Background/Purpose:

Pneumonia, remains as a main cause of mortality in patients with SLE. In patients without autoimmune disease, scores as Pneumonia Severity Index (PSI) and CURB-65 can identify patients at low risk of complications. These indexes has not been validated in patients with SLE. 

The aim of this study was to stablish the rate of complications of patients with SLE and pneumonia, identify specific risk factors associated with bad prognosis in these patients and describe the clinical behavior of PSI and CURB-65 to identify patients at low risk of complications.

Methods:

We retrospectively reviewed medical records of patients with SLE (ACR criteria) and pneumonia who attended the emergency room in a single tertiary care center, (January 2010 – March 2015). We collected laboratory and clinical data: demographics, treatment, risk factors commonly associated with poor outcome in pneumonia, and disease activity (SLEDAI-2K). Patients were followed for 30 days after discharge. A composite negative outcome were defined as need of mechanical ventilation (MV), septic shock or death during follow up.

Statistics. Initially we conducted a univariate analysis, using Student T, Mann-Whitney U, Chi squared or Fisher’s exact test as appropriate. Thereafter we conducted a multivariate analysis (logistic regression) with all variables with a p value ≤ 0.10.

Results:

We included 163 patients (121 women, 74%), who presented 194 episodes of pneumonia. At evaluation the mean ± SD age was 34.6 ± 12.4 years. Time since diagnosis of SLE 7.1 ± 8.2 years. SLEDAI 2K was 8 ± 5.6.

Duration of hospitalization was 10.7 ± 7.2 days. Fifty seven (29%) patients had the composite outcome: 13 (7%) septic shock, 50 (26%) needed MV and 12 (6%) patients died.

Variables who reach clinical significance in the univariate analysis were: body mass index (BMI) (higher risk for patients with BMI lower than 19), chronic kidney disease, congestive heart failure, use of methotrexate and use of tacrolimus. Clinical variables with significant association were respiratory rate (RR), heart rate (HR), body temperature, oxygen saturation (detected by pulse oximetry), leukopenia, confusion, changes in pH and procalcitonin levels.

In the multivariate analysis remain significant: low BMI (beta exponential 4.02, 95% CI 1.4 – 11.6, p=0.010), RR (beta exp 1.08, 95% CI 1.02 – 1.15, p=0.006), HR (beta exp 1.04, 95% CI 1.02 – 1.06, p=0.001), body temperature (beta exp 0.46, 95% CI 0.28 – 0.75, p=0.002) and arterial pH (beta exp 0.001, 95% CI 0.00001 – 0.079, p=0.002).

Twenty nine patients (14.9%) had zero points in CURB65 (low risk), 4 of them (13.8%) had an adverse outcome. A score of zero in CURB65 had a sensitivity of 93% and specificity of 18% for identifying patients without complications. Sixty patients were classified as Class I in PSI, 7 of them had a negative outcome. Class I in PSI had a sensitivity of 88% and specificity of 39% for identifying patients without complications.

Conclusion:

In this study none SLE associated variable was significant associated with bad prognosis in patients with SLE and pneumonia. PSI and CURB65 are in general accurate but misclassified about 10% of patients as low risk, so clinical judgement must complement these tools.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/risk-factors-for-complicated-pneumonia-in-systemic-lupus-erythematosus-sle/