Variability in Telavancin Cross-Reactivity among Vancomycin Immunoassays

Abstract

Telavancin is a semisynthetic lipoglycopeptide with a dual mechanism of action against Gram-positive pathogens. Two brief reports have suggested potential cross-reactivity of telavancin with the vancomycin particle-enhanced turbidometric immunoassay (PETIA). The purpose of this study was to evaluate several commercially available vancomycin immunoassays (fluorescence polarization [FPIA], enzyme-multiplied immunoassays [EMIT], PETIA, and chemiluminescent immunoassay [CMIA]) for cross-reactivity with telavancin. Seven sites were selected to analyze serum samples for vancomycin. Each site received a set of samples (n = 18) which combined drug-free serum with telavancin, 7-OH telavancin metabolite, or vancomycin. Immunoassays demonstrating potential cross-reactivity were further evaluated by sending a duplicate sample set to multiple laboratories. Cross-reactivity was defined as the percent theoretical concentration (reported concentration/theoretical concentration × 100). No cross-reactivity was seen with FPIA or EMIT. Within the theoretical concentration range of 5 to 120 μg/ml of telavancin, the Synchron PETIA system reported vancomycin concentrations ranging from 4.7 to 54.2 μg/ml compared to vancomycin concentrations from 1.1 to 5.6 μg/ml for the Vista PETIA system. The Architect CMIA system reported vancomycin concentrations in the range of 0.27 to 0.97 μg/ml, whereas Advia Centaur XP CMIA reported vancomycin concentrations between 1.6 and 31.6 μg/ml. The Architect CMIA immunoassay had the lowest percent cross-reactivity (0.8 to 5.4%), while the Synchron PETIA immunoassay demonstrated the highest percent cross-reactivity (45.2 to 53.8%). Telavancin samples measured by liquid chromatography-mass spectroscopy were within 93.9 to 122% of theoretical concentrations. Vancomycin concentrations were not measured in any 7-OH telavancin-spiked sample. Vancomycin concentrations measured by liquid chromatography-mass spectroscopy were within 57.2 to 113% of theoretical concentrations. PETIA and CMIA measured vancomycin concentrations in telavancin-spiked samples. Significant variability in percent cross-reactivity was observed for each platform regardless of immunoassay method.

INTRODUCTION

Immunoassays are commonly used to monitor vancomycin serum concentrations in clinical practice. The current immunoassay technology for determining serum drug concentration uses vancomycin-specific antibodies and enzymatic reactions which cause quantitative changes in solution color, fluorescence, or turbidity. Commercially available vancomycin immunoassays vary by specific methodology and include fluorescence polarization immunoassays (FPIA), enzyme-multiplied immunoassays (EMIT), particle-enhanced turbidimetric immunoassays (PETIA), and chemiluminescent immunoassays (CMIA). These immunoassays have demonstrated a potential for cross-reactivity with nonvancomycin moieties (e.g., vancomycin crystalline degradation product 1, or CDP-1) (1).

Telavancin is a lipoglycopeptide antibacterial agent originally derived from vancomycin. It exhibits concentration-dependent bactericidal effects via a dual mechanism of action that combines the inhibition of cell wall synthesis and the disruption of membrane barrier function. Telavancin is approved in the United States and Canada for the treatment of adult patients with complicated skin and skin structure infections caused by susceptible Gram-positive pathogens (2). In Europe, telavancin has been approved for treatment of methicillin-resistant Staphylococcus aureus nosocomial pneumonia when other alternatives are unsuitable. Telavancin recently was approved in the United States for hospital-acquired and ventilator-associated bacterial pneumonia caused by susceptible isolates of Staphylococcus aureus (methicillin-susceptible and -resistant isolates), reserved for use when alternative agents are not suitable (2). The recommended dosage regimen for telavancin is 10 mg/kg of body weight intravenously infused over 60 min every 24 h in patients with normal renal function (e.g., creatinine clearance of >50 ml/min).

In healthy adult subjects and patients, the 10 mg/kg dosage regimen results in mean steady-state peak plasma concentrations ranging from 101 to 116 μg/ml (2–4). With an elimination half-life of approximately 8 h, the mean trough plasma concentrations of telavancin ranged from 8 to 11 μg/ml in subjects and patients with normal renal function (3, 4). Additionally, telavancin has a 7-OH metabolite, THRX-651540, which achieves peak plasma concentrations of ∼0.5 μg/ml (5).

