Peritoneal Dialysis Guidelines 2019 Part 2: Main Text (Position paper of the Japanese Society for Dialysis Therapy)

Abstract

Background

This article is a duplicated publication from the Japanese version of “2019 JSDT Guidelines for Peritoneal Dialysis” with permission from the Japanese Society for Dialysis Therapy (JSDT). This clinical practice guideline (CPG) was developed primarily by the Working Group on Revision of Peritoneal Dialysis (PD) Guidelines of the Japanese Society for Dialysis Therapy. Recently, the definition and creation process for CPGs have become far more rigorous; traditional methods and formats no longer adhere to current standards. To improve the reliability of international transmission of our findings, CPGs are created in compliance with the methodologies developed by the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) working group. Part 2 of this PD guideline is the first CPG developed by our society that conforms to the GRADE approach.

Methods

Detailed processes were created in accordance with the Cochrane handbook and the GRADE approach developed by the GRADE working group.

Results

Clinical question (CQ)1: Is the use of renin-angiotensin system inhibitors (RAS inhibitors), such as angiotensin-converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB), effective in PD patients?

Recommendation: We suggest the usage of RAS inhibitors (ACEI and ARB) in PD patients (GRADE 2C).

CQ2: Icodextrin or glucose solution: which is more useful as a dialysate among patients with PD?

Recommendation: We suggest using icodextrin when managing body fluids in PD patients (GRADE 2C).

CQ3: Is it better to apply or not apply mupirocin/gentamicin ointment to the exit site?

Recommendation: We suggest not applying mupirocin/gentamicin ointment to the exit sites of PD patients (GRADE 2C).

CQ4: Which surgical approach is more desirable when a PD catheter is placed, open surgery or laparoscopic surgery?

No recommendation.

CQ5: Which administration route of antibiotics is better in PD patients with peritonitis, intravenous or intraperitoneal?

Recommendation: We suggest intraperitoneal administration of antibiotics in PD patients with peritonitis (GRADE 2C).

Note: The National Insurance does not currently cover intraperitoneal administration.

CQ6: Is peritoneal dialysis or hemodialysis better as the first renal replacement therapy in diabetic patients?

No recommendation.

Conclusions

In the future, we suggest that society members construct their own evidence to answer CQs not brought up in this guideline, and thereby show the achievements of Japan worldwide.

Background

This clinical practice guideline (CPG) was developed primarily by the Working Group on Revision of Peritoneal Dialysis Guidelines of the Japanese Society for Dialysis Therapy (JSDT). To date, the society has developed CPGs and their revised versions either alone or in concert with other organizations, which have been translated into English and disseminated worldwide. These guidelines have been formatted in the form of a textbook, and their content has been useful to dialysis therapists who comprise the members of our society. In part 1, we have included textbook-like information in line with the first edition of the peritoneal dialysis (PD) guidelines in 2009. However, the definition and creation process for CPGs have become far more rigorous, and traditional methods and formats no longer adhere to the recent current standards. Since 2012, the Guidelines and Procedures Creative Working Group (formerly the Guidelines Committee) has considered the methods for the creation of guidelines and has formulated a policy document discussing the methods to be used in the creation of guidelines [1]. Furthermore, to improve the reliability of international transmission of our findings, this policy document dictates that CPGs are to be created in compliance with the methodologies developed by the GRADE working group. Part 2 of this PD guideline is the first CPG developed by our society that conforms to the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach.

In the field of PD where evidence is often scarce, the development of a CPG in accordance with the GRADE approach is difficult. Furthermore, the little evidence that is rarely transmitted from within Japan raises doubts as to whether the guidelines developed on such a foundation can truly be said to be Japanese. Thus, there are several CQs that had to be removed due to an overwhelming lack of evidence. At present, we are proud that this CPG is based on the evidence we were able to gather, and was formulated through appropriate methodical analysis. While our recommendations are academically sound, some of the involved procedures were not listed in the insurance coverage. Hence, we would like to emphasize that these recommendations are merely academic advices. Furthermore, due to both lack of evidence and certainty, none of our recommendations were under the “strong” category. In other words, the society would like to make it very clear that the procedures not recommended are by no means unadvised or rejected. No two patients are alike, and yet, the recommendation categories were determined using current statistical methods that necessarily paint all patients with the same broad brush. Thus, as a clinical course, clinicians using these guidelines must reject our weak suggestions for methodologies not recommended after considering the unique pathologies of their individual patients. As seen in Section 3, clinicians regularly opt for non-recommended methodologies in the real clinical setting even when a strongly recommended course of action exists. When it comes to weakly suggested courses of action, a great deal of variance exists between cases with relatively similar methodologies. In other words, we would like to emphasize that the “recommendations” listed here are not mandates to be performed in all scenarios.

