Reversion from prediabetes to normoglycaemia and risk of cardiovascular disease and mortality: the Whitehall II cohort study

Abstract

Aims/hypothesis: Reversion from prediabetes to normoglycaemia is accompanied by an improvement in cardiovascular risk factors, but it is unclear whether this translates into a reduction in risk of cardiovascular disease (CVD) events or death. Hence, we studied the probability of reversion from prediabetes to normoglycaemia and the associated risk of future CVD and death using data from the Whitehall II observational cohort study.

Methods: Three glycaemic criteria for prediabetes (fasting plasma glucose [FPG] 5.6-6.9 mmol/l, 2 h plasma glucose [2hPG] 7.8-11.0 mmol/l, and HbA_1c 39-47 mmol/mol [5.7-6.4%]) were assessed in 2002-2004 and 2007-2009 for 5193 participants free of known diabetes at enrolment. Among participants with prediabetes in the first examination, we calculated the probability of reversion to normoglycaemia by re-examination according to each glycaemic criterion. Poisson regression analysis was used to estimate and compare incidence rates of a composite endpoint of a CVD event or death in participants with prediabetes who did vs did not revert to normoglycaemia. Analyses were adjusted for age, sex, ethnicity and previous CVD.

Results: Based on the FPG criterion, 820 participants had prediabetes and 365 (45%) of them had reverted to normoglycaemia in 5 years. The corresponding numbers were 324 and 120 (37%) for the 2hPG criterion and 1709 and 297 (17%) for the HbA_1c criterion. During a median follow-up of 6.7 (interquartile range 6.3-7.2) years, 668 events of non-fatal CVD or death occurred among the 5193 participants. Reverting from 2hPG-defined prediabetes to normoglycaemia vs remaining prediabetic or progressing to diabetes was associated with a halving in event rate (12.7 vs 29.1 per 1000 person-years, p = 0.020). No association with event rate was observed for reverting from FPG-defined (18.6 vs 18.2 per 1000 person-years, p = 0.910) or HbA_1c-defined prediabetes to normoglycaemia (24.5 vs 22.9 per 1000 person-years, p = 0.962).

Conclusions/interpretation: Most people with HbA_1c-defined prediabetes remained prediabetic or progressed to diabetes during 5 years of follow-up. In contrast, reversion to normoglycaemia was frequent among people with FPG- or 2hPG-defined prediabetes. Only reversion from 2hPG-defined prediabetes to normoglycaemia was associated with a reduction in future risk of CVD and death.

Keywords: 2 h Plasma glucose; Cardiovascular disease; Fasting plasma glucose; HbA1c; Mortality; Normoglycaemia; Prediabetes; Reversion.

Fig. 1

Rate ratios (RRs) of an event (CVD or death) for reverting from prediabetes to normoglycaemia vs not reverting (a) or for decreasing 1 SD in glycaemic measure over 5 years from phase 7 to phase 9 (b). Grey: unadjusted RR; light blue: adjusting for age and sex; dark blue: further adjusting for previous CVD. The RR for 1 SD decrease (b) is further adjusted for baseline glycaemia in all the analyses. The x-axis is on a natural logarithmic scale