A less granular perspective was available for diabetes medications in replication cohort, but the use of any diabetes medication was borderline significant ( P = 0.05) with a higher proportion of intermittent carriers reporting the use of any diabetes medications.
While we did not have specific data regarding who in the replication cohort used DPP4 inhibitors, a medication that attenuated the effect of %HbA1c in the primary cohort, when use of any oral agent, the best approximation of use of DPP4 inhibitors available in this cohort, was forced into the multivariable model, the effect of %HbA1c on the odds of intermittent carriage was attenuated slightly (aOR = 1.24; CI: 1.09–1.41) but remained highly significant, while the effect of oral medication use remained non-significant (aOR = 1.44; CI: 0.80–2.59) (data not shown).