Advances in Peritoneal Dialysis, Vol. 18, 2002

Kinetic Analysis ofIcodextrin Interference withSerum Amylase Assays

Run Wang,1 Valerie Leesch,2 Paul Turner,2 James B.Moberly,1 Leo Martis1

From: 1Baxter Healthcare Corporation, Renal Division,McGaw Park, and 2Technology Resources, Round Lake,Illinois, U.S.A.

Patients treated with Extraneal peritoneal dialysissolution (Baxter Healthcare Corporation, Deerfield,IL, U.S.A.) have a significant decrease in serum amy-lase activity. The decline is reported to be due to in-terference of icodextrin in a routinely used laboratoryassay. The present study was designed to investigatethe kinetics of icodextrin interference in the amylaseactivity assay and to determine whether assay inter-ference can account for the total decline in amylaseactivity.

Plasma obtained from healthy volunteers wasspiked with 0, 0.21, 0.71, and 3.6˚mg/mL icodextrin.Amylase activity was determined using Sigma kit577-10 (Sigma Diagnostics, St.˚Louis, MO, U.S.A.).Amylase activity in plasma samples spiked with3.6˚mg/mL icodextrin was also monitored while vary-ing the concentration of the substrate (ET-G7-PNP)from the assay˚kit.

Amylase activity decreased with increasingamounts of icodextrin and decreasing amounts of as-say substrate. A 72.6% decrease in amylase activitywas observed in samples spiked with 3.6˚mg/mLicodextrin as compared with samples withouticodextrin at a substrate level similar to that in theassay kit (0.71˚mmol/L). Double reciprocal and Dixonplots indicate competitive inhibition of amylase ac-tivity by icodextrin.

Icodextrin functions as a competitive inhibitor inthe assay for amylase activity, as predicted by thestructural similarities between icodextrin and theamylase assay substrate. The degree of icodextrin in-terference suggests that the entire decline in amylaseactivity observed in patients using Extraneal can beaccounted for qualitatively by icodextrin interference.The amylase activity decline in patients treated withExtraneal is an artifact attributable to assayinterference.

Key wordsIcodextrin, amylase, polyglucose, Extraneal, assayinterference

IntroductionAmylase, an enzyme produced by the salivary glands,pancreas, and other human tissues, digests polysac-charides specifically at internal α-1,4˚glucosidic bonds(1). Amylase typically initiates the hydrolysis ofpolysaccharides in the mouth as carbohydrates areingested, and it further degrades the oligosaccharidesin the digestive tract to maltose (G2), maltotriose (G3),and other small oligosaccharides. Both the pancreaticand the salivary isoforms of the enzyme are found inthe circulation as serum α-amylases. Although thefunction of circulating amylase is not well understood,measurement of serum amylase activity is useful as adiagnostic tool, because markedly elevated serumamylase levels are associated with pancreatitis˚(2).

Administration of Extraneal peritoneal dialysissolution (Baxter Healthcare Corporation, Deerfield,IL, U.S.A.) results in a significant decrease in mea-sured serum amylase activity (3—7). Levels of serumamylase activity decline 70%˚— 90% within 1˚weekof Extraneal administration and remain low (butstable) during continued administration. Upon discon-tinuing Extraneal, serum amylase levels return tobaseline level. No adverse clinical effect has been at-tributed to the decline in measured serum amylaseactivity.

Previous reports (4,7) attribute the decline in se-rum amylase activity to assay interference caused bythe presence of icodextrin, or its metabolites, or both,in the serum samples. The current study was designedto investigate the kinetics of that assay interferenceand to determine whether assay interference can ac-count for the total decline in amylase activity.

Materials and methodsBlood was collected from volunteers and allowed tostand at room temperature for 1˚hour to coagulate.

97

Serum was obtained by centrifugation at 3950g for30˚minutes at 4°C. Serum samples were stored frozenat —80°C until use. An amylase assay reagent kit(no.˚577—10: Sigma Diagnostics, St.˚Louis, MO,U.S.A.), employing the substrate 4,6-ethylidene(G7)-p-nitrophenyl(G1)-α,D-maltoheptaside (ET-G7-PNP),was reconstituted with distilled deionized water inthree different proportions to obtain three levels ofsubstrate concentration. Upon mixing with serum andicodextrin, the concentrations of substrate in the finalassay samples were 0.55˚mmol/L (low), 0.71˚mmol/L(medium), and 1.43˚mmol/L (high). The medium sub-strate level was similar to that indicated in the instruc-tions for the Sigma assay˚kit.

Icodextrin (ML Laboratories, Birchwood, U.K.)was prepared as a stock solution (100˚mg/mL) in dis-tilled deionized water. Icodextrin was spiked intoserum samples to yield final concentrations of 0, 0.21,0.71, or 3.6˚mg/mL, where the highest level was simi-lar to levels of icodextrin (minus the primary metabo-lites maltose, maltotriose, and maltotetraose)typically found in plasma samples from patients usingExtraneal. Amylase reagent, icodextrin stock solution,serum and distilled deionized water were combinedin the wells of a 96-well plate (#3635: Corning Costar,Corning, NY, U.S.A.) and mixed gently. Absorbancemeasurements were recorded at 405˚nm and 750˚nmevery 20˚seconds for 20˚minutes using a Spectra-max˚250 plate reader (Molecular Devices Corp.,Sunnyvale, CA, U.S.A). Absorbance at 750˚nm wassubtracted from that at 405˚nm to correct for effectssuch as solution turbidity. The plate reader was equili-brated at 37°C for all assays, and all assays were con-ducted on the same˚day.

Results and discussionFigure˚1 contains representative data showing absor-bance as a function of enzymatic reaction time. Theeffects of varying concentrations of icodextrin (0, 0.21,0.71, and 3.6˚mg/mL) are shown for a single substrateconcentration (0.71˚mmol/L).

Following a brief lag phase, absorbance increasedlinearly. The slope of the response in the linear por-tion of the absorbance—time curve was used to deter-mine the enzymatic reaction rate. Similar data wereobtained at ET-G7-PNP substrate concentrations of0.55˚mmol/L and 1.43˚mmol/L.

Using those data, the effects of icodextrin and sub-strate concentrations on serum amylase activity were

Wang et al

calculated (Table˚I). For the medium substrate con-centration (0.71˚mmol/L), amylase activity decreasedby 72.6% in the presence of 3.6˚mg/mL icodextrin ascompared with the absence of icodextrin. For the highsubstrate concentration (1.43˚mmol/L), the decreasein amylase activity due to 3.6˚mg/mL icodextrin wasless (58.7%). Analyses of the relationship of the en-zymatic rate—1 versus the substrate concentration—1

(double reciprocal plot) and of the enzymatic rate—1

versus the icodextrin concentration (Dixon plot) indi-cate that the inhibition of amylase activity due toicodextrin is competitive [that is, the intercept for thedouble reciprocal plot is near zero, indicating com-petitive inhibition (Figure˚2)]. Those results accordwith previous results reported by Schoenicke et al (7)and Grzegorzewska et al˚(4).

Interference with the amylase activity assay byicodextrin is also expected because icodextrin and thesubstrates commonly used in the amylase activity as-say both have identical 1,4-α-D-glucosidic bonds. Asillustrated in Figure˚3, amylase hydrolyzes 1,4-α-D-glucosidic bonds in the substrate (ET-G7-PNP) to formG2-, G3-, and G4-PNP fragments. The α-glucosidasein the assay kit further hydrolyzes G2-PNP andG3-PNP to yield glucose and p-nitrophenol, the latterproducing a yellow color that can be measured at405˚nm. Increase in the absorbance at 405˚nm is di-rectly proportional to amylase hydrolytic activity inthe sample. When icodextrin is present, amylase hy-drolyzes 1,4-α-D-glucosidic bonds in both the sub-strate and icodextrin, leading to direct competition foramylase hydrolysis activity. Therefore, the amount ofamylase activity in the sample appears to be less thanif icodextrin were not present in the assay system.

