CVD is considered as a multifactorial disease which has a high mortality rate. Abnormal levels of blood lipids are among the risk factors for CVD(Reference Stamler and Neaton1, Reference Imamura, Doi and Arima2). The cholesterol-lowering effect of probiotics has been the centre of attention in numerous studies. The term probiotics denotes ‘for life’ and refers to micro-organisms which have positive effects on human health. Micro-organisms presiding in food, especially lactic acid bacteria, were believed to have beneficial health advantages from long ago. As probiotics, certain species among bifidobacteria and lactobacilli can help balance the intestine's microflora. This balance is formed by increasing the number of helpful bacteria and reducing pathogenic bacteria in the intestine. Factors such as inadequate nutrition, antibiotics, stress, ageing and infections can disturb this balance. Whenever this state of equilibrium is upset within the intestine, the harmful bacteria find a chance to thrive and multiply. The growth of these bacteria causes digestive ailments and illnesses(Reference Drisko, Giles and Bischoff3). In addition, balanced microflora can improve the barrier function of the gut, influence protective functions of the intestinal mucosa and modify its immune response(Reference Goldin and Gorbach4–Reference Shi and Walker6). Releasing antimicrobial substances in the body, resisting colonisation of harmful bacteria, stopping pathogens from attaching to the gastrointestinal tract, breaking up destructive elements and boosting the activity of brush-border enzymes are among the actions of probiotics.
In their study of the Maasai tribes, Mann & Spoerry noted for the first time that consumption of fermented milk reduced serum cholesterol concentration(Reference Mann and Spoerry7). In vitro experiments with some strains of bifidobacteria and lactobacilli reported that they were able to assimilate cholesterol if bile acids were present(Reference Gilliland, Nelson and Maxwell8–Reference Klaver and Van Der Meer11).
Probiotic bacteria can ferment indigestible carbohydrates in the food and produce SCFA in the gut. This can cause a reduction in the systemic levels of blood lipids(Reference Pereira and Gibson12). Some of these bacteria can assimilate cholesterol directly from the gastrointestinal tract or deconjugate bile salts and hence have an effect on the process of cholesterol absorption(Reference Gilliland, Nelson and Maxwell8, Reference Tahri, Grill and Schneider10). Some studies have attributed the hypocholesterolaemic effects of fermented milk to the sphingolipids present in yoghurt and the fatty acids in milk fat(Reference Fabian and Elmadfa13, Reference Kiessling, Schneider and Jahreis14).
Sphingolipids can be found in lipid-rich structures, such as cellular membranes, and especially lipoproteins. They have an effect on the process of cholesterol metabolism and transport.
Even though some human and animal studies have suggested a moderate cholesterol-lowering property of some fermented dairy products, clinical studies on human subjects have not been conclusive(Reference Fabian and Elmadfa13–Reference Greany, Bonorden and Hamilton-Reeves20). The contradictory outcome is most probably due to different experimental designs, different strains of lactic bacteria, and intakes of fermented milk as a confounding factor, thereby leading to difficulty in evaluating the results.
The evidence about the effect of probiotic yoghurt on lipid profile has shown conflicting results. Furthermore, in the studies that were conducted, usually probiotic and conventional yoghurt-consuming groups were compared. In the present study, in addition to the two yoghurt groups, a third group which did not use fermented milk was considered as a control group. The aim of the present study was to investigate the effect of probiotic yoghurt containing Lactobacillus acidophilus La5 and Bifidobacterium lactis Bb12 and conventional yoghurt on the lipid profile in healthy women.
Subjects and methods
The present study was a randomised clinical trial. A total of ninety female volunteers aged between 19 and 49 years, with cholesterol levels less than 6·2 mmol/l, TAG levels less than 2·3 mmol/l and a BMI less than 30 kg/m2 were recruited from the personnel and students of Tehran University of Medical Sciences (TUMS) through an advertisement. Exclusion criteria were: smoking; having kidney, liver or inflammatory intestinal disease, thyroid, diabetes, disorders, immunodeficiency diseases, lactose intolerance; taking supplements, taking medication; consuming probiotic yoghurt or any other probiotic-containing products in the last 2 months; being elite athletes, pregnant or breast-feeding.
