Peritoneal Dialysis International, Vol. 30, pp. 611–618doi:10.3747/pdi.2009.00161

0896-8608/10 $3.00 + .00Copyright © 2010 International Society for Peritoneal Dialysis

611

EFFECTS OF SOY CONSUMPTION ON SERUM LIPIDS AND APOPROTEINS INPERITONEAL DIALYSIS PATIENTS: A RANDOMIZED CONTROLLED TRIAL

Hadi Tabibi,1 Hossein Imani,2 Mehdi Hedayati,3 Shahnaz Atabak,4 and Leila Rahmani5

Department of Human Nutrition,1 Faculty of Nutrition Sciences and Food Technology,2 National Nutrition andFood Technology Research Institute; Prevention & Treatment of Obesity Research Center,3 Research

Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences & HealthServices; Department of Nephrology,4 Shahid Modares Hospital,5 Shahid Beheshti

University of Medical Sciences & Health Services, Tehran, Iran

Correspondence to: H. Tabibi, Department of Human Nu-trition, Faculty of Nutrition Sciences and Food Technology,National Nutrition and Food Technology Research Institute,46 West Arghavan St., Farahzadi Blvd., Shahrak Qods, P.O. Box19395-4741, Tehran, Iran.

hadtabibi@yahoo.comReceived 7 August 2009; accepted 4 December 2009.

♦♦♦♦♦ Background: Lipid abnormalities, particularly high serumconcentration of lipoprotein(a) [Lp(a)], are one of the majorrisk factors for cardiovascular disease (CVD) in peritonealdialysis (PD) patients. The present study was designed toinvestigate the effects of soy consumption on serum lipidsand apoproteins, especially Lp(a), in PD patients.♦♦♦♦♦ Methods: This study was a randomized clinical trial inwhich 40 PD patients (20 males, 20 females) were randomlyassigned to either the soy or the control group. Patients inthe soy group received 28 g/day textured soy flour (con-taining 14 g of soy protein) for 8 weeks, whereas patientsin the control group received their usual diet, without anysoy. At baseline and the end of week 8 of the study, 5 mL ofblood was collected from each patient after a 12- to 14-hourfast and serum triglyceride, total cholesterol, low densitylipoprotein-cholesterol (LDL-C), high density lipoprotein-cholesterol (HDL-C), apoprotein B100 (apo B100), apopro-tein AI (apo AI), and Lp(a) were measured.♦♦♦♦♦ Results: In the present study, serum Lp(a) concentra-tions were above the normal range in 86% of the PD pa-tients. Mean serum Lp(a) concentration was reducedsignificantly, by 41%, in the soy group at the end of week 8compared to baseline (p < 0.01); the reduction was also sig-nificant compared to the control group (p < 0.05). Duringthe study, mean serum Lp(a) concentration did not changesignificantly in the control group. There were no significantdifferences between the two groups in mean changes inserum triglyceride, total cholesterol, HDL-C, LDL-C, apoB100, or apoAI.♦♦♦♦♦ Conclusion: The results of our study indicate that soy con-sumption reduces serum Lp(a) concentration, which is a

risk factor for cardiovascular disease in peritoneal dialysispatients.

Perit Dial Int 2010; 30:611–618 www.PDIConnect.comepub ahead of print: 8 Apr 2010 doi:10.3747/pdi.2009.00161

KEY WORDS: Serum lipids; apoproteins; lipopro-tein(a); soy.

The most important cause of mortality in patients withchronic renal failure, including dialysis patients, is

cardiovascular disease (CVD). The frequency of CVD in di-alysis patients has been reported as high as 3 – 45 timesthat observed in the general populations and approxi-mately 50% of deaths in these patients are related toCVD (1,2). In peritoneal dialysis (PD) patients, one ofthe major risk factors for CVD is lipid abnormalities, in-cluding low serum concentrations of high density lipo-protein-cholesterol (HDL-C) and apoprotein (apo AI),and high serum concentrations of triglycerides, totalcholesterol, low density lipoprotein-cholesterol (LDL-C),apoprotein B100 (apo B100), and particularly lipopro-tein(a) [Lp(a)] (3–10). Serum Lp(a) concentrationhigher than 30 mg/dL, a prevalent lipid abnormality inPD patients, is an important risk factor for CVD (3,5,9–15). At the present time, blood lipid normalizing drugssuch as statins, nicotinic acid, fibric acid derivatives, etc.are used to treat lipid abnormalities and prevent CVD inPD patients; however, no effective treatment to reduceLp(a) in these patients is known so far (3).