Additionally, because we received input that information highly desired by society members should be incorporated in the form of CQs, we collected challenge CQs and outside opinions to a certain extent. In particular, we received the following CQs: (i) whether hemodialysis (HD) or PD is better for patients with chronic renal failure who require dialysis, (ii) what type of PD treatment is best for these patients who select PD, (iii) what sorts of harm or burden exists for each of these treatment options, (iv) which has the best survival rate, and (v) whether or not the length of hospitalization or hospital visits should be reduced. Since the individual circumstances of patients can vary significantly, PD has already been established as a standard treatment worldwide, and studies comparing HD and PD monotherapy do not exist. We have judged these questions to be difficult to answer. We speculate that these sort of comparative trials are unlikely to be properly conducted in the future. In comparison to HD, in which a comparatively consistent level of treatment is maintained, for PD (as we discuss in Peritoneal Dialysis Outcomes and Practice Patterns Study (PDOPPS)), the incidence of peritonitis, treatment continuation rates, and dropout rates vary significantly from facility to facility. There is also the issue of which level of PD to be considered in comparison with HD. To improve the treatment level and achieve uniformity, we believe that discussion is necessary after the guidelines has been disseminated. After considering the matter among our working group, we have decided to limit to only one challenge CQ (CQ6) due to reasons including an overwhelming lack of evidence. Although we considered a total of 7 CQs, there was not enough evidence to answer CQ7. Hence, we included only six of them. In the future, we suggest that society members construct their own evidence and answer CQs not brought up in this iteration of the guidelines. Thereby, this may showcase the achievements of Japan worldwide.

Summary of recommendations

• CQ1: Is the use of renin-angiotensin system inhibitors (RAS inhibitors), such as angiotensin-converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB), effective in peritoneal dialysis patients? The conceptual diagram of clinical questions is depicted in Fig. 1.

  Recommendation: We suggest the usage of RAS inhibitors (ACEI and ARB) in PD patients (GRADE 2C).

• CQ2: Icodextrin or glucose solution: which is more useful as a dialysate among patients with PD?

  Recommendation: We suggest using icodextrin when managing body fluids in PD patients (GRADE 2C).

• CQ3: Is it better to apply or not apply mupirocin/gentamicin ointment to the exit site?

  Recommendation: We suggest not applying mupirocin/gentamicin ointment to the exit sites of PD patients (GRADE 2C).

• CQ4: Which surgical approach is more desirable when a PD catheter is placed, open surgery or laparoscopic surgery?

  No recommendation.

• CQ5: Which administration route of antibiotics is better in peritoneal dialysis patients with peritonitis, intravenous or intraperitoneal?

  Recommendation: We suggest intraperitoneal administration of antibiotics in PD patients with peritonitis (GRADE 2C).

  Note: The National Insurance does not currently cover intraperitoneal administration.

• CQ6: Is PD or HD better as the first renal replacement therapy in diabetic patients?

  No recommendation.

Methods

About our guideline creation approach

The definition of the CPG used herein and the overall construction of this document were in accordance with the CPG creation policy developed by the Japanese Society for Dialysis Therapy [1,2,3]. Detailed processes were created in accordance with the Cochrane handbook and the GRADE approach developed by the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) working group [2, 4, 5] (Tables 1 and 2).

