Keywords

1 Introduction

Peritoneal dialysis (PD) is the dominant modality for home dialysis in end-stage kidney failure (ESKF), although its uptake varies enormously worldwide, ranging between 2 and 74 % of dialysis populations [1]. Despite proven economic advantage [2–4], improved quality of life [5, 6], higher levels of satisfaction with treatment [7], early survival advantage [8, 9], delayed need for vascular access procedures [10, 11], reduced blood transfusion requirements [12], reduced hepatitis virus transmission rates [13] and better preservation of residual renal function [14, 15], PD is a greatly underutilised dialysis modality [16], and there is decreased uptake across North America, Australia and New Zealand, with greater variability within Europe [16–18]. Infection risk and concern about inferior outcomes are the most commonly cited reasons for preferential uptake of haemodialysis (HD) [16, 18]. Despite this impression, research indicates that HD and PD patients have similar overall infection risk [19] and that improvements in PD outcome have outperformed those seen with in-centre HD [20]. Nonetheless, the complications of PD represent barriers to its widespread implementation, and their management and prevention are key to maintaining PD technique and overcoming clinicians’ prejudices.

2 Infectious Complications

2.1 Peritonitis

2.1.1 Epidemiology and Risk Factors

PD -related peritonitis is the most frequent, serious complication of PD and is the most common reason for transfer to HD [21]. PD peritonitis contributes to about 20 % of PD technique failure [21] via increased risks of catheter removal and permanent HD transfer. Long-term peritonitis damages the peritoneal membrane , resulting in ultrafiltration failure and dialysis inadequacy and may contribute to the most feared complication—encapsulating peritoneal sclerosis [22]. PD-related peritonitis increases mortality risk, accounting for 16 % of PD deaths [23–25] and increases morbidity in terms of hospitalisation and reduction in residual renal function [26].

Rates of PD-related peritonitis vary enormously across different centres and countries. Reported rates range from 0.06 to 1.66 episodes per patient-year [27], although the literature suffers from a paucity of well-performed studies dominated by single-centre reports. Even within the same country, there is considerable variation in the rates of PD peritonitis regardless of centre size. The Australian and New Zealand Dialysis and Transplant (ANZDATA) Registry has demonstrated a tenfold variation in PD peritonitis rates among centres [21] (Fig. 10.1), epidemiological studies performed in Scotland and the UK irrespective of centre size, patient-to-staff ratio or duration of PD training time [28]. Poor PD outcomes reflect variations in clinical practice and deviations from international guidelines, particularly with respect to prophylaxis practices with exit-site mupirocin and antifungal therapy during episodes of bacterial peritonitis [21].

Fig. 10.1
figure 1

Peritoneal dialysis rates by treating centre in Australia and New Zealand in 2011, as captured by the ANZDATA registry. (ANZDATA 2012 Annual Report)

Full size image

Contributing to this observed variation in peritonitis rates are inconsistencies in the definitions adopted by heterogeneous studies. The International Society of Peritoneal Dialysis (ISPD) has standardised diagnostic criteria along with the definitions for recurrent, relapsing, repeat, refractory and catheter-related infections [24]. Peritonitis patients presenting with cloudy effluent should be presumed to have peritonitis, which is confirmed by obtaining effluent cell count, differential and culture. Peritonitis should, however, always be included in the differential diagnosis of any PD patient with abdominal pain or fever, even if the effluent is clear. An effluent cell count with white blood cells numbering more than 100/µL following a dwell time of at least 2 h with at least 50 % polymorphonuclear neutrophilic cells reflects significant peritoneal inflammation and peritonitis is the most likely cause. The ISPD guidelines emphasise the percentage of polymorphonuclear cells rather than the absolute number of white cells to diagnose peritonitis and endorse an empirical approach to antibiotic therapy largely irrespective of the initial Gram stain as it is frequently negative or misleading [24, 29]. The role of the Gram stain is primarily to identify the presence of yeast or other fungal elements and thereby prompt early initiation of antifungal therapy and removal of the Tenckoff catheter [24]. Effluent samples should be inoculated into two blood culture bottles at the bedside and brought within 6 hours to the laboratory. Identification of causative organisms is not only important for determining antibiotic sensitivities and guiding antibiotic selection, but also for assisting in elucidating the source of contamination and risk stratifying the patient with regards to relapsing, recurrent and repeat infection. Questioning the patient about lapses in technique and, in particular, contamination or disconnection, may frame re-education attempts following resolution of the infection. Likewise, clinical features that suggest a gastroenterological source, such as recent endoscopy, constipation and the presence of localised tenderness suggestive of appendicitis or cholecystitis, may indicate the presence of an underlying surgical issue. The catheter should be inspected for evidence of exit-site and tunnel infection .

There is currently inadequate evidence to recommend the use of flow cytometry or multiple enzyme-linked immunosorbent assay to distinguish between Gram-positive and Gram-negative infections [28]. This novel development, however, suggests a future possibility for point of care testing and the emergence of more timely and targeted peritonitis therapy [30].

