Lipid-altering effects of a dietary supplement tablet containing free plant sterols and stanols in men and women with primary hypercholesterolaemia: a randomized, placebo-controlled crossover trial

This randomized, placebo-controlled, crossover trial assessed the lipid-altering efficacy of a dietary supplement (tablet form) providing 1.8g/day free (non-esterified) plant sterols and stanols versus placebo for 6 weeks as part of a therapeutic lifestyle changes (TLC) diet in 32 men and women with primary hypercholesterolaemia. Mean ± SE baseline (end of a 5-week TLC diet lead-in) lipid concentrations (mmol/1) were total cholesterol (TC), 5.88 ± 0.08; non-high-density lipoprotein cholesterol (non-HDL-C), 4.71 ± 0.09; low-density lipoprotein cholesterol (LDL-C), 4.02 ± 0.08; HDL-C, 1.17 ± 0.06 and triglycerides (TGs), 1.51 ± 0.12. Differences from control in responses (plant sterol/stanol — control) were significant (p < 0.05) for LDL-C (− 4.9%), non-HDL-C (− 3.6%) and TC (− 2.8%). HDL-C and TG responses were not significantly different between treatment conditions. These results indicate that 1.8g/day free plant sterols/stanols administered in a tablet produced favourable lipoprotein lipid changes in men and women with hypercholesterolaemia.


Introduction
The guidelines of the National CholesterolEducation Program (NCEP) recommend incorporation of 2g/day of plant sterols or stanols( phytosterols or phytostanols) intot he therapeutic lifestyle changes (TLC)diet as adietaryadjunct for individuals who do nota chieve their low-density lipoprotein cholesterol (LDL-C) treatment targets with diet alone (NCEP ExpertPanel 2002). Alarge body of evidence supports the efficacyo fp lant sterol/stanol-enriched products providing 1.0-3.0 g/dayp lant sterols/stanols for lowering LDL-C,n on-high-density lipoprotein (HDL)-C and total cholesterol (TC) concentrations (Food and Drug Administration 2000;NCEP Expert Panel 2002;Maki et al. 2003;Demonty et al. 2009; Sá nchez-Munize ta l. 2009).R esults from studies in subjects with hypercholesterolaemia have generally indicated that HDL-Ca nd triglyceride (TG)concentrations are unchanged by consumption of plant sterols or stanols, althoughs ome evidence suggests that plant stanols(and presumablysterols)may lower the TG concentration in subjects with hypertriglyceridemia (Plat et al. 2009).
Because of their structurals imilarity to cholesterol, plant sterols and stanolsc ompete with cholesterol for incorporation into micelles, as well as for transport across the brush border by the NiemanPick C1 -Like 1t ransporter (von Bergmann et al. 2005; Huie ta l. 2008; Jones2 008).I na ddition, the accumulation of plant sterols or stanolsi nt he enterocyte appearst o trigger thep roductiono fa denosine triphosphate binding cassette transportersG 5a nd G8, which function to transport sterols, including cholesterol, out of the enterocyte, into the intestinal lumen (von Bergmann et al. 2005;Hui et al. 2008;Jones 2008). The net result of these mechanisms is to reduce intestinal cholesterol absorption (Ikeda and Sugano 1998;Jones 2008).C onsequently,hepatic cholesterol content is reduced, inducing an up-regulation of LDL receptor expression,therebyenhancing the removal of LDLa nd othera polipoprotein B-containing lipoprotein particles from the circulation (Jones 2008).
Various commercially available products contain plant sterols and/or stanols in free and/or esterified forms including margarine-type spreads, yogurta nd yogurt-based drinks, orange juice and dietary supplements in capsule or tablet form. Most clinical trials examining the effects of plant sterols/stanols on lipid concentrations have administered food forms (Ostlund et al. 1999;Maki et al. 2001Maki et al. , 2003Nestel et al. 2001;Volpe et al. 2001;de Graaf et al. 2002;Gremaud et al. 2002;Pouteau et al.2003;Berger et al. 2004;Thomsene ta l. 2004;Doornbos et al. 2006;Abumweis et al. 2008), while fewer havei nvestigated the effects of tablet and capsule forms (McPherson et al. 2005;Goldberg et al. 2006;Acuff et al. 2007;Carre ta l. 2009). The use of plant sterol/stanol capsules or tablets offersapractical option compared with traditional food applications because it providesa vehiclet hatc an be easily incorporated into a cholesterol-lowering regimenw ithout impacting dietarymacronutrient distribution.
Recently,t he US Food and Drug Administration reviewedt he published data on phytosterols and risk for coronaryh eartd isease( CHD;F ood andD rug Administration 2010).Intheir review theyconcluded that 'the available scientific evidence for the cholesterol-lowering effectso fp hytosterolsi nd ietary supplements shows thatf ormulation of thes upplementp roducti sa ni mportant factor in the effectiveness of the product'.T hey further concluded that 'the totality of available scientific evidence for the cholesterol-lowering effects of nonesterified phytosterols in dietarysupplements is inconsistent'. The agency requested submission of additionald ata to demonstrate thec holesterol-lowering efficacyo fn onesterified phytosterols consumeda si ngredients in dietary supplements.T his trial was undertakent o assess the efficacyo fadietarys upplement in at ablet providing at otalo f1 .8 g/day plant sterol/stanol in a non-esterified form, as partofaNCEPTLC diet, for improving the lipoprotein lipid profile in men and women with primaryh ypercholesterolaemia. Similar doses of non-esterified plant sterol/stanol administered in food products haveb een shown to significantly lower LDL-C levels (Jones et al. 1999;de Graaf et al. 2002).D aily plant sterol intakei nt he US diet is , 200 mg (range 150 -450 mg in various We stern populations; Ostlund 2002),w hich is insufficient to materially alter cholesterol levels.A ccordingly,t he dosage of 1.8 gw as chosen with the expectation that total plant sterol/ stanol intake in study subjects would be , 2.0 g/day from ac ombination of diet and the supplementp rovided. Ad oseo f1 .7 g/dayn onesterified sitostanol-containing phytosterols administered in margarine for 30 daysr esulted in a1 5.5% LDL-C reduction (Jones et al. 1999), and 1.8 g/d administered in chocolate for 4weeks decreased LDL-Cby10.3% (de Graaf et al. 2002).

