Impact of Pharmacist-Managed Erythropoiesis-Stimulating Agents Clinics for Patients With Non–Dialysis-Dependent CKD

Original Investigation Impact of Pharmacist-Managed Erythropoiesis-Stimulating Agents Clinics for Patients With Non–Dialysis-Dependent CKD Sherrie L. Aspinall, PharmD, MSc,1,2,3 Francesca E. Cunningham, PharmD,1 Xinhua Zhao, PhD,2 Joy S. Boresi, PharmD,4 Ivy Q. Tonnu-Mihara, PharmD, MS,5 Kenneth J. Smith, MD, MS,6 Roslyn A. Stone, PhD,2,7 and Chester B. Good, MD, MPH,1,2,3,8 on behalf of the ESA Clinic Study Group* Background: Erythropoiesis-stimulating agents (ESAs) are associated with serious adverse events, and maintaining hemoglobin levels within a narrow range can be difficult. We examined the quality of ESA prescribing and monitoring in pharmacist-managed ESA clinics versus usual care in patients with non– dialysisdependent chronic kidney disease (NDD-CKD). Study Design: Historical cohort. Setting & Participants: Outpatients receiving ESAs for NDD-CKD at 10 Veterans Affairs Medical Centers with both pharmacist-managed ESA clinics (n ⫽ 314) and physician-based care (ie, usual care; n ⫽ 91) and 6 sites with usual care only (n ⫽ 167) on January 1, 2009, were followed up for 6 months. Predictor: Type/site of care (ie, pharmacist-managed ESA clinic, usual care at ESA clinic site, usual-care site). Outcomes: Primary outcomes were proportion of hemoglobin values in the target range of 10-12 g/dL, ESA dose, and frequency of hemoglobin monitoring. Factors associated with hemoglobin values out of target range were identified using multinomial logistic regression. Results: More hemoglobin values were in the target range in pharmacist-managed ESA clinics (71.1% vs 56.9% for usual-care sites; P ⬍ 0.001). The average 30-day dose of darbepoetin was 163 ␮g in pharmacist-managed ESA clinic patients versus 240 ␮g in usual-care site patients and 258 ␮g in usual-care patients at ESA clinic sites. For epoetin, corresponding average 30-day doses were 44,890 versus 47,141 and 57,436 IU. Veterans in pharmacist-managed ESA clinics had more hemoglobin measurements on average (5.8 vs 3.6 in usual-care sites and 3.8 in usual care at ESA clinic sites; P ⫽ 0.007). In the multinomial model, usual care was associated with hemoglobin levels out of target range, whereas heart failure and diabetes were associated with values in range. Limitations: We could not assess whether different hemoglobin targets were used by usual-care providers. Conclusions: Relative to usual care, pharmacist-managed clinics provided improved quality of ESA dosing and monitoring for patients with NDD-CKD. Am J Kidney Dis. 60(3):371-379. Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. INDEX WORDS: Chronic kidney disease; anemia; erythropoietin.

Editorial, p. 340

E

rythropoiesis-stimulating agents (ESAs) are widely used in managing the anemia of chronic kidney disease (CKD), but have potential for serious adverse events.1-3 In addition, adjusting the dosage to maintain hemoglobin levels within target range is challenging.4 Pharmacists have successfully managed patients with anemia and CKD in programs in which From the 1Veterans Affairs Center for Medication Safety, Hines, IL; 2Center for Health Equity Research and Promotion, Veterans Affairs Pittsburgh Healthcare System; 3University of Pittsburgh, School of Pharmacy, Pittsburgh, PA; 4St Louis Veterans Affairs Medical Center, St Louis, MO; 5Veterans Affairs Long Beach Healthcare System, Long Beach, CA; 6Division of Clinical Modeling and Decision Sciences, 7Graduate School of Public Health, and 8School of Medicine, University of Pittsburgh, Pittsburgh, PA. * A list of the members of the ESA Clinic Study Group appears in the Acknowledgements. Received December 13, 2011. Accepted in revised form April 14, 2012. Originally published online May 28, 2012. Am J Kidney Dis. 2012;60(3):371-379

they generally were responsible for obtaining laboratory tests, adjusting ESA doses, recommending iron therapy, and educating patients.5-13 Although these studies reported improvement in outcomes, they had small sample sizes from single sites and were uncontrolled or pre-post comparisons. Whether the benefit of pharmacist involvement persists with more patients and across multiple facilities is unclear. In this study, we compare the quality of ESA prescribing and monitoring for patients with non–dialysis-dependent CKD (NDD-CKD) Address correspondence to Sherrie L. Aspinall, PharmD, MSc, Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System, 7180 Highland Dr (151C-H), Pittsburgh, PA 15206. E-mail: [email protected] Published by Elsevier Inc. on behalf of the National Kidney Foundation, Inc. This is a US Government Work. There are no restrictions on its use. 0272-6386/$0.00 http://dx.doi.org/10.1053/j.ajkd.2012.04.013

371

Aspinall et al

in Veterans Affairs (VA) Medical Centers with and without pharmacist-managed ESA clinics.

METHODS Study Setting and Population Pharmacists from 16 VA Medical Centers volunteered to participate; they were from 10 sites with pharmacist-managed ESA clinics and 6 medical centers with physician-based care (ie, usual care). Each site included a maximum of 50 randomly selected outpatients who were receiving ESAs for NDD-CKD on January 1, 2009. Patients were followed up through June 30, 2009. A maximum was specified because we did not want sites with larger patient populations to overly influence results. During this time, most sites (8/10) with pharmacist-managed ESA clinics had patients who were not followed up in that clinic; these patients were classified as receiving “usual care at ESA clinic sites.” NDD-CKD was defined as estimated glomerular filtration rate (eGFR) ⬍60 mL/min/1.73 m2. Veterans who were “dual managed” (ie, VA and non-VA providers co-managing the ESA) at usual-care sites were included; this type of care is not permitted in pharmacistmanaged ESA clinics. Because only 4%-6% of patients in each group were new to ESA therapy (ie, started ESA therapy in December 2008), our analysis focuses on patients who were receiving ESA treatment on a long-term ongoing basis. No patient was excluded based on the number of ESA doses received or hemoglobin tests ordered. Institutional review boards for participating sites and the VA Center for Medication Safety approved the study.

