Comparative effectiveness of achieving pharmacokinetic and pharmacodynamic parameters and safety of two vancomycin dosing regimens in patients with orthopaedic infection

Introduction

Bone and joint infections, including chronic osteomyelitis, periprosthetic and implant‑associated infections, represent one of the most challenging problems in modern traumatology and orthopaedics. Their successful management requires a combination of radical surgical debridement and prolonged systemic antimicrobial therapy that takes into account the aetiology of the infection and ensures effective drug concentrations in biofilms and bone tissue [1]. The glycopeptide antibiotic vancomycin has for decades maintained a stable position in the treatment of infections caused by methicillin‑resistant staphylococci, and is also widely used empirically when coagulase‑negative staphylococci – the dominant pathogens in periprosthetic complications – are suspected [2, 3].

According to current concepts, the key pharmacokinetic/pharmacodynamic (PK/PD) parameter predicting the clinical efficacy of vancomycin is the ratio of the 24‑hour area under the concentration‑time curve to the minimum inhibitory concentration (AUC₂₄/MIC). Modern clinical guidelines, based on the 2020 consensus, position the target AUC₂₄/MIC range of 400–600 h·µg/mL as the optimal balance between bactericidal activity and the risk of nephrotoxicity [4, 5]. At the same time, the traditional surrogate marker – trough concentration within 15–20 µg/mL – has been recognised as a less accurate predictor of outcome, prompting a shift towards routine AUC calculation in clinical practice [6].

However, despite the universal nature of these recommendations, their extrapolation to specific patient populations is fraught with several uncertainties. First, the target AUC₂₄/MIC corridor of 400–600 h·µg/mL has been predominantly validated for bloodstream infections caused by S. aureus with MIC ≤1 µg/mL [7]. For coagulase‑negative staphylococci, in particular S. epidermidis, for which EUCAST has set a susceptibility breakpoint of 4 µg/mL, the PK/PD target required for eradication has been insufficiently studied [8]. Moreover, attempts to achieve the target AUC₂₄/MIC at high MIC values (≥2 µg/mL) inevitably require dose escalation, which significantly increases the risk of vancomycin‑induced nephropathy [9].

Second, the paradox of modern pharmacotherapy lies in the fact that the registered instructions for vancomycin preparations in the Russian Federation differ substantially in their dosing approaches. Most of them limit the maximum daily dose to 2 grams, which, under the standard regimen (1 g every 12 hours), often fails to achieve target AUC₂₄/MIC values in patients with preserved renal function. At the same time, alternative regimens that allow increased maximum single and daily doses are specified in the instructions of only two registered products, and their comparative efficacy and safety remain a matter of debate.

Patients with bone and joint infections represent a particular clinical cohort. Surgical stress response, intraoperative blood loss, haemodilution, and hypoproteinaemia can significantly modify the volume of distribution and clearance of vancomycin, while the pursuit of high vancomycin doses to achieve bactericidal concentrations in bone tissue increases the risk of acute nephropathy [2, 10, 11]. Thus, for a patient with orthopaedic infection, the contradiction lies between the probability of subtherapeutic concentrations at the infection site and the risk of vancomycin‑induced nephropathy. The absence of direct comparative studies of the 750 mg every 8 hours versus 1000 mg every 12 hours regimens does not allow an evidence‑based choice between the strategies of “more frequent” or “higher” doses.

Objective – to compare the effectiveness of achieving the target area under the pharmacokinetic curve and the safety of two different vancomycin dosing regimens in patients with bone and joint infections.

Materials and methods

This prospective observational study included 20 patients. Inclusion criteria were: female sex, age 35 to 65 years, presence of indications for vancomycin administration, planned debridement surgery for chronic osteomyelitis and/or implant‑associated infection and/or septic arthritis, infection localisation in long bones or large joints of the upper or lower extremity.