A case series suggested telavancin concentrations are detectable with a vancomycin PETIA (Synchron LX system; Beckman Coulter, Inc., Brea, CA, USA) (6). The authors reported 4 patients receiving telavancin with detectable vancomycin concentrations ranging from 5.5 to 49.9 μg/ml. A subsequent in vitro study using telavancin-spiked serum samples confirmed these results using the same PETIA (Synchron LX) (7). Evans et al. also demonstrated that a second PETIA (Dimension Vista; Siemens Healthcare Diagnostics, Inc., Newark, DE) could detect vancomycin concentrations in telavancin-spiked samples (7). Neither study evaluated the potential cross-reactivity with other commercially available immunoassays (FPIA, EMIT, or CMIA) or the metabolite of telavancin.

Our study was conducted sequentially with two objectives: (i) identify commercially available vancomycin immunoassays with potential cross-reactivity for telavancin or 7-OH telavancin, and (ii) validate those immunoassays demonstrating potential cross-reactivity by testing multiple sites in duplicate.

(This work was presented, in part, at the 24th European Congress of Clinical Microbiology and Infectious Diseases [ECCMID], Barcelona, Spain, May 2014).

MATERIALS AND METHODS

Laboratory sites.

Laboratory sites were eligible for inclusion if they were clinical or reference laboratories servicing either a hospital or clinic-based population and assayed vancomycin serum concentrations onsite using an enzyme immunoassay methodology. Each site was a clinical laboratory which performed vancomycin assays for routine patient care. The seven laboratories included stated that the reagent kits used with their analyzer platform were those supplied by the manufacturer of the platform.

Description of immunoassays. (i) Phase one.

Seven immunoassays utilizing 4 different methods were included. FPIA (n = 1), EMIT (n = 2), PETIA (n = 2), and CMIA (n = 2) were evaluated in phase one (Table 1). Single pooled human serum samples were spiked with the following drugs and concentrations: telavancin, 0, 5, 10, 20, 40, 60, and 100 μg/ml; 7-OH telavancin, 0, 0.25, 0.5, 1, and 2 μg/ml; and vancomycin, 0, 5, 10, 20, 40, and 60 μg/ml. In phase one, a single spiked serum sample of every drug concentration was sent to be assayed at each participating laboratory. Sites were blinded to sample content and concentration, and each sample was analyzed singly and reported per individual laboratory protocol to investigators as a random vancomycin concentration. A set of spiked serum samples at each drug concentration also was assayed by liquid chromatography-mass spectroscopy (LC-MS).

TABLE 1.

Description of immunoassays evaluated in first phase of the study

Method	Immunoassaya
FPIA	Integra 800; Roche Diagnostics, Indianapolis, IN
EMIT	Olympus Au 680 Analyzer; Olympus America, Inc., Center Valley, PA
EMIT	Roche Modular System: P-Module; Roche Diagnostics, Indianapolis, IN
PETIA	UniCel DxC 800 Synchron; Beckman-Coulter, Inc., Brea, CA
PETIA	Dimension Vista; Siemens Healthcare Diagnostics, Inc., Tarrytown, NY
CMIA	Advia Centaur XP; Siemens Healthcare Diagnostics, Inc., Tarrytown, NY
CMIA	Architect i2000SR System; Abbott Diagnostics Division, Chicago, IL

^aEach participating laboratory in phases one and two reported their lower limit of quantitation for vancomycin in serum as <1.4 μg/ml for Integra FPIA, <5.0 μg/ml for Olympus and P-Module EMIT, <3.5 or 4.0 μg/ml for Synchron PETIA, <0.8 μg/ml for Vista PETIA, <0.67 or <0.7 μg/ml for Centaur CMIA, and <0.2 or <0.24 μg/ml for Architect CMIA.

Phase two.

