Vancomycin in peritoneal dialysis: Clinical pharmacology considerations in therapy

Intraperitoneal vancomycin is the first-line therapy in the management of peritoneal dialysis (PD)-related peritonitis. However, due to the paucity of data, vancomycin dosing for peritonitis in patients on automated peritoneal dialysis (APD) is empiric and based on clinical experience rather than evidence. Studies in continuous ambulatory peritoneal dialysis (CAPD) patients have been used to provide guidelines for dosing and are often extrapolated for APD use, but it is unclear whether this is appropriate. This review summarizes the available pharmacokinetic data used to inform optimal dosing in patients on CAPD or APD. The determinants of vancomycin disposition and pharmacodynamic effects are critically summarized, knowledge gaps explored, and a vancomycin dosing algorithm in PD patients is proposed.


Authors Authors
Vancomycin is often selected as empiric first line therapy for suspected  in peritoneal dialysis (PD) related peritonitis. However, data on vancomycin dosing in various PD 75 modalities are limited, especially for automated peritoneal dialysis (APD). The paucity of well-designed 76 pharmacokinetic studies has led to vancomycin dosing guidelines for PD patients that are based on 77 limited information resulting in the possibility of achieving sub-or supra-therapeutic trough 78 concentrations in this special patient population.(1) 79 80

PRINCIPLES OF VANCOMYCIN THERAPY 81
Vancomycin is a tricyclic glycopeptide antibiotic with broad spectrum activity against Gram-82 positive bacteria. It is effective for the treatment of Gram-positive infections including peritonitis and is 83 the drug of choice for methicillin-resistant Staphylococcus aureus (MRSA). Vancomycin is poorly 84 absorbed following oral administration. Therefore, it is commonly administered as an intravenous 85 infusion, except in peritoneal dialysis where the route is preferentially intraperitoneal. Approximately 86 50% of vancomycin is protein-bound in plasma with a variable volume of distribution ranging between 87 0.4-1 L/kg in the non-PD population.(2, 3) An initial distribution half-life ranging from 30 minutes to 1 88 hour followed by a mean terminal elimination half-life ranging from 6-12 hours were determined 89 following intravenous dosing in patients with normal renal function.(3) Metabolism is negligible and 90 elimination occurs primarily through glomerular filtration, such that advanced renal disease substantially 91 reduces the clearance of vancomycin resulting in an elimination half-life of about 7.5 days compared to 92 4-6 hours in normal patients. This means that in patients with kidney failure, the dosing of vancomycin 93 must be adjusted.(4, 5) 94 The Clinical and Laboratory Standards Institute (CLSI)  during PD for absorption into the plasma. Based upon a single dose study of six non-infected subjects on 130 PD, vancomycin has a lower dialysate to plasma ratio than urea and creatinine at two hours.(11) There 131 is no correlation between vancomycin PD clearance and dialysis adequacy (Kt/V) following an 132 intravenous dose in patients on APD.(12) 133 Teicoplanin, a glycopeptide antibiotic with a similar molecular structure (1,564 g/mol) and 134 spectrum of activity to vancomycin, was studied in non-infected adults on continuous ambulatory 135 peritoneal dialysis (CAPD).(13) The absolute bioavailability (Fip) was calculated using dialysate drug 136 concentration (corrected for amount remaining in the cavity) and drug amount sampled, which was then 137 plotted against a total dwell time of five hours. Teicoplanin systemic bioavailability, reflecting transfer 138 from the peritoneal space, was directly related to dwell time. Furthermore, the consistency in 139 absorption increased with time suggesting that complete and less variable bioavailability with 140 teicoplanin can be achieved with longer dwell times. Vancomycin possess the desired physiochemical properties as a drug candidate for 164 intraperitoneal administration in APD patients. In addition, with its well-established stability in PD fluids, 165 bioavailability is adequate as long as sufficient dwelling time is allowed for drug absorption. However, 166 the appropriate duration of the dwell time has not been well studied. Hence, it is crucial to monitor 167 vancomycin levels frequently to adjust dosing to get therapeutic concentrations in each individual 168 patient. 169