The ISPD guidelines state that rates of culture-negative peritonitis should not be greater than 20 % of episodes and could be further improved by culturing the sediment after centrifuging 50 mL of effluent [24]. The species cultured is useful for prognostication purposes. Relapsing peritonitis is an infection with the same organism or a sterile episode occurring within 4 weeks of completion of therapy, whereas recurrent peritonitis involves a different causative organism, still within 4 weeks of completion of therapy. Relapsing and recurrent peritonitis complicate 14 and 5 % of peritonitis episodes, respectively, and carry increased risk of catheter removal and permanent transfer to HD therapy [31]. By contrast, repeat peritonitis, defined as infection caused by the same organism after 4 weeks of completion of therapy, has more benign implications [32]. The detection of bacterial fragments in PD effluent following an episode of peritonitis may predict relapse or repeat peritonitis but has not yet been adopted into clinical practice [33].

Rates of peritonitis escalate where units deviate from ISPD guidelines [34]. A rate of 1 episode every 18 months (0.67/year at risk) has been deemed acceptable [24], although units should strive to improve beyond this. PD peritonitis rates as low as 0.36 episodes per patient-year (1 episode every 33 patient-months at risk) are considered achievable with adoption of best practice [29]. Departures from ISPD guidelines predict inferior PD outcomes and reduced technique survival [34, 35]. In Australia, the most recent official overall peritonitis rate is 0.43 episodes per patient-year (1 episode per 28 patient-months; Fig. 10.2). Surveys however indicate poor adherence (< 50 %) to evidence-based practices, such as administering prophylactic antibiotics at the time of Tenckoff catheter insertion, prescribing topical antimicrobial prophylaxis [36] or selecting appropriate antibiotics for treatment of peritonitis [37]. PD peritonitis rates in other parts of the world are likewise suboptimal [24, 34]. Although implementation of ISPD guideline recommendations into clinical practice remains suboptimal, the “Call to Action” initiative in Australia has demonstrated that the systematic adoption of standardised unit protocols based on ISPD Guidelines, education of young nephrologists in PD management, establishment of national peritonitis registry, introduction of a national key performance indicator project based on benchmarked peritonitis rates and conduct of ongoing surveillance of PD practice and patient outcomes [34] has achieved dramatic reductions in PD peritonitis rates across Australia [21] (Fig. 10.2).

Fig. 10.2
figure 2

Improved rates of peritoneal dialysis (PD)-related peritonitis following the “Call to Action” initiative and the launch of the PD Academy educational program. (ANZDATA Registry annual report 2012)

Full size image

Risk factors for PD peritonitis include advanced age, frailty and comorbidity along with lower socioeconomic status and indigenous racial origin [21, 28]. Smoking and obesity increase risk of infection generally and PD peritonitis more specifically, while pets and rural living emphasise the importance of good hygiene to successful PD technique [21]. Preserved residual renal function and the prior use of HD also increase risk [35, 38]. Patient preference for PD predicts technique success [39], while depression and anxiety predict higher rates of PD peritonitis [29]. In their discussion of PD peritonitis risk factors, Cho and Johnson [28] emphasise that there is no high-level evidence that modifying these risk factors will lead to reduced peritonitis rates, nor that for patients with nonmodifable risk factors increased home support, increased training frequency or more intensive infection prophylaxis mitigate peritonitis risk .

2.1.2 Empirical Management

The mainstay of peritonitis management is the timely initiation of empirical antimicrobial agents that are likely to eradicate the most common causative organisms, as endorsed by ISPD. Empiric antibiotics must cover both Gram-positive and Gram-negative organisms and should be based on local antimicrobial susceptibility data [24, 33, 35]. Vancomycin or cephalosporins may be used for Gram-positive organism cover along with third-generation cephalosporin, aminoglycoside or carbopenam for Gram-negative organism cover. First-generation cephalosporins, such as cefazolin or cephalothin, demonstrate generally equivalent outcomes to glycopeptides (e.g. vancomycin), although glycopeptide regimens were more likely to achieve a complete cure (3 studies, 370 episodes: risk ratio (RR) 1.66, 95 % confidence interval (CI) 1.01–2.72) [40]. On the other hand, cephalosporin administration may be associated with a lower risk of selecting for multiresistant organisms [41]. Short-term use of gentamicin (< 5 days) has not been shown to be associated with more rapid decline of residual renal function [26, 42]. This factor, together with the risk of ototoxicity, should, however, be considered during prolonged courses of more than 1–2 weeks duration where alternative agents should be sought [40].

Intraperitoneal (IP) administration of antibiotics is superior to intravenous (IV) dosing for treating peritonitis [39]. Intermittent versus continuous IP antibiotic dosing results in comparable clinical outcomes [40]. Rapid exchanges in automated peritoneal dialysis (APD) may lead to inadequate time to establish effective dialysate concentrations of antibiotics [24], although it is presently unknown whether or not APD patients with peritonitis should be temporarily switched from APD to continuous ambulatory peritoneal dialysis (CAPD) for the duration of their peritonitis treatment . One retrospective, observational study reported no differences in peritonitis-related relapse rates, catheter removal rates or death in 239 PD patients continued on APD during peritonitis treatment compared with 269 patients managed on CAPD [43]. Although further research in the area is clearly warranted, the ISPD Guidelines recommend that APD patients treated for PD-related peritonitis using an intermittent IP dosing regimen should dwell their antibiotic-loaded dialysis fluids for at least 6 h to facilitate adequate antimicrobial concentrations and effect [24].

