CA 19-9 is associated with poor glycemic control in diabetic patients: role of insulin resistance.

BACKGROUND

CA19-9 is considered a tumor marker. Reports have suggested higher CA19-9 levels in diabetic patients even with no malignancy. Our aim was to reveal the relation of CA19-9 with glycemic control in diabetic and non-diabetic subjects. For the first time we considered diabetes as an outcome based on a cut-off where the association of CA19-9 with diabetes is considerable.

METHODS

The study was carried out at an outpatient metabolism clinic. A total of 422 consecutive participants were enrolled. Subjects with prior diagnosis of any cancer, renal, endocrine, or hepatic problems were not included. Age, gender, and medications as well as weight, height, and BMI were recorded. Creatinine, fasting plasma glucose (FPG), post-prandial plasma glucose (PPPG), fasting plasma insulin, HbA1c, C-peptide, HOMA-IR, and CA 19-9 were measured.

RESULTS

In all, 266 (63.03%) subjects had recently been diagnosed with diabetes. CA19-9 was significantly higher in the diabetic group (16.73 +/- 13.83 vs. 11.93 +/- 11.42, p < 0.001). BMI, waist circumference, FPG, and HbA1c were higher in quartiles with greater CA19-9 levels. Number of diabetic subjects in each quartile had a stepwise increase (48%, 56%, 72%, and 77%, p < 0.01). FPG, PPPG, HbA1c, and HOMA-IR were directly correlated with CA19-9 levels independent of age, gender, and BMI. We presume the 10.83 U/mL value for CA19-9 to be the optimal cut-off in indicating diabetes status (sensitivity: 0.63, specificity: 0.55).

CONCLUSIONS

Otherwise normal diabetic subjects have greater CA19-9 values. CA19-9 should be interpreted with regard to diabetes status. We suggest that CA19-9 levels above 10.83 U/mL in the absence of other pathologies are in favor of glycemic impairments. CA19-9 values greater than 34.30 U/mL may accompany an 84% frequency of diabetic subjects especially in settings such as referral metabolism clinics. CA19-9 values of less than 6.46 U/mL are likely to rule out the presence of diabetes, especially while testing in general population.