Studyd esign
This was arandomized crossover study consisting of a 5-week diet plus single-blind placebo lead-in, followed by two double-blind 6-week treatment periods during which subjects received either dietarys upplement tablets (four tablets daily,two tablets with each of two meals)p roviding 1.8 g/day non-esterifiedp lant sterols/stanols or the same numbero fm atching placebo tablets.T he study was conducted at two clinical researchc entres (Provident ClinicalR esearch in Addison, IL,U SA, and Bloomington, IN, USA) according to GoodC linical Practice Guidelines, the Declaration of Helsinki and the US 21 Code of Federal Regulations. Informed consentf or the study was obtained from all subjects before protocol-specific procedures were carried outa nd subjects were informed of their rightst ow ithdraw from the study at any time.

Subjects
Mena nd women of age 21 -79y ears, inclusive, each with af asting LDL-C level $ 3.4 and , 5.7 mmol/l, and in good general health on the basis of medical history and routine laboratoryt ests were eligible for the study.I ndividuals were excluded from participation if they had abody mass index of . 42.0 kg/m 2 , fasting bloodg lucose $ 7.0 mmol/l or diabetes mellitus, restingbloodpressure $ 160 mm Hg systolic and/or $ 100 mm Hg diastolic, or CHD or aCHD risk equivalent as defined by the NCEPAdult Treatment PanelI II (NCEP ExpertP anel 2002).A dditional exclusion criteria included ahistoryofextreme dietary habits, eating disorders, alcoholism, cancer or any clinically important cardiovascular disorders.U se of any medications, dietarys upplements or fortified foodsw ith lipid-altering effects, includings terola nd stanol products, were excludedf or at least 4w eeks prior to study entryaswas the use of weight-loss drugs or programmesa nd recent body weight change . 4.5 kg.