Some studies, conducted mainly in nonuremic pa-tients, indicate that soy consumption reduces serum con-centrations of total cholesterol, LDL-C, triglycerides, andapo B100 and raises serum HDL-C concentration, particu-larly in hyperlipidemic patients (16–26). In addition, lim-ited studies have been conducted to assess the effect of

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

612

TABIBI et al. NOVEMBER 2010 – VOL. 30, NO. 6 PDI

asked to continue their usual diet and to avoid consum-ing soy-containing products during the study. Subjectswere advised not to change their dietary habits, physi-cal activities, or drug regimens. All CAPD patients wereusually dialyzed with four 2-L exchanges of 1.5% and/or 2.5% glucose-containing dialysis solutions, except for1 patient who used 4.25% glucose-containing dialysissolutions. At baseline and the end of week 8, 5 mL ofblood was obtained from each patient after a 12- to14-hour fast. Blood samples were centrifuged at2500 rpm for 15 minutes to separate sera. Serum con-centrations of triglycerides, total cholesterol, LDL-C, andHDL-C were measured enzymatically using commercialkits (Pars Azemoon, Tehran, Iran) with the aid of aHitachi 717 automated analyzer (Boehringer MannheimDiagnostics, Mannheim, Germany). Serum concentra-tions of apo B100, apo AI, and Lp(a) were determined byimmunoturbidimetric methods using commercial kits(Roche Diagnostics, Mannheim, Germany) with the aidof a Cobas Mira automated analyzer (Roche Diagnostics).Serum albumin concentration was measured by thebromcresol green method using commercial kits (ParsAzemoon) with the aid of a Hitachi 717 automated ana-lyzer. The coefficient of variation for serum lipids,apoproteins, and albumin was less than 3%. Serum tumornecrosis factor-alpha (TNF-α) concentration was deter-mined with the use of enzyme-linked immunosorbentassay (ELISA) kits (Bender MedSystems, Vienna, Aus-tria). The coeff icient of variation for serum TNF-αwas 6%.

Patients were weighed at baseline and at the end ofweek 8. In addition, because dietary confounding fac-tors could affect blood lipids and apoproteins, it wasnecessary to compare the two groups regarding thesefactors during the study. Therefore, dietary intakes ofsubjects were assessed using a 24-hour dietary recall for3 consecutive days at baseline and at the end of weeks 4and 8. Patients’ diets were analyzed by Food ProcessorII software (ESHA Research, Salem, OR, USA). To ascer-tain patients’ compliance, we provided each patient witha fixed number of soy packets and instructions to returnthe unconsumed packets at the end of the study. The de-gree of compliance for each patient was determined ac-cording to the number of returned packets and the dataof the patients with compliance less than 90% were notanalyzed.

STATISTICAL METHODS

Statistical analysis was performed using the Statisti-cal Package for the Social Sciences (SPSS, Inc., Chicago,IL, USA) version 12.0. Chi-square test was used to com-

soy consumption in lowering serum Lp(a) concentration(18–20,22,27,28), and the majority of these studies,which were randomized controlled trials (18,20,22,27),have shown that soy protein does not significantly re-duce serum Lp(a) (18–20,22,27). These studies were con-ducted mainly in nonrenal failure individuals with normalserum level of Lp(a). However, Meinertz et al. showed ina randomized controlled trial that alcohol-extracted soyprotein in normolipidemic non renal failure individualsmarkedly lowered serum Lp(a), especially in individualswith serum Lp(a) level higher than 30 mg/dL (28). Ac-cording to the available literature, no such study in thisfield has been done yet in PD patients. Therefore, thepresent study was designed to investigate the effects ofsoy consumption on serum lipids and apoproteins, es-pecially Lp(a), in PD patients.

METHODS

This study was a non-blinded randomized clinical trial.The minimum sample size estimated for each group was20 at a power (1 – β) of 90% and α = 0.05 for a two-armparallel study with two-tailed testing to detect a differ-ence of 41 mg/dL in serum Lp(a) concentration, with apooled standard deviation of 40.1 mg/dL — obtainedfrom Tonstad et al.’s study (18). Forty patients (20 men,20 women) undergoing continuous ambulatory PD(CAPD) in the age range of 18 – 83 years (mean ± SD:52 ± 15 years) were recruited from the PD units ofModares Hospital and Shafa Clinic in Tehran, Iran. Pa-tients enrolled in this study did not have thyroid dis-orders or hyperphosphatemia (serum phosphorus>5.5 mg/dL). In addition, all patients were asked aboutsoy consumption at baseline and subjects that had regu-larly used soy in their diets within 3 months before thestart of the study were excluded. The study protocol wasapproved by the Ethics Committee of the National Nutri-tion and Food Technology Research Institute of Iran.Written informed consent was obtained from all patientsbefore initiating the study.

