Area-Under-Curve–Guided Versus Trough-Guided Monitoring of Vancomycin and Its Impact on Nephrotoxicity: A Systematic Review and Meta-Analysis

Supplemental Digital Content is Available in the Text.

Background:

Conventionally, vancomycin trough levels have been used for therapeutic drug monitoring (TDM). Owing to the increasing evidence of trough levels being poor surrogates of area under the curve (AUC) and the advent of advanced pharmacokinetics software, a paradigm shift has been made toward AUC-guided dosing. This study aims to evaluate the impact of AUC-guided versus trough-guided TDM on vancomycin-associated nephrotoxicity.

Methods:

A systematic review was conducted using PubMed, Embase, Web of Science, Cumulative Index to Nursing and Allied Health Literature, Google scholar, and Cochrane library databases; articles published from January 01, 2009, to January 01, 2021, were retrieved and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist. Studies that evaluated trough-guided or AUC-guided vancomycin TDM and vancomycin-associated nephrotoxicity were included. Random-effects models were used to compare the differences in nephrotoxicity.

Results:

Of the 1191 retrieved studies, 57 were included. Most studies included adults and older adults (n = 47, 82.45%). The pooled prevalence of nephrotoxicity was lower in AUC-guided TDM [6.2%; 95% confidence interval (CI): 2.9%–9.5%] than in trough-guided TDM (17.0%; 95% CI: 14.7%–19.2%). Compared with the trough-guided approach, the AUC-guided approach had a lower risk of nephrotoxicity (odds ratio: 0.53; 95% CI: 0.32–0.89). The risk of nephrotoxicity was unaffected by the AUC derivation method. AUC thresholds correlated with nephrotoxicity only within the first 96 hours of therapy.

Conclusions:

The AUC-guided approach had a lower risk of nephrotoxicity, supporting the updated American Society of Health-System Pharmacists guidelines. Further studies are needed to evaluate the optimal AUC-derivation methods and clinical utility of repeated measurements of the AUC and trough levels of vancomycin.

INTRODUCTION

Vancomycin has been used in clinical practice for more than 50 years¹ and is the first-line treatment for methicillin-resistant Staphylococcus aureus (MRSA) infections. Initially, the clinical utilization of vancomycin was low owing to its adverse effects, including infusion reactions, ototoxicity, and nephrotoxicity. However, the increasing global prevalence of MRSA and penicillin-resistant Streptococcus pneumoniae infections¹ has led to a marked increase in the use of vancomycin.

Vancomycin is a glycopeptide antibiotic that displays time-dependent bactericidal activity. The percentage of time over minimum inhibitory concentration (%Time > MIC) was once considered the primary pharmacokinetic/pharmacodynamic parameter that correlated with efficacy of time-dependent antimicrobials.¹ However, Moise-Broder et al² subsequently demonstrated that the ratio of the steady-state 24-hour area under the concentration–time curve (AUC24) to the MIC (AUC24/MIC) correlated with clinical outcomes in patients with MRSA infections rather than %Time > MIC.

The 2009 consensus guideline on vancomycin therapeutic drug monitoring (TDM) advocated trough-guided monitoring as a surrogate for AUC because it was “accurate and practical.” A higher trough target of 15–20 mg/L was recommended in patients with complicated S. aureus infections to improve clinical outcomes, although the evidence in support of the safety of sustained higher trough concentrations at the time of publication was limited.³

Thereafter, several systematic reviews have substantiated the studies that support the association of a 2-fold to 3-fold increased risk of nephrotoxicity with higher trough levels.^4,5 Clark et al⁶ reported that a trough target of 15–20 mg/L could lead to higher vancomycin AUC exposure than was necessary for the treatment of invasive MRSA infections and could predispose patients to an increased risk of nephrotoxicity. Notably, all patients with a trough of >9 mg/L achieved a vancomycin AUC of ≥400 mg ×h/L.⁶ Moreover, vancomycin dose adjustment for a specified trough target as a surrogate marker of AUC is reportedly inadequate owing to the substantial interpatient variability of the AUCs.⁷

The recently updated American Society of Health-System Pharmacists vancomycin TDM guideline⁸ recommends a target AUC between 400 and 600 mg × h/L because of vancomycin's toxicodynamics and acute kidney injury (AKI) risk. This minimizes possible nephrotoxicity while maintaining treatment efficacy for invasive MRSA infections. In support of this recommendation, a study reported that AUC monitoring resulted in a significantly lower rate of nephrotoxicity (adjusted odds ratio (OR), 0.514; 95% confidence interval [CI], 0.332–0.79; P = 0.003) than trough monitoring.⁹ Before the release of the updated guideline, a survey¹⁰ conducted among pharmacists in the United States highlighted several key issues related to the utilization of AUC-guided monitoring of vancomycin. First, approximately three-quarters of academic medical centers continued to use trough-guided monitoring. Costs and unfamiliarity with AUC-guided monitoring, training requirements, and the unclear benefit of AUC-guided monitoring, including its effects on clinical efficacy and outcomes and nephrotoxicity rates, were barriers to its implementation. Regarding nephrotoxicity, a meta-analysis¹¹ conducted in 2019 included 2 studies that compared AUC-guided versus trough-guided monitoring directly and reported that nephrotoxicity rates were significantly lower in AUC-guided monitoring (OR, 0.68; 95% CI, 0.46–0.99). Thereafter, several studies have evaluated the risk of adverse effects associated with AUC-guided versus trough-guided vancomycin TDM.

