
In the case of PAH, evidence suggests that SSc-PAH patients have a worse response to therapy when compared to idiopathic PAH. New therapies (phosphodiesterase type 5 inhibitors and endothelin receptor antagonists) have improved its prognosis but not that much and thus, this is still a severe manifestation of the disease and a frequent cause of death. An early combination schedule of treatment has been suggested to be better in terms of prognosis. However, more studies are required to demonstrate and standardize this strategy of treatment [63].
Interstitial lung disease constitutes the most severe manifestation of the disease and is in fact the first cause of death in these patients (Figure 4). Therefore, it is crucial to perform a regular screening of this involvement and an early treatment when diagnosed. Patients showing the following criteria would warrant an immunosuppressive treatment: (1) either an extent of lung disease >20% on High-resolution computed tomography (HRCT) or an indeterminate extent (disease extent not readily classifiable as minimal or severe; HRCT extent is 10-30%) of disease plus an FVC < 70%, (2) patients experiencing a significant decrease in pulmonary functional assessment during the follow-up (FVC > 10% or DLco >15% or both, whatever the extent of lung involvement is for 12 months). Currently, the management of SSc-ILD is largely confined to immunomodulvation. Non-selective immunosuppressants such as cyclophosphamide followed by mycophenolate mofetil and azathioprine are still the most widely used medications in SSc-ILD. Several alternative approaches may be considered, including B cell depletion therapies (rituximab), anti-TGF-β antibody, tyrosine kinase inhibitors (imatinib, dasatinib), anti-IL-6 antibody, anti-IL-13 antibody, pirfenidone and haematopoietic stem cell transplantation (HSCT). Finally, lung transplantation may be limited to those patients, with severe SSc-ILD, unresponsive to pharmacologic therapy [64]. It is important to remember that, although often used, during the first stages of treatment, prednisone doses over 15 mg a day can be dangerous in order to trigger a scleroderma renal crisis.

Renal involvement
Scleroderma renal crisis occurs during the rapid progression of skin thickening in the early stages of dcSSc (<5 years after disease onset). Several case series published during the past 20 years and a 2013 systematic literature review has estimated that SRC develops in 5-15 and 15% of patients with dcSSc, respectively. Interestingly, the incidence and prevalence of SRC seem to be decreasing over time, possibly as a result of early recognition and management of SRC risk factors and early signs and symptoms in patients with dcSSc [65]. The introduction of ACE inhibitors (captopril) reduced dramatically its frequency as the cause of death since early 1990s of the past century (Figure 4). Besides, its incidence is decreasing but due to unknown reasons. The extended use of ACE inhibitors prescribed for other reasons (i.e. arterial hypertension or heart failure) has not been found as the cause of this decreasing incidence.

Heart involvement
Despite the fact that definite SSc for cardiac involvement does not exist, we could categorize its involvement in five major groups: pericarditis with or without cardiac tamponade, ischemic cardiopathy (documented myocardial infarction, angina, ischemic alterations in myocardial perfusion SPECT or requirement of coronary revascularization, surgical or percutaneous), pacemaker bearing regardless of the time of arrhythmia, sudden death and congestive heart failure [66]. As it is more recognized nowadays, its ratio is increasing, but the real challenge in the years to come shall be to distinguish the real scleroderma involvement from a cardiac SSc-non-related involvement. Anyway, we can hypothesize today that this is related to SSc in younger patients without classical cardiovascular risk factors.

Gastrointestinal involvement
The gastrointestinal tract is the most affected organ after the skin and can be affected from the oral cavity to the anus. About 90% of all patients will be affected during follow-up. This involvement can result in a decreased quality of life more often than a direct cause of death. In fact, it has been a rare cause of death over time. The few fatal cases have been related to those with severe intestinal involvement leading to malabsorption and secondary starvation. We cannot forget the possible role of the oesophageal involvement in the development of interstitial lung disease [67].

SSc-non-related causes of death
Among SSc-non-related causes of death, we can find three major diseases: cancer, infections and cardiovascular diseases ( Reprinted from Rubio-Rivas et al. [43], with permission from Elsevier. Table 6. SSc-non-related causes of death in medical literature [43].

Cancer
A higher standardized incidence ratio (SIR) of cancer in these patients not only compared to the general population that could be related to immunosuppressive treatment but also to the self-nature of the disease has been described [68]. In fact, those cancers diagnosed within the first 3 years after the diagnosis of scleroderma have been suggested to be classified as SSc-related causes. Cancer among SSc patients with RNA polymerase III antibodies has been reported to be in a close temporal relationship to the onset of SSc (first 36 months after the onset of SSc), which supports the paraneoplastic phenomenon in this subset of patients [69]. Thus, it is recommended to rule out this possibility at the time of the diagnosis, although protocol for this purpose has not been standardized so far. Cancers most frequently found in SSc patients are those from breast, blood, lung, gastrointestinal tract, genitourinary tract and skin and, out of these, those most related to the presence of RNAp III antibodies were breast cancer, skin cancer and genitourinary cancer [69].

