Tuberculose em unidade de terapia intensiva: análise descritiva em um hospital de referência HIV/AIDS na Amazônia Brasileira

Abstract

Tuberculosis (TB) is recognized as a serious public health problem. In Amazonas state, in 2014, TB mortality rate was 3.5/100,000 habitants, and the proportion of TB /HIV co-infection was 16.2%. Data in the literature on severe cases of TB treated in intensive care units (ICU) are scarce. This study aimed to describe socio-demographic, clinical, hematology, laboratory, radiological and variables associated with mortality in patients with TB in the ICU. It is a descriptive retrospective study of cases treated for TB in the ICU of Tropical Medicine Heitor Vieira Dourado Foundation (FMT-HVD), Amazonas, from 2011 to 2014. We considered TB cases patients with smear or positive culture; or/and clinical, clinical and radiological; or/and histopathological criteria (autopsy) compatible with TB; or/and basic cause of death records in the Mortality Information System (SIM) with mention of TB; or/and TB diagnosis at discharge ICU. The information was obtained through electronic medical records and ICU admission books. During the study period 131 patients were included. The mean age was 36.4 years, with a predominance of males (71%). TB/HIV co-infection was observed in 91.6% of cases, of these, 87.3% had CD4 cell counts below 200 cells / mm³ and 48.9% were antiretroviral therapy (ART). Anti-TB treatment for up to 30 days prior to ICU admission was mentioned in 56 patients (49.1%). 99 patients who were smear, 21.2% were positive, while 87 who underwent culture, 27 (31.3%) confirmed TB. One hundred and two died (mortality 77.9%), the majority (93.1%) in less than 28 days of stay in the ICU. Among HIV-positive patients were more frequent hypochromic anemia (p = 0.04), and less frequent neutrophil (p = 0.04) and leukocytosis (p = 0.04). Early death (<28 days) occurred more often in patients without sputum (p = 0.05), the extrapulmonary form (p = 0.06) in female patients with anemia (p = 0.03), Syndrome Acute Respiratory Distress (ARDS) (p = 0.02) and low adherence to ART (p = 0.03). Death was more frequent in patients with lower CD4 count (p = 0.009), use of mechanical ventilation (MV) (p = <0.0001), hypoalbuminemia (p = 0.005), and thrombocytopenia (p = <0.0001). In conclusion, most patients were diagnosed with TB by clinical criteria (75.6%). The smear and culture were underperforming. Low laboratory confirmation was associated with a high proportion of patients coinfected with HIV and to high mortality rates. Therefore, it is suggested the implementation of clinical protocols that streamline diagnosis and adoption of appropriate treatment for TB patients treated in the ICU.