Vancomycin is an antibiotic glycopeptide discovered in the 1950s and approved by Food and Drug Administration (FDA) to treat severe infections caused by gram-positive bacteria, especially methicillin-resistant Staphylococcus aureus (MRSA).12 Despite more than 60 years of clinical use, questions still need to be answered regarding strategies to individualize therapy, minimize toxicity, and prevent increased bacterial resistance.3

Vancomycin therapeutic level monitoring is highly recommended due to its narrow therapeutic index,4 wide interpatient and intrapatient pharmacokinetic variability along with exposure-dependent nephrotoxicity.5 Therapeutic Drug Monitoring (TDM) rests on pharmacokinetic/pharmacodynamic (PK/PD) studies, which embrace three PK/PD indices: ratio of maximum concentration to minimum inhibitory concentration (MIC), ratio of area under the concentration-time curve to MIC and time concentration above MIC.6 Vancomycin antimicrobial effect is dependent on the ratio of the area under the plasma concentration-time curve during 24 h (AUC0–24) over the minimum inhibitory concentration by broth microdilution (AUC0–24/MICBMD), and a value between 400 and 600 mg.hour/L is indicated as the best target to treat MRSA invasive infections.7

As vancomycin quantification is mandatory for TDM, robust assays are required to ensure adequate, effective, and safe serum concentrations.8 Although immunoassays are the most commonly used for vancomycin measurement in clinical laboratories, chromatography methods coupled with tandem mass spectrometry are considered the gold standard for their high specificity and reproducibility.9 Even though immunoassays enable simple and fast measurements by consolidated analytical systems and commercial kits, inconsistent accuracy and precision may occur due to interference of cross-reacting substances, such as the biologically inactive vancomycin crystalline degradation product, CDP-1.10

Vancomycin is mainly excreted by glomerular filtration. Serum half-life in patients with normal renal function is approximately 6 h, however, renal impaired patients have a prolonged half-life of up to 200 h.411 The prolonged exposure of the drug to body temperatures allows CDP-1 to accumulate.12 Since CDP-1 has chemical structures similar to vancomycin, there are several reports about cross-reaction with anti-vancomycin antibodies in non-specific immunoassays, leading to falsely elevated vancomycin levels.12-16

Infectious diseases are the second leading cause of death in patients with chronic kidney disease undergoing hemodialysis (HD), preceded only by cardiovascular disorders.11 The skin disruption by HD vascular access is the main route of infection, leading to bacteremia commonly caused by MRSA.11 Vancomycin is widely prescribed for patients undergoing HD, and TDM is recommended by the latest consensus guideline published by Ryback et al.17 Therefore, it is crucial to elucidate if currently available immunoassays are susceptible to CDP-1 interference.

This study aims to compare two immunoassays, Chemiluminescent Microparticle Assay (CMIA) and Kinetic Interaction of Microparticles in Solution (KIMS), with Liquid Chromatography coupled with Tandem Mass Spectrometry (LC-MS/MS) for vancomycin measurements from patients undergoing HD to avoid dose adjustment based on falsely elevated vancomycin serum levels.

The CMIA evaluation consisted of 64 serum samples, obtained from 42 patients, measured for vancomycin concentration as part of the TDM routine. Of the 42 patients, 28 (66.7 %) were men, and the mean age was 58 years (range: 30–84 years); 38 were receiving continuous renal replacement therapy (CRRT), 2 were receiving intermittent HD (IHD), 1 patient changed from CRRT to IHD and 1 was receiving peritoneal dialysis. The KIMS assessment consisted of 54 samples from 23 patients, measured for vancomycin concentration as part of the TDM routine. There were 17 (73.9 %) men, and the mean age was 75 years (range: 34–89 years); 14 were receiving CRRT, 5 were receiving IHD, 1 patient changed from CRRT to IHD and 3 were receiving peritoneal dialysis. The frequency of prescribed vancomycin doses for studied samples in the CMIA and KIMS cohorts is listed in Table 1. Vancomycin doses were administered every 12 h in 73 % and 63 % and every 24 h in 22 % and 33 % of CMIA and KIMS cohorts, respectively. In both CMIA and KIMS evaluations, all patients used high-flux membrane dialysis, except one who was on peritoneal dialysis.

