By Dr Pang Ching Wai, CMC
Introduction: In the last two decades, the perception of HIV infection and Acquired Immune Deficiency Syndrome (AIDS) has changed from an almost uniformly fatal condition into a manageable chronic condition. Since its introduction in the 1990s, anti-retroviral therapy has improved the outcome of the HIV-infected patient significantly and has reduced the incidence of illness associated with the acquired immunodeficiency syndrome. HIV-infected patients, because of their improvedlife expectancy, have become more common among ICU admissions. In this study, we aimed to describe the etiologies of ICU admissions of HIV-infected patients in Hong Kong over a 5 years period and to examine whether there was any prognostic factor which could effectively predict length of stayand mortality.
Methods: Medical records of all HIV-infected patients who were admitted to ICUs of 7 tertiary hospitals in Hong Kong from 1stOct 2008 to 30thSept 2013 were retrieved. Their demographics, clinical characteristics and parameters as well as treatment status were extracted. They were analyzed for any association with day 30 mortality, ICU length of stay and hospital length of stay.
Results: During the study period, there were 104 ICU admissions identified, in which 36.5% of patients were newly diagnosedto have HIV infections after hospital admissions. The commonest causes of ICU admissions were respiratory failure (39.4%) and sepsis (30.8%). The mortality at day 30 counting from the date of ICU admissions was 25%. Serum albumin level on ICU admission (odds ratio [OR] =0.805; 95% CI =0.702-0.924; P=0.002) and APACHE II score (OR =1.089; 95% CI=1.017-1.165; P=0.014) were significantly associated with Day 30 mortality in multiple logistic regression. CD4 count was not significantly associated with Day 30 mortality(OR, 1.000; 95% CI, 0.995-1.005; P=0.926). A cut-off point of serum albumin of 18.5g/Lhad a positive predictive value of 90.48% of day 30 mortality (sensitivity of 75.00%, specificity of 76.92%).
Conclusions:During the period 2008-2013, the day30 mortality of HIV-infected patients who had been admitted to ICU was 25%. Respiratory failure was the leading cause of ICU admission in Hong Kong. Day 30 mortality was significantly associated with serum albumin level and APACHE II score. Serum albumin was found to be a valid tool for predicting day 30 mortality with a cut-off point of 18.5g/L. CD4 count was not statistically significant in its relationship with day 30 mortality.