Original Paper

Effect of Fermented Milk (Yogurt) Containing Lactobacillus Acidophilus L1 on Serum Cholesterol in Hypercholesterolemic Humans
James W. Anderson, MD and Stanley E. Gilliland, PhD Metabolic Research Group, VA Medical Center, University of Kentucky, Lexington, and Department of Animal Science, Oklahoma State University, Stillwater, Oklahoma Key words: lactobacilli, fermented milk, yogurt, cholesterol, hypercholesterolemia
Objective: Two controlled clinical studies were performed to examine effects of consumption of one daily serving of fermented milk (FM) (yogurt) on serum lipids. Methods: In the first study, subjects were randomly allocated to FM containing Lactobacillus acidophilus L1 of human origin or to FM containing L. acidophilus ATCC 43211 of swine origin. In this single-blind study, subjects consumed one 200 ml serving of FM daily for 3 weeks. The second study was a double-blind, placebo-controlled, cross-over study. Subjects completed a 4-week first treatment, had a 2-week washout, and completed a second 4-week treatment. In the second study subjects consumed FM containing L. acidophilus L1 or placebo FM over 4 weeks. Results: In the first study, FM containing L. acidophilus L1 was accompanied by a 2.4% (p 0.05) reduction of serum cholesterol concentration. In the second study, strain L1 reduced serum cholesterol concentration by 3.2% (p 0.05) in the first treatment period. In the second treatment period there were no significant changes in serum cholesterol concentration. Combined analysis of the two L1 treatment studies demonstrated a 2.9% (p 0.01) reduction in serum cholesterol concentration. Conclusion: Since every 1% reduction in serum cholesterol concentration is associated with an estimated 2% to 3% reduction in risk for coronary heart disease, regular intake of FM containing an appropriate strain of L. acidophilus has the potential of reducing risk for coronary heart disease by 6 to 10%.

INTRODUCTION
Since the early studies of Mann and Spoerry [1] there has been an increasing interest in the cholesterol-lowering properties of fermented milk (FM) (yogurt) products, and numerous studies have focused on the potential hypocholesterolemic activity of FMs in humans [2 8]. From these seven clinical studies, six reported a reduction in serum cholesterol concentration associated with FM intake. The following median reductions in serum cholesterol concentration were obtained: 1 week, 5.4%; 2 weeks, 6.6%; 3 weeks, 6.9%; and 4 weeks, 6.4%. These findings of cholesterol-lowering activity of FM products were also confirmed in experimental animals [9 12]. These and other studies strongly suggest that FMs have an important cholesterol-lowering potential. However, as was already demonstrated from the clinical study of Jaspers and

colleagues [4], different strains of lactic acid bacteria may have different effects on serum cholesterol concentrations. Although the underlying cholesterol-lowering mechanism(s) has not yet been sufficiently elucidated, the metabolism of cholesterol and/or bile salts by the bacteria incorporated in the FM product may be responsible for the cholesterol-lowering effect. Gilliland and colleagues [1316] made in vitro comparisons of isolates of intestinal L. acidophilus to assess their indigenous capacities to tolerate bile, to deconjugate bile salts, and to assimilate cholesterol. Furthermore, when an isolate of L. acidophilus from a pig (strain ATCC 43121) with high in vitro activity (i.e., bile resistance, bile salt hydrolysis and cholesterol assimilation) was administered to cholesterolfed pigs, serum cholesterol values were significantly lower than values of controls [16]. No such cholesterol reduction was observed upon feeding a L. acidophilus strain with poor bile

Address reprint requests to: James W. Anderson, MD, Chief Endocrine-Metabolic Section, VA Medical Center, Cooper Drive Division (111C), Lexington, KY 40511.

Journal of the American College of Nutrition, Vol. 18, No. 1, 4350 (1999) Published by the American College of Nutrition 43

Fermented Milk Lowers Serum Cholesterol

tolerance and low cholesterol-assimilation capacity in vitro [16]. Because strains of L. acidophilus of human origin may be more effective as dietary adjuncts in humans, Buck and Gilliland [13] characterized several strains of L. acidophilus of human origin for their in vitro activity towards bile salts and cholesterol. The L. acidophilus strain L1 showed high in vitro activity with respect to bile resistance, bile salt hydrolysis and cholesterol assimilation. Therefore, these studies were designed to examine the cholesterol-lowering effects of FM containing L. acidophilus L1. In the first study, L. acidophilus L1 was compared with the pig strain ATCC 43121. In the second study L1 was compared with a FM containing S. thermophilus, but containing no L. acidophilus. Since a 1% reduction in serum cholesterol is associated with a 2 to 3% reduction in estimated risk for coronary heart disease [17], even small reductions in serum cholesterol concentration associated with intake of FM containing L. acidophilus would be important. Table 1. Characteristics of the Subjects
Group Study I L1 FM ATCC FM Study II L1 FM Placebo
* M/F: Male/Female. ** Mean (SEM).