Table 1 Grades and definitions of evidence quality in CPGs [1]

Table 2 Meanings of recommendation strengths for the guideline user [1]

Fig. 1

Conceptual diagram of clinical questions (CQs)

For Japanese terminology, we referred to the “Minds CPG Creation Procedures” created by the Minds (Japan Council for Quality Health Care, EBM Medical Information Division) [6, 7]. However, our creation methodology was developed in line with the GRADE approach.

Audience of this PD CPG

The target audience for this CPG is doctors who treat patients requiring PD therapy.

Results/recommendations

CQ1: Is the use of RAS inhibitors, such as ACEI and ARB, effective in peritoneal dialysis patients?

Recommendation

We suggest the use of RAS inhibitors (ACEI and ARB) in PD patients (GRADE 2C).

About this CQ

RAS inhibitors, such as ACEI and ARB, have renoprotective effects and reduce proteinuria. These are also drugs often used to treat chronic kidney disease. Previous studies have shown that preserving residual renal function in PD patients is beneficial to their prognosis. Herein, we examined whether or not the administration of RAS inhibitors to PD patients was effective.

While ACEI and ARB act via different mechanisms, we have grouped them together as RAS inhibitors for the purpose of this systematic review (SR) and have collectively considered them in SR1.1. The market share of ARB in Japan far outweighs that of ACEI. Hence, we determined that considering each separately was not reflective of the reality. Conversely, we have considered the matter in SR1.2 anticipating that we would receive questions as to whether ACEI or ARB are more effective for PD patients (Tables 3 and 4).

Table 3 Evidence profile: SR1.1. Comparison of ACEI and ARB with other drugs

Table 4 Evidence profile: SR1.2. Comparison of ACEI with ARB (effect of ARB to ACEI)

Basis of our judgment

We investigated SR1.1 and SR1.2; however, we were unable to determine which of ACEI and ARB was superior to the other for SR1.2. Consequently, we collectively treated ACEI and ARB as RAS inhibitors and have used SR1.1, which considers their effectiveness as the basis for our recommendation.

The certainty of evidence was assessed to be low critical importance, and the certainty of the evidence in general was also judged to be low.

Many of the cases involved RAS inhibitors being administered to patients prior to the introduction of PD. If one were to say that this pattern of administration is beneficial, nearly no cases would involve patient rejection of the drug. Additionally, previous studies have already clarified the relationship between residual renal function and the prognosis of PD patients. For this SR, we used urine volume as an index of residual renal function (a few studies that used eGFR as an index were also identified). Maintenance of urine volume is an appropriate index through which the patient can judge whether or not their residual renal function is being maintained. With regard to values (and intentions), we have determined that pressing issues of uncertainty or significant sources of variance most likely do not exist.

In terms of critical outcomes (i.e., total mortality, technical survival, and cardiovascular events), all studies had rather short observation periods of between 1 and 2 years in length. Consequently, the number of events was low, no differences were observed between the groups, and any effects of the interventions used remained unclear. Thus, we considered residual renal function including urine volume in our CPG panel. This was related to both patient prognosis and quality of life (QOL) as critical outcomes, and it could therefore be used in the judgment of recommendations. The intervention increased the mean difference of urine volume by 142.5 mL, and the lower limit of the 95% CI by 25.42 mL. Although there was no statistical difference, we assessed that the intervention was probably superior based on the result of the reduction of the relative risk of anuria to 0.70 [95% CI; 0.46–1.02]. Additionally, a study in subgroup analyses evaluating ACEI (Li 2003 [8], see Additional file 1: Appendix 6.1.2) had a low risk of bias. Despite the existence of only one study, a post-intervention mean difference in the GFR of 0.72 [0.52–0.90] allowed us to assess that an effect was present. With regard to ARB, we assessed that intervention with RAS inhibitors was superior based on our judgment that an effect existed for preserving urine volume and GFR.

In terms of cost, the patient need not pay for these interventions under the Japanese insurance system. From the perspective of rising healthcare costs, inexpensive generic alternatives for both ACEI and ARB have circulated widely throughout the market. We also do not believe cost to be much of an issue in this particular case.

CQ2: Icodextrin or glucose solution: which is more useful as a dialysate among patients with PD?

Recommendation

When it comes to the management of body fluid in PD patients, we suggest the use of icodextrin (GRADE 2C).

