Background Systemic lupus erythematosus (SLE) is frequently not reported in death certificates of lupus patients, despite its known role as an underlying and/or immediate cause of death. Possible reasons may be insufficient access to patients' medical records at time of death (including details on their medical history) and/or physicians' unawareness of the contribution of SLE to death.
Objectives We aimed to analyze the extent and predictors of non-reporting of SLE in death certificates of 90 deceased SLE patients regularly followed-up in a routine academic setting at our Department.
Methods We retrospectively observed 90 SLE patients (68 females) deceased within the 2002–2011 period. All patients were ≥18 years of age and Croatian residents at the time of death, fulfilling ≥4 classification criteria of the American College of Rheumatology (ACR). We identified patients with SLE listed as a cause of death in the death certificate. An extensive set of variables was compared between patients with and without SLE reported in the certificate: demographics, ACR criteria at time of death and damage according to the Systemic Lupus Erythematosus International Collaborating Clinics (SLICC)/ACR index and its components at the time of death. We also compared the proportion of in-hospital deaths and autopsies performed. Frequencies were compared using the χ2 and Fisher's exact test, and continuous variables using the t-test and Mann-Whitney U test. Variables associated with reporting of SLE in the death certificate in the univariate analysis were included in a multivariate logistic regression model.
Results SLE was reported in death certificates of 41/90 (46%) patients. Patients with SLE not reported in their death certificates were older at death (62±14 vs. 53±15 years) and diagnosis (53±14 vs 42±18 years) and had a longer time from their last visit at our Department to death (0.80±1.00 vs. 0.34±0.66 years), compared to patients with SLE listed in the death certificate (p<0.05). They also had a lower proportion of renal disorder (20/49 vs. 29/41), cardiovascular and pulmonary damage (18/49 vs. 28/41 and 7/49 vs. 13/41, respectively), and died less frequently in hospital (28/49 vs. 35/41) and due to infections (4/49 vs. 26/41) (p<0.05). Conversely, these patients had a higher frequency of malignancy as a feature of damage (17/49 vs. 6/41) and a cause of death (14/49 vs. 1/41) (p<0.05). Only patients without SLE listed in their death certificate accrued gastrointestinal damage (7/49 vs. 0/41, p=0.015), hence this type of damage could not be included in the multivariate logistic model. In the multivariate model, the presence of infection as a cause of death was the single variable related to (non-)reporting of SLE (OR 0.053; 95% CI 0.012–0.237) (Table 1).
Conclusions Non-reporting of SLE in death certificates of lupus patients may be an obstacle towards assessing the true extent of SLE-related mortality, calling into question the reliability of vital statistics data extracted only from death certificates. Infections as causes of death and gastrointestinal damage may influence reporting of SLE in death certificates.
Calvo-Alen J et al. Rheumatology 2005;44:1186–9.
Disclosure of Interest None declared