The Value of Therapeutic Drug Monitoring of Vancomycin in Treatment in a Tertiary Hospital, Cho Ray Hospital, Viet Nam

Aims: To evaluate the value of therapeutic drug monitoring (TDM) of vancomycin in clinical practice. Methods: To review retrospectively on 292 hospital cases (111 females, 181 males) treated with vancomycin, July to October 2014, at ChoRay Hospital. The main evaluating parameters were TDM criteria for vancomycin (dose, dosing interval, times of monitoring), trough level, dose adjustment, renal function follow-up, minimum inhibitory concentration of infectious agents, clinical response. Results: Two hundred seventy-five patients (94.2%) received routine dose of 1 g vancomycin per IV infusion time. Dosing interval was given correctly to estimated glomerular filtration rate (eGFR) level 80.8% (235/291). The 1 st monitoring after 9 th dose was in 139 cases (47.6%). Trough level was lower than 10 mg/L in 86 patients (29.5%), higher than 20 mg/L in 96 (32.9%), and 110 in optimal range 10-20 mg/L (37.7%). Age and eGFR were 2 independent predictors for trough level. Original Research Article Hoang et al.; BJMMR, 16(6): 1-16, 2016; Article no.BJMMR.26806 2 Dose adjustment were done in 6.9% (6/86) patients  10 mg/L, 20.8% (20/96) ones >20 mg/L, and 11.8% (13/110) ones 10-20 mg/L. Vancomycin concentrations in young patients were lower than those in elderly ones with OR = 5.9 [95%CI: 2.6 – 14.0], p = 0.0001. Response sensitivity was 69.3% (13/19) for dose reduction, and 83.3% (5/6) for dose increase. Dose adjustment did not make change in trough level compared to unadjusted ones. Nephrotoxicity rate was found as 8.4%. Treatment failure was 50% in patients with trough concentration/minimum inhibition concentration ratio ≤ 10 compared to 15% in ones with higher ratio > 10, p = 0.034. The failure rate was highest in patients received vancomycin  7 days (22/70: 31.4%), OR: 4 (2.0-7.7) p=0.002. The clinical AUC/MIC ratio cut-off, 190 mg/L/day, had 75.9% and 66.7%, respectively for sensitivity and specificity to predict the success result in treatment. Conclusion: The criteria of TDM on vancomycin were not applied strictly, especially for dosing intervals, dosing adjustment and follow-up thereafter. The clinical pharmacodynamics of vancomycin is dependent on both concentration and duration of treatment.


Vancomycin has been used as a first-line antibiotics for treatment of infection caused by methicillin-resistant
Staphylococcus aureus (MRSA) [1]. The target concentration of vancomycin efficiently has become more important due to facing to the rapid increase of MRSA infections. In addition the therapeutic window of trough vancomycin level is narrow, as 10-15 mg/L to avoid development of resistant [2] or 15-20 mg/L for more resistant strains [3], but < 20 mg/L to minimize risk of nephrotoxicity [4]. Thus the therapeutic drug monitoring (TDM) is strongly recommended when using vancomycin. There are many guidelines for vancomycin TDM suggested by several organizations. A systematical review on 635 records related to vancomycin TDM selected out 12 clinical practice guidelines achieving high quality recommendations [5].
The aims of this study were to evaluate the practical value of vancomycin TDM when applying in routine treatment. The specific objectives were the compliance to criteria in vancomycin TDM (dosing intervals, dose adjustment, monitoring frequency on kidney function and vancomycin concentrations), nephrotoxicity of vancomycin, and clinical response (pharmacodynamics of vancomycin) of vancomycin treatment.

