• -

  Sufficient dialysate volume (>50 ml) is extracted using an aseptic technique.

• -

  The sample is transported in a sterile container without preservative.

• -

  Part of the collected sample is inoculated in blood culture bottles (5−10 ml/aerobe and anaerobe bottle). A tube is to be reserved for direct Gram staining of the sample.

• -

  Until transport to the laboratory, the samples are to be stored at under 4 °C (except the blood culture bottles, which are to be kept at room temperature).

• -

  Whenever possible, the samples are to be processed within 6 h. A delay of over 12 h can reduce the reliability of the results.

• -

  From the practical perspective, a number of studies have shown that direct inoculation of the dialysate sample in blood culture bottles simplifies the procedure and affords successful results in a short period of time.

• -

  In centers with high negative culture rates, the results can be improved by centrifuging the effluent and culturing the resulting sediment in both solid media and inoculated in blood culture bottles.

If the cultures prove negative after 3−5 days using blood culture bottles, but the clinical evidence is strongly suggestive of PI, subcultures should be made on agar plates incubated at 37 °C in an atmosphere with 5−10% CO2 and in an anaerobic atmosphere, during an additional 3–4 days, in order to identify difficult-to-grow microorganisms.

The samples of peritoneal effluent are processed in two ways (Fig. D2):

- Study of cellularity – Leukocyte count: Analysis of the leukocyte count in dialysate requires a sample dwell time of at least two hours14. There is no evidence that samples obtained after very long peritoneal dwell times may give rise to overdiagnosis, though it is advisable to individualize the interpretation of cellularity in samples obtained from empty peritoneal cavities or in cases with very small drainage volumes. A low leukocyte count in the presence of marked turbidity suggests the associated presence of fibrin, other proteins, lymph or hemoperitoneum (Table D5).

The leukocyte differential formula is useful in situations where the dialysate sampling conditions are suboptimal due to short dwells in the abdominal cavity, or the use of a diluted sample. In these situations, the presence of >50% of PMNs supports the diagnosis. A relative eosinophil count of >10% establishes the diagnosis of eosinophilic peritonitis. Mild to moderate eosinophilia in PI can be seen in sporadic cases56, particularly in fungal or parasitic infections, as well as in association to chemical irritants (e.g., vancomycin), or the use of icodextrin. However, the presence of marked eosinophilia does not suggest an infectious origin of the condition.

The reactive strips commonly used for urine testing may be useful for the early diagnosis of PI57. These strips detect the presence of leukocytes, though in qualitative terms and without providing a differential count, with a sensitivity of close to 100%, a specificity of 95–97%, and a positive predictive value of approximately 95%. Their advantages include the immediate obtainment of results, and low cost. The study of the peritoneal effluent sediment based on the modified automated blood count is widely employed, since it is reliable, rapid and inexpensive. Conventional cytology (Giemsa and Papanicolaou staining) is even more reliable, and allows us to assess atypical presentations (e.g., neoplasms), but such techniques are expensive and should only be used where specifically indicated.

- Bacteriological study. The study of a dialysate sample subjected to Gram staining is very recommendable, despite its low sensitivity (which can be incremented if the sediment is resuspended after centrifugation of the original sample), since it is strongly orientative when positive14. In particular, the presence of yeast forms is usually easy to detect with this method58.

Microbiological culture identifies the causal microorganism and guides the best therapeutic option.

The advisability of strict adherence to the recommendations of the present guide or to those of other clinical practice guides referred to PI is debatable, given the scarcity of firm evidence on the questions raised. Most recommendations are based on the opinion (even when of a collegiate and fundamented nature) of groups of experts, who can be more or less accurate in the same way as any other groups of experts. It may prove difficult to convince a team with extensive experience in PD to follow the recommendations of professionals who are no more experienced than themselves, and which may relate to questions under clinical and sociosanitary circumstances that differ from those of their own particular setting. Furthermore, there is no evidence that strict adherence of each and every one of the recommendations of the guides is able to improve the outcomes, though there are clear signs — particularly from the Anzdata setting44,68 — that structured interventions based on the principles contained in the guides are able to improve the control of PI, and initiatives have been made seeking to confirm these signs69.

Overall, this Committee expresses the following OPINIONS (ungraded):

• 1

  Any clinical guide on the prevention of PI with quality criteria (directed by groups of recognized experts and using standardized methods) can provide useful guidance for the prevention and management of PI.

• 2

  The strength of the recommendations is conditioned by the quality of the evidence, but also by the consistency of the results of previous studies on the topic. In general, recommendations based on quality evidence and/or strong recommendations should be followed.

• 3

  Suggestions, based more on opinion than on evidence, should be considered by those centers lacking clear experience in the subject at hand, while experienced groups must have more freedom to adapt to their particular circumstances.

• 4

  The maximum benefits are expected more from global adherence to the clinical practice guides than from the adherence to concrete recommendations of such guides.

In recent years there has been growing interest in the application of strategies for continuous quality improvement (CQI) in healthcare practice, and PD has been no exception in this sense43,70,71. These strategies imply a global and continuous approach to problems (in our case, the general outcomes of a PD program), identifying and correcting the weak points in order to be able to detect and mitigate the weaknesses of the system. This process must follow a series of steps that include the correct definition of objectives, the designing of clear and concise protocols, adaptation to the available resources and, fundamentally, the monitoring of results71. In the concrete case of PI, any initiative for improvement is critically dependent upon detailed and up-to-date knowledge of the incidence of this complication. The ISPD recommends monitoring the events on a standardized and iterative or continuous basis11. This approach allows periodic comparison of the results, and the identification of effective preventive practices to ensure continuous improvement of the outcomes.

The ISPD guide of 2016 addresses the impact of continuous quality improvement strategies upon the incidence of PI in PD to only a limited extent14. The apparent reason for this is the lack of publications on the subject, and the existing studies are moreover of an observational and uncontrolled nature, and with evident risk of publication bias. Nevertheless, the studies coincide in detecting very significant reductions in the incidence of PI72,73. Other studies, involving prospective and more orthodox designs, have ratified these results70,74.