Patients were randomly allocated to either the soy orthe control group using a block randomization. Since thestudy period in the majority of investigations of the ef-fects of soy consumption on blood lipids and apoproteinswas between 4 and 12 weeks, mainly 6 weeks, 8 weekswas therefore chosen as the study period. Subjects inthe soy group received 28-g packets of raw textured soyflour (containing 14 g of soy protein and 61 mg ofisoflavonoids, including genistein, daidzein, andglycitein) and were asked to cook and consume one soypacket instead of one serving of meat (i.e., 60 g of meat)at dinner for 8 weeks. Patients in the control group were

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

613

PDI NOVEMBER 2010 – VOL. 30, NO. 6 EFFECTS OF SOY ON LIPID PROFILE IN PD PATIENTS

pare qualitative variables between the two groups. Be-cause all quantitative parameters had normal distribu-tion according to the Kolmogorov–Smirnov test, we usedthe t-test and the paired t-test to compare parametersbetween and within groups respectively. In addition, be-cause dietary factors were measured three times duringthe study, analysis of variance for repeated measure-ments was used to compare the data between varioustimes. Results are expressed as mean ± SD and differ-ences are considered significant at p ≤ 0.05.

RESULTS

Of the 40 CAPD patients initially enrolled, 2 subjectsin the soy group were withdrawn because of noncompli-ance and 2 patients in the control group due to lack ofcooperation.

The baseline characteristics of patients did not differsignificantly between the two groups (Table 1). In addi-tion, body mass index did not significantly change withineach group during the study.

There were no significant differences in mean dietaryintake of total energy, protein, carbohydrate, total fat,saturated fatty acids, monounsaturated fatty acids, poly-unsaturated fatty acids, or cholesterol between the twogroups at baseline or at the end of weeks 4 and 8. Inaddition, these factors did not significantly changewithin each group during the study (Table 2).

There were no significant differences between the twogroups for dietary intake of vegetable or animal proteinat baseline. Mean dietary intake of vegetable protein wassignificantly higher in the soy group at the end of weeks 4and 8 compared to the control group (p < 0.01), whereasmean dietary intake of animal protein in the soy groupwas significantly lower (p < 0.05; Table 2). There wereno significant differences between the two groups in

TABLE 1The Baseline Characteristics of CAPD Patients

in the Soy and Control Groups

Soy group Control groupCharacteristic (n=18) (n=18)

Age (years)a 51.5±17 53.0±13BMI (kg/m2)a 25±5 27±5Sex

Men 9 (50%) 9 (50%)Women 9 (50%) 9(50%)

Smokers 2 (11%) 2 (11%)Diabetics 4 (22%) 7 (39%)Type of dialysis solution

1.5% glucose 10 (56%) 5 (28%)2.5% glucose 4 (22%) 3 (17%)4.25% glucose 0 (0%) 1 (6%)1.5%+2.5% glucose 4 (22%) 9 (50%)

Intake of drugGemfibrozil 3 (17%) 3 (17%)Statins 5 (28%) 7 (39%)

BMI = body mass index.a Age and BMI are presented as mean±SD.

TABLE 2The Dietary Factors Affecting Serum Lipids and

Apoproteins in the Soy and Control Groupsa

Constituent Baseline Week 4 Week 8

Energy (kcal/day)Soy 1036±336 1131±316 1071±291Control 1065±381 1162±335 1060±449

Total protein (g/day)Soy 43±9 45±10 43±9Control 41±15 44±12 40±15

Vegetable protein (g/day)Soy 17±4 29±8b,c 29±8b,c

Control 17±7 20±7 18±8Animal protein (g/day)

Soy 26±7 15±5c,d 14±4c,d

Control 24±16 23±12 22±15Carbohydrate (g/day)

Soy 140±49 148±49 149±49Control 140±55 154±58 141±70

Fiber (g/day)Soy 12±5 15±4.5 15.5±4b,c

Control 10±4 12±5 10±4Total fat (g/day)

Soy 36±18 43±14 36±12Control 40±20 44±15 40±20

SAFA (g/day)Soy 11±7 13±5 11±3Control 12±6 13±5 12±4

MUFA (g/day)Soy 12±7 14±5 12±5Control 13±6 14±5 14±8

PUFA (g/day)Soy 10±5 13±6 11±5Control 12±8 12±6 11±7

Cholesterol (mg/day)Soy 182±148 173±131 139±87Control 150±109 154±109 116±97

SAFA = saturated fatty acids; MUFA = monounsaturated fattyacids; PUFA = polyunsaturated fatty acids.a n = 18 for all values.b p < 0.01 versus the control group.c p < 0.01 versus baseline.d p < 0.05 versus the control group.All values are presented as mean±SD.