Monitoring of serum antibiotic levels (vancomycin, gentamicin) during treatment of PD peritonitis has not been clearly demonstrated to result in improved efficacy or safety, but is often performed [44–46]. Re-dosing is generally advised when serum vancomycin levels fall below 15 µg/mL and serum gentamicin levels fall below 0.5 µg/mL.

It is imperative that antifungal prophylaxis, in the form of either nystatin or daily fluconazole, is administered during antibiotic therapy based on previous randomised controlled studies that such an approach reduces the risk of subsequent fungal peritonitis [47–50]. Unfortunately, registry data suggest that less than one in ten patients in Australia and New Zealand receive antifungal prophylaxis and risk-adverse outcomes following severe fungal peritonitis [35, 49].

Once culture results and sensitivities are known, antibiotic therapy should be adjusted to appropriate specific therapy. Efficacy of therapy should be assessed on clinical grounds; most patients with PD peritonitis show considerable improvement with 48 h of commencement of therapy. Repeated cell counts of ≥ 1090/mm2 within dialysis effluent predict treatment failure and catheter removal is indicated [51]. Refractory peritonitis, defined as failure of PD effluent to clear after 5 days of appropriate antibiotic treatment, should be treated with immediate catheter removal, as per ISPD Guidelines [24]. Prolonged attempts to treat refractory peritonitis with antimicrobial agents but without catheter removal result in extended hospital stay, peritoneal membrane damage and increased risks of fungal peritonitis and death. Catheter removal is also indicated in all cases of fungal peritonitis and many cases of Pseudomonas and relapsing peritonitis [31, 49, 52]. Following catheter removal and transfer to HD, patients who subsequently return to PD have comparable peritonitis-free, technique and patients survival rates to those PD patients who experienced peritonitis and did not have a catheter removed [53]. The survival of such patients was also comparable to those of PD patients who permanently transferred to HD following catheter removal for peritonitis [53]. Thus, patients who transfer to HD following a severe peritonitis episode should not be discouraged from returning to PD. The optimal timing of return to PD is not known at present .

For PD peritonitis that does respond promptly to IP antibiotic therapy (typically 70–80 % of cases), the ISPD guidelines endorse a minimum duration of antimicrobial therapy for peritonitis of 2 weeks for mild infections but 3 weeks for moderate-to-severe infection (e.g. Staphylococcus aureus, Gram-negative organism, enterococci, polymicrobial).

2.1.3 Microbiology

2.1.3.1 Gram-Positive Peritonitis

Historically, Gram-positive organisms and particularly Coagulase-negative Staphylococcus account for the majority of PD peritonitis cases. Such infections are typically of milder severity and generally reflect touch contamination and a break in technique. The introduction of disconnect systems rather than standard spike systems particularly improved the rates of Gram-positive contamination-related PD peritonitis [54–58]. Such infections typically respond rapidly to antibiotic therapy and may be appropriate for outpatient therapy. In some units, there is a very high rate of methicillin-resistance, which may necessitate the use of vancomycin as empiric therapy. Relapsing coagulase-negative peritonitis is suggestive of biofilm formation, which can be addressed through catheter replacement under antibiotic coverage as a single procedure [28]. A recent Cochrane systematic review demonstrated that, based on a single small study, simultaneous catheter removal and replacement was better than urokinase at reducing treatment failure rates (RR 2.35, 95 % CI 1.13–4.91) in the setting of relapsing or persistent peritonitis [40]. Streptococcus and Enterococcus peritonitis tend to present with more severe and painful infection and may reflect a metastatic source such as gastrointestinal tract, genitourinary tract, exit-site or tunnel infection, or dental abscess. Touch contamination should also be considered. Enterococccal infections carry a high risk of catheter removal (52 %), permanent transfer to HD (52 %) and death, which may be averted by timely removal of the PD catheter [24, 28]. Ampicillin remains the antibiotic of choice in vancomycin-resistant enterococcal (VRE) infection, but linezolid or quinupristin/dalfpristin may be required if ampicillin resistance is detected. S. aureus causes severe peritonitis and is frequently accompanied by exit-site or tunnel infection. S. aureus infection with concurrent exit-site or tunnel infection is frequently refractory and requires catheter removal and a rest period of at least 2 weeks off PD. Methicillin-resistant S. aureus infections are typically refractory and carry a high risk of permanent transfer to HD. Both vancomycin and rifampicin can be used intraperitoneally .