PATIENTS AND STUDY METHODS
Cho Ray Hospital is tertiary general hospital, located in Ho Chi Minh City, Viet Nam, having around 49 departments (including 11 in Surgery, 6 in Neurology, 11  The general content of vancomycin TDM guidance in Cho Ray Hospital can be briefly described as the initial/maintenance doses as 1000 mg/60 kg body weight/per IV infusion time; dosing intervals based on estimated glomerular filtration rate (eGRF): Q 8-12h for eGFR ≥ 50 ml/min/1.73 m 2 , Q 24h for eGFR 35 -49.9 eGFR ≥ 50 ml/min/1.73 m 2 , Q 48h for eGFR 25 -34.9 eGFR ≥ 50 ml/min/1.73 m 2 , and Q72h or defining based on daily level of serum vancomycin for eGFR < 25 eGFR ≥ 50 ml/min/1.73 m 2 ; frequency of routine monitoring on a weekly basis at minimum requirement for patients with stable renal function, for cases with unstable renal function, hemodynamically compromised or at risk for nephrotoxicity, monitoring should be repeated at least 1-3 times weekly. The target trough concentrations were 10-15 mg/L for soft and skin tissue infections, abscess, cellulitis with MIC < 1 mg/L, 15-20 mg/L for soft and skin tissue infections, abscess, cellulitis with MIC ≥ 1 mg/L or complicated infections (endocarditis, osteomyelitis, bacteremia, prosthetic joint infection, or pneumonia), and 20-25 mg/L for infections involving central nervous system (bacterial meningitis).
This retrospective study reviewed 292 medical case reports of patients treated with vancomycin and having at least one monitoring measurement of vancomycin concentration during treatment time. All medical case reports were from all clinical wards of hospital from June to September 2014. The main parameters investigated were patient anthropometrics (age, gender, clinical wards, date of admission, date of discharge), final disease diagnosis on discharge, all details relating to vancomycin TDM as treatment dose per IV perfusion time, rate of IV perfusion, dosing interval, blood creatinine and eGFR related to the initial dose given, time of the first monitoring vancomycin concentration (counting by treatment days and by times of drug given from the first dose of vancomycin), repeat frequency on kidney function evaluation (number of times for blood creatinine and eGFR investigation), dose adjustment for vancomycin after the first trough concentration results, repeat frequency on monitoring of trough concentrations in the whole treatment course; MIC results if available; and the patient status on hospital discharge.
The patient status on hospital discharge was classified into 2 groups: success (better or well improved) and failure (unimproved, worsened, or died).

Vancomycin Assay
Vancomycin drug concentrations were analyzed by an in vitro chemiluminescent microparticle immunoassay (CMIA) in human serum (ARCHITECT i1000SR Immunoassay Analyze, Abbott Diagnostics, Chicago, IL, USA), at the Biochemistry department, Cho Ray Hospital. All drug samples were performed within the same day of collection day. The inter-assay coefficients of variation for vancomycin in ARCHITECT i1000SR were 5.0% at 6.1 mg/L, 4.9% at 18.6 mg/L, and 5.1% at 33.1 mg/L. The limit of detection of vancomycin was  2 mg/L.

Study Approval
The study was approved by the Ethical Committee of Cho Ray Hospital on 20-November 2015. All hospital case records were reviewed at the Hospital Files Room belonged to the General Planning Room of Cho Ray Hospital.

Statistical Analysis
Investigating data from hospital case records were transferred to study record form per each study case. All data was analyzed with the SPSS software (PASW Statistics 18). The descriptive data were presented in frequency, percentage, mean, standard of deviation, and range from minimum to maximum value. Chi-square analysis with or without odd ratio estimation were applied for testing significant difference in qualitative data. The t-test was used for comparison quantitative data between two independent variables. Multivariable binary logistic regression analysis was used to define the independent predictor for vancomycin concentration or risk factor for treatment failure from the monovariable risk factors/predictors. All statistical tests were considered as significantly different if p value < 0.05.
For evaluation of the predictors of trough concentrations of vancomycin, the concentrations were investigated in 3 subclasses as:  10 mg/L versus > 10 mg/L;  20 mg/L versus > 20 mg/L, and optimal range [10 -20 mg/L] versus "out of range" [< 10 mg/L or > 20 mg/L].
The trough concentration/MIC ratio of vancomycin was used to study on the concentration-dependent pharmacodynamics of vancomycin.
It was because of the lack of the total area under the concentration curve of vancomycin, AUC/MIC ratio of vancomycin could not be evaluated and investigated against the treatment results in this retrospective study. We suggested an alternative parameter, called as "clinical AUC/MIC ratio" of vancomycin. The clinical AUC/MIC ratio of vancomycin was defined as the multiple of trough concentration/MIC ratio by duration of vancomycin treatment (days). This parameter was investigated as the both concentrationdependent and time-dependent pharmacodynamics of vancomycin against treatment result.
Receiver Operating Characteristic (ROC) Curve Analysis was applied for determination of cut-off value for clinical AUC/MIC ratio for prediction of treatment result (success versus failure).

