A Randomized Multicenter Clinical Trial Comparing Isosmolar Icodextrin with Hyperosmolar Glucose Solutions in CAPD

Kidney International, Vol. 46 (1994), pp.496—503

A randomized multicenter clinical trial comparing isosmolar Icodextrin with hyperosmolar glucose solutions in CAPD

CHANDRA D. MI5TRY, R Goji, ELIZABETH PEERS, and the MIDAS STUDY GROUP

Institute of Nephrology, Cardiff Royal Infirmary, Cardiff; Wales, and Manchester Royal Infirmaiy, Manchester, and Innovata Biomed Ltd., St. Albans, England, United Kingdom

A randomized multicenter clinical trial comparing isosmolar Icodex- Continuous ambulatory peritoneal dialysis (CAPD) has become trin with hyperosmolar glucose solutions in CAPD. The osmotic effec- established as an alternative to hemodialysis and is utilized in tiveness of a large molecular weight glucose polymer fraction (Icodextrin) as a novel "colloid" osmotic agent in peritoneal dialysis was established, about half the dialysis population in the United Kingdom [11. but the long-term safety remained undetermined. A randomized, con-Currently, glucose is the only osmotic agent in use but it is trolled multicenter investigation of Icodextrin in ambulatory peritoneal associated with several long-term complications. These are pri- dialysis (MIDAS) was undertaken to evaluate the long-term safety and marily related to rapid absorption through the peritoneum result- efficacy by comparing daily overnight (8 to 12 hr dwell) use of isosmolar Icodextrin (282 mOsmlkg) with conventional 1.36% (346 mOsmlkg) and ing in ultrafiltration of short duration and metabolic complica- 3.86% (484 mOsmlkg) glucose exchanges over six months. Two hundredtions such as hyperglycemia, hyperinsulinemia, hyperlipidemia and nine patients were randomized from 11 centers, with 106 allocated to and obesity [2]. In addition, there has been growing concern that receive Icodextrin (D) and 103 to remain on glucose (control group; C);the hyperosmolality and low pH of this solution may damage the 138 patients completed the six month study (71 C, 67 D). All patients were divided into weak (1.36%) or strong (3.86%) subgroups based on their use peritoneum and, thereby threaten its viability as a dialyzing of glucose solutions overnight during the pretreatment baseline period. membrane [3, 4]. Thus, the need for a more physiological solution The mean (± SEM) overnight ultrafiltration (UF) with D was 3.5 times containing an osmotic agent that would provide sustained ultra- greater than 1.36% glucose at eight hours [527 36 vs. 150 47 ml; 95%filtration and minimize the metabolic complications has led to an confidence interval (CI) for the difference +257 to +497 ml; P < 0.0001] and 5.5 times greater at 12 hours (561 44 vs. 101 48 ml, 95% CI for extensive but hitherto unsuccessful search for alternatives [5]. the difference +329 to +590; P < 0.0001) and no different from that of More recently, using a large molecular weight glucose polymer 3.86% glucose at eight hours (510 48 vs. 448 60 ml, 95% CI for thefraction derived from hydrolyzed corn starch, we have successfully difference —102 to +226 ml; P =0.44)and at 12 hours (552 44 vs. 414 adapted the physiological phenomenon of colloid osmosis to 78 ml, 95% CI for the difference —47 to +325 ml; P =0.06).The biochemical profiles were no different in the two groups except for a small generate ultrafiltration in peritoneal dialysis [6]. An isosmolar fall in serum sodium (140 to 136 mmol/liter) and chloride (103 to 99 solution of glucose polymer produces sustained ultrafiltration for mmol/liter) concentrations in the Icodextrin group. The mean serum up to 12 hours with lower transperitoneal absorption and poten- maltose increased from a pre-dialysis value of 0.04 g/liter to a steady state tial calorie load compared to glucose solutions. Furthermore, the level of 1.20 g/liter within two weeks and remained stable throughout the study. This was not associated with any adverse clinical effects and the lack of hyperglycemia and hyperinsulinemia associated with its overall CAPD-related symptom score was significantly better for D than use may offer some long-term metabolic advantages and possibly C. This study demonstrates that the daily overnight use of an isosmolar simplify management of diabetic patients [7]. Icodextrin solution was safe and effective up to six months and could Although only a fraction of glucose polymer load is absorbed, replace the overnight use of hyperosmolar glucose solutions. Longer term its metabolism is nevertheless incomplete. The absorbed polymer data will be necessary to establish further safety and efficacy. is rapidly hydrolyzed by amylase to oligosaccharidcs and eventu- ally to the disaccharide, maltose; further metabolism is limited by the absence of maltase activity in the human circulation [8, 9]. Other members of the MIDAS Study Group are: C.B. Brown, S. Smith, Even though substantial amounts of maltase activity have been DL. Edwards (Northern General Hospital, Sheffield); B.J.R. Junor, A. demonstrated in a variety of extraintestinal tissues [10, 111, Gordon (Western Infirmary, Glasgow); M. McMillan, M. Robertson (Stobhill Hospital, Glasgow); J. Michael, J. McKain (Queen Elizabeth including the kidneys [12, 131, maltose accumulation is known to Hospital, Birmingham); M. Raftery, J. Peters (Royal London Hospital, occur in renal failure [6]. In a short-term study, using a single London); E.J, Clutterbuck, M, Clemenger (Hammersmith Hospital, Lon- overnight exchange, a 36-fold increase in maltose level in CAPD don); J. Walls, C. Orton (Leicester General Hospital, Leicester); T.HJ. patients was not associated with any adverse effects [6, 14]. Goodship, J. Grieves (Royal Victoria Hospital, Newcastle-Upon-Tyne); J. However, the long-term consequences remained unknown and Olubodun (Nigeria); F.G. Jackson (Freeman Hospital, Newcastle-Upon- Tyne); D. Dharmasena, 0. Hourahane (Cardiff Royal Infirmary); D.J. cause for concern. Howarth (Manchester Royal Infirmary); R.N. Boyes, L.M. Clisby, Y. So far the experience with glucose polymer has been confined Beran (Innovata Biomed Ltd). to a small number of patients in a single center [6, 14]. The Received for publication September 20, 1993 present study (Multicentre Investigation of Icodextrin in Ambu- and in revised form February 25, 1994 latory Peritoneal Dialysis—MIDAS) was therefore designed to Accepted for publication February 28, 1994 investigate the long-term safety and efficacy in a larger, more © 1994 by the International Society of Nephrology heterogeneous, group of patients representative of the CAPD