2.1.3.2 Gram-Negative Peritonitis

The most common causes of Gram-negative peritonitis episodes are Pseudomonas, E. coli and Klebsiella and clinically manifest as severe peritonitis with high rates of hospitalisation, catheter removal and permanent transfer to HD. Catheter removal is indicated if there is accompanying catheter infection. A review of 210 episodes of Enterobacteriaceae peritonitis found that recent antibiotic use and concurrent exit-site infection were predictors of this type of peritonitis with 10 % of patients dying within a month of peritonitis onset [59]. This emphasises the need for aggressive and appropriate therapy to prevent mortality and catheter loss. An assessment should be made for constipation, diverticulitis and colitis, which allow the transmural migration of coliforms. Gram-negative organisms have a high risk of relapse due to biofilm formation and require a longer duration of therapy. Pseudomonas infections should always be managed with two antipseudomonal antibiotics and must be continued for 2 weeks while the patient is temporarily transferred to HD. Stenotrophomonas peritonitis may follow the use of carbapenems, fluoroquinolones and late-generation cephalosporins, which select for this multiresistant organism. Therapy must be prolonged (3–4 weeks) and utilise two drugs; usually a combination of trimethoprim/sulphamethoxazole, ticarcillin/clavulanate and/or oral minocycline [24].

2.1.3.3 Polymicrobial Peritonitis

Polymicrobial infections , particularly those involving the presence of anaerobic organisms, carry a high risk of death and should prompt surgical evaluation for the possibility of underlying intra-abdominal pathology, such as cholecystitis, ischaemic bowel, appendicitis or abscess. Nevertheless, most recent reports suggest that polymicrobial peritonitis is associated with a relatively low incidence of catastrophic surgical pathology, ranging from 2.8 to  9 % [60–63]. Underlying surgical peritonitis should be suspected if patients present with haemodynamic instability, sepsis, lactic acidosis or elevations in peritoneal fluid amylase. The Gram stain may identify a mixed bacterial population and should prompt early surgical opinion, abdominal imaging and management with ampicillin, metronidazole and aminoglycoside in the recommended IV doses. Early catheter removal should be considered .

2.1.3.4 Fungal Peritonitis

Fungal peritonitis occurs in 1–23 % of peritonitis episodes [64]. Risk factors include multiple episodes of bacterial peritonitis , particularly polymicrobial, and recent (within 1 month) treatment with broad-spectrum antibiotics in the absence of adequate fungal prophylaxis [49, 60]. Outcomes following fungal peritonitis are generally poor with a risk of death as high as 25 % [51]. Fungal peritonitis necessitates immediate removal of the catheter and empirical management with amphotericin B or flucytosine and thereafter based on culture and susceptibility results [24] .

2.1.3.5 Tuberculous Peritonitis

Peritonitis due to mycobacteria is a rare occurrence, but should be considered when peritonitis persists or relapses despite antimicrobial therapy, in patients with systemic features and when the peritoneal effluent demonstrates a lymphocytosis. Outcomes are poor with very high rates of catheter loss (80 %) and significant mortality (40 %) [40]. Smears should be examined for acid fast bacteria with Ziehl-Neelsen stain but smear-negative disease is common. Although examination of dialysate effluent for mycobacterial DNA and/or adenosine deaminase is useful, exploratory laparoscopy with biopsy of the peritoneum has a higher diagnostic yield and should be considered when tuberculous peritonitis is suspected. Treatment reflects general tuberculosis protocols with the avoidance of ethambutol due to increased risk of optic neuritis in end-stage renal failure. Typical regimens include rifampicin, isoniazid, pyrazinamide and ofloxacin. Catheter removal is usually advised .

2.1.3.6 Culture-Negative Peritonitis

As previously indicated, rates of culture-negative peritonitis can be minimised by improved PD fluid sampling and culture techniques and should be below 20 % in all PD units. A history of previous antibiotic use is a recognised risk factor and should be sought on presentation. Outcomes following culture-negative peritonitis are relatively benign with higher rates of cure with antibiotics alone, less need for hospitalisation, less mortality, less catheter removal and increased maintenance of PD modality [65, 66]. Special culture techniques may identify lipid-dependent yeast, Mycobacteria, fungi, Campylobacter, Legionella and other fastidious bacteria. In clinical practice, if a patient is improving clinically on empirical therapy, it can be continued for duration of 2 weeks provided the effluent clears rapidly .

2.1.4 Outcomes Following PD Peritonitis

The majority of patients who suffer an episode of PD peritonitis will respond to therapy and continue with this dialysis modality. Studies indicate 80–85 % of peritonitis episodes are successfully treated [39]. In certain areas, including Australia, rates of suboptimal outcomes are higher and include a 14 % risk of relapse, 5 % risk of recurrence, 22 % risk of catheter removal and 18 % rate of permanent HD transfer [35, 49, 50, 52, 67, 68]. In their recent Cochrane Review, Ballinger and colleagues considered catheter removal to be equivalent to treatment failure [40]. Other bodies regard the need for catheter removal as a marker of the severity of the episode [69]. Certainly, a delay in catheter removal is associated with high rates of transfer to permanent HD and should be considered early in peritonitis cases caused by S. aureus [35, 69], Pseudomonas [35, 69], Enterococci [68], fungi [49] and multiple organisms [35, 60]. In their study of patients with severe peritonitis requiring catheter removal, Szeto and colleagues concluded that PD can be resumed in only a small number of patients and, when successful, predicts both patient and technique survival [69]. Other studies have likewise reported a low rate (about 20 %) [70] of successful reinsertion and resumption of PD following peritonitis. The timing of reinsertion is not clear but anecdotal recommendations endorsed by the Caring for Australasians with renal impairment (CARI) guidelines range from simultaneous removal and reinsertion to waiting a minimum of 3 weeks [71].