Little additional evidence has been forthcoming since the publication of the ISPD guide of 201614. An Anzdata document68 has analyzed the potential measures for consistently reducing the incidence of PI in PD, from a continuous quality improvement approach. In particular, the document underscores the convenience of the following:

• 1

  Facilitation of the awareness and disclosure of poor outcomes in comparison with other Units

• 2

  Identification of the needless variations in updated good practices

• 3

  Adaptation of the clinical practice guides to improve those areas with poor outcomes

In addition, it is recommended that scientific societies, patient associations and the public institutions should collaborate to:

• 1

  Define and establish operating and outcome indicators in the PD Units

• 2

  Conduct periodic audits

• 3

  Compare outcomes and practices between Units in order to adopt and implement improvements

• 4

  Favor the existence of registries, which should include data reflecting the level of care quality in the PD Units

• 5

  Establish a minimum body of professional standards

• 6

  Favor training programs, especially for residents in training and nurses dedicated to PD

• 7

  Identify the evidence shortcomings in clinical practice and adopt measures to correct them

Makhija et al.75 recently published an analysis of the economical benefits of a continuous quality improvement program, fundamentally resulting from a reduction in the incidence of PI. Their estimations, based on real data, included an investment return of 169% (the cost of carrying out the program minus the savings derived from the reduction of the number of events). The main measures of the continuous quality improvement process were improvement of the nurse/patient ratio, the adoption of exit site care protocols, standardized patient training guides, ongoing training and certification of nurses dedicated to PD, home visits and support, and the monitoring of results, with the adoption of corrective actions when needed. This study evidences the need for the organization and health authorities to be motivated to economically invest in actions for improvement43.

Another study, also from Colombia76, analyzed the incidence of early PI (in the first 90 days) in a national structure (3525 patients from 49 Units), where care was organized around highly standardized and evidence-based practices. The study reported very acceptable infection rates (0.23 episodes per patient and year), which the authors attributed to the integral and standardized prevention strategies used. The retrospective design of the study, the absence of well-defined prior objectives, and the lack of a control group weaken the conclusions drawn, however.

Despite the lack of direct evidence, it seems reasonable to assume that correct insertion and early care of the peritoneal catheter will lessen the ulterior risk of CRI and, therefore, of PI. It is advisable to carefully select the cutaneous insertion zone and exit site of the catheter (avoiding folds or friction zones), prepare the surgical field in accordance with the general surgical principles, and observe strict asepsis during insertion. Alternative strategies have been developed for patients in which insertion in conventional zones poses a problem (extreme obesity, stomas). These strategies make PD feasible, but their benefits in terms of the prevention of infections have been generally disappointing100.

The technique for construction of the catheter exit site is another potentially relevant issue. The cutaneous orifice should be relatively well adjusted to the catheter itself, and should be located at least 1 cm (and preferably 2−4 cm) from the surface external Dacron cuff (in catheters with two cuffs). Some experiences101 have led to the suggestion that orienting the catheter exit site downwards and avoiding suture stitches close to the orifice reduces the risk of CRI and of secondary PI11,25, though the evidence in this respect is inconclusive102,103.

The risk of CRI, and secondarily of PI, may also depend on the quality of care after insertion of the catheter, over both the short and long term.

The ISPD guides of 2016 does not go much into detail regarding asepsis during peritoneal dialysis exchange, and largely makes reference to previous documents on the subject11. There is general agreement among the professionals dedicated to PD on the advisability of correct hand hygiene during the exchange procedure. It is also recognized that such hygiene is often gradually neglected by patients over time — this representing a motivation for home visiting and retraining programs78,85,86. Although the different elements of adequate hand hygiene in dialysis exchange have been reviewed in depth139, there is very little controlled information on the specific impact of different factors (timing and duration of hygiene, appropriate type of soap, hand drying method, role of antiseptics or the wearing of gloves, etc.) upon the risk of PI. Some of the points to be remembered are:

• -

  Washing should last at least 15 s140,141, though some groups extend this to a minimum of 30 s.

• -

  The risk of contamination and transmission increases when the hands are not dried142, and it has been seen in patients on PD that wet hands hold up to 100 times more bacteria than hands that have been towel dried for 15 s143,144. The ideal drying method (towel, paper towel or hot air) has not been established. When a towel is used, drying ideally should last about 15 s, and this time should be extended to 45 s when a hot air stream is used, for drying of the hands.

• -

  Failure to remove rings when washing significantly reduces the efficacy of hand hygiene145.

• -

  Long native145 as well as artificial nails146 reduce washing efficacy and thus should be avoided as far as possible.

• -

  The usual soaps do not contain antimicrobial agents. They are able to eliminate dirt, but different studies show that the residual bacterial count is substantial147.

• -

  There are different types of antibacterial soaps. Some of them (such as those based on triclosan) are associated to significant bacterial resistance rates139.

• -

  It is strongly advisable to use alcoholic disinfectants (usually based on ethanol, isopropanol or n-propanol) after hand washing. These products offer antibacterial activity against grampositive and gramnegative microorganisms, including methicillin-resistant Staphylococcus aureus. A number of studies have shown that gels and solutions based on water-alcohol formulations improve the results of washing with water and soap148–151. In patients on PD, it has been reported that the sequence of rubbing the hands with 70% ethanol after washing with soap is more effective than either measure applied separately151. Although disinfection with water-alcohol formulations is an adequate alternative in cases where the patient lacks adequate elements for hand washing, it is the majority opinion of the panel of experts (64%) that the use of these products ideally should follow correct hand washing152.

• -

  The wearing of sterile gloves is an option in patients with problems for correct disinfection of the hands (including different allergies or skin diseases), though the results of this practice have not been established.

No information is available on how the conditions of the zone of exchange affect the risk of peritonitis. Traditionally, an exclusively dedicated, clean and well ventilated and illuminated room with access to running water has been considered ideal. However, many patients perform dialysis exchange in bedrooms or other shared zones, with no evidence that this circumstance increases the risk of infection.

The ISPD guides of 2016 indicates that the wearing of a facemask during the exchange is optional14. This decision is due to the fact that the evidence on the subject is scarce, weak and contradictory153,154. A related issue is the influence of dental and periodontal health upon the risk of PI. Patients on dialysis have higher periodontal disease rates than the general population, and this complication has been associated to the presence of systemic inflammatory markers155. A retrospective study involving 75 patients on PD156 has evidenced that those subjects who maintain higher standards of oral hygiene have significantly lower PI rates.

