Recreational drug inhalation and socioenvironmental determinants of pulmonary infection in patients with HIV: a hospital-based case control study

BACKGROUND: Pulmonary infections cause substantial morbidity and mortality among people with HIV. Our objective was to evaluate common urban exposures among those living with HIV which may increase risk of pulmonary infection. METHODS: We enrolled 75 adult patients infected with HIV and admitted to San Francisco General Hospital into a hospital-based case-control study: 38 cases with pulmonary infection and 37 controls admitted for diagnoses other than pulmonary infection. We administered questionnaires to assess inhaled drug exposures, outdoor and indoor pollution, and occupational exposures. We fit multivariable logistic regression models with pulmonary infection as the dichotomous outcome and each exposure as a primary predictor in separate models, including potential confounders based on biological plausibility and backward selection criteria. RESULTS: Participants were middle-aged (median age 48 years) and predominantly male (77%). Poverty was prevalent, with 79% unemployed and 27% unstably housed or homeless. Median CD4 count was lower in cases compared with controls (88 vs. 300 cells/µL, p = 0.005). Most cases were admitted with either bacterial pneumonia (76%) or Pneumocystis pneumonia (16%), and the most common admission diagnosis among controls was non-pulmonary infection (43%). 59% of participants had smoked tobacco cigarettes within the month prior to enrollment while 40% had recently smoked marijuana. In adjusted logistic regression models, tobacco smokers had six times higher odds of pulmonary infection (aOR 6.43, 95%CI 1.70-24.4) compared with non or former smokers, while those who had recently smoked marijuana experienced a 79% reduction in odds of pulmonary infection (aOR 0.21, 95%CI 0.06-0.78) compared with no recent marijuana smoking. Unstable housing or homelessness predicted a three-fold increased odds of pulmonary infection (aOR 3.22, 95%CI 1.01-10.3), whereas those sleeping on higher floors were at reduced odds of pulmonary infection, with each floor level above street-level predicting a 40% reduced odds of pulmonary infection (aOR 0.60, 95%CI 0.41-0.86). Finally, opening windows for ventilation was associated with reduced pulmonary infection odds (aOR 0.051, 95%CI 0.004-0.64). CONCLUSIONS: Pulmonary infection odds were increased with tobacco smoking and homelessness; and decreased among recent marijuana smokers, those sleeping on higher floors and opening windows for ventilation.