A 1-week pre-adjustment period was designated, during which all subjects had to refrain from taking yoghurt or any other fermented food. Subjects were randomly assigned into three groups, each group consisting of thirty individuals. The first and second groups consumed 300 g daily of probiotic and conventional yoghurt, respectively. The third group, as the control group, did not consume any fermented and probiotic products for the duration of the study (6 weeks). The volunteers were told not to alter their exercise routine or regular diet, and not to consume any yoghurt other than the one provided to them by the researcher. They were also asked to refrain from consuming any other probiotic and fermented products. The study was triple blind for the yoghurt-consuming groups. That is, in addition to the subjects and the investigator, the evaluator of the results was also not aware which treatment any particular subject received. Necessary arrangements were made so that every week the subjects would receive a week's supply of their probiotic or conventional yoghurts directly from the factory. Of the subjects, two had to be excluded from the study at week 1 because of taking antibiotics: one from the probiotic group and one from the control group.
Both the probiotic and conventional yoghurts that were supplied contained Lactobacillus bulgaricus and Streptococcus thermophilus. The probiotic yoghurt was also enriched with adding cultures of B. lactis Bb12 and L. acidophilus La5 (Chr Hansen, Denmark). Direct Vat Set (DVS) cultures were used. Microbiological analysis of the probiotic yoghurt showed that it contained 3·9 × 107 colony-forming units (cfu) of both B. lactis Bb12 and L. acidophilus La5. The analysis on conventional yoghurt confirmed the presence of 106–107 cfu of S. thermophilus and L. bulgaricus. The yoghurts' acidity was pH 4·3. The fat content was 2·5 %, comparable in both yoghurt types.
Information on food consumption (through 3 d dietary records), fasting blood samples and anthropometric measurements were collected at three intervals: at the beginning of the trial (T1), at the end of week 3 (T2) and at the end of week 6 (T3). Compliance with the yoghurt consumption guidelines at home was monitored once per week through phone interviews.
At each of the three intervals, body weights were measured (digital floor scale; Seca, Hamburg, Germany) with 0·1 kg accuracy without shoes and with minimum clothing. Weight was measured in the fasting state. The subjects' heights were measured, with 0·1 cm accuracy, with non-stretchable tape (Seca). BMI was determined by dividing body weight by height squared (kg/m2).
At the end of each interval 3 d dietary records were taken from each volunteer, who was instructed to write down the types and amounts of foods eaten. The dietary record comprised of one weekend and two non-consecutive weekdays. The amount of food eaten by each subject was estimated from household measures and from photographs of portion sizes(Reference Ghafarpour, Houshiar-rad and Kianfar21). Through weekly follow-ups by phone, and through periodical visits of the subject at each interval to the Nutrition Department at TUMS, a nutritionist checked the subjects' compliance with the study protocol and assessed dietary records in person.
The subjects were directed to report at TUMS at the end of each interval for blood tests. The blood sample was drawn from the antecubital vein in the arm. The blood was drawn after a 14 h fast, in the morning. Serum total cholesterol and TAG concentrations were measured enzymically with Parsazmun's kits (DiaSys, Germany) using cholesterol oxidase-p-aminophenazone (CHOD-PAP) and glycerol phosphate oxidase-p-aminophenazon (GPO-PAP)(Reference Artiss, Zak, Rifai and Warnick22, Reference Cole, Klotzsch, McNamara, Rifai, Warnick and Dominiczak23). HDL-cholesterol and LDL-cholesterol were also measured enzymically with Parsazmun's kits (DiaSys, Germany)(Reference Rifai, Bachorik, Alberts, Burtis and Ashwood24).
