Nephrotoxicity Comparison of Two Commercially Available Generic Vancomycin Products

Abstract

To date, no comparative clinical studies have investigated the effects of different vancomycin products on nephrotoxicity. The objective of this single-center, retrospective, matched-cohort study was to investigate the impact of two different vancomycin products on the development of nephrotoxicity. The study population included adults receiving a single vancomycin product, from either Pfizer or Hospira, for their entire course of therapy. Patients were matched based on underlying nephrotoxicity risk factors. Secondary outcomes included the need for renal replacement therapy, length of hospital stay, and in-hospital mortality. One-hundred forty-six matched pairs (n = 292) were included, and they had no significant differences in demographics, comorbid conditions, severity of illness, or vancomycin-associated nephrotoxicity risk factors. The frequency of nephrotoxicity was 8.9% in the Pfizer group and 11.0% in the Hospira group as defined by the 2009 consensus vancomycin guidelines (P = 0.56), 17.1% in the Pfizer group and 13.0% in the Hospira group as defined by the Acute Kidney Injury Network (AKIN) (P = 0.33), and 10.3% in the Pfizer group and 11.6% in the Hospira group as defined by RIFLE (risk, injury, failure, loss, and end-stage renal disease) criteria (P = 0.71). There were no differences between groups in regard to nephrotoxicity by any definition or in secondary outcomes. In multivariate analysis of overall nephrotoxicity risk factors, the type of vancomycin product was not independently associated with increased odds of developing nephrotoxicity according to the RIFLE criteria. Based on our results, there are no discernible differences between Pfizer and Hospira vancomycin products in the frequency of nephrotoxicity. Confirmation of these results with other types of vancomycin and different patient populations is warranted.

INTRODUCTION

The therapeutic use of vancomycin is associated with the development of nephrotoxicity, and several factors are associated with an increased risk of vancomycin-associated nephrotoxicity (1, 2). One potential nephrotoxicity risk factor not reported in the currently available literature is the vancomycin product manufacturer or brand. Concern for discordant in vitro and in vivo characteristics among vancomycin products has been raised due to conflicting potency results in animal studies (3, 4), a case report suggesting that clinical failure in an episode of methicillin-resistant Staphylococcus aureus (MRSA) bacteremia was driven by use of a generic vancomycin product that subsequently resolved when switched to the branded product (5), and the Food and Drug Administration (FDA) drug approval and generic manufacturing processes, which allow some variance in the percentage of active drug and do not require proof of in vivo efficacy or safety (FDA Code of Federal Regulations, Title 21, vol. 5, part 320) (6). However, no clinical studies have investigated the role that different types of vancomycin products may have on efficacy or toxicity.

A higher rate and extent of vancomycin-associated nephrotoxicity was observed at the Detroit Medical Center in early 2013. The Detroit Medical Center did not consistently track or use a single type of vancomycin product during this period. Given the ambiguous current evidence and lack of clinical data, there was concern that nephrotoxicity risk was associated with the type of vancomycin product used. Therefore, we conducted this retrospective, matched-cohort study to investigate the effect of specific vancomycin products on nephrotoxicity.

MATERIALS AND METHODS

The study's primary objective was to investigate the association of two different generic vancomycin products, those manufactured by Pfizer and Hospira, with nephrotoxicity. Three nephrotoxicity definitions were used: the 2009 consensus vancomycin guideline (1) and Acute Kidney Injury Network (AKIN) (7, 8) and RIFLE (risk, injury, failure, loss, and end-stage renal disease) criteria (7, 8) (Table 1). Secondary objectives were to determine if there was a difference between groups in the requirement for hemodialysis or continuous renal replacement therapy, hospital length of stay, and in-hospital all-cause mortality.

TABLE 1.

Primary outcome definitions

Definition	Criteria
Consensus vancomycin guideline (1)	Serum creatinine increase of ≥0.5 mg/dl or 50% from baseline for ≥2 readings
AKIN (7, 8)a
Stage 1	Serum creatinine increase of ≥0.3 mg/dl or ≥1.5 × baseline
Stage 2	Serum creatinine increase of ≥0.5 mg/dl or ≥2 × baseline
Stage 3	Serum creatinine increase of ≥3 × baseline or acute increase of 0.5 mg/dl if serum creatinine is ≥4 mg/dl
RIFLE (7, 8)a
Risk	Serum creatinine increase of ≥1.5 × baseline or creatinine clearance decrease of >25%
Injury	Serum creatinine increase of ≥2 × baseline or creatinine clearance decrease of >50%
Failure	Serum creatinine increase of ≥3 × baseline or creatinine clearance decrease of >75%

^aSerum creatinine increase must occur for ≥2 readings within 48 h.