496 MIDAS Study Group: Icodextrin in CAPD 497 population in the United Kingdom (UK). Since the isosmolarfundoscopy) were made on each patient before and after six glucose polymer (termed Icodextrin hereafter) solution was highlymonths of treatment. effective over long dwell exchanges, a single daily overnight use of Patients were asked to report all medical events which occurred Icodextrin with daytime exchanges of conventional glucoseduring the study. These were elicited by open questioning. In seemed most appropriate. This randomized controlled trial com-addition, a list of 12 common symptoms in CAPD patients (pain pared Icodextrin with glucose solutions during the overnight dwellduring exchange, abdominal fullness, night cramps, loss of appe- of 8 to 12 hours over six months, in CAPD patients from 11 renaltite, constipation, diarrhea, thirst, difficulty maintaining correct units in the UK. The trial was conducted over the period of Marchweight, shortness of breath, ankle swelling, tiredness and visual 1991 to February 1992. disturbance) were graded on a scale 0 to 3 (0 =none,1 =mild, 2 =moderateand 3 severe) at each visit during the study, and a Methods composite score determined for each patient.

Patients Composition of dialysate Patients were eligible for the study if they were aged 18 years or Icodextrin was manufactured by M.L. Laboratories plc., and over and established on CAPD for at least three months usingdialysis solution constituted by Baxter and Fresenius in a manner standard three to four exchanges, with no more than one "strong"identical to conventional glucose solutions. The composition of hypertonic glucose bag (3.86% glucose) per 24 hours. The patientsIcodextrin solution (mmol/liter) was: sodium 133, chloride 97, were also free of peritonitis and mechanical drainage complica-lactate 40, calcium 1.75, magnesium 0.25, Icodextrin 12.5. The tions for at least one month prior to the study. Icodextrin contained glucose polymer chain length varying be- Approval was obtained from local ethical committees at eachtween 4 to 250 glucose units with weight average molecular weight center and written informed consent given by every patient. (Mw) of 16,200 and number average molecular weight (Mn) of 5,800. The total osmolality of 7.5% Icodextrin solution was 282 Randomization andtreatment mOsmlkg and the pH 5.3. The osmolalities of commercially available glucose solutions Patients were randomized into Icodextrin (D) or control glu- supplied by Baxter/Fresenius were 346 mOsmlkg for 1.36%/1.5% cose (C) groups by telephone from a single office (Innovataand 484 mOsmlkg for 3.86%14.25%; pH 5.2. Biomed) at the first visit. Patients in the Icodextrin group used a 1.5 or 2 liter dialysis bag containing 7.5% Icodextrin, in place of Statistical analysis their usual bag of glucose solution for the overnight dwell, while The primary efficacy analysis of overnight ultrafiltration volume the control group continued their usual CAPD exchanges. Theand CAPD symptomatology was on an intent-to-treat basis using dialysis regimen remained unchanged except for three specialthe last values carried forward, thereby including all data available study weeks when all patients underwent 12-hour overnight from withdrawn patients. Secondary analyses, excluding patients dwells. These were weeks 4, 13, and 21. The patients were allowedwho violated the entry criteria (9 C, 4 D) and excluding those who to change the concentration or the number of daytime glucose also had variable bag regimens during the study (39 C, 45 D) were exchanges to maintain adequate fluid balance. All changes wereundertaken, but are not presented here as the results were recorded. qualitatively identical to the primary analysis. Patients continued with their usual medication and no specific The main parameter for efficacy was the volume produced after alterations to diabetic treatment were recommended other thanthe overnight dwell at 12 hours (special weeks 4, 13 and 21) and the usual stringent care of diabetic control. All changes during theat 8 hours (weeks 3, 12 and 20). The median value over the week study were recorded. Patients could withdraw permanently at anywas taken for each patient to avoid undue influence of outlying time, and were allowed up to four weeks temporary withdrawalvalues (weeks with 3 values or fewer were excluded, and the (holidays abroad, intra-abdominal complications). Peritonitis perequivalent "last value" used), and the mean of the medians in se was not a reason for withdrawal. each group were compared by repeated measures analysis of variance. Patients were subdivided into a "weak bag group" or a Monitoring "strong bag group" according to whether they used 1.36% or In each center, a Nurse Monitor was employed to co-ordinate3.86% glucose solutions for the overnight dwell, respectively, the study. The study treatment began at visit 2; patients were seenduring the baseline period (between randomization and the start every other week for the first two months and every four weeks forof treatment). the remaining time on the study. Body weight, blood pressure and For safety, adverse events, clinical and laboratory measure- changes in medication were recorded at each visit. Patients kept aments, all available data were included in the analyses. diary throughout the study and recorded the weights of bags All tests were two-sided and conducted at the 5% level of before and after each overnight dwell, as well as the duration ofsignificance. Significance levels for laboratory data were adjusted exchanges. For the three special weeks, bag weights were recordedfor multiple testing by a modified method of Bonferroni [15]. for all exchanges. Results Blood samples for hematology, standard biochemistry (including osmolality) were collected at each visit; serum samples for Of the 209 patients randomized from the 11 centers, 106 were estimation of Icodextrin and oligosaccharides were obtainedallocated to Icodextrin (D) and 103 to glucose (C). Of the total, 26 immediately prior to the study and at one, three and six months.were diabetic (11 C, 15 D). The patients in the C and D groups Fasting lipids, residual creatinine clearance, CXR, ECG andwere well matched for age, sex, primary renal disease, duration of ophthalmological assessment (visual acuity, slit lamp examination,dialysis and medical complications. The peritonitis rate prior to 498 MIDASStudy Group: Icodextrin in CAPD