Therefore, an updated systematic review and meta-analysis of AUC-guided versus trough-guided TDM of vancomycin was conducted. The primary objective aimed to determine if there was a difference in the rates of vancomycin-associated nephrotoxicity between AUC-guided and trough-guided monitoring. The secondary objective was the process outcomes such as AUC derivation methods and time periods of monitoring and nephrotoxicity.

MATERIALS AND METHODS

Protocol and Registration

The protocol for this review was registered on PROSPERO (available at:https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=189633). The review conformed to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) checklist.¹²

Information Sources and Search Strategy

Five electronic databases, including PubMed, Embase, Web of Science, Cumulative Index to Nursing and Allied Health Literature Plus with Full Text, and the Cochrane Register of Controlled Trials, were comprehensively searched to identify potential studies. Search concepts included “vancomycin,” “area under curve,” “Bayes Theorem,” “drug monitoring,” and “acute kidney injury.” MeSH terms for the aforementioned terms were combined with the relevant keywords (limited to title/abstract) in the various databases, which were accessed and searched on November 01, 2020. Only studies in English were considered for this review. In addition, relevant articles were manually searched, and the aforementioned terms were searched on Google Scholar as well on January 01, 2021. The complete search strategy for each database is presented in Supplemental Digital Content 1 (see Table, http://links.lww.com/TDM/A642).

Definitions The definition of nephrotoxicity in each study was captured and classified into “the 2009 guideline,” Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease (RIFLE),¹³ Pediatric RIFLE (pRIFLE),¹⁴ AKI Network (AKIN),¹⁵ or Kidney Disease Improving Global Outcomes (KDIGO) Criteria (Table 1).¹⁶

TABLE 1.

Classifications of Nephrotoxicity

Classification	Definition
2009 guideline3	A minimum of 2 or 3 consecutive documented increases in serum creatinine concentrations (defined as an increase of 0.5 mg/dL or a ≥50% increase from baseline, whichever is greater) after several days of vancomycin therapy
RIFLE13	Fulfilling the “R” of the RIFLE criteria. Risk is defined as an increase of serum creatinine with 50% corresponding to a decrease in glomerular filtration rate, relative to baseline, of >25% or a urine output of <0.5 mL·kg−1·h−1 for >6 h
pRIFLE14	Fulfilling the “R” of the pRIFLE criteria. Risk is defined as a decrease in estimated creatinine clearance as calculated by the Schwartz equation, relative to baseline, of >25% or a urine output of <0.5 mL·kg−1·h−1 for >6 h
AKIN15	An abrupt (within 48 h) reduction in kidney function currently defined as an absolute increase in serum creatinine ≥0.3 mg/dL (≥26.4 μmol/L), a percentage increase in serum creatinine ≥50% (1.5-fold from baseline), or a reduction in urine output (documented oliguria of less than 0.5 mL·kg−1·h−1 for >6 h)
KDIGO16	AKI is defined as an increase in serum creatinine by ≥ 0.3 mg/dL (≥26.5 µmol/L) within 48 h or an increase in serum creatinine to ≥1.5 times baseline, which is known or presumed to have occurred in the past 7 days or a urine volume of <0.5 mL·kg−1·h−1 for 6 h

RIFLE, Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease; KDIGO, Kidney Disease Improving Global Outcomes; AKIN, Acute Kidney Injury Network; pRIFLE, Pediatric Risk, Injury, Failure, Loss, End Stage Renal Disease.

Inclusion and Exclusion Criteria

The inclusion criteria were as follows: (1) experimental or observational studies published from January 01, 2009; (2) studies published in the English language; (3) patients of any age who received intermittent intravenous vancomycin for at least 48 hours; (4) nephrotoxicity was a primary or secondary objective/outcome of the study and was clearly defined; and (5) trough-guided monitoring with the 2009 consensus guideline-recommended therapeutic targets. The following publication/study types were excluded if patients had received renal replacement therapy or vancomycin through nonintravenous routes (ie, oral or orthopedic implants), computer-simulated and animal studies, systematic reviews, meta-analyses, review papers, letters, commentaries, case and case series, and studies with duplicated data/data sets.

Study Selection and Data Collection Process

Articles identified from the search of each database were imported into EndNote version X9 (Clarivate Analytics, Philadelphia, PA). Duplicated articles were removed using the “remove duplicates” function in EndNote; thereafter, any remaining duplicated articles were manually removed. Subsequently, the remaining records were exported into Microsoft Excel version 15.25 (Microsoft, Redmond, WA) for screening. Two investigators (A.L.S. and S.F.H.W.) screened the title and abstracts of the potentially eligible studies and examined if they met the predefined inclusion and exclusion criteria. Nonconcordance was reviewed by 2 investigators (H.Z. and C.H.T.), and the final list of selected articles was included in the review. Data were extracted from Excel by 2 investigators (A.L.S. and S.F.H.W.), and a third investigator (J.J.B.S.) checked random samples (10 of 38 studies) of the extracted data to ensure its accuracy.

Data Extraction

Patient characteristics included in the studies and details of vancomycin-associated nephrotoxicity outcomes were recorded. Vancomycin information, including total daily maintenance and cumulative doses and the duration of its usage, were collected. Regarding AUC information, the AUC determination method, AUC timing, target range and proportion of patients within range, and mean/median AUC achieved during the study were collected. Regarding trough information, trough definition and timing, target range and proportion of patients within range, and mean/median target achieved were collected. Nephrotoxicity-associated variables reported in the univariate and multivariate analyses were recorded.