Domestic pets are a recognized potential cause of PI157–159, with important morbidity and mortality, as evidenced by some studies157. A recent review159 suggests a number of contamination routes: pets playing or resting on the cycler, resulting in contamination, or biting of the catheter or contamination of the hands and oropharynx of the patient. It seems reasonable to establish barriers to prevent pets from accessing the zone in which dialysis exchange is carried out, and to maximize the measures of hygiene on the part of the patient.

The ISPD guides of 2016 recommends the use of double bag (“flush before fill”) systems as the best connection option for exchanges in PD. This technique is warranted by the experience of three decades and by conclusive meta-analyses47–49. The evidence supporting the use of double bag systems versus double-connection Y-set systems is not so conclusive, but is nevertheless probably sufficient. This latter comparison is of scant relevance in Spain, since double-connection systems are little used in this country.

Since 2016 there has been no new information to modify the criteria described above. Some devices have returned to the concept of ultraviolet (UV) disinfection161, though without evidence supporting the results. On the other hand, a recent randomized trial162 has compared the results of an incremental PD regimen (3 exchanges) versus a conventional regimen involving four exchanges in 139 incident patients on PD. A fewer number of connections was associated to a lesser risk of PI, though statistical significance was not reached.

In its day, the fewer number and greater concentration in time of the dialysis connections raised considerable interest in the possibility that automated peritoneal dialysis (APD) might be able to reduce the PI rates to below those associated with manual PD. The first randomized studies appeared to confirm this possibility163. However, in the first decade of this century, large observational registries documented similar PI rates with both modalities164–166. This situation led the ISPD guides of 2016 to not consider the indication of APD for reducing the risk of PI14. The reason for the discrepancy with respect to the older studies could be related to differences in the connection systems (use of threaded connectors versus pin systems, use of Y-set systems in manual PD, etc.).

Following publication of the guide of 2016, a new observational study was carried out by the BrazPD working group167, which once again evidenced similar risk with manual and automated PD.

The use of so-called biocompatible solutions has become consolidated in the last decade, based on their real or purported benefits in terms of preservation of the peritoneal membrane, residual kidney function, and peritoneal defense mechanisms. With regard to this latter aspect, a significant number of clinical trials have been published, with contradictory results51,168–176. In view of this scenario, and based on meta-analyses that report equivocal results and considerable heterogeneity50,52,177, the ISPD guides of 2016 does not recommend these solutions as an instrument for the prevention of PI in PD14.

Since the publication of the ISPD guides of 2016, a new clinical trial randomized 78 prevalent patients on PD to biocompatible or classical solutions (buffered with lactate), with a follow-up of two years178. The incidence of peritonitis was found to be lower in the biocompatible solution group. A recent meta-analysis (that does not include the previously mentioned study of Farhat)179 essentially confirmed the impossibility of establishing a firm recommendation on this subject.

The ISPD guides of 2016 issues no statement on the effect of peritoneal glucose or icodextrin load upon the incidence of PI. A recent analysis of the BalANZ trial180 identified no correlation between peritoneal glucose load (stratified by median values) and the risk of infection, the type of microorganism isolated, or the clinical course. With regard to icodextrin, data from meta-analyses179,181 evidence no apparent association to the risk of PI.

The frequency of disruption of asepsis in the course of treatment with PD is not clear though, in view of the many connections which the patient performs each year, it can be presumed that such disruptions are not rare. A previous study offered a series of guidelines (ungraded opinions) to deal with these situations, including40:

• -

  The patient should be instructed during training on the potential risks associated with failure to adopt measures of protection after contamination, including immediate reporting of the event to the reference center.

• -

  Sampling for peritoneal effluent culture (optional).

• -

  Prophylactic coverage against grampositive and gramnegative microorganisms.

• -

  Any positive isolation should be dealt with as PI, even in the presence of a clear effluent, but should not be counted as such, due to the lack of an inflammatory response. If the isolated organism corresponds to a common microorganism of the skin, a confirmatory sample should be obtained before starting antibiotherapy.

• -

  If contamination through contact (including direct contact or the detection of bag rupture) was not followed by infusion, replacement of the transference line may suffice.

• -

  In the event of any doubt as to whether infusion occurred, assume that it did occur.

A retrospective study covering a 15-year period182 compiled 548 episodes of contamination. When the latter was only related to inappropriate contacts, with the transference system closed and without the infusion of dialysate (“dry” contamination), no PI episodes were recorded. In contrast, 5.6% of the so-called “wet” contaminations were followed by infection. Notoriously, the administration of antibiotic prophylaxis greatly reduced the risk of infection (to 0.5%). Although the design of this study resulted in a low level of evidence, its results directly support the data of Bender et al.40.

Since the introduction of the ISPD guides of 2016, no controlled data on the subject have been published, though the nomenclature on accidental contamination (“wet” versus “dry”) appears to have been accepted183. Likewise, no information is available on the type of antibiotics, the administration route or the duration of prophylaxis, in these cases.

Invasive gastrointestinal procedures (including endoscopies and biliary tract interventions) and gynecological techniques manipulate areas located close to the peritoneal cavity; consequently, they may constitute a source of contamination and PI. Although the available information is scarce and with a low level of evidence, the ISPD guides of 2016 contemplates certain aspects that may help to establish a series of guiding criteria14:

• -

  Biliary tract, lower gastrointestinal and gynecological procedures pose a defined risk of secondary PI, though this risk has not been quantified.

• -

  The administration of antibiotic prophylaxis could reduce the risk of PI in this context184,185.

• -

  In contrast, uncomplicated upper gastrointestinal procedures do not appear to imply a risk of PI185.

• -

  There is no evidence regarding a specific risk of PI following urological instrumentations, though most centers administer antibiotic prophylaxis based on the potential risk of bacteremia. There is no evidence supporting the routine administration of antibiotic prophylaxis in the case of bladder catheterization.