This was a single-center, retrospective, matched-cohort study conducted at Sinai-Grace Hospital, a community teaching hospital with 404 licensed beds in Detroit, MI. The study protocol was approved by the Detroit Medical Center and Wayne State University institutional review boards. The study included patients who were 18 to 99 years of age and received vancomycin for at least 48 h. Patients must have received a single vancomycin product, manufactured by either Pfizer or Hospira, for their entire course of therapy. All vancomycin was prepared by the pharmacy department. Only the Pfizer vancomycin product was available for preparation from 20 May to 8 July 2013, and only the Hospira product was available from 15 July to 25 November 2013. Pharmacy personnel ensured that only the selected product was available during each respective period. Patients were included only if their vancomycin course started and stopped within the respective periods. They must have had at least two serum creatinine measurements: one prior to or within 24 h of the first vancomycin dose and one within 48 h of the last vancomycin dose. Patients were excluded if they had elevated serum creatinine levels compared to their baseline levels at the time of vancomycin initiation (at least two consecutive measurements of ≥0.3 mg/dl above baseline), received intermittent hemodialysis or continuous renal replacement therapy prior to the first vancomycin dose, or did not receive a scheduled maintenance dose due to moderate to severe renal insufficiency. All vancomycin was dosed by the pharmacy department according to population pharmacokinetic studies and subsequently adjusted to target recommended serum concentrations (1, 9, 10).

Data collected included patient demographics, comorbid conditions (including Charlson comorbidity index), vancomycin regimen (product, indication, daily dose, duration, and trough concentration), concomitant nephrotoxins, concomitant antimicrobial regimens (daily dose and duration), and every serum creatinine measurement during the vancomycin therapy (11). Vancomycin trough concentration was defined as a steady-state serum concentration obtained within 2 h prior to a maintenance dose. Creatinine clearance was calculated according to the Cockcroft-Gault formula using ideal body weight (12). The following medications or mediation classes were considered nephrotoxins: aminoglycosides, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, acyclovir, amphotericin B, calcineurin inhibitors, intravenous contrast dye, loop diuretics, nonsteroidal anti-inflammatory drugs, polymyxins, and vasopressors.

Patients from the Pfizer and Hospira groups were matched based on five nephrotoxicity risk factors (2): creatinine clearance at vancomycin initiation (<40, 40 to 69, and ≥70 ml/min), initial vancomycin maintenance dose (≤2,000, >2,000 to <4,000, and ≥4,000 mg/day), vancomycin duration (≤3, and >3 to <7, ≥7 days), acute severity of illness (severe sepsis or septic shock versus absence of severe sepsis or septic shock), and number of concomitant nephrotoxins (0, 1 to 2, and ≥3) (13).

Categorical variables were compared using the chi-square test, and continuous variables were compared using the Mann-Whitney U test or t test as appropriate. Bivariate logistic regression was performed to identify independent risk factors for vancomycin-associated nephrotoxicity defined by RIFLE criteria. Variables with P values of <0.1 in the bivariate analysis were included in the multivariate analysis. P values of ≤0.05 were considered significant. Statistical analyses were performed with SAS software version 9.2 (Cary, NC). A 10% absolute difference in the proportion of patients experiencing nephrotoxicity was determined by the study investigators to represent a clinically significant difference between types of vancomycin product. Therefore, with 146 patients in each group, this analysis had an 89% power to identify an increase in rates of nephrotoxicity from 5% to 15% and a 78% power to identify an increase in rates from 10% to 20% with a given product at an α value of 0.05.

RESULTS

A total of 657 patients were identified and screened for eligibility. Of the 250 patients screened in the Pfizer cohort, 146 met inclusion criteria. Of the 407 patients screened in the Hospira cohort, 298 met inclusion criteria. The primary reasons for ineligibility in both cohorts were receipt of vancomycin for <48 h, receipt of hemodialysis at baseline, and serum creatinine elevations prior to therapy initiation. One hundred forty-six patients in the Hospira cohort were matched to the 146 patients in the Pfizer cohort. One hundred twenty-six patients were matched on all five variables, while 20 patients were matched on four variables.

There were no significant differences in baseline demographics or underlying risk factors for nephrotoxicity (Table 2). The most common vancomycin indications, empirical or definitive, were pneumonia (47%), skin and skin structure infections (27%), and bone and joint infections (8%). Eighty-eight patients (60%) in the Pfizer cohort compared to 105 patients (72%) in the Hospira cohort received a vancomycin loading dose (≥20 mg/kg) (P = 0.04). There were no other significant differences in the vancomycin indications, dosing regimens, or initial or overall serum vancomycin exposures.