Evaluation of Clinical Performance of TDM for Vancomycin
Two hundred and ninety-two patient cases were included in retrospective review, in which 111 females (38%  Table 4 showed only 2 independent risk factors for low concentration of vancomycin ≤ 10 mg/L: eGFR affecting in the negative way to have lesser cases with vancomycin ≤ 10 mg/L (eGFR < 60, OR < 1); and age groups affecting in the positive way to have more cases with vancomycin ≤ 10 mg/L (age higher, OR >1), p < 0.05 for both.

Patients with trough vancomycin concentrations in optimal range [> 10 and ≤ 20 mg/L]
Ten hundred and ten patients had vancomycin in optimal range, > 10 and  20 mg/L, in first monitoring time, accounted percentage of 37.7%. The factors affecting to vancomycin concentration, [>10-20 mg/L], was evaluated in Table 7.      Among 19 cases with dose treatment adjusted, 6 patients reached concentrations in range of 10-20 mg/L (6/19: 31.6%) and 13 still be out of target (68.4%)-including 4 cases with low concentrations  10 mg/L and 9 with high ones > 20 mg/L. Thus, the dose adjustment did not give any difference effect compared to group without dose intervention in above (p = 0.95).
The binary logistic regression analysis revealed only vancomycin treatment time ≤ 7 days as independent risk factor for low (poor) incidence rate of success treatment result with OR = 0.1 (95% CI: 0.02 -0.5), p = 0.005 compared to the longer one (Table 9).
There were 66 patients had done the MIC for S. aureus detected in pathological samples. The trough concentration of vancomycin was not different in 4 levels of MIC, but the ratio between concentration over MIC was different at cut-off point as value of 30, p = 0.015 (Table 10).

The Clinical AUC/MIC Ratio (Combination of Time -Dependent and Concentration-Dependent Pharmacodynamics [Killing Action] of Vancomycin)
The clinical AUC/MIC ratio was defined as multiple of trough concentration/MIC ratio with duration of treatment of vancomycin (days). The ROC showed the area under the curve (AUC) of clinical AUC/MIC of 0.727 [0.54 -0.92], p = 0.015. Cut-off value suggested as 190 mg/L/day (Fig. 1). The sensitivity and specificity of this clinical AUC/MIC cut-off value was 75.9% and 66.7%, respectively for the prediction of success result in treatment (Table 11).

Clinical Performance of Vancomycin Monitoring
The initial or maintenance dose of vancomycin is guided as 15-20 mg/kg [11], but clinical routine use as 0.