• -

  No optimum antibiotic prophylaxis regimen has been established, though it seems reasonable for coverage to include enterobacteria and intestinal anaerobes. The coverage of yeasts does not appear to be indicated on a systematic basis, but may be considered in situations of risk.

• -

  Likewise, there is no evidence supporting the tendency of some centers to void the peritoneal cavity for the procedure.

Since the publication of the guide of 2016, some further evidence of interest has emerged. A randomized trial recorded no apparent benefit of antibiotic prophylaxis with ceftazidime in patients on PD subjected to colonoscopy186. More recently187, a large retrospective observational study recorded secondary PI in 3.8% of the colonoscopic procedures performed in patients on PD. No patient receiving antibiotic prophylaxis suffered this complication. Notoriously, the incidence of PI in patients subjected to polypectomy was three times higher than among patients that did not undergo this procedure. In another less weighted study188, 26 gynecological procedures were complicated by 7 episodes of PI. The risk of infection was significantly lower among the patients that received antibiotic prophylaxis — though the sample size limited the reliability of the results obtained.

A recent article has re-evaluated this issue, advocating antibiotic prophylaxis also before upper gastrointestinal tract endoscopies, as well as voiding of the peritoneal cavity, in order to improve the peritoneal defenses183. However, this was an opinion document that did not in itself contribute evidence.

Transient bacteremia is common after dental procedures and can lead to PI. The high prevalence of periodontal disease among patients on PD may increase this risk155. It seems reasonable to administer antibiotic prophylaxis at least in procedures associated with a greater bacteremia risk (extractions, periodontal cleaning, endodontic treatments, dental implants), but the possible benefits of this strategy have not been clearly established. Overall, this suggestion is supported by 92% of the members of the panel of experts.

Although PI due to fungal microorganisms is infrequent (1–14% of all cases of PI), it constitutes one of the most feared complications of PD, due to its high mortality rate and associated failures of the technique189–193. The fact that between 56–89% of all such episodes appear after bacterial infection (whether peritoneal or otherwise) treated with antibiotics189,190,194 has generated recurrent interest over the last 30 years regarding the convenience of prescribing antifungal prophylaxis in these clinical contexts. Although multiple studies with low levels of evidence have been published, randomized trials195,196 and meta-analyses119,120 with positive outcomes have served as the basis of the ISPD guides of 2016 for supporting this measure. The guide does not include suggestions on other concrete clinical contexts, or on which antifungals are to be preferred or the duration of prophylaxis in these cases — though it does appear to acknowledge a certain advantage of azole drugs over nystatin14. Eighty-four percent of the members of the panel of experts share this same opinion.

No new evidence of interest has appeared since the publication of the ISPD guide. A descriptive observational study193 has not contributed anything new. A meta-analysis96 in turn has confirmed the previous impressions, with no further evidence. Two recent surveys3,197 describe great variability in the adoption of this practice in different parts of the world, ranging from 0 to 70%, as well as variable criteria for the use of such treatment. There is a notorious lack of information as to which antibiotic treatments should combine antifungal prophylaxis.

In addition to the measures commented up to this point, many other strategies for the prevention of PI have been proposed. In general, these approaches are based on observational studies that detect associations between a given factor and the risk of infection (Table D1)5,177,198–202. As is logical, little can be done about the non-modifiable risk factors, apart from disadvising PD in those patients who present such factors. Correct patient selection is therefore of great theoretical importance for reducing the posterior incidence of PI. However, with the exception of obvious conditions (e.g., patients with active inflammatory bowel disease or a history of diverticulitis), it is very difficult to predict whether a concrete individual will suffer frequent PI episodes, and the existing risk models are generally weak and inconsistent36,165,203. This Committee considers that it is more logical to adopt an individualized approach and to focus on correct training than to exclude patients simply because they belong to a certain risk group, where risk at an individual level is hard to establish.

In principle, focusing attention on the modifiable risk factors177 seems more promising, though in practice it faces important difficulties. The association of some of these factors (e.g., smoking) to the risk of PI is controversial. There is also no evidence that the correction of other factors (such as obesity or depression) is able to reduce the risk of infection. However, in other cases, the available evidence does deserve some consideration.

The prevention of enteric PI is a challenge for PD units, in view its high frequency and clinical aggressivity. Many of these infections originate from pre-existing gastrointestinal disorders. For years it has been speculated that factors which weaken the antimicrobial intestinal barrier can increase the risk of enteric PI. In this regard, the ISPD guides of 2016 briefly comments on the convenience of avoiding abnormal peristalsis — whether reduced (constipation, intestinal paresis) or accelerated (gastroenteritis) — though stressing the weakness of the evidence14. A relationship has been postulated between hypokalemia and infection due to enteric organisms204,205, though here again there is no direct evidence to suggest that correction of the disorder reduces the risk of this complication. A single study, with a low level of evidence206, has related the regular use of laxatives to a lesser risk of PI.

At the time of drafting of the ISPD guides of 2016, two observational studies207,208 related low 25-OH-vitamin D3 levels to an increased risk of PI.

A number of studies of interest have appeared since the publication of the guide of 2016, though with limited levels of evidence. A meta-analysis of 6 observational studies (totaling 3613 patients) concluded that the evidence relating hypokalemia to the risk of PI is very inconsistent209. On the other hand, two retrospective observational studies210,211 have detected an association between treatment with gastric acid secretion inhibitors and the risk of PI. In the first of these studies, the observed association was limited to infections caused by enteric microorganisms, and it proved much more consistent with H2 receptor antagonists than with proton pump inhibitors210. Lastly, two observational studies212,213 have related the existence of sustained overhydration to a risk of PI due to enteric organisms in patients on PD. In the more statistically potent of these studies211, the relationship was apparently consistent, though there was evidence of a confounding factor in the form of the presence of malnutrition.

An observational study214 has related plasma vitamin D levels to the risk of PI. Notoriously, the association did not depend on the administration of oral vitamin D supplements. A more recent meta-analysis215 including 5 studies with patients on PD, 11 studies with patients on hemodialysis, and one study with both types of treatment, demonstrated a lowered general risk of infection among the patients with normal or elevated vitamin D levels. In addition, the patients administered oral vitamin D supplements likewise showed a reduced risk of this type of complication. The significance of this meta-analyses in relation to the topic of this document is limited, due to the heterogeneity of the analyzed studies.