Caution should be used in the interpretation ofassay results using icodextrin spiked into serum orplasma samples. Pre-incubation of the sample mayreduce the amount of icodextrin owing to its hydroly-sis by amylase, thereby decreasing the degree oficodextrin interference. Our experimental results showthat approximately 25% of icodextrin was hydrolyzedto G2, G3, and G4 after 2˚hours incubation at 37°C inplasma. As expected, the reduction in amylase activ-ity was found to be 72% at the time of mixing and57% and 46% after 1 and 2˚hours of incubation, re-spectively, for the same spiked icodextrin concentra-tion. In addition, measured amylase activity may varyif the substrate concentration in the assay kit used ishigher or lower, because the degree of icodextrin in-

98

terference depends on substrate concentration. Forexample, the assay kit produced by BoehringerMannheim (catalog no.˚1555685: BoehringerMannheim, Mannheim, Germany) has a substrate levelof approximately 3.55˚mmol/L, which is much higherthan that of the Sigma kit. As a result, the degree oficodextrin interference may be much lower or unob-servable when the Boehringer assay kit is used.

ConclusionIcodextrin functions as a competitive inhibitor in theassay for amylase activity, as predicted by the struc-tural similarities between icodextrin and the amylase

assay substrate. The degree of icodextrin interferencesuggests that the entire decline in amylase activityobserved from patients using Extraneal can be ac-counted for by icodextrin interference. Therefore, theamylase activity decline in patients treated withExtraneal is an artifact due to assay interference.

Icodextrin Interference in Amylase Assay

FIGURE 1 Representative data showing absorbance as a function of enzymatic reaction time. Each assay was performed in duplicate.mM˚= mmol/L.

TABLE I Effects of icodextrin and substrate concentrations onserum amylase activity

Icodextrina % Dropb Substratec % Dropb

concentration amylase concentration amylase(mg/mL) activity (mmol/L) activity

0.21 8.2% 0.51 76.8%0.71 34.4% 0.71 72.6%3.6 72.6% 1.43 58.7%

a Substrate concentration was 0.71˚mmol/L for the assays withicodextrin concentrations at 0, 0.21, 0.71, and 3.6˚mg/mL.b The percentage activity drop was calculated as compared withsamples without icodextrin.c The icodextrin concentration was 3.6˚mg/mL.

FIGURE 2 Double reciprocal and Dixon plots. mM˚= mmol/L.

99

References1 Pieper—Bigelow C, Strocchi A, Levitt MD. Where

does serum amylase come from and where does itgo? Gastroenterol Clin North Am 1990; 19:793—810.

2 Salt WB, Schenker S. Amylase its clinical signifi-cance: a review of the literature. Medicine 1976; 55:269—89.

3 Mistry CD. Glucose polymer as an osmotic agent in

FIGURE 3 Schematic showing the principle of the amylase activityassay and the representative structure of icodextrin.

continuous ambulatory peritoneal dialysis [MDThesis]. London: University of London; 1988:1—300.

4 Grzegorzewska AE, Antczak—Jedrzejczak D, MariakI. Polyglucose dialysis solution influences serumactivity of amylase and of lipase differently. AdvPerit Dial 2000; 16:113—18.

5 Bajo MA, Selgas R, Hevia C, et al. Icodextrin 7.5%dialysate for diurnal exchange in patients treated withCCPD [Abstract]. Perit Dial Int 1999; 19(Suppl 1):S49.

6 Mistry CD, Gokal R. Single daily overnight (12-hdwell) use of 7.5% glucose polymer (Mw 18700; Mn7300) +0.35% glucose solution: a 3-month study.Nephrol Dial Transplant 1993; 8:443—7.

7 Schoenicke G, Grabensee B, Plum J. Interference oficodextrin with serum amylase activity measurement[Abstract]. J˚Am Soc Nephrol 1999; 10:A1163.

Corresponding author:Run Wang, PHD, Advanced Research and Development,Renal Division, Baxter Healthcare Corporation,1620˚Waukegan Road, MPGR-E2, McGaw Park, Illinois60085 U.S.A.

Wang et al

Advances in Peritoneal Dialysis, Vol. 18, 2002

Declining Trend ofPeritoneal Dialysis:A˚Single-Center Experience

From: University of Texas Medical Branch, Galveston,Texas, U.S.A.

Mahendra Agraharkar, Sharon Henry, Dora Martinez,Becky Bonds

Peritoneal dialysis (PD), despite being advantageousto patient, physician, and society, has failed to showthe growth it deserves. On the contrary, PD utiliza-tion has declined. Over the past several years, we havenoticed a decline in the number of our home dialysispatients. When compared to the national trend, wefind our trend to be not significantly different fromother centers across the country. A similar trend hasalso been noticed in Canada. Although several rea-sons may exist for the decline, we intend to concen-trate on local factors.

In the first quarter of 1996, we had a total of46˚adult and pediatric end-stage renal disease (ESRD)patients on PD. That number decreased to 23 at theend of fourth quarter of the year 2001. The losses inour program far exceeded the gains. We lost our pa-tients mainly to in-center hemodialysis (ICHD) andto transplantation. Peritonitis and membrane failureremained the major grounds for the loss to ICHD.

In our center, geographic location and a lack ofstructured pre-ESRD education probably played amajor role in the decline. Many of our patients arefrom distant counties that have a contract with Uni-versity of Texas Medical Branch for providing healthcare to their indigent population. However, once thosepatients develop complications, the counties rely onthe expertise of local physicians and nephrologists.

Key wordsHome dialysis, declining trend, in-center hemodialysis

IntroductionSince peritoneal dialysis (PD) was made available asan alternative to hemodialysis (HD), the PD modalityhas failed to show the growth it deserves, especiallyconsidering that the survival rate and the advantagesto patients and nephrologists favor PD (Figure˚1). Thetrend extends worldwide; it is not limited to North

America (1—3). Although PD utilization is decreasingworldwide, the United States happens to be amongthe countries that use the PD modality the least˚(4).

As PD utilization continues to decrease worldwide,we notice a similar trend at the University of TexasMedical Branch (UTMB). Our institution has a con-tract with several counties in the state of Texas forproviding health care to their indigent populations.For that reason, many patients in our program comefrom distant areas; unfortunately, in most cases, theyare not from the adjacent counties. The adjacent coun-ties are catered to by the institutions and nephrologistsfrom Houston metropolis.

The better way to provide quality health care todistant patients is to offer home dialysis. For that rea-son, UTMB had one of the largest home dialysis pro-grams. Over the past several years, the home dialysisprogram has seen a steep decline in patients on PD.We decided to study the reasons for that decline.

Patients and methodsPatients added to our program were automaticallyentered into the Baxter PD Census (BPDC), BaxterHealthcare Corporation, Deerfield, IL, U.S.A. We re-trieved the data for our institution from the BPDC andperformed a comparison with the data from the total

FIGURE 1 The point prevalence of end-stage renal disease patientsin the United States by modality, depicting the trend during the1990s. The x˚axis represents the years and the y˚axis shows theabsolute count (∞1000).

Num

ber

of p

atie

nts

(tho

usan

ds)

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

USRDS 2000

of patients (thousands)

HemodialysisPeritoneal dialysisTransplant

250

200

150

100

50

0

HemodialysisPeritoneal dialysisTransplant

101

BPDC. The BPDC data is gathered quarterly and in-cludes both adult and pediatric end-stage renal dis-ease (ESRD) patients. For the present study, werecorded total additions and losses to the program.Among the losses, losses to in-center hemodialysis(ICHD), losses due to death or transplant, and lossesdue to other causes were recorded. Patients lost toICHD were further grouped into loss due to peritoni-tis, catheter loss due to infection or other catheter-related problems, membrane failure or failure to meetadequacy criteria, psychosocial problems, and othermedical problems. The adult and pediatric populationswere later separately analyzed.