Description of Pharmacist-Managed ESA Clinics Pharmacists’ scope of practice allowed them to dose and monitor ESA therapy. Patients at most sites were referred to the pharmacist-managed ESA clinic by a medical provider. Study clinics had been operational since at least August 1, 2008, and evidence-based protocols were in place for the management of patients receiving ESAs.

Data Sources and Data Collection At each site, outpatients with an active ESA prescription on January 1, 2009, were identified by searching the Veterans Health Information Systems and Technology Architecture or their local data warehouse. Pharmacists used the VA electronic medical record, Computerized Patient Record System (CPRS), to ascertain the ESA indication; only veterans who received an ESA for NDD-CKD were included in this study. All patient-level data were collected retrospectively from the CPRS. Baseline data included patient demographics, eGFR, and comorbid conditions. Study period data included specific ESAs (ie, darbepoetin or epoetin), iron and antihypertensive medications prescribed (including dose, dosing instructions, and start/stop dates), and dates and results of hemoglobin, hematocrit, transferrin saturation, serum iron, and ferritin tests. Dates for death, start of hemodialysis therapy, and discharge from the clinic also were recorded, if applicable.

Primary and Secondary Outcome Measures The primary outcome was the quality of ESA prescribing, which we defined operationally as the proportion of hemoglobin values within the target range of 10-12 g/dL. At the time of the study, the US Food and Drug Administration (FDA) recommended hemoglobin levels of 10-12 g/dL in patients with CKD.14 This guidance was consistent with the 2007 update of the NKF-KDOQI (National Kidney Foundation’s Kidney Disease Outcomes Quality Initiative) guidelines, which recommend a hemoglobin target generally in the 372

range of 11-12 g/dL.15 Because published data suggest that “higher” ESA doses and fluctuations in hemoglobin level may contribute to adverse outcomes,4,16 we also compared ESA dosage and intraand interpatient variance in hemoglobin values as quality of prescribing measures. The secondary outcome was the quality of ESA monitoring, which we defined operationally as the average number of hemoglobin and iron tests per patient during the 6-month study period based on the 2006 NKF-KDOQI guidelines. Specifically, those guidelines recommend monitoring hemoglobin levels at least monthly and iron status tests at least every 3 months.17

Clinical Outcomes Dates of emergency department visits and hospitalizations for adverse events potentially related to ESA treatment (eg, myocardial infarction, stroke, and heart failure) and episodes of uncontrolled hypertension (defined as systolic blood pressure ⬎160 mm Hg or diastolic blood pressure ⬎100 mm Hg) as documented in outpatient records during the study period were recorded.

Statistical Analysis Our primary comparison of interest was the quality of care provided by pharmacist-managed ESA clinics versus usual-care sites, but we also evaluated usual care at sites with an ESA clinic. Baseline patient characteristics were compared using ␹2 tests for categorical variables or t tests for continuous variables. To describe the quality of ESA prescribing, we summarized the proportion of hemoglobin values that were within, less than, and greater than the target range of 10-12 g/dL. Random-effect multinomial models with clustering at the patient and site levels were used to compare the groups. To assess provider response to values outside the target range, we calculated and compared the average daily ESA dose during the 30 days before and after each hemoglobin measurement to ascertain whether the ESA dose was increased after a hemoglobin value less than target or held/decreased after a value greater than target. Random-effect logistic regression models with clustering at the patient and site levels were used to compare the groups. To compare ESA doses, we calculated patient-level average 30-day doses during the study period. Given the wide range and skewness of doses, we log transformed the data and tested for group differences using randomeffect linear regression models with clustering at site-level; we displayed the averaged 30-day doses of darbepoetin and epoetin in each group by kernel density plots (smoothed histograms). Finally, fluctuations in hemoglobin values within and between patients were assessed separately for each group using a random-effects linear model with clustering at the patient level. To describe the quality of ESA monitoring, we assessed the frequency and average number of hemoglobin and iron studies per patient during the 6-month period. Iron studies included measurement of ferritin, serum iron, or transferrin saturation. Mean values were compared using random-effect Poisson models. We also evaluated whether hemoglobin test was frequency associated with being in target range. For each type/site of care, we used random-effect Poisson models with clustering at the site level to compare the total number of hemoglobin tests in patients with at least one hemoglobin value out of target range with patients with no out-of-range values. For the clinical outcome measures of adverse events and uncontrolled hypertension, we estimated rates (per 180 person-days). The study was not powered to detect differences in adverse events by type of care. A random-effect multinomial logistic regression model with clustering at patient and site levels was used to identify factors associated with hemoglobin values less than and greater than target. Hemoglobin values within target range (ie, 10-12 g/dL) were chosen as the base category for these comparisons, and 2 binary logits (⬍10 vs 10-12 Am J Kidney Dis. 2012;60(3):371-379

Pharmacist-Managed ESA Clinics g/dL and ⬎12 vs 10-12 g/dL) were estimated simultaneously. Data management was done using SAS software (SAS Institute, Inc, www.sas.com), version 9.2, and all models were run in Stata (StataCorp LP, www.stata.com), version 11.

