Normalisation of dyslipidaemia, a modifiable risk factor, can prevent or reduce the risk of CVD, which is a leading cause of death worldwide(Reference Pearson, Blair and Daniels1, Reference Kostis2).
In order to treat dyslipidaemia, different pharmacologic and non-pharmacologic approaches have been used. Statins are the first-line therapy for management of increased LDL cholesterol (LDL-C), despite adverse effects such as gastrointestinal symptoms and muscle aches(Reference Clark3). Although fibrates may decrease serum TAG and increase serum HDL cholesterol (HDL-C), more powerful drugs for normalisation of these lipids need to be investigated.
Conjugated fatty acids are well-known hypolipidaemic agents(Reference Wang, Nagao and Inoue4, Reference Wilson, Nicolosi and Saati5) documented for their effect on lipid metabolism(Reference Turpeinen, Bärlund and Freese6). PSO (Punica granatum) consists of about 80 % conjugated octadecatrienoic fatty acids, with a high content of 9-cis, 11-trans, 13-cis acid or punicic acid (PA), one of the isomers of conjugated linolenic acid (CLN)(Reference Lansky and Newman7); animal studies report controversial results for the hypolipidaemic role of PSO(Reference Arao, Wang and Inoue8, Reference Yang, Leung and Cao9). Although Arao et al. (Reference Arao, Wang and Inoue8) showed that PSO supplementation in obese hyperlipidaemic Otsuka—Long–Evans–Tokushima fatty rats reduces TAG accumulation, Yang et al. (Reference Yang, Leung and Cao9) demonstrated that PSO does not alter serum cholesterol concentration.
Considering the lack of data on the effect of PSO on lipid profiles in human subjects, the present study investigated the effect of PSO treatment on lipid profiles of hyperlipidaemic subjects.
Subjects and methods
Subjects and study design
This is a parallel, randomised, double-blind and placebo-controlled study. Subjects were recruited from an endocrine clinic in sequential fashion. Inclusion criteria were age over 20 years, not pregnant or lactating, having no diagnosed allergy or liver dysfunction, BMI ≤ 35 kg/m2, serum total cholesterol >5·2 mmol/l and serum TAG >1·65 mmol/l.
Subjects were randomly assigned to one of two groups, the PSO (n 25) and the placebo (n 26) groups; randomisation was stratified for lipid reducing drug consumption and blind for the subjects and all investigators, except for the randomiser. At baseline, fasting blood samples were taken, and the subjects were given capsules containing either 400 mg PSO or placebo for 4 weeks; capsules were identically packaged to be indistinguishable. Subjects consumed one capsule of either PSO or placebo twice daily, and they were telephoned weekly to evaluate compliance to medication and to ensure that no side effects occurred. Adherence was determined by capsule count after 4 weeks, and the individuals with < 80 % consumption were excluded.
It was recommended that the subjects keep their regular lifestyle, including diet and physical activity during the course of study. Questionnaires were used to collect demographic data, medical history at baseline, and medication and supplement use at baseline and 4 weeks later. The modifiable activity questionnaire and three 24 h dietary recalls were collected at baseline and 4 weeks later to assess physical activity and dietary intake. BMI and waist:hip ratio were calculated at baseline and at the end of 4 weeks. Body composition including percentage of lean and fat mass was measured by bioelectrical impedance analysis, using Bodystat 1500MDD at baseline and after 4 weeks.
The present study was conducted according to the guidelines of the Declaration of Helsinki, and all procedures involving human subjects were approved by the ethical committee of the Research Institute for Endocrine Sciences of the Shahid Beheshti University of Medical Sciences. Written informed consent was obtained from all the subjects.