Number of subjects 14 15 21 19

M/F*

Age** (years) 49 (2) 51 (3) 55 (3) 58 (3)

3/11 6/9 9/12 9/10

Five subjects withdrew voluntarily because of personal reasons such as schedule conflicts. The remaining subjects in both studies, whose medical history, medications and serum lipids were consistent with inclusion and exclusion criteria, were divided into two experimental groups (Table 1). The initial subject characteristics did not significantly differ between the two groups (Table 1) for both studies.

MATERIALS AND METHODS

Subjects
Male and female subjects with primary hypercholesterolemia (type IIa or IIb lipoprotein patterns) were recruited for both studies. Initial serum cholesterol values ranged in the first study between 4.69 and 7.75 mmol/L, and in the second study between 5.40 and 8.32 mmol/L and fasting serum triglyceride concentrations were 5.65 mmol/L for all subjects. Subjects with familial or secondary hypercholesterolemia, diabetes mellitus or hypothyroidism were excluded from the studies. Subjects had not received hypocholesterolemic medication within the last 3 weeks, and had not had a myocardial infarction, angioplasty or stroke within the last 3 months. They were not treated with anticoagulants, immunosuppressants, corticosteroids or estrogens, and they did not have thyroid replacement. They were lactose tolerant and had not consumed fresh FM in the last 2 weeks. Females who were pregnant or had plans to become pregnant during the time course of the studies or who were lactating were excluded. Subjects had no recent history of alcohol abuse. Subjects were prescreened to determine if their complete blood count and blood chemistry panel (albumin, alkaline phosphatase, CO2, bilirubin, BUN, calcium, GGT, globulin, glucose, LDH, potassium, SGOT, SGPT, sodium, total protein and uric acid), as well as TSH and urinalysis, were within the normal range. All subjects gave informed consent to a protocol approved by the Institutional Review Board of the University of Kentucky. For the first study 30 subjects enrolled, and one subject withdrew consent after the screening visit. Table 1 provides the gender and average ages for subjects. Forty-eight subjects enrolled in the second study, and eight subjects were not randomized. Three failed the screen with elevated TSH levels.

Study Design
Study I. Study I was a single-blind, random-allocation, parallel study to compare serum lipid effects of two different FM products over a 3-week period; FM containing human L. acidophilus L1 (L1 FM) was compared with swine L. acidophilus strain ATCC 43121 (ATCC FM). Subjects, randomly allocated to one of two groups, received either 200 g of L1 FM or ATCC FM after each evening meal. Prior to the first FM consumption period, two fasting lipid measurements were made at baseline. Fasting serum lipid measurements were made at weeks 2 and 3, and subjects had weekly body weight measurements. At these weekly visits, subjects received fresh FM and adherence to fermented products was assessed. Subjects were further questioned about any symptoms or medication changes since their last visit. Study II. Study II was a double-blind, random-allocation, placebo-controlled cross-over study of 4-weeks duration with a 2-week wash-out period, to compare the effects on serum lipids of FM containing the L. acidophilus L1 strain (L1 FM) with a FM not containing these active bacteria (placebo FM). Subjects were randomly allocated to one of two groups receiving either 200 g placebo FM or L1 FM product after each evening meal for 4 weeks. In this study after the 2-week washout period, subjects received the other FM product for the final 4 weeks. Prior to the first FM consumption period, three fasting serum lipid measurements were made at baseline and two fasting lipid measurements were done before starting the second FM treatment. Fasting serum lipid measurements were made at week 2, 3 and 4 of each FM treatment period, and subjects had weekly body weight measurements. At these weekly visits, subjects received fresh FM and the adherence to fermented products

44

VOL. 18, NO. 1

Fermented Milk Lowers Serum Cholesterol

was assessed. Subjects were further questioned about symptoms or medication changes since their last visit.