RESULTS Participant Characteristics Except for age, racial, and ethnicity differences between veterans managed in ESA clinics and usualcare sites, baseline patient characteristics were similar in all 3 groups (pharmacist-managed ESA clinics, usual-care sites, and usual care at ESA clinics; Table 1). Mean patient age ranged from 73.9-78.4 years, and most were men (96.7%-98.1%). Similar proportions of patients were dual-managed in usual-care sites (22.2%) and usual care at ESA clinic sites (17.6%). Patients at sites with pharmacist-managed ESA clinics were mainly from the South (67% in pharmacistmanaged ESA clinic patients and 83.5% in usual-care patients at ESA clinic sites), whereas usual-care patients were from the Northeast and Midwest (44.9% and 55.1%, respectively). Quality of ESA Prescribing Compared with both usual-care groups, more hemoglobin values were in the target range of 10-12 g/dL in pharmacist-managed ESA clinics (71.1% vs 56.9% at usual-care sites [P ⬍ 0.001] and 51.7% in usual care at ESA clinic sites; Table 2). Fewer hemoglobin values were ⬎12 g/dL in ESA clinic patients (9.6%) versus usual-care site patients (22.1%; P ⬍ 0.001) and patients managed by usual care at ESA clinic sites (16.5%). For hemoglobin values ⬍10 g/dL, ESA clinic pharmacists increased the dose of darbepoetin or epoetin 57.7% of the time compared with 28.6% in usual-care sites (P ⫽ 0.009) and 30.4% in usual care at ESA clinic sites (Table 3). For hemoglobin values ⬎12 g/dL, pharmacists withheld or decreased the ESA dose 61.1% of the time compared with 34.3% (P ⫽ 0.09) in usual-care sites and 18.6% in usual care at ESA clinic sites. ESA dosage distributions ranged widely for the 3 types of care (Fig 1). Average 30-day ESA doses were lower in patients managed by pharmacists in ESA clinics. Using the raw data, the average 30-day dose of darbepoetin was 163 ␮g in pharmacist-managed ESA clinic patients versus 240 ␮g at usual-care sites and 258 ␮g for usual-care patients at ESA clinic sites. For epoetin, the corresponding average 30-day doses were 44,890 versus 47,141 and 57,436 IU. Due to skewness, we compared the groups using log-transformed data. For darbepoetin, geometric mean values were 115 ␮g in pharmacist-managed ESA clinic patients versus 156 ␮g at usual-care sites (P ⫽ 0.049) and 181 ␮g for usual-care patients at ESA clinic sites. For epoetin, the correspondAm J Kidney Dis. 2012;60(3):371-379

ing geometric mean values were 29,677 versus 32,856 (P ⫽ 0.6) and 39,258 IU. Intrapatient variances in hemoglobin values during the study period were similar in pharmacist-managed ESA clinic and usual-care site patients (0.66 and 0.71, respectively), but interpatient variance was lower for patients in the former group (0.51) compared with usual care (1.89). For usual-care patients at ESA clinic sites, both intrapatient (1.03) and interpatient variances (1.46) were higher than those for pharmacistmanaged ESA clinic patients. Quality of ESA Monitoring During the 6-month study period, pharmacistmanaged ESA clinic patients had an average of 5.8 ⫾ 2.6 (SD) hemoglobin measurements versus 3.6 ⫾ 3.0 at usual-care sites and 3.8 ⫾ 2.8 for usual-care patients at ESA clinic sites (P ⫽ 0.007; Table 4). The average number of iron studies per patient was 2.2 ⫾ 1.6 in pharmacist-managed ESA clinics versus 1.0 ⫾ 1.1 in usual-care sites and 1.1 ⫾ 1.2 in usual care at ESA clinic sites (P ⬍ 0.001). We evaluated whether the number of hemoglobin values was associated with being in target range. Of 550 patients with at least one hemoglobin test, 66.2% had at least one hemoglobin value out of target range and 33.8% had all hemoglobin values within target range. Patients who ever had hemoglobin values out of range had on average 1-2 more hemoglobin tests than patients who always had hemoglobin values in the target range across the 3 groups (pharmacist-managed ESA clinic group, average of 6.1 vs 5.1 tests, P ⬍ 0.001; usual-care group, 4.3 vs 3.1, P ⫽ 0.006; and group with usual care at ESA clinic sites, 4.9 vs 2.4, P ⬍ 0.001). The difference in number of tests (ie, ⬃1 test) is similar between the ESA-clinic and usual-care sites (P ⫽ 0.5). Compared with usual-care patients, relatively more patients in pharmacist-managed ESA clinics (64.7% vs 32.3%; P ⫽ 0.08) had at least one transferrin saturation measured during the study period, and somewhat more of those tested (47.8% vs 37.0%; P ⫽ 0.7) had a transferrin saturation value ⬍20%. For usual-care patients at ESA clinic sites, 37.4% had at least one transferrin saturation and 52.9% of patients tested had a transferrin saturation ⬍20%. Similar proportions of patients in pharmacist-managed ESA clinics and usual-care sites received oral (76.8% vs 72.5%; P ⫽ 0.3) and intravenous iron (6.1% vs 4.2%; P ⫽ 0.4). For usual-care patients at ESA clinic sites, 58.2% received oral iron and 4.4% received intravenous iron. Clinical Outcomes Adverse-event rates (per 180 patient-days of ESA therapy) were clinically similar across the 3 types of care (Table 5). Antihypertensive treatment was similar 373

Aspinall et al Table 1. Baseline Characteristics of Patients by Type/Site of Care Pharmacist-Managed ESA Clinic (n ⴝ 314)

Usual-Care Site (n ⴝ 167)