Table 1. Patient characteristics Table 2. Number of patients withdrawn before and during treatment Control Icodextrin Control Icodextrin No. patients randomized 103 106 BeforeDuringBeforeDuringTotal Males 67 71 Loss of ultrafiltration 1 1 2 3 7 Females 36 35 Cather related problems 0 1 1 3 5 Age years 1 8 15 Mean SD 55 15 55 14 Adverse events (e.g. 0 6 peritonitis) Range 22—82 18—77 Withdrawals > 4 weeks 0 5 0 3 8 70.5 14.2 71.4 13.4 Weight kg (Mean SD) Patient request 1 2 2 4 9 CAPD profile 0 2 0 1 3 Duration of CAPD therapy months 27.6 25.6 24.7 19.8 Non-compliance Transplant 0 7 3 6 16 (mean sn) Transfer to hemodialysis 0 2 0 1 3 Exchange vol Other 2 2 1 0 5 2.0 liters 91 89 1.5 liters 12 17 Total 4 28 10 29 71 32 31 Overnight hypertonic exchanges (No. 32 39 patients) Episodes of peritonitis/patient months 1:21 1:15 Residual renal function C. > 2 mi/mm 37 38 Medical history (No. patients) Obstructive airways disease 8 10 Hypertension 35 45 Coronary heart disease 14 20 +329 to +590; P < 0.0001) and no different from that of 3.86% Peripheral vascular disease 2 5 G at eight hours (510 48 vs. 448 60 ml, 95% CI for the Primary renal disease difference —102 to +226 ml; P = and at 12 hours (552 44 Glomerulonephritis 16 17 0.44) Polycystic kidney disease 10 10 vs. 414 78 ml, 95% CI for the difference —47 to +325 ml; P = Hypertension 24 24 0.06).The numbers of patients in each of these groups are shown Chronic pyelonephritis 10 11 in Figure 1. Extending the dwell time from 8 to 12 hours tended Diabetes mellitus 8 11 to reduce net UF with glucose but to increase it with Icodextrin Heredofamilial diseases 2 0 Unknown 33 33 (Fig. 1). In contrast to Icodextrin, 9 to 41% of patients on glucose experienced net overnight reabsorption of intraperitoneal fluid (negative UF; Fig. 2). The mean UF per week was relatively constant in both groups but between-patient variability was higher the study was, however, 1.4 times higher in the D group (Table 1).in the C compared to the D group throughout the study. Sixty-seven percent of the CAPD population studied used at least The overnight use of Icodextrin did not influence the daytime one 3.86% glucose ("strong") exchange per 24 hours. The overallUF with glucose solutions: in weeks 4, 13 and 21 the daytime UF and overnight usage of strong glucose was similar in the twovolumes were (mean SEM) 352 46, 331 51 and 331 51, groups. respectively, in the C group and 33155, 361 62 and 311 66, Of the C group, there were 54 and 35 in the weak and strongrespectively, in the D group (NS). The daytime use of 3.86% subgroups, respectively, while in the D group there were 46 andglucose exchanges was lower in the D patients compared to the C 30. A few patients had no consistent pattern of weak or strong bagin the weak bag group (1.2 0.3 vs. 2.4 0.3 bags/wk), but the use (4 C, 9 D) or had not completed diaries (6 C, 11 D) orsituation was reversed in the strong bag group (2.70.31 vs. 1.9 withdrew during the baseline period (4 C, 10 D), were not 0.3 bags/wk), despite similar overnight UF with Icodextrin. allocated to either of these subgroups. Solute transport. The transperitoneal transport as assessed by A total of 71 patients withdrew during the study, of whom 14 (4serial measurement of solute plasma concentration shows that C, 10 D) did so between randomization and starting treatment; 57small molecular weight solutes such as potassium, urea, creati- (28 C, 29 D) withdrew during treatment. The most commonnine, calcium, phosphate, uric acid, as well as large Mw solutes, reasons for withdrawal were transplantation and cardiovascularsuch as albumin and total protein, were similar in the two groups events (Table 2). When major and minor adverse events werebut a small, statistically significant fall in serum sodium and assessed for relationship to Icodextrin treatment, no specificchloride levels was observed irs the D group compared to the C association was observed except in one patient with cardiac failure(Table 3). in whom Icodextrin mediated ultrafiltration was lower than 3.86% Lipids. Lipid profiles were similar in both groups at entry. In the glucose and may have contributed to fluid overload. Dextrin group, however, the cholesterol, triglycerides and LDL One hundred and thirty-eight patients (71 C, 67 D) completedlevels fell by 6 to 10% of the pretreatment values, but these the six month study (66% of randomized; 72% of those whochanges did not achieve statistical significance. started treatment). Safety Efficacy Ultrafiltration,The mean (± SEM) overnight ultrafiltration (UF) Icodextrin absoiption and serum levels. The total carbohydrate with D was 3.5 times greater than 1.36% glucose (G) at eightabsorbed from the eight hour overnight exchanges was deter- hours [527 36 vs. 150 47 ml; 95% confidence interval (CI) formined in 18 (7 C, 11 D) patients, by subtracting the amount of the difference +257 to +497 ml; P < 0.0001] and 5.5 times greatercarbohydrate in the dialysate at the end of the dwell from that at 12 hours (561 44 vs. 101 48 ml, 95% CI for the differenceinfused. All control patients used strong bags overnight. During MIDAS Study Group: Icodextrin in CAPD 499