VANCOMYCIN CLEARANCE DURING PERITONEAL DIALYSIS 171
Vancomycin elimination following an intraperitoneal dose is governed by its total body 172 clearance. Total body clearance is the sum of clearances contributed from elimination organs, mainly 173 kidneys, in the case of vancomycin, and is defined as the volume of plasma cleared of vancomycin per 174 time unit. Elimination processes in PD patients include those originating from residual kidney function 175 (RKF), other non-renal sources plus the drug cleared through PD. Total body clearance is especially 176 important as it controls the overall exposure of vancomycin for the given bioavailability achieved from a 177 dwell. Dialytic clearance is defined as the volume of plasma that has been cleared of vancomycin (i.e. 178 removed from systemic circulation into the peritoneal space) by PD per unit time. Figure  Continuous ambulatory peritoneal dialysis typically employs short dwell times (4-6 hours), which 187 may not be sufficient to reach equilibration between the dialysate and plasma. Studies in non-infected 188 adult CAPD patients report dialytic clearances ranging between 1.2-2.4 mL/min, which account for 20-189 25% of the total plasma clearance. (15,22,23) In patients with peritonitis, vancomycin dialytic clearance 190 increases to 3.8 mL/min following a less-than five-hour exchange. (24)  To date, there has only been one study exploring intravenous vancomycin disposition in subjects 206 on APD.(12) The primary objective was to characterize vancomycin pharmacokinetic parameters in 207 adults without peritonitis after a single intravenous dose. Following the intravenous administration of 15 208 mg/kg, subjects received three cycle treatments over the course of eight hours followed by two 8-hour 209 off-cycler dwells for a total of 24 hours. A 2-liter 2.5% dextrose dialysate prescription was used during 210 and off-cycler dwell. The plasma half-life was 11.6 hours following an on-cycler exchange consisting of 211 three 2-hour dwells. When the same patients were removed from the cycler and allowed to dwell for 7-212 8 hours, the plasma half-life increased to 62.8 hours. Although vancomycin was not dosed 213 intraperitoneally in this study, rapid decline in the plasma half-life support the contribution of APD in the 214 removal of drug. Clearance values did not largely differ from those on CAPD. Approximately 30% of 215 vancomycin was removed by APD relative to the total plasma clearance, which is close to the proportion 216 reported in patients on CAPD. Although intraperitoneal vancomycin administration is recommended by 217 guidelines in patients with PD peritonitis, this intravenous administration study provides a valuable 218 insight towards drug clearance during APD.(29) It should be noted that intravenous administration of 219 vancomycin may not be adequate to achieve effective antibacterial concentrations in the 220 peritoneum. (30)  221 The current International Society for Peritoneal Dialysis (ISPD) guideline recommends 222 supplemental dosing in order to achieve plasma vancomycin troughs above 15 mg/L when administered 223 intermittently. Alternatively, temporarily switching to CAPD is another option for APD patients who 224 develop peritonitis, but is not always feasible. In patients on APD, leveraging the long dwell to 225 appreciate optimal vancomycin transfer is appropriate to ensure adequate time to achieve and sustain 226 therapeutic levels. 227