Usual Care at ESA Clinic Site (n ⴝ 91)

Age (y) Age group ⬍65 y 65-84 y ⱖ85 y

73.9 ⫾ 10.9

78.4 ⫾ 8.8

76.2 ⫾ 12.0

77 (24.5) 177 (56.4) 60 (19.1)

17 (10.2) 103 (61.7) 47 (28.1)

15 (16.5) 54 (59.3) 22 (24.2)

Male Race Unknown White Black Other

308 (98.1)

162 (97.0)

88 (96.7)

40 (12.7) 165 (52.5) 106 (33.8) 3 (1.0)

8 (4.8) 143 (85.6) 12 (7.2) 4 (2.4)

11 (12.1) 52 (57.1) 26 (28.6) 2 (2.2)

Hispanic ethnicity Yes No Unknown

20 (6.4) 262 (83.4) 32 (10.2)

1 (0.6) 157 (94.0) 9 (5.4)

10 (11.0) 71 (78.0) 10 (11.0)

257 (81.8) 6 (1.9)

134 (80.2) 2 (1.2)

76 (83.5) 0 (0.0)

0.7 0.7

30 (9.6) 1 (0.3) 7 (2.2) 19 (6.1)

19 (11.4) 0 (0.0) 8 (4.8) 6 (3.6)

8 (8.8) 0 (0.0) 4 (4.4) 5 (5.5)

0.5 0.9 0.2 0.3

103 (32.8) 204 (65.0) 295 (93.9) 44 (14.0) 153 (48.7) 10 (3.2) 4 (1.3) 31 (9.9)

64 (38.3) 102 (61.1) 162 (97.0) 16 (9.6) 89 (53.3) 12 (7.2) 4 (2.4) 17 (10.2)

27 (29.7) 60 (65.9) 87 (95.6) 8 (8.8) 47 (51.6) 3 (3.3) 1 (1.1) 8 (8.8)

0.2 0.4 0.2 0.2 0.4 0.06 0.5 0.9

29.4 ⫾ 12.5

32.2 ⫾ 12.2

28.7 ⫾ 12.9

0.02b

133 (42.4) 181 (57.6)

61 (36.5) 106 (63.5)

44 (48.4) 47 (51.6)

0.2 0.2

0 (0.0)

37 (22.2)

16 (17.6)

NA ⬍0.001b,e

20 (6.4) 212 (67.5) 37 (11.8) 45 (14.3)

75 (44.9) 0 (0.0) 92 (55.1) 0 (0.0)

7 (7.7) 76 (83.5) 8 (8.8) 0 (0.0)

Characteristic

Concomitant indications for ESAc None Chemotherapy-induced anemia Anemia of chronic disease HCV treatment–related anemia Myelodysplastic syndrome Other Comorbid conditions Heart failure Diabetes Hypertension Active cancer (nonmelanoma) Coronary artery disease History of DVT History of pulmonary embolism History of MI eGFR (mL/min/1.73 m2)d ESA Darbepoetin Epoetin Dual care of ESA Census region Northeast South Midwest West

Pa

⬍0.001 ⬍0.001

0.5 ⬍0.001b

⬍0.001b

Note: Values for continuous variables given as mean ⫾ standard deviation; values for categorical variables shown are number (percentage). Abbreviation: DVT, deep vein thrombosis; eGFR, estimated glomerular filtration rate; ESA, erythropoiesis-stimulating agent; HCV, hepatitis C virus; MI, myocardial infarction. a 2 ␹ tests for categorical variables and t tests for continuous variables. P values in the table are for pairwise comparisons of pharmacist-managed ESA clinics versus usual care. b Statistically significance difference (P ⬍ 0.05) for pairwise comparison of usual care versus usual-care at ESA clinic sites. c Totals can be ⬎100% because patients can have more than one concomitant indication. d n ⫽ 532. e Statistically significance difference (P ⬍ 0.05) for pairwise comparison of pharmacist-managed ESA clinics versus usual-care at ESA clinic sites.

374

Am J Kidney Dis. 2012;60(3):371-379

Pharmacist-Managed ESA Clinics Table 2. Proportion of Hemoglobin Values in Specified Ranges

Hemoglobin range ⬍10 g/dL 10-12 g/dLd ⬎12 g/dL

Pharmacist-Managed ESA Clinic (n ⴝ 1,807)

Usual Care Site (n ⴝ 606)

Usual Care at ESA Clinic Site (n ⴝ 346)

Pa

349 (19.3) 1284 (71.1) 174 (9.6)

127 (21.0) 345 (56.9) 134 (22.1)

110 (31.8) 179 (51.7) 57 (16.5)

⬍0.001b,c 0.03b — ⬍0.001b

Note: Analysis by random-effect multinomial model with clustering at the patient and site levels; hemoglobin values within target range (ie, 10-12 g/dL) are the base category for these comparisons. We also tested group differences for proportions of hemoglobin values within target range using random-effect logistic regression model with clustering at patient and site levels: P ⬍ 0.001 for pharmacist-managed clinics versus usual-care sites and usual care at ESA clinic sites; no differences between the latter 2 groups. Abbreviation: ESA, erythropoiesis-stimulating agent. a Except where indicated, P values are for pairwise comparisons of the pharmacist-managed ESA clinics versus usual-care sites. b P ⬍ 0.05 for pairwise comparison of pharmacist-managed ESA clinic versus usual care at ESA clinic sites. c P for overall group differences. d Reference group.