A8hrdwell B l2hrdwell 800 800

600 600

400 400

200 200

0 0 3 12 20 4 13 21 Week Week 800 800

600 600

400 400

Fig. 1. Net overnight ultrafiltration with glucose 200 200 and Icodextrin exchanges. The volume is expressed as the mean of the medians in each group with last values carried forward. Symbols 0 0 are: (LI) 1.36% glucose; ()Icodextrin;(11) 3 12 20 4 13 21 3.86%glucose. Error bars =SEM; < 0.001. Numbers within bars represent patients in each Week Week group.

A 8hrdwell B l2hrdwell a)> 40 40 C) — ,30 — 30 20 20 100 -—-- 010 3 12 20 4 13 21 Week Week >a) 40 40 C) 30 30 cCw 20 20 .g2U- 10 DQ. 10 0 I' 0[]L'.1 J Fig.2. Percentage of patients with negative ultrafiltration during overnight dwell. Numbers of 3 12 4 13 21 20 patientsare as in Fig. 1. Symbols are: (LI) 1.36% glucose; ()Icodextrin;()3.86% Week Week glucose. glucoseexchanges 86% of the initially infused carbohydrate was Figure 3 shows circulating levels of the total Icodextrin fractions absorbed compared with 20% of Icodextrin. The mean (± SEM)(maltose two high molecular weight fractions; G2-HMW) and mal- carbohydrate absorption was lower with Icodextrin (29 5g) tose (G2) alone over a period of six months. The mean level of than with 3.86% glucose (62 5g; P < 0.01). Icodextrin and its metabolites (total Icodextrin) increased from a 500 MIDASStudy Group: Icodextrin in CAPD

Table3.Biochemical profile during glucose (C) and Icodextrin (D) exchanges (mean SEM) Solute Pre Week 2 Week 4 Week 8 Week 12 Week 16 Week 20 Week 24 p Weight kg C 70.5 1.4 71.0 1.5 70.81.5 71.0 1.6 71.0 1.7 71.1 1.7 71.6 1.8 71.41.9 D 71.41.4 70.9 1.5 71.01.5 70.7 1.6 71.0 1.6 70.5 1.7 70.3 1.7 70.1 1.6 NS Hbg/dl C 9.8±0.2 9.6±0.2 9.7±0.2 9.7±0.2 9.7±0.2 9.8±0.2 9.9±0.2 9.9±0.2 D 10.1 0.2 9.90.2 9.8 0.2 9.9 0.2 9.9 0.2 10.20.2 10.30.3 10.4 0.2 NS Sodium C 139 0.3 139 0.4 139 0.4 139 0.4 138 0.4 139 0.4 138 0.4 138 0.4 mmol/liter D 140 0.3 137 0.4 136 0.3 137 0.3 136 0.3 136 0.3 136 0.3 136 0.3 <0.05 Chloride C 103 0.5 102 0.5 103 0.5 103 0.6 103 0.6 103 0.7 102 0.7 103 0.7 mmol/liter D 103 0.7 99 0.7 99 0.7 100 0.7 100 0.5 99 0.6 98 0.7 99 0.6 <0.05 Urea C 22.90.5 23.3 0.5 22.9 0.5 23.20,6 22.8 0.6 23.00.6 22.80.6 22.9 0.6 mmol!liter D 22.00.6 21.80.6 22.4 0.7 21.40.6 21.7 0.6 21.90.7 22.50.6 22.5 0.7 NS Cr pmol/liter C 1008 23 1013 24 1018 25 1023 26 1031 27 1034 27 1044 27 1052 30 D 974 27 943 31 961 32 931 28 935 28 950 30 966 31 974 32 NS Phosphate C 1.75 0.05 1.700.04 1.75 0.05 1.720.04 1.750.05 1.75 0.05 1.72 0.05 1.80 0.05 mmol/liter D 1.71 0.04 1.650.04 1.69 0.05 1.630.05 1.630.05 1.62 0.05 1.620.041.66 0.04 NS Albumin C 36 0.5 35 0.5 36 0.5 36 0.5 360 0.5 36 0.5 35 0.5 36 0.5 giliter D 36 0.4 35 0.4 35 0.4 35 0.4 35 0.5 36 0.4 35 0.5 36 0.5 NS Plasma osmolality C 310 1.1 310 1.0 309 1.0 309 1.2 308 1.1 310 1.1 310 1.2 311 1.4 mOsm/kg D 310 1.0 311 1.1 312 1.4 311 1.1 311 1.1 313 1.2 313 1.1 313 1.3 NS

