• lupus erythematosus
• systemic
• epidemiology
• autoimmune diseases

Systemic lupus erythematosus (SLE) is a systemic autoimmune disease characterised by immune inflammation.1 Autoimmune haemolytic anaemia (AIHA) is a pathological state in which antibodies attack red blood cells. AIHA and SLE shared genetic and environmental risk factors and pathophysiological mechanisms.2 AIHA is clearly over-represented in patients with SLE and often occurs before a diagnosis of SLE.3 However, at present, studies on the incidence of SLE in patients with AIHA are scarce. Therefore, we explored the correlation between AIHA and SLE risk in a nationwide, population-based, matched cohort study.

In the 2003–2013 Taiwanese National Health Insurance Database, we identified patients newly diagnosed with AIHA between 2005 and 2012 (online supplemental figure S1). We selected age-matched and sex-matched (1:20) non-AIHA cohort from one million representative populations. From this cohort, we further selected a comparison group via propensity-score matching (PSM, 1:2) for age, sex, comorbidities and possible confounders using the greedy algorithm (online supplemental methods). Ultimately, we identified 713 patients with AIHA and 1416 PSM-matched individuals without AIHA, with balanced baseline characteristics between groups (online supplemental table S1). Before PSM, we examined the risk of SLE associated with AIHA using the Cox proportional regression analysis after adjusting for demographics, medical utilisation and comorbidities at baseline shown as HR with 95% CIs. Sensitivity analyses were conducted using various definitions of SLE based on SLE treatment or after exclusion of patients with secondary AIHA. After PSM, we estimated the association between AIHA and SLE incidence was estimated using the conditional Cox model. The cumulative incidence of SLE was significantly higher in the AIHA group than in the control group (p<0.001) at the end of the follow-up period before PSM (online supplemental figure 2A) and after PSM (online supplemental figure 2B). Before PSM, the incidence of SLE in the AIHA cohort was 172.92 times higher than that in the non-AIHA group (403.13 vs 2.33 per 100 000 person-months), and the risk of SLE was increased in the patients with AIHA (HR, 155.38; 95% CI, 95.42 to 253.00, table 1). In different Cox regression models, the risk of SLE was consistently increased in patients with AIHA (online supplemental table S2). Women and younger age groups were also associated with an increased risk of SLE (online supplemental table S2). Sensitivity analyses revealed consistent results (online supplemental table S3). After PSM, the incidence of SLE in the AIHA cohort was 64.76 times higher than that in the non-AIHA cohort (409.07 vs 6.32 per 100 000 person-months) and patients with AIHA had an increased risk of SLE (HR, 54.67; 95% CI, 22.33 to 133.89, table 1).

SLE may affect the blood system, leading to one or more lineages of haemocytopaenia. Early recognition of SLE features in patients with haemocytopaenia may lead to a different management strategy.4 Recently, Zhu et al showed a significant association between idiopathic thrombocytopaenia and SLE risk (HR=17.4) in PSM-matched populations.5 The present study used a similar method to assess the risk of SLE in patients with AIHA and demonstrated a high association (HR=54.7) between AIHA and SLE risk using PSM-matched populations. A study has reported that the incidence of anaemia in patients with SLE with active disease is relatively high.6 The mechanisms of SLE anaemia mainly include immunological and non-immunological factors, of which AIHA is the most common immunological factor.3

In conclusion, patients with AIHA had a significantly higher risk of SLE than non-AIHA individuals. Clinicians should conduct early monitoring of SLE in patients with AIHA and provide relevant education for patients with AIHA. Further research is needed in the future to clarify the possible mechanisms of these correlations.