in all groups (96.2% in pharmacist-managed ESA clinics, 98.8% in usual-care sites, and 92.3% in usual care at ESA clinic sites), as was the mean number of antihypertensive medications per patient (4.3 ⫾ 2.2 in pharmacist-managed ESA clinics, 4.1 ⫾ 2.1 at usualcare sites, and 4.4 ⫾ 1.9 in usual care at ESA clinic sites). Multinomial Model of Factors Associated With Hemoglobin Level Compared with patients at pharmacist-managed ESA clinic sites, patients at usual-care sites had a significantly higher risk of hemoglobin values ⬍10 g/dL (odds ratio [OR], 2.13; P ⫽ 0.009; Table 6) and ⬎12 g/dL (OR, 3.93; P ⬍ 0.001); usual-care patients at ESA clinic sites also were at higher risk of both

outcomes (ORs of 2.72 for hemoglobin level less than target and 2.93 for hemoglobin level greater than target, P ⬍ 0.001 for each). Relative to patients with baseline eGFR ⱖ30 mL/min/1.73 m2, those with eGFR ⬍15 mL/min/1.73 m2 had increased risk (OR, 2.58; P ⫽ 0.001) of hemoglobin level ⬍10 versus 10-12 g/dL. Patients with heart failure were less likely to have hemoglobin values less than the target range than those without heart failure (OR, 0.75; P ⫽ 0.03), and patients with diabetes were less likely to have hemoglobin values ⬎12 g/dL than those without diabetes (OR, 0.63; P ⫽ 0.02).

DISCUSSION Our study compared the quality of ESA prescribing and monitoring in pharmacist-managed ESA clinics

Table 3. Proportion of Tests With an Increased ESA Dose Following Hemoglobin ⬍10 g/dL or Withheld/Decreased ESA Dose Following Hemoglobin ⬎12 g/dL by Type/Site of Care ESA

Pharmacist-Managed ESA Clinic

Darbepoetin Epoetin Total

67/137 (48.9) 109/189 (57.7) 176/305 (57.7)

Usual-Care Site

Usual Care at ESA Clinic Site

Pa

Increased ESA Dose Following Hemoglobin Level <10 g/dL 12/42 (28.6) 18/64 (28.1) 30/105 (28.6)

18/39 (46.2) 10/54 (18.5) 28/92 (30.4)

0.2 0.01b 0.009b

Withheld or Decreased ESA Dose Following Hemoglobin Level >12 g/dL Darbepoetin Epoetin Total

37/60 (61.7) 43/76 (56.6) 80/131 (61.1)

17/44 (38.6) 20/67 (29.9) 37/108 (34.3)

4/23 (17.4) 4/20 (20.0) 8/43 (18.6)

0.2b 0.2 0.09b

Note: Values shown are n/N (percentage). Based on average daily dose calculated from the ESA dosing 30 days before and after (including test date) the hemoglobin test date. Hemoglobin tests are included in both the darbepoetin and epoetin rows for the small number of patients (n ⫽ 13) who switched medications within 30 days of a test date; the row with the total category includes each hemoglobin test once. Tests were excluded for patients who were observed fewer than 7 days before or after the test date. Assessments were done separately for each row. Abbreviation: ESA, erythropoiesis-stimulating agent. a P values are for pairwise comparisons of pharmacist-managed ESA clinics versus usual-care sites in random-effect logistic regression model with clustering at patient and site levels. b P ⬍ 0.05 for pairwise comparison of pharmacist-managed ESA clinic versus usual care at ESA clinic site. Am J Kidney Dis. 2012;60(3):371-379

375

Aspinall et al

Figure 1. Kernel density plots of averaged 30-day erythropoiesis-stimulating agent (ESA) doses of darbepoetin and epoetin. Averaged 30-day dosage was calculated using averaged patient daily dosage ⫻ 30. Vertical lines denote means for each type of care.

with usual care in a relatively large number of patients from multiple sites during the same period. We evaluated several measures of the quality of prescribing because the mechanism behind the increased risk of morbidity and mortality associated with higher hemoglobin levels is unclear.18-20 The proportion of hemoglobin values in the target range in the pharmacistmanaged ESA clinics (71.1%) in our study is similar to that reported by Bucaloiu et al5 (69.8%), and both are significantly greater than the proportions in physician-managed patients (56.9% and 43.9%, respectively). Other studies have reported greater achievement of target hemoglobin or hematocrit values after implementing pharmacist-managed programs.9,10 In our study, there also were significant differences between the pharmacist-managed ESA clinics and usualcare sites in the proportions of hemoglobin values greater than the target range. This is remarkable given consistent reports from the FDA and others about increased risk of serious cardiovascular events associated with higher targeted hemoglobin levels.1-3,14,21 In addition, the proportion of test results ⬎12 g/dL that were followed by withholding or decreasing the ESA dose was greater in the pharmacist-managed ESA clinics than in usual-care sites (61.1% vs 34.3%, 376

respectively); the difference was borderline significant, perhaps due to a small sample size. Physicians may have been treating to higher targeted hemoglobin levels, although improved quality of life at hemoglobin values higher than FDA recommendations has been questioned.3,20,22 We also assessed intra- and interpatient variability in hemoglobin values because of potential association between hemoglobin level variability and adverse outcomes.4,23 In addition, a number of factors makes it difficult to maintain hemoglobin levels within the relatively narrow range of 10-12 g/dL.4,24,25 We found that intrapatient variances in hemoglobin values were similar in patients at pharmacist-managed ESA clinics and usual-care clinics during the 6-month study period, but that interpatient variances were higher in usual-care patients. This suggests more consistent management across patients in the pharmacist-managed ESA clinics. Fewer providers are involved in the pharmacist-managed clinics, and protocols are in place for managing patients receiving ESAs. When we evaluated ESA dose as a measure of ESA prescribing quality, we found that lower epoetin and darbepoetin doses were used in the pharmacistmanaged ESA clinics versus usual care; however, the Am J Kidney Dis. 2012;60(3):371-379

