Review of Antibiotic Dosing with Peritonitis in APD

Abstract

Peritonitis is the leading cause of transfer from peritoneal dialysis (PD) to hemodialysis (HD). It is also the leading cause of hospitalization of PD patients. The usual treatment of peritonitis for automated PD (APD) patients consists of antibiotics given once daily in the long dwell. However, the once-daily antibiotic dosing recommendations are based primarily on studies with continuous ambulatory PD (CAPD) regimens. Published studies on antibiotic dosing in APD are very limited. We will review the scant literature on this topic. It is possible that extrapolating once-daily dosing from CAPD to APD may lead to underdosing. There is a need for further pharmacokinetic studies of antibiotic dosing in APD.

Peritonitis has been referred to as the “Achilles heel” of peritoneal dialysis (PD). Peritonitis rates were high in the early days of PD. Published data from the late 1970s demonstrated very high peritonitis rates, using primitive connectology. During the first 2 years of PD at the University of Missouri, the peritonitis rate was 1 episode every 2.2 patient (pt) months (1). As connectology improved, rates rapidly dropped but were still problematic. National data from the Australia and New Zealand (ANZ) registry have demonstrated that widespread implementation of clinical practice guidelines (International Society for Peritoneal Dialysis [ISPD] and Kidney Health Australia-Caring for Australians with Renal Impairment [KAHCARI]) can dramatically reduce peritonitis rates on a national level. In 2008, the rates of peritonitis in Australia and New Zealand were 0.62 and 0.8 episodes/yr, respectively, which fell to 0.41 and 0.47 episodes/yr, after widespread implementation of the clinical practice guidelines (2,3).

Even as peritonitis rates have decreased over time, infection is still the leading cause of transfer from PD to hemodialysis (HD) in the United States (4,5). The 2017 United States Renal Data System (USRDS) reveals that all-cause hospital admissions were slightly higher for HD than PD (1.73 vs 1.69 admissions/pt year), but that admissions for infection are more common in PD vs HD (0.56 vs 0.44 episodes/pt year) and infection is the leading cause of hospital admission for PD patients (6). Infections in PD patients account for one third of the admissions and are higher than cardiovascular (CV) causes, in contrast to in-center HD patients. This makes the prevention and treatment of peritonitis a top priority among PD practitioners.

The empiric treatment recommendations for peritonitis are: gram-positive coverage with first-generation cephalosporin or vancomycin; gram-negative coverage with third-generation cephalosporin or an aminoglycoside (7). Intraperitoneal (IP) dosing is preferred over intravenous dosing of antibiotics, unless signs and symptoms of systemic sepsis are present (7,8). Unfortunately, data regarding antibiotic dosing for PD peritonitis are extremely limited, especially for APD (8). Li et al. in recent guidelines point out that extrapolating dosing from continuous ambulatory PD (CAPD) to automated PD (APD) may lead to underdosing (7). This paper will review pertinent issues with PD peritonitis and the antibiotics commonly prescribed for peritonitis with a focus on their use in APD to highlight the deficits of the research in this area.

First Generation Cephalosporins, Specifically Cefazolin

There are very limited papers studying dosing of first-generation cephalosporins in APD as shown in Table 1. Perhaps because of this, the most recent International Society for Peritoneal Dialysis (ISPD) guidelines (7) eliminated the table for APD antibiotics and only showed dosing for CAPD as opposed to the guidelines in 2010 (13–15). The most recent guidelines suggest that continuous dosing of cephalosporins in APD may be preferable (7). This recommendation may have been made due to the paucity of pharmacokinetic data available on intermittent dosing of first-generation cephalosporins in APD. Tosukhowong et al. state because cefazolin has no post-antibiotic effect, it is given continuously (16). When a drug concentration is less than the minimum inhibitory concentration (MIC), it loses bactericidal and bacteriostatic effects. Therefore, once-daily dosages are not recommended for cefazolin. However, in the US, once-daily dosing is a common practice.

