Aim: To compare HbA_1c and fasting plasma glucose assessment, with the 2-h oral glucose tolerance test as reference, in screening for diabetes in people with turberculosis. Methods: Individuals (N=268) with newly diagnosed smear-positive tuberculosis were screened for diabetes at a tertiary hospital in Lahore, Pakistan. Diabetes diagnosis was based on WHO criteria: thresholds were ≥48 mmol/mol (≥6.5%) for HbA_1c and ≥7.0mmol/l for fasting plasma glucose. Results: The proportion of participants diagnosed with diabetes was 4.9% (n =13) by oral glucose tolerance test, while 11.9% (n =32) and 14.6% (n =39) were diagnosed with diabetes using HbA_1c and fasting plasma glucose criteria, respectively. The area under the receiver-operating characteristic curve was 0.79 (95% CI 0.64 to 0.94) for HbA_1c and 0.61 (95% CI 0.50 to 0.73) for fasting plasma glucose, with a borderline significant difference between the two tests (P=0.07). Conclusions: HbA_1c and fasting plasma glucose performed equally in terms of diagnosing new diabetes cases in individuals with tuberculosis, but the proportion of participants falsely classified as positive was higher for fasting plasma glucose. This may be explained by acute blood glucose fluctuations when using fasting plasma glucose. HbA_1c may be a more reliable test in individuals with transient hyperglycaemia.