Pharmacist-Managed ESA Clinics Table 4. Monitoring of Hemoglobin and Iron During Study Period by Type/Site of Care Pharmacist-Managed ESA Clinic

Usual-Care Site

Usual Care at ESA Clinic Site

Pa

Hemoglobin Monitoring No. of hemoglobin results/patient 0 1 2-3 4-5 ⱖ6

0 (0.0) 13 (4.1) 47 (15.0) 89 (28.3) 165 (52.5)

13 (7.8) 27 (16.2) 59 (35.3) 32 (19.2) 36 (21.6)

9 (9.9) 13 (14.3) 23 (25.3) 24 (26.4) 22 (24.2)

Mean no. of hemoglobin results/patient

5.8 ⫾ 2.6

3.6 ⫾ 3.0

3.8 ⫾ 2.8

0.007b

Iron Monitoring No. of iron studies/patientc 0 1 2 3⫹

19 (6.1) 89 (28.3) 101 (32.2) 105 (33.4)

72 (43.1) 50 (29.9) 29 (17.4) 16 (9.6)

36 (39.6) 25 (27.5) 22 (24.2) 8 (8.8)

Mean no. of iron studies/patient

2.2 ⫾ 1.6

1.0 ⫾ 1.1

1.1 ⫾ 1.2

⬍0.001b

Note: Values for categorical variables shown are number (percentage); values for continuous variables given as mean ⫾ standard deviation. Abbreviation: ESA, erythropoiesis-stimulating agent. a P values are for pairwise comparisons of pharmacist-managed ESA clinics versus usual-care sites in random-effect Poisson regression models for count data. b P ⬍ 0.05 for pairwise comparison of pharmacist-managed ESA clinics versus usual care at ESA clinic site. c Iron studies include ferritin, serum iron, or transferrin saturation.

difference was not statistically significant for epoetin because of wide dose ranges. Other pharmacistmanaged anemia programs also have reported using lower ESA doses.5,6,10 Using lower doses obviously results in lower costs for the payer, but patients who receive higher doses may be poor responders; in one study, those with a poor initial response to darbepoetin had higher rates of death and a composite end point of death from any cause/cardiovascular events.19 Table 5. Distribution of Potentially ESA-Related Adverse Events by Type/Site of Care

Adverse Event

Thromboembolica Heart failurea Uncontrolled hypertensionb

PharmacistManaged ESA Clinic (n ⴝ 51,077 patient-days)

Usual-Care Site (n ⴝ 27,521 patientdays)

Usual Care at ESA Clinic Site (n ⴝ 13,094 patientdays)

6 (0.02) 18 (0.06) 185 (0.66)

7 (0.05) 9 (0.06) 73 (0.48)

3 (0.04) 7 (0.10) 50 (0.69)

Note: Values shown are n (rate per 180 patient-days). Multiple events occurred in the same patients. Abbreviation: ESA, erythropoiesis-stimulating agent. a Event resulted in emergency department visit or hospitalization. Thromboembolic events include myocardial infarction, stroke, deep vein thrombosis, and pulmonary embolism. b Systolic blood pressure ⬎160 mm Hg or diastolic blood pressure ⬎100 mm Hg. Am J Kidney Dis. 2012;60(3):371-379

Although our primary comparison of interest was pharmacist-managed ESA clinics versus usual-care sites, we also assessed the quality of ESA prescribing and monitoring in usual-care patients at sites with ESA clinics. Patients in the pharmacist-managed ESA clinics had baseline characteristics similar to usualcare patients at those sites, but the quality of ESA prescribing and monitoring was significantly better in patients managed by pharmacists at ESA clinics. Consequently, essentially all patients who are treated with an ESA for CKD-related anemia now are managed in the pharmacy-run clinics at those sites. Our multinomial model supports an association between type of care and hemoglobin level ⬍10 or ⬎12 g dL. For those receiving usual care compared with care in a pharmacist-managed ESA clinic, the OR for having a hemoglobin level ⬍10 versus 10-12 g/dL was about 2, and the OR of having a hemoglobin level ⬎12 g/dL was almost 4. ORs for having a hemoglobin level lower or higher than the target range for usual care at ESA clinic sites compared with pharmacist-managed care in an ESA clinic were 2.7 or greater. Compared with eGFR ⱖ30 mL/min/1.73 m2, patients with eGFR ⬍15 mL/min/1.73 m2 were more likely to have a hemoglobin level ⬍10 g/dL. Although no patient was receiving dialysis, it is possible that those with worse kidney function retained more fluid; hemodilution could have contributed to lower values.4 Last, patients with heart failure and diabetes 377

Aspinall et al Table 6. Multinomial Model of Factors Associated With Hemoglobin Level Hb <10 vs 10-12 g/dL

Hb >12 vs 10-12 g/dL

Variable

aOR (95% CI)

P

aOR (95% CI)

P

Type of care Pharmacist-managed ESA clinic Usual care Usual care at ESA clinic site

1.00 (reference) 2.13 (1.21-3.74) 2.72 (1.61-4.61)

⬍0.001 — 0.009 ⬍0.001

1.00 (reference) 3.93 (2.21-6.99) 2.93 (1.67-5.15)

⬍0.001 — ⬍0.001 ⬍0.001

eGFR ⬍15 mL/min/1.73 m2 15-29 mL/min/1.73 m2 ⱖ30 mL/min/1.73 m2

2.58 (1.46-4.55) 1.04 (0.70-1.52) 1.00 (reference)

0.003 0.001 0.9 —

1.03 (0.54-1.99) 0.82 (0.54-1.23) 1.00 (reference)