Quality Assessment

The quality of the included studies were independently assessed by 2 investigators (A.L.S. and J.J.S.B.). Discrepancies in scoring were discussed with a third investigator (H.Z.), and the final score was consensually derived. The Newcastle–Ottawa scale was used to assess study quality.¹⁷ Each study was assigned a score ranging from 0 to 9; the higher scores represented a higher study quality. In this review, only studies with a score of ≥7 were included in the meta-analysis. This cutoff was consensually selected by the authors to ensure low bias when pooling potentially heterogeneous data into the meta-analysis.

Statistical Analyses

Cochrane Review Manager software (Review Manager (RevMan) [Computer program]. Version 5.4, The Cochrane Collaboration, 2020.) and OpenMeta [Analyst] (Version for Windows 10) were used for the statistical analyses. Heterogeneity (I²) was assessed using the χ² test (P < 0.1 and I² of >50% indicated significant heterogeneity). The preplanned analysis conformed to the following: OR with 95% CI was estimated using random-effects models (Mantel–Haenszel and inverse-variance methods) because of significant heterogeneity and varied definitions of nephrotoxicity in the included studies (dichotomous data).

RESULTS

An initial search conducted on November 01, 2020, identified 2107 studies. An additional search conducted on January 01, 2021, on Google scholar identified 5 studies. Figure 1 illustrates the flowchart for the inclusion of the studies. Of the 2112 retrieved studies, 57 were included in the review.

FIGURE 1.

PRISMA 2020 flow diagram for new systematic reviews depicting the selection process of studies included in the systematic review and meta-analysis.

Among the 57 included studies (Table 2), 4 made direct comparisons between AUC-guided and trough-guided cohorts^9,18–20; 10 reported both AUC-guided and trough-guided monitoring^{21–28,64,65}; 2 reported only AUC-guided monitoring^29,30; and 41 reported only trough-guided monitoring.^{31–63,66–73} Four studies had vancomycin doses adjusted based on both AUC-guided and trough-guided monitoring; 48 had vancomycin doses adjusted based solely on trough-guided monitoring; and 5 had vancomycin doses titrated based solely on AUC-guided monitoring. Most studies were from the United States (n = 34) and Asia-Pacific region (n = 18). Most studies were conducted in the adult and older adult populations (n = 49); 9 studies and 1 study was conducted in the pediatric and neonatal populations, respectively.

TABLE 2.