Table 1

First Generation Cephalosporins: Cefazolin

Author/year	Type of study	Number of patients & peritonitis yes/no	Pharmacokinetic data	Outcome
Fielding 2002 (9)	Retrospective review of 60 peritonitis episodes in 40 patients	40 APD with peritonitis	No	– 78.3% episodes resolved – 82.9% of gram-positive resolved – 50% of gram-negative resolved – Relapse rate 8% (pts whose catheters were not removed) – States this was effective therapy
		Cefazolin 1.5 g IP QD in long dwell
Manley 2000 (10)	Prospective observational study	10 APD no peritonitis	Yes	– Dosing of 15–20 mg/kg cefazolin adequate
		Cefazolin 15 mg/kg QD in long dwell
Troidle 1999 (11)	Prospective observational study patients with peritonitis	40 APD with peritonitis	No	– 55% episodes resolved with QD cefazolin – States therapy is not optimal
	Cefazolin 2 g QD in long dwell
Peerapornratana 2017 (12)	Prospective observational study	6 APD no peritonitis	Yes	– Adequate cefazolin in plasma >8 mg/L at 24 hours
	Cefazolin 20 mg/kg in short dwells over 10 hours on cycler dextrose No last fill Patients with dry day

APD = automated peritoneal dialysis; IP = intraperitoneal; QD = daily; pts = patients.

It may be that with a single dose given in the long dwell of APD, dialysate cephalosporin levels are low during the short nighttime cycles. Cephalosporins are bactericidal when the concentration exceeds the MIC of bacteria, usually 1 – 8 mg/L. Concentration of the antibiotic at the site of infection is most relevant for killing bacteria (17). Whether this has a clinical impact on relapsing/repeat peritonitis and biofilm formation is unclear. Biofilm may be the cause of both relapsing and repeat peritonitis (7). Peritoneal dialysis catheters have been shown to be colonized by bacterial biofilms and may lead to recurrent infections (18–20). Biofilms have an innate lack of antibiotic susceptibility compared with the planktonic form of the organism (21). The MIC is the concentration of an antibiotic to eradicate the planktonic form of the bacteria (21). However, the MIC concentration is ineffective on organisms in their biofilm state (21). Eradication of the organism in the biofilm state may take 100 – 1,000 times the concentration of the antibiotic MIC level to be effective (21). The minimum biofilm eradication concentration (MBEC) is defined as the minimum concentration of an antimicrobial that eradicates biofilm (22). Girard has proposed testing for MBEC levels when patients have a longstanding or recurrent peritonitis (20). This requires further investigation.

Manley et al. demonstrated that high dialysate flow rates (> 5.50 mL/min) vs low dialysate flow rates (< 5.50 mL/min) lead to an increased rate of cefazolin clearance in North American PD patients (23). Automatic PD prescriptions typically use 12 – 16 L of dialysate per day compared with 6 – 10 L/day in CAPD (24). With most cycler prescriptions, the dwell time is 1 – 1.5 hours per cycle at night. It is unclear if this is adequate to allow diffusion from plasma to dialysate (site of the infection) at sufficiently high levels. What is clear is that CAPD and APD prescriptions are very different and we cannot conclude what works in CAPD will work APD.

Vancomycin

Vancomycin is routinely given intermittently intraperitoneally (IP) for peritonitis for APD and CAPD. This is a very convenient approach and is usually dosed every 3 – 7 days. However, Manley et al. point out this may lead to low peritoneal levels of vancomycin at the end of the rapid cycles and daily dosing may be preferable (25). Despite the routine use of vancomycin in APD patients, there are limited studies as shown in Table 2. It is not clear what serum vancomycin trough level is necessary for peritonitis resolution. Dialysate levels of vancomycin are considerably lower than serum levels. Fish et al. demonstrated on the 5th day, after receiving a 30 mg/kg dose of vancomycin, 98% of the 48 patients had vancomycin serum concentrations > 12 mg/L (26). However, 23% (8 CAPD patients and 3 APD) of those patients had a vancomycin effluent level (< 4 mg/L) at the end of a 4-hour dwell. The authors suggest that smaller more frequent doses of vancomycin may be preferable and serum vancomycin levels do not guarantee therapeutic levels within the dialysate. They also acknowledge that with APD the vancomycin levels may be even lower as the dwells in APD are shorter than 4 hours (26).