0.6 0.9 0.3 —

Age (/1 y) Female Race White Black Other/unknown Heart failure Diabetes mellitus Coronary artery disease

1.00 (0.99-1.02) 3.67 (0.96-14.1) 1.00 (reference) 1.29 (0.83-1.99) 1.52 (0.82-2.80)

0.7 0.06 0.3 — 0.3 0.2

1.00 (0.98-1.02) 1.36 (0.38-4.86) 1.00 (reference) 0.97 (0.60-1.56) 1.41 (0.74-2.70)

0.9 0.6 0.5 — 0.9 0.3

0.58 (0.40-0.84) 0.78 (0.54-1.14) 0.65 (0.28-1.51)

0.004 0.2 0.3

0.83 (0.56-1.22) 0.63 (0.42-0.94) 0.66 (0.27-1.65)

0.3 0.02 0.4

Note: Analysis by random-effect multinomial logistic regression model with clustering at the site and patient levels. Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; eGFR, estimated glomerular filtration rate; ESA, erythropoiesisstimulating agent; Hb, hemoglobin.

were less likely to have hemoglobin levels out of the target range, possibly as a result of more frequent monitoring due to health status concerns. Although rates of thromboembolic events, heart failure, and uncontrolled hypertension were similar in the usual-care and pharmacist-managed ESA clinic patients, the study was not designed to test these clinical outcomes. Uncontrolled hypertension was the most common adverse event, and more episodes of uncontrolled hypertension were detected in patients in pharmacist-managed ESA clinics than in usual care. However, patients in pharmacist-managed ESA clinics were seen more frequently, and providers had additional opportunities to measure blood pressure. Our study was comprehensive in its assessment of the quality of ESA prescribing and monitoring, but has limitations. First, geographic differences in the locations of pharmacist-managed clinics and usualcare sites could be associated with prescribing and monitoring practices. However, management was similar for usual-care patients at sites with and without ESA clinics. Second, we could not assess whether different hemoglobin targets were used by usual-care providers (the target was 10-12 g/dL in all patients with NDD-CKD in the pharmacist-managed ESA clinics during the study period). Third, in the usualcare patients who were co-managed by VA and non-VA providers, we were able to collect only hemoglobin and iron study results that were recorded in the VA 378

medical record; however, results did not change when we ran the analyses excluding these patients (data not shown). Finally, although results initially may not appear to be generalizable outside the VA, several previous studies involved pharmacists successfully managing patients with anemia and CKD in the private sector,5,9,10,13 and pharmacists can be part of multidisciplinary teams in outpatient nephrology and dialysis clinics, as well as in hospital-based settings. Relative to usual care, pharmacist-managed ESA clinics provided improved quality of ESA dosing and monitoring for patients with NDD-CKD. New FDA recommendations are to consider initiating ESA therapy only when hemoglobin level is ⬍10 g/dL and to use the lowest effective dose to prevent blood transfusions in patients with anemia and CKD. Nevertheless, results of this study are still applicable. When a physician decides to initiate ESA therapy, the pharmacist and physician should work together to determine the lowest acceptable hemoglobin level for the patient to minimize the symptoms of anemia and avoid transfusions. Then the pharmacist would continue to be responsible for closely monitoring the patient and adjusting the ESA dose to achieve these objectives, as well as to avoid the high hemoglobin values that have been associated with serious adverse events. Pharmacist-managed ESA clinics are structured to realize these new goals and facilitate positive outcomes. Am J Kidney Dis. 2012;60(3):371-379

Pharmacist-Managed ESA Clinics

ACKNOWLEDGEMENTS The members of the ESA Clinic Study Group include Jeanette Kreutzer, PharmD, and Sarah Providence, PharmD (VA Pittsburgh Healthcare System); Muriel Burk, PharmD (Pharmacy Benefits Management Services-Hines VA Medical Center); Nicole M. Dolder, PharmD (W.G. “Bill” Hefner Medical Center); Mabel Dea, PharmD (VA Long Beach Healthcare System); John Huang, PharmD, Sharon Jung Tschirhart, PharmD, and Prudence Hofmann, PharmD (South Texas Health Care System); Sonya Wilmer, PharmD (Michael E. DeBakey VA Medical Center); Thomas Lynch, RPh (Syracuse VA Medical Center); Isabel Devlin, PharmD, Mark Heller, PharmD, and Susan Rourke-Webb, PharmD (North Florida/South Georgia Veterans Health System); Carol Rudo, PharmD, Christopher Gallagher, PharmD, and Rosetta Corbett (VA Maryland Health Care System); Leah Flowers, PharmD, and Jon Jackson, PharmD (Jackson VA Medical Center); David Jansen, PharmD (St. Louis VA Medical Center); Douglas Covey, PharmD, and Joseph Fierro, PharmD (James A. Haley VA Medical Center); Robert Lana, PharmD, Karen Smith, PharmD, and Kimberly Allison, PharmD (VA Hudson Valley Health Care System); Matthew Lechtenberg, PharmD, and Virgil Angleton, PharmD (Eastern Kansas Health Care System); Monica Schaefer, PharmD, and Ann Ungerman, PharmD (Kansas City VA Medical Center); Michaela Hrdy, PharmD, Kristen Cook, PharmD, and Katherine Miller, PharmD (VA Nebraska-Western Iowa Health Care System). Selected results were presented as a poster during the American College of Clinical Pharmacy Annual Meeting, Pittsburgh, PA, October 19, 2011. The authors acknowledge the support of Ms Lucy Pandey and Mr Ken Bukowski, program analysts with the VA Center for Medication Safety, for establishing the database and entering data for the study. Support: There was no specific funding support for the work. However, these findings are the result of work supported in kind by the VA Center for Medication Safety/VA Pharmacy Benefits Management Services, Hines, IL; VA Pittsburgh Healthcare System, Pittsburgh, PA; and the other VA medical centers that participated in the study. The views expressed in this report are those of the authors, and no official endorsement by the Department of Veteran Affairs or the United States Government is intended or should be inferred. Financial Disclosure: The authors declare that they have no relevant financial interests.