About this CQ

In the management of PD, peritoneal ultrafiltration volume and residual renal function are very important factors that relate to body fluid management and PD continuation period. The icodextrin dialysis solution does not use glucose as an osmotic agent. Because it is a high molecular weight glucose polymer, it is very difficult for it to be absorbed through the peritoneum. Therefore, it has superior ultrafiltration effects as a dialysis solution. In particular, in patients categorized as high transporter in peritoneal equilibrium tests, icodextrin dialysis solution was effective even in cases with poor clearance with the glucose dialysis solution. This was highly useful for fluid management associated with continuation of PD. Conversely, there were concerns about lowering residual renal function as well as the occurrence of rashes. In Japan, the icodextrin dialysis solution entered the market in 2003. Despite the 15 years that have passed since then, in the most current version of PD guidelines for Japan (published in 2009), it has not been discussed. Here, we have carried out a systematic review of the benefits and adverse effects of icodextrin dialysis solution vs. glucose dialysis solution monotherapy.

Basis of judgment

With regard to the effectiveness of icodextrin dialysate (interventional group) compared to the glucose dialysis solution monotherapy (control group), all-cause mortality had a risk ratio (RR) of 0.75 [0.33–1.71]. Among 1000 patients under study, the number of mortalities decreased by seven patients (18 patients decreased to 19 patients increased); thus, no significant difference was observed. However, icodextrin therapy had the superior point estimate. Next, with regard to episodes of uncontrolled fluid overload (judged to be an important outcome), the icodextrin group had an RR of 0.31 [0.12–0.82]. Among 1000 patients, the number of those with episodes decreased by 73, suggesting a significant superiority of the icodextrin group. While a significant difference was also not observed for peritoneal ultrafiltration volume, the icodextrin group had larger values. Ultimately, the icodextrin group had fewer withdrawals for technical survival, and an RR of 0.57 [0.29–1.12]. Among 1000 patients, the withdrawals decreased by 32 patients (9 patients increased to 52 patients decreased). Thus, the result itself was quite good while no significant difference was observed. Conversely, a reduction in urine volume and a loss of residual renal function were not observed with an increase in peritoneal ultrafiltration volume. In our CPG panel, it was pointed out that in this evidence, the outcome resulted in poor body fluid management. Furthermore, it would be better if we limited our recommendations to patients with poor body fluid management even though the evidence did not only target patients with poor body fluid management. Conversely, the opinion was also emerged that it may be possible to perform body fluid management with only highly concentrated glucose solutions. Ultimately, we decided to append “When it comes to the management of body fluid in PD patients” to our recommendation and to take care to prevent the incorporation of excessive use protocols such as ones in which the icodextrin dialysis solution was used for all patients.

Additionally, while peritoneal function was determined to be a critical outcome in our panel committee meeting, the importance of various outcomes was further discussed. Peritoneal function was judged to be next to the overall survival rate and technical survival, which was of comparatively low importance. Thus, this outcome was downgraded to an important status at the time of the CPG panel meeting.

With regard to adverse effects, peritonitis had an RR of 0.95 [0.79–1.15], and a difference was not observed between the interventional group and the control group. For rashes, the RR was 1.84 [0.48–7.09], and point estimates revealed that the intervention group had more occurrences, although a significant difference was not observed. However, most of these rashes were described in studies that were conducted soon after the icodextrin dialysis solution was introduced into the market (Wolfson 2002 [9], Finkelstein 2005 [10], see Additional file 1: Appendix 6.2.2). In comparatively newer reports (Lin 2009a [11], Chang 2016 [12], see Additional file 1: Appendix 6.2.2), these occurrences have clearly decreased in frequency, and they occurred at virtually the same rate as that associated with the glucose dialysis solution. A recent study by Baxter International showed that of 537 patients, in whom icodextrin dialysis solution was used, 27 (5%) developed rashes and eight (1.5%) developed exfoliative dermatitis. A response by Terumo indicated that since their product went on sale in December 2014, only one case of whole-body exfoliative dermatitis occurred in 2015, a decidedly small percentage. A case report by Nanamatsu et al. on icodextrin dialysis solution allergies also indicated that they occur most often between 7 and 10 days after beginning the treatment, and nearly all cases resolve quickly after stopping treatment [13]. From the above, we determined that the occurrence of undesirable effects due to the use of icodextrin dialysis solution is by no means common.