TABLE 2.

Baseline characteristics

Characteristic	Data fora:			P valueb
	Overall population (n = 292)	Pfizer drug recipients (n = 146)	Hospira drug recipients (n = 146)
Age (yr)	60.3 ± 18.1	60.1 ± 18.2	60.6 ± 18.0	0.79
Male	143 (49)	67 (46)	76 (52)	0.29
African American	258 (88)	132 (90)	126 (86)	0.27
Total body weight (kg)	81 (66–98)	77 (64–100)	82 (70–97)	0.80
Initial creatinine clearance (ml/min)	58 (39–81)	57 (38–81)	62 (40–79)	0.54
Charlson comorbidity index	2 (1–4)	2 (1–4)	2 (1–4)	0.18
Diabetes	99 (34)	45 (31)	54 (37)
Chronic obstructive pulmonary disease (COPD)	68 (23)	37 (25)	31 (21)
Heart failure	58 (20)	34 (23)	24 (16)
History of cerebrovascular accident (CVA)	50 (17)	29 (20)	21 (14)
Chronic kidney disease	41 (14)	16 (11)	25 (17)
Severe sepsis/septic shock	53 (18)	28 (19)	25 (17)	0.65
Concomitant nephrotoxins				0.71
0	92 (32)	46 (32)	46(32)
1–2	190 (65)	94 (64)	96 (66)
≥3	10 (3)	6 (4)	4(3)
Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers	98 (34)	48 (33)	50 (34)	0.80
Acyclovir	3 (1)	2(1)	1 (1)	0.56
Aminoglycoside	2 (1)	2 (1)	0	0.16
Amphotericin B	0	0	0	–
Calcineurin inhibitors	1 (0)	1 (1)	0	0.32
Colistin	0	0	0	–
Intravenous contrast	71 (24)	32 (22)	39 (27)	0.42
Loop diuretics	74 (25)	44 (30)	30 (21)	0.06
Nonsteroidal anti-inflammatory agents	18 (6)	8 (6)	10 (7)	0.63
Vasopressors	7 (2)	3 (2)	4 (3)	0.70
Vancomycin loading dose of ≥20 mg/kg	193 (66)	88 (60)	105 (72)	0.04
Vancomycin loading dose (mg/kg)	24 (22–25)	23 (23–25)	24 (22–25)	0.85
Initial maintenance dose (mg)	2,000 (1,250–2,813)	2,000 (1,250–2,500)	2,000 (1,750–2,250)	0.45
Median maintenance dose (mg)	2,000 (1,500–3,000)	2,000 (1,500–3,000)	2,250 (1,500–3,000)	0.34
Cumulative vancomycin dose (mg)	9,250 (5,500–14,250)	9,000 (5,063–13,375)	9,625 (6,250–14,688)	0.27
Vancomycin duration (h)	96 (72–156)	96 (72–156)	96 (72–153)	0.42
Median vancomycin trough concn (mg/liter)	14.6 (11.3–18.1)	14.5 (11.3–18.3)	14.6 (10.4–18.4)	0.48
Vancomycin trough of >20 mg/liter	52 (18)	26 (18)	26 (18)	1.00

^aData are mean ± SD, number (%), or median (interquartile range).

^bP value for difference between Pfizer and Hospira vancomycin products.

There were no significant differences between cohorts with respect to the primary outcome, nephrotoxicity, by any of the three study definitions used (Table 3). Nephrotoxicity according to the 2009 consensus vancomycin guideline definition occurred in 13 (8.9%) Pfizer and 16 (11%) Hospira patients (P = 0.56). According to the AKIN definition, nephrotoxicity occurred in 25 (17.1%) Pfizer and 19 (13%) Hospira patients (P = 0.33). According to the RIFLE criteria, nephrotoxicity occurred in 15 (10.3%) Pfizer and 17 (11.6%) Hospira patients (P = 0.71). There were no significant differences between groups in the need for renal replacement therapy, hospital length of stay, or in-hospital mortality (Table 3). Neither receipt of Hospira nor receipt of Pfizer vancomycin product was associated with nephrotoxicity in either bivariate or multivariate analysis (Table 4).

TABLE 3.