Peritoneal transport characteristics are dramatically altered by an episode of severe peritonitis with marked decline in ultrafiltration and increase in D/P creatinine ratio at 4 h, but dialysis adequacy and nutritional status can be maintained regardless of this [69]. It is interesting to that note that, in general, Asian patients enjoy better post-PD peritonitis outcomes than do Caucasian patients [34, 37, 69].

Epidemiological studies of cases of PD peritonitis report an association with PD peritonitis and mortality with highest risk in the first 30 days but extending until 120 days following an episode of infection [25, 27]. Low serum albumin predicts technique failure and death in patients on PD [72], which likely reflects its role as an acute phase reactant and marker of inflammation but may also reflect malnutrition. Hypoalbuminaemia also predicts catheter loss in PD peritonitis [73].

2.2 Exit-Site and Tunnel Infections

Exit-site infection is suggested by the presence of erythematous skin surrounding the catheter and definite in the presence of purulent drainage. A positive culture in the absence of these clinical findings may reflect simple colonisation rather than infection. Thus, the diagnosis requires experience and clinical judgement. Tunnel infection may occur with or without accompanying exit-site infection and manifests as erythema, oedema and tenderness, although is frequently clinically occult. Tunnel infection may be confirmed with the use of ultrasound. Aggressive management is recommended for S. aureus and Pseudomonas exit-site infection as there is often concomitant tunnel infection. Empirical antibiotic therapy may be initiated immediately and should always cover S. aureus or may be deferred until culture and sensitivity results become available to guide therapy. Oral antibiotic therapy is as effective as intraperitoneal antibiotic therapy. Gram-positive exit-site and tunnel infection can be managed with a first-generation cephalosporin, such as cephalexin. Alternatives include clindamycin, doxycycline and minocycline. These drugs do not require dose adjustment for renal failure. In slowly resolving or severe S. aureus infection, rifampicin 600 mg daily should be added. Pseudomonas exit-site infections are particularly difficult to treat and require prolonged duration of therapy. Oral fluoroquinolones as monotherapy may be a reasonable treatment option for mild cases. However, severe, refractory or recurrent pseudomonal exit-site infection requires a second antipseudomonal drug. Antibiotic therapy should be continued until the exit site appears normal. ISPD guidelines recommend a duration of 2 weeks for most exit-site infections, which is extended to 3 weeks in the case of pseudomonal exit-site infection [24]. Progression to peritonitis or concomitant peritonitis is an indication for catheter removal.

2.2.1 Prevention of Peritoneal Dialysis-Related Peritonitis and Exit-Site Infection

Practices to reduce infection risk in PD patients include a number of interventions already mentioned, including appropriate selection of patients, training and retraining of patients and staff education to increase professional confidence in PD and to champion its role in renal replacement therapy. The role of hand hygiene cannot be overemphasised. Patient education should teach routine exit-site care with water and antibacterial soap or non-cytotoxic antiseptics. Furthermore, units should establish ongoing surveillance of their infection rates and perform root cause analysis of all episodes of peritonitis [34]. Nasal carriage status of S. aureus should be sought and documented on all patients entering a PD program and eradication attempted. Nasal carriage of S. aureus is associated with an increased risk of catheter exit-site infection and its eradication with mupirocin has been shown to improve rates of both exit-site infection and PD peritonitis [74, 75].

There is currently no evidence to support the use of any particular catheter other than the standard silicone Tenckoff catheter for the prevention of peritonitis . The use of double-cuff catheters initially showed promise, but with the widespread adoption of prophylactic intranasal and exit-site ointments their role has diminished [28]. Prior to catheter placement, proper bowel preparation and skin cleansing, including removal of hair where indicated, can improve rates of peritonitis and exit-site infection [27, 76]. Catheter placement by an experienced surgeon with prophylactic single-dose IV antibiotics decreases the risk of subsequent infection [77]. There is some evidence that vancomycin may be superior to cephalosporin [27]. Newly inserted catheters should be immobilised but sutures at the exit site increase infection and are contraindicated [75].

Disconnect systems utilising twin-bag and Y-sets are superior to spiking of dialysis bags, and this is supported by randomised control trial evidence and systematic review [77, 78]. There are equivalent rates of PD peritonitis among APD and CAPD patients [79].

