Publications

BACKGROUND

Readmissions after hospitalization for acute myocardial infarction (AMI) are common. However, the few currently available AMI readmission risk prediction models have poor-to-modest predictive ability and are not readily actionable in real time. We sought to develop an actionable and accurate AMI readmission risk prediction model to identify high-risk patients as early as possible during hospitalization.

METHODS AND RESULTS

We used electronic health record data from consecutive AMI hospitalizations from 6 hospitals in north Texas from 2009 to 2010 to derive and validate models predicting all-cause nonelective 30-day readmissions, using stepwise backward selection and 5-fold cross-validation. Of 826 patients hospitalized with AMI, 13% had a 30-day readmission. The first-day AMI model (the AMI "READMITS" score) included 7 predictors: renal function, elevated brain natriuretic peptide, age, diabetes mellitus, nonmale sex, intervention with timely percutaneous coronary intervention, and low systolic blood pressure, had an optimism-corrected C-statistic of 0.73 (95% confidence interval, 0.71-0.74) and was well calibrated. The full-stay AMI model, which included 3 additional predictors (use of intravenous diuretics, anemia on discharge, and discharge to postacute care), had an optimism-corrected C-statistic of 0.75 (95% confidence interval, 0.74-0.76) with minimally improved net reclassification and calibration. Both AMI models outperformed corresponding multicondition readmission models.

CONCLUSIONS

The parsimonious AMI READMITS score enables early prospective identification of high-risk AMI patients for targeted readmissions reduction interventions within the first 24 hours of hospitalization. A full-stay AMI readmission model only modestly outperformed the AMI READMITS score in terms of discrimination, but surprisingly did meaningfully improve reclassification.

BACKGROUND

Alcohol use has been shown to accelerate disease progression in experimental studies of simian immunodeficiency virus in macaques, but the results in observational studies of HIV have been conflicting.

METHODS

We conducted a prospective cohort study of the impact of unhealthy alcohol use on CD4 cell count among HIV-infected persons in southwestern Uganda not yet eligible for antiretroviral treatment (ART). Unhealthy alcohol consumption was 3-month Alcohol Use Disorders Identification Test-Consumption positive (≥3 for women, ≥4 for men) and/or phosphatidylethanol (PEth-an alcohol biomarker) ≥50 ng/mL, modeled as a time-dependent variable in a linear mixed effects model of CD4 count.

RESULTS

At baseline, 43% of the 446 participants were drinking at unhealthy levels and the median CD4 cell count was 550 cells/mm (interquartile range 416-685). The estimated CD4 cell count decline per year was -14.5 cells/mm (95% confidence interval: -38.6 to 9.5) for unhealthy drinking vs. -24.0 cells/mm (95% confidence interval: -43.6 to -4.5) for refraining from unhealthy drinking, with no significant difference in decline by unhealthy alcohol use (P value 0.54), adjusting for age, sex, religion, time since HIV diagnosis, and HIV viral load. Additional analyses exploring alternative alcohol measures, participant subgroups, and time-dependent confounding yielded similar findings.

CONCLUSION

Unhealthy alcohol use had no apparent impact on the short-term rate of CD4 count decline among HIV-infected ART naive individuals in Uganda, using biological markers to augment self-report and examining disease progression before ART initiation to avoid unmeasured confounding because of misclassification of ART adherence.

BACKGROUND

Residual systemic inflammation persists despite suppressive antiretroviral therapy (ART) and is associated with non-AIDS clinical outcomes. We aimed to evaluate the association between ART adherence and inflammation in Ugandans living with HIV who were predominantly receiving nevirapine-based ART with a thymidine analog backbone and were virologically suppressed by conventional assays.

METHODS

Plasma concentrations of interleukin-6 (IL-6), D-dimer, soluble (s)CD14, sCD163, and the kynurenine/tryptophan ratio, in addition to CD8 T-cell activation, were measured at baseline and 6 months after ART initiation in treatment-naive adults who achieved an undetectable plasma HIV RNA (<400 copies/mL) at their 6-month visit. Adherence was measured through medication event monitoring system and calculated as the ratio of observed/prescribed device openings per participant. We fit adjusted linear regression models to estimate the association between ART adherence and the log-transformed plasma concentrations of inflammatory biomarkers.

RESULTS

We evaluated 282 participants (median age, 35 years; 70% women). The median (interquartile range) adherence was 93% (84-98). In the adjusted analyses, for every 10% increase in average ART adherence, we found a 15% [P < 0.0001; 95% confidence interval (CI), -21.0 to -7.9], 11% (P = 0.017; 95% CI, -18.3 to -2.0), and 3% (P = 0.028; 95% CI, -5.0 to -0.3) decrease in IL-6, D-dimer, and sCD14, respectively.

CONCLUSIONS

Higher ART adherence was associated with lower levels of biomarkers of inflammation, immune activation, and coagulopathy among Ugandans living with HIV who achieved viral suppression shortly after ART initiation. This suggests that ART adherence could have biological consequences beyond viral suppression. Whether ART adherence optimization in virologically suppressed individuals could reduce residual inflammation remains unknown.