Impaired Fasting Glucose and Diabetes Have Same Risk of Retinopathy

The 10-year prevalence of retinopathy was nearly identical among individuals with impaired fasting glucose as among those with overt diabetes at baseline in a longitudinal French study of 700 adults.

The findings provide further evidence that calls into question the current definition of diabetes, which was based on glycemic thresholds for retinopathy and nephropathy derived from cross-sectional studies conducted in the 1990's.

Dr. Pascale Massin said at the annual meeting of the European Association for the Study of Diabetes (EASD) that recent studies using more sensitive methods for assessing retinopathy are showing that, just as with cardiovascular disease, the relationship between blood glucose and retinopathy varies across a continuum.

The study population for the Data from an Epidemiological Study on the Insulin Resistance Syndrome (DESIR) comprised 5,212 men and women who were 30-65 years of age at baseline in 1994-1996. Recruited from 10 French Social Security health examination centers, the participants in the main DESIR study underwent health examinations at entry and 3, 6, and 9 years later. At 10 years, 733 participants were invited to return for a special examination to detect retinopathy using a nonmydriatic camera.

Of the 700 study participants for whom the eye exam findings were evaluable, 235 had diabetes. Of those, 36 were already receiving treatment for diabetes prior to the study, while the other 199 were found to have at least one fasting plasma glucose level of 126 mg/dL or higher during the study.

Two other age-, sex-, and center-matched groups were selected from the main DESIR study population: 238 who had normal fasting glucose (defined at the time as less than 110 mg/dL) throughout the study, and 227 who had impaired fasting glucose (IFG, defined as 110-125 mg/dL) at some point during the study.

Retinopathy was defined as microaneurysms, retinal hemorrhages, exudates, cotton wool spots, and/or neovascularization based on three photographs per eye: one centered on the macula, one temporal to the macula, and one on the optic disk.

Of the 44 people who were classified as having retinopathy, 19 had diabetes (including 5 being treated at baseline), 19 had IFG at least once during the study, and 6 had normal glucose levels throughout the 9 years of the main DESIR study.

The retinopathy was mild in most patients, with just microaneurysms and/or retinal hemorrhages.

Baseline characteristics that were significantly predictive of retinopathy included systolic blood pressure (142 mm Hg in those who developed retinopathy vs. 137 mm Hg in those who did not), mean fasting plasma glucose (130 vs. 106 mg/dL), and hemoglobin A1c (6.4% vs. 5.7%).

The two factors most strongly associated with retinopathy at 10 years were fasting plasma glucose and HbA1c during the first 9 years of follow-up. Triglyceride and C-reactive protein levels also were associated with retinopathy, but the significance level was lower, she said.

The overall prevalence of retinopathy at 10 years was 3.6%. The prevalence of retinopathy was 2.5% among those who remained normoglycemic, 8.4% among those with IFG, and 8.1% among the patients with diabetes.

The positive predictive value (PPV) for retinopathy was stronger for baseline HbA1c than for fasting plasma glucose.

For a baseline HbA1c cutoff of 6.0%, the PPV was 6.8%; for an HbA1c cutoff of 6.5%, the PPV was 15.9%. For a baseline fasting glucose cutoff of 108 mg/dL, the PPV was 8.6%, and for a cutoff of 117 mg/dL, the PPV was 17.4%.

The levels at which the PPV increased -- a fasting plasma glucose of 117 mg/dL and an HbA1c of 6.5% -- could provide a rationale for the choice of thresholds for the definition of hyperglycemia associated with 10-year retinopathy risk, and also may be indicative of diabetes risk.

These findings are in line with those of a recent analysis of three population-based cross-sectional studies, which found that the prevalence of retinopathy increased with plasma glucose levels, with no clear cutoff (Lancet 2008; 371:736-43).

The authors of that study, Dr. Tien Y. Wong of the Centre for Eye Research Australia, University of Melbourne, and associates, concluded that, "The current FPG cutoff of [126 mg/dL] used to diagnose diabetes did not accurately identify people with and without retinopathy. These findings suggest that the criteria for diagnosing diabetes could need reassessment."