Exit-site infection and peritonitis caused by S. aureus and other Gram-positive organisms are prevented by the use of topical exit-site mupirocin. A number of studies have demonstrated efficacy, while meta-analysis data reveal reductions in the rates of peritonitis by 40–66 % and exit-site infection by 62–77 % [74]. Hence, the use of topical exit-site mupirocin is endorsed by the ISPD [29]. Gentamicin and Polysporin Triple ointment were also found to be effective in preventing bacterial exit-site and peritoneal infection at the cost of increased rates of fungal infection at these sites [80–82]. The ISPD guidelines recommend against their use and raise concern over the particular risk of inducing gentamicin resistance, which remains a useful drug for treatment of PD peritonitis [27]. Likewise, mupirocin resistance has been documented and it is expected that high-level resistance will eventually result in clinical failure or unacceptable relapse rate. The HONEYPOT trial proposed that medical grade, antibacterial honey has antimicrobial properties without inducing antimicrobial resistance and may have a role in preventing exit-site infections and hence PD peritonitis [83]. This multicentre open-label trial randomly assigned 371 patients to either topical exit-site honey or intranasal mupirocin. The rate of infection within the honey group was equivalent to that seen in the mupirocin group. Diabetics randomised to honey, however, had increased rates of exit-site and peritoneal infection and use of honey was poorly tolerated and greater numbers of patients withdrew from this arm compared to the mupirocin group. Thus, current evidence does not support the use of honey in PD patients.

Fungal peritonitis is predicted by the use of antibiotics for both bacterial peritonitis and nonperitoneal infections [49]. Fungal prophylaxis with oral fluconazole or nystatin has demonstrated benefit and should be routinely employed, particularly during prolonged courses of antibiotics for conditions, such as foot ulcer or osteomyelitis.

Severe constipation and diarrhoea may precipitate episodes of PD peritonitis via the transmigration of microorganisms across the bowel wall [84]. Hypomotility and gastroparesis likewise increase risk, and constipation is a common clinical problem that is often poorly recognised by the chronic PD patient. These disorders should be sought and managed appropriately to prevent PD peritonitis . Gastrointestinal pathology, such as cholecystitis, ischaemic bowel, diverticulitis and colitis, can cause an enteric peritonitis, which is frequently polymicrobial. Many nephrologists consider inflammatory bowel disease to be a contraindication to PD. The occurrence of surgical issues such as these may be an indication for catheter removal and transfer to HD.

Invasive procedures, such as colonoscopy, hysteroscopy and performance of dental work, can also lead to peritonitis at a rate of up to 6 in every 100 patients [85]. Ampicillin prophylaxis eradicated all cases of post-procedural peritonitis, although the difference was not statistically significant due to the relatively low event rates. The ISPD guidelines recommend emptying the abdomen of fluid prior to the procedure and giving consideration to pre-procedural antibiotic prophylaxis [27].

Combining and implementing all of the above prevention strategies in “bundle-of-care” programs delivered by experienced units with regularly trained doctors, nurse and patients can help to ensure adherence to evidence-based best practices, thereby effectively reducing peritonitis rates in continuous cycles of quality improvement [76].

The BalANZ trial recently demonstrated a potential role for neutral pH, low glucose degradation product (GDP), and “biocompatible” PD fluids in preventing PD peritonitis [86, 87]. This multicentre open-label randomised controlled trial assigned 185 incident PD patients with residual renal function to pH-neutral, low GDP dialysis solution or conventional dialysis solution for 2 years. The biocompatible group exhibited significantly longer times to first peritonitis episode and lower rates of peritonitis as well as longer times to the development of anuria. This same study also observed that use of novel dialysis solutions resulted in shorter peritonitis-related duration of hospitalisation, suggesting that biocompatible solutions reduced both the likelihood and severity of peritonitis. This finding has not been upheld in a recent Cochrane review of biocompatible dialysis fluids, but identified limitations in trial heterogeneity, definitions of peritonitis and high rates of attrition bias in included studies other than the BalANZ trial [88].

3 Noninfectious Complications

3.1 Catheter Complications

The success of chronic PD depends upon safe and permanent access to the peritoneal cavity [89]. A review of recent literature reveals catheter failure rates of up to 35 % at 1 year [90]. Catheter-related problems contribute to a significant proportion of failed PD cases and necessitate transfer to HD within the first year in up to 20 % of cases [91, 92]. Prevention, early recognition and appropriate management of these complications are important to avoid patient morbidity and disillusionment with the PD technique, which may undermine a patient’s willingness to persevere with PD [91].

A number of different variations to the standard Tenckoff catheter have been developed, including variation on the number of cuffs (one vs two), the design of the subcutaneous pathway (bent or “swan neck” vs straight) and the profile of the intraperitoneal portion (straight vs coiled). Systemic review and meta-analysis data reveal an advantage in favour of straight compared to coiled catheters [90]. There is inadequate evidence to recommend single versus double cuff catheters [89]. The use of a “swan neck” catheter is associated with the lowest rates of drainage dysfunction [90] and is endorsed by the ISPD [89].

Presurgical evaluation should include assessment for pre-existing abdominal wall herniation, as this will likely worsen when subjected to increased intraperitoneal pressures. A history of prior abdominal surgeries and most particularly abdominal catastrophe should be sought [89]. Skin and bowel preparation has been covered elsewhere. Insertion technique, either by laparoscopic route or open procedure, does not predict infectious outcome nor catheter survival [77]. The experience of the surgeon, however, is critical in optimising catheter outcomes [34, 89]. The catheter tip should sit deep in the pelvis and ideally within the left lower quadrant to minimise risk of incomplete drainage [89]. Some groups describe an advanced laparoscopic technique with rectus sheath tunnelling, prophylactic adhesiolysis and prophylactic omentopexy to fix redundant omentum to the upper abdomen via a suture [93]. They report reduction in the rate of catheter flow complications to < 1 % compared with 12 % with standard laparoscopic technique [93]. The advanced laparoscopic technique may be employed in patients with risk of catheter malfunction, such as prior abdominal surgery [91]. ISPD guidelines suggest catheter survival of > 80 % at 1 year is a reasonable goal [89] .