Pomegranate seed oil capsule preparation
Pomegranate seed oil (PSO) capsules (Pometane) were obtained from Vitane Pharma Gmbh, Inc. (Wolfratshausen, Germany), and the raw material for PSO was obtained from Sarouneh Company (Urmia, Iran). Chemical analysis of PSO showed the major forms of fatty acids to be: PA (18 : 3) 72·0 %; oil acid (18 : 2) 10·7 %; linol acid (18 : 2) 8·4 %; palmitic acid (16 : 0) 4·0 %; stearic acid (18 : 0) 2·6 %; trace amounts ( < 1 %) of other fatty acids, with total amount of 7·3 % of SFA, 11·7 % of MUFA and 81 % of PUFA; crude PSO contained the antioxidant vitamin E of 36·90 mg/100 g and was kept at cold temperature to avoid oxidation until formulated in capsules.
Biochemical analysis
Fasting blood samples were taken after 12.00–14.00 hours overnight fasting at baseline and 4 weeks after intervention, and they were centrifuged to separate serum, for biochemical analysis, conducted at the endocrine research laboratory on the day of blood collection. Fasting plasma glucose was measured by the enzymatic colorimetric method using a glucose oxidation kit. Serum total cholesterol and TAG levels were assayed using enzymatic colorimetric tests with cholesterol esterase and cholesterol oxidise and glycerol phosphate oxidise, respectively; HDL-C was measured by the immunoturbidimetry method after precipitation of apo B containing lipoproteins with phosphotungstic acid (Pars Azmoun Company, Tehran, Iran). LDL-C was calculated from serum total cholesterol, TAG and HDL-C, according to Friedewald's equation(Reference Friedewald, Levy and Fredrickson10); it was not calculated when serum TG >4·52 mmol/l. Insulin and oxidised LDL were measured by ELISA kits (Mercodia AB Company, Uppsala, Sweden); insulin resistance was calculated by the homeostasis model for insulin resistance (HOMA-IR) model, using the formula:
Statistical method
To have a clinically significant effect on lowering serum cholesterol(Reference Law, Wald and Thompson11), we based our sample calculation to detect a difference of 200 mg/l and an sd of 200 mg/l in endpoint cholesterol concentrations between the PSO and placebo groups, and we found that a sample size of 21 was sufficient with 95 % CI and 90 % power in each group. Fifty-one hyperlipidaemic subjects were enrolled to compensate for eventual study dropouts.
SPSS (version 16.0; SPSS, Inc., Chicago, IL, USA) was used for all statistical analyses. Differences between the two groups at baseline were tested with Student's t test and the Mann–Whitney test. Paired Student's t test and Wilcoxon ranked test were used to compare the baseline and 4-week values in each group. To distinguish treatment effect between the groups, following adjustment of their baseline values, analysis of covariance test was used.
Results
Forty-five of fifty-one patients enrolled in our study were eligible for final analysis, as six withdrew and did not provide 4-week blood samples. Mean values for baseline body weight and age were 74·2 (sd 10·0) and 75·7 (sd 12·2) kg, and 51 (sd 9) and 55 (sd 9) years for the PSO and placebo groups, respectively. No significant differences between the groups were seen for age, sex, weight, height, consumption of lipid lowering drugs or n-3 supplements and smoking. All the subjects maintained good health throughout the study, without any major adverse events.
A comparison of dietary intake and physical activity variables at baseline and after 4 weeks between the PSO and placebo groups revealed no significant differences. Energy intake and dietary cholesterol decreased in the placebo group during the treatment period (P < 0·05).