The immunoassays which demonstrated cross-reactivity in phase one were further evaluated by finding two additional laboratories which used the same immunoassay. Pooled human serum samples were spiked with the following drugs and concentrations: telavancin, 0, 5, 10, 20, 40, 60, 80, 100, and 120 μg/ml; vancomycin, 0, 10, 20, and 40 μg/ml. In phase two, duplicate spiked serum samples at every drug concentration were sent to be assayed at each participating laboratory. Sites were blinded to sample content and concentration, and each sample was analyzed singly and reported per individual laboratory protocol to investigators as a random vancomycin concentration. Additionally, a duplicate set of aliquoted serum samples at each drug concentration was assayed via LC-MS.

Sample preparation.

Serum samples were prepared using commercially available pooled human serum from healthy donors (BP2657-100; Fisher Scientific International Inc., Richmond, British Columbia; two lots were used: 133995 [used in phase one] and 125405 [used in phases one and two]). Telavancin and its 7-OH metabolite, THRX-651540, were provided by Theravance, Inc. (South San Francisco, CA). Samples were spiked individually in phase one with 10 μg/ml and 5 μg/ml concentrations prepared through serial dilution of 100 μg/ml and 60 μg/ml samples, respectively. In phase two, stock solutions of each sample concentration were made, and an aliquot was used to prepare individual samples. The stock vancomycin solution (V2002-100; Sigma-Aldrich, St. Louis, MO) (1 mg/ml) was prepared by dissolving weighed powder in a 0.9% NaCl solution. Stock telavancin solution (1 mg/ml) and 7-OH metabolite (1 mg/ml) were prepared by dissolving weighed powder in sterile 0.9% NaCl–0.2% HCl solution. The stock solutions of vancomycin and telavancin were further diluted to individual tested concentrations. Samples were aliquoted into red-top Vacutainers, frozen, and shipped within 24 h of preparation on dry ice via priority ground/air transportation to the laboratory sites.

LC-MS analysis.

Quantitative analysis of samples for confirmation of theoretical concentrations was performed with a qualified liquid chromatography-tandem mass spectroscopy (LC-MS/MS) method at Theravance, Inc. (South San Francisco, CA). Samples were analyzed on an API 4000 LC-MS/MS (AB Sciex, Framingham, MA). In phase one, telavancin and vancomycin LC-MS standards were prepared in human serum at ranges of 0.025 to 125 μg/ml for telavancin, 0.0125 to 125 μg/ml for vancomycin, and 0.025 to 12.5 μg/ml for 7-OH telavancin. In phase two, LC-MS standards were prepared in human serum at ranges from 0.050 to 125 μg/ml for telavancin and 0.050 to 50 μg/ml for vancomycin. Deuterated telavancin was used as the internal standard. Samples were prepared by adding 50 μl of serum sample with 250 μl of 850 ng/ml internal standard in water and 0.2% formic acid. Samples were centrifuged, 180 μl of supernatant was extracted and combined with 60 μl of acetonitrile, and 10 μl was injected into the column. A Thermo C₁₈ column (2.1 by 50 mm, 5 μm particle size) at a flow rate of 0.25 ml/min was used in phase one with the following retention times: telavancin, 1.77 min; vancomycin, 1.74 min; and 7-OH telavancin, 1.73 min. A Waters Xbridge Phenyl column (2.1 by 50 mm, 5 μm particle size) at a flow rate of 0.25 ml/min was used in phase two with the following retention times: telavancin, 1.53 min; vancomycin, 1.33 min. Mobile phase A contained 0.2% formic acid in high-performance liquid chromatography (HPLC)-grade water, and mobile phase B contained 0.2% formic acid in acetonitrile. The interday coefficients of variation ranged from 1.2% to 14.5% for telavancin and 7.3% to 27.1% for vancomycin. The lower limit of quantitation was 50 ng/ml for telavancin and vancomycin and 25 ng/ml for 7-OH telavancin (phase one only).

Assessment of cross-reactivity.

Results are reported as a percentage of theoretical concentration (reported concentration/theoretical concentration × 100). In phase two of the study, means, standard deviations, and percent theoretical concentrations are reported for each immunoassay group.

RESULTS

Phase one.

The results of telavancin-spiked samples in phase one are summarized in Table 2. Results are presented as percent theoretical (reported concentration/theoretical concentration × 100). The FPIA system did not consistently report vancomycin concentrations in telavancin-spiked samples. The 40 and 100 μg/ml samples reported vancomycin concentrations as 3.75% and 1.5% theoretical, respectively, but this finding was not replicated for any other spiked sample. Both EMIT systems reported no detectable vancomycin concentrations.