In CMIA evaluation, vancomycin concentrations ranged from 7.30 to 58.50 μg/mL with CMIA and from 7.39 to 56.23 μg/mL with LC-MS/MS. CMIA and LC-MS/MS measurements were highly correlated, with a Spearman rank correlation coefficient value of 0.840 (p < 0.001, 95 % CI 0.749‒0.900). No deviation of linearity was observed (p = 0.81). Regression analysis is shown in Fig. 1A and the results are summarized in Table 2. No systematic bias was observed, as the intercept was not significantly different from 0 μg/mL (95 % CI: −0.6931 to 5.909). Nonetheless, the slope of 0.82 (95 % CI 0.6873‒0.9671) showed a small proportional bias. The mean difference between vancomycin concentrations measured by CMIA and LC-MS/MS was −1.0 μg/mL (−10.38 to 11.26). Bland-Altman plots are shown in Fig. 2A. Most of the differences between both methods were within the ±1.96 standard deviation range. LC-MS/MS measured levels were, on average, 2.64 % higher than CMIA (range: −57.88 % to 39.48 %). The difference between values obtained was within 20 % of the mean for 67 % of the samples (43/64), following ISR.

Fig. 1.

Passing Bablok regression plots comparing vancomycin concentrations measured by (A) LC-MS/MS and CMIA and (B) LC-MS/MS and KIMS. The solid line represents the regression line, the dashed lines represents the confidence interval for the regression line, and the dotted line represents an identity line (x = y). CMIA, Chemiluminescent Microparticle Assay; KIMS, Kinetic Interaction of Microparticles in Solution; LC-MS/MS, Liquid Chromatography coupled with Tandem Mass Spectrometry.

(0.18MB).

Fig. 2.

Bland-Altman plots comparing vancomycin concentrations measured by (A) LC-MS/MSCMIA and CMIA and (B) LC-MS/MS and KIMS. CMIA, Chemiluminescent Microparticle Assay; KIMS, Kinetic Interaction of Microparticles in Solution; LC-MS/MS, Liquid Chromatography coupled with Tandem Mass Spectrometry .

(0.19MB).

In the KIMS assessment, vancomycin concentrations ranged from 7.10 to 70.40 μg/mL with KIMS and from 7.90 to 75.87 μg/mL with LC-MS/MS. KIMS and LC-MS/MS measurements revealed a high correlation, with Spearman rank correlation coefficient value of 0.926 (p < 0.001, 95 % CI 0.875 to 0.957), and no significant deviation of linearity was observed (p = 0.49). Fig. 1B represents the regression analysis and the results are summarized in Table 3. The intercept was not significantly different from 0 μg/mL (95 % CI −1.3759 to 2.689), then no systematic bias was observed. The slope was 0.89 (95 % CI 0.8032 to 0.990), representing a small proportional bias. The mean difference between vancomycin concentrations measured by KIMS and LC-MS/MS was −2.28 μg/mL (−15.34 to 2.75). Bland-Altman plots are shown in Fig. 2B. The difference between both methods was within the ±1.96 standard deviation range. LC-MS/MS measured levels were, on average, 8.81 % higher than KIMS (range: −13.40 % to 35.04 %). The difference between values obtained was within 20 % of the mean for 76 % of the samples (41/54).

Vancomycin TDM for serious MRSA infections in patients undergoing hemodialysis is strongly recommended by the current international consensus.17 Given the narrow vancomycin AUC range for therapeutic efficacy and minimal nephrotoxicity, AUC-target is the most accurate and optimal way to guide vancomycin dosing.517 The AUC0–24/MIC is the PK/PD parameter that best describes the vancomycin antimicrobial activity, and a value of AUC0–24/MIC 400‒600 has been shown to be the PK/PD target. Although Bayesian PK models can accurately estimate the AUC0–24 with a single point in the dose interval, and this is the current guidelines’ recommendations for patients not on renal replacement therapy, this is not validated for patients under dialysis. Therefore, current recommendations is still based on a single point evaluation for patients under replacement renal therapy.17

Given the relevance of serum concentration monitoring in patients undergoing HD, the vancomycin assay's accuracy must be guaranteed. Differences between methods of vancomycin measuring have been reported in the last years: previous studies reported CDP-1 cross-reaction in different immunoassays, which are the most applied assays at routine laboratories.81012,14-16,21-25 In addition to CDP-1 cross-reaction, other interferers may also occur, such as impurities present in commercial vancomycin,26 immunoglobulin M,27 paraproteins, and rheumatoid factor,28 dialyzer flow rate2930 and the lack of technique standardization between immunoassays.10