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

614

TABIBI et al. NOVEMBER 2010 – VOL. 30, NO. 6 PDI

terms of dietary intake of fiber at baseline or at the endof week 4, but mean dietary intake of fiber was signifi-cantly higher in the soy group at the end of week 8 com-pared to the control group (p < 0.01; Table 2).

At baseline, 86% of patients had serum Lp(a) higherthan 30 mg/dL and there was no significant differencebetween the two groups in serum Lp(a) concentration.At the end of week 8, serum Lp(a) concentration de-creased significantly in the soy group compared to base-line (p < 0.01), whereas no significant change wasobserved in the control group. The reduction in serumLp(a) concentration in the soy group was significantcompared with the control group (p < 0.05; Table 3). Pat-terns of change in individual serum Lp(a) concentrationsduring the study in the soy and the control groups areshown in Figures 1 and 2.

No significant changes were observed in serum con-centrations of triglyceride, total cholesterol, HDL-C,LDL-C, or apo B100 within each group during the study(Table 3). Serum apo AI concentration was decreased sig-nificantly in the soy (p < 0.05) and control (p < 0.01)groups at the end of week 8 compared to baseline butserum apo AI reduction was not significantly differentbetween the two groups (Table 3).

Serum concentration of TNF-α did not significantlychange within each group during the study (Table 3).Serum albumin was increased significantly in the twogroups at the end of week 8 compared to baseline (p <0.01); however, the increase in serum albumin was notsignificantly different between the two groups (Table 3).

DISCUSSION

High serum Lp(a) concentration is a common lipiddisorder in PD patients (3,5,14,15) and serum Lp(a) con-centration higher than 30 mg/dL constitutes a major riskfactor for CVD (9–14).

At baseline, approximately 86% of all the patients inboth groups had serum Lp(a) concentration higher than30 mg/dL. During the 8 weeks, mean serum Lp(a) de-creased significantly, by 41%, in the soy group: from78 mg/dL at baseline to 46 mg/dL at the end of the study.The reduction was significant compared with the con-trol group. In addition, at baseline, approximately 94.5%of the patients in the soy group had serum Lp(a) higherthan 30 mg/dL, but at the end of the study, 72.5% of thepatients in the soy group had serum Lp(a) higher than30 mg/dL. In 83% of the patients in the soy group, serumLp(a) concentration was decreased at the end of thestudy compared with baseline.

So far, limited studies have been conducted to assessthe effect of soy in lowering serum concentration of

Lp(a); the majority of these studies showed soy proteindid not reduce serum Lp(a) significantly (18–20,22,27).These studies were conducted mainly in non renal fail-ure individuals with normal serum level of Lp(a); so wemay not expect soy protein to reduce serum level of Lp(a)significantly. However, Meinertz et al. showed that al-cohol-extracted soy protein in normolipidemic non renalfailure individuals markedly lowered serum Lp(a), espe-cially in individuals with serum Lp(a) level higher than30 mg/dL (28). This result is in agreement with our study.

TABLE 3Serum Concentrations of Lipids, Apoproteins, Tumor

Necrosis Factor-Alpha (TNF-α), and Albumin inthe Soy and Control Groupsa

Serum parameter Baseline Week 8 Changed

Lp(a) (mg/dL)Soy 78±48 46±26b –32±43e

Control 56±26.5 52±28 –4±28.5Triglyceride (mg/dL)

Soy 176±81 197.5±131 21±131Control 203±157 189±130 –14±86

Total cholesterol (mg/dL)Soy 188±38 190±57 2±48Control 187±59 181±54 –6±34

LDL-C (mg/dL)Soy 89±19 89±30 –0.5±23Control 87±32 86±33 –1±15

HDL-C (mg/dL)Soy 42±9 43±9.5 1±8Control 37±8 42±15 5±12

Apo B100 (mg/dL)Soy 105±29 101±23 –4±18Control 114±36 98±31 –16±32

Apo AI (mg/dL)Soy 165±26 150±19c –15±23Control 163.5±15 148.5±22b –15±22

TNF-α (ng/L)Soy 3.5±2.5 5.5±4 1.3±4.5Control 5.5±4 7±7 1.5±8

Albumin (g/dL)Soy 4±0.5 4.5±0.4b 0.5±0.6Control 4±0.5 4.4±0.4b 0.4±0.5

Lp(a) = lipoprotein(a); LDL-C = low density lipoprotein cho-lesterol; HDL-C = high density lipoprotein cholesterol; Apo =apoprotein.a n = 18 for both groups.b p < 0.01 versus baseline.c p < 0.05 versus baseline.d Changes reflect week 8 values minus baseline values.e p < 0.05 versus the control group.All values are mean±SD.

