Reply to Chu et al

To the Editor—We thank our colleagues Chu et al for their interest in our article [1] and for their comments and questions [2]. We carefully rereviewed the results that we reported. To address their question about the odds ratio (OR) for obstructive coronary artery disease (OCAD) between participants with and without latent tuberculosis infection (LTBI) that increased from 3.2 in the unadjusted model (calculated from Table 2) to 4.96 in the adjusted model (Table 4), we used the same notation that Chu et al present in their letter. Thus, we refer to a regression model with 3 variables: outcome variable Y, independent variable X1, and independent variable X2. The outcome variable is OCAD and the independent variables are LTBI and atherosclerotic cardiovascular disease (ASCVD) risk score. The association between Y and X1 does not have to decrease after adjusting for X2, even when X1 and X2 are both positively correlated with Y. For example, this may occur when X2 is a composite variable such as the ASCVD risk score, which is derived from a number of ASCVD risk factors (eg, age, sex, race, systolic blood pressure, diastolic blood pressure, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, history of diabetes mellitus, and tobacco use).

In Table 4, we reported the results for a 4-predictor model based on a priori considerations. However, for this response, we also conducted a 2-predictor model with only ASCVD risk score and LTBI. The OR for LTBI was 4.78 (95% confidence interval, 1.1–21.8; P = .043), which is very close to 4.96 in the 4-predictor model. The 2-predictor model that includes both ASCVD risk score and LTBI status is also the best model after model selection using the Akaike information criterion. Thus, we do not believe that the higher adjusted OR for LTBI can be explained by including site and human immunodeficiency virus status in the model. Chu et al also mentioned a possible nonlinear relationship between ASCVD risk score and OCAD. We tried to fit the model using polynomial regression. The higher-order terms were not significant, which indicates that there is no significant evidence that the relationship is nonlinear. We agree that the sample size of our study may not have been large enough to detect more subtle differences in OCAD across groups, a limitation that we acknowledged in our article. Deriving our own ASCVD risk score system using the study dataset was out of the scope of our study and would require a much larger sample size, which could be considered in future studies. In summary, we believe our data suggest a clinically relevant association between LTBI and OCAD that is independent of global ASCVD risk. Larger studies are needed to validate our findings.

Notes

Disclaimer. The contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health or the institutions with which the authors are affiliated. The funding source had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.

Financial support. This work was supported in part by the National Center for Advancing Translational Sciences (grant KL2 TR001426 to M. A. H.) and the National Heart, Lung, and Blood Institute (grant K23 HL123341 to C. T. L.) at the National Institutes of Health.

Potential conflicts of interest. C. J. F. reports grants from Gilead Sciences Inc, ViiV Healthcare, Janssen, Amgen, Merck, Cytodyn, AbbVie, Atea, Ansun, and Merck and personal fees from Clinical Care Options outside the submitted work. C. T. L. reports grants from Medtronic Foundation and Gilead Sciences and personal fees from Esperion Therapeutics outside the submitted work. All other authors report no potential conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.