Lúpus Eritematoso Sistêmico na Atenção Hospitalar Terciária: avaliação de fatores clínicos e laboratoriais associados à infecção

Abstract

Introduction: Systemic lupus erythematosus (SLE) is a disease with high morbidity and mortality. About 25% of deaths in this population are secondary to infectious complications. The identification of infection in SLE patients is a major challenge in clinical practice because it presents clinical manifestations that are confused with active disease. Markers such as C-reactive protein (CRP) and procalcitonin showed limitations in differentiating infection from disease activity in these patients. The neutrophil-lymphocyte (NLR) and platelet-lymphocyte (PLR) ratios showed promising results for evaluation of disease activity, but they were still poorly explored in order to differentiate infection from disease activity. The objective of the present study was to evaluate clinical markers and the performance of CRP, NLR and PLR in the diagnosis of infection in a group of Brazilian patients with SLE admitted in a Tertiary Hospital. Methods: One hundred and four hospitalizations of SLE patients were retrospectively evaluated and allocated into three groups: infected (26), infected and disease activity (41), disease activity without infection (37). They were compared in relation to clinical and laboratory variables. The disease activity was measured by the modified SLEDAI2K score. Results: The frequency of fever, hypotension and tachypnea was higher in infected patients. CRP, NLR and PLR also differed among the groups with higher values observed in the infected group, intermediary in the mixed group and lower in the group with isolated activity - CRP [56 (12 - 133) mg / dL vs 26 (3 - 59) mg / dL vs 15 (25-38) mg / dL, p = 0.002], NLR [7.9 (4.3 - 19.9) vs. 4.0 (1.9 - 8.5 ) vs. 3.1 (2.14.9), p = 0.005], PLR [270 (146-616) vs. 227 (138-520) vs. 134 (104-250), p = 0.025]. The CRP value of 20 mg / dl showed a 67% sensitivity, 67% specificity (E), positive predictive value (PPV) of 78% and negative predictive value (NPV) of 53% for diagnosis of infection, considering 64% the prevalence of infection in the sample. The NLR and PLR presented the following measures of accuracy for the cohort points of 3.5 (S 65%, E 58%, PPV 73%, NPV 48%) and 151.4% (S 71%, E 53%, PPV 73% and NPV 51%), respectively. Fever, tachypnea, total leukocyte and PLR were independently associated with infection. The combined use of clinical and laboratory variables showed superior accuracy to the isolated use for diagnosis of infection in SLE patients. Conclusion: NLR and PLR presented inferior performance than CRP for the diagnosis of infection in patients with SLE. However, the combined use of the proposed flowchart has potential for future use in clinical practice, especially considering the satisfactory accuracy of the method and its low cost and accessibility.