ResultsAt the beginning of the year 1996, we had a total of46 PD patients. That number decreased to 21 at theend of the fourth quarter of the year 2001, a decre-ment of more than 50% (Figure˚2). At the end of the4th quarter of the year 1996, the gross gain was 33˚pa-tients against the loss of 35˚patients. The trend con-tinued into the year 1997, with a total gain of18˚patients and a total loss of 22. In the years 1998,1999, 2000, and 2001, the gains were 15, 14, 13, and12˚patients respectively, and the losses were 21, 17,13, and 9˚patients respectively (Figure˚3).

Among the losses, 48% were lost to ICHD; 36%,to transplant; 9%, to death; and 7%, to miscellaneouscauses [Figure˚4(A)]. Among the adult PD patients,67% were lost to ICHD as compared with 29% amongthe pediatric patients. Renal transplantation was per-formed in 19% of the adult PD patients as comparedwith 64% of the pediatric patients. Of the adult PDpatients, 11% died while receiving PD; only 2.4% ofthe patients in the pediatric PD program died[Figure˚4(B)]. In comparing our figures to the overallBPDC [Figure˚4(C)], we noted that, although we lostmore patients to ICHD, the difference was marginal(48% at UTMB vs. 41% in BPDC). More of our pa-tients underwent transplantation, and that differencewas significant (36% at UTMB vs. 15% in BPDC).

Among the UTMB patients lost to ICHD, the in-dications for the switch were mainly peritonitis (37%)and the need to provide adequate dialysis as per theDialysis Outcomes Quality Initiative (DOQI) guide-lines (34%) [Figure˚5(A)]. In comparison, the BPDCdata revealed lower percentages for peritonitis (23%)and failure meet adequacy criteria (21%) as the rea-son for switching to ICHD [Figure˚5(B)].

Agraharkar et al

DiscussionEarlier studies had revealed that the survival rate issimilar among patients on HD and PD, and some stud-ies even claimed better survival on HD (5—10). How-ever, recent studies have clearly established a survivaladvantage for patients on PD as compared with pa-tients receiving ICHD (11,12). Patients on PD alsoseem to have an enhanced sense of satisfaction whencompared with ICHD patients, which suggests thatPD provides a better quality of life (13—15). Yet, de-spite the advantages, use of the PD modality is de-clining. The probable factors for the decline are mainlythe proliferation of HD facilities, the nature of the in-cident ESRD patients (older and frailer), and miscon-ceptions about the ability of those incident patients tobe maintained on PD for long durations (16,17).

The UTMB is located on Galveston Island, whichis situated approximately 40˚miles southeast of themetropolitan city of Houston (Figure˚6). The popula-tion of the island of Galveston is approximately60,000. The adjacent counties to the north of GalvestonCounty on the mainland form the southern part of thegreater Houston area, which has several university andmajor hospital—based nephrology groups and manypracticing nephrologists. However, many countiessurrounding the Houston metropolitan area have acontract with UTMB to provide medical care to theindigent populations of those counties.

Patients from the indigent population either con-sult a physician too late or (more commonly) presentto the nearest hospital when in need of emergent di-alysis and are then transferred to UTMB for manage-ment of renal failure. That situation perpetually resultsin initiation of HD and subsequent placement in anoutpatient dialysis facility closer to their residence. Ifsuch patients are referred early, they are offered di-alysis education and a choice to select the dialysismodality. If they decide to select home dialysis, theyare offered assistance for their stay on Galveston Is-land for the duration of the training period. Most pa-tients who opt for PD are not hesitant to travel longdistance for their monthly visits.

Loss to ICHDAs mentioned earlier, our center provides health careto several distant counties. The patients are indigent,and once they develop complications, some of themtend to go to the nearest hospital. Because of the dis-tance involved, many delay going to hospital; but, as

102

the severity of their complications increases, theypresent to the nearest hospital. Not all of the hospi-tals may have nephrologists who are comfortable intreating PD patients especially when those patientshave complications related to PD. We had to switchsome of our patients to ICHD when the criteria foradequacy of peritoneal dialysis could not be met. Our

loss to ICHD was therefore probably disproportion-ately high.

Loss to transplantationThe transplantation rate among our PD patients is quitehigh. As cited earlier, our loss to transplantation is36%, which is more than twice the BPDC rate of 15%.

Peritoneal Dialysis Experience at UTMB

FIGURE 2 The point prevalence of end-stage renal disease patients treated by peritoneal dialysis at the University of Texas MedicalBranch. The lower segment of each bar represents the pediatric patients and the upper segment represents the adult patients. The x˚axisrepresents quarters of years, beginning with the first quarter of the year 1996 and ending with the last quarter of the year 2001. The y˚axisrepresents the absolute number of patients.

1996 1997 1998 1999 2000 2001

103

We can only speculate on the reason: perhaps the pa-tients appear so healthy that they go on the list morequickly and receive a transplant graft faster. A largeproportion are pediatric patients, who have a highertransplantation rate of 64% [Figure˚4(B)].

Failure to meet adequacy criteriaWe have noted that patients on automated peritonealdialysis (APD) have a tendency to avoid manual ex-

FIGURE 3 Additions to, and losses from, the peritoneal dialysisprogram at the University of Texas Medical Branch. The dottedline represents the number of patients exiting the program. Thex˚axis represents quarters of years, beginning with the first quarterof the year 1996 and ending with the last quarter of the year 2001.The y˚axis represents the absolute number of patients.

FIGURE 4 (A)˚The percentage of patients exiting the peritoneal dialysis program at the University of Texas Medical Branch (UTMB) forvarious reasons. (B)˚The number of patients exiting the pediatric peritoneal dialysis program at UTMB for various reasons. (C)˚Thepercentage of patients exiting the Baxter national peritoneal dialysis census for various reasons.

A C

TRANSPLANT

36%

B

TRANSPLANT

15%

Agraharkar et al

FIGURE 5 (A)˚The reasons for switching from the peritoneal di-alysis program at the University of Texas Medical Branch to in-center hemodialysis. The numbers are percentages. (B)˚The reasonsfor switching from a peritoneal dialysis program to in-center he-modialysis per the Baxter national peritoneal dialysis census. Thenumbers are percentages.

OtherMedicalLosses

10%

Other MedicalLosses, 7%

A

B

104

changes. We noted that, the more manual exchangesadded, the lower the satisfaction (and hence compli-ance) among the patients especially among thosewho are employed.

Inadequate gainsOur center had lacked a structured pre-ESRD educa-tion program. Studies have shown that structured pre-ESRD education helps more patients to select thedialysis modality of their choice in a timely manner(18). We concur with those authors, because we haveseen sustained growth since we started pre-ESRD edu-cation in our institution. Also, one important causefor the loss of patients to ICHD was catheter failure.Such losses were not included in the BPDC data, be-cause those patients were never enrolled into our PDprogram owing to primary failure or non function ofthe PD catheter after the surgery.

PRIMARY CATHETER FAILURE

We have data only from March 1997 until the end of2001 on primary catheter failure. During that time,46˚patients were referred for placement of a PD cath-eter. In 33˚patients, the catheter was successfullyplaced, for a success rate of 72% and a failure rate of

28%. Some patients were sent for catheter placementdespite a prior history of abdominal surgery.

ConclusionWe analyzed our data to determine the cause of a de-clining trend in the utilization of PD at our institution.We conclude that the decline is in keeping with a glo-bal trend. The decline in our PD program is mainlydue to a lack of aggressive recruitment: fewer patientswere added to the program than left the program. Al-though our findings are similar to those of the BPDC,our losses are due mainly to ICHD and to transplanta-tion, as opposed to ICHD and death as recorded in theBPDC. The discrepancy between our data and theBPDC data may be due to the inclusion of pediatricpatients, who constitute a sizable proportion of ourtotal PD population and who are likely to have a highertransplantation rate and lower mortality. Loss to ICHDwas due mainly to peritonitis and failure to meet theDOQI criteria for adequacy. Another speculation forthe lower death rate may be the switch to ICHD as thepatient becomes sicker. The death rate would then bereported under ICHD, which may also explain ourhigher rate of loss to ICHD. The distance from ourpatients residences to our hospital probably plays amajor role in our losing them to ICHD. Finally, pri-mary catheter failure was also a significant factor forpoor recruitment to our PD program.