Details of the Included Studies

Author, Year, Country, and Study Design	Total No. of Patients	Primary Monitoring Method	Baseline Renal Function	Nephrotoxicity Definition	Period of Nephrotoxicity Studied	Patients with Nephrotoxicity, n (%)	NOS*
			SCr (mg/dL) CrCl (mL/min) eGFR (mL/min/1.73 m2)
Adults and older adults, comparing AUC vs trough
Finch et al, 2017, United States, retrospective, quasi-experimental study9	1280	AUC, trough	SCr: 0.9	2009 guideline, AKIN, RIFLE	NR†	2009: 108 (8.43)	8
						AKIN: 238 (18.6)
						RIFLE: 238 (18.6)
Neely 2018, United States, prospective cohort study18	252	AUC, trough	SCr: 0.83	2009 guideline	Initiation of vancomycin until 72 h post completion	8 (3.17)	9
			CrCl: 134.3
Oda et al, 2020, Japan, retrospective cohort study19	74	AUC, trough	CrCl: 85.8	AKIN, RIFLE	During vancomycin therapy	AKIN: 17 (23.0)	8
						RIFLE: 14 (18.9)
Muklewicz et al, 2021, United States, retrospective, pre–post quasi-experimental study20	636	AUC, trough	SCr: 0.9 ± 0.4	2009 guideline,‡ AKIN and RIFLE	NR†	2009: 59 (9.27)	8
						AKIN: 111 (17.5)
						RIFLE: 82 (12.9)
Adults and older adults, reported AUC and trough
Lodise et al, 2009, United States, retrospective cohort study21	166	Trough	CrCl: 72.4 ± 28.9	2009 guideline	Initiation of vancomycin until 72 h post completion	21 (12.6)	8
Hale et al, 2017, United States, retrospective cohort study22	97	Trough	SCr: 0.7 (0.5–0.9)	2009 guideline	NR†	9 (9.27)	7
			CrCl: 102.0 (74.5–120)
Liu et al, 2017, United States, retrospective analysis of RCT23	214	Trough	CrCl: 104.9	AKIN	NR†	39 (18.2)	6
Chavada et al, 2017, Australia, retrospective cohort study24	227	Trough	NR†	2009 guideline	Within first 14 days of vancomycin therapy	20 (8.81)	5
Zasowski et al, 2018, United States, retrospective cohort study25	323	AUC	SCr: 0.88 (0.69–1.18)	2009 guideline‡	Initiation of vancomycin until 72 h post completion	20 (6.19)	9
Mogle et al, 2018, United States, retrospective cohort study26	46	AUC	NR†	2009 guideline	Within first 7 days of vancomycin therapy	6 (13.0)	7
Lodise et al, 2020, United States, prospective cohort study27	212	AUC	NR†	2009 guideline,‡ RIFLE	Initiation of vancomycin until 48 h post completion	2009: 60 (28.3)	5
						RIFLE: 55 (25.9)
Liu et al, 2020, Australia, retrospective analysis of RCT28	291	Trough	SCr: 0.97 (0.75–1.63) eGFR: 71.8 (38.4–98.7)	RIFLE, KDIGO	NR	KDIGO 50 (17.1)	6
Adults and older adults, reported AUC only
Chattaweelarp et al, 2020, Thailand, retrospective cohort study29	131	AUC	CrCl: 30.9 ± 29.5	2009 guideline‡	Initiation of vancomycin until 48 h post completion	17 (12.9)	5
Vali et al, 2020, United Kingdom, prospective cohort study30	243	AUC	eCrCl: 69.9	AKIN	NR†	18 (7.40)	6
Adults and older adults, reported trough only
Minejima et al, 2011, United States, prospective cohort study31	227	Trough	MDRD: 115 CKD-EPI: 96.7	2009 guideline, AKIN, RIFLE	NR†	2009: 17 (7.5) AKIN: 43 (18.9) RIFLE: 42 (18.5)	7
Horey et al, 2012, United States, retrospective cohort study32	270	Trough	SCr: 1.2 ± 0.4	2009 guideline	Initiation of vancomycin until 72 h post completion	34 (12.5)	9
			CrCl: 66.8 ± 29.4
Hall et al, 2013, United States, retrospective cohort study33	336	Trough	SCr: 0.9	2009 guideline	During vancomycin therapy	78 (23.2)	8
			CrCl: 81
Moh'd et al, 2014, United States, retrospective cohort study34	579	Trough	SCr: 0.88 ± 0.31	2009 guideline	During vancomycin therapy	154 (26.5)	8
			CrCl: 100.4 ± 45.0
Burgess et al, 2014, United States, retrospective cohort study35	191	Trough	NR†	2009 guideline‡	Within first 7 days of vancomycin therapy	23 (12.0)	6
Rostas et al, 2014, United States, retrospective cohort study36	80	Trough	NR†	2009 guideline	Initiation of vancomycin until 72 h post completion	5 (6.25)	8
Moenster et al, 2014, United States, retrospective cohort study37	139	Trough	CrCl: 73.8	2009 guideline‡	NR†	36 (25.8)	8
Meaney et al, 2014, United States, retrospective cohort study38	125	Trough	CrCl: 84.6 ± 27.6	2009 guideline‡	48 h after initiation of vancomycin until 72 h post completion	17 (13.6)	9
Dong et al, 2015, China, retrospective cohort study39	90	Trough	SCr: 0.9 ± 0.27	2009 guideline	During vancomycin therapy	14 (15.5)	8
He et al, 2016, China, retrospective cohort study40	174	Trough	NR†	2009 guideline	Initiation of vancomycin until 72 h post completion	38 (21.8)	8
Hammoud et al, 2016, United States, retrospective cohort study41	498	Trough	SCr: 0.87 ± 0.38	2009 guideline	During vancomycin therapy	62 (12.4)	7
Brown et al, 2017, United States, retrospective cohort study42	162	Trough	SCr: 1.1	2009 guideline	NR†	7 (4.32)	7
			CrCl: 82.4
Takahashi et al, 2017, Japan, retrospective cohort study43	197	Trough	SCr: 0.64 ± 0.22	2009 guideline‡	NR†	33 (16.7)	7
Anderson et al, 2017, United States, retrospective cohort study44	455	Trough	SCr: 0.9	AKIN	NR†	77 (16.9)	7
			CrCl 115.1
Imai et al, 2017, Japan, retrospective cohort study45	592	Trough	SCr: 0.74 ± 0.35	2009 guideline‡	During vancomycin therapy	87 (14.6)	7
			CrCl: 91.2 ± 42.9
Park et al, 2018, Korea, retrospective cohort study46	315	Trough	SCr: 0.68 (0.55–0.92)	2009 guideline	NR†	48 (15.2)	8
Liang et al, 2018, China, prospective cohort study47	370	Trough	SCr: 0.73	AKIN	During vancomycin therapy	49 (12.1)	8
			CrCl 87.4
Robertson et al, 2018, United States, retrospective cohort study48	169	Trough	SCr: 0.79 ± 0.24	2009 guideline‡	Initiation of vancomycin until 72 h post completion	17 (10.0)	9
Wang et al, 2019, China, retrospective cohort study49	65	Trough	SCr 0.85 ± 0.47	2009 guideline	Initiation of vancomycin until 72 h post completion	20 (30.7)	7
Molina et al, 2019, United States, retrospective cohort study50	394	Trough	SCr: 0.75 ± 0.3	AKIN	Initiation of vancomycin until 48 h post completion	103 (26.1)	9
Almeida et al, 2019, Brazil, prospective cohort study51	98	Trough	SCr: 0.70 (0.58–0.96), eGFR: 118.20 (96.3–133.8)	AKIN	Initiation of vancomycin until 72 h post completion	19 (19.3)	9
Yahav et al, 2019, Israel, retrospective cohort study52	285	Trough	SCr: 1.3	RIFLE	Initiation of vancomycin until 96 h post completion	75 (26.3)	7
Okada et al, 2019, Japan, retrospective cohort study53	150	Trough	CrCl: 94.5	KDIGO	Initiation of vancomycin until 48 h post completion	17 (11.3)	8
Sussman et al, 2020, United States, retrospective cohort study54	99	Trough	SCr: 0.9 ± 0.23	RIFLE	NR†	28 (28.2)	8
Zamoner et al, 2020, Brazil, prospective cohort study55	135	Trough	GFR: 106 ± 25.7	KDIGO	NR†	37 (27.4)	8
Haruki et al, 2020, Japan, retrospective cohort study56	272	Trough	SCr: 0.72	RIFLE	Initiation of vancomycin until 48 h post completion	45 (16.5)	8
			CrCl: 60.5
Xu et al, 2020, China, retrospective cohort study57	341	Trough	SCr: 0.76 ± 0.28	KDIGO	Initiation of combination until 48 h post completion	55 (16.1)	9
Sharma et al, 2020, United States, retrospective cohort study58	1130	Trough	SCr 0.8 ± 0.30	AKIN	NR†	93 (8.23)	7
Liu et al, 2015, China, retrospective cohort study59	124	Trough	SCr: 0.94 ± 0.40	AKIN	Initiation of vancomycin until 72 h post completion	36 (29.0)	8
			CrCl: 99.27 ± 9.69
Huang et al, 2018, China, retrospective cohort study60	50	Trough	SCr: 0.8 ± 0.37	2009 guideline	Initiation of vancomycin until 72 h post completion	12 (24.0)	8
Pan et al, 2020, China, retrospective cohort study61	204	Trough	SCr: 0.85 (0.71–1.04)	KDIGO	Initiation of vancomycin until 48 h post completion	47 (23.0)	8
Adolescents and adults, reported trough only
Suzuki et al, 2015, Japan, retrospective cohort study62	63	Trough	SCr: 0.65 ± 0.28	2009 guideline‡	Between 2 and 10 days after initiation of vancomycin therapy	13 (20.6)	6
			CrCl: 113.1 ± 40.6
Hays et al, 2019, United States, retrospective cohort study63	50	Trough	SCr: 0.6	2009 guideline	NR†	20 (40.0)	8
Pediatrics, reported AUC and trough
Le et al, 2015, United States, retrospective cohort study64	680	Trough	SCr: 0.4 (0.3–0.5)	2009 guideline	NR†	40 (5.88)	8
Holsen et al, 2017, United States, retrospective cohort study65	93	Trough	SCr: 0.58 eGFR: 79.3	pRIFLE	Initiation of combination until 72 h post completion	18 (19.3)	8
Pediatrics, reported trough only
Geerlof et al, 2014, United States, retrospective cohort study66	102	Trough	SCr: 0.36 (0.1–0.9)	2009 guideline	Initiation of vancomycin until 72 h post completion	7 (6.86)	8
Sinclair et al, 2014, United States, retrospective cohort study67	175	Trough	eGFR: 174.72	pRIFLE	Initiation of vancomycin until 72 h post completion	24 (13.7)	9
McQueen et al, 2016, United States, retrospective cohort study68	185	Trough	NR†	2009 guideline	NR†	28 (15.1)	5
Nuhait et al, 2018, United States, retrospective cohort study69	248	Trough	SCr: 0.45 ± 0.1	KDIGO	NR†	4 (1.61)	8
			0.44 (0.38–0.5)
Woldu et al, 2018, United States, retrospective cohort study70	291	Trough	SCr: 0.34 (0.3–0.5)	pRIFLE	Initiation of vancomycin until 72 h post completion	19 (6.52)	9
Feiten et al, 2019, Brazil, retrospective cohort study71	110	Trough	SCr: 0.3 (0.23–0.4)	pRIFLE	During vancomycin therapy	13 (11.8)	8
			CrCl: 127.16 (91.33–158.32)
Hsu et al, 2019, United States, retrospective/prospective quasi-experimental study72	386	Trough	SCr: 0.35	2009 guideline‡	Initiation of vancomycin until 72 h post completion	33 (8.54)	8
Neonates, reported trough only
Reilly et al, 2019, United States, retrospective cohort study73	212	Trough	SCr: 0.61 ± 0 0.32, urine output: 3.6 ± 0.92 mL·kg−1·h−1	Increase in baseline SCr by 50% or reduction in urinary output by 50%	During vancomycin therapy	17 (8.01)	8