Table 4. Major adverse events Control Icodextrin

CCl) Myocardial infarction 4 2 0 Cardiac failure 1 1 C.) 2 Cl) Cardiovascular accident Pneumonia — 1 C Uncontrolled hypertension — 1 Pulmonary embolus — 1 Multiple arterial emboli 1 — Peritonitis episodes/patient month Pre-MIDAS 1:21 1:15 0 4 8 12 16 20 24 During the study 1:16 1:13 Week Fig.3. Semm levels of total Icodextrin fractions (—U—)andmaltose (G2; ——D—'—)(mean SEM). of dry weight. The mean number of non-CAPD related medical events (that is, all other medical events not classified as pertaining to CAPD) in the C and the D groups were 3.1 and 2.6 per patient, respectively. Ophthalmological assessments before and after baseline value of 0.35 0.02 g/liter to a steady state level of 4.87treatment did not show any pattern of abnormality in the D group 0.17 glliter and remained stable throughout the study. Thecompared to the C. serum maltose followed an identical pattern and rose from 0.04 The peritonitis rate during the study was higher in the Icodex- 0.004 glliter to a steady state level of 1.20 0.043 g/liter. trin group, consistent with the higher rate at entry, but there was Osmolality. The use of Icodextrin solution resulted in a slight no significant difference between the two groups (Table 4). rise in serum osmolality compared to the control group but this change was not statistically significant (Table 3). Diabetic patients Liver ,funcsion tests. Liver function tests performed during the A subgroup analysis of diabetic patients showed a similar study remained normal in both groups. pattern of changes with respect to ultrafiltration, biochemistry, Medical profile and symptomatology. Most patients toleratedpolymer accumulation and symptom profile to non-diabetic pa- their treatment well. Major adverse events were similar in the twotients. The overall diabetic control and insulin requirements were groups, and were mainly related to pre-existing cardiovascularnot different in the two groups. disease (Table 4). Two deaths occurred during the study, both in the C group; one patient died in the D group a week after Discussion withdrawal from the study, due to cerebrovascular accident This study, involving approximately 5%ofthe UK CAPD (CVA). The numbers of patients receiving erythropoietin therapypopulation, was the first large long-term, randomized controlled were similar in the two groups (24 C, 25 D). study undertaken to compare the efficacy and safety of a new There was a significant (P < 0.05) reduction in the overallosmotic agent, Icodextrin, with conventional glucose solutions. It CAPD-related symptom score at the end of the six month studydemonstrated that the overnight use of Icodextrin, which is period in the D group compared to the C (Figure 4). Theisosmolar, was well tolerated and produced ultrafiltration superior symptomatic improvement was particularly noted in abdominalto 1.36% glucose, and no difference in the magnitude to hyper- fullness, tiredness, shortness of breath and effective maintenancetonic 3.86% glucose solution was found. pathways oftransperitonealabsorptionosmoticagents.The cause forthisremainsuncertainbutitprobablyrelatestodifferent tion comparedtothecontrolsisofsomeinterest.Theprecise However, amuchsmallerbetween-patientvariationinultrafiltra- group, ontheotherhand,ispredictablebasisoflarge The reduceddaytimeuseof3.86%exchangesintheweakbag bag perweek)inthepatientsusingIcodextrinstrong was asmallriseindaytimeuseof3.86%glucoseexchanges(one was particularlymarkedfor1.36%exchanges,butalsooccurred in study. Symbolsare:(fl)1.36%glucose;(•) [19—22]. Inpreviousstudies,thetransperitonealabsorptionof ular absorptionisrelativelyconstantthroughthelymphaticsystem subject toconsiderablevariability[17,18],whereasmacromolec- absorption ofsmallMwglucosethroughtheperitoneal"pores"is tion ofultrafiltrationratesandsolutetransportcharacteristics. influence onperitonealfunctionasjudgedbysequentialestima- improvement inovernightultrafiltrationwithIcodextrin. for theuseofhypertonicexchangesandrestrictionfluidintake. group. Thismaybeduetobetween-centervariationinthecriteria on averagenodifferentfrom3.86%glucosesolutions,butthere sensation ofabdominalfullnessintheIcodextringroup. sustained