The data were analysed using ANOVA. ANOVA was used to identify any differences among the three groups or among the three intervals within a group. The Kolmogorov–Smirnov test was used to test the normality of the distribution of variables. For distributions that were not normal the non-parametric test of Kruskal–Wallis was used to compare the groups, and the data were expressed in medians and interquartile ranges of the original variable. Otherwise, the data were expressed as means and standard deviations. Multiple comparisons were conducted by the Bonferroni post hoc test. The statistical tests were conducted using SPSS (version 11.5; SPSS, Inc., Chicago, IL, USA). Food Processor II (ESHA, Salem, OR, USA) was used in performing nutrient calculations for the 3 d dietary records. The database of this software is built upon the Nutrient Database Bank for Standard Reference from the US Department of Agriculture and other sources. The database was modified with reference to the existing national Iranian food composition table, developed by the Iranian National Institute of Nutrition and Food Technology. In developing the regional food composition database, biological variation, fortification and enrichment of food items (such as rice, wheat, dairy products, oils, etc) were taken into account. This program is used for analysing national nutrition surveys in Iran. P < 0·05 was considered as statistically significant.
The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Ethics Committee at TUMS. Written informed consent was obtained from all subjects.
The mean values and standard deviations of weight, height, age and BMI of the volunteers at the beginning of the study did not indicate any significant differences among the three study groups. The mean age of the conventional and probiotic groups was 32 and 35·5 years, respectively, and 34·7 years for the control group.
ANOVA did not show any significant change in weight and BMI of the subjects at the three intervals of the study within a group (Table 1). ANOVA did not show any difference among the three groups at the beginning of the study for intakes of total fat, SFA, MUFA, PUFA and dietary fibre, based on the 3 d dietary records. There was no significant change in the intakes of these parameters during the study, within any group (Table 1). The Kruskal–Wallis test showed no statistically significant change in total energy and cholesterol intakes at the beginning of the study among the three groups. There was no significant difference in total energy or cholesterol intake throughout the study within a group (Table 1).
T1, data collection at the beginning of the study; T2, data collection after 3 weeks; T3, data collection after 6 weeks.
* Median and interquartile ranges.
ANOVA showed that there was no statistically significant difference among the three groups in blood lipid parameters at baseline. Neither was there any statistically significant difference throughout the study for blood lipid parameters within any group (Table 2). In the probiotic yoghurt group, a descending trend in total:HDL-cholesterol ratio of 12·5 % (P>0·05) was observed.
T1, blood sample collection at the beginning of the study; T2, blood sample collection after 3 weeks; T3, blood sample collection after 6 weeks.
Comparing the mean differences between baseline (T1) and 6 weeks (T3) for TAG and LDL-cholesterol showed no statistically significant difference among the three groups (Table 3). The post hoc test showed that the decrease (T1 − T3) in total cholesterol was significant for both the conventional and probiotic yoghurt groups (P = 0·014 and P = 0·002, respectively) compared with the control group. The increase in the mean difference (T1 − T3) concentration of HDL-cholesterol was significant only for probiotic yoghurt compared with the control group (P = 0·009). For the total:HDL-cholesterol ratio the post hoc test showed that the decrease was significant for both the conventional and probiotic yoghurt groups compared with the control group (P = 0·006 and P = 0·000, respectively) (Table 3).
Mean value was significantly different from that of the control group: * P < 0·05, ** P < 0·01.
In the present study there were no statistically significant differences between the three groups in terms of weight and BMI. It is possible that by introducing yoghurt to the diet of subjects, a change in energy intake and, consequently, weight gain occurs. We did not find any significant change in weight, and BMI within any group during the study.
Though we did not observe statistically significant changes in nutrient intakes during the course of the study, there was a descending trend in the intake of total energy and SFA in the control group, compared with the two yoghurt-consuming groups (Table 1). The reason for the observed trend could be that consumption of yoghurt as a side-dish is part of a dietary habit in Iran and yoghurt is one of the main items in the daily household basket. For the two yoghurt-consuming groups, their dietary yoghurt was substituted by the experimental products. But for the control group, yoghurt consumption was restricted and it was withdrawn from their diet. Although an equal amount of energy, through other foods, was introduced to the diet of the control group to compensate for the yoghurt restriction, in terms of fat content, it could have resulted in a reduction in energy and saturated fat intake in the control group.