Adequate knowledge of the local microbiota conditions the initial treatment of PI. In its defect, the ISPD14 suggests administering vancomycin or a first-generation cephalosporin to cover grampositive microorganisms, and an aminoglycoside or third-generation cephalosporin to cover gramnegative microorganisms. However, some aspects need to be considered:

• -

  For different reasons (Table T2), multiple initial treatment protocols have been developed for PI in PD. The scarce available comparative evidence between different strategies, together with local factors, limit the scope of any attempt to establish general recommendations. The high prevalence of allergy (real or purported) to betalactam antibiotics – which are the basis of most of the treatment regimens – often makes it necessary to have secondary protocols. In these cases, there are effective alternatives for the treatment of infections due to both grampositive (vancomycin, teicoplanin, daptomycin or linezolid) and gramnegative microorganisms (aztreonam, aminoglycosides, quinolones).

  Table T2.

  Causes of variability in the antibiotic treatment protocols for peritoneal infection.

  Diagnostic efficiency of the Unit (quality of Microbiology)

  Local etiological spectrum

  Antibiotic susceptibility of the local microbiota

  Residual kidney function

  Clinical context (limiting factors or factors that complicate the response)

  Antibiotic allergies or intolerance

  Diversity of PD procedures (automated PD and others)

  Antibiotic versatility regarding administration routes and dosing intervals

  Restrictive versus proactive vision of the use of new antibiotics

  Economic resources and local availability of antibiotics

  Scarcity of controlled, quality data (evidence)
• -

  Although short cycles of aminoglycosides are agreed to be scantly toxic, the natural tendency, provided the microorganism is sensitive, is to continue with the initial treatment regimen. For this reason, it seems reasonable to limit the use of this drug family in patients with residual kidney function. In contrast, vancomycin offers a very convenient empirical therapeutic profile that compensates its potential toxicity, and the incidence of resistances is relatively low in Spain.

• -

  Gram staining is of limited sensitivity. Furthermore, even if grampositive or gramnegative bacteria are detected, this does not rule out the possible growth in culture of other microorganisms. In any case, gram staining does help to define the initial antibacterial management strategy. Its greatest usefulness is in the early detection of the presence of yeasts58, which would radically modify the management approach.

A review of the literature posterior to the guide of 2016 has only revealed two relevant contributions. In a small single-center randomized trial, Xu et al.218 showed that oral moxifloxacin or intraperitoneal ceftazidime as an adjuvant to intraperitoneal vancomycin offers equivalent resolution rates (78% versus 80%; p = 0.80). In a more recent multicenter randomized trial219, the treatment success rate (82.6% versus 81.1%; p = 0.4), as well as other secondary outcome variables, were found to be very similar on comparing intraperitoneal cefepime versus cefazolin with ceftazidime via the intraperitoneal route. According to the authors, this demonstrated the efficacy of cefepime monotherapy versus the conventional protocols.

The main advantages and inconveniences of the different antibiotic groups as initial empirical treatment for PI are reported in Table T3. This table clearly evidences why vancomycin and cephalosporins are the preferred choices. Carbapenems and daptomycin are excellent alternatives for covering gramnegative and grampositive microorganisms, though the cumulative experience in this case is more limited; as a result, many centers regard them as options held in reserve, in view of the efficacy still shown by the more traditional drug associations, which remain the first choice.

In overall terms, this Committee agrees with the position of the ISPD guides of 201614 regarding the convenience of withdrawing the peritoneal catheter in cases of refractory PI. Although there is no controlled evidence, the general experience is that this measure protects the life of the patient and reduces suffering, while its capacity to preserve the peritoneal membrane with a view to continuing PD is less evident253. We wish to underscore the equivocal nature of the definition of refractory PI according to the ISPD, since it literally contemplates the normalization of peritoneal cellularity by day 5 of specific treatment. We suggest that the term “refractory” be reserved for this context, but consider that cases showing a tendency towards improvement (clearly subsiding pain, negative control cultures, and a clear decrease in cell count with progression towards a monocytic profile) should be regarded as slowly resolving PI. In these circumstances, the decision to withdraw the catheter should be established on an individualized basis, after considering alternative strategies (change/addition of an antimicrobial agent, monitoring of antibiotic levels), assessing the clinical condition of the patient, and considering the risk-benefit ratio of the measure.

The ISPD guides of 2016 are deliberately ambiguous in establishing recommendations on the management of relapsing, recurrent or repeated PI14, specifying that, in any of these circumstances, withdrawal of the peritoneal catheter should be considered. While in agreement with this recommendation, this Committee wishes to make some observations:

• -

  In first relapses of PI, we recommend the checking of possible breaches in the conditions of treatment of the original infection, including potential response-limiting factors (catheter dependency, abdominal disease), poor adherence, non-optimum antibiotherapy or underdosing.

• -

  In first relapses of PI characterized by low aggressivity, the panel of experts are unanimous in suggesting that the best option is a new antibiotic cycle, prolonging its duration for at least three weeks and/or administering antimicrobials with good action against biofilms — with preference over direct withdrawal of the catheter.

• -

  In first relapses of PI characterized by aggressive behavior, or in second relapses, we recommend considering withdrawal of the peritoneal catheter.

• -

  In the case of recurrent or repeated PI, the first step should be to check adherence to the technique on the part of the patient or caregiver.

• -

  In the case of recurrent PI due to enteric organisms, it is essential to carefully evaluate the presence of underlying abdominal disease.

The ISPD guides of 2016 do not include a specific section on the approach to catheter-related PI, though multiple observations relevant to this situation are made. It is assumed that stricto sensu, fungal, refractory or relapsing PI should be included in this category, since it is considered that the catheter had been colonized, and that keeping it in place will perpetuate the infection.

In this aspect, this Committee agrees with the recommendations of the ISPD to consider withdrawal of the catheter in fungal, refractory or relapsing PI (see above). The aims are to accelerate recovery, mitigate suffering, reduce mortality and generate a setting in which the patient, if willing, can again attempt treatment with PD. Following withdrawal of the catheter, a period of time (at least two weeks) should be allowed until insertion of a new catheter. However, when the catheter is removed during a period of remission of PI, simultaneous withdrawal and reinsertion can be decided. This strategy in general affords good results in PI due to CNS14, and has been attempted (with less consistent success) in PI due to more aggressive species such as Staphylococcus aureus272 or Pseudomonas spp.273. This Committee supports this strategy in PI due to CNS, but not in the other mentioned cases.

