Survival and cancer risk in an unselected and complete Norwegian idiopathic inflammatory myopathy cohort

Abstract

Objective

To utilise an exposed/unexposed cohort strategy for mortality and cancer analyses across unselected and complete cohorts of patients with idiopathic inflammatory myopathy (IIM) resident in south-east Norway (denominator population 2.6 million), between 2003 and 2012.

Method

IIM cases were identified by comprehensive searches through patient administrative databases followed by manual chart review. Polymyositis (PM) and dermatomyositis (DM) cases were classified by the Peter and Bohan and/or Targoff diagnostic criteria and sporadic inclusion body myositis (sIBM) by the European NeuroMuscular Centre (ENMC) criteria from 1997 and/or 2011. Every patient was matched for sex, age and residential area with 15 unexposed/non-IIM individuals drawn from the national population registry.

Results

Total mortality in the IIM cohort was 27% (87/326). Standardized mortality rate (SMR) was higher in DM (2.6) than PM (2.4) and sIBM (1.7). IIM-related causes of death were frequent (64%) and included cancer (all IIM subsets), aspiration (sIBM), pulmonary complications (PM/DM) and infections (PM/DM). Multivariate analyses identified age at diagnosis (PM and sIBM), positive anti-SSA (PM), cancer (DM), and DLCO < 60% (DM) as independent mortality risk factors. Cancer risk was increased in DM (standard incidence rate 2.0) and PM (SIR = 1.3), but not in sIBM (SIR = 0.9). Ovarian cancer was more prevalent in DM than in the general population (8.3% vs 1.1%).

Conclusion

Our results suggest that mortality rates and cancer risk remain elevated in DM, and to a lesser degree also in PM. Mortality rate was also increased in sIBM, but some deaths appeared to be due to potentially preventable causes.

Introduction

Idiopathic inflammatory myopathies (IIM) are chronic, systemic disorders of unknown aetiology. They are defined by progressive loss of striated muscle tissue and include three major clinical syndromes: dermatomyositis (DM), polymyositis (PM) and sporadic inclusion body myositis (sIBM) [1], [2], [3], [4], [5]. PM and DM cause symmetrical and proximal weakness, while sIBM is more asymmetric and involves the quadriceps and finger flexors [6], [7], [8]. Myositis specific auto-antibodies (MSA) are frequent in PM and DM [9] and appear promising as markers for clinical subtypes [10], [11], [12] and cancer-associated myositis [9], [11].

New classification criteria for the entire IIM group are in progress [12], but to date DM and PM are usually classified by the 1975 Peter and Bohan diagnostic criteria [13], [14], or the revised 1997 criteria suggested by Targoff et al. [15] Classification of sIBM is mostly by the 1995 Griggs pathology criteria [16] or the more clinically based 1997 European Neuromuscular Centre (ENMC) criteria [17] last updated in 2011 [18].

Studies from the 1970s indicated increased mortality in PM/DM [19], [20], [21], [22], [23]. Later work have confirmed this, but shown large variation in mortality rates, probably due to differences in patient selection, classification, loss to follow-up and treatment-related issues [4], [5]. Peter and Bohan criteria were applied in eight studies on mortality [3], [4], [24], [25], [26], [27], [28], [29] but only one of them assessed an unselected population-based cohort [3]. This study, which is the largest mortality study to date, identified 248 PM/DM cases diagnosed in Finland from 1969 to 1985 through hospital discharge searches and subsequent case assignment by chart review. Mortality rates in this study were assessed by life tables [3], and not by matched population controls [26]. Older age at diagnosis, cardiopulmonary disease, cancer and the presence of MSA (including anti-amino-acyl-tRNA synthetase auto-antibodies, anti-signal recognition particle antibody, anti-155/140 and anti-CADM-140 antibodies) have all been identified as poor prognostic factors in PM/DM [5], [10], [30], [31], [32].

Little is known about mortality in sIBM. Long-term follow-up studies from Europe have not found reduced life expectancy [6]. In contrast, a recent multi-national study on selected patients, with no detailed case definitions, reported increased mortality, with an estimated standardized mortality ratio (SMR) as high as 6.58 [31].

Associations between PM/DM and solid tumours have been shown in population-based studies with different case assignment strategies [30], [33], [34], [35], [36], [37]. The largest of these studies was a pooled analysis of data from Sweden, Denmark and Finland [35], [36], [38]. It included 618 DM and 914 PM patients identified by ICD-7 or ICD-8 hospital discharge diagnoses, but less than one-third of these cases had their PM/DM diagnosis verified by chart review [30]. Standardized incidence rates (SIR) were estimated using National cancer registry data as reference and found to be 3.0 in DM and modestly increased, at 1.3 in PM [30]. In a recent study on biopsy-proven IIM cases from Australia increased cancer risk was observed across all IIM subsets, including sIBM [36]. The most common PM/DM associated malignancies appear to be of ovarian, lung, gastrointestinal, breast or haematological origin [30], [34], [35], [36], [38].

Data on survival and cancer across all the IIM diagnoses in the general population are rather limited. There are few studies based on unselected groups and few undertaken on sIBM. In the current study, based on sound case definitions, we assess survival and SMR, causes of death and risk factors for mortality, cancer frequency and cancer types seen in an unselected IIM cohort that includes all PM, DM and sIBM patients resident in south-east Norway (with a denominator population of 2.64 million) between January 2003 and December 2012. All patients with IIM were coupled with 15 age-, sex- and residential area-matched unexposed non-IIM individuals, from the national population registry, utilising an unexposed/exposed cohort strategy.