Aside from this, no variance in values was observed. Similarly, there is no difference between these treatments in terms of the cost borne by the patient under Japanese healthcare insurance, and we therefore believe the choice to have little financial impact.

Consequently, we have assessed that in terms of effect, the use of icodextrin dialysis solution is superior to that of glucose dialysis solution monotherapy (Table 5).

Table 5 Evidence profile

CQ3: Is it better to apply or not apply mupirocin/gentamicin ointment to the exit site?

Recommendation

We suggest not applying mupirocin/gentamicin ointment to the exit sites of PD patients (GRADE 2C).

About this CQ

Peritonitis is a common reason for early withdrawal from PD treatment. In order for the patient to continue their PD treatment in a stable manner, it is important to prevent peritonitis and exit site infection, which is the most common cause of peritonitis. In western countries, application of antibiotic ointment to the exit site is often recommended. However, neither mupirocin nor gentamicin ointment is covered by insurance in Japan. Existing guidelines and SRs on the effect of mupirocin ointment on exit site infections and peritonitis include investigations where application of mupirocin to the exit site has been grouped together with the application to the nasal cavity, researches on patients who carry Staphylococcus bacteria in their nasal cavity, and observational trials in addition to RCTs. In this CQ, we examined whether the application of mupirocin ointment or gentamicin ointment is superior to not applying any kind of antibiotic ointment to the exit site for preventing exit site infection. Additionally, we looked exclusively at RCTs that targeted all PD patients, who do and do not carry bacteria in their nasal cavity, to determine whether mupirocin or gentamicin is better at preventing infection.

Basis of judgment

Under the International Society for Peritoneal Dialysis (ISPD) recommendations for peritonitis, the 2016 version on prevention and treatment strongly recommends mupirocin administration because the systematic reviews listed in the ISPD recommendation also indicated that mupirocin was evidently very effective. In these systematic reviews, nasal cavity application and observational studies showed that mupirocin administration was highly effective (see Additional file 1: Appendix). However, the SR of the results of RCTs on exit site application in this guideline showed that the effect of mupirocin administration was far smaller than that evaluated in ISPD recommendations. In a comparison between applications of mupirocin ointment to the exit site and controls, mupirocin ointment was superior in terms of point estimates of all outcomes. Conversely, no significant differences were observed between the two drugs in any of the outcomes in a comparison of mupirocin and gentamicin ointments. At our panel meetings, these results were highlighted as evidence that implementation of the antibiotics application in Japan may improve treatment results.

After collecting a great deal of evidence about the adverse effects of these treatments, the issue of the emergence of resistant bacteria due to long-term usage of antibiotic ointments arose in our CPG panel meetings. To date, the mupirocin resistance rates of MRSA following mupirocin ointment use have been reported at 2.2 to 81%. Additionally, gentamicin resistance rates of 17.7 to 70.9% have already been observed in skin and soft tissue infections. Thus, long-term use of either of these treatments results in fear of the appearance or increase of resistant strains.

The PD peritonitis incidence in Japan is 0.20 events/patient-year [14], lower than the figure cited in the ISPD recommendations (0.50). Even if application of mupirocin or gentamicin ointment is effective in western countries, current exit site management protocols in use in Japan which do not involve the application of antibiotic ointments may even be said to be superior than those in many other countries, leading us to believe that there is not much benefit to be gained from their application. Furthermore, considering the cost and issue of antibiotic resistance, we have decided on a recommendation that suggests that mupirocin/gentamicin ointment not be applied. Finally, preventative application of mupirocin and gentamicin ointment to the exit site is not covered by insurance (Tables 6 and 7).

Table 6 Evidence profile: SR3.1. Comparison of mupirocin ointment with a control treatment

Table 7 Evidence profile: SR3.2. Comparison of mupirocin and gentamicin ointments

CQ4: When a PD catheter is placed, which surgical approach is desirable, open surgery or laparoscopic surgery?