References1 United States Department of Health and Human

Services, Public Health Service, National Institutes ofHealth, National Institute of Diabetes and Digestiveand Kidney Diseases. Treatment modalities. In:USRDS 2000 Annual Data Report. USRDS 1997Annual Data Report. Bethesda: United States RenalData System; 2000: 69—76.

2 Statistics Canada, Canadian Institute for HealthInformation. Volume˚1: Dialysis and Renal Trans-plantation. In: Canadian Organ Replacement Register2000 Annual Report. Ottawa: Canadian Institute forHealth Information; 2000: 1—28.

3 Gokal R, Mallick NP. Peritoneal dialysis. Lancet1999; 353:823—8.

4 Nissenson AR, Prichard SS, Cheng IK, et al. Non-medical factors that impact on ESRD modality selec-tion. Kidney Int 1993; 43(Suppl 40):S120—7.

5 Gokal R, Baillod R, Bogle S, et al. Multi-centre studyon outcome of treatment in patients on continuousambulatory peritoneal dialysis and haemodialysis.

FIGURE 6 The geographic location of the island of Galveston andthe counties in the state of Texas. The shaded counties representthe areas referring patients to the University of Texas MedicalBranch.

Peritoneal Dialysis Experience at UTMB

105

Nephrol Dial Transplant 1987; 2:172—8.6 Nissenson AR, Gentile DE, Soderblom RE, Oliver

DF, Brax C. Morbidity and mortality of continuousambulatory peritoneal dialysis: regional experienceand long-term prospects. Am˚J Kidney Dis 1986;7:229—34.

7 Rotellar C, Black J, Winchester JF, et al. Ten yearsexperience with continuous ambulatory peritonealdialysis. Am˚J Kidney Dis 1991; 17:158—64.

8 Churchill DN. Comparative morbidity among hemo-dialysis and continuous ambulatory peritoneal dialy-sis patients. Kidney Int 1993; 43(Suppl 40):S16—22.

9 Bloembergen WE, Port FK, Mauger EA, Wolfe RA.A comparison of death between patients treated withhemodialysis and peritoneal dialysis. J˚Am SocNephrol 1995: 6:184—91.

10 Foley RN, Parfrey PS, Harnett JD, et al. Mode ofdialysis therapy and mortality in end-stage renaldisease. J˚Am Soc Nephrol 1998; 9:267—76.

11 Fenton SS, Schaubel D, Desmeules M, et al. Hemo-dialysis versus peritoneal dialysis: a comparison ofadjusted mortality rates. Am˚J Kidney Dis 1997; 30:334—42.

12 Collins AJ, Hao W, Xia H, et al. Mortality risks ofperitoneal dialysis and hemodialysis. Am˚J KidneyDis 1999; 34:1065—74.

13 Evans RW, Manninen DL, Garrison LP Jr, et al. Thequality of life of patients with end-stage renal dis-ease. N˚Engl J Med 1985; 312:553—9.

14 Gutman RA, Stead WW, Robinson RR. Physicalactivity and employment status of patients on mainte-nance dialysis. N˚Engl J Med 1981; 304:309—13.

15 Simmons RG, Abress L. Quality-of-life issues forend-stage renal disease patients. Am˚J Kidney Dis1990; 15:201—8.

16 Blake PG, Finkelstein FO. Why is the proportion ofpatients doing peritoneal dialysis declining in NorthAmerica? Perit Dial Int 2001; 21:107—14.

17 Oreopoulos DG, Blake P. Declining utilization ofperitoneal dialysis: time to stop imposing our biaseson the patients and let them be dialyzed with themodality of their choice. ASAIO˚J 2001; 47:312—15.

18 Golper TA, Vonesh EF, Wolfson M, Baudoin M,Schreiber MJ. The impact of pre-ESRD education ondialysis modality selection [Abstract]. J˚Am SocNephrol 2000; 11:A1223.

Corresponding author:Mahendra Agraharkar, MD, FACP, Nephrology Division, De-partment of Medicine, 4.200˚John Sealy Annex, Univer-sity of Texas Medical Branch, 301˚University Boulevard,Galveston, Texas 77555-0562 U.S.A.

Agraharkar et al

Advances in Peritoneal Dialysis, Vol. 18, 2002

High Peritoneal Transport:A Blessing or Curse?

From: Division of Nephrology, Department of InternalMedicine, University of Missouri, Columbia, Missouri,U.S.A.

C. Gentiana Voinescu, Ramesh Khanna, Karl D. Nolph

High transporters are defined based on the perito-neal equilibration test. Peritoneal transport ratechanges over time, inflammation and angiogenesisaffecting the total pore area. Factors influencing theneovascularization process are described.

High transporters have distinctive clinical andlaboratory features. The incidence of high transport-ers varies among different populations.

Unfortunately, high transporters have the worstclinical outcomes. Mechanisms proposed to explainthe adverse prognosis including hypoalbuminemia,chronic fluid overload, malnutrition, and chronic in-flammation ar e discussed.

We suggest dividing baseline high transportersinto two groups: sick and healthy high trans-porters. The two types of high transporters have dif-ferent baseline characteristics and different clinicaloutcomes. Hopefully, further studies will better de-fine the appearance of the two groups of hightransporters.

Key wordsHigh transporters, high peritoneal transport, hypo-albuminemia, malnutrition-inflammation-atheroscle-rosis syndrome

IntroductionSince its description in the mid-1980s by Twardowskiet al (1), the peritoneal equilibration test (PET) hasbeen widely used to characterize peritoneal dialysis(PD) patients according to peritoneal transport status.Based on the dialysate-to-plasma ratio of creatinine(D/P Cr) or the dialysate-to-initial-dialysate ratio ofglucose (D/D0), or both, patients can be classified ashigh, high-average, low-average, and low transport-ers. High transporters have the highest rates of creati-nine diffusion and glucose absorption. After a 4-hourdwell time, they achieve a D/P Cr ratio above 0.8,while their D/D0 decreases below˚0.3.

Discussion

Peritoneal membrane evolution over timeAmong the various transport resistances that consti-tute the peritoneal membrane, the vascular compo-nent represented by the capillary wall is the mostimportant barrier to solute transport (2). Thus, theo-retically, the transport capacity of the peritoneal mem-brane depends on the two characteristics of thevascular site: the effective peritoneal surface area (to-tal number of pores present in perfused capillaries)and the membrane permeability (size of the pores).Over time, the effective peritoneal surface area canincrease, as seen with neoangiogenesis, leading tohigher mass transfer-area coefficients (MTCs) for lowand high molecular weight molecules. The intrinsicmembrane permeability can potentially be altered bya change in the diameter of the large pores, but clini-cal changes in transport are predominantly in totalpore area (2). A recent study (3) tried to address therelative importance of peritoneal fibrosis and angio-genesis in peritoneal membrane dysfunction. The au-thors compared the response to intraperitoneallipopolysaccharide-induced inflammation in twogroups of genetically altered animals. In one group,production of decorin, a proteoglycan inhibitor oftransforming growth factor-β, reduced collagen pro-duction. In the other genetically altered group, thehigh production of angiostatin showed a significantreduction in vessel density. They demonstrated bet-ter ultrafiltration preservation in animals producingabundant angiostatin as compared with animals pro-ducing high decorin, suggesting that the blood vesseldensity, and not the collagen concentration, is respon-sible for peritoneal membrane dysfunction.