3.2 Inflow Pain

Pain on dialysate infusion may occur in the absence of infection and is a common occurrence in PD patients. The pain typically diminishes with the duration of the dwell period and usually improves with increasing time on PD as the peritoneal membrane adapts. Some patients, however, experience severe and persistent pain that necessitates the discontinuation of PD. The pain is attributed to the acidic pH of conventional lactate-buffered dialysate (usually pH 5.2–5.5) and bioincompatible hypertonic, high glucose concentration and dialysate temperature as well as catheter tip position. In their systematic review of biocompatible PD fluids , Cho and colleagues concluded that there was a significant reduction in inflow pain with the adoption of bicarbonate-buffered pH neutral, low glucose dialysate [94] and this was supported by meta-analysis [88]. Slowing the rate of infusion may also alleviate pain. It has been hypothesised that other sequelae of non-physiologic dialysate fluids, such as loss of peritoneal mesothelial cell viability and function, compromised peritoneal immune function, promotion of fibrosis and vascular remodelling within the peritoneal membrane are also reversible with the adoption of biocompatible fluids [94].

3.3 Outflow Failure

Outflow failure is defined as the incomplete recovery of instilled dialysate fluid within a reasonable time frame (30–45 min), which may be precipitated by constipation, catheter migration, intraluminal catheter obstruction by thrombus or fibrin, catheter kinking or catheter occlusion (e.g. by redundant omentum or adhesions). This issue complicates approximately 10 % of PD cases. It is not explained by catheter type (straight vs coiled intraperitoneal segment) [90], although may be amenable to advanced laparoscopic technique (described above) [93]. A kinked catheter usually demonstrates resistance to both inflow and outflow of dialysate and may be identified by plain abdominal radiograph. Likewise, the presence of constipation can be assessed by abdominal imaging. Catheter malposition is usually apparent within days of first using the catheter, while omental occlusion complicates PD several weeks after catheter placement. Physical examination can exclude leakage as a differential to outflow failure.

Management of outflow failure depends primarily on the cause identified. Liberal use of laxatives including suppositories and enemas can be used to treat constipation, and the resumption of bowel movement cures the majority of cases of outflow failure [95]. Intraluminal instillation of heparin and thrombolytics [96] may resolve both inflow and outflow obstruction, and guide-wire manipulation [97–99] can be considered when there is radiographic evidence of migration. Studies report an initial good success rate (85 %) but warn this is short-lived and prone to recurrence in the long term [97]. Laparoscopic repositioning and/or replacement of a nonfunctioning catheter remains a valuable recourse for long-term patency [97, 99] and is usually required if a catheter fails to flip down or unblock after 2–3 days of aperients and when no other cause has been identified. Occluded catheters may need to be managed by adhesiolysis, omentectomy or catheter replacement, as necessary .

3.4 PD Fluid Leakage

Leakage of dialysate may occur around the catheter (manifest as an accumulation of high glucose-containing fluid around the PD catheter exit site , subcutaneous swelling, genital oedema and/or apparent ultrafiltration failure), into the genitalia via a patent processus vaginalis, or into the pleural cavity (manifesting as dyspnoea, weight gain and ultrafiltration failure). The incidence (5 %) appears higher in CAPD , presumably related to upright posture and increased pressures on the abdominal wall but is widely under-reported [100]. Risk factors for pericatheter leaks include weak abdominal wall musculature following pregnancy or multiple abdominal surgeries, early initiation of PD following catheter placement , use of large intraperitoneal exchange volumes and factors impairing wound healing (e.g. diabetes mellitus, obesity, corticosteroid therapy, etc.) [95, 100]. There is no consensus on catheter choice for the avoidance of this complication but the use of a laparoscopic technique may reduce its incidence [95, 101]. Evidence favours a 14-day rest period following surgical catheter placement to allow postoperative healing [100], but urgent-start PD may employ low volumes in the supine position using a cycler [102].

Management of pericatheter leaks consists of decreased upright posture, temporary adoption of nocturnal APD, reduction in dialysis volumes and surgical repair. Genital oedema responds to surgical ligation of the patent processus vaginalis. Pleural leaks are more resistant to intervention and often necessitate transfer to HD [103, 104]. Some centres report success with chemical pleurodesis using either talc [105, 106] or tetracycline [107]. Systemic volume overload and congestive cardiac failure should be excluded [103, 104].

