Skip Navigation LinksHome > November 2014 - Volume 33 - Issue 11 > Predicting Mortality in HIV-infected Children Initiating Hig...
Pediatric Infectious Disease Journal:
doi: 10.1097/INF.0000000000000454
HIV Reports

Predicting Mortality in HIV-infected Children Initiating Highly Active Antiretroviral Therapy in a Resource-deprived Setting

Nugent, James MPH, MD*; Edmonds, Andrew MSPH, PhD; Lusiama, Jean MD; Thompson, Deidre MPH; Behets, Frieda MPH, PhD§; for the Pediatric HIV Care and Treatment Group

Collapse Box


Background: While highly active antiretroviral therapy (HAART) programs have been scaled up across sub-Saharan Africa, no prognostic models for the prediction of mortality risk for children initiating HAART are widely available. Current clinical prediction tools for human immunodeficiency virus (HIV)-infected children are derived from pre-HAART data and therefore cannot predict mortality for children initiating HAART. The purpose of this study was to develop a mortality risk scoring system for HIV-infected children beginning HAART in a resource-deprived setting.

Methods: Observational data from HIV-infected children initiating HAART from December 2004 through March 2012 in Kinshasa, Democratic Republic of Congo, were analyzed. Cox proportional hazards models were constructed to assess associations between demographic and clinical characteristics at the time of HAART initiation and mortality. Each child received a model-based risk score predicting mortality after HAART initiation.

Results: By 31 March 2012, 1010 children had started HAART. One hundred three children (10.2%) died at a median of 5.3 months post-HAART initiation, yielding a mortality rate of 3.4 deaths per 100 child-years. The final mortality prediction model included undernutrition, low CD4 count, HIV symptoms, and low total lymphocyte count. These factors were highly predictive of mortality in the study population (C statistic = 0.79) and performed well when applied to the validation population (C statistic = 0.77).

Conclusions: Mortality among children starting HAART in resource-deprived settings can be predicted using a simple scoring system incorporating several readily available factors. Identifying predictors of mortality will help clinicians target modifiable risk factors, such as undernutrition, which are not directly addressed by HAART.

© 2014 by Lippincott Williams & Wilkins, Inc.


Article Tools


Article Level Metrics

Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.