Adult respiratory distress syndrome in systemic lupus erythematosus: causes and prognostic factors: a single center, retrospective study

Abstract

To determine the causes and prognostic factors of Adult Respiratory Distress Syndrome (ARDS) in patients with systemic lupus erythematosus (SLE). Methods: Among 544 Korean SLE patients, who were followed in the Lupus Clinic of the Catholic Medical Center from 1993 to 1997, patients diagnosed as ARDS were examined retrospectively. During the study period, non-SLE patients with ARDS were investigated and then compared to SLE patients with ARDS in terms of clinical variables. Results: Nineteen patients with SLE were revealed to have a history of ARDS (3.5%) and 13 (68.4%) of 19 patients died. Death related to ARDS was found in 34.2% of all deaths (n = 38) from SLE during the study period. The frequency and causes of ARDS were as follows; 9 sepsis or bacteremia (47.4%), 2 miliary tuberculosis (10.5%), 2 invasive pulmonary aspergillosis (10.5%), 2 acute pulmonary alveolar hemorrhage syndrome (10.5%), 1 acute lupus pneumonitis (5.3%), 1 massive hemorrhage due to placenta previa (5.3%), 1 aspiration pneumonitis (5.3%), 1 disseminated intravascular coagulation associated with systemic vasculitis (5.3%). The main organisms in sepsis were gram negative bacilli (61.5%) The median steroid dose administered 1 month before ARDS was significantly higher in patients (n = 13) with infectious ARDS compared to those (n = 6) with ARDS due to other causes (P = 0.038). Comparison of the laboratory and clinical variables between the survivors (n = 6) and the deceased (n = 13) showed that the survivors had lower SLAM indices at presentation (P = 0.004) and APACHE (Acute Physiology, Age, Chronic Health Evaluation) III scores within 24 h after diagnosis of ARDS (P = 0.024) than the deceased. The APACHE III scores correlated well with the SLAM indices (r = 0.615, P = 0.007). Non-SLE patients with ARDS during the study period were selected for comparison to SLE patients with ARDS. Age at the onset of ARDS was younger in SLE (n = 19) compared to non-SLE (n = 190) (P < 0.001). Duration from ARDS onset to death was shorter in SLE patients (P < 0.001). The mortality from ARDS tended to be higher in SLE patients (P = NS). The first-day APACHE III score was significantly higher in deceased SLE patients (n = 13) compared to deceased non-SLE patients (n = 105) (P = 0.001). Conclusions: ARDS was a common premortem event of SLE and showed a high fatality rate in SLE. The most common cause of ARDS in Korean patients with SLE was sepsis by gram negative bacilli. ARDS in SLE developed at a younger age, and progressed more rapidly compared to ARDS in general. The SLAM index and APACHE III score could be useful to predict the prognosis of ARDS in SLE.