^*NOS Newcastle–Ottawa scale.

^†NR, Not Reported. Mean reported in mean ± SD; median reported in median (interquartile range) or (range lowest– highest).

^‡Study did not include requirement of consecutive increases in SCr for their definition of nephrotoxicity.

SCr, serum creatinine; CrCl, creatinine clearance; eGFR, estimated glomerular filtration rate; RIFLE, Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease; KDIGO, Kidney Disease Improving Global Outcomes; AKIN, Acute Kidney Injury Network; pRIFLE, Pediatric Risk, Injury, Failure, Loss, End Stage Renal Disease.

Most studies (n = 50) reported the baseline renal function of their cohorts, albeit with different parameters, such as serum creatinine (SCr), creatinine clearance, and estimated glomerular filtration rate.

The definitions of nephrotoxicity were varied; approximately half of the studies incorporated the 2009 guideline definition either fully or only the increase in SCr cutoff being specified, and 12 studies excluded the requirement of consecutive increases in SCr. The remaining studies used the various definitions of AKI, such as AKIN, RIFLE, pRIFLE, and KDIGO. Five studies^{9,19,20,27,31} included 2 or more definitions of nephrotoxicity and accordingly reported the respective nephrotoxicity rates. Compared with the 2009 guideline definitions, the AKIN and RIFLE criteria of nephrotoxicity had higher reported nephrotoxicity rates.

Nineteen studies did not define the actual period of nephrotoxicity occurrence, whereas 24 included the period of nephrotoxicity from vancomycin initiation either alone or in combination with another antibiotic till 48–96 hours post completion of vancomycin therapy. Nine studies considered nephrotoxicity during the period of vancomycin use, 3 included specified periods of 7 or 14 days, whereas only 2 studies considered nephrotoxicity 48 hours after initiation.

Forty-eight studies^{9,18–22,25,26,31–34} with^36–61 scores^63–67 of^69–73 ≥7 on the Newcastle–Ottawa scale were considered for the meta-analysis. Of these, 10 included patients aged <18 years (neonates–adolescents).^{63–67,69–73} These pediatric studies performed trough-guided monitoring, which precluded the comparison between trough-guided versus AUC-guided vancomycin TDM. Consequently, only 38 studies were included in the final meta-analysis.