A cross-sectional study (4) evaluating the histol-ogy of parietal peritoneal tissue obtained by biopsy in22 continuous ambulatory peritoneal dialysis (CAPD)patients found a strong correlation between the rela-tive capillary area and the D/P˚Cr. That finding indi-cated that increased capillary surface area is actuallyinvolved in the mechanism of high transport duringPD. The same group of researchers analyzed the rela-

107

tionship between vascular endothelial growth factor(VEGF), known for its angiogenic action, and perito-neal membrane transport status (5). The positive cor-relation between the growth factor and peritonealtransport suggested that VEGF might be responsiblefor the hypervascularization associated with a highperitoneal transport rate. Free plasma VEGF levelshave been found to be high in pre-dialysis patientsand may contribute to vascular endothelial dysfunc-tion (6). Leptin, another angiogenic substance, hasbeen implicated in the peritoneal hypervascularizationthat characterizes high transporters˚(7).

With the initiation of PD, the peritoneal membraneundergoes structural and functional changes. Duringthe first 2˚weeks of PD, in 35.3% of patients, a slightincrease in peritoneal transport characteristics tendsto occur (8). For that reason, Rocco et al recommendperforming the baseline PET one month after the ini-tiation of˚PD.

Long term, and in the absence of recurrent epi-sodes of peritonitis, peritoneal transport tends to re-main stable in most patients (9). However, severalstudies have shown that high transporters tend slowlytoward a decrease in peritoneal membrane transportover time, while the low and low-average groups dem-onstrate reciprocal changes (10,11).

IncidenceThe incidence of baseline high transport status variesamong different populations.

The CANUSA study (12) documented an inci-dence of 15.3% high transporters among patients ini-tiating PD in the United States and Canada. Thatincidence is similar to the incidence reported in north-ern European (13) and Latin (14) populations. A muchhigher incidence (around 50%) has been found inAsian Indians (15) and in Greeks (16). In Australiansand New Zealanders, the incidence of high transport-ers commencing PD is 32% or higher (17).

Clinical and laboratory features specific for hightransportersCompared with other transport types, high transport-ers have more rapid increases in D/P˚Cr (18). As aresult, small-solute clearance targets can easily bereached unless the patient is very large. The glucoseosmotic gradient dissipates relatively rapidly owingto the high rate of glucose absorption. During longdwells, net ultrafiltration (UF) ceases when the glu-

Voinescu et al

cose osmotic gradient declines to the point where lym-phatic reabsorption exceeds the rate of ultrafiltration.Maximum UF is captured within the first 2˚hours ofa dwell (18); long dwells (>˚4˚— 6˚hours) may havelittle or negative UF. As a result, in the absence ofresidual renal function, patients may run into prob-lems with volume control and hypertension. To maxi-mize clearances and UF, short-dwell therapies suchas nightly intermittent peritoneal dialysis (NIPD) ordaily automated peritoneal dialysis (DAPD) arerecommended.

The high transporter group includes higher pro-portions of children (19), older men, and patients withdiabetes (12).

Acid—base status might also be influenced by peri-toneal membrane transport characteristics. In a cross-sectional study of 143 stable PD patients (20), lactateand dialytic base gains were significantly higher inhigh transporters, leading to increased pH and bicar-bonate levels.

Several studies have found lower serum albuminin the high transporters as compared with other groups(12,21,22). That observation was linked to higher di-alysate protein losses (12,21) and fluid overload owingto poor UF. The gastrointestinal tract might contrib-ute to an additional protein loss. Protein-losing enter-opathy, determined by the fecal clearance ofα1-antitrypsin, is more frequent in patients with highurea and creatinine MTCs (23).

Unfortunately, high transporters have the worstclinical outcomes. They are at increased risk of death(2-year survival is 71% versus 91% in low transport-ers in the CANUSA study) and technique failure(12—14,24).

Clinical outcomes and proposed mechanismsTo explain the adverse prognosis found in high trans-porters, various mechanisms have been postulated(25).

It is now widely recognized that hypoalbumin-emia is associated with higher mortality ratesamong patients on dialysis (26). On the other hand,hypoalbuminemia is more common in high trans-porters. Is the high peritoneal transport status lead-ing to hypoalbuminemia? Can hypoalbuminemiaalone explain the increased mortality risk seen inhigh transporters?

Although protein loss is greater in the high trans-port group after initiation of PD, that high loss cannot

108

explain why some of the patients have lower baselineserum albumin levels (12). Moreover, it has not yetbeen demonstrated that hypoalbuminemia per se isthe direct cause of high mortality, rather than being amarker of underlying comorbidities (25).

Some researchers have related adverse outcomesin high transporters to poor UF and chronic fluid over-load. Although the thesis is not fully proved, there aredata to support it: high transporters have higher bodyweights and diastolic blood pressures (13). Patientswith symptomatic fluid retention on CAPD are 3.7times more likely to be high transporters rather thanlow transporters (27).

Repeated use of hypertonic exchanges may in-crease the carbohydrate load, possibly suppressingappetite and reducing protein intake. Additionally,increased dialysate protein losses may contribute tomalnutrition and further high risk of mortality. Butdoes the literature contain evidence to support mal-nutrition in high transporters?

Although initial cross-sectional studies showedlower net protein catabolic rate (nPCR), lean bodymass (LBM), and daily creatinine production in thehigh transporters, suggesting malnutrition (21), fur-ther studies did not confirm those observations(12,24,28,29). Aside from hypoalbuminemia, theCANUSA study (12) showed no differences amongthe transport groups in other estimates of nutritionstatus, including subjective global assessment, LBM,and PCR. When nutrition status was comprehensivelyassessed by means of 24-hour dietary recall togetherwith a nutrition index (calculated based on eight clini-cal, biochemical, and anthropometric parameters), nocorrelation with peritoneal transport rate was found(28).

Those results lead us to examine an intriguinghypothesis: that high transport status, hypoalbumin-emia, and adverse outcomes are the result of underly-ing comorbidities, perhaps related in many patients tothe MIA (malnutrition, inflammation, atherosclero-sis) syndrome (30).

Through cytokine dysregulation, the chronic in-flammation that accompanies uremia might causehigher peritoneal transport (via neovascularizationand maybe other mechanisms) in the peritoneal mem-brane, and an increased risk of cardiovascular dis-eases and hypoalbuminemia (30). However, theliterature data to support that hypothesis are conflict-ing. A cross-sectional study (31) performed on 39˚PD

patients found no correlation between peritonealtransport rate and chronic inflammation as measuredby blood levels of hyaluronan, interleukin-1β, tumornecrosis factor˚α, and C-reactive protein (CRP). Onthe other hand, the same group later reported (11)that during the first year on PD, changes in perito-neal transport rate appeared to be linked to inflam-mation (as suggested by a high CRP) and a decreasein residual renal function (RRF). Those findings arecompatible with the possibility that two types of hightransporters might exist at baseline: one subgroup withevidence of chronic inflammation, lower baselineRRF, and a faster loss of RRF during the first year;and another subgroup that would include patients di-agnosed with a high transport status at baseline, with-out much evidence of inflammation and with better,preserved˚RRF.

Evidence for a major role of inflammation andneovascularization comes from a prospective studyby Chung et al (32). They followed 213 PD pa-tients and found the same poor survival in somebaseline high transporters as was documented byprevious studies. The 2-year patient survival in thebaseline high transport group was significantlylower as compared with the other groups combined(57.1% vs. 79.5%, p˚= 0.009). However, when pa-tients with comorbid diseases were censored fromthe analysis, the survival in the baseline high trans-porters without comorbid conditions was not dif-ferent compared with the other transport groupscombined. That study brings weight to the hypoth-esis that high transporters can be grouped into twotypes: with and without comorbid diseases associ-ated with inflammation.

The vascular density hypothesis: sick andhealthy high transportersBased on current literature evidence, we suggest di-viding baseline high transporters into two groups(Figure˚1). Sick high transporters have evidence ofinflammation before initiating dialysis. Through itsneoangiogenic action, inflammation leads to abundantperitoneal capillaries. Patients at risk of falling intothis category display baseline low serum albumin,higher CRP, lower RRF, and lower protein equivalentof nitrogen appearance (PNA). The baseline lowerRRF further predisposes to inflammation (11), whichin turn accelerates the loss of RRF. For these patientswith comorbid diseases, evidence of chronic inflam-

High Peritoneal Transport

109

mation, and decreased RRF at baseline, a high perito-neal transport status is a predictor of increased mor-tality and morbidity while on˚PD.