Pharmacokinetics of Trough Concentration of Vancomycin
The percentage 37.7% (110/292 cases) of patients reaching optimal range of trough concentration, 10-20 mg/L, in the first regimen of vancomycin in this study was equivalent to those revealed in previous prospective studies in Viet Nam [12,13]. Both retrospective study and prospective studies showed 30 -60% of cases needed dose adjustment after the first regimens of vancomycin following to experience or guidance. The American Thoracic Society (ATS) suggested an initial empirical vancomycin dose of 15 mg/kg every 12 hours in adults with normal renal function (equivalent to 1 gram/12 hrs.) [144]. However, based on pharmacokinetics of vancomycin, that empirical dose of vancomycin could not produce trough concentration of 15-20 mg/L as proved by this study. The proportion of predicted optimal vancomycin trough concentration in range 10-20 mg/L during the first 4 days of therapy was 22%, and 77% in lower concentrations (< 10 mg/L) [15]. Pediatric pharmacokinetics of vancomycin in Canadian also reported the proportion of 60% trough concentrations fell within therapeutic range 5-10 mg/L up to the fifth day of vancomycin treatment [16].
A recent retrospective pharmacokinetic analysis of serum levels of vancomycin determined during routine monitoring of 46 patients in a medical ICU concluded that standard dosages of vancomycin led to a 33% risk of not achieving the recommended area under the concentrationtime curve over 24 h/MIC breakpoint for S.
aureus [17]. The dose of 15 mg/kg of vancomycin given every 8 hours in children gave around 60% of cases reaching concentration of 5-15 mg/L [18]. In conclusion, the first vancomycin treatment doses, based on experience or monogram-based guidance, did not get the trough concentrations in target range, thus the dose adjustment must be considered as important point for any cases treated with vancomycin.
Up to now, it seems no data indicating that achieving the optimal concentration is easy and safe by the dose adjustment. The increase or decrease of dose in adjustment caused the proportional response in drug concentration. This study showed that comparing to the first concentrations, sensitivity of vancomycin concentration decreased as 69.3% with dose reduction, and 83.3% with dose increase, p = 0.03. Although having the parallel response of concentrations to the dose adjustment, but it could not be enough to bring concentrations to the target of optimal range. This study showed the failure of reaching to the optimal range of vancomycin concentrations, 10-20 mg/L, in 19 cases with dose treatment adjusted, 6 patients reached concentrations in range of 10-20 mg/L (31.6%) and 13 still be out of target (68.4%). Comparing to 37 cases without dose adjustment but receiving the 2 nd concentration monitoring, 11 cases in optimal range, 10-20 mg/L, (29.7%) and 26 others out of range concentration (70.3%). Thus, the dose adjustment did not give any positive effect (p = 0.95). Thomson et al. [15], using population pharmacokinetic model suggested the new high dose guidance of vancomycin to get 55% chance having optimal concentrations, as 1250 and 1500 mg per 12 hours for renal clearance 90-110 ml/min and > 110 ml/min, respectively. It means the higher dose for higher renal clearance. This study revealed the odd ratio for having high concentration ≥ 20 mg/L in patients with eGFR < 60 was 3.3 (95% CI: 1.9 -5.6), p = 0.0001. The odd ratio (OR) for low vancomycin concentrations in young patients compared to elderly ones was 5.9 [95% CI: 2.6 -14.0], p = 0.0001. In conclusion, the talent in management of dose adjustment is a challenge of good combination of all factors such as population pharmacokinetic model, eGFR level, age of patients.
In this study, the independent predictors affecting to vancomcycin had been evaluated in details by classification of concentrations into 3 subclass analysis. The first subclass comparison between concentrations ≤ 10 mg/L versus > 10 mg/L 2 independent factors were found as eGFR < 60 ml/min/1.73 m 2 (p=0.001) and older age >40 yrs. (p=0.01) in prefer to have the high vancomycin concentrations > 10 mg/dL. The second one between concentrations ≤ 20 mg/L versus > 20 mg/L 2 independent factors were the same as eGFR <60 ml/min/1.73 m 2 (p=0.0001) and elderly age > 60 yr. (p = 0.017). And the third one between optimal concentrations, > 10 -≤ 20 mg/L, versus"out of range" concentrations, ≤ 10 or >20 mg/L, only one factor was age affecting to vancomycin concentration. The odd ratio (OR) of capacity of "out of range" concentration (≤ 10 or > 20 mg/L) in young patients ( 40 yr.) compared to elderly ones (> 40 yr.) was 2.5 [95%CI: 1.3 -4.8], p = 0.0009. In conclusion, the main factors affecting to trough concentration of vancomcyin were eGFR < 60 ml/min/1.73 m2, older age > 40 yr., elderly age > 60 yrs. These factors should be considered in both initial doses and adjusting doses of vancomycin. This finding was in agreement with the population pharmacokinetic model for initial higher doses (1250-1500 mg vancomycin) for patients with high renal clearance > 90 ml/min suggested by Thomson et al. [15]. Thus, in clinical treatment, using high initial doses of vancomycin is needed for young patients with normal kidney function.

Vancomycin and Nephrotoxicity
Role of vancomycin monitoring in preventing nephrotoxicity was guided in ASHP Therapeutic Position Statements 2008 [2]. The frequency of monitoring may be more to patients targeting to produce sustained trough concentrations of 15-20 mg/L or higher, or who are at risk of toxicity, such as receiving concurrent nephrotoxins. In this study, 151 cases having trough concentrations > 15 mg/L, but only 27 cases received 2 nd monitoring.