TABLE 2.

Reported vancomycin concentrations in spiked telavancin samples (phase one)

Theoretical telavancin concn (μg/ml)	Reported vancomycin concna (% theoretical)
	FPIA (Integra)	EMIT		PETIA		CMIA
		Olympus	P-Module	Synchron	Vista	Architect	Centaur	LC-MS
0	BQL	BQL	BQL	BQL	BQL	BQL	1.3	BQL
5	BQL	BQL	BQL	BQL	BQL	0.2 (4.0)	1.1 (22.0)	5.9 (118.8)
10	BQL	BQL	BQL	BQL	BQL	0.2 (2.0)	2.4 (24.0)	12.2 (122.0)
20	BQL	BQL	BQL	8.9 (44.5)	1.2 (6.0)	0.5 (2.5)	4.8 (24.0)	23.4 (117.0)
40	1.5 (3.75)	BQL	BQL	16 (40.0)	2.3 (5.8)	0.6 (1.5)	7.5 (18.8)	46.3 (115.8)
60	BQL	BQL	BQL	25.9 (43.2)	2.4 (4.0)	0.9 (1.5)	15.5 (25.8)	71.3 (118.8)
100	1.5 (1.5)	BQL	BQL	38.5 (38.5)	4.6 (4.6)	0.9 (0.9)	25.5 (25.5)	110.0 (110.0)

^aBQL, below the quantitation limit.

The PETIA systems consistently demonstrated detectable vancomycin concentrations in telavancin-spiked samples above a concentration of 10 μg/ml. The Synchron DxC reported concentrations in the range of 38.5 to 44.5% theoretical, while the Vista 1500 reported concentrations in the range of 4.0 to 6.0% (Table 2).

Both CMIA systems consistently reported vancomycin concentrations in telavancin-spiked samples. The Architect i2000 system demonstrated cross-reactivity in the range of 0.9 to 4.0%, and the Advia Centaur XP system demonstrated cross-reactivity in the range of 18.8 to 25.8%.

LC-MS-analyzed samples were 110 to 122% of theoretical telavancin concentrations. 7-OH telavancin metabolite samples did not have detectable vancomycin concentrations with any vancomycin immunoassay in the tested range (0.25 to 2.0 μg/ml) except for one FPIA control sample with no spiked drug, which reported a value of 1.4 μg/ml.

Phase two.

Based on the results from phase one, two PETIA (Synchron and Vista) and two CMIA (Architect and Advia Centaur XP) systems were included for further analysis. Twelve reference laboratories (n = 3 per immunoassay), including the original four tested in phase one, were evaluated. The mean concentration ± standard deviation (% theoretical) for each immunoassay is summarized in Table 3. Within the theoretical concentration range of 5 to 120 μg/ml of telavancin, the Synchron PETIA reported the highest range of vancomycin concentrations, from 4.7 to 54.2 μg/ml. The Vista PETIA reported concentrations ranging from 1.1 to 5.6 μg/ml (Table 3). The Architect CMIA reported vancomycin concentrations in the range of 0.27 to 0.97 μg/ml, and the Advia Centaur XP CMIA reported vancomycin concentrations between 1.6 and 31.6 μg/ml.

TABLE 3.

Reported vancomycin concentrations in spiked telavancin samples (phase two)