Exclusion criteria were factors that could significantly affect pharmacokinetic characteristics: body weight less than 50 kg or more than 100 kg, limb amputation, reduced renal excretory function (serum creatinine above 80 µmol/L), presence of any diagnosed urinary tract disease, administration of drugs or combinations with high nephrotoxic potential, moderate anaemia (haemoglobin below 90 g/L), severe hypoproteinaemia (total protein below 60 g/L), severe comorbid disease or clinically unstable condition – sepsis, septic shock, decompensation of any chronic diseases.

According to standard local protocols adopted at the clinic, all patients perioperatively received empiric intravenous antimicrobial therapy with vancomycin and cefoperazone + [sulbactam] routinely. Intravenous vancomycin infusion was started intraoperatively after collection of tissue biopsies for microbiological examination. Patients were divided into two groups of 10 persons each according to the vancomycin dosing regimen. In Group 1, the loading dose of vancomycin (“Vankoter AF”, powder 750 mg, manufacturer “AlPharma”, RF) was 1500 mg, maintenance – 750 mg every 8 hours; patients received the drug for 7 days. In Group 2, the loading dose of vancomycin (“Vancomycin”, lyophilisate 1000 mg, manufacturer “Belmedpreparaty”, RB) was 2000 mg, maintenance – 1000 mg every 12 hours, also for 7 days. The loading dose infusion was administered over 90–120 minutes; maintenance infusions lasted 60 minutes.

In addition, according to local protocols, patients received analgesics (tramadol, paracetamol, ketoprofen), prophylaxis of venous thromboembolic complications (dalteparin sodium, dabigatran etexilate), and prophylaxis of gastropathy and gastrointestinal bleeding (omeprazole).

On day 3 of therapy, serum vancomycin levels were measured. Vancomycin concentrations in serum were determined using a Shimadzu Prominence‑I LC‑2030C liquid chromatograph (Japan) by high‑performance liquid chromatography with photometric detection in gradient elution mode. Blood was drawn for the peak vancomycin concentration (V1) – one hour after the end of infusion, and for the trough concentration (V2) – immediately before the infusion of the next antibiotic dose. The efficacy criterion was achievement of the target 24‑hour area under the pharmacokinetic curve (AUC₂₄) within 400–600 h·µg/mL. AUC₂₄ values below 400 h·µg/mL were considered insufficient, while those exceeding 600 h·µg/mL were regarded as excessive.

The AUC₂₄ (24‑hour area under the pharmacokinetic curve) was calculated under the assumption that vancomycin clearance follows first‑order kinetics (Fig. 1) [12].

Fig. 1. Graphical representation of the method for calculating the area under the pharmacokinetic curve (AUC) according to Deng Y et al. 2023

The first‑order elimination rate constant k (≈ 1/h) was calculated using the formula k = –ln(V2/V1)/Δt; then the elimination half‑life was calculated as t₁/₂ = –ln(0.5)/k. For the AUC calculation, the following data were used: infusion duration, time interval between infusions, V1 and V2 values, and the number of infusions per day. The vancomycin concentration V0 (peak concentration at the end of infusion) was determined by extrapolating V1 over time using the estimated k constant. The total AUC for a single dosing cycle was calculated as the sum of two areas, a and b, where area a = (V3+V0)/number of infusions × infusion duration (h); area b = (V0–V3)/k. In our case, the concentration V3 corresponding to the start of the next dose infusion coincided with V2. AUC₂₄ was defined as the sum of AUCs over the number of doses received by the patient per day; thus, in Group 1, AUC₂₄ = 3 × AUC, and in Group 2, AUC₂₄ = 2 × AUC.

Safety was assessed by the absence of cutaneous and post‑injection adverse reactions, decreased diuresis, laboratory signs of vancomycin nephrotoxicity and haematological toxicity. Complete blood count and biochemical blood tests were performed preoperatively and on day 7 after surgery. The observation period was 7 days (until the end of the initial empiric antibiotic course), but in one patient it was extended to 13 days (until discharge) due to the development of an adverse drug reaction.