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

615

PDI NOVEMBER 2010 – VOL. 30, NO. 6 EFFECTS OF SOY ON LIPID PROFILE IN PD PATIENTS

Our study is the first investigation on the effect ofsoy consumption on serum Lp(a) in PD patients, approxi-mately 86% of whom had serum Lp(a) level higher than30 mg/dL. High serum levels of Lp(a) in PD patients maybe due to increased hepatic synthesis of Lp(a) followingloss of amino acids and proteins through PD. Some stud-ies have indicated that increased urinary excretion ofprotein results in increased hepatic synthesis of Lp(a)(5,15,29–31). In addition, there is an inverse relation-ship between Lp(a) level and serum albumin (32,33). Onthe other hand, increased hepatic synthesis of Lp(a),which is an acute-phase protein, may be due to increasedproduction of cytokines in dialysis patients (29,34).

In the present study, serum concentration of albuminincreased significantly in both the soy and the controlgroup but no significant difference was observed be-tween the two groups in the level of serum albumin atbaseline or at the end of the study (week 8). Therefore,the reduced concentration of Lp(a) in the soy group can-not be attributed to a rise in serum albumin level be-cause this rise occurred in the control group as well.

Cytokines such as TNF-α stimulate hepatic synthesisof Lp(a) in dialysis patients (29). In our study, soy con-sumption for 8 weeks had no significant effect on serumconcentration of TNF-α. Therefore, the reduced concen-tration of Lp(a) in the soy group cannot be attributed toa change in serum TNF-α concentration. In contrast toour study, Huang et al. showed that consuming soy milkfor 10 weeks significantly reduced TNF-α level in meno-pausal women (35). This disparity may involve the factthat Huang et al.’s research lacked a control group andwas conducted in nondialysis subjects.

The textured soy flour was composed substantially ofprotein, fiber, isoflavonoids, including phytoestrogens,and some carbohydrates. Therefore, the effect of thetextured soy flour on serum Lp(a) is related to these com-ponents. In the present study, the effect of soy consump-tion on serum Lp(a) concentration may be because ofthe direct effect of soy phytoestrogens on Lp(a) synthe-sis; some studies have indicated that estrogen may re-duce serum Lp(a) by 30% (22,34) through affecting thesynthesis of Lp(a) (22). Thus, soy phytoestrogens mayplay a role similar to that of estrogen in reducing thesynthesis of Lp(a). In addition, some studies have shownthat soy protein (without isoflavonoids) reduces serumLp(a) concentration (28) but it has not been establishedso far that soy fiber can reduce serum Lp(a).

Hypertriglyceridemia is common in PD patients (3,29).At baseline, mean serum triglyceride concentrationswere 176 and 203 mg/dL in the soy and control groupsrespectively. Most patients had triglyceride levels in thenormal range. The normality of serum triglyceride inthese patients may be partly due to patients with hightriglyceride levels taking gemfibrozil. In this study, soyconsumption for 8 weeks did not alter serum triglycer-ide levels significantly. This finding is in agreement withthose of some previous studies (16,18,22,23,36–38) butdiffers from findings of several other studies (17,19–21,24,25). To observe the effects of soy on serum triglycer-ide concentration, it is necessary to conduct such stud-ies on hypertriglyceridemic patients. In the presentstudy, mean serum triglyceride was within the normalrange for the soy group; therefore, soy consumption wasnot expected to have any effect on serum triglyceridelevels in these normotriglyceridemic patients.

Figure 1 — Patterns of change in individual serum lipopro-tein(a) [Lp(a)] concentrations in the soy group during the8-week study.

Figure 2 — Patterns of change in individual serum lipopro-tein(a) [Lp(a)] concentrations in the control group during the8-week study.

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

616

TABIBI et al. NOVEMBER 2010 – VOL. 30, NO. 6 PDI

Serum total cholesterol, LDL-C, and apo B100 havebeen reported to be higher than normal in PD patients(29,39). At baseline, mean serum total cholesterol,LDL-C, and apo B100 were within the normal ranges forboth the soy and the control groups. The normality ofserum total cholesterol, LDL-C, and apo B100 in thesepatients may be due partly to patients with high con-centrations of total cholesterol or LDL-C taking statins,including lovastatin.

In our study, soy consumption for 8 weeks had no sig-nificant effect on serum total cholesterol, LDL-C, or apoB100. In agreement with these findings, some previousstudies have shown soy consumption causes no changesin serum total cholesterol (16,36–38), LDL-C (21,26,36–38), or apo B100 (16,18). In contrast, some studieshave indicated that soy reduces the level of serum totalcholesterol (17–25), LDL-C (16–19,22–25,38,40–42),and apo B100 (17,19,20,23,38,40).