Table 2

Vancomycin

Author/year	Type of study	Number of patients and peritonitis yes/no	Pharmacokinetic data	Outcomes
Manley 2001 (25)	Prospective observational	10 APD no peritonitis	Yes	– Serum vancomycin and dialysate vancomycin levels above MIC only during 1st & 2nd dwell.
	Vancomycin 15 mg/kg given IV			– In APD vancomycin clearance is significant & daily IP dosing needed to provide adequate dialysate concentrations.
Fish 2012 (26)	Prospective observational	19 APD with peritonitis	Yes	– Serum vancomycin >12 mg/L 98% pts – Dialysate vancomycin <4 mg/L 23% pts – Conclude serum levels alone are not to be used to predict dialysate level due to low correlation coefficient – Suggest smaller more frequent dosing may be preferable
	Vancomycin 30 mg/kg IP Day 5 checked serum & dialysate vancomycin levels after a 4-h dwell
Mulhern 1995 (27)	Retrospective review Vancomycin 15 mg/kg IV Q week x 4	10 APD 21 CAPD All with peritonitis	Yes	– 9 episodes relapsed – Cumulative trough of <12 mg/L or initial day 7 trough <9 mg/L were predictive of relapse
Stevenson 2015 (28)	Retrospective cohort	3 APD 27 CAPD All with peritonitis	Yes	– Vancomycin levels were similar in patients achieving cure vs no cure – Conclude outcomes not associated with serum levels
Blunden 2010 (29)	Retrospective observational Vancomycin 25 mg/kg for anuric (increased by 25% for non-anuric	120 APD 267 CAPD All with peritonitis	Yes	– Vancomycin level did not predict cure or relapse of gram-positive or culture-negative peritonitis
Schaefer 1999 (30)	Prospective observational	152 pediatric pts with 166 episodes of bacterial peritonitis	No	– No breakdown between outcomes of CAPD and APD pts
	Vancomycin 30 mg/L continuous dosing vs initial loading dose 15 mg/kg followed by 2nd dose of 30 mg/kg after 7 days	Both APD and CAPD utilized (no breakdown given on number of APD vs CAPD)		Overall: – No difference in relapse rate continuous vs intermittent vancomycin – Eradication of causative organism more frequent in continuous vs intermittent at both 60 hours (p<0.001) and 7 days (p=0.004)

APD = automated peritoneal dialysis; IV = intravenous; MIC = minimum inhibitory concentration; IP = intraperitoneal; pts = patients; CAPD = continuous ambulatory peritoneal dialysis; pts = patients; Q week = weekly.

Mulhern demonstrated that patients who relapsed had lower serum vancomycin levels (7.8 ± 0.6 mg/L during relapse vs 13.7 ± 0.9mg/L during relapse-free episodes p = 0.0004) (27). In contrast, Stevenson showed no relationship between vancomycin levels and peritonitis cure, relapse or repeat; however, 79% of these patients were on CAPD (28). Another paper on both CAPD and APD (69% CAPD and 31% APD) patients showed no relationship between vancomycin serum level and peritonitis outcome (29).

A recent paper looked retrospectively at a group of 35 Canadian patients with 58 episodes of coagulase-negative Stapholycoccus (CNS) peritonitis (31). No episodes of relapsing peritonitis were found in patients with a mean serum vancomycin trough of > 16 mg/L. This is higher than the 15 mg/L recommended by ISPD (7). Of interest during a second episode of CNS peritonitis, higher vancomycin levels were not helpful in preventing further relapse/repeat peritonitis. This may suggest that if a CNS biofilm becomes established, further vancomycin treatment is unlikely to effect a cure (31).

Vancomycin-resistant S. aureus is a serious concern within the medical community. A vancomycin trough of < 10 mg/L may predict not only the risk of treatment failure but also the development of vancomycin-resistant S. aureus (32). This was from a consensus paper on therapeutic monitoring of vancomycin in adult patients. It was not studied directly with PD peritonitis.