In the case of PI related to infection of the exit site or tunnel of the peritoneal catheter, the ISPD guides of 2016 are not particularly explicit. It is necessary to refer to the ISPD guides of 201725 on catheter-related infection. In this scenario, the suggestion is to consider withdrawal of the catheter, allowing a minimum time of two weeks before inserting a new catheter. This strategy is particularly convenient in the case of infection produced by aggressive and/or persistent organisms such as Staphylococcus aureus, non-fermenting gramnegative bacilli or enterobacteria (Serratia spp., Proteus spp. and others)272. This Committee agrees with this approach, underscoring that the conservative management of these infections is possible, provided there are no complicating elements (refractory or slowly resolving PI, wall abscesses, or refractory or recurrent catheter-related infection).

There is no recent evidence to modify the considerations of the ISPD guides of 2016 14. The observational studies published by Szeto274 and Burke et al.275 remain as the only valid references for guiding the management of these cases. Following these publications, we have only identified two small single-center observational and descriptive studies that offer no further information of relevance276,277.

• -

  We recommend each PD Unit to develop a Continuous Quality Improvement program in order to reduce the incidence of PI (1C) (Question P1)

• -

  We suggest periodic meetings of the multidisciplinary teams in charge of the continuous quality improvement program in order to analyze the outcomes (2C) (Question P1)

• -

  We recommend the patient training and monitoring principles proposed by the documents of the ISPD in order to reduce the PI rates (1C)(Question P2)

• -

  We recommend the nursing staff in charge of patient training and monitoring to be specialized (1C) (Question P2)

• -

  We suggest that visits to the home of the patient, if possible, beginning before the start of treatment with PD, may help reduce the risk of PI (2C) (Question P2)

• -

  We suggest that retraining under indication could help reduce the incidence of PI in situations of risk (2C) (Question P2)

• -

  Based on the available evidence, this Committee does not endorse programmed retraining for reducing the incidence of PI (2C) (Question P2)

• 1

  We recommend the administration of prophylaxis with systemic antibiotics immediately before insertion of the peritoneal catheter (1A)(Question P3)

• 2

  We suggest that each PD program should choose its own antibiotic prophylaxis regimen, considering the local incidences and antibiotic sensitivities (2D) (Question P3)

• 3

  We suggest vancomycin as a preferential option, though first or second generation cephalosporins may afford equivalent results (2B) (Question P3)

• 4

  We recommend that the risk of PI should not be a consideration when deciding which insertion technique or peritoneal catheter design to be used in a patient on PD (1A)(Question P3)

• 5

  We recommend regular monitoring and care of the exit site of the peritoneal catheter, if possible on a daily basis, as an indirect way to prevent PI (1C)(Question P4)

• 6

  We recommend early treatment of CRI, in order to prevent progression to PI (1C) (Question P4)

• 7

  We suggest the use of immobilization systems to avoid repeated traction upon the peritoneal catheter, at least during the post-implantation healing phase (2C) (Question P4)

• 8

  We suggest that each center should decide the type of soap or topical antiseptic for care of the catheter exit site, in accordance with the conditions of each patient (2C) (Question QP4)

• 9

  We suggest that each center should decide the type of care or dressing of the catheter orifice (open or closed), in accordance with the conditions of each patient, as there is no evidence generally supporting any specific type for preventing CRI and PI (2C) (Question P4)

• 10

  We recommend the administration of topical mupirocin or gentamycin during care of the exit site of the peritoneal catheter, as a way to reduce the risk of CRI and PI (1B)(Question P5)

• 11

  We suggest the convenience of periodic monitoring of the possible appearance of bacterial resistances to mupirocin or gentamycin among the bacterial flora (particularly Staphylococcus aureus) colonizing the nasal passages and the peri-catheter area (2B) (Question P5)

• 12

  This Committee provides no recommendation on the convenience of daily versus intermittent administration of the mentioned antimicrobials (ungraded) (Question Q5)

• 13

  This Committee considers that the evidence does not warrant the treatment of nasal carriers of Staphylococcus aureus to reduce the direct risk of PI, though treatment may prove to be of benefit through reduction of the risk of CRI (2D) (Question P5)

• 1

  We recommend that PD always should be performed in a clean environment, preferably with nearby access to running water (1C)(Question P6)

• 2

  We recommend correct hygiene before performing PD, with washing of the hands using common or antibacterial soap, followed by rubbing of the hands with a water-alcohol solution (1C)(Question P6)

• 3

  We recommend avoiding physical contact between domestic pets and the materials used for PD. In addition, patients should maximize asepsis and disinfection if coming into contact with pets (1C) (Question P6)

• 4

  This Committee suggests the convenience of wearing a facemask for exchange, particularly in patients with poor dental and periodontal health (2C) (Question P6)

• 1

  We recommend the use of systems with disconnection and purging before filling as the option of choice for connection in PD (1A)(Question P7)

• 2

  We suggest that double bag systems are preferable to the double-connection Y-set systems (2B)(Question P7) (little used in Spain)

• 3

  We recommend that the prevention of PI should not be included among the indications for treatment with APD (1B)(Question P8)

• 4

  We suggest that the use biocompatible solutions has no confirmed effect upon the risk of PI (2A)(Question P9)

• 1

  We recommend that patients should be specifically instructed during training in PD on how to act in the event of accidental disruption of asepsis (1C)(Question P10)

• 2

  We suggest that “wet” contamination or contamination of an uncertain nature should be dealt with by replacing the connection line and administering antibiotic coverage via the oral or parenteral (intravenous or intraperitoneal) route, and covering both grampositive and gramnegative organisms, with a minimum duration of two days (2C)(Question P10)

• 3

  We recommend that “dry” contamination at least should dealt with by replacing the connection line (2D)(Question P10)