Baseline high transporters without associatedcomorbidities or inflammation the healthybaseline high transporters most likely have a ge-netically inherited increased vascular density. Theirlack of significant inflammation might be quantifiedby normal/low serum albumin, normal CRP, betterRRF, and more normal PNA. Upon initiation of PD(lower portion of Figure˚1, below dashed line) and inthe absence of angiogenic stimuli, the healthy hightransporters and the non high transporters (assumingthat they don t have other associated comorbidities)have better survival and less evidence of inflamma-tion (no elevation in CRP, normal/low serum albu-min). On the other hand, over time, their peritonealvascular density can increase and change their out-comes if they acquire inflammation and neo-vascularization. The change in solute transport isheralded by an increase in D/P˚Cr. When that hap-pens, their mortality and morbidity (malnutrition, ath-

erosclerosis) approaches the level seen in baselinehigh transporters with evidence of pre-dialysisinflammation.

ConclusionsWe speculate that there are two types of high trans-porters: sick and healthy high transporters. Di-viding the high transporters into two subgroups andacknowledging their dynamics might help to explainthe controversial results seen in various studies. Itmight also explain why different populations havehigher proportions of baseline high transporters [asseen in Australians, New Zealanders (17), Greeks (16),and Asian Indians (15)]; why Greeks don t have thepoor outcomes (16) documented in most other stud-ies; and why, by 6˚— 12˚months of follow-up, only22.4% of Australian high transporters had maintainedsignificant RRF (17).

For future studies, it might be important to docu-ment whether baseline high transport status is associ-ated with chronic inflammation or comorbidities (orboth) and whether the high transport status was ac-quired at baseline or during follow-up. The differencescan be quite significant. In one retrospective study(10), for example, 16˚patients were defined as hightransporters based on baseline PET. The final PETrecorded 8˚high transporters. Of the initial 16˚hightransporters, only 3˚patients remained for the finalcount. The rest (5˚patients) acquired the high perito-neal status over time.

For both subgroups, short dwell-time therapiessuch as NIPD or DAPD maximize small-solute clear-ances and UF.

Further studies will hopefully better define theappearance of the two types of high transporters.

References1 Twardowski ZJ, Nolph KD, Khanna R, et al. Perito-

neal equilibration test. Perit Dial Bull 1987;7:138—47.

2 Krediet RT. Evaluation of peritoneal membraneintegrity. J Nephrol 1997; 10:238—44.

3 Margetts PJ, Gyorffy S, Kolb M, et al.Antiangiogenic and antifibrotic gene therapy in achronic infusion model of peritoneal dialysis in rats.J˚Am Soc Nephrol 2002; 13:721—8.

4 Numata M, Nakayama M, Pecoits—Filho RF, et al.Increased peritoneal capillary surface area in CAPDpatients with high peritoneal transport rate. J˚Am Soc

Figure 1 The vascular density hypothesis. CRP˚= C-reactive pro-tein; RRF˚= residual renal function; PNA˚= protein equivalent ofnitrogen appearance.

Voinescu et al

110

Nephrol 2001; 12:436A—7A.5 Pecoits—Filho RF, Araujo MR, Lindholm B, et al.

Vascular endothelial growth factor: relationship toinflammation and peritoneal solute transport rate inperitoneal dialysis patients [Abstract]. J˚Am SocNephrol 2001; 12:438A.

6 Harper SJ, Downs L, Tomson CR, et al. Elevatedplasma vascular endothelial growth factor levels innon-diabetic predialysis uraemia. Nephron 2002;90:341—3.

7 Pecoits—Filho RF, Heimb r ger O, Lindholm B, et al.Leptin sensitivity correlates to high solute transportrate in peritoneal dialysis patients [Abstract]. J˚AmSoc Nephrol 2001; 12:438A.

8 Rocco MV, Jordan JR, Burkart JM. Changes in peri-toneal transport during the first month of peritonealdialysis. Perit Dial Int 1995; 15:12—17.

9 Blake PG, Abraham G, Sombolos K, et al. Changesin peritoneal membrane transport rates in patients onlong term CAPD. Adv Perit Dial 1989; 5:3—7.

10 Lo WK, Brendolan A, Prowant BF, et al. Changes inthe peritoneal equilibration test in selected chronicperitoneal dialysis patients. J˚Am Soc Nephrol 1994;4:1466—74.

11 Chung SH, Heimb r ger O, Stenvinkel P, et al. Asso-ciation between inflammation and changes in re-sidual renal function and peritoneal transport rateduring the first year of dialysis. Nephrol Dial Trans-plant 2001; 16:2240—5.

12 Churchill DN, Thorpe KE, Nolph KD, KeshaviahPR, Oreopoulos DG, Pag D for the Canada—U.S.A.(CANUSA) Peritoneal Dialysis Study Group. In-creased peritoneal membrane transport is associatedwith decreased patient and technique survival forcontinuous peritoneal dialysis patients. J˚Am SocNephrol 1998; 9:1285—92.

13 Wang T, Heimb r ger O, Waniewski J, Bergstr m J,Lindholm B. Increased peritoneal permeability isassociated with decreased fluid and small-soluteremoval and higher mortality in CAPD patients.Nephrol Dial Transplant 1998; 13:1242—9.

14 Cueto—Manzano AM, Diaz—Alvarenga A, Correa—Rotter R. Analysis of the peritoneal equilibration testin Mexico and factors influencing the peritonealtransport rate. Perit Dial Int 1999; 19:45—50.

15 Agarwal DK, Sharma AP, Gupta A, et al. Peritonealequilibration test in Indian patients on continuousambulatory peritoneal dialysis: does it affect patientoutcome? Adv Perit Dial 2000; 16:148—51.

16 Passadakis PS, Thodis ED, Panagoutsos SA, SelisiouCA, Pitta EM, Vargemezis VA. Outcome for continu-ous ambulatory peritoneal dialysis patients is not

predicted by peritoneal permeability characteristics.Adv Perit Dial 2000; 16:2—6.

17 Australia and New Zealand Dialysis and TransplantRegistry. 24th Annual Report. Woodville, SouthAustralia: ANZDATA; 2002: 57.

18 Twardowski ZJ. Nightly peritoneal dialysis. Why,who, how, and when? ASAIO Trans 1990; 36:8—16.

19 Mactier RA, Khanna R, Moore H, Russ J, Nolph KD,Groshong T. Kinetics of peritoneal dialysis in chil-dren: role of lymphatics. Kidney Int 1988; 34:82—8.

20 Kang DH, Yoon KI, Lee HY, Han DS. Impact ofperitoneal membrane transport characteristics onacid—base status in CAPD patients. Perit Dial Int1998; 18:294—302.

21 Nolph KD, Moore HL, Prowant B, et al. Continuousambulatory peritoneal dialysis with a high flux mem-brane. ASAIO˚J 1993; 39:904—9.

22 Diaz—Alvarenga A, Abasta—Jimenez M, Bravo B,Gamba G, Correa—Rotter R. Serum albumin and bodysurface area are the strongest predictors of the perito-neal transport type. Adv Perit Dial 1994; 10:47—51.

23 Aguilera A, Bajo MA, Codoceo R, et al. Protein-losing enteropathy is associated with peritonealfunctional abnormalities in peritoneal dialysis pa-tients. Perit Dial Int 2000; 20:284—9.

24 Szeto CC, Law MC, Wong TY, Leung CB, Li PK.Peritoneal transport status correlates with morbiditybut not longitudinal change of nutritional status ofcontinuous ambulatory peritoneal dialysis patients: a2-year prospective study. Am˚J Kidney Dis 2001;37:329—36.