Patient primary data were recorded in an MS Excel spreadsheet. Statistical analysis was performed using IBM SPSS Statistics version 27. Normality of data distribution was tested with the Shapiro–Wilk test. Since not all measured variables followed a normal distribution, the median (Me) and interquartile range (Q1–Q3) were used as measures of central tendency; statistical significance was assessed using the Mann–Whitney U test and Pearson’s χ² test, with p <0.05 considered statistically significant. Correlation was evaluated using Spearman’s rank correlation coefficient (r), and the strength of correlation was interpreted according to the Chaddock scale.

Results

At baseline, the patients in both groups were comparable in anthropometric characteristics. Although the median age in Group 1 was somewhat lower than in Group 2: 46 (38–61) years and 62 (50–63) years, the difference was not statistically significant (p=0.22). Height was 163 (160–170) cm and 164 (162–167) cm, and body weight was 70 (60–85) kg and 71 (63–77) kg in the first and second groups, respectively.

C‑reactive protein, a dynamic and sensitive marker of inflammatory activity and one of the possible causes of acute kidney injury [13–15], was within the range typical for patients with chronic orthopaedic infection without systemic spread; intergroup differences were not statistically significant. Among the main blood cell counts and biochemical parameters, statistically significant differences were found only for platelet count and total serum protein, but these values were within reference intervals and could not substantially affect drug pharmacokinetics. Serum creatinine and creatinine clearance at inclusion were comparable between the groups (Table 1). Operation duration and blood loss volume also did not differ significantly between Group 1 and Group 2, amounting to 170 (135–208) min vs. 127 (75–160) min (p=0.145) and 500 (400–700) mL vs. 450 (300–600) mL (p=0.529), respectively.

Intergroup comparison of pharmacokinetic parameters revealed a statistically significant difference in trough vancomycin concentration (Table 2). The minimum AUC₂₄ value was 237 h·µg/mL. The maximum AUC₂₄ – 1028 h·µg/mL – was recorded in the single patient who developed nephropathy. In addition, two other patients had near‑maximal AUC₂₄ values (927 and 1023 h·µg/mL), but showed no signs of nephropathy.

Table 1. Preoperative laboratory parameters in the study groups

Notes: CRP – C‑reactive protein; * – statistically significant differences.

Table 2. Pharmacokinetic parameters in the study groups

Note: * – statistically significant differences.

On day 7 after surgery and after initiation of antibiotic therapy, no statistically significant differences were observed between the groups in the laboratory parameters studied (Table 3). Complete blood count showed no signs of drug‑induced haematological toxicity. The detected anaemia of mild to moderate severity, as well as hypoproteinaemia, were interpreted as post‑haemorrhagic changes. A slight increase in C‑reactive protein compared with preoperative levels, with normal leucocyte count, was attributed to a non‑specific inflammatory response to surgical trauma. Creatinine and its clearance did not show significant changes in the study groups.

Table 3. Postoperative laboratory parameters in the study groups

Note: CRP – C‑reactive protein.

On day 7 after surgery, a statistically significant (p=0.042) moderate correlation (r = 0.458) was found between vancomycin trough concentration and serum creatinine level. No correlation was found between creatinine clearance and vancomycin serum concentrations or AUC₂₄, which may be due to the small number of observations in the groups.