The effects of PSO on biochemical, anthropometric, body composition and blood pressure variables are given in Table 1; PSO administration resulted in a decrease in serum TAG concentration (2·75 (sd 1·40) v. 3·45 (sd 1·56) mmol/l, P < 0·01) and a non-significant increase in HDL-C concentration (1·38 (sd 0·44) v. 1·25 (sd 0·39) mmol/l, P = 0·081) compared with baseline values; TAG:HDL-C ratio decreased significantly within the PSO group (5·73 (sd 4·55) v. 7·49 (sd 4·95), P < 0·031). Comparison of treatment effects between the groups with baseline values as covariates revealed lower values for cholesterol:HDL-C (5·45 (sd 1·51) v. 5·89 (sd 1·43), P < 0·05) and higher, but non-significant HDL-C concentrations (1·38 (sd 0·44) v. 1·25 (sd 0·26) mmol/l, P = 0·059) in the PSO group as compared with the placebo group, respectively. Other lipid profile variables including cholesterol, LDL-C, oxidised LDL and LDL-C:HDL-C ratio revealed no significant differences within or between the groups. Fat mass, BMI and waist:hip ratio decreased in the PSO group, but not significantly. PSO had no effect on systolic and diastolic blood pressure within or between the groups (data not shown). Means for differences were significant in HDL-C concentration (0·13 v. − 0·02 mmol/l) and cholesterol:HDL-C ratio ( − 0·42 v. 0·01) when the PSO group was compared with the placebo group (P < 0·05 for both) (Fig. 1).
Table 1 Anthropometric, body composition, blood pressure and biochemical variables at baseline and after 4 weeks in the pomegranate seed oil (PSO) and placebo groups†
(Mean values and standard deviations)
LDL-C, LDL cholesterol; HDL-C, HDL cholesterol; Ox-LDL, oxidised LDL; HOMA-IR, homeostasis model for insulin resistance.
Mean values were significantly different within the groups using paired t test: *P < 0·009, **P < 0·031, ***P < 0·001.
† All PSO variables were not significant at baseline, in comparison to placebo.
‡ Analysis of covariance with 4-week values as dependent variables, baseline values as covariates and treatment group as a fixed factor.
Fig. 1 Mean differences compared with baseline in the pomegranate seed oil (PSO; 
) and placebo groups (■). Ox-LDL-C, oxidised LDL; TC, total cholesterol; HDL-C, HDL cholesterol, LDL-C, LDL cholesterol; HOMA-IR, homeostasis model for insulin resistance. *P < 0·05.
Discussion
In the present study, 4 weeks of PSO treatment decreased serum TAG concentration, TAG:HDL-C ratio and cholesterol:HDL-C ratio in the PSO group, as compared with the placebo group.
Insignificant differences in physical activity levels and daily diets of the subjects throughout the study duration indicate that any changes in variables were probably due to PSO treatment. Since a target of 2·11 mmol/l or less for serum TAG in elderly women has been proposed(Reference Mazza, Tikhonoff and Schiavon12), the decrease in serum TAG from 3·45 (sd 1·56) to 2·75 (sd 1·40) mmol/l may be clinically valuable. Data indicate that each mg/dl increase in HDL-C corresponds to a 2–3 % risk reduction of CHD(Reference Gotto13); hence our HDL-C increase of 50 mg/l in the present study reflects a 10–15 % risk reduction of CHD. Besides the lipid profile, we investigated TAG:HDL-C, cholesterol:HDL-C and LDL-C:HDL-C as lipid-related indices. The TAG:HDL-C ratio predicts CVD and metabolic syndrome(Reference Hadaegh, Khalili and Ghasemi14), explaining its crucial role in treatment guidelines; therefore, 1·8° reduction of the TAG:HDL-C ratio in the present study is another significant finding. High cholesterol:HDL-C ratios ( ≥ 4) are also associated with coronary artery disease(Reference Nair, Carrigan and Curtin15), and we have found a significant difference in cholesterol:HDL-C ratio after adjusting for baseline values, in PSO group v. placebo group.
A recent study in healthy human subjects reported Trichosanthes kirilowii seeds, naturally occurring sources of PA, had no significant effect on serum lipids(Reference Yaun, Wahlqvist and Yuan16) contrary to our finding which may be attributable to the pre-existence of hyperlipidaemia in the present study. Current data on the effect of PSO on lipid concentration and metabolism are controversial(Reference Arao, Wang and Inoue8, Reference Yang, Leung and Cao9, Reference Yamasaki, Kitagawa and Koyanagi17). Similar to our findings, the liver accumulation of TAG was reduced by PSO administration(Reference Arao, Wang and Inoue8, Reference Koba, Imamura and Akashoshi18); a plausible mechanism, demonstrated in an in vitro study, that 9-cis, 11-trans, 13-cis CLN suppressed cellular synthesis of TAG in HepG2 cells(Reference Arao, Yotsumoto and Han19).