Studyp roducts anddiet instruction
Following the diet and the single-blind placebo leadin, subjects were randomly assigned to receive, in a double-blind manner, eithera ctive (non-esterified sterol/stanol tablets, 0.45 gp er tablet) or control Freeplants terol/stanol lipid effects (matched placebo tablets, the main ingredient of whichw as lactose) study productf or 6w eeks (treatment period I) andt hen cross over to receive the opposite productfor 6weeks (treatment period II). Subjects were instructed to swallow, with water or another beverage, four active or controlt ablets daily (two with eacho ft wo meals) at consistent times each day.C ompliance with active and controlt ablets was assessed by counting unused study product returned to the clinic. Subjectswere instructed to maintain their usual physical activity patterns throughoutthe trial.
Beginning at week 2 5a nd throughout the study, subjectsw erec ounselledt of ollowt he weight maintenance version of the TLC diet recommended by theN CEP, andh andoutsw erep rovidedt o reinforce the diet instructions (NCEP ExpertP anel 2002).T oevaluate compliance with the TLC diet, diet records were completed on 3c onsecutive days( 2 weekdays and 1weekend day) at baselineand the end of each treatment period using standardized forms that included instructions for obtaining complete information such as the use of condiments, brand names andcooking methods.Daily intakes of energy and key nutrients were calculatedf rom these records utilizing Food ProcessorS QL Software (version 10.4, EHSA Research, Salem,O R, USA). Body weight was measured at each clinicv isit throughout the trial.

Laboratorym easurements
Fasting (9 -15h)p lasma lipid profile measurements from blood samples collected in duplicate at baseline (weeks 2 1a nd 0) anda tt he end of each treatment period (weeks 5and 6, 11 and12) were conducted by the EMH Reference Laboratory(Elmhurst, IL, USA) according to the Standardization Program of the Centersf or Disease Control andP reventiona nd the National Heart, Lung,a nd Blood Institute. Lipoprotein lipida ssessments includedT C, non-HDL-C (calculateda sT C-HDL-C), LDL-C,H DL-C and TG in mg/dl and converted to mmol/l with conversion factorsof0.02586 for cholesterol and0.01129 for TG. LDL-C concentration was calculated using Friedewald's( 1972) equation as follows: LDL-C ¼ TC 2 HDL-C 2 TG/5. Because this equation is notv alid when the TG concentration is above4 .51 mmol/l, LDL-C values were not calculated in the few instances when subjects had valuesint his range.

Statistical analyses
Statistical analyses were completed in SAS version 9.2 (SAS Institute, Cary, NC, USA).E fficacya nalyses were performed on the sample that includeda ll subjects who were randomized andp rovided at least onep ost-randomization fasting lipid profile during eacht reatment condition. The sample size for this study was selected for 90% power (5% a -level) to detecta7% difference between treatments in the changefrom baselineLDL-C concentration. Baseline characteristicsf or subjectsi nt he two treatment sequences were compared using analysis of variance (continuous variables) or Fisher'sexacttest (categorical variables). Repeated measures analysis of covariance (using SAS PROC MIXED) was used to compare lipid, dietaryi ntake and vital signs response variables (changes or percent changes from baseline) for the two treatment conditions (active and control) using the baselinev alue as ac ovariate.T he initial model included subjecta sarandom effect and terms for treatment condition, sequencea nd treatment by sequence interaction. If the treatment by sequence interaction term for av ariable was not statistically significant ( p . 0.05), it was droppedf rom the final model.
Examination of responses by sequences uggested that no material differences were present that would bring into question the appropriateness of pooling data from the two sequence groups.R esiduals from the final model were examined to assess normality,a nd if clear evidence of non-normalityw as present, rank transformations were employed in the final models. An additional analysis of covariancewas conducted using the results of Kris-Etherton and Yu 's (1997) predictive equation for LDL-C changes with response to dietary alteration with and without an adjustment for dietary cholesterol intake basedo nH egsted et al.'s (1993) equation accordingt od iet record results. This was done in order to assess possible confounding by dietaryc hanges during the treatment periods.S afety was assessed by the evaluation of treatment emergent adversee vents and changes in vital signs measurements. McNemar's test was used fors tatistical comparisons between treatment conditions for categorical variables.