• 4

  We recommend that patients on PD scheduled for invasive biliary, lower gastrointestinal or gynecological procedures should receive appropriate antibiotic prophylaxis (1C)(Question P11)

• 5

  We suggest that the provided antibiotic prophylaxis should be active against common enterobacteria and intestinal anaerobes (2C) (Question P11)

• 6

  We suggest that antibiotic prophylaxis is not necessary in endoscopic upper gastrointestinal tract explorations (2C) (Question P11)

• 7

  We suggest that the available information does not allow us to establish the convenience of voiding the peritoneal cavity for the abovementioned procedures (2C) (Question P11)

• 8

  We suggest that routine antibiotic prophylaxis be considered for dental procedures, especially those with a greater bacteremic potential (ungraded) (Question P11)

• 9

  We recommend that patients on PD administered oral or parenteral antibiotic treatment should receive oral antifungal prophylaxis during the treatment period (1B)(Question P12)

• 10

  We suggest that azoles may be the most appropriate choice for this purpose, though nystatin may be an adequate alternative (2C)(Question P12)

• 11

  We suggest the convenience of maintaining adequate patient plasma vitamin D levels (with the prescription of supplements if needed), in order to reduce the general risk of infection in patients subjected to PD (2D)(Question P13)

• 1

  We recommend that antibiotherapy be started as soon as possible when PI is suspected (1C)(Question T1)

• 2

  We recommend adequate coverage of grampositive and gramnegative microorganisms as part of the initial PI treatment protocol, in all cases (1C) (Question T1)

• 3

  We recommend adjusting the empirical treatment protocol for PI to the local epidemiology (1C) (Question T1)

• 4

  In the case of centers without sufficient experience of their own, we recommend vancomycin or a first-generation cephalosporin to cover grampositive microorganisms, and a third-generation cephalosporin or an aminoglycoside to cover gramnegative microorganisms (1B) (Question T1)

• 5

  We suggest limiting the use of aminoglycosides in patients with PI and significant residual kidney function (2C) (Question T1)

• 6

  We suggest the start of treatment for PI using antibiotic doses at the upper limit of the range recommended by the Summary of Product Characteristics, taking into account the residual kidney function of the patient (2C)

• 7

  We suggest the convenience of having alternative antibiotic protocols for patients with a history of allergies or intolerances to antibiotics, particularly betalactam drugs (2C) (Question T1)

• 1

  We recommend peritoneal lavage in the case of very turbid peritoneal drainage and/or intense abdominal pain, with the aim of providing pain relief and improving the mechanics of PD (1C)(Question T2)

• 2

  We suggest the addition of heparin to the dialysate in patients with a turbid effluent, in order to maintain good mechanics of PD (2C)(Question T2)

• 1

  We recommend the intraperitoneal route as the preferred option for the treatment of PI on an ambulatory basis, provided the level of training of the patient is adequate (1B)(Question T3)

• 2

  In hospitalized patients, we likewise suggest the intraperitoneal route as the preferred option, though the intravenous route is also considered adequate – the decision being conditioned to the pharmacokinetic properties of the antibiotic employed and the clinical context (2B)(Question T3)

• 3

  We recommend that the treatment of PI should not be based on monotherapy via the oral route (1C)(Question T3)

• 4

  We recommend that intraperitoneal aminoglycosides should be administered on an intermittent basis for the treatment of PI in PD (1B) (Question T4)

• 5

  We recommend that the intraperitoneal administration of vancomycin be prescribed on either a continuous or an intermittent basis (1B) (Question T4)

• 6

  We suggest that whenever possible, the plasma vancomycin and aminoglycoside levels should be monitored, especially in patients with residual kidney function (2B) (Question T4)

• 7

  We suggest that the intraperitoneal administration of betalactams and carbapenems be prescribed on either a continuous or an intermittent basis – the decision being conditioned by the normal dosing interval of each antibiotic (2C) (Question T4)

• 8

  In patients on APD with PI treated using intraperitoneal antibiotics, we recommend the administration of vancomycin and betalactams on a continuous basis in order to avoid subtherapeutic drug levels in the dialysate during the short nocturnal dwells (1B) (Question T5)

• 9

  When choosing an intermittent intravenous, intraperitoneal or oral antibiotic administration regimen in patients on APD, we recommend the application of general criteria similar to those used in CAPD (1B) (Question T5)

• 10

  In patients on APD with PI, we recommend that in the case of intermittent intraperitoneal administration, the latter should be carried out in the course of a long exchange, with a dwell of at least 6 h (1B) (Question T5)

• 11

  In patients with a nocturnal intermittent PD regimen (dry day) to be treated via the intraperitoneal route, we suggest the temporary introduction of a diurnal exchange to facilitate administration (intermittent or continuous) of the antimicrobial treatment (2D) (Question T5)

• 12

  We suggest that the practice of transferring patients on APD to CAPD during the treatment of PI via the intraperitoneal route is not justified by therapeutic reasons, though it may be justified on a practical basis (2D) (Question T5)

• 1

  We recommend a minimum duration of 14 days for the antibiotic treatment of PI in PD (1B) (Question T6)

• 2

  We suggest that it may be advisable to prolong treatment at least three weeks in PI due to Staphylococcus aureus, Corynebacterium spp., Enterococcus spp., gramnegative bacteria and mixed flora, with the participation of enteric organisms (2B) (Question T6)

• 3

  We suggest prolongation of the duration of treatment for at least one additional week when clinical criteria indicate an increased risk of a complicated course of PI (2C) (Question T6)

• 4

  We recommend the use of two synergic antibiotics for the treatment of PI due to Pseudomonas spp. and Stenotrophomonas spp. (1B) (Question T7)

• 5

  We recommend the use of antibiotic combinations for the treatment of polymicrobial PI, in relation to the isolated microorganisms (1B) (Question T7)

• 6

  In aggressive enterococcal PI treated with vancomycin or ampicillin, we suggest the addition of un intraperitoneal aminoglycoside, provided the antibiogram evidences a synergistic potential (2B) (Question T7)

• 1

  For PI due to coagulase-negative staphylococci, we recommend treatment with vancomycin or first or second generation cephalosporins during a minimum period of two weeks (1B) (Question T8)