These controversial findings may be due mainly todifferences in baseline serum levels of total cholesterol,LDL-C, and apo B100. A meta-analysis on previous stud-ies showed that soy reduces the levels of total choles-terol and LDL-C in hypercholesterolemic patients but notin normocholesterolemic individuals (25). As mentionedearlier, the majority of patients in our study had serumtotal cholesterol, LDL-C, and apo B100 levels within thenormal ranges; therefore, we cannot expect soy con-sumption to reduce serum total cholesterol, LDL-C, orapo B100 in these patients.

In PD patients, the level of HDL-C is lower than thenormal range (3,29,39). In our study, soy consumptioncaused no significant change in serum HDL-C concentra-tion. This finding is similar to those of most previousstudies (16–18,20,21,23,24,36,37,40), whereas a fewstudies have indicated that soy increases serum HDL-C(22,41).

At baseline, mean serum apo AI was within the nor-mal range (>110 mg/dL) for both groups. The serum con-centration of apo AI was reduced significantly in the soyand the control groups. This result shows the observeddecrease in the soy group is not due to soy consumptionsince it was also seen in the control group, which did notreceive soy. The only probable explanation for this re-duction is that serum lipids and apoproteins undergosome changes in normal range over time. This finding isin agreement with those of previous studies (16,17,20,40), confirming the ineffectiveness of soy consump-tion on serum apo AI concentration.

In the present study, 2 patients in the soy group werewithdrawn because of noncompliance. We did not con-tinue to follow these patients and so we cannot performan intent-to-treat analysis; this is a weakness of our study.

In conclusion, this study indicates that soy consump-tion reduces serum Lp(a), which is a risk factor for CVDin peritoneal dialysis patients. Future studies should bedone to clarify the mechanisms whereby soy consump-tion can reduce elevated serum Lp(a) concentration inperitoneal dialysis.

DISCLOSURES

The authors affirm that no financial conflict of inter-est exists.

ACKNOWLEDGMENTS

This study was supported by the National Nutrition and FoodTechnology Research Institute of Iran.

The authors thank the staff of the peritoneal dialysis unitsin Modares Hospital and Shafa Clinic for their invaluable as-sistance, and the staff of the research laboratory of theResearch Institute for Endocrine Sciences and the researchlaboratory of National Nutrition and Food Technology ResearchInstitute for their technical assistance. The authors also grate-fully acknowledge the cooperation of the participating pa-tients, without whom this investigation would not have beenpossible.

REFERENCES

1. Singh AK, Brenner BM. Dialysis in the treatment of renalfailure. In: Kasper DL, Braunwald E, Fauci AS, Hauser SL,Longo DL, Jameson JL, eds. Harrison’s Principles of In-ternal Medicine. 16th ed. New York: McGraw-Hill; 2005:1663–7.

2. Jungers P, Nquyen Khoa T, Massy ZA, Zingraff J, LabrunieM, Descamps-Latscha B, et al. Incidence of atheroscleroticarterial occlusive accidents in predialysis and dialysis pa-tients: a multicentric study in the Ile de France district.Nephrol Dial Transplant 1999; 14:898–902.

3. Wanner C. Altered lipid metabolism and serum lipids inrenal disease and renal failure. In: Kopple JD, Massary SG,eds. Kopple and Massary’s Nutritional Management ofRenal Disease. 2nd ed. Philadelphia: Lippincott Williams& Wilkins; 2004: 42–55.

4. O’Neal D, Lee P, Murphy B, Best J. Low-density lipopro-tein particle size distribution in end-stage renal diseasetreated with hemodialysis or peritoneal dialysis. Am J Kid-ney Dis 1996; 27:84–91.

5. Shoji T, Nishizawa Y, Nishitani H, Yamakawa M, Morii H.High serum lipoprotein(a) concentrations in uremic pa-tients treated with continuous ambulatory peritoneal di-alysis. Clin Nephrol 1992; 38:271–6.

6. Fytili CI, Progia EG, Panagoutsos SA, Thodis ED, PassadakisPS, Sombolos KI, et al. Lipoprotein abnormalities in he-modialysis and continuous ambulatory peritoneal dialy-sis patients. Ren Fail 2002; 24:623–30.

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

617

PDI NOVEMBER 2010 – VOL. 30, NO. 6 EFFECTS OF SOY ON LIPID PROFILE IN PD PATIENTS

7. Kimak E, Berger B, Solski J, Janicka L, Ksiazek A. Com-parison of lipid and lipoprotein profiles in long-termchronic ambulatory peritoneal dialysis (CAPD) in elderlypatients with chronic renal failure (CRF). Int Urol Nephrol2002; 33:203–4.