overalongerperiod[7],whichmayexplainthereduced a lowerrateofultrafiltrationattheonsetdialysisbutbeing ultrafiltration kineticsofIcodextrindifferfromglucoseinhaving and carbohydrateloadareevengreater[6].Furthermore,the 3.86% glucosewheretheadditionalbenefitsintermsofosmolar effectively replaceallovernightglucoseexchanges,especially reabsorption. This supportsourprevious viewthatitcould ultrafiltration associatedwith Icodextrin solutionledtonegligible some patientsusinga3.86%solution.Incontrast,thesustained overnight dwellinasubstantialproportionofpatients[7,161. This intraperitoneal volume(negativeultrafiltration)duringthelong glucose solutions(2to4hr)resultedinnetreabsorptionof Fig. The overallovernightultrafiltrationachievedbyIcodextrinwas In thisstudy,thelong-termuseofIcodextrinhadnodetectable As expected,theshortdurationofeffectiveultrafiltrationwith score < 0.05betweengroups. Symptom 4. Theoverallsymptomscoreineachgroupbeforeandattheendof 0 a 0) Week Icodextrin. MIDAS StudyGrout:IcodextrininCAPD Error bars= SEM, lality significantlyduringthecourseofstudy. 3.5 mmol/liter,whichwasinsufficienttoaffecttheserumosmo- tration ofIcodextrinandallitsmetabolitesamountedtolessthan to hyperosmolality.Inthepresentstudy,combinedconcen- molar concentrationperunitmassandthusarelesslikelytolead molecular sizesgreaterthanglucose,generateamuchlower generation ofmaltoseandisomaltose.Thesefractions,having total serumosmolalityisnegligibleduetoitslowabsorptionand osmotic agents[37,38],therelativecontributionofIcodextrinto further study. non-enzymatic tissueglycationisunknownandthesubjectof maltose [36].Whethertheelevatedlevelofiscapable important ratelimitingstepintheintracellularmetabolismof impermeability ofthehumancellularmembranerepresentsan maltose diffusesreadilyacrosssomecellwalls,buttherelative storage ortissuedepositionisunlikely.Ithasbeenshownthat isomaltase defect[34,35],andsothepossibilityofglycogen renal failurethereisnoevidenceofanintracellularmaltase/ larly transportedmaltosewouldbecompletelymetabolized.In present intracellularlywithinlysozymes[31—33].Thus,intracellu- lular accumulationofmaltoseinuremia,theenzymeisabundantly maltase [30].Althoughalackofrenalleadstoextracel- readily metabolizedbyintra-andextracellularamylase consisting mainlyofonetofouraglucosidicbondswhichare relatively undegradable,Icodextrinisahighlysolubleproduct polysaccharides suchasdextranandhydroxyethylstarchwhichare that thismayoccurwithIcodextrin.However,incontrastto the dailyuseofIcodextrin.Themajoritythemhavebeenusing open natureofthestudy.However,itisrelevantthatatend profile, butnofirmconclusioncouldbedrawnfromthisduetothe patients. Duringthisperiodnoadverseclinicaleffectswere with thishypothesis. which isinconsistentwithlymphaticuptake[6,7,14].Asimilar accumulation ofdextran[26—28]andhydroxyethylstarch[29], it forupto24monthswithoutadverseeffects. the studymostpatients(72%)ingroupoptedtocontinuewith tients intheIcodextringroupreportedanimprovedsymptom ascribed toIcodextrinoritsmetabolites.Interestingly,thepa- represent thelongestexposureofthesesubstancesinuremic uremic patients. there arenodetailedmetabolicstudiesofmaltoseavailablein observed. However,incontrasttonormalsubjects[8,23—25], [6, 7]maybeinsufficienttoaccountforthetotalclearance of maltose(3.5ml/min)duringdaytimeglucoseexchangesalone tients asthepreviouslycalculatedsmalltransperitonealclearance that someextrarenalmaltosemetabolismoccursinuremicpa- products andtheirperitoneal/metabolicclearances.Itispossible librium isreachedbetweentheabsorbedIcodextrin,degradation longer durationofthisstudy.Thisimpliesthatadynamicequi- Furthermore, steadystatelevelsweremaintaineddespitethe reach steadystatelevelsformaltoseandlargerfractions[141. ence bothinthemagnitudeofriseandtimeintervalto safety analysisinthisstudy.Itisconsistentwithpreviousexperi- breakdown productsrepresentsanimportantcomponentofthe finding inthesubgroupofpatientsthisstudyisaccordance Icodextrin hasvariedbetween24and28%over8to12hours, In contrasttopoorlymetabolizedsmallmolecularweight There hasbeenconcernexpressedregardingpossibletissue The levelsofmaltoseandoligosaccharidesoversixmonths The patternofsystemicaccumulationIcodextrinandits 501 502 MIDAS Study Group: Icodextrin in CAPD