Primary and secondary outcomes of patients receiving vancomycin product

Outcome	Results with vancomycin from (no. [%]):		P value
	Pfizer drug recipients (n = 146)	Hospira drug recipients (n = 146)
Primary, by definition
Consensus vancomycin guideline	13 (8.9)	16 (11)	0.56
AKIN
All stages	25 (17.1)	19 (13)	0.33
Stage 1	12 (8.2)	3 (2.1)
Stage 2	12 (8.2)	12 (8.2)
Stage 3	1 (0.7)	4 (2.7)
RIFLE
All stages	15 (10.3)	17 (11.6)	0.71
Risk	6 (4.1)	8 (5.5)
Injury	8 (5.5)	5 (3.4)
Failure	1 (0.7)	4 (2.7)
Secondary
Renal replacement therapy required	0 (0)	1 (0.7)	0.26
In-hospital mortality	5 (3.4)	9 (6.2)	0.27
Length of stay (mean days [range])	10 (6–18)	11(7–17)	0.35

TABLE 4.

Predictors of nephrotoxicity

Variable	Bivariate analysisa			Multivariate analysis
	OR	95% CI	P value	OR	95% CI	P value
Receipt of Hospira vancomycin	1.15	0.55–2.40	0.71	1.37	0.59–3.17	0.47
Concomitant piperacillin-tazobactam	5.17	2.29–11.66	<0.001	3.97	1.66–9.50	0.002
Vancomycin trough of >20 mg/liter	4.54	2.09–9.90	<0.001	3.83	1.59–9.21	0.003
Severe sepsis/septic shock	4.51	1.54–13.26	0.003	3.10	0.99–9.74	0.053
Receipt of ≥2 concomitant nephrotoxins	4.22	1.96–9.06	<0.0001	3.78	1.61–8.86	0.002
Vancomycin loading dose of ≥20 mg/kg	0.52	0.25–1.10	0.08	0.44	0.19–1.06	0.066

^aOR, odds ratio; CI, confidence interval.

DISCUSSION

In this study, no association was found between type of vancomycin product and nephrotoxicity. The two groups were well matched with regard to nephrotoxicity risk factors. Given that vancomycin-associated nephrotoxicity is often related to multiple factors, we performed a multivariate analysis to identify variables independently associated with nephrotoxicity, including the vancomycin product (2). Type of vancomycin product was not independently associated with nephrotoxicity in this multivariate analysis.

Several variables in this study were associated with nephrotoxicity in multivariate analysis (Table 4). All of these variables were previously reported in the literature (2, 14, 15). Receipt of any concomitant nephrotoxin, vasopressors, or aminoglycosides or of an increasing number of nephrotoxins has been associated with nephrotoxicity (2, 16-19). In our study, 68% of patients received at least one concomitant nephrotoxin; however, the association with increased odds of nephrotoxicity remained significant only if patients received at least two concomitant nephrotoxins. Nephrotoxicity was analyzed as a function of number of nephrotoxins, as the use of aminoglycosides and vasopressors, the agents specifically associated with nephrotoxicity, was low in our study. Although not a designed focus of our study, we found that the association between receipt of concomitant piperacillin-tazobactam and vancomycin-associated nephrotoxicity was consistent with recent reports (14, 15). There were 108 total patients on concomitant vancomycin and piperacillin-tazobactam. Nephrotoxicity occurred in 23 of these patients (21.3%) according to RIFLE criteria. In comparison, 66 patients received concomitant cefepime and vancomycin, and nephrotoxicity occurred in 3 patients (4.5%) according to RIFLE criteria (odds ratio, 0.46; 95% confidence interval, 0.15 to 1.35; P = 0.15).

This study had limitations. First, it was a retrospective study conducted at a single center. Second, only two vancomycin products were included. After our investigation was conducted, Lewis and colleagues (20) reported that the mass spectrometry fingerprints of Pfizer and Hospira products are similar to each other but different from the product from APP Pharmaceuticals; therefore, the results of our study may not be applicable to other vancomycin products. Third, 20 patients were matched on four of five matching variables despite our intent to match all patients on five variables. Fourth, the two vancomycin products were used sequentially in time rather than concomitantly. Therefore, additional patient and treatment factors may have been unaccounted for. Additionally, the 96-h median vancomycin therapy duration in our study was shorter than the duration previously associated with nephrotoxicity (2). A longer duration of therapy may be needed to observe possible differences in nephrotoxicity, if any, between products. However, the perceived increase in potential vancomycin-associated nephrotoxicity cases seen at our center was characterized by rapid-onset fulminate renal failure, usually occurring within the first 72 h of exposure, which would have been detected in this analysis.

In conclusion, the type of vancomycin product received, Pfizer or Hospira, was not associated with an increased risk for nephrotoxicity. Our study suggests that the practice of avoiding the purchase of certain vancomycin products may not be necessary to avoid development of toxicity in patients. However, confirmation of these results through studies with other types of vancomycin products and patient populations is warranted.