REFERENCES 1. Pfeffer MA, Burdmann EA, Chen C-Y, et al. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med. 2009;361(21):2019-2032. 2. Drueke TB, Locatelli F, Clyne N, et al. Normalization of hemoglobin level in patients with chronic kidney disease and anemia. N Engl J Med. 2006;355(20):2071-2084. 3. Singh AK, Szczech L, Tang KL, et al. Correction of anemia with epoetin alfa in chronic kidney disease. N Engl J Med. 2006;355(20):2085-2098. 4. Kalantar-Zadeh K, Aronoff GR. Hemoglobin variability in anemia of chronic kidney disease. J Am Soc Nephrol. 2009;20(3): 479-487. 5. Bucaloiu ID, Akers G, Bernudez MC, et al. Outpatient erythropoietin administered through a protocol-driven, pharmacistmanaged program may produce significant patient and economic benefits. Manag Care Interface. 2007;20(6):26-30. 6. Clapp SE, Bardo JA, Chrymko MM. Implementation of a pharmacist-managed clinic for patients receiving erythropoietinstimulating agents. Am J Health-Syst Pharm. 2008;65(15):14581463.

Am J Kidney Dis. 2012;60(3):371-379

7. Gilmartin C. Pharmacist’s role in managing anemia in patients with chronic kidney disease: potential clinical and economic benefits. Am J Health-Syst Pharm. 2007;64(13)(suppl 8):S15-S22. 8. Gilreath JA, Sageser DS, Jorgenson JA, Rodgers GM. Establishing an anemia clinic for optimal erythropoietic-stimulating agent use in hematology-oncology patients. J Natl Compr Canc Netw. 2008;6(6):577-584. 9. Joy MS, Candiani C, Vaillancourt BA, Chin H, Hogan SL, Falk RJ. Reengineering clinical operations in a medical practice to optimize the management of anemia of chronic kidney disease. Pharmacotherapy. 2007;27(5):734-744. 10. Kimura T, Arai M, Masuda H, Kawabata A. Impact of a pharmacist-implemented anemia management in outpatients with end-stage renal disease in Japan. Biol Pharm Bull. 2004;27(11): 1831-1833. 11. To LL, Stoner CP, Stolley SN, Buenviaje JD, Ziegler TW. Effectiveness of a pharmacist-implemented anemia management protocol in an outpatient hemodialysis unit. Am J Health-Syst Pharm. 2000;58(21):2061-2065. 12. Stemer G, Lemens-Gruber R. Clinical pharmacy activities in chronic kidney disease and end-stage renal disease patients: a systematic literature review. BMC Nephrol. 2011;12:35-47. 13. Walton T, Holloway KP, Knauss MD. Pharmacist-managed anemia program in an outpatient hemodialysis population. Hospital Pharmacy. 2005;40(12):1051-1056. 14. Food and Drug Administration (FDA). Alert. 11/8/2007. Information for healthcare professionals: erythropoiesis stimulating agents. http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm126481.htm. Accessed June 17, 2011. 15. National Kidney Foundation. KDOQI Clinical Practice Guideline and Clinical Practice Recommendations for Anemia in Chronic Kidney Disease: 2007 update of hemoglobin target. Am J Kidney Dis. 2007;50(3):471-530. 16. Singh AK. What is causing the mortality in treating the anemia of chronic kidney disease: erythropoietin dose or hemoglobin level? Curr Opin Nephrol Hypertens. 2010;19(5):420-424. 17. National Kidney Foundation. KDOQI Clinical Practice Guidelines and Clinical Practice Recommendations for Anemia in Chronic Kidney Disease. Am J Kidney Dis. 2006;47(suppl 3):S1-S145. 18. Fishbane S, Besarab A. Mechanism of increased mortality risk with erythropoietin treatment to higher hemoglobin targets. Clin J Am Soc Nephrol. 2007;2(6):1274-1282. 19. Solomon SD, Uno H, Lewis EF, et al. Erythropoietic response and outcomes in kidney disease and type 2 diabetes. N Engl J Med. 2010;363(12):1146-1155. 20. Palmer SC, Navaneethan SD, Craig JC, et al. Metaanalysis: erythropoiesis-stimulating agents in patients with chronic kidney disease. Ann Intern Med. 2010:153(1):23-33. 21. Food and Drug Administration (FDA). Alert. 11/16/2006. Information for healthcare professionals: erythropoiesis stimulating agents. http://www.mtppi.org/fda_adv.pdf. Accessed June 17, 2011. 22. Fishbane S. The role of erythropoiesis-stimulating agents in the treatment of anemia. Am J Manag Care. 2010;16:S67-S73. 23. Yang W, Israni RK, Brunelli SM, Joffe MM, Fishbane S, Feldman HI. Hemoglobin variability and mortality in ESRD. J Am Soc Nephrol. 2007;18(12):3164-3170. 24. DeNicola L, Conte G, Chiodini P, et al. Stability of hemoglobin levels during the first year of epoetin treatment in patients with chronic kidney disease. Clin J Am Soc Nephrol. 2007;2(5):938-946. 25. Ebben JP, Gilbertson DT, Foley RN, Collins AJ. Hemoglobin level variability: association with comorbidity, intercurrent events, and hospitalizations. Clin J Am Soci Nephrol. 2006;1(6): 1205-1210.

379