In one case in Group 1, nephropathy presumably associated with vancomycin was diagnosed. The adverse reaction developed in a 51‑year‑old female patient, body weight 86 kg, height 174 cm, admitted for debridement surgery of the lower limb. The patient had no urinary tract disease. Notably, the duration of surgery was prolonged (330 minutes) with substantial blood loss (1200 mL). On day 3, V1 and V2 levels were 56.2 µg/mL and 30.6 µg/mL, respectively, and AUC₂₄ was 1028 h·µg/mL. Urgent blood tests revealed moderate anaemia – haemoglobin 81 g/L, hypoproteinaemia – total protein 63 g/L, serum creatinine 147 µmol/L, creatinine clearance decreased to 55 mL/min, i.e. a decline in creatinine clearance of more than 50 % from baseline, which meets current criteria for vancomycin‑induced acute kidney injury [16]. Vancomycin therapy was discontinued; antibiotic therapy was adjusted according to microbiological results of intraoperative samples; symptomatic treatment was administered. The patient was discharged on day 13 after surgery with positive dynamics.

Discussion

Our study did not show statistically significant differences either in the median AUC₂₄ or in the frequency of achieving the target range between the study groups. However, exceeding the target range was more common than achieving target values: in Group 1 by a factor of 3.5, in Group 2 by a factor of 1.5. A similar trend towards exceeding the target pharmacokinetic level when attempting to intensify the dosing regimen has been previously described by us [2, 6].

Failure to achieve target AUC₂₄ was observed in 10 % and 20 % of patients in Groups 1 and 2, respectively. This is substantially lower than if we had used trough serum vancomycin concentration as the target marker; in that case, 60 % of patients in Group 1 and 100 % in Group 2 would not have reached the target.

In our study, a statistically significant relationship between vancomycin pharmacokinetic parameters and renal function indicators was found only between creatinine level and trough vancomycin concentration, which may be related to the small sample size and homogeneity of the compared groups. This is consistent with earlier studies that emphasised the greater importance of vancomycin trough concentrations as predictors of acute kidney injury [17]. More recent studies provide arguments for greater reliability of vancomycin dose adjustment based on the area under the pharmacokinetic curve [16, 18].

Measured serum concentrations and/or calculation of the area under the pharmacokinetic curve, as well as its ratio to MIC, appear to be accessible and relatively simple methods for selecting the optimal dosing regimen and monitoring its efficacy and safety. It is assumed that AUC‑based dosing of vancomycin may more effectively prevent vancomycin‑induced nephropathy [19].

To optimise dosing, the use of the 24‑hour area under the concentration‑time curve to MIC ratio (AUC₂₄/MIC) is also recommended, as it predicts vancomycin efficacy with the least likelihood of nephrological and haematological toxicity. In a study on the treatment of methicillin‑resistant S. aureus (MRSA) bacteraemia, an AUC₂₄/MIC ratio of 400 h·µg/mL or higher was associated with the greatest clinical efficacy. The upper threshold beyond which the risk of nephrotoxicity significantly increases is considered to be 600 h·µg/mL. These values are usually achievable provided that the MIC does not exceed 1 µg/mL [20]. It is also known that the minimum level of this parameter for vancomycin efficacy against infections caused by other pathogens may be lower: for coagulase‑negative staphylococci – at least 300 h·µg/mL; for enterococcal bacteraemia, the target AUC₂₄/MIC should be no less than 389 h·µg/mL [21, 22]. However, to date, no data demonstrating an outcome dependence on the AUC₂₄/MIC ratio have been obtained for coagulase‑negative staphylococci [23]. Thus, the need for dose adjustment to achieve the required AUC₂₄/MIC ratio relies largely on evidence from studies of specific bloodstream infections. Moreover, the MIC value itself can vary considerably (1.5–2‑fold) depending on the method used – e‑test or broth microdilution [16]. It should be noted that in many cases it is quite problematic to demonstrate the significance of high vancomycin concentrations in improving clinical outcomes [24].

Meanwhile, the widespread use of vancomycin for the treatment of orthopaedic infections, such as osteomyelitis, periprosthetic and implant‑associated infections, encourages persistent searches for therapeutic strategies and objective dose‑adjustment parameters that allow better antibiotic penetration into bone tissue at the infection site and achievement of an effective AUC/MIC ratio. It should be borne in mind that drug penetration into bone is approximately 40 % of the serum level, but this figure can vary greatly depending on vascularisation and bone type (cancellous or cortical) [25, 26]. There is evidence that vancomycin penetration into the peri‑implant area and the osteomyelitis focus is lower than into intact bone tissue [27], which may necessitate maintaining plasma vancomycin concentrations or AUC at the upper end of the therapeutic range, including through intensified dosing.

