More Frequent Hemoglobin Measurements and More Frequent Epoetin Alfa Titrations Are Both Associated with Increased Epoetin Alfa Dose but Not More Time in Target Hemoglobin Range,

Abstract 4194

For the past 20 years, management of hemoglobin (Hb) level via erythropoiesis-stimulating agents (ESAs), such as epoetin alfa (EPO) has been the norm in the care of end-stage renal disease (ESRD) patients. Recently, questions about patient safety and new economic bundling rules have brought renewed focus on appropriate Hb targets and how they relate to the use of EPO. Important issues related to Hb control are the frequency of Hb measurements, Hb value, the frequency of changes in EPO dose (titrations), and total EPO dose. Current practice aims to achieve greater control of Hb by more frequent measurement and finer EPO dose titrations. No studies have definitively shown that levels of titration and measurement frequency lead to better clinical outcomes or more efficient use of EPO. One large retrospective analysis found increasing Hb measurements and EPO dose titration frequency decreased patient variability around the facility-level Hb mean (Khan et al, 2011). However, a second study found that frequent titrations have been shown to be the most important driver of Hb cycling—potentially dangerous large fluctuations in Hb levels (Fishbane and Berns, 2005). The issue remains controversial.

We conducted a retrospective database analysis in order to quantify the frequency of Hb dose titrations and measurements, and to evaluate their relationship to ESA dose and the ability to keep patients within a Hb range of 10–12 g/dL. Data from prevalent (≥ 120 days), adult (>18 years old) hemodialysis patients dialyzing at DaVita dialysis clinics ≥ 3 times per week between January 1, 2009 and December 31, 2010 were included.

For every session, patients' stable dosing periods were defined as any series of at least 3 doses during which the dose did not change more than 10% from first dose in the series. Dose holds were defined as ≥ 3 consecutive sessions with 0 EPO dose. An EPO dose of 0 lasting <3 consecutive sessions was ignored when identifying stable periods and holds. Doses not categorized as part of a stable period or a dose hold were considered transition periods. A dose titration was defined as a difference of > 10% between any of the following: the mean dose of 2 consecutive stable periods; the mean dose of stable period and next/previous dose in transition period; and 2 consecutive doses within a transition period. Transitions involving a dose hold were counted as 1 titration if the patient then returned to a dose within 10% of the mean of the prior period; or 2 titrations otherwise. Similar rules were developed for the minority of patients who received EPO once or twice per week. Correlations among measures (titrations/patient-month and Hb measurements/patient-month with mean monthly EPO dose and time with Hb in range) were assessed using Pearson product-moment correlation and linear regression (adjusted for racial composition, percentage of dialysis vascular access types, percentage with comorbidities, and mean age, vintage, and BMI) at the physician and the facility levels. Time in range was defined as total patient time in range/total patient-time. Facility level analyses were weighted by facility load (minimum of 300 patient-months) and are reported here.

Information from 81,464 patients at 1,336 facilities was assessed. The mean number of titrations was 13.6 ± 2.83 (mean ± SD) per patient per year and the mean number of Hb measurements was 36.3 ± 8.24 per patient per year. The mean percent of patient-time in range among these facilities was 57.1% ± 5.8%. At the facility level, after adjustment for case mix factors (mean age, vintage, and BMI at facility, access types, and racial composition) the annual number of dose titrations per patient and Hb measurements per patient were not associated with the percent of patient-months within a 10–12 g/dL Hb range (p=0.12 and p=0.47; respectively). An analysis of ESA utilization showed that an increase of 1 titration per patient at the facility level was associated with an extra 18,000 U (p<0.001) per patient.

In this retrospective study of >80,000 US dialysis patients over 2 years, neither increased Hb measurements nor increased EPO titrations were associated with improving patient-time in Hb range. However, increased EPO titrations were associated with significantly increased EPO utilization. Although these associations do not demonstrate causality (or the direction of potential causality), they do underscore the need to assess titration practices.