Clinical Pharmacodynamics of Vancomycin
There was no difference in trough concentration of vancomycin in both 1 st and 2 nd monitorings between success cases and failure cases in this study. Feffres MN et al. in a retrospective study on 102 patients with health-care-associated MRSA pneumonia also found no significant differences between survivors and non-survivors in terms of trough serum vancomycin concentrations and AUCs [26]. The ratio between trough concentration of vancomycin over MIC of 10 times seemed to be a trend for prediction of treatment result in this study with OR of 5.7, but unfortunately not reaching to significant difference (p = 0.06). No relationships were found between peak concentrations, trough concentrations, or pharmacodynamic parameters (e.g., peak/MIC, time above the MIC, or AUC/MIC) and organism eradication or overall patient outcome [25].
The concept of concentration-dependent killing is simplified with pharmacodynamics parameter as peak/MIC ratio ≥ 10 times for aminoglycosides and fluoroquinolones [28,29]. The parameter for optimal response for time-dependent killing antibiotics (beta-lactams, clindamycin, linezolid) as the time of drug concentrations above MIC is ≥ 50% of the dosing interval [28,29]. This study showed that the trough concentration of vancomycin/MIC ratio ≥ 10 times was accounting 91% (60/66 cases) and 100% of trough vancomycin concentrations were higher than MIC over the whole treatment time. Both these parameters showed that vancomycin has combining action on both concentrationdependent killing and time-dependent killing antibiotics as suggested from previous papers [28,[30][31][32]. The duration of bacteremia was 2 days in nafcillin group and 5 days in vancomycin group [33]. In patients received vancomycin for endocarditis, the duration of bacteremia was even more prolonged with mean of 7 days [34]. The apparently slower bacterial killing action of vancomycin than previous studies has been recognized [30]. Bacteremia due to heterogeneous vancomycin-intermediate S. aureus (hVISA) was reported so long as 39 days [35].
The multivariable binary logistic regression analysis revealed only the duration of treatment of vancomycin was the only independent predictor for treatment success (OR = 0.1 [95% CI: 0.02 -0.5], p = 0.005). This OR of 0.1 showed that the chance for treatment success in patients treated ≤ 7 days was 10 times lesser than whom who had the longer treatment (≥ 8 days).
The theoretical value of AUC/MIC ratio was not easily applied in clinical treatment because of the lack of peak concentration of vancomycin. Thus, we proposed a similar parameter called as "clinical AUC/MIC ratio". The clinical AUC/MIC ratio was defined as multiple of trough concentration/MIC ratio with duration of treatment of vancomycin (days). The ROC showed the area under the curve (AUC) of clinical AUC/MIC ratio of 0.727 [0.54 -0.92], p = 0.015. Cut-off value suggested as 190 mg/L/day (Fig. 1). The sensitivity and specificity of this clinical AUC/MIC ratio cut-off was 75.9% and 66.7%, respectively for the prediction of success result in treatment (Table 11).
The clinical AUC/MIC ratio was suggested in this study with aiming to find out some clinical parameter for prediction the outcome of treatment. This finding was in agreement with hypothesis that vancomycin acts by both parameters: concentration-dependent and timedependent actions [30,28,31,32]. This result will encourage clinical doctors in trying to prolong more treatment days of vancomycin as good as possible, may be more than 10 days, before saying about the failure of this luxury drug. The properly use of vancomycin according to both pharmacokinetics and pharmacodynamics will be supported by result of this study. The widespread resistance of S. aureus against vancomcycin has continued by times. The Centers for Disease Control and Prevention (CDC) in May-2015 announced the criteria for classification of vancomycin-intermediate S. aureus (VISA) with minimum inhibitory concentration (MIC) for vancomycin of 4-8 µg/ml, and vancomycinresistant S. aureus (VRSA) with vancomycin MIC ≥16 µg/ml [36].

CONCLUSION
The therapeutic drug monitoring of vancomycin has been applied in routine clinical treatment. However, the guidance criteria of this TDM are not followed strictly, especially for dosing intervals, dosing adjustment and follow-up after the first monitoring. The changes in dose adjustment affected directly to drug concentrations afterwards, but not yet helping to reach into the target of optimal range, 10 -20 mg/L. This is still a big challenge requirement the deep understanding of physicians about the pharmacokinetics and pharmacodynamics of vancomycin. The predictors for vancomycin concentrations were eGRF, age. Nephrotoxicity rate was found as 8.4%, often associated with high trough vancomcycin concentration > 30 mg/L. The clinical AUC/MIC ratio of 190 mg/L/day was the cut-off value for the prediction of success result in treatment with sensitivity and specificity of 75.9% and 66.7%, respectively.

CONSENT
It is not applicable.

COMPETING INTERESTS
Authors have declared that no competing interests exist.