25 Blake PG. What is the problem with high transport-ers? Perit Dial Int 1997; 17:317—20.

26 Lowrie EG, Lew NL. Death risk in hemodialysispatients: the predictive value of commonly measuredvariables and an evaluation of death rate differencesbetween facilities. Am˚J Kidney Dis 1990; 15:458—82.

27 Tzamaloukas AH, Saddler MC, Murata GH, et al.Symptomatic fluid retention in patients on continu-ous peritoneal dialysis. J˚Am Soc Nephrol 1995;6:198—206.

28 Cueto—Manzano AM, Espinosa A, Hernandez A,Correa—Rotter R. Peritoneal transport kinetics corre-late with serum albumin but not with the overallnutritional status in CAPD patients. Am˚J Kidney Dis1997; 30:229—36.

29 Harty JC, Boulton H, Venning MC, Gokal R. Isperitoneal permeability an adverse risk factor formalnutrition in CAPD patients? Miner ElectrolyteMetab 1996; 22:97—101.

30 Bergstr m J, Lindholm B. Malnutrition, cardiac

High Peritoneal Transport

111

disease, and mortality: an integrated point of view.Am˚J Kidney Dis 1998; 32:834—41.

31 Wang T, Heimb r ger O, Cheng HH, Bergstr m J,Lindholm B. Does a high peritoneal transport ratereflect a state of chronic inflammation? Perit Dial Int1999; 19:17—22.

32 Chung SH, Chu WS, Lee HA. Peritoneal transport

characteristics, comorbid diseases and survival inCAPD patients. Perit Dial Int 2000; 20:541—7.

Corresponding author:Ramesh Khanna, MD, Division of Nephrology, MA436Health Sciences Center, One Hospital Drive, Columbia,Missouri 65212 U.S.A.

Voinescu et al

Antibiotic ProphylaxisBefore Peritoneal DialysisCatheter InsertionAtul Katyal, Adit Mahale, Ramesh Khanna

From: University of Missouri—Columbia, Division ofNephrology, Columbia, MO, U.S.A.

The use of antibiotic prophylaxis before peritonealdialysis (PD) catheter insertion has been a matter ofgreat interest for both the surgical and the nephrol-ogy community. No uniform consensus exists on thetiming, duration, or choice of antibiotic prophylaxis.The exact incidence of early postoperative peritonitisis unknown. The impact of the use of antibiotic pro-phylaxis in the prevention of early PD peritonitis andin long-term catheter survival is not clear. In that re-spect, many retrospective and prospective studies havebeen undertaken in the past, and they have shownconflicting results. Based on extensive data from thesurgical and nephrology literature, and also based onour experience, we present a review of the use of anti-biotic prophylaxis before peritoneal dialysis catheterinsertion.

Key wordsCatheter, antibiotic, prophylaxis

IntroductionThe true incidence of early postoperative peritonitisafter peritoneal catheter insertion is unknown.Postoperative infections can lead to infection of thetunnel and exit site, and may cause recurrent peritonitisand catheter loss.

Discussion

Principles of antibiotic useInitial use of antibiotic prophylaxis before or at thetime of peritoneal dialysis catheter placement hasbeen based mainly on data extrapolated from similarsurgical interventions (1—4). Surgical wounds havebeen traditionally classified into clean operationswithout foreign implants, clean operations withforeign body implants, clean contaminated wounds,contaminated wounds, and dirty wounds (1,5). The

incidence of wound infection increases progressivelyfrom clean (2.9%), through clean contaminated(3.9%) and contaminated (8.5%), to dirty/infected(12.6%) (6). High-risk patients (Table˚I) withmultiple medical problems have a higher incidenceof wound infections (5—8).

The operation for insertion of a peritoneal dialysiscatheter can be classified as clean surgery withimplantation of a foreign body. Patients requiring suchcatheters are at high risk, especially those with multiplemedical problems and malnutrition. The use of low-pH, high glucose, lactate-containing peritoneal dialysissolutions has been shown to reduce the phagocyticactivity of macrophages in the peritoneum (9).

The surgical literature contains abundant and con-vincing evidence for the use of antibiotic prophylaxisfor clean operations utilizing foreign body implants(1,3,5,6,10). In fact, the current trends even advocatethe use of antibiotic prophylaxis in clean surgery with-out implants (11).

Studies in peritoneal dialysis patientsNumbers of published studies have looked at whetherthe use of antibiotic prophylaxis before peritonealdialysis catheter insertion makes any difference tothe incidence of early postoperative peritonitis.

Using the Cox proportional hazards model, theU.S. Renal Data System (USRDS) 1992 Data Report(12) showed, in 3366˚patients, no difference betweenpatients receiving antibiotic prophylaxis and those notreceiving it. But the study was retrospective, and thetiming and duration of antibiotic administration wasnot clearly defined.

Golper et al, in the Network˚9 study (13), showeda 39% reduction in the risk of peritonitis and a 38%reduction in combined peritonitis and exit-site/tunnelinfection. No difference was seen in the time to firstperitonitis.

Lye et al (14), using single-dose cefazolin andgentamicin, showed no significant benefit of antibioticprophylaxis. Bennett—Jones et al (15), using

Advances in Peritoneal Dialysis, Vol. 18, 2002

113

gentamicin at the time of catheter insertion, showed areduction in the incidence of peritonitis from 46% (noantibiotic) to 8% (antibiotics). Sardegna et al (16), ina retrospective study, showed significant benefit usingmultiple antibiotics in a pediatric dialysis population.

Wikdahl et al (17), in a prospective study usingintravenous (1.5˚g, 0.5˚— 2˚hours before surgery) andintraperitoneal (250˚mg in 1-L bag starting peri-operatively) cefuroxime (a second-generation cepha-losporin), showed a beneficial role for preoperativeantibiotics. Active therapy was associated with a lowerincidence of microbial growth in the dialysate (0 of18˚patients vs. 6 of 20˚patients, p˚= 0.021) during a10-day follow-up period.

In a recent large prospective randomized trialconducted over a 6-year period (18), 221˚patientswere randomly assigned either to group˚I (vanco-mycin, 1˚g intravenously, 12˚hours before the pro-cedure), group˚II (cefazolin, 1˚g intravenously,3˚hours before the procedure), or group˚III (no an-tibiotics). At 2˚weeks, the groups given antibiot-ics showed a significantly lower incidence ofperitonitis (1%, 7%, and 12% for groups˚I, II, andIII, respectively). All the catheters in that studywere placed peritoneoscopically.

Timing of antibiotic administrationThe timing of the prophylactic antibiotic adminis-tration before a surgical procedure has been exten-

sively studied in several prospective randomized tri-als (2,4,19). Most studies have concluded that, to pre-vent infections, antibiotics should already be presentin adequate concentrations when the tissues are ex-posed to contaminating bacteria. The rate of infec-tion increases for each hour that antibiotics aredelayed after the start of the operation (2). The ex-planation for that observation is that, when an inci-sion is made, the body s inflammatory processimmediately mobilizes and begins to isolate thewound. By the time the operation is completed, thewound contaminated by the skin microbial flora hasbeen completely isolated, and systemic antibioticscannot reach the wound.

Duration of antibiotic prophylaxisThe total duration of antibiotic prophylaxis is stillbeing debated, although the use of antibiotic therapybeyond 24˚hours has shown no benefit and may leadto a rise in antibiotic resistance (20—22). A single-dose intravenous antibiotic before surgery has beenshown to be effective in most procedures ofapproximately 0˚— 2˚hours (1,2,5). Procedures lastinglonger than 3˚— 4˚hours may require an additionaleffective dose (5).

When compared with one shot antibioticregimens, longer courses of antibiotics have notresulted in any statistical difference with regard to theincidence of postoperative infection (23). The focusis on maintaining adequate antibiotic levels in thetissues continuously from the time of incision to a timeafter wound closure sufficient to cover the lag phaseof the contaminating organisms (approximately6˚hours for a 2-hour operation), but not longer (2).