One study suggests a metabolic pathway for 9-cis, 11-trans, 13-cis CLN is its conversion to 9-cis, 11-trans conjugated linoleic acid (CLA) in a rat model; this conversion coefficient is about 12 % for PSO(Reference Tsuzuki, Kawakami and Abe20). Also two recently published articles have demonstrated that PA from T. kirilowii seed oil incorporated into different tissues and serum metabolises to CLA in rats and human subjects(Reference Yuan, Sinclair and Xu21, Reference Yuan, Yuan and Li22). Hence, the physiological activities of PSO may be attributable to 9-cis, 11-trans CLA derived endogenously or to PA itself. Scientific evidence reveals the inconsistent effects of 9-cis, 11-trans CLA on lipid profiles(Reference Tricon, Burdge and Kew23–Reference McLeod, LeBlanc and Langille26), alterations that may occur mostly through the carnitine palmitoyltransferase activity-enzyme involved in fatty acid β-oxidation, as mitochondrial and peroxisomal carnitine palmitoyltransferase activity is reported to increase with PSO(Reference Koba, Imamura and Akashoshi18) and CLN(Reference Koba, Akahoshi and Yamasaki27); therefore, there is no conclusive evidence to confirm the effect of CLN and CLA on lipid profiles, but it does seem that the isomer type and amount of CLN and CLA could be determining factors.
In the present study, no significant changes were observed in body composition (fat or lean mass), BMI or waist:hip ratio following treatment. Since energy intake, a main contributor to weight gain, did not change during the 4 weeks, decrease in body composition indices is attributable either to PSO or to 9-cis, 11-trans CLA. In vivo studies on weight management and body composition found a reduction on perirenal, epididymal and omental adipose tissues following PA supplementation(Reference Arao, Wang and Inoue8, Reference Koba, Imamura and Akashoshi18).
In the present study, we found no beneficial relationship between insulin profiles and PSO supplementation. A recent animal study showed that PSO reduced the risk of diabetes by lowering glucose and insulin concentrations and insulin sensitivity(Reference McFarlin, Strohacker and Kueht28), a result also consistent for specific isomers of CLA, as shown by the Risérus et al. (Reference Risérus, Vessby and Arnlöv24) study which reported that administration of 3 g of 9-cis, 11-trans CLA over 12 weeks yielded a 15 % decrease in insulin sensitivity.
The present study has a few limitations. Serum PSO and CLA concentrations were not measured, which may further have facilitated assessment of compliance of the subjects by measuring PSO metabolites. Our sample size was small, and patients with mixed hyperlipidaemia were enrolled; study duration was 4 weeks and only one blood sample was obtained following intervention. Another limitation might be that we powered our study based on a 200 mg/l difference in total cholesterol between the PSO and placebo groups; this, however, was not achieved, and despite changes that were observed, one may speculate that the present study is underpowered.
To conclude, the present study found that PSO consumption in hyperlipidaemic subjects did not alter cholesterol and LDL-C but did reduce TAG and TAG:HDL-C ratio during 4-week study period. Further studies of larger samples with longer durations need to confirm the effects of PSO and related mechanisms.
Acknowledgements
The authors would like to thank the subjects who participated in the present study. The present study was supported by a financial grant (no. 234) from the Research Institute of Endocrine Sciences, Shahid Beheshti University Medical Sciences and Pharmacy and Pharmaceutical Sciences Research Center, Tehran University of Medical Sciences, Iran. None of the authors had any personal or financial conflicts of interest. P. M. was involved in design, implementation and analysis; M. R. F. was involved in preparation of PSO capsules and implementation; G. A. was involved in review of the literature, implementation and analysis; A. S. supervised of technical issues in the preparation of PSO and design; F. A. supervised overall project, design, implementation and analysis.