This study aimed to assess the performance of two different immunoassays for vancomycin measurement compared to LC-MS/MS in patients undergoing HD. As stated earlier, several studies reported the performance of different immunoassays in measuring vancomycin levels. Nevertheless, none of them compared CMIA and KIMS with HPLC-MS/MS specifically in patients receiving HD. For that, 64 samples from 42 patients analyzed by CMIA and 54 samples from 23 patients analyzed by KIMS were included.

Most patients in CMIA and KIMS cohorts were men (67 % and 74 %, respectively), and CRRT was more frequent than IHD in both groups (88 % and 65 %, respectively). Prescribed vancomycin doses every 12 h were more frequent than once a day for CMIA and KIMS cohorts. A large range of concentrations was evaluated by CMIA and KIMS, and measurements were highly correlated with LC-MS/MS, with no deviation of linearity. Both CMIA and KIMS demonstrated acceptable correlation and agreement, although, the slight proportional bias indicates that whereas concentrations increase, differences between immunoassays and LC-MS/MS increase proportionally, but not biased to a clinically significant degree. Therefore, the higher the concentration, the greater the difference between methods. Therefore, although not significant, this accumulation may contribute to the difference between methods. No systematic bias was observed, and the difference between methods was within the ±1.96 standard deviation range. Following ISR, the difference between values was within 20 % of the mean for 67 % of the samples for CMIA and for 76 % for KIMS. However, the range of observed differences (−10.38 to 11.26 and −15.34 to 2.75 for CMIA and KIMS, respectively) seems insufficient to affect clinical decisions on vancomycin dosing.

High-flux membranes present increased surface area and higher permeability, allowing clearance of middle to large molecules, such as vancomycin.2930 Since the chemical structure of CDP-1 and vancomycin are very close, dialytic clearance characteristics must be the same.1430 As CDP-1 tends to be removed by high-permeability dialyzers, its serum levels should not be sufficient to interfere with vancomycin measurements by immunoassays. The possible removal of CDP-1 by high-permeability dialyzers and the use of monoclonal antibodies may explain the performance of both CMIA and KIMS described in this study.

Our study has some limitations that must be acknowledged. First, although LC-MS/MS method has been validated according to the guidelines proposed by the American FDA for validation of bioanalytical methods,20 which includes specificity and sensitivity testing, we do not have a CDP-1 standard to perform analyze the presence of this compound. However, the chromatographic conditions used are similar to those in other studies,3132 hence, we expect that interferences will be unlikely. Second, atenolol-d7 was used as IS since there is no availability of a deuterated analogue of vancomycin on the market. Nonetheless, this has unlikely affected our results, since atenolol-d7 is a compound with a similar retention time and will never be present in patient samples. Its use as an internal standard for vancomycin has already also been described in other studies.32 Third, we did not test the analyte stability in the matrix. However, −20 °C frozen were analyzed in no more than 15 days, and it has been demonstrated that vancomycin remains stable for up to 6-months at −20 °C.33 Finally, although we included a number higher than the minimum recommended of 40 samples, according to CLSI analytical guideline,18 a larger sample size could potentially increase the power to detect small but statistically significant differences.

This is the first evaluation of CMIA and KIMS performance compared to LC-MS/MS in patients undergoing HD. Both CMIA and KIMS showed acceptable correlation and agreement.The small proportional bias noted in our study indicates that whereas concentrations increase, differences between immunoassays and LC-MS/MS increase proportionally. However, the range of observed differences seems not to be of clinical relevance.

Our study suggests that neither CMIA nor KIMS undergo cross-reactivity from CDP-1 for vancomycin measurement in samples obtained from patients undergoing hemodialysis, although CDP-1 concentrations were not measured. The possible removal of this compound by high-permeability dialyzers and the use of monoclonal antibodies may explain the performance of both CMIA and KIMS. In summary, CMIA and KIMS commercial assays provide accurate vancomycin measurements in patients undergoing HD.

The study was approved by Institutional Research Ethics Committees. All patients enrolled in the study, or their guardians agreed to participate by a free and informed consent form. The patient's anonymity was ensured.