Subjects anddemographics
Sixty-two subjects were screened, 32 of whom entered and completed the first treatment period (Figure 1). One subject withdrew consent after completing partof the second treatment period while receiving the active study product, butw as retainedi nt he efficacy analyses. Demographic and baseline characteristics of the subjects are presented in Ta ble I. Subjects were predominantly female (59%), of non-Hispanic White race/ethnicity (91%) and non-smokers( 91%).P articipants had am ean age of 57.6 yearsa nd am ean body mass index of 27.4 kg/m 2 .C ompliance with active and controltablets was 98.0% and 98.4% of the expected tablets, respectively ( p ¼ 0.586).

Lipids
Fasting plasma lipids at baseline, percent changes from baseline for control anda ctive treatments and K.C. Maki et al. differences in responses between conditions are shown in Ta ble II. Differencesf rom controli nr esponses (plant sterol/stanol 2 control) were significant ( p , 0.05)f or LDL-C ( 2 4.9%), non-HDL-C ( 2 3.6%)a nd TC ( 2 2.8%). HDL-C andT G responses were not significantly different between treatment conditions.

Vital signs,b odyw eight and safety
There were no statistically significant or clinically relevant changes in vital signs, body weight or clinical laboratoryvalues. Mean bodyweight changed , 0.5 kg andm eans ystolica nd diastolic bloodp ressures changed # 1mmHgt hroughout the study.A dverse events assessed by non-leading questions at eachclinic visitw ere reported by eight (12.5%) of the subjects during the active period and seven (10.9%) of the subjects during the control period. The most common adversee vents were related to the respiratorys ystem (rhinitis and sinusitis). None of the adversee vents were serious and all but one adversee vent (moderate endometrial hyperplasia experienced by one subject when receiving the placebo) were reported to be mild. Increased appetite reported by one subject during the control period was consideredb yt he study physician to be possibly related to study product. Allo ther adversee vents were classifiedb yt he study physicians as not related or unlikely to be related to the study products.