However, the precise cause of hyponatremia in the IcodextrinLtd. (Sheffield), and Alison Scrimgeour for data handling and statistical group remains uncertain. One possible explanation may be that analysis, and Linda Prior for administrative support. accumulation of Icodextrin and its metabolites, though insufficient Reprint requests to Dr. C.D. Mist.'y, Institute of Nephrolo, Cardiff Royal to affect total serum osmolality (isosmolar), are relatively imper-Infirmary, Newport Road, Cardiff: CF2 JSZ Wales, United Kingdom meable across the cell membrane (with high refection coefficient) and may induce osmotic flow by a mechanism similar to isosmolar References "colloid" osmosis. Perhaps the flow of water from intracellular to 1. RAINE AEG, MARGREITER R, BRUNNER FP, EHRICH JHH, GEERING extracellular compartment could lead to dilutional hyponatremia. W, LANDAIS P, LOIRAT C, MALLICK NP, SELWOOD N, TUFVESON F, Alternatively, according to the recent "three-pore" theory of VALDERRABM4O F: Report on management of renal failure in Europe, transperitoneal fluid and solute transport proposed by Stelin and XXII, 1991. Nephrol Dial Transplant (Suppi 2):7—35, 1992 2. HAIN H, KESSEL M: Aspect of new solutions for peritoneal dialysis. Rippe [39], up to 50%oftransperitoneal ultrafiltration induced by Nephrol Dial Transplant 2:67—72, 1987 glucose (radius 3 A) occurs via the transcellular pores (4 to 6 A) 3. DUWE AK, VAS SI, WEATHERHEAD JW: Effect of the composition of not coupled to solute transport. Thus, a substantial sieving of peritoneal dialysis fluid and chemiluminescence, phagocytosis and sodium gives rise to hyponatremic dialysate in the early phase of bacterial activity in vitro. Infect Immun 33:130—135, 1981 4. D.zo-Buxo JA, FARMERCD,CI-IANDLER JT, WALKER PJ, BURGESS glucose based dialysis. In contrast, the large molecular weight WP: CCPD—"wet" is better than "dry." Pent Dial Bull 7 (Suppl):S22, Icodextrin (radius 20 to 30 A) probably mediates a large propor- 1987 tion of transcapillary ultrafiltration through the small intracellular 5. TWARDOWSKI ZJ, KHANNA R, NOLPH KD: Osmotic agents and pores (40 to 55 A) with relatively less sieving of sodium and ultrafiltration in peritoneal dialysis. Nephron 42:92—101, 1986 therefore larger dialysate loss. Nevertheless, the small reduction 6. MISTRY CD, MALLICK NP, GOKAL R: Ultrafiltration with isosmotic solution during long peritoneal dialysis exchanges. Lancet ii:178—182, in serum sodium, although statistically significant, does not appear 1987 to be clinically relevant as no adverse effects have been attributed 7. MISTRY CD: Glucose polymer as an osmotic agent in continuous to it in this study or during the subsequent long-term use of 24 peritoneal dialysis (MD thesis). London, University of London, 1989, months (unpublished observation). pp 1—300 The maltose and related fractions do not affect glucose metab- 8. WESER E, SLEISENGERMH:Metabolism of circulating disaccharides in man and in the rat. J C/in Invest 46:499—505, 1967 olism nor lead to hyperinsulinemia [24,40]. It is therefore possible 9. VAN HANDLE E: Trehalase and maltase in the serum of vertebrates. that this may have a favorable impact on the lipid profile. Comp Biochem Physiol 26:555—561, 1968 Although, the cholesterol levels showed some reduction during 10. BITFENCOURT H, SLEISENGERMH,WESER E: Studies of serum and the study, this did not reach statistical significance. It seems that tissue maltase in the rat. Gastroentrology 57:410, 1968 11. DREYFUS JC, ALEXANDRE Y: Electrophoretic characterization of acid a single isosmolar exchange per day is not sufficient to produce a and nuteral amylo-al,4 glucosidase (acid maltase) in human tissues detectable change, and perhaps multiple exchanges over a longer and evidence for electrophoretic variants in acid maltase deficiency. period of observation are required. The results of the MIDAS Biochem Biophys Res Commun 48:914, 1972 study in the subgroup of patients with diabetes will be the subject 12. OHNEDA A, YAMAGATA 5, Tsursusi K, FUJIWARA H: Distribution of maltose intravenously administered to rabbits and its metabolism in of a separate report. the kidney. Tohokv J Exp Med 112:141—154, 1974 This large multicenter, randomized controlled study has dem- 13. STEVENSON FK: The disaccharidase activity of a membrane fraction onstrated that an isosmolar Icodextrin solution was well-tolerated obtained from rabbit cortex. Biochim Biophys Acta 266:144—153, 1972 over six months and was as effective as 3.86% glucose over long14. MISTRY CD, GOKAL R: A single daily overnight (12 hr dwell) use of dwell exchanges. In its present form, a daily overnight use would 7.5% glucose polymer (Mw 18,700; Mn 7,300) +0.35% glucose solution: A 3 month study. Nephrol Dial Transplant 8:443—447, 1993 obviate the need for hypertonic glucose exchanges, particularly15. Sin. Z: On probabilities of rectangles in multivariate student 3.86% glucose, but would be equally effective in any long dwell distributions: Their dependence on correlations. Ann Math Stat 42: regimes such as the daytime use in automated peritoneal dialysis 169—175, 1971 (APD). In addition, the safety aspect of this study marks an 16. PYLE WK, Popovici-i RP, MONCRIEF JW: Mass transfer evaluation in important advance towards the ultimate objective of developing a peritoneal dialysis, in CAPD Update, edited by MONCRIEF JW, Popo- VICH RP, New York, Masson, 1981, pp. 35—52 physiological solution suitable for all CAPD exchanges. Our17. STEIN G, RIPPE B: A phenomenological interpretation of the variation previous work suggested that an isosmolar combination of small in dialysate volume with dwell time in CAPD. Kidney Int 38:465—472, and large molecular agents is effective in producing optimal 1990 ultrafiltration for a given dwell time [41]. For shorter daytime 18. PYLE WK: Mass transfer in peritoneal dialysis. (Ph.D. dissertation) University of Texas, 1981 exchanges (4 to 6 hr) an isosmolar combination of only a quarter19. Rir'PE B, STEIN G, AHLMEN J: Lymph flow from the peritoneal cavity of the concentration of currently used Icodextrin and a half the in CAPD patients, in Frontiers in Peritoneal Dialysis, edited by MAHER 1.36% glucose may be sufficient. It would have the advantage of JF, WINCHESTER JF, New York, Field, Rich and Assoc. Inc., 1986, pp. replacing all the currently used hyperosmolar exchanges with 24—30 20. DAUGIRDAS JT, ING TS, GANDHI VC, HANO JE, CHEN W, YUAM L: physiological solutions containing half the daily calorie load Kinetics of peritoneal fluid absorption in patients with chronic renal without exceeding the current level of maltose accumulation. failure. J Lab C/in Med 95:351—361, 1980 21. FLESSNER MF, DEDRICK RL, FENSTERMACHER JD, BLASBERG RG, SIEBER SM: Peritoneal absorption of macromolecules, in Frontiers in Peritoneal Dialysis, edited by MAHER JF, WINCHESTER JF, New York, Acknowledgments Field, Rich and Assoc. Inc., 1986, pp. 41—46 22. SPENCER PC, FARRELL PC: Solute and water transfer kinetics in This study was entirely supported by M. L. Laboratories plc. and conducted CAPD, in Continuous Ambulatoiy Peritoneal Dialysis, edited by Goi by Innovata Biomed Limited; Innovata Biomed Ltd. is a subsidiary of M.L. R, Edinburgh, Chirchill Livingstone, 1986, pp. 38—55 Laboratories plc. We thank Dr. J. Fox and Mrs. C. Joseph (Alta Bioscience, 23. YOUNG JM, WESER E: The metabolism of circulating maltose in man. University of Birmingham) for undertaking the Icodextrin assay, S-Cubed J C/in Invest 50:986—991, 1971 MIDAS Study Group: Icodextrin in CAPD 503