Recommendation

No recommendation.

About this CQ

The performance of an appropriate catheter placement surgery is a prerequisite for the smooth implementation of PD. There are three methods used to insert a catheter: open surgery, laparoscopic surgery, and percutaneous insertion. There are no available devices for percutaneous catheter insertion in Japan, so either open or laparoscopic surgery is practiced in most facilities. Laparoscopic procedures generally require a smaller incision in the skin than open procedures, thereby leaving smaller scar, and are also less painful than open surgeries after the procedure. Its most significant benefits are shorter postoperative recovery periods and hospital stays and that patients can return to their daily activities quicker. Additional merits of laparoscopic surgery include the following: the surgeon can insert the catheter while directly observing its tip as it goes in; procedures to prevent catheter coiling, like omentopexies and omentectomies, can be performed; the ability to prevent catheter migration by securing the catheter inside the abdominal cavity; additional procedures, such as adhesiolysis for patients with abdominal cavity adhesion, can be performed; and hernia repair surgeries can be simultaneously performed on patients with inguinal hernias. However, the increased risk of complications unique to laparoscopic surgery include increased cost due to a need for general anesthesia, specialized equipment and particular facilities, the need for a well-versed surgeon in laparoscopic procedures, organ damage that may occur during trocar insertion, bleeding that may occur on removal, subcutaneous emphysema or gas embolism/pulmonary embolism associated with pneumoperitoneum, and organ damage due to limited vision. While several studies have compared the postoperative outcome of open and laparoscopic surgery, the total number of cases investigated is rather low, and there is not enough evidence as to which of the two methods is superior. Thus, we selected this CQ since we noticed that there is a need to clarify which catheter insertion surgery is better.

Basis of judgment

The survival rate for laparoscopic surgery compared to open surgery had an RR of 1.07 [0.98–1.16] (low evidence certainty). Additionally, the relative risk of catheter migration in laparoscopic surgery compared to the open approach was 0.42 [0.18–0.96] (moderate evidence certainty). The RR was 0.99 [0.43–2.29] for exit site and tunnel infections (low evidence certainty). Because the only analysis in which a statistically significant difference was found included studies in which the catheter was anchored inside the abdominal cavity, we carried out a subgroup analysis on the presence or absence of catheter anchoring. Significantly fewer incidences of catheter migration occurred in laparoscopic procedures with catheter anchoring than in open procedures (RR 0.09 [0.01–0.68]). However, we found no significant difference between laparoscopic procedures without catheter anchoring and open procedures (RR 0.58 [0.27–1.25]). Furthermore, only pain was evaluated with respect to QOL, and the certainty of that evidence was very low. From the above, we judged that there are some, albeit small, desirable effects associated with use of laparoscopic surgery. Regarding hospital stay duration, we found that open surgery was associated with shorter stays than laparoscopic surgery, but only two studies discussed this, and one of them, whose weighting was 91% in the meta-analysis (Wright 1999 [16], see Additional file 1: Appendix 6.4.2), indicated an average hospitalization time of 2.4 days for open surgery and 3.1 days for laparoscopic surgery. These results differ greatly from those seen in Japan, and we decided not to use them as a basis for our judgment.

Conversely, in terms of adverse effects, early complications (presence of leaks) had an RR of 0.86 [0.25–3.02], and late complications (hernia) had an RR of 0.75 [0.21–2.72]. Additionally, while laparoscopic surgery requires general anesthesia, we considered it to be associated with nearly no adverse effects, and therefore, the undesirable effects arising from it were judged to be quite small.

Regarding patient preference for this CQ, because open surgery can be performed under local or spinal anesthesia while laparoscopic surgery requires general anesthesia, the anesthesia-associated risk is higher for the latter procedure but it also involves less pain. However, because open procedures are also minimally invasive, it was pointed out in our CPG panel meetings that the merits of laparoscopic surgery are also quite small. As for cost, laparoscopic surgery for the purpose of PD catheter surgery is not covered by insurance in Japan. The cost of laparoscope use is borne by the facility in which it is used.

