Evaluation of point-of-care algorithms to detect diabetes during screening for latent TB infection.

ABSTRACT

BACKGROUND: Individuals with both diabetes mellitus (DM) and TB infection are at higher risk of progressing to TB disease.OBJECTIVE: To determine DM prevalence in populations at high risk for latent TB infection (LTBI) and to identify the most accurate point-of-care (POC) method for DM screening.METHODS: Adults aged ≥25 years were recruited at health department clinics in Hawaii and Arizona, USA, and screened for LTBI and DM. Screening methods for DM included self-report, random blood glucose (RBG), and POC hemoglobin A1c (HbA1c). Using HbA1c ≥6.5% or self-reported history as the gold standard for DM, we compared test strategies to determine the most accurate method while keeping test costs low.RESULTS: Of 472 participants, 13% had DM and half were unaware of their diagnosis. Limiting HbA1c testing to ages ≥30 years with a RBG level of 120-180 mg/dL helped identify most participants with DM (sensitivity 85%, specificity 99%) at an average test cost of US$2.56 per person compared to US$9.56 per person using HbA1c for all patients.CONCLUSION: Self-report was insufficient to determine DM status because many participants were previously undiagnosed. Using a combination of POC RBG and HbA1c provided an inexpensive option to assess DM status in persons at high risk for LTBI.