AUC-Guided Versus Trough-Guided Monitoring

Four studies compared nephrotoxicity rates between AUC-guided and trough-guided monitoring.^9,18–20 The occurrence of nephrotoxicity was significantly lower in the AUC-guided approach than in the trough-guided approach (Fig. 2) [OR, 0.53 (95% CI, 0.32–0.89), P = 0.02]. A moderate amount of heterogeneity was noted between these studies (I² = 45%, P = 0.14).

FIGURE 2.

Forest plot comparing the risk of nephrotoxicity in AUC-guided versus trough-guided monitoring of vancomycin.

The overall vancomycin exposure and both the cumulative and maintenance doses were lower in AUC-guided monitoring (Table 3) generally lower as well. Patients in this group were noted to have lower vancomycin exposure as evidenced from the lower vancomycin trough levels.

TABLE 3.

Dose and Trough Level Differences Between AUC-Guided and Trough-Guided Cohorts in Studies With Head-To-Head Comparisons

Study	AUC-Guided Monitoring	Trough-Guided Monitoring	P
Finch et al, 20179 median (IQR)	CD 7000 (5000–9250) mg	CD 7500 (5438–10,250) mg	0.001
	T 12.0 (8.4–15.7) mg/L	T 15.0 (10.8–19.5) mg/L	<0.001
Neely et al, 201818 mean (min–max)	MD 1663 (700–4300) mg/d	MD 1818 (275–2760) mg/d	—
	22.3 (9.5–85.7) mg·kg·−1d−1	24.0 (7.4–52.6) mg·kg·−1d−1
	T 10.2 (3.5–29.6) mg/L	T 14.4 (3.8–27.2) mg/L
Oda et al, 202019 median (range)	MD 31.8 (15.4–82.9) mg·kg·−1d−1	MD 35.0 (11.9–76.9) mg·kg·−1d−1	0.926
	T 11.2 (4.7–18.5) mg/L	T 16.7 (15.0–19.5) mg/L	<0.001
Muklewicz et al, 202120 mean ± SD	MD 2114.3 ± 748.4 mg/d	MD 2105.0 ± 717.3 mg/d	0.872
	24.1 ± 7.3 mg·kg·−1d−1	26.6 ± 8.2 mg·kg·−1d−1	<0.001
	T 12.4 ± 5.3 mg/L	T 13.2 ± 6.2 mg/L	0.079

CD, cumulative dose; MD, maintenance dose; T, trough levels; IQR, interquartile range.

The pooled rate of nephrotoxicity from studies that used AUC-guided monitoring (n = 6) was 6.2% (95% CI, 2.9–9.5, I² = 85.98%) (see Figure 1A, Supplemental Digital Content 2, http://links.lww.com/TDM/A643).^9,19,24,25 This is lower than that in studies that used trough-guided monitoring (n = 36), wherein the pooled rate of nephrotoxicity was 17.0% (95% CI, 14.7–19.2, I² = 89.26%) (see Figure 1B, Supplemental Digital Content 2, http://links.lww.com/TDM/A643).

Owing to possible differences in nephrotoxicity rates based on the different definitions of nephrotoxicity, a sensitivity analysis was performed based on the 2009 guideline definition alone. The pooled rates of nephrotoxicity from the AUC-guided and trough-guided studies^{9,17,19,24,25} were 6.1% (95% CI, 2.6–9.5, I² = 88.62%) and 15.0% (95% CI, 12.4–17.7, I² = 87.21%), respectively. A trend toward lower nephrotoxicity rates in AUC-guided cohorts was observed (see Figures 2A, B, Supplemental Digital Content 2, http://links.lww.com/TDM/A643).

AUC-Guided Monitoring

Of the 14 studies that reported AUC-guided monitoring, 6 were included in the meta-analysis; of which, 3 studies each used Bayesian software and PK equations, respectively, to calculate the AUC. Regarding the nephrotoxicity rates, the pooled estimates were similar, 4.2% (95% CI, −0.4–8.8) and 7.5% (95% CI, 5.9–9.0), respectively (see Figures 3A, B Supplemental Digital Content 2, http://links.lww.com/TDM/A643).

Examination of the Classification and Regression Tree (CART)–derived cutoffs reported in the included studies revealed that the upper limits of the CART-derived break point for AUCs ranged from >563 mg × h/L to ≥1300 mg × h/L. A recent analysis of the CAMERA2 trial [27] examined the relationship of AUC and nephrotoxicity and reported that linear increases in nephrotoxicity rates were observed after an AUC_{24–48 h} of approximately 300 mg × h/L.

Trough-Guided Monitoring

Most studies that used trough-guided monitoring used the established cutoffs of 15–20 mg/L from the 2009 guideline to assess the risk of nephrotoxicity in their analysis. Of the 47 included trough-guided studies, 37 (78.7%) reported a correlation between trough-guided monitoring and nephrotoxicity outcomes.

Significant Predictors of Nephrotoxicity From Multivariate Analyses

Of the 38 studies included in this meta-analysis, the most common nephrotoxicity-associated predictors were use of concurrent nephrotoxins (n = 26, 68.4%), higher vancomycin trough concentration (n = 26, 68.4%), and duration of vancomycin use (n = 12, 31.6%) (Fig. 3). Notably, 4 of 5 studies (80%) with AUC-guided monitoring and 26 of 36 studies (72.2%) with trough-guided monitoring identified vancomycin AUC and higher vancomycin trough levels as nephrotoxicity predictors, respectively.

FIGURE 3.

Frequency of significant predictors of nephrotoxicity from multivariate analysis.

