Abiding by the recent definitions of outcomes and measurement methods8, our experience confirms that peritonitis still represents a serious cause of morbidity and mortality in PD patients. In fact, peritonitis required hospitalization in more than 50% of the cases and accounted for approximately 35% of the transfer to HD and 7% of deaths.
Our overall peritonitis rate was far below the threshold of 0.4 episode/patient-year proposed by the ISPD recommendations8. This overall good rate might be associated with the early start of in-hospital PD training (mainly the day after the catheter insertion) that patients received in our center9,12.
Significantly, nearly 60% of the incident patients never experienced peritonitis, while the peritonitis-free survival at 3 months and at one year of treatment was above 90 and 80%, respectively.
Furthermore, more than two-thirds of the episodes were successfully treated by antibiotic therapy, conversely catheter removal was necessary in less than one-fifth of the events. Those results should reassure PD candidates that even if peritonitis may represent a severe complication, in most of the cases it is manageable by medical treatment.
The empiric antibiotic therapy (first-generation cephalosporin plus aminoglycosides) resulted effective in almost 65% of the cases. Although the initial treatment was successful in the most cases of peritonitis caused by gram-positive bacteria, a second-line antimicrobial therapy was required in almost two-thirds of the episodes due to gram-negative microorganisms. These results drew our attention to act soon on: (1) an investigation regarding Gram-negative antibiotic sensitivities at our center, (2) a tobramycin pharmacokinetic study in CAPD.
Whenever empiric antibiotic treatment fails, the identification of the organism and its antibiotic sensitivities indicate the possible source of infection and help physicians to guide the choice of therapy. In our experience, the rate of culture-negative peritonitis was approximately 25%, substantially greater than the suggested ISPD benchmark of 15%8.
Infectious culture-negative peritonitis may result from recent antibiotic exposure, suboptimal sample collection or culture methods, or unusual organisms.
Thus, we promptly revised the method of culturing PD effluent aiming at enhancing the yield of peritoneal fluid culture and reducing the time needed for a positive result. Two significant changes that we recently implemented were the following: (1) centrifugation of 50 ml of PD effluent followed by resuspension of the sediment in 5 ml and subsequent inoculation in blood-culture bottles; (2) routinely culture of PD dialysate in both aerobic and anaerobic media.
Culture-negative peritonitis compared with positive-culture episodes showed a more benign outcome in terms of hospitalization, catheter removal, HD transfer and death. Our results are in accordance with those of the large report from Australia and New Zealand Dialysis and Transplant Registry13, while are in contrast with Szeto’s data14. However, in the latter study only a low proportion of patients (< 5%) received the antibiotics at the onset of culture-negative peritonitis, while there was a significative delay in the timely administration of antimicrobial therapy in most of the patients, suggesting that if antibiotic treatment is promptly initiated, culture-negative peritonitis may likely show more favorable clinical outcomes than culture-positive episodes.
Furthermore, nearly all patients that required catheter removal as a rescue treatment for peritonitis, were permanently transferred to HD. Thus, the development of strategies aimed either at maintaining patients or shifting them back to PD are urgently needed. The use of simultaneous removal and replacement of peritoneal catheter in relapsing peritonitis after resolution of clinical signs and normalization of effluent WBC count has been proposed15,16. A similar approach could be also implemented in a case of a refractory peritonitis if it shows a partial response to the antibiotic therapy17. In the remaining cases, whenever feasible, the replacement of the PD catheter should be advocated as soon as possible to avoid long transit on HD18.
In our center pre-PD peritonitis percentage was extremely low (less than 1%) due to the double purse-string used for catheter placement that in most cases allowed the patient to initiate the peritoneal exchanges within 24 hours from catheter insertion9.
Cause-specific peritonitis that carry an inauspicious prognosis are enteric and catheter-associated peritonitis. We recorded a relatively low rate of those two entities in comparison to other series.
As far as catheter-related peritonitis is concerned, in our unit the extensive use of mini-invasive surgical technique in refractory tunnel infections could have prevented the occurrence of secondary peritonitis in some cases19,20; while the employ of simultaneous removal and replacement of peritoneal catheter in peritonitis arising in conjunction with the exit-site or tunnel infections could have decrease the associated HD transfer rate21.
The outcome of enteric peritonitis has not changed despite the many improvements in the practice of PD22,23. In our series these episodes were associated with death in above 60% of the patients. It is possible that early diagnosis of surgical peritonitis and exploratory laparotomy might improve outcomes23. Thus, in case of high suspicion of enteric peritonitis and no improvement after 48 hours from catheter removal together with broad-spectrum antibiotic therapy (including coverage for anaerobic bacteria), we are now aiming at carrying out a fast surgical referral to promptly evaluate the need of either a diagnostic laparoscopy or an exploratory laparotomy.
Considered altogether, relapsing, recurrent and repeat peritonitis accounted for approximately 20% of all episodes. Previous studies indicated that relapsing and recurrent as well as repeat peritonitis represent distinct clinical entities24–26.
Notably, our data confirmed that compared with generic episode, repeat peritonitis do not possess a higher rate of complications. Conversely, relapsing and recurrent peritonitis were significantly more likely to be complicated by catheter removal and death showing approximately a double probability.
Thus, the causes of these two entities should be deeply investigated in the future, aiming at devising adequate preventive measures to decrease the associated-morbidity and mortality.
The limitations of our study are mainly related to its single-center and retrospective nature. However, in comparison with registry studies, these data should be more complete, and their accuracy easier to be verified.
Despite the previous drawbacks, the paper includes long-term follow-up of a relatively large and unselected PD cohort with data collection performed trough the support of an electronic database.
In summary, our study adds additional evidence to some relevant aspects concerning PD-associated peritonitis. Although representing the main cause of HD transfer cause and a potentially severe complication, more than 50% of incident PD patients will never experience peritonitis. Furthermore, entities such as relapsing, recurrent and enteric peritonitis carry a worse outcome than generic peritonitis. Thus, these episodes should be promptly addressed and aggressively treated. Conversely, culture-negative peritonitis showed a better prognosis as compared to positive-culture episodes. However, aiming at the prevention of microbial resistances, the reduction of culture-negative cases is necessary to minimize patient exposure to empiric therapy with multiple antibiotics. In our experience first-line empirical antibiotic therapy, constituted by intraperitoneal cefazolin in association with tobramycin, was effective in almost 65% of the cases. Nevertheless, the efficacy of this regimen was unsatisfactory in Gram-negative microorganisms requiring further microbiological sensitivities investigations as well as pharmacokinetic evaluation.
In conclusion, the use of standardized definitions followed by a defined protocols for peritonitis treatment based on the sources of the bacteria involved may enhance clinical practice and allows comparative studies.