Dear Editor,

I read the article with interest which was published in Plos One in 2013. [1] It is one of the very few studies that has studied the effect of glycemic control during anti-TB treatment among TB-Diabetes patients. I would like to congratulate the authors for their research work. This has also been defined as one of the agendas for research into TB-Diabetes by Harries AD et al. [2]

In this study, the exposed group are those with uncontrolled glycemia / Diabetes during anti-TB treatment; unexposed group are those with controlled glycemia / Diabetes. Authors intend to assess whether there is association between uncontrolled glycemia and unfavourable TB treatment outcomes. The authors have concluded that they could not confirm the association between glycemic control during anti-TB treatment and TB treatment outcomes.

The operational definition for glycemic control among TB-Diabetes patients used by the authors may be the reason for the same. Differential missclassification bias resulted in apparent lack of association in the study. Let me explain this further.

The references cited to support the operational definition have been misinterpreted by the authors. The cut offs used by the authors (100 mg/dl for FBG; 140 for PPBG or RBG) are cut offs for normal and intermediate (pre-diabetes) blood glucose values when these tests are used for diagnosis. [3-5] The cut off I am quoting (130 mg/dl for FBG; 180 for PPBG) is the cut off for glycemic control to monitor treatment of Diabetes patients during follow up. [6] In other words, the cut off for fasting blood glucose (FBG) should have been 130mg/dl instead of 100 mg/dl used by the authors.The cut off post prandial blood glucose or random blood glucose (PPBG or RBG) should have been 180 mg/dl instead of the 140 mg/dl used by the authors.

As a result, those with controlled glycemia were wrongly classified as uncontrolled glycemia (exposed). There is a possibility that if they had used the correct cut off for glycemic control (130 from FBS and 180 for PPBS/RBS), the difference in unsuccessful treatment outcomes among TB-Diabetes with and without glycemic control would have been significant. Because of their use of very strict cut offs and resulting differential misclasssification bias there was an apparent lack of association when one does in fact exist possibly.

A reanalysis after changing the cut offs and definition of exposed and unexposed would be interesting. I understand that among all TB-Diabetes patients (n=667) due to operational issues there were many TB-Diabetes patients with unknown glycemic control status (n=427). There were 240 TB-Diabetes patients with known glycemic status and a reanalysis among these 240 patients is worth a try.

I wish the authors all the best in their endeavors.


References

1. Nandakumar K, Duraisamy K, Balakrishnan S, Sunilkumar M, Jaya Sankar S, Sagili KD, et al. Outcome of Tuberculosis Treatment in Patients with Diabetes Mellitus Treated in the Revised National Tuberculosis Control Programme in Malappuram District, Kerala, India. Wilkinson RJ, editor. PLoS One. 2013;8: e76275. doi:10.1371/journal.pone.0076275

2. Harries AD, Murray MB, Jeon CY, Ottmani S-E, Lonnroth K, Barreto ML, et al. Defining the research agenda to reduce the joint burden of disease from diabetes mellitus and tuberculosis. Trop Med Int Health. 2010;15: 659–63. doi:10.1111/j.1365-3156.2010.02523.x

3. Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia. Summary of Technical Report and Recommendations. Geneva, Switzerland; 2006.

4. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2006;29 Suppl 1: S43–8.

5. Operational guidelines. National programme for prevention and control of cancer, diabetes, cardiovascular disease and stroke. New Delhi; 2010.

6. American Diabetes Association. Standards of medical care in Diabetes - 2015. Diabetes Care. 2015;38: S1–S93.