The combination of didanosine (ddI) and tenofovir (TDF) has potential advantages, but because of several pitfalls (unexpected decreases in CD4+ T cells, increased risk of pancreatitis) its use has been questioned. Since anecdotal cases of transient insulin-dependent diabetes mellitus were seen in our clinic in patients on ddI + TDF-containing regimens, we explored the rate of this complication in more detail. Retrospective analysis of plasma glucose levels in patients who completed 12 months of treatment with three different triple antiretroviral regimens including ddI + TDF, TDF, or ddI was done. Patients taking antidiabetic drugs and/or those with baseline glucose levels >125 mg/dl were excluded. Weight, age, concomitant antiretrovirals, and ddI dose were assessed. At 12 months without treatment changes, fasting glucose levels were compared to baseline. A multivariate analysis was performed to evaluate which variables were associated with glucose elevations. A total of 177 HIV-infected patients were assessed (78 on ddI + TDF, 42 on TDF, and 57 on ddI). Mean baseline features were well balanced between groups for age (mean, 39 years), gender (78% male), CD4+ count (mean, 507 cells/mm3), weight (mean, 67 kg), and glucose level (mean, 95 mg/dl). There were only significant differences between groups for baseline viral load and protease inhibitor (PI) use (13% in the ddI + TDF arm vs. 7% and 9% in the TDF and ddI arms, respectively). At 12 months, 60% of the patients in the ddI + TDF arm were taking ddI 250 mg/day and the rest were on ddI 400 mg/day. At 12 months, hyperglycemia was significantly more frequent in the ddI + TDF arm (33%) when compared to patients on TDF or ddI separately (5% and 10%, respectively). In the multiple linear regression analysis, a lower weight (beta -0.35; 95% CI -0.67 to -0.03; p = 0.033) and use of ddI + TDF (beta: 13.05; 95% CI: 0.2 to 26; p = 0.047) were independently associated with a higher risk of developing hyperglycemia. The risk of hyperglycemia is increased in patients treated with ddI + TDF, particularly in those with lower weight. As high ddI exposure has been associated with endocrine pancreatic dysfunction and diabetes, ddI "overdosing" as result of concomitant TDF use and low weight might explain our findings. These results add a further note of caution to the use of TDF and ddI in combination.