Infections
Risk factors associated with infections in patients diagnosed with scleroderma include oesophageal (increased risk for aspiration pneumonia) and interstitial lung disease (increased risk for pneumonia), severe Raynaud's phenomenon or calcinosis (risk for localized superinfections) and the use of specific treatments for the management of the disease. Bacterial infections due to Gram-positive bacilli have been described, especially in patients with severe Raynaud or calcinosis. In patients receiving immunosuppressive treatment, especially corticosteroids, Nocardia sp. and Mycobacteria sp. must be taken into account. A few viruses such as Epstein-Barr virus and CMV have been described as triggering the onset of scleroderma, and Parvovirus B19 DNA has been detected in patients who have scleroderma, but the clinical correlate of this finding is unclear. Finally, among fungi, Pneumocystis jirovecii pneumonia has been reported in some patients who have scleroderma [70]. Aspergillus sp. has been rarely reported in scleroderma, but in any patient under cellular immunosuppression has to be taken into account as well [71].

Cardiovascular disease
Scleroderma, as other autoimmune diseases, shows up as an inflammatory background that leads to the fibroblast activation. This background is more visual in the first stages of the skin or lung involvement, but it is thought to happen elsewhere. Thus, scleroderma itself can be understood as a new cardiovascular risk factor not only involved in the development of microvascular disease but also of macrovascular disease [72]. Atherosclerosis has been found to be increased in patients with SSc in all territories: coronary arteries, carotid arteries, cerebrovascular vessels and peripheral arteries [72]. This is the most controversial group of diseases in terms of classification since it is a challenge to differentiate whether it is an SSc-related or a SSc-non-related event. Although sometimes undistinguishable, currently, the clinical context has been hypothesized to aid for this purpose. Thus, in young patients without other classical cardiovascular risk factors (smoking behaviour, diabetes mellitus, arterial hypertension, hyperlipidaemia, obesity), we state that a particular event should be classified as SSc-related.
According to the latest studies [43], the big picture when talking about causes of death should be that SSc-related death is estimated nowadays in 56.7% of all deaths. Among them, representing lung death 57%, heart death 28.2%, renal death 11.7% and gastrointestinal death 6.4% ( Table 7). In contrast, SSc-non-related death is estimated in 43.3% of all patients and among them being cancer, infections and cardiovascular disease the leading cause of SSc-non-related death.

A temporary pattern of SSc-related causes of death
In general, we could state that early death within the first years after the diagnosis of scleroderma is primarily due to the autoimmune disease itself, and late death is due to SSc-non directly related causes. Besides, this progression is currently even more notorious since data from the Spanish Registry show the fact that beyond 10 years after diagnosis, 83% of all deaths are due to SSc-non-related causes, supporting the idea that by struggling with the disease in the first years could save quite a few deaths due to the self-disease [54,55].  Among these SSc-related causes, pulmonary death (due to ILD or PAH) has been and still currently is the leading SSc-related cause of death in all stages of the disease. In contrast, renal death was the second cause of the death in the past and most of all in the dcSSc subset and within the first 5 years after SSc diagnosis, but, in the last years, we have been witnesses of an important decrease of renal death rate. Heart death is at first sight more present nowadays and within the early stages of the disease but possibly due to a better understanding and knowledge of this involvement. Gastrointestinal death has been the cause of death only inisolated cases over time (Table 8).
Thus, it is expected to see an increasing rate of SSc-non-related causes in the years to come, mainly cancer and cardiovascular causes. Among the SSc-related causes, cardiovascular causes will be the cornerstone and the challenge will be to distinguish SSc-related and SScnon-related cardiovascular events.

Risk factors of poor outcome
To date, several risk factors have been identified related to poor prognosis, sometimes reported as a result of univariate analysis and sometimes as a result of multivariate analysis [43].  Taking into account the number of citations from the different studies (Figure 5), more cited risk factors would be an older age at diagnosis, dcSSc subset, male gender and visceral involvement (most of all lung, heart and renal involvement).
It is not easy to quantify the overall risk attributed to any of these factors since they have been described in different ways, but by meta-analysing those described homogeneously (Figure 6), we could state a hazard ratio for kidney involvement 4.22 (3.42-5.19 [43].
New risk factors are required in order to identify those patients with worse prognosis who could get some benefits in terms of a more aggressive therapy and/or closer follow-up.
Recently, the mode of onset has been evaluated as a potential risk factor, finding a worse prognosis in those patients with an onset in the form of non-Raynaud's phenomenon, with the only exception of arthralgia (data not yet released).
Anyway, the risk should be individualized and accordingly lead to the decision-making for every patient. Thus, it should be our aim to create prognosis scales based on these known risk factors.  with permission from Elsevier [43].