As indicated from our CPG panel meetings, the balance between the benefits and adverse effects of laparoscopic surgery is rather uncertain, and given the fact that in Japan, circumstances can differ greatly from facility to facility; we determined that a proactive recommendation or disrecommendation of laparoscopic surgery will only increase confusion in the clinical setting (Table 8). Thus, we have decided to put forth no recommendation on this issue.

Table 8 Evidence profile

CQ5: Which administration route of antibiotics is better in PD patients with peritonitis, intravenous or intraperitoneal?

Recommendation

We suggest intraperitoneal administration of antibiotics in PD patients who have developed peritonitis (GRADE 2C).

Note: The national insurance does not currently cover intraperitoneal administration.

About this CQ

The main objectives of this CQ are (1) to clarify the superiority or inferiority of intraperitoneal administration in comparison with intravenous administration and (2) to clarify whether continuous or intermittent intraperitoneal administration is better.

However, after conducting a literature search in accordance with a process that is discussed below, we found that the two RCTs that were suitable for analysis contained data that could not present a valid answer to the second question above. Specifically, one of the papers [17] used both vancomycin (VCM) and tobramycin (TOB) in both groups, and compared intraperitoneal continuous administration (VCM 20 mg/L + TOB 4 mg/L mixed into each PD bag) and daily administration (VCM 0.5 or 1.0 g + TOB 1.0 mg/kg BW given as intravenous injection once daily). The other study [11] compared intraperitoneal and intravenous initial VCM administration (loading dose; 1 g/body); after initial treatment in both groups, the patients received intraperitoneal VCM (25 mg/L). Based on these reasons, we decided that answering second question of CQ5 was not possible.

Both groups received the same intraperitoneal VCM regimen after initial dosage in the latter study [18]. Therefore, it may be thought that the control group cannot be considered to have undergone “intravenous administration.” However, we judged that increasing the number of patients covered by our analysis was of great benefit here, so we incorporated this study and attempted to answer first question of CQ5.

Basis of judgment

Despite the lack of a statistically significant difference in PD withdrawal, based on the fact that intraperitoneal administration caused a decrease of 130 patients/1000 patients (430 decreased to 170 increased), and given that there are no reports that indicate that it has more adverse effects than intravenous administration, we judged the intervention (intraperitoneal ) to be superior (Table 9).

Table 9 Evidence profile (effects of intravenous administration in comparison to intraperitoneal administration)

However, when the lack of insurance coverage for intraperitoneal administration is taken into account, as well as the fact that the benefits of antibiotics other than VCM have not been clarified, we decided that at present, it is prudent to issue this recommendation with certain conditions attached. Additionally, many facilities are now already practicing intraperitoneal administration of antibiotics following the ISPD guidelines. For this reason, we decided to weakly recommend the procedure instead of not recommending it. Of course, this recommendation does not negate or reject intravenous administration. At our CPG panel meetings, an opinion was put forth that the addition of insurance coverage for intraperitoneal administration is necessary, especially in the case of future at-home PD treatment.

CQ6: Is PD or HD better as the first renal replacement therapy in diabetic patients?

Recommendation

No recommendation.

About this CQ

While there are many studies that consider the relative merits of HD and PD in patients with diabetes, almost all of them are observational studies, and various results have been reported. The only one RCT [19] targets a population with a mixture of diabetic and non-diabetic patients. There are no analysis of diabetes patients alone. Consequently, we had to carry out our SR of diabetes patients using observational studies.

Basis of judgment

Because these studies are all observational, both the risk of bias and the possibility that adjustment for confounding factors was insufficient are serious. For this reason, the quality of the evidence was downgraded to “very low.” Of the papers we considered here, many cited a low mortality for HD (Table 10). It was pointed out that at the very least, there is no evidence actively supporting PD. In addition, no studies have compared urine volume, residual renal function, and glycemic control. All the studies examined here were unclear regarding the details of their treatment conditions (type of dialysis solution, etc.) and research subjects (age, etc.). In particular, the treatment methods of HD varied greatly. Thus, we ultimately determined in our CPG panel meetings that we are unable to issue a recommendation on this matter.

Table 10 Evidence profile