IMPACT OF RESIDUAL KIDNEY FUNCTION (RKF) AND TREATMENT OUTCOME 229
Residual kidney function in PD patients will have a profound effect for hydrophilic drugs 230 removed exclusively through renal filtration. Enhanced drug clearance from RKF may have implications 231 to treatment outcomes in patients with PD-related peritonitis. Therefore, patients with greater RKF may 232 require higher or more frequent antibiotic dosing. 233 The importance of RKF on the outcome of PD-related peritonitis in patients treated with 234 antibiotics has been discussed for more than ten years, but the data describing this relationship are still 235 scarce and controversial. The ISPD 2010 update on PD-related infections has previously recommended a 236 25% increase in antibiotic dose in patients with a daily urine output of over 100 mL.(31) This 237 recommendation has been removed in the updated 2016 guideline, which reflects the lack of evidence 238 to support this empiric recommendation.(29) In a retrospective study examining the impact of RKF on 239 vancomycin concentrations, the influence of RKF was found to not have a significant impact. (32)  240 Vancomycin concentrations appeared lower in patients who were non-anuric across both modalities 241 even though a 25% higher dose was administered to those with RKF. This however was concluded to not 242 be statistically significant. Similar results have been published showing no difference in treatment 243 outcomes in non-anuric and anuric patients treated with cefazolin and gentamicin. (33)  244 In contrast, a recent study investigating the relationship between RKF and PD-related peritonitis 245 treatment outcomes was able to explain treatment failures related to the remaining degree of renal 246 function.(34) Treatment failure in those with Gram-positive and culture-negative peritonitis were found 247 to be significantly higher for patients with a urinary creatinine clearance greater than 0-5 mL/min 248 compared to those who were anuric. Significantly higher relapse and recurrence were observed in those 249 patients with Gram-positive or culture-negative infections and creatinine clearances greater than 5 250 mL/min. Cefazolin and vancomycin were the main antibiotics used in the study. These observations may 251 be useful when attempting to understand the impact of RKF on treatment outcomes and raise the 252 question as to whether patients with RKF greater than 5 mL/min were under-dosed with antibiotic in 253 previous studies. 254 In patients treated with vancomycin, RKF may account for 10-23% of the total body clearance in 255 PD.(12, 22) Studies examining the impact of RKF on vancomycin clearance, exposure, and treatment 256 outcomes in PD-related peritonitis are limited. Interestingly, for the subset of patients with a glomerular 257 filtration rate greater than 5 mL/min, RKF accounted for 39-84% of the total vancomycin clearance.(12) 258 It would appear that the impact from RKF has a substantial effect on the total clearance of vancomycin. 259 Thus, the recent 2016 ISPD recommendation of removing the 25% dosage increase to account for RKF is 260 unclear as most of the studies cited accounted for a dosage increase for those who were non-anuric.(32, 261 35) In the absence of additional studies, dosage adjustments to account for RKF may still be appropriate 262 as there is a substantial contribution observed on the total vancomycin clearance. For now, we can only 263 speculate that the resulting impact in treatment failure for Gram-positive peritonitis may be associated 264 with higher drug clearance values in patients with creatinine clearances greater than 5 mL/min. 265

THERAPEUTIC DRUG MONITORING AND PHARMACODYNAMIC RESPONSE 267
Vancomycin therapeutic drug monitoring is critical for patients with peritonitis and is routinely 268 performed because 1) the concentration plays the key component for the effect and 2) the initial 269 antibiotic dose is needed to target the maximum effect in order to allow proper eradication and 270 prevention of resistance. Moreover, the treatment window timeframe is crucial for patients. Hence, 271 appropriate plasma sampling during this timeframe is important, but may be difficult as the turnaround 272 time for assay results is a rate-limiting factor in achieving desired therapeutic drug levels. Furthermore, 273 not only is it important to ensure that the initial dose is sufficient, but also if that initial dose is able to 274 maintain therapeutic effect throughout treatment. Yet, current clinical practice is based on empirical 275 decisions, which may not reflect the most optimized regimen for patients on PD. 276 The traditional role of plasma trough concentration monitoring has been conflicting in the PD 277 population. Unlike the established optimal plasma trough levels of 10-15 mg/L for uncomplicated 278 infections or 15-20 mg/L for complicated infections, there is substantial interpatient variability for those 279 patients on PD. Higher rates of PD-related peritonitis relapse have been associated with a cumulative 4-280 week plasma trough below 12 mg/L when compared to those maintained above that threshold.(36) In 281 this study, vancomycin was given intravenously where plasma levels were maintained above 12 mg/L 282 rather than the current 15 mg/L recommendation by the ISPD. The type of modality did not differ 283 among the outcome groups, however vancomycin clearance and RKF information were not reported 284 which may have contributed to variability in the plasma concentration. On the other hand, data from a 285 single-center study involving 34 PD patients experiencing PD-related peritonitis showed no relationship 286 between plasma vancomycin levels measured during the first week and PD-related peritonitis 287 outcomes.(37) Here, CAPD was reportedly the most frequent modality (80%) used with an average 288 residual creatinine clearance of 2.8 mL/min/1.73m 2 . Vancomycin was dosed based on ISPD 289 recommendations and plasma levels were maintained above 15 mg/L. Of these 34 PD patients with 290 confirmed Gram-positive infections, 43% of cases were associated with coagulase-negative 291 Staphylococcus ssp. while only 11% of cases were due to MRSA. In total, although the frequency and 292 level of vancomycin measurement was not associated with adverse clinical events during the first week 293 of treatment, the number of patients studied may be too small to draw a firm conclusion. 294 Pharmacokinetic sources of variability can be explained in part due to varying exchanges provided by the 295 patient's PD modality, impact from RKF, and peritoneum physiology affecting drug absorption. In 296 addition, the pharmacodynamics-or bacterial susceptibility measured by its MIC-contributes to the 297 variability in clinical response, which may not be explained due to vancomycin pharmacokinetics alone. 298 Taken together, vancomycin shows substantial interindividual variability in clinical response for 299 patients treated for PD-related peritonitis. Table 3 gives an overview of the 300 pharmacokinetic/pharmacodynamic factors to be considered at the time of TDM of vancomycin in 301 patients on both CAPD and APD regimens. 302

