Gender disparity in fistula use at initiation of hemodialysis varies markedly across ESRD networks-Analysis of USRDS data.

BACKGROUND

Gender disparities had been noted in the care of women with end stage renal disease (ESRD) in the early 2000's, including less frequent initiation of hemodialysis utilizing a fistula but more recent data have not been examined and underlying factors have not been extensively studied.

STUDY

DESIGN

Data from the United States Renal Data System (USRDS) were examined, including 202,999 hemodialysis patients. Only those who had received prior nephrology care were included. Multiple logistic regression was used, adjusted for possible confounders, including age, race, cause of ESRD, BMI, height, history of alcohol or drug abuse, medical comorbidities, ability to ambulate, time of nephrology care, type of insurance, and ESRD network.

RESULTS

The odds of arteriovenous fistula (AVF) use at initiation of hemodialysis were significantly lower in women compared to men (OR = 0.69, 95% CI 0.67-0.71, P < 0.0001). The gender gap in AVF use at initiation was highest in New York and the upper Midwest (networks 2 and12) and smallest for Southern California and the Pacific Northwest and Alaska (18 and 16). Gender disparity was more pronounced for black women, with odds ratios for AVF use at initiation of dialysis (OR = 0.66, 95% CI 0.62-0.69), P < 0.0001 as compared to non-black (OR 0.70, 95% CI 0.68-0.73), P ≤ 0.0001.

LIMITATIONS: Limitations include use of USRDS data. Data misclassification or errors in data reporting may exist and certain comorbid conditions may be underreported. Data regarding rate of primary fistula non-function are also not available.

CONCLUSION

Adjusted odds ratio for AVF use was significantly lower in women compared to men, independent of time of nephrology care and other predictors. The gender disparity was most pronounced for black women and also varied from 20% to 46% lower odds for AVF use in women for different ESRD networks, after controlling for possible confounding variables, suggesting that practice based factors may be of importance in explaining this important finding.