Choice and route of antibiotic administrationThe first-generation cephalosporins continue to be themost frequently used antibiotics in the context ofperitoneal dialysis catheter placement. Cefazolin, withits prolonged half-life in renal failure (normal: 90˚—150˚minutes) and its time to peak serum concentrationof 0.5˚— 2˚hours (intramuscular administration), offersa valuable coverage. The effective dose should begoverned by the patient s weight. The intravenousroute of administration is preferred.

The role of topical antimicrobials and oralantibiotics in prophylaxis remains to be determined.The routine use of vancomycin as a prophylactic agentshould be discouraged to avoid the development of

TABLE I Factors associated with increased risk of infection

Systemic factorsDiabetesImmunosuppressionObesityExtremes of ageMalnutrition and protein depletionRecent surgeryMassive transfusionMultiple (3 or more) preoperative comorbid medical

diagnosesASA class 3, 4, or 5

Local factorsForeign body or implantsNo bath with chlorhexidine soap before surgeryElectrocauteryInjection with epinephrineWound drainsHair removal with razorPrevious irradiation of site

ASA˚= American Society of Anesthesiologists.

Katyal et al

114

resistant micro-organisms such as vancomycin-resistantenterococci and staphylococci. Maintaining a local,up-to-date hospital analysis of the microbialsusceptibilities of wound isolates is also important fordetecting important shifts in patterns of resistance.Changes can then be made accordingly.

ConclusionTo summarize, reports conflict regarding the use ofprophylactic antibiotics for peritoneal dialysis cath-eter replacement. The largest prospective clinical trialin peritoneal dialysis patients (18) showed a signifi-cant benefit of antibiotic prophylaxis in reducing earlypostoperative peritonitis. The surgical literature alsocontains abundant and convincing evidence for theuse of antibiotic prophylaxis in clean operations uti-lizing foreign body implants.

The timing of antibiotic administration is crucialto its effectiveness, and the antibiotic should be givenimmediately before surgery so that adequate tissueconcentration has been achieved at the time of inci-sion. A single-dose first-generation cephalosporin,given intravenously 1˚— 2˚hours before the procedure,is a good first-line prophylactic antibiotic. At the Uni-versity of Missouri—Columbia, we use cefazolin 1.5˚g,intravenously, 2˚hours before the procedure.

Use of vancomycin as a routine prophylactic agentshould be discouraged.

According to USRDS, only 43% of patients re-ceive antibiotic prophylaxis before catheter insertion.Whether the widespread use of prophylactic antibiot-ics will make a significant difference to the overallincidence of peritonitis remains to be seen (24). Toachieve good results, a well-thought-out prophylacticregimen needs to be supplemented by exquisite surgi-cal technique and competent post-surgicalmanagement.

References1 Nichols RL. Surgical wound infection. Am J Med

1991; 91:54S—64S.2 Sanderson PJ. Antimicrobial prophylaxis in surgery:

microbiological factors. J Antimicrob Chemother1993; 31(Suppl B):1—9.

3 Ulualp K, Condon RE. Antibiotic prophylaxis forscheduled operative procedures. Infect Dis ClinNorth Am 1992; 6:613—25.

4 Burke JF. The effective period of preventive antibi-otic action in experimental incisions and dermallesions. Surgery 1961; 50:161—8.

5 Woods RK, Dellinger EP. Current guidelines forantibiotic prophylaxis of surgical wounds. Am FamPhysician 1998; 57:2731—40.

6 Haley RW, Culver DH, Morgan WM, White JW,Emori TG, Hooton TM. Identifying patients at highrisk of surgical wound infection. A simple multivari-ate index of patient susceptibility and wound con-tamination. Am J Epidemiol 1985; 121:206—15.

7 Culver DH, Horan TC, Gaynes RP, et al. Surgicalwound infection rates by wound class, operativeprocedure, and patient risk index. National Nosoco-mial Infections Surveillance System. Am J Med1991; 91(Suppl 3B):152S—7S.

8 Nichols RL. Postoperative wound infection. N Engl JMed 1982; 307:1701—2.

9 Thomas S, Schenk U, Fischer FP, Mettang T,Passlick—Deetjen J, Kuhlmann U. In vitro effects ofglucose polymer—containing peritoneal dialysis fluidson phagocytic activity. Am J Kidney Dis 1997;29:246—53.

10 Platt R, Zaleznik DF, Hopkins CC, et al. Periopera-tive antibiotic prophylaxis for herniorrhaphy andbreast surgery. N Engl J Med 1990; 322:153—60.

11 Lewis RT, Weigand FM, Mamazza J, Lloyd—SmithW, Tataryn D. Should antibiotic prophylaxis be usedroutinely in clean surgical procedures: a tentativeyes. Surgery 1995; 118:742—6.

12 United States Department of Health and HumanServices, Public Health Service, National Institutesof Health, National Institute of Diabetes and Diges-tive and Kidney Diseases. USRDS 1992 Annual DataReport. Catheter related factors and peritonitis risk inCAPD patients. Bethesda: United States Renal DataSystem; 1992.

13 Golper TA, Brier ME, Bunke M, et al. Risk factorsfor peritonitis in long-term peritoneal dialysis: theNetwork 9 peritonitis and catheter survival studies.Academic Subcommittee of the Steering Committeeof the Network˚9 Peritonitis and Catheter SurvivalStudies. Am J Kidney Dis 1996; 28:428—36.

14 Lye WC, Lee EJC, Tan CC. Prophylactic antibioticsin the insertion of Tenckhoff catheters. Scand J UrolNephrol 1992; 26:177—80.

15 Bennet—Jones DN, Martin J, Barrat AJ, Duffy TJ,Naish PF, Aber GM. Prophylactic gentamicin in theprevention of early exit-site infections and peritonitisin CAPD. Adv CAPD 1988; 147—50.

16 Sardegna KM, Beck AM, Strife CF. Evaluation ofperioperative antibiotics at the time of dialysis catheterplacement. Pediatr Nephrol 1998; 12:149—52.

17 Wikdahl AM, Engman U, Stegmayr BG, SorenssenJG. One-dose cefuroxime i.v. and i.p. reduces micro-bial growth in PD patients after catheter insertion.

Antibiotic Prophylaxis Before PD Catheter Insertion

115

Nephrol Dial Transplant 1997; 12:157—60.18 Gadallah MF, Ramdeen G, Mignone J, Patel D,

Mitchell L, Tatro S. Role of preoperative antibioticprophylaxis in preventing postoperative peritonitis innewly placed peritoneal dialysis catheters. Am JKidney Dis 2000; 36:1014—19.

19 Classen DC, Evans RS, Pestotnik SL, Horn SD,Menlove RL, Burke JP. The timing of prophylacticadministration of antibiotics and the risk of surgical-wound infection. N Engl J Med 1992; 326:281—6.

20 Wittman DH, Condon RE. Prophylaxis of postopera-tive infections. Infection 1991; 19:S337—44.

21 Fukatsu K, Saito H, Matsuda T, Ikeda S, Furukawa S,Muto T. Influences of type and duration of antimicro-bial prophylaxis on an outbreak of methicillin-resis-tant Staphylococcus aureus and on the incidence of

wound infection. Arch Surg 1997; 132:1320—5.22 Fabian TC, Croce MA, Payne LW, Minard G,

Pritchard FE, Kudsk KA. Duration of antibiotictherapy for penetrating abdominal trauma: a prospec-tive trial. Surgery 1992; 112:788—94.

23 Dellinger EP, Wertz MJ, Lennard ES, OreskovichMR. Efficacy of short-course antibiotic prophylaxisafter penetrating intestinal injury. A prospectiverandomized trial. Arch Surg 1986; 121:23—30.

24 Vas S. Changing picture of peritonitis in peritonealdialysis. Am J Kidney Dis 2000; 36:1057—8.

Corresponding author:Atul Katyal, MD, Division of Nephrology, MA436 HealthSciences Center, University of Missouri, One HospitalDrive, Columbia, Missouri 65212 U.S.A.

Katyal et al