Third Generation Cephalosporins, Specifically Ceftazidime

Table 3 shows the limited data on the use of third- generation cephlosporins in APD. Of the 4 papers listed, 1 was simulated data (Monte Carlo) (33). The Monte Carlo analysis demonstrated the inability of intermittent dosing with ceftazidime to achieve an IP antibiotic level of > 5 x MIC. The levels were < 5 x MIC for 75% – 100% of the time. The same was true with intermittent dosing of cefazolin (33).

Table 3

Third-Generation Cephalosporins: Ceftazidime

Author/year	Type of study	Number of patients & peritonitis yes/no	Pharmacokinetic data	Outcome
Hota 2011 (33)	Monte Carlo analysis	1,000 virtual pts – CAPD		– Continuous or intermittently dosed ceftazidime did not meet AUC/MIC ratio
	Modeled time IP antibiotic concentration > 5 x MIC & ratio of AUC/MIC occurred. – Used 2005 ISPD dosing for 70 kg man	(5 exchanges) – APD (52-hour cycles night & 27-hour exchanges during day)		– Continuous dosing CAPD & APD ceftazidime gave greater than 5 x MIC (50% treatment time)
	Intermittent dose: 15 mg/kg Continuous dose: LD 500 mg/L & MD of 125 mg/L			– Intermittent dosing failed to meet this 75%–100% of time
Peerapornratana 2017 (12)	Prospective observational study; 20 mg/kg in short dwells over 10 hours on cycler No last fill Patients with dry day	6 APD without peritonitis	Yes	– Adequate ceftazidime in plasma >8 mg/L at 24 hours
Kim 2011 (34)	APD 3 dwells on cycler with 2-day dwells. Sampled at mid-point & end of each dwell both blood and dialysate Simulated dosing of 15 mg/kg IV & 20 mg/kg IP	11 APD without peritonitis	Yes	– Ceftazidime IV 15 mg/kg or 20 mg/kg IP QD is adequate
Sisterhen 2006 (35)	Prospective non-randomized pilot study of one 24-h period 125 mg/L was given with each exchange	5 pediatric APD pts without peritonitis	Yes	– Continuous dosing of IP ceftazidime without a loading dose achieved adequate serum & dialysate levels

IP = intraperitoneal; MIC = minimal inhibitory concentration; AUC = area under the curve; ISPD = International Society for Peritoneal Dialysis; LD = loading dose; MD = maintenance dose; CAPD = continuous ambulatory peritoneal dialysis; APD = automated peritoneal dialysis; IV = intravenous; pts = patients.

Third generation antibiotics are almost always given in the long dwell daily. The guidelines recommend an intermittent once-daily dose (7). However, this daily dose was studied in CAPD but is widely extrapolated to APD. Very little data support this approach. More studies are needed.

Aminoglycosides, Specifically Gentamicin and Tobramycin

Aminoglycosides initially bind to the outside of the bacteria in a concentration-dependent manner, causing damage to the membrane and possible cell death. The antibiotic is then taken into the cell, where it attaches to ribosomes; this impairs the reading of messenger ribonucleic acid and kills the bacterium. This second bactericidal phase is not concentration-dependent. It is known as the post-antibiotic effect (PAE). Because of the PAE, aminoglycosides are suited for intermittent dosing (36).

Intermittent dosing of aminoglycosides may be more appropriate than continual dosing due to the phenomenon of adaptive resistance (AR). Adaptive resistance occurs in Pseudomonas aeruginosa and other gram-negative bacilli (37) and refers to the organism's resistance to the bactericidal effect of an aminoglycoside antibiotic (38). The drug refractoriness is enhanced by continuous exposure to the drug but can be reversed after several hours in a drug-free environment (38). Because of adaptive resistance, longer dosing intervals with aminoglycosides may improve their efficacy (37).