8. Ong-Ajyooth L, Sirisalee K, Shayakul C, Vareesangthip K,Vasuvattakul S, Vanichakarn S, et al. Comparison of lipidabnormalities in continuous ambulatory peritoneal dialy-sis and hemodialysis patients. Transplant Proc 1994; 26:2077–9.

9. Siamopoulos KC, Elisaf MS, Bairaktari HT, Pappas MB,Sferopoulos GD, Nikolakakis NG. Lipid parameters includ-ing lipoprotein (a) in patients undergoing CAPD and he-modialysis [Abstract]. Perit Dial Int 1995; 15:342–7.

10. Garcia–Lopez E, Carrero JJ, Suliman ME, Lindholm B,Stenvinkel P. Risk factors for cardiovascular disease in pa-tients undergoing peritoneal dialysis. Perit Dial Int 2007;27(Suppl 2):S205–9.

11. Superko HR. Did grandma give you heart disease? The newbattle against coronary artery disease. Am J Cardiol 1998;82:34Q–46Q.

12. Spinler SA, Cziraky MJ. Lipoprotein(a): physiologic func-tion, association with atherosclerosis, and effects of lipid-lowering drug therapy. Ann Pharmacother 1994; 28:343–51.

13. Djurovic S, Berg K. Epidemiology of Lp(a) lipoprotein: itsrole in atherosclerotic/thrombotic disease. Clin Genet1997; 52:281–92.

14. Dahlen GH, Stenlund H. Lp(a) lipoprotein is a major riskfactor for cardiovascular disease: pathogenic mechanismsand clinical significance. Clin Genet 1997; 52:272–80.

15. Heimburger O, Stenvinkel P, Berglund L, Tranaeus A,Lindholm B. Increased plasma lipoprotein(a) in continu-ous ambulatory peritoneal dialysis is related to perito-neal transport of proteins and glucose. Nephron 1996; 72:135–44.

16. Wong WW, Smith EO, Stuff JE, Hachey DL, Heird WC,Pownell HJ. Cholesterol-lowering effect of soy protein innormocholesterolemic and hypercholesterolemic men.Am J Clin Nutr 1998; 68(6 Suppl):1385S–9S.

17. Lichtenstein AH, Jalbert SM, Adlercreutz H, Goldin BR,Rasmussen H, Schaefer EJ, et al. Lipoprotein response todiets high in soy or animal protein with and without iso-flavones in moderately hypercholesterolemic subjects.Arterioscler Thromb Vasc Biol 2002; 22:1852–8.

18. Tonstad S, Smerud K, Hoie L. A comparison of the effectsof 2 doses of soy protein or casein on serum lipids, serumlipoproteins, and plasma total homocysteine in hyper-cholesterolemic subjects. Am J Clin Nutr 2002; 76:78–84.

19. Jenkins DJ, Kendall CWC, Jackson CJ, Connelly PW, ParkerT, Faulkner D, et al. Effects of high- and low-isoflavonesoyfoods on blood lipids, oxidized LDL, homocysteine, andblood pressure in hyperlipidemic men and women. Am JClin Nutr 2002; 76:365–72.

20. Teixeira SR, Potter SM, Weigel R, Hannum S, Erdman JWJr, Hasler CM. Effects of feeding 4 levels of soy protein for

3 and 6 wk on blood lipids and apolipoproteins in moder-ately hypercholesterolemic men. Am J Clin Nutr 2000; 71:1077–84.

21. Desroches S, Mauger JF, Ausman LM, Lichtenstein AH,Larmarche B. Soy protein favorably affects LDL size inde-pendently of isoflavones in hypercholesterolemic men andwomen. J Nutr 2004; 134:574–9.

22. Nilausen K, Meinertz H. Lipoprotein(a) and dietary pro-teins: casein lowers lipoprotein(a) concentration as com-pared with soy protein. Am J Clin Nutr 1999; 69:419–25.

23. Hoie LH, Morgenstern ECA, Gruenwald J, Graubaum HJ,Busch R, Luder W, et al. A double-blind placebo-controlledclinical trial compares the cholesterol-lowering effects oftwo different soy protein preparations in hypercholester-olemic subjects. Eur J Nutr 2005; 44:65–71.

24. Wang Y, Jones PJH, Ausman LM, Lichtenstein AH. Soy pro-tein reduces triglyceride levels and triglyceride fatty acidfractional synthesis rate in hypercholesterolemic subjects.Atherosclerosis 2004; 173:269–75.

25. Anderson JW, Johnstone BM, Cook-Newell ME. Meta-analysis of the effects of soy protein intake on serum lip-ids. N Engl J Med 1995; 333:276–82.