The present results showed no statistically significant difference in lipid profile parameters within any group throughout the study. There was a lowering effect (T1 − T3 mean difference) due to the consumption of both conventional and probiotic yoghurts on total cholesterol and total:HDL-cholesterol ratio compared with the control group. We did not observe any significant differences in lipid profile between the two yoghurt groups. We did, however, observe significant differences when compared with the control group. At baseline, the PUFA:SFA ratio was 1·03 for both the conventional and probiotic yoghurt groups. This ratio was 0·95 for the control group. In the control group, throughout the study the ratio increased due to a decrease in SFA and the mean value of PUFA remained almost constant at T1 and T3 intervals. The increase observed in plasma total cholesterol and total:HDL-cholesterol ratio in the control group could be due to the increase in the intake of cholesterol in this group. In the yoghurt-consuming groups, too, an increase in the intake of cholesterol was observed, but there was a decrease in PUFA:SFA ratios. Despite these changes, there was an improvement in plasma total cholesterol and total:HDL-cholesterol ratio in the yoghurt-consuming groups. Consumption of fermented milk could account for the reduction in cholesterol concentration in the present study. Some early studies on fermented milk also showed cholesterol-lowering effects of these products(Reference Hepner, Fried and St Jeor25). Adding L. acidophilus to infant formula resulted in a lower serum cholesterol concentration compared with the control formula(Reference Harrison and Peat26).
In a study by Schaafsma et al. a significant reduction of serum cholesterol concentrations was seen with daily consumption of fermented milk containing two strains of L. acidophilus after 21 d(Reference Schaafsma, Meuling and Bouley27). The yoghurt also had 2·5 % fructo-oligosaccharides, which may have affected serum cholesterol concentrations. In a study by Agerbaek et al. on the effect of a milk product, fermented with Enterococcum faecium and two strains of S. thermophilus, on lipid profile, a significant reduction in LDL-cholesterol concentration was found after 6 weeks of daily consumption of a fermented milk product compared with a placebo (milk acidified with an organic acid)(Reference Agerbaek, Gerdes and Richelsen28). Long-term consumption of the same product resulted in LDL-cholesterol reduction similar to the placebo group, indicating that both low-fat milk and fermented milk products may have some hypocholesterolaemic effects(Reference Richelsen, Kristensen and Pedersen29).
The significant decrease in the mean difference of total:HDL-cholesterol ratio in both yoghurt groups compared with the control in the present study could be the consequence of decreases in the total cholesterol level as well as increases in HDL-cholesterol concentration after consuming fermented milk for 6 weeks(Reference Kiessling, Schneider and Jahreis14, Reference Schaafsma, Meuling and Bouley27).
The observed effect in both yoghurt groups could be because of sphingolipids in yoghurt and the distribution of milk fat(Reference Fabian and Elmadfa13, Reference Kiessling, Schneider and Jahreis14). Short- and long-term animal studies with rats have shown that sphingolipids reduce plasma cholesterol(Reference Imaizumi, Tominaga and Sato30, Reference Kobayashi, Shimizugawa and Osakabe31). Some studies have shown that sphingolipids have an LDL-cholesterol- and total cholesterol-reducing effect and an increasing effect on HDL-cholesterol(Reference Vesper, Schmelz and Nikolova-Karakashian32). Other studies have indicated that sphingolipids have a reversing effect on the LDL-:HDL-cholesterol ratio(Reference Smedman, Gustafsson and Berglund33). Other studies suggested that the increase in HDL-cholesterol and decrease in total:HDL-cholesterol ratio is because of SFA (12 : 0 and 14 : 0) in milk fat(Reference Samuelson, Bratteby and Mohsen34–Reference Temme, Mensink and Hornstra36). The conventional and probiotic yoghurts that we used in the present study contained 2·5 % fat. Therefore, sphingolipid contents in both yoghurts should be similar. It should be noted, however, that another potential source of sphingolipids is cell membranes of bacteria themselves, which were higher in the probiotic yoghurt. This could have led to a higher sphingolipids content in the probiotic yoghurt and as a result an increase in HDL-cholesterol.