If used for more than a brief period, vestibular and ototoxicity may become an issue. Aminoglycosides have different levels of toxicity in the order of gentamicin > tobramycin > netilmicin (39). Gentamicin and tobramycin can be toxic to both the cochlea (hearing loss) and the vestibular apparatus (ataxia, vertigo, and oscillopsia), and the damage may be acute or chronic (40). Both ototoxicity and vestibular toxicity have been reported in PD with the use of gentamicin, with some data indicating cumulative effect with repeated courses (41 42 43-44). Empiric dosing as initial treatment of peritonitis is likely safe but long-term use is more problematic. The 2016 ISPD guidelines provide intermittent and continuous dosing recommendations for both gentamicin and tobramycin (7). The studies cited in the 2016 guidelines were all done on CAPD patients (7). In the current literature, studies on APD dosing are scant as shown in Table 4.

Table 4

Aminoglycosides: Gentamicin and Tobramycin

Author/year	Type of study	Number of patients & peritonitis yes/no	Pharmacokinetic data	Audiometry	Outcome
Fielding 2002 (9)	Retrospective review: Cefazolin 1.5 g IP QD and Gentamicin IP 0.6 mg/kg QD x 3 days (then as indicated) (UO<500 mL) or 1.5 mg/kg (UO>500 mL)	40 APD with peritonitis	No	No	– 78.3% successfully treated – 82.9% gram-positive infections resolved – 50% gram-negative infections resolved
Tang 2014 (45)	Retrospective cohort study Gentamicin 0.6 mg/kg Gentamicin level (day 2) If level <0.5 mg/L another dose given Total course 5 days	7 APD with peritonitis	Yes	No	– Gentamicin levels did not predict efficacy of treatment except for polymicrobial peritonitis
Manley 2000 (10)	Tobramycin 0.6 mg/kg IV QD	10 APD without peritonitis	Yes	No	– Tobramycin half life shorter on cycler (p<0.001) – Both serum & dialysate levels were above MIC throughout 24-h period. – Modeling indicates IP dosing should be 1.5 mg/kg on day 1, then 0.5 mg/kg thereafter

IP = intraperitoneal; UO = urine output; APD = automated peritoneal dialysis; MIC = minimal inhibitory concentration; IV = intravenous.

Conclusion and Recommendations

Few data are available on the pharmacokinetics of many commonly used antibiotics when given intraperitoneally in APD, yet this is the form of dialysis that most patients utilize in the western world. There are many areas where more research is needed. Pharmacokinetic and efficacy data are needed on continuous vs intermittent dosing, increased dosing for residual renal failure, effect of membrane transport status on dosing, and particularly dosing in APD vs CAPD.

Research is also needed to evaluate drug pharmacokinetics during peritonitis. A paper by Imada in 1988 demonstrated that CAPD patients with peritonitis who received IP doses of cefotaxime or cephalothin had lower levels of antibiotic in the dialysate than those who did not have peritonitis (46). The serum levels of cefotaxime were higher in patients with peritonitis than those without peritonitis. The serum level of cephalothin was not different between the 2 groups (46). The authors went on to say that high concentrations of antibiotic in the peritoneum can only be maintained by adding antibiotic to each CAPD exchange (46). Since peritoneal inflammation affects IP drug pharmacokinetics and absorption is increased into the blood, studies must also be done at different stages of peritonitis as the inflammation is resolving.

Also needed is more research into MIC vs MBEC to potentially reduce relapsing and repeat peritonitis. We urge further research on these areas to enhance treatment of peritonitis and improve outcomes. Nephrologists, nurses, and pharmacists should collaborate in designing and carrying out these studies. Both the ISPD and industry should help with funding of these studies.

In the absence of more data on dosing for APD, authors should carefully monitor levels of vancomycin, ensuring that the trough levels stay at 15 μg/mL as a minimum. Careful assessment of response to the antibiotics by repeated measurement of effluent cell count is useful in following the course of therapy of patients on APD in particular. In some circumstances, conversion from APD to CAPD may be ideal to ensure adequate delivery of antibiotics, but this is often impractical. Another option for peritonitis that is not resolving quickly may be to convert from intermittent dosing of antibiotics to continuous dosing while remaining on APD.