Glycemic control paradox: Poor glycemic control associated with higher one-year and eight-year risks of all-cause hospitalization but lower one-year risk of hypoglycemia in patients with type 2 diabetes.

OBJECTIVE

The relationship between glycemic control and adverse outcomes found in a population with diabetes has seldom been evaluated in patients with type 2 diabetes. We explored the association between hemoglobin A1c (HbA1c) and hospitalization risks within one-year and eight-year follow-up periods.

METHODS

We conducted a retrospective cohort study on 57,061 patients with type 2 diabetes from National Diabetes Case Management Program during 2002-2004 in Taiwan. HbA1c at baseline and in-hospital mortality, all-cause and cause-specific hospitalization over one year and eight years were analyzed.

RESULTS

After multivariate adjustment, one-year risk was higher for cases with HbA1c level <6%, 9-10%, ≥10% versus 6-7% for all-cause hospitalization (hazard ratio [HR]: 1.11, 95% confidence interval [CI]: 1.03-1.20; 1.08, 1.01-1.16, and 1.19, 1.12-1.26, respectively) and for ≥10% for diabetes-related hospitalization (1.68, 1.46-1.92). Yet each 1-step increment in HbA1c category (<6.0, 6.0-6.9, 7.0-7.9, 8.0-8.9, 9.0-9.9 and ≥10.0%) showed linkage with lower risk of hypoglycemia hospitalization (0.81, 95% CI: 0.74-0.88). For eight-year risk, subjects with HbA1c level <6%, and ≥10% were more likely to have in-hospitality mortality (1.16, 1.03-1.31, and 1.23, 1.11-1.35, respectively). Each 1-step increment in HbA1c category showed an association with higher risks of all-cause and diabetes-related hospitalization (1.04, 1.03-1.05, and 1.15, 1.14-1.17, respectively).

CONCLUSIONS

Higher HbA1c level correlated with lower one-year risk due to hypoglycemia hospitalization but increased one-year and eight-year risks due to all-cause and diabetes-specific hospitalization among Chinese people with type 2 diabetes in Taiwan. Future study must ascertain how to meet HbA1c targets and improve outcome without risk to this population.