CONSIDERATIONS FOR INTRAPERITONEAL DOSING 304
Clinicians should consider dwell times that achieve substantial equilibrium between the 305 peritoneum compartment and the systemic circulation. The reported bioavailabilities in literature are 306 dwell-time specific and may not be applicable in all patient-specific situations. Therefore, considering 307 the transfer half-life between the dialytic compartment and systemic circulation can be useful to 308 understand the time that it takes to reach equilibrium (i.e., steady-state). This may take up to 15 hours 309 considering a transfer half-life of 3 hours.(19) In this situation, dosing during the long-dwell interval may 310 provide adequate drug absorption to achieve therapeutic concentrations in plasma in patients on APD. 311 The bioavailability of vancomycin significantly increases during PD-related peritonitis. Plasma 312 concentrations as high as 40 mg/L have been reported following a 6 hour dwell using recommended 313 intraperitoneal doses of vancomycin in PD-related peritonitis.(14, 16) Alternatively, plasma 314 concentrations as low as 10 mg/L have been reported following a 6 hour dwell using a 500 mg 315 intraperitoneal dose in PD-related peritonitis.(38) Regardless of the PD modality, absorption does not 316 largely change between CAPD or APD based on the equilibration half-lives reported. (12, 19, 20) 317 In patients with PD peritonitis on APD, doses of 15-20 mg/kg together with dwell times ranging 318 from 10-15 hours may be more appropriate than the targeted concentration strategy mentioned above. 319 TDM should also be performed to evaluate therapeutic and toxic concentration fluctuations and to 320 maintain concentrations above 15 mg/L as recommended by the ISPD guidelines. 321

FUTURE RESEARCH AND DOSING GUIDELINES IN AUTOMATED PERITONEAL DIALYSIS 323
Empiric Gram-positive management using vancomycin for PD-related peritonitis in patients on 324 APD is summarized in figure 2. This algorithm accounts for RKF and suggests a dosage increase of 20% 325 for those who are non-anuric with a creatinine clearance greater than 5 mL/min based on observational 326 outcome studies.(34) In addition, monitoring plasma vancomycin concentrations 48 hours post-dose 327 would be appropriate based on previous experience. As such, re-dosing would be necessary to maintain 328 the targeted 15 mg/L concentration. During this time, adjustments to antibiotic therapy should be 329 guided by the microbiology or susceptibility report. This should be practiced together with routine TDM 330 at appropriate sampling times to rationally select the effective dose for each patient. Pharmacometric 331 modeling and simulation could help to increase the knowledge on vancomycin dose exposure response 332 relationship and propose optimal dosing and TDM strategies in PD patients. 333 As above recommendations are based on limited evidence, dedicated studies are needed to 334 support them. Table 4

Proposal for Critical Research Areas of Needed Research for Vancomycin Therapy in Peritoneal Dialysis
▪ Effect of APD on peritoneal and plasma levels during rapid cycles ▪ Peak concentration following absorption from the long-dwell Increasing the dwell time enhances vancomycin bioavailability. Peritoneum and dialysate properties should be considered as these both affect the 542 rate and extent of absorption following an intraperitoneal dose. Following dosing and an appreciable dwell time, vancomycin is eliminated by PD, 543 renal, and non-renal sources. These processes make up the total body clearance of vancomycin. 544 This illustration is a derivative of "Simple squamous epithelium", "Arteries", "Arterial circulation" and "Bubble" by Servier Medical Art 545