26. Weggemans RM, Trautwein EA. Relation between soy-associated isoflavones and LDL and HDL cholesterol con-centrations in humans: a meta-analysis. Eur J Clin Nutr2003; 57:940–6.

27. Dent SB, Peterson CT, Brace LD, Swain JH, Reddy MB,Hanson KB, et al. Soy protein intake by perimenopausalwomen does not affect circulating lipids and lipoproteinsor coagulation and fibrinolytic factors. J Nutr 2001; 131:2280–7.

28. Meinertz H, Nilausen K, Hilden J. Alcohol-extracted, butnot intact, dietary soy protein lowers lipoprotein(a) mark-edly. Arterioscler Thromb Vasc Biol 2002; 22:312–16.

29. Lacour B, Massy Z, Drueke TB. Lipid metabolism. In: MassrySG, Glassock RJ, eds. Massry & Glassock’s Textbook ofNephrology. 4th ed. Philadelphia: Lippincott Williams &Wilkins; 2001: 1346–56.

30. Ross EA, Shah GM, Kashyap ML. Elevated plasma lipopro-tein (a) levels and hypoalbuminemia in peritoneal dialy-sis patients. Int J Artif Organs 1995; 18:751–6.

31. Anwar N, Bhatnagar D, Short CD, Mackness MI, DurringtonPN, Prais H, et al. Serum lipoprotein (a) concentrations inpatients undergoing continuous ambulatory peritonealdialysis. Nephrol Dial Transplant 1993; 8:71–4.

32. Irish AB, Simons LA, Savdie E, Hayes JM, Simons J. Lipo-protein(a) levels in chronic renal disease states, dialysisand transplantation [Abstract]. Aust N Z J Med 1992; 22:243–8.

33. Yang WS, Kim SB, Min WK, Park S, Lee MS, Park JS. Athero-genic lipid profile and lipoprotein(a) in relation to serumalbumin in haemodialysis patients [Abstract]. Nephrol DialTransplant 1995; 10:1668–71.

34. Scanu AM. Atherothrombogenicity of lipoprotein(a): thedebate. Am J Cardiol 1998; 82:26Q–33Q.

35. Huang Y, Cao S, Nagamani M, Anderson KE, Grady JJ, Lu

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from

618

TABIBI et al. NOVEMBER 2010 – VOL. 30, NO. 6 PDI

LJ. Decreased circulating levels of tumor necrosis factor-alpha in postmenopausal women during consumption ofsoy-containing isoflavones. J Clin Endocrinol Metab 2005;90:3956–62.

36. Olyaei AJ, Demattos AM, Bennett WM. Prescribing drugsin renal disease. In: Brenner BM, ed. Brenner & Rector’sthe Kidney. 6th ed. Philadelphia: W.B. Saunders; 2000:2632–4.

37. Engelman HM, Alekel DL, Hanson LN, Kanthasamy AG,Reddy MB. Blood lipid and oxidative stress responses tosoy protein with isoflavones and phytic acid in postmeno-pausal women. Am J Clin Nutr 2005; 81:590–6.

38. Chen ST, Ferng SH, Yang CS, Peng SJ, Lee HR, Chen JR.Variable effects of soy protein on plasma lipids in hyper-lipidemic and normolipidemic hemodialysis patients. AmJ Kidney Dis 2005; 46:1099–106.

39. Prichard SS. Metabolic complications of peritoneal dialy-sis. In: Daugirdas JT, Blake PG, Ing TS, eds. Handbook ofDialysis. 3rd ed. Philadelphia: Lippincott Williams &Wilkins; 2001: 405–10.

40. Chen ST, Chen JR, Yang CS, Peng SJ, Ferng SH. Effect ofsoya protein on serum lipid profile and lipoprotein con-centrations in patients undergoing hypercholesterolaemichaemodialysis. Br J Nutr 2006; 95:366–71.

41. Taku K, Umegaki K, Sato Y, Taki Y, Endoh K, Watanabe S.Soy isoflavones lower serum total and LDL cholesterol inhumans: a meta-analysis of 11 randomized controlled tri-als. Am J Clin Nutr 2007; 85:1148–56.

42. Allen JK, Becker DM, Kwiterovich PO, Lindenstruth KA,Curtis C. Effect of soy protein-containing isoflavones onlipoproteins in postmenopausal women. Menopause 2007;14:106–14.

This single copy is for your personal, non-commercial use only. For permission to reprint multiple copies or to order presentation-ready copies

for distribution, contact Multimed Inc. at marketing@multi-med.com

at HIN

AR

I BA

ND

1 on Decem

ber 10, 2010 w

ww

.pdiconnect.comD

ownloaded from