Discussion
The results of the present study indicate that 1.8 g/day unesterified plant sterols/stanols administered orally in tablets resulted in significant reductions in atherogenic lipoprotein lipids (LDL-C, non-HDL-C and TC)i n individuals with hypercholesterolaemia. The LDL-C lowering of 4.9% was within the expected range, albeit at the lower end, based on meta-analyses (Katan et al. 2003;Abumweis et al. 2008;Demonty et al. 2009). Accordingt oc alculations from am eta-analysis, weighted meanp ercentage reductions in LDL-C rangef rom , 5% at 1.0 g/day phytosterol to , 11% at 3.0 g/day phytosterol with no appreciableincrease in response at higher dosagesu pt o9 .0 g/day (Demonty et al. 2009).
Although LDL-C has long been consideredt he principal lipoprotein determinant of atherosclerosis, resultsfrom observationalstudies and clinical trials of lipid-altering therapies have consistentlys hown that the non-HDL-C leveli sabetterp redictor of CHD event risk than LDL-C alone, irrespective of whether the TG concentration is elevated (Cui et al. 2001;Bittner 2003;Pischon et al. 2005;Ridker et al. 2005  Freeplants terol/stanol lipid effects Liuetal. 2006;Milleretal. 2008). Non-HDL-C was reduced by 3.6% in the present trial.
Most prior clinical trials havebeen conducted using plant sterols and stanols provided in foods (Ostlund et al. 1999;Maki et al. 2001Maki et al. , 2003Nestel et al. 2001;Volpe et al. 2001;de Graaf et al. 2002;Gremaud et al. 2002;Pouteaue ta l. 2003;Berger et al.2 004;Thomsene ta l. 2004;D oornbose ta l. 2006;Abumweis et al.2 008). Relatively few studies have investigated plant sterols/stanols in tablet or capsule forms,b ut the available results supportt he ability of plant sterols ands tanolst or educe LDL-C concentrations in some supplement forms (McPherson et al. 2005;Goldberg et al.2 006;Acuff et al. 2007;Carr et al. 2009). Goldberg et al. (2006) demonstrated that tablets providing 1.8 g/day soy stanolsi na dditiont o the subjects' usual statin therapy,r educed LDL-C by 9.1%. McPherson et al. (2005) reported areduction in LDL-Co f1 0.4% aftera dministrationo fs tanol lecithin tablets providing 1.26 gstanols/day.T he use of plant sterols andstanolsincapsules or tablets offers apractical option compared to traditional phytosterolcontaining food vehicles by virtue of being easily incorporated into ac holesterol-lowering regimen without altering the macronutrient distribution. This feature might increase complianced uring long-term use comparedwith the types of dietarychanges needed to incorporate phytosterol-containing foods .
Because of the hypothesized mechanisms of action: (1) competing with cholesterol for incorporation into micelles and( 2) up-regulation of the production of sterolt ransportp roteins from the enterocyte into the intestinal lumen, it is recommendedt hat sterol or stanol products be consumed with meals to ensure adequate availability of bile acids for micelle formation .I nt he presents tudy,a ll subjects were advised to take the tablets twice daily with meals. It was believed at one time that plant sterols would be more effective when consumedw ith diets containing higherlevelsofcholesterol or fat (Mussner et al. 2002;Berger et al. 2004).Sinceonly10-20% of cholesterol passing through the intestine daily is of dietaryorigin, sterols and stanolsappear to be equally effective when consumed with low-fat diets such as the weight maintenance version of the TLC dietu sed in the current study (Hallikainen and Uusitupa 1999;Jones et al. 1999;Hallikainen et al. 2000;Katan et al. 2003;Abumweiseta l. 2008;Demonty et al.2 009).
Additional research is needed to further evaluate the effects of this and otherplant sterol/stanol products on lipid levels in othert ypes of dyslipidemia, particularly Fredrickson Type IIb (mixed) dyslipidemia. Also of interest wouldb ea ne valuation of the effect on cholesterol biosynthesis of thei ntakeo fp lant sterol/stanol tablets during the day (only with lunch) versusint he evening (only with dinner).
Interventiont rialsh aves hown that each 1% reduction in LDL-C (or non-HDL-C)l owerst he risk of am ajor cardiovascular event by , 1% over a period of 5years (Grundy et al. 2004;Robinson et al. 2005Robinson et al. , 2009.H owever, the cardiovascular benefit of maintaining low levels of atherogenic lipoprotein cholesterol levelso ver decades mayb el arger than would be predicted on the basis of resultsfrom shorttermcholesterol-lowering intervention trials. Each 1% reductioni nL DL-C or non-HDL-Cm ay be associated with as mucha sa3% reduction in CHD event risk if maintained over an extended period (Cohenetal. 2006).Thus, the changes in atherogenic lipoprotein cholesterol observed in the presents tudy are clinicallyr elevant.

Conclusions
In conclusion,i ncorporation of ad ietarys upplement tabletc ontaining 1.8 g/day unesterified plant sterols/ stanolsinto the NCEPTLC diet produced favourable changes in apolipoprotein B-containing lipoprotein lipids in individuals with hypercholesterolaemia and would be expected to reduce the risk for cardiovascular diseasei fc onsumed over an extendedp eriod of time (Miettinen andGylling 2004 Notes: HDL-C, high-density lipoprotein cholesterol;LDL-C, low-density lipoprotein cholesterol; SEM, standarderror of the mean; TC, total cholesterol; TG, triglycerides; *Results for both treatment sequenceswerepooled. Adjusting for differences in dietaryintake did not significantly alter results; † Baseline ¼ average of values at weeks 2 1a nd 0; control and active, % D ¼ percentage change from baseline to the average of valuesatthe last 2weeks of each treatment period (weeks 5and 6, weeks11and 12); difference in response ¼ active%change 2 control % change; ‡ To convertf rom mmol/l to mg/dl, for cholesterol multiply by 38.7 and for TG multiply by 88.6.
for assistance with statistical analysis and Rachel Hubacher andN icholas Shera, both formerly with Provident Clinical Research/Biofortis North America, for technical assistance.
Declaration of interest:T his trialw as funded by Pharmavite, LLC, Northridge, CA, USA. KCM, ALL, MSR and MRDwere, at the time this study was conducted,employees of Provident Clinical Research, whichh as receivedr esearch grants upport and consulting fees from Pharmavite,L LC,t he manufacturero ft he product studied. BHJ, ES andJ RB are employees of Pharmavite, LLC, the manufacturer of the product studied.