The purpose of this study is to prospectively identify blood markers that identify chronic hemodialysis patients at risk for early (<90 days) and late (>= 1 year) mortality

The annual incidence of patients initiating chronic hemodialysis (HD) will double (~100K→~200K) over the next decade. Despite several advances, HD mortality rates remain extremely high and have not dramatically changed over the past several years. The United States Renal Data System (USRDS) is the largest epidemiological database of patients with end-stage renal disease (ESRD) throughout the US, and studies utilizing the USRDS and similar studies emanating from large dialysis providers have served as the benchmark for guiding the care of all dialysis patients in the US While these databases have been an excellent resource for testing hypotheses, lack of concomitant blood samples has prevented a large-scale clinical trial testing of hypotheses necessitating biological samples. To overcome this limitation, we established a 'USRDS-like' clinical trial with the largest US dialysis provider to prospectively acquire research quality demographic data and concomitant biological samples to test two specific aims:

AIM 1: HD patients are profoundly vitamin D deficient, yet ~60% of incident hemodialysis patients in the US are not treated with active vitamin D. Vitamin D is linked with cardiovascular health, and although higher serum vitamin D levels are associated with improved survival in different settings (e.g., cancer, infection), no large study has examined whether baseline vitamin D levels are associated with HD mortality. Measuring serum vitamin D levels among HD patients is sporadic since the utility of this measurement as it relates to dialysis mortality is unclear.

Hypothesis: Higher vitamin D levels at initiation of HD are associated with lower risk of 90-day and 1-year mortality among patients who do not receive vitamin D therapy.

AIM 2: Among diabetic non-HD patients, blood levels of HbA1c predict all cause and cardiovascular related mortality. Diabetes is the leading cause of ESRD in the US (~50% of all ESRD patients) and mortality rates of diabetics on HD are ~10-20% higher than non-diabetics. Leading Nephrology thought leaders and regulatory bodies (tied to reimbursement and quality control) suggest HbA1c levels must be routinely checked in diabetic HD patients. Importantly, the association between HbA1c levels and all-cause and cardiovascular mortality among HD patients is unclear (in contrast to non-HD data), and erythropoietin therapy and shortened red-cell survival (common in HD patients) render measurements of HbA1c less meaningful.

Hypothesis: Elevated HbA1c levels (among diabetics) at the initiation of HD are associated with increased 90 day mortality, however, HbA1c measurements performed thereafter, e.g., at 90 days following the initiation of HD, are not associated with subsequent 1 year survival.

These aims are examined in a prospective cohort study (of incident HD patients living throughout the US (>1000 centers) with detailed demographic data updated daily, and research quality blood samples collected every 3 months from the initiation of chronic HD. We initiated this "USRDS-like" study to overcome one of the largest limitations of existing epidemiological dialysis datasets - lack of concomitant biological samples in a nationally representative cohort. The study will be available for data sharing and patient protection measures are in place.

Inclusion Criteria:

• End stage renal disease on hemodialysis at Fresenius Medical Center
• New initiation of hemodialysis

Exclusion Criteria:

• Death with 14 days of initiation of hemodialysis