Furthermore, the results from some studies that used probiotic capsules instead of dairy products as a vehicle for administrating probiotics did not support the cholesterol-lowering effect of probiotic bacteria(Reference Greany, Bonorden and Hamilton-Reeves20, Reference Lewis and Burmeister37). Rather, they supported the claim that a dairy product is a more effective medium of administering probiotic bacteria. It was proposed that orotic acid and 3-hydroxyl-methyl glutarate in milk, independently, could have cholesterol-lowering effects.
In the present study, other mechanisms could be responsible for the hypocholesterolaemic effect that we observed in the group that consumed probiotic yoghurt. It has been proposed that when probiotics settle in the gut, they ferment indigestible carbohydrate from food. Their action raises the SCFA in the gut. SCFA are produced from peptides, polysaccharides, proteins and oligosaccharides, mainly by anaerobic bacteria, and are the final product of bacterial activity in the gastrointestinal tract. These large molecules are depolymerised by a variety of hydrolytic enzymes that are produced by bacteria and allow the organisms to ferment their sugar content. SCFA can lower the lipids in blood through blocking the synthesis of hepatic cholesterol and/or through redirecting plasma cholesterol toward the liver. By producing bile acids through deconjugating the bile salts in the small intestine, probiotics prevent micelle production. When cholesterol enters the enterohepatic circulation, it is dealt with in the same way. Probiotics, by using hydroxy steroid dehydrogenase and conjugated bile acid hydrolase enzymes, break down the bile acid and hydrolyse bile salts(Reference Gilliland, Nelson and Maxwell8, Reference Klaver and Van Der Meer11, Reference Gilliland and Walker15). By doing so, the enterohepatic circulation of bile acids will be disrupted(Reference De Boever and Verstraete38–Reference Doncheva, Antov and Softova40). Hydroxy methyl glutarate CoA (HMG CoA) is another compound that helps probiotics block HMG CoA reductase activity, which is a rate-limiting enzyme and is involved in the endogenous production of cholesterol. Probiotic bacteria reduce the absorption of cholesterol in the intestine by binding and hence incorporating it into the cell membrane. Cholesterol can also be assimilated during growth. For the mentioned mechanisms to operate, lactic acid bacteria must remain alive when passing through the gastrointestinal tract. This seemed to be true only in the probiotic yoghurt group, but not in the conventional yoghurt group. All of the above-mentioned activities help with the cholesterol-lowering actions of yoghurt.
In the present study, changes in lipid profile were observed in both probiotic and conventional yoghurt-consuming groups compared with the control group. We did not observe significant differences between lipid profile parameters in the two yoghurt groups. The fact that changes in lipid profile were significant in the yoghurt-consuming groups compared with the control group, and not with each other, leads us to draw the conclusion that any effect is due to the consumption of fermented milks and not necessarily the probiotic yoghurt. The present study showed that both products had a positive effect on the lipid profile of the subjects.
The present study was supported by a grant (no. 6242) from the Vice-chancellor for Research, TUMS, Iran. We are grateful to the Research and Development Division of Iran Dairy Industry Corporation (IDIC-Pegah), in Tehran, that provided the dairy products for the present study.
I. E. initiated the study, supervised the project, and assisted in interpretation and in writing the manuscript. H. S.-Y. conducted the study, carried out statistical analysis and interpretation and wrote the manuscript. A. D. supervised the study and assisted in writing the manuscript. M. J. supervised the laboratory analysis. R H. supervised the statistical analysis. M. C. assisted in conducting the study.
The authors declare that there are no conflicts of interest.