Theoretical telavancin concn (μg/ml)	Reported vancomycin concna (mean ± SD, in μg/ml [% theoretical])
	PETIA		CMIA		LC-MS
	Synchron	Vista	Architect	Centaur
0	BQL	BQL	BQL	BQLb	BQLc
5	BQL	1.1 ± 0.21d (22.7)	0.27 ± 0.06e (5.4)	1.6 ± 0.28 (32.3)	5.7 ± 0.1 (113.6)
10	4.7 ± 0.6f (47.3)	1.2 ± 0.22g (11.8)	0.33 ± 0.05h (3.3)	3.0 ± 0.19 (29.6)	9.7 ± 0.3 (97.3)
20	10.3 ± 1.3 (51.5)	1.5 ± 0.48i (7.3)	0.47 ± 0.03 (2.3)	6.0 ± 0.29 (30.0)	22.0 ± 2.5 (110)
40	20.4 ± 1.8 (51.1)	2.1 ± 0.46 (5.3)	0.60 ± 0.02 (1.5)	10.3 ± 0.63 (25.8)	43.6 ± 4.7 (108.9)
60	29.6 ± 2.9 (49.3)	3.0 ± 0.37 (5.0)	0.74 ± 0.04 (1.2)	15.1 ± 0.72 (25.1)	56.4 ± 2.9 (93.9)
80	39.8 ± 4.4 (49.7)	4.0 ± 0.39 (5.0)	0.85 ± 0.05 (1.1)	21.6 ± 1.68 (27.0)	80.8 ± 4.4 (101)
100	53.8 ± 3.7 (53.8)	4.9 ± 0.32 (4.9)	0.90 ± 0.08 (0.9)	24.9 ± 1.14 (24.9)	103.2 ± 6.9 (103.2)
120	54.2 ± 4.5 (45.2)	5.6 ± 0.5 (4.7)	0.97 ± 0.03 (0.8)	31.6 ± 1.43 (26.3)	124.5 ± 16.3 (103.8)

^aBQL, below the quantitation limit.

^bOne sample was 0.09 μg/ml.

^cOne sample was 50 ng/ml.

^dThree of 6 samples were BQL.

^eTwo of 6 samples were BQL.

^fTwo of 6 samples were BQL.

^gTwo of 6 samples were BQL.

^hOne of 6 samples was BQL.

ⁱOne of 6 samples was BQL.

The Architect CMIA has the lowest percent theoretical values (0.8 to 5.4%), while the Synchron PETIA demonstrated the highest percent theoretical values (45.2 to 53.8%). Standard deviation values also were higher for the PETIA systems than the CMIA systems. Except for the Synchron PETIA, immunoassays typically demonstrated higher percent theoretical concentrations in the lower range of 5 to 10 μg/ml versus the higher range of 20 to 120 μg/ml. A nonlinear relationship was observed between spiked telavancin concentrations and reported vancomycin concentrations. The Architect, Synchron, and Vista immunoassay systems reported undetectable concentrations for 11/54 samples in the range of 5 to 20 μg/ml. LC-MS telavancin samples were within 93.9 to 113.6% of theoretical concentrations.

Vancomycin concentrations.

Vancomycin concentrations for both study phases are summarized in Tables 4 and 5. In phase one, vancomycin concentrations determined by LC-MS were within 57 to 93% of theoretical concentrations (Table 4). In phase two, vancomycin concentrations determined by LC-MS were within 94 to 113% of theoretical concentrations (Table 5). Overall there was good agreement between reported concentrations from laboratories and those determined by LC-MS. In phase one, vancomycin concentrations between 20 and 60 μg/ml reported from clinical laboratories were within 90 to 126% of LC-MS concentrations and 80 to 117% of theoretical concentrations. The 5 μg/ml vancomycin samples were reported as nondetectable by three of seven laboratories, with a mean concentration of 3.8 in the remaining laboratories. The 10 μg/ml vancomycin samples were undetectable in 2/7 laboratories with a mean concentration of 5.3 μg/ml in the remaining laboratories. In phase two, the mean reported vancomycin concentrations for each immunoassay were within 87 to 110% of LC-MS-determined concentrations and 92 to 104% of theoretical concentrations.

TABLE 4.

Reported vancomycin concentrations in spiked vancomycin samples (phase one)

Theoretical vancomycin concn (μg/ml)	Reported vancomycin concn (% theoretical/% LC-MS)							LC-MS
	FPIA (Integra)	EMIT		PETIA		CMIA
		Olympus	P-Module	Synchron	Vista	Architect	Centaur
0	1.4	BQLa	BQL	BQL	BQL	BQL	1.2	BQL
5	5.8 (116/172)	BQL	BQL	BQL	2.3 (46/68.2)	1.2 (24/35.6)	5.9 (118/175)	3.37 (67.4)
10	4.5 (45/79)	7.2 (72/126)	BQL	BQL	4.5 (45/78.7)	4.7 (47/82.2)	5.8 (58/101)	5.72 (57.2)
20	19.1 (96/114)	16.1 (81/96)	18.6 (93/111)	16.1 (81/96)	18.2 (91/109)	16.2 (81/97)	19 (95/114)	16.7 (83.5)
40	41.5 (104/119)	35.4 (89/101)	41.6 (104/119)	31.8 (80/91)	36.2 (91/103)	33.3 (83/95)	43 (108/123)	35 (87.5)
60	70 (117/126)	52 (87/94)	60.3 (101/108)	50.1 (84/90)	50 (83/90)	56.4 (94/101)	61.6 (103/111)	55.6 (92.7)

^aBQL, below the quantitation limit.