• 2

  For the frequent recurrences or relapses of PI due to coagulase-negative staphylococci, this Committee suggests individualization of the management strategy. If peritoneal catheter replacement is decided, withdrawal and placement should be made simultaneously or with a brief time interval, once remission has been achieved (2B) (Question T8)

• 3

  We recommend maintaining the treatment of PI due to Staphylococcus aureus for at least three weeks (1C) (Question T8)

• 4

  We recommend thorough screening of simultaneous catheter-related infection in PI due to Staphylococcus aureus, and the consideration of withdrawal of the peritoneal catheter in PI due to Staphylococcus aureus associated to extensive soft tissue infection or wall abscesses caused by the same microorganism (1B) (Question T8)

• 5

  We recommend treatment of PI due to methicillin-sensitive strains of Staphylococcus aureus with a first generation cephalosporin or vancomycin, while methicillin-resistant strains should be treated with vancomycin or daptomycin (1C) (Question T8)

• 6

  We suggest the addition of an aminoglycoside in PI due to Staphylococcus aureus with a clinically aggressive presentation (2C) (Question T8)

• 7

  We suggest the treatment of streptococcal PI using vancomycin or betalactams (ampicillin or cephalosporin) for at least two weeks from the initially prescribed empirical protocol (2C) (Question T8)

• 8

  We recommend the treatment of enterococcal PI using ampicillin or vancomycin (plus an aminoglycoside in the case of synergism) for three weeks, with alternative antibiotic treatment according to susceptibility in the case of resistant strains (linezolid, daptomycin) (1C)

• 9

  We suggest the treatment of PI due to Corynebacterium with vancomycin during at least three weeks (2C) (Question T8)

• 10

  We recommend the treatment of PI due to Enterobacteriaceae during at least three weeks (1C) (Question T8)

• 11

  We suggest the use of synergic antibiotic combinations (cephalosporin and aminoglycoside) or carbapenem as treatment for PI due to Enterobacteriaceae with a clinically aggressive presentation (2C) (Question T8)

• 12

  We recommend the treatment of PI due to Pseudomonas spp. with two antibiotics exhibiting antipseudomonal activity during at least three weeks, with withdrawal of the peritoneal catheter in the case of concomitant infection of the tunnel or exit site (1B) (Question T8)

• 13

  We recommend the treatment of polymicrobial infections due to non-enteric species for two or three weeks, conditioned by the isolate of greatest aggressivity (1C)(Question T8)

• 14

  We recommend the treatment of polymicrobial PI with the isolation of enteric organisms using broad spectrum antibiotics covering grampositive species, enterobacteria and anaerobes, during at least three weeks (1C) (Question T8)

• 15

  We recommend abdominal imaging tests and/or surgical assessment in enteric polymicrobial PI, and in infections caused by Enterobacteriaceae or enterococci with a clinically aggressive, torpid or refractory behavior (1C) (Question T8)

• 16

  We recommend that PI with negative culture and rapid remission be regarded as grampositive infection, maintaining treatment for two weeks (1C) (Question T8)

• 17

  In PI with negative culture and a slow response to empirical treatment, we suggest switching to alternative therapy no more than beyond day 5, and to consider withdrawal of the peritoneal catheter (2C) (Question T8)

• 18

  We recommend withdrawal of the peritoneal catheter in PI due to yeasts (1C)(Question T9)

• 19

  We recommend treatment with antifungals according to antibiotic susceptibility during at least two weeks after withdrawal of the peritoneal catheter, in PI due to yeasts (1C)(Question T9)

• 20

  We suggest that an azole drug (in sensitive strains) or an echinocandin (alone or in combination until antimicrobial susceptibility has been established) offers a better efficacy-toxicity profile than amphotericin B, alone or combined with flucytosine, in PI due to yeasts (2D)(Question T9)

• 21

  We recommend combination therapy for the required duration in PI due to mycobacteria (1C)(Question T10)

• 22

  We recommend withdrawal of the peritoneal catheter in PI due to mycobacteria, except in clinically mild cases or infections with a rapid response to tuberculostatic drugs (1C)(Question T10)

• 23

  We suggest systematic screening of infection due to mycobacteria, in PI with negative culture and a torpid or refractory clinical course (2B)(Question T10)

• 1

  We recommend withdrawal of the peritoneal catheter in cases of totally refractory PI (1C) (Question T11)

• 2

  We suggest that in slowly resolving PI (decreasing activity on day 5 of treatment), the optimization of antimicrobial treatment should be decided before considering withdrawal of the catheter (2C) (Question T11)

• 3

  We recommend withdrawal of the peritoneal catheter in aggressive relapses of PI, or in second relapses of any PI (1C) (Question T11)

• 4

  We suggest prolonged and optimized antimicrobial treatment for the initial management of non-aggressive first relapses of PI (2C) (Question T11)

• 5

  We recommend investigation of the possible causes of recurrent PI, including poor adherence to the technique or occult abdominal conditions (1C) (Question T11)

• 6

  We suggest considering withdrawal of the peritoneal catheter within the strategy for dealing with recurrent PI (2C) (Question T11)

• 7

  We recommend withdrawal of the peritoneal catheter in PI associated to catheter-related infection produced by the same microorganism, in the following cases: refractory PI, slowly resolving PI, relapsing PI, PI associated to wall abscess, and PI associated to catheter-related infection characterized by slow resolution or recurrence (1C)(Question T12)

• 8

  We suggest considering withdrawal of the catheter in the rest of cases of simultaneous infection of the catheter produced by the same microorganism (2C)(Question T12)

• 9

  We suggest a period of at least two weeks between withdrawal of the catheter and the reinsertion of a new catheter, in PI associated to catheter-related infection produced by the same microorganism (2C)(Question T12)

• 10

  We recommend a period of at least two weeks between withdrawal of the catheter due to PI and the insertion of a new catheter (1C)(Question T13)

• 11

  We recommend that insertion of a new catheter should not be attempted if there are persistent signs of peritoneal irritation (1C)(Question T13)

• 12

  We suggest that simultaneous withdrawal and insertion of a peritoneal catheter in PI is feasible, provided the infection is in complete remission and is caused by non-aggressive microorganisms (2B)(Question T13).