DISCUSSION

Studies published after the 2009 vancomycin guideline have largely examined the effect of trough levels and nephrotoxicity. Studies using AUC-guided dosing or reporting AUC-based parameters have emerged since 2017; hence, the evidence for AUC-guided monitoring is yet to be fully established in clinical practice. Therefore, although a vancomycin AUC of ≥400 mg × h/L is the PK–PD parameter that correlates satisfactorily with clinical efficacy,^2,3 clinical implementation and utilization of AUC-guided monitoring remain in the nascent stage. The geographical distribution of the studies included in this review is reflective of the global prevalence of MRSA infections.⁷⁴ Most of the data were from studies conducted in the adult and older adult populations; therefore, the current PK–PD target indices may be more appropriate for the general adult population than the pediatric or neonatal populations.

Comparison of AUC-guided versus trough-guided monitoring of vancomycin revealed that patients being managed through AUC-guided monitoring had lower exposure to vancomycin; this may partially explain the lower nephrotoxicity rates. This meta-analysis suggested that compared with trough-guided monitoring, AUC-guided monitoring reduced the overall rates of nephrotoxicity. The significant amount of heterogeneity observed may have been contributed by other factors such as the concomitant use of nephrotoxins. Prudent supportive measures, including adequate hydration and regular assessment of patients' clinical status and need for vancomycin, should be incorporated into clinical care protocols to ensure safe vancomycin use. Separate comparisons of nephrotoxicity rates between AUC-guided and trough-guided cohorts demonstrated a trend toward lower nephrotoxicity rates in studies with AUC-guided monitoring. This meta-analysis complements the research of Aljefri et al¹¹ and the additional data from and the analysis in this study substantiates the lower nephrotoxicity rates associated with AUC-guided monitoring. Moreover, in a previous meta-analysis on AUC-guided monitoring and AKI,¹⁰ AUC data were mainly derived from the first 48 hours of vancomycin therapy.

Examination of the included studies that reported AUC and timing of AUC-guided TDM revealed similar trends especially when AUC-guided TDM was used within the first 96 hours of therapy. Nonetheless, knowledge of the utility of repeated AUC measurements and the impact of this practice on nephrotoxicity rates remains limited; thus, further studies are required.

The heterogeneity noted in the meta-analysis may be attributable to the varied definitions and length of monitoring used across studies for vancomycin-associated nephrotoxicity. The 2020 vancomycin monitoring guideline suggested that an increase in SCr of ≥0.3 mg/dL over a 48-hour period, adapted from the AKIN/KDIGO definitions,^15,16 could be used as an indicator for vancomycin nephrotoxicity. A more sensitive definition such as the AKIN criteria may lead to higher reported rates of nephrotoxicity; 3 of the included studies^9,20,31 that reported both the old 2009 guideline and the AKIN criteria demonstrated a consistently higher prevalence of nephrotoxicity when the AKIN criteria was used. The sensitivity analysis revealed some differences in the crude rates; however, overall, the trend of lower nephrotoxicity rates in AUC-guided studies was observed. The use of a standardized definition for the assessment of vancomycin-associated nephrotoxicity in future studies will facilitate more meaningful comparisons of AUC-guided versus trough-guided TDM. Regarding the AUC target range for vancomycin, the latest American Society of Health-System Pharmacists guideline recommends targeting an AUC between 400 and 600 mg × h/L.⁸ A safe upper limit is yet to be conclusively established from the variable AUC cutoffs for nephrotoxicity based on CART analyses of the different studies, thereby indicating the prospects for further research.

The benefit of any 1 method of AUC determination cannot be elucidated as the pooled rates of nephrotoxicity were similar between studies that used Bayesian Software or PK equations. Therefore, based on available evidence, concurring that the Bayesian software approach is the preferred AUC determination method, as suggested by the 2020 guideline, is difficult.⁸ The costs incurred and training needed to implement the use of Bayesian Software may require justification by other potential benefits such as the need for single blood draws and faster optimization of therapy. Thus, the optimal method of AUC determination is beyond the scope of this review. However, regarding the benefit of AUC-guided monitoring over trough-guided monitoring, this review suggests that either method would be appropriate.

Therefore, although AUC-guided monitoring for vancomycin is being emphasized, a trough level may have clinical utility for the subsequent monitoring of patients who require a protracted course of vancomycin. Previous studies that examined the relationship between troughs and nephrotoxicity^4,5,76 reported an association between higher trough levels and increased risk of nephrotoxicity. Trough levels may fulfil the clinical need to monitor accumulation and alert the clinician to potential nephrotoxicity. The literature for the optimal monitoring of vancomycin for patients on a longer course of vancomycin remains limited; therefore, future studies could address whether trough-guided or AUC-guided monitoring beyond the first week of therapy would be more sensitive in detecting nephrotoxicity. Furthermore, cost-effectiveness analysis could be performed to assess if AUC-guided monitoring is superior in improving patient outcomes.

In the studies that included only the pediatric and neonatal populations, the overall nephrotoxicity rates were generally lower than that in studies including adults. Of the 2 studies that reported both AUC and trough values, both demonstrated an association of higher trough levels and nephrotoxicity, whereas only 1 demonstrated an association between higher AUC exposure and nephrotoxicity outcomes. Le et al have highlighted several useful PK indices regarding vancomycin monitoring in pediatrics, namely, that a trough of ≥15 mcg/mL and AUC ≥800 mg × h/L increased the risk of nephrotoxicity. Moreover, a CART analysis found that an elevated AUC >1063 mg× h/L was an important nephrotoxicity-predicting break point. This AUC break point is generally higher than that in several studies including adults.^24,26,27,29 However, the study of Le et al⁶⁴ excluded patients who received concomitant nephrotoxic medications; hence, the tolerance of exposure may have been high.