TABLE 5.

Reported vancomycin concentrations in spiked vancomycin samples (phase two)

Theoretical vancomycin concn (μg/ml)	Reported vancomycin concna (mean ± SD, in μg/ml [% theoretical/% LC-MS])
	PETIA		CMIA		LC-MS
	Synchron	Vista	Architect	Centaur
0	BQL	BQLb	BQL	BQLc	BQL
10	9.2 ± 0.8 (92/95)	10.0 ± 0.5 (100/104)	10.2 ± 0.3 (102/106)	9.3 ± 0.5 (93/97)	9.6 ± 0.2 (96)
20	20.6 ± 1.1 (103/91)	20.8 ± 1.0 (104/92)	20.8 ± 0.5 (104/92)	19.7 ± 1.3 (98/87)	22.5 ± 3.5 (113)
40	41.3 ± 4.9 (103/110)	38.4 ± 2.1 (96/102)	39.2 ± 1.3 (98/105)	38.5 ± 2.9 (96/103)	37.5 ± 3.2 (94)

^aBQL, below the quantitation limit.

^bOne sample was 1.3 μg/ml.

^cTwo samples were 0.33 and 0.14 μg/ml.

DISCUSSION

This study investigated telavancin cross-reactivity with vancomycin immunoassays. In the first phase of the investigation, the FPIA Integra system and two EMIT assays (P-module and Olympus AU) do not appear to significantly cross-react with physiologic concentrations of telavancin (5 to 100 μg/ml) (Table 2). The PETIA systems (Synchron and Vista) and the two CMIA systems (Architect and Advia Centaur XP) also demonstrated consistent cross-reactivity with telavancin. Samples spiked with 7-OH telavancin did not demonstrate any cross-reactivity over the tested concentration range (0.25 to 2 μg/ml).

In the second phase of the investigation, the Synchron PETIA reported the highest percent theoretical values (∼50%) among the 4 platforms tested, which was consistent with previous literature (6, 7). The Advia Centaur XP CMIA reported the next highest level (24.9 to 32.3%). The Vista PETIA and Architect CMIA reported lower percent theoretical values (4.7 to 22.7% and 0.8 to 5.4%, respectively). With Vista PETIA and the two CMIA systems, percent theoretical is not constant across the range of 5 to 120 μg/ml and tends to be higher at lower concentrations. The percent theoretical was higher with lower concentrations in Vista PETIA, Advia Centaur CMIA, and Architect CMIA platforms but not with the Synchron PETIA. This suggests a nonlinear relationship between spiked telavancin concentrations and reported vancomycin concentrations. This relationship may be due to telavancin sample concentrations approaching the lower limit of quantitation of vancomycin in these immunoassays. In the range of 5 to 10 μg/ml telavancin, the Architect, Synchron, and Vista immunoassay systems did not consistently report vancomycin concentrations.

The vancomycin results of phases one and two support that the studied immunoassays and clinical sites accurately determine vancomycin concentrations compared to LC-MS. In phase two of the study, LC-MS and reported vancomycin concentration were consistent with theoretical values for vancomycin-spiked samples. However, in the phase one study, at a concentration of 5 to 10 μg/ml, LC-MS and reported vancomycin concentrations were consistent with each other but not consistent with theoretical vancomycin values (Table 4). This may be due to error during sample preparation, differences in sample preparation in phases one and two, and/or differences between assays in the lower limit of quantitation. Some of the immunoassays' lower limit of quantitation approached theoretical concentrations (e.g., <5.0 μg/ml for the EMIT assays and <3.5 or 4.0 for Synchron PETIA).