3.5 Abdominal Wall Herniation

Abdominal wall herniation can be a troublesome complication of CAPD and its risk is increased by female gender, parity, small body size, increasing age, longer time on PD, autosomal dominant polycystic kidney disease, diabetes and prior abdominal surgery [108–110]. Sites of occurrence include the inguinal canal and patent processus vaginalis, umbilicus, linea alba and site of prior abdominal incisions [111]. Symptoms include swelling and disfigurement and can be complicated by intestinal obstruction, bowel incarceration and strangulation. Diagnosis is made by physical examination and ultrasound imaging. Management consists of surgical repair employing a polypropylene mesh prosthesis, which allows the resumption of PD within several days of hernia repair, usually via low-volume supine rapid cycling PD and graduated return to the former PD regimen [110, 111]. The Bargman protocol for postoperative management of PD following hernia repair avoids interim transfer to HD while avoiding underdialysis and re-herniation [112].

3.6 Intestinal Perforation

Intestinal perforation can complicate catheter implantation due to direct injury or may occur later due to bowel wall erosion and ulceration. In patients with advanced vascular disease, intestinal perforation may be precipitated by ischaemia of the bowel wall. It is an uncommon complication that reflects the experience of the surgeon. Clues to presentation are the occurrence of polymicrobial peritonitis , bloody or feculent dialysate and diarrhoea following dialysate instillation. Management includes cessation of PD with catheter removal and surgical repair of the bowel under antibiotic coverage [95].

3.7 Haemoperitoneum

Bloody peritoneal dialysate is an infrequent occurrence and may reflect a range of intra-abdominal events with both benign and harmful significance. It should be remembered that a very small amount (< 1 ml of blood) can make peritoneal fluid appear blood tinged and that in the absence of PD many of these events may be clinically silent.

Benign causes of haemoperitoneum include menstrual bleeding, likely secondary to ovulation, retrograde menstruation and endometriosis. Rapid flushes and the instillation of heparin can prevent obstruction of the PD catheter due to clots. Mild and spontaneously resolving bleeding can also follow catheter manipulation or insertion. To date, there is no evidence favouring coiled catheters over straight tip catheters with respect to the complication of catheter-related haemoperitoneum.

Intra-abdominal pathology, such as rupture of liver cysts and splenic injury, may cause intra-abdominal bleeding, sometimes with surgical implications. Retroperitoneal events including rupture of kidney cysts may uncommonly cause haemoperitoneum accompanied by haematuria [113]. Intra-abdominal malignancy including liver carcinomatosis [114] and renal cell carcinoma [115] may potentially cause haemoperitoneum in PD. This possibility can be evaluated further with PD fluid cytology and computed tomography (CT) of abdomen.

3.8 Encapsulating Peritoneal Sclerosis

Encapsulating peritoneal sclerosis (EPS; formerly sclerosing encapsulating peritonitis) is a complication of PD characterised by persistent, intermittent or recurrent adhesive bowel obstruction with peritoneal fibrosis and malnutrition. Its incidence varies from 0.3 to 3.3 % and increases with time on PD with occurrences as high as 6.4 and 19.4 % in Australia and 2.1 and 5.9 % in Japan at 5 and 8 years, respectively [22]. Mortality rates have been reported to be as high as 50 % but more recent data from an Australian study suggest a considerably lower mortality risk [116]. The ISPD guidelines emphasise that EPS is infrequent and its risk of occurrence should not time limit the delivery of PD [22]. In this way, it is comparable to the risk of infectious endocarditis or osteomyelitis in the HD population. Risk factors for EPS include time on dialysis [22, 117, 118], bioincompatible dialysate [120], dialysate contamination [119, 120, 122], catheter type and episodes of severe peritonitis [116–118, 121].

Clinical manifestations vary widely and there are no reliable biochemical or radiological screening tests. It is frequently recognised following the cessation of PD. Patients present with clinical features of bowel obstruction, including anorexia, nausea, vomiting and weight loss. Other presentations include haemoperitoneum and sterile nonresolving PD peritonitis. Inflammatory markers may be present, such as raised serum C-reactive protein concentrations, anaemia and hypoalbuminaemia. Suggestive CT features include peritoneal calcification, bowel thickening, bowel tethering and bowel dilatation. Laparotomy is required for a definitive diagnosis [22].

Membrane transport characteristics may reflect a decline in ultrafiltration capacity and increase in peritoneal membrane small solute transport over time, similar to that seen following episodes of severe peritonitis (see above) and with long-term PD [120]. However, screening for a rise in transport characteristics is not helpful in risk-stratifying patients, as EPS can occur in patients with slow transport characteristics.

Management consists of cessation of PD, removal of the catheter and transfer to HD. Nutritional support via parenteral nutrition may be necessary and many patients will recover with conservative therapy. Drug therapies include corticosteroids, tamoxifen and immunosuppression, although the evidence for these treatments is scant. Surgical enterolysis by an experienced surgeon may improve symptom burden and survival [22] .

4 Conclusion

Peritonitis is a major complication that undermines the significant lifestyle, survival and economic benefits of PD. It represents a major disincentive to uptake this dialysis modality and has profound morbidity, mortality and healthcare consequences. Central to the management of peritonitis is the adoption of appropriate prophylaxis strategies, continuous quality improvement programs and the implementation of evidence-based practice and retraining programs. Noninfectious complications of PD are likewise amenable to thoughtful presurgical evaluation and management based on best practice. Further collaborative research is required in this area to overcome the barriers to maintaining long-term PD.