Kidney Week

Abstract: TH-PO777

Factors Associated with Gastrointestinal Bleeding in ESRD Patients

Session Information

• Dialysis: Epidemiology, Outcomes, Clinical Trials - Non-Cardiovascular - I
  November 02, 2017 | Location: Hall H, Morial Convention Center
  Abstract Time: 10:00 AM - 10:00 AM

Category: Dialysis

• 607 Dialysis: Epidemiology, Outcomes, Clinical Trials - Non-Cardiovascular

Authors

• Blanchard, Thomas C, Fresenius Medical Care North America, Waltham, Massachusetts, United States

Thomas C Blanchard,

• Nash-Hampton, Teresa A, Fresenius Medical Care North America, Waltham, Massachusetts, United States

Teresa A Nash-Hampton,

• Goulatia, Ashish, Fresenius Medical Care North America, Waltham, Massachusetts, United States

Ashish Goulatia,

• Ray, Jay, Fresenius Medical Care North America, Waltham, Massachusetts, United States

Jay Ray,

• Usvyat, Len A., Fresenius Medical Care North America, Waltham, Massachusetts, United States

Len A. Usvyat,

• Maddux, Dugan, Fresenius Medical Care North America, Waltham, Massachusetts, United States

Dugan Maddux,

• Maddux, Franklin W., Fresenius Medical Care North America, Waltham, Massachusetts, United States

Franklin W. Maddux,

Background

Gastrointestinal (GI) bleeding is a serious problem among patients with End Stage Renal Disease (ESRD), affecting this population at rates almost two orders of magnitude higher than in the general population (Yang et al., 2012). We aimed to characterize the demographic, treatment, and laboratory parameters in dialysis patients with GI bleeds and build prediction models that could identify patients at a higher risk of experiencing a GI bleed.

Methods

We analyzed data on all dialysis patients treated at Fresenius Medical Care North America dialysis clinics as of December 2016. We used a logistic regression with 40 variables including demographics, comorbidities, treatment parameters and laboratories to identify factors highly associated with GI bleeds and measured the effect size. Variables on treatment and laboratory parameters were determined from mean values at 90 to 183 days prior to December 2016.

Results

We studied data from 141,973 dialysis patients, of which 0.7% were diagnosed with a GI bleed. We found that lower hemoglobin, calcium, and transferrin saturation (TSAT) values were associated with GI bleeds. Variability in calcium and hemoglobin (using standard deviation) was also associated with GI bleeds. Further, GI bleeds were more common among those diagnosed with ulcers and with older patients. When tested on a held-out set of 50% of the data, the model was able to accurately predict which patients have a GI bleed with a receiver operating characteristic area under the curve of 0.81.

Conclusion

Our analysis identifies several factors associated with GI bleeds in dialysis patients, which may be useful for predicting patients who are likely to experience a GI bleed. Further research validating the accuracy of the prediction of patients with GI bleed is needed to confirm these findings.

Funding

• Commercial Support –