Background:  Lactic acidosis and severe hyperlactatemia are rare but serious complications of ART. To identify risk factors for hyperlactatemia/lactic acidosis among HIV-infected adults, we conducted a case-control study, including 19 centers in 11 countries in Europe, the Americas, and Australia.

Methods:  Lactic acidosis was defined as arterial blood pH <7.35, bicarbonate <20 mmol/L, and lactate above normal; hyperlactatemia was defined as 2 consecutive blood lactate readings >5 mmol/L, irrespective of acid-base status. We randomly selected 2 controls per case from patients >16 years, matched by center and calendar year. Conditional logistic regression was used to identify risk factors.

Results:  Between 1997 and 2005, 110 cases (49 with lactic acidosis) were diagnosed and 220 controls were included:  36.4% of cases and 18.2% of controls were women (p <0.001); median age was 42.4 years for cases (inter-quartile range [IQR] 36.0 to 52.5) and 40 for controls (IQR 35.0 to 47.1) (p = 0.013); 41.8% of cases and 30.8% of controls were non-white (p = 0.014). In the univariate analysis cases were more likely to be receiving stavudine (d4T) (odds ratio [OR] 3.67, 95% confidence interval [CI] 2.20 to 6.14) or didanosine (ddI) (OR 5.54, 95%CI 3.04 to 10.10) at the time of the event, whereas controls were more likely to be exposed to zidovudine (AZT) or lamivudine (3TC). An even stronger association was exposure to a d4T/ddI-based combination (OR 25.31, 95%CI 7.43 to 86.23) compared with AZT/3TC-based combination. Case status was associated with shorter duration of exposure to both dideoxynucleosides. In a multivariable model adjusting for age, gender, and current CD4 count, hyperlactatemia and lactic acidosis remained associated with shorter duration on ddI (adjusted OR 8.39, 4.54 and 3.09 for patients exposed for <12, 12 to 24, and >24 months, respectively). In addition, age >40 years (adjusted OR 2.6, 95%CI 1.08 to 6.29), female gender (adjusted OR 5.97; 95%CI 1.92 to 18.5), and advanced HIV-induced immunosuppression are independently associated with hyperlactatemia and lactic acidosis (adjusted OR 5.66, 9.53 and 7.36 for patients with 200 to 349, 100 to 199, and <100 CD4, respectively).

Conclusions:  Lactic acidosis and hyperlactatemia were strongly associated with exposure to dideoxynucleosides. The additional associations with female gender, advanced immunosuppression, and possibly ethnicity have important consequences for choice of ART regimens in developing countries. The association with shorter duration of exposure argues against a mechanism depending on cumulative exposure and may support the hypothesis of a particular susceptibility in some patients.