One case series and one in vitro study previously demonstrated the Synchron and Vista PETIA can report vancomycin concentrations in telavancin samples (6, 7). Gelfand et al. initially reported the potential cross-reactivity of telavancin and the Synchron PETIA based on serum samples drawn from patients receiving telavancin therapy (6). Evans et al. performed an in vitro analysis similar to the present study which demonstrated the Synchron PETIA reports theoretical vancomycin concentrations of approximately 61%, while the Vista PETIA reports 6 to 15% theoretical concentrations (7). Their results are consistent with our findings.

Our study has several strengths compared to the previous reports. We evaluated several other vancomycin immunoassays (e.g., FPIA, EMIT, and CMIA) for cross-reactivity between telavancin and vancomycin. The metabolite of telavancin was evaluated for cross-reactivity among the immunoassays. Spiked serum sample concentrations for all compounds were confirmed via LC-MS. Several clinical laboratories which used the same immunoassay and were blinded to sample content confirmed and validated our cross-reactivity findings. However, the generalizability of these results may be limited by the use of pooled human serum spiked with drug versus clinically obtained serum concentrations.

Commercial vancomycin immunoassays function by measuring vancomycin-specific antibody binding to marker-labeled vancomycin. Even if two immunoassays share a similar general enzyme technique, each specific immunoassay utilizes its own analyzer platform and reagent testing kit provided by the manufacturer. This may explain why the two CMIA systems (Architect and Advia Centaur XP) and PETIA systems (Synchron and Vista) report substantially different values with the same spiked telavancin concentrations (Table 2).

Currently, no studies have evaluated the potential mechanisms for these telavancin results with vancomycin assays. Telavancin is structurally similar to, but distinct from, vancomycin. The primary structural difference is a hydrophobic side chain (decylaminoethyl) attached to the vancosamine sugar and the phosphonomethyl aminomethyl group on the 4 position of amino acid 7 (8). The observed cross-reactivity for certain immunoassays may be due to an inability to discriminate this structural difference. Since the mechanism is unknown, it remains unclear why certain immunoassay systems (e.g., CMIA and PETIA) report higher vancomycin concentrations with the same spiked telavancin concentration. The various magnitudes and degrees of cross-reactivity among immunoassays likely are due to differences in the underlying immunosorbent methodology and/or reagent manufacturer. All of these considerations should be evaluated during the development and quality assurance phase of all current and future vancomycin immunoassays to ensure cross-reactivity with telavancin is not present.

The previous reports by Gelfand et al. and Evans et al. suggested that the Synchron vancomycin immunoassay could be used to monitor serum telavancin concentrations (6, 7). The present study suggests this is not be feasible for two reasons. First, using immunoassays in this manner would require the extrapolation of the reported vancomycin concentration to a theoretical telavancin concentration with a scaling factor. Because each immunoassay reacts differently with telavancin, the clinician would have to be certain as to the exact immunoassay used by their institution to apply the specific scaling factor. Significant confusion could arise between laboratories and clinicians regarding which specific immunoassay is available at their institution (i.e., there are multiple systems utilizing PETIA). Additionally, laboratories often change immunoassays. Second, telavancin has been shown to achieve mean steady-state trough concentrations between 7 and 16 μg/ml, depending on renal function (2–4). We demonstrate that the vancomycin immunoassays are the least able to detect telavancin consistently within the range of 5 to 10 μg/ml. The Synchron system is unable to reliably detect spiked samples at a concentration of 10 μg/ml (2/6 samples were reported to be below the quantitation limit). Additionally, the Synchron results had the highest standard deviation values of the immunoassay platforms. Therefore, predicting telavancin concentrations with the Synchron PETIA would be particularly inaccurate and unreliable within the physiologic range of trough plasma concentrations. In addition, trough plasma concentrations of telavancin are more likely to be associated with a larger likelihood of variability due to different rates of elimination among patients (4). This variability in percent theoretical in addition to the possible nonlinear relationship between concentrations suggests these immunoassays cannot accurately predict telavancin concentrations.

Overall, we demonstrate and verify the potential cross-reactivity of telavancin with vancomycin immunoassays. We also demonstrate this potential cross-reactivity varies significantly by specific immunoassay regardless of the underlying methodology.