Currently, for most vancomycin preparations registered in the Russian Federation, very narrow dosing limits are established, with a maximum daily dose of 2 grams. Nevertheless, the official package insert of one product (RU No. LSR‑000126/09) contains instructions on the possibility of using a daily dose of up to 3 grams for bacterial meningitis in adults, and another (RU No. LP‑No. (003043) – (RG‑RU)) bases dosing recommendations on patient body weight and serum concentration monitoring, limiting only the maximum single dose to 2 grams.

There is no absolutely “safe” concentration corridor for vancomycin in which the probability of toxicity is completely absent. However, an AUC₂₄ above 600 h·µg/mL or a trough concentration above 15 µg/mL are associated with an increased risk of nephrotoxicity [28, 29]. Based on these benchmarks, patients in Group 2 had a lower risk of nephropathy.

Considering the case of nephropathy in one of the patients included in our study, attention should be paid to the presence of an additional risk factor for acute kidney injury – significant blood loss, which can adversely affect renal excretory function [10, 30, 31]. However, in another case with massive intraoperative blood loss (3100 mL), no nephropathy developed, nor in two other cases where AUC₂₄/MIC was 926 and 1023 h·µg/mL. We believe that, apart from pre‑existing renal impairment, other risk factors cannot be obligate predictors of acute nephropathy. At the same time, their presence, and especially their combination, should increase clinician vigilance and determine the strategy for dynamic monitoring during treatment.

Various methods for assessing renal function and calculating vancomycin doses have been proposed. In particular, the feasibility of using new pharmacokinetic models in clinical practice has been demonstrated [32]. Since the calculation of the 24‑hour area under the pharmacokinetic curve can be performed using different methods, their results may differ and affect the AUC₂₄/MIC ratio [33, 34]. The present study has a number of limitations that should be considered when interpreting the results. The small sample size (only 20 patients) limits statistical power and does not allow definitive conclusions about the presence or absence of differences between the groups. The inclusion of only women with narrow selection criteria (age 35–65 years, normal body weight, preserved renal function, absence of severe comorbidity) reduces the generalisability of the results to men, elderly patients, obese individuals, or those with impaired excretory function. AUC₂₄ was calculated using a simplified two‑point method rather than full logarithmic or Bayesian modelling, which may introduce systematic error. Different manufacturers’ products (“Vankoter AF” and “Vancomycin”) were used in the groups, adding additional uncertainty when comparing regimens. Finally, patients received concomitant nephrotoxic drugs (NSAIDs, cephalosporins), which does not allow unequivocally separating the effect of vancomycin itself on the development of adverse events.

Conclusion

Both vancomycin dosing regimens examined showed comparable efficacy in achieving the target area under the pharmacokinetic curve. At the same time, the more intensive vancomycin regimen is characterised by an increased risk of nephrotoxicity. The target AUC₂₄/MIC ratio cannot be strictly standardised because this calculated parameter depends on the pathogen species, the vancomycin minimum inhibitory concentration and the method of its determination, the method used to calculate the area under the pharmacokinetic curve, and the site of infection.

Measurement of trough serum vancomycin concentration may in many situations be sufficient for predictive assessment of nephrotoxicity risk. Calculation of the AUC₂₄/MIC ratio based on two serum concentrations in adult patients is advisable in older age groups, in patients with deviations in body weight and/or height from population averages, and also when long‑term therapy is planned in patients with a known pathogen and vancomycin MIC.

Given the limitations of the study, the data obtained should be considered preliminary and require confirmation in larger and more diverse patient cohorts.