Abstract: FR-PO521

Functional Iron Deficiency and Incident Diabetes Risk in US Veterans with Pre-Dialysis CKD

Session Information

Category: Chronic Kidney Disease (Non-Dialysis)

  • 304 CKD: Epidemiology, Outcomes - Non-Cardiovascular

Authors

  • Cho, Monique E., Veterans Health Administration, Salt Lake City, Utah, United States
  • Hansen, Jared, Veterans Health Administration, Salt Lake City, Utah, United States
  • Peters, Celena B., Veterans Health Administration, Salt Lake City, Utah, United States
  • Sauer, Brian C., Veterans Health Administration, Salt Lake City, Utah, United States
Background

While the association between iron overload and diabetes is well established, the possible adverse metabolic effect of iron deficiency, both functional and absolute, has not been investigated in any population, including those with CKD.

Methods

We performed a historical cohort study using the national data from the Veterans Affairs Informatics and Computing Infrastructure. We identified a non-diabetic, pre-dialysis CKD cohort (MDRD eGFR <60 mL/min/1.73m2) with at least one set of iron indices between 2006-2015. The clinical characteristics were determined from the ICD-9 codes and laboratory data during the baseline period, defined as the year preceding the first available iron indices. Patients with diabetes, ESRD, genetic and chronic disorders affecting iron metabolism were excluded. The cohort was divided into 4 iron groups based on the joint quartiles (Q) of transferrin saturation (Tsat) and ferritin: FID, 1st Tsat Q + 3rd−4th ferritin Qs; Low Iron (LI) reflecting absolute iron deficiency, 1st Tsat+ferritin Qs; High Iron (HI), 4th Tsat+ferritin Qs; and Reference (R), 2nd−3rd Tsat+ferritin Qs. Incident diabetes was determined by the ICD-9 codes. Matching weights were used to determine the effect of iron status on incident diabetes, using R as the comparison group.

Results

Of the 1,159,371 Veterans with CKD, 68,728 met the inclusion criteria. The mean±SD for age and eGFR were 74±12 years and 43±11 mL/min/1.73 m2, respectively. The median (IQR) Tsat and ferritin values were 18 (13, 23)% and 98 (53, 150) ng/mL. Of the study cohort, 44% could not be categorized into any of the 4 iron groups. In the remaining 38,428 Veterans, the prevalence for FID, HI, LI, and R were 12%, 21%, 20%, and 47%, respectively. After matching weights were implemented, the covariates were evenly distributed among the iron groups. During the mean±SD follow-up period of 4.0±2.7 years, only FID was associated with increased risk for incident diabetes [Risk Ratio, RR (95% CI), 1.20 (1.09, 1.33)], after successful matching for age, sex, race, BMI, hyperlipidemia, CV history, albumin, hemoglobin, and eGFR. The RRs for HI and LI were 0.98 (0.90, 1.07)] and 1.04 (0.95, 1.13), respectively.

Conclusion

FID, but not absolute iron deficiency, is associated with increased risk for incident diabetes in CKD. Further studies are needed to investigate the underlying mechanisms.

Funding

  • Veterans Affairs Support