24. YoUNG EA, FLETCHER iT, CIOLErFI LA, HOLLRAH LA, WESER E: 33. HERS HG: a-Glucosidase deficiency in generalised glycogen storage Metabolism of parenteral glucose oligosaccharides in man. J Parenter disease (Pompe's disease). Biochem J 86:11, 1963 Enteral Nutr 5:369—377, 1981 34. MCNAIR A, OLsEN J: Disaccharidase activity in chronic renal failure. 25. YOUNG EA, DRUMMOND A, COOL DA, CloLErn LA, CRAIN M, Acta Med Scan 195:93, 1974 TRAYLOR J, WESER R: The effect of insulin on the metabolism 35. MISTRY CD, Fox JE, MALLICK NP, GOKAL R: Circulating maltose and of parenteral maltose in man. J Gun Endocrinol Metab 50:764—772, isomaltose in chronic renal failure. Kidney mt 32 (Suppl 22):S210— 1980 S214, 1987 26. BERGONZI G, PATIES C, VASSALLO G, Z.rG1wDI A, PolsErn PG, 36. EARL DP: An eighteenth century suggestion for peritoneal dialysis. mt BAi.occtil S, FONTANA F, ScitP1oNL L: Dextran deposits in tissues of JArtif Organs 3:67—68, 1980 patients undergoing haemodialysis. Nephrol Dial Transplant 5:54—58, 37. Ri'jA RM, KRAMER MS, MANCHANDA R, LAZARO N, ROSENBAUM JL: 1990 Peritoneal dialysis with fructose dialysate—Prevention of hyperglycae- 27. BULL JP, RICKrn-rs C, SQUIRE JR, MAYCOCK W, SPOONER WD'A, mia and hyperosmolaity. Ann Intern Med 79:511—517, 1973 MOLLISON PL, PATrERSON JCS: Dextran as a plasma substitute. 38. MArrY E, DOLKART R, LAMEIRE N: Potential hazards with the use of Lancet i:134—143, 1949 glycerol dialysate in diabetic CAPD patients. Pent Dial Bull 7:16—20, 28. VICKERY AL: The fate of dextran in tissues of the acutely wounded. A 1987 study of the histological localization of dextran in tissues of Korean 39. STELIN G, RIPPE B: A phenomenological interpretation of the varia- battle casualties. Am J Pathol 32:161—178, 1956 tion in dialysate volume with dwell time in CAPD. Kidney mt 29. THOMPSON WL, FUKUSHIMA T, RUTHERFORD RB, WALTON RP: 38:465—472, 1990 Intravascular persistence, tissue storage and excretion of hydroxyethyl 40. MISTRY CD, MALLICK NP, GOKAL R: The use of large molecular starch. Surg Gynecol Obstet 13 1:965—972, 1970 weight glucose polymer as an osmotic agent in continuous ambulatory 30. GAMKLOU R, SCHERSTEN T: Activity of a-amylase and a 1, 4-glucosi- peritoneal dialysis, in Advances in Continuous Ambulatory Peritoneal dase in human liver tissue. Scan J Clin Invest 30:201—207, 1972 Dialysis 1986, edited by KHANNA R, NOLPH KD, PROWANT B, TWAR- 31. LUNDQUIST A, OCKERMAN PA, SCHERSTEIN B: Fine needle aspiration DOWSKI ZJ, ORE0P0UL05 DG, Toronto, University of Toronto Press, biopsy of the liver in healthy adults. Activity of enzyme of glycogen 1986, pp. 7—11 utilisation and glucose production. Enzymol Biol Gun 10:8, 1989 41. MISTRY CD, Go.i. R: Can ultrafiltration occur with a hyposmolar 32. ToRtas HN, OLAVARRIA JM: Liver a-glucosidases. J Biol Chem solution in peritoneal dialysis?: The role for "colloid" osmosis. Gun 239:2427—2434, 1964 Sci 85:495—500, 1993