A head-to-head comparison between trough-guided versus AUC-guided monitoring in the pediatric and neonatal populations is lacking. Therefore, further studies are needed to evaluate the effect of AUC-guided monitoring and nephrotoxicity rates relative to the trough-guided monitoring. Moreover, current CART-derived thresholds and outcome data of toxicity are solely based on retrospective studies, and the application of these thresholds should be validated in prospective studies; prospectively collected data on nephrotoxicity outcomes may better elucidate the link between these PK–PD parameters and nephrotoxicity, providing a more accurate AUC upper limit.

In this review, potential variables that could affect vancomycin-related nephrotoxicity rates included the use of concomitant nephrotoxins, such as diuretics and aminoglycosides, and patients' comorbidities, age, and weight. Some studies have attempted to account for nephrotoxins by excluding patients on concurrent nephrotoxins.^45,46 Clinical states such as hypotension and chronic kidney diseases should prompt clinicians to be more prudent in monitoring for vancomycin-related nephrotoxicity. Age and weight may affect nephrotoxicity rates; patients who were older or heavier had higher nephrotoxicity rates. Accounting for all potential confounding variables in the assessment of vancomycin-related nephrotoxicity is challenging; therefore, future studies should consider the inclusion and adjustment for these variables when evaluating the role of trough-guided versus AUC-guided TDM. The low percentage of an AUC threshold nephrotoxicity predictor may be because of the limited number of studies investigating this variable.

A recently published review highlighted the limitations and challenges of AUC-guided monitoring.⁷⁹ Notably, studies evaluating AUC-guided monitoring were deemed to be limited because of methodological weaknesses and inconsistencies. In our review, the evidence was mainly derived from retrospective studies and indicated that nephrotoxicity rates can be affected by varied confounders. Nonetheless, based on the published evidence and pooled nephrotoxicity rates in this meta-analysis, patients managed by AUC-guided monitoring apparently have consistently lower nephrotoxicity rates than in trough-guided monitoring. Furthermore, the review⁷⁹ argues against the supposed benefit that AUC-guided monitoring lowers nephrotoxicity rates. Two quasi-experimental studies^30,80 that did not report a reduction in nephrotoxicity rates with AUC-guided monitoring in comparison with trough-guided monitoring were reviewed. The study of Vali et al³⁰ was not included in the meta-analysis because it had a Newcastle–Ottawa score of 6. The study of Meng et al⁸⁰ was not included in the review because it did not meet the study's inclusion criteria of at least 48 hours of vancomycin use. Therefore, owing to the exclusion of these studies, this review's conclusion differs from that of the study by Jorgensen et al. This suggests that there are clinical scenarios where AUC-guided monitoring may be comparable with trough-guided monitoring for vancomycin-related nephrotoxicity, such as shorter courses of vancomycin. Well-designed, prospective, randomized, controlled trials are needed to address this crucial research gap.

This review has certain limitations. First, most of the included studies were retrospective in design; hence, determining a causal relationship between vancomycin drug levels and nephrotoxicity was difficult. Second, a significant heterogeneity in the definitions of nephrotoxicity and its assessment was observed. Therefore, a standardized definition and assessment of vancomycin-related nephrotoxicity should be explored in future studies. Third, in this review, vancomycin TDM-related data pertaining to patients with renal dysfunction or those on renal replacement therapy and patients who received vancomycin through continuous infusion were limited. Future studies should evaluate these parameters. Fourth, this review did not address the effects of intermittent versus continuous dosing of vancomycin studies using the continuous infusion method. Review by Chu et al⁸¹ suggested that continuous vancomycin infusion may have lower nephrotoxicity rates compared with intermittent dosing. Therefore, future reviews should consider evaluating nephrotoxicity rates between institutions using continuous vancomycin dosing and Bayesian methods. Fifth, the number of vancomycin levels were not reported in most studies. The ability of the Bayesian method for vancomycin TDM to estimate AUC from a single vancomycin level is a key advantage of this method. Future studies should report the number of vancomycin levels per patient to allow the comparison of potential benefits of reduced blood drawing in institutions that use Bayesian methods for vancomycin titration. Sixth, the initial search was conducted 2 years ago, and new studies have been published on this topic. The search was rerun using the study's search strategy Supplemental Digital Content 1 (see, http://links.lww.com/TDM/A642) on September 14, 2022, and 3 additional studies^82–84 published after January 01, 2021, that assessed nephrotoxicity between the AUC-guided versus trough-guided approach were identified. However, dosing regimens were not explicitly mentioned in these studies; hence, they did not fulfil the inclusion criteria of this review. Nonetheless, these newer studies support the trend of lower nephrotoxicity rates in patients undergoing AUC-guided monitoring.

CONCLUSIONS

AUC-guided monitoring may be associated with a lower risk of vancomycin-related nephrotoxicity than trough-guided monitoring in the adult and older adult population. AUC-guided vancomycin TDM should be combined with proper patient assessment, regular review of the need for vancomycin continuation, and close monitoring with supportive care. Institutional implementation of AUC-guided vancomycin TDM should consider costs and staff training requirements for its appropriate use. Current research gaps include the optimal frequency of monitoring for AUC-guided vancomycin TDM and lack of studies on using AUC-guided dosing in the pediatric and neonatal populations.

Supplementary Material

tdm-45-519-s001.docx ^{(29KB, docx)}

tdm-45-519-s002.docx ^{(292.1KB, docx)}

tdm-45-519-s003.pdf ^{(247.4KB, pdf)}