Trends in Timing of and Glycemia at Initiation of Second-line Type 2 Diabetes Treatment in U.S. Adults
Research design and methods: We included data from 199,042 adults with type 2 diabetes in the US Veterans Affairs Health Care System initially treated with metformin monotherapy from 2005 to 2013. We used multivariable Cox proportional hazards and linear regression to estimate associations of year of metformin monotherapy initiation with time to second-line diabetes treatment over five years of follow-up (primary outcome) and with hemoglobin A1c (HbA1c) at the time of second-line diabetes treatment initiation (secondary outcome).
Results: The cumulative five-year incidence of second-line medication initiation declined from 47% among metformin initiators in 2005 to 36% in 2013 counterparts (P<0.0001) despite a gradual increase in mean HbA1c at the end of follow-up (from 6.94±1.28% to 7.09±1.42%, trend P<0.0001). Compared to metformin monotherapy initiators in 2005, adjusted hazard ratios for five-year initiation of second-line diabetes treatment ranged from 0.90 [95% CI 0.87, 0.92] in 2006 metformin initiators to 0.68 [0.66, 0.70] in 2013 counterparts. Among those receiving second-line treatment within five years of metformin initiation, HbA1c at second-line medication initiation increased from 7.74±1.66% in 2005 metformin initiators to 8.55±1.92% in 2013 counterparts (trend P<0.0001).
Conclusions: We observed progressive delays in diabetes treatment intensification consistent with therapeutic inertia. Process of care interventions early in the diabetes disease course may be needed to reverse adverse temporal trends in diabetes care.