Measurements
Body weight was determined by the use of a calibrated balance. Blood was drawn after an overnight fast for measurement of serum cholesterol [18], HDL-cholesterol [19], and triglyceride concentrations [20]. Serum LDL-cholesterol levels were calculated as previously described [21]. Nutrient intake was estimated by analysis of three-day diet records using a computerized nutrient data base (Nutritionist IV, 1994, NSquared Computing, San Bruno, CA). The concentrations of viable L. acidophilus cells for both strains in the FMs were determined on the first and seventh day of each weekly FM batch by plating serial dilutions prepared in physiological saline with Rogosa agar (Difco). The viable counts of L. acidophilus determined after 72 h of anaerobic incubation, were expressed as CFU/g FM product.

Diet Instructions
All subjects were instructed to maintain usual dietary habits throughout the study and to complete 3-day diet records, which were to be returned at each baseline and after each FM treatment period. While on FM treatment, subjects were asked not to use sweet Acidophilus milk and not to incorporate high fiber items such as preserves into FM.

Fermented Milk (FM) Preparation

All FM products were weekly produced under sanitary food preparation conditions at the Department of Animal Science (Dairy Section) of the University of Kentucky by staff of the Dutch dairy company Campina Melkunie. Three products were used in the two studies. All products were produced using demineralized water containing 12% (w/v) skimmed milk powder and 0.4% (w/v) gelatin (Bloom 240). The L1 FM was inoculated with 1.0% (v/v) L. acidophilus L1 and 0.1% (v/v) S. thermophilus MUH34 culture. The ATCC FM was inoculated with 1.0% (v/v) L. acidophilus ATCC 43121 starter culture and 0.1% (v/v) S. thermophilus MUH34 culture. Placebo FM was solely inoculated with 1.0% (v/v) S. thermophilus MUH34 culture. All FM products were fermented at 37C until a final pH of ca. 4.6 was reached. Fermentations times were as follows: L1 FM, 18 hours; ATCC FM, 18 hours; and placebo, 24 hours. All batches of fresh products of L1 FM product contained viable L1 counts exceeding 1 107 colony forming units (CFU)/g. All cultured milks closely resembled yogurts available for purchase in food stores. Freshly prepared FMs were dispensed into 200 g-cups, packaged and stored at 7C prior to delivery to experimental subjects on the next 2 days.

Statistics
Change scores (treatment values minus baseline) among different treatments were compared by a one way analysis of variance (ANOVA). Change scores between groups also were compared using two-tailed T-tests assuming equal variance. Baseline and treatment values within groups were compared using paired T-tests. A p value of 0.05 was considered to be statistically significant.

RESULTS
Study I: Serum Lipid Changes
Serum Cholesterol. Table 2 presents serum lipid and body weight responses to a daily intake of 200 g of L1 FM or ATCC FM. Serum cholesterol concentrations in the L1 FM group decreased significantly by 2.4% after treatment (p 0.05). There was a small, but nonsignificant, reduction of serum cholesterol concentration in the ATCC FM group after treatment.

Table 2. Serum Lipida (mmol/L) and Body Weighta (kg), Responses in Hypercholesterolemic Subjects to Fermented Milk Containing Human L. acidophilus L1 (L1 (L1 FM) and to Fermented Milk Containing Swine L. acidophilus ATCC 43121 (ATCC FM)
L1 FMb Measurement Baseline Cholesterol LDL-cholesterol HDL-cholesterol Triglycerides Body weight
a b

ATCC FMc Change, % 2.4 (1.1) 2.6 (2.2) 3.9 (1.6) 3.2 (4.6) 0.6 (0.2) Baseline 6.38 (0.23) 4.21 (0.16) 1.50 (0.14) 1.47 (0.14) 72.1 (3.2) Treatment 6.30 (0.21) 4.15 (0.17) 1.40 (0.12)** 1.64 (0.20) 71.8 (3.1) Change, % 0.9 (1.7) 1.1 (2.6) 5.9 (1.7) 16.5 (9.6) 0.9 (0.2)

Treatment 6.12 (0.20)* 4.06 (0.19) 1.12 (0.08)* 2.06 (0.26) 75.7 (4.8)

6.27 (0.19) 4.18 (0.19) 1.17 (0.08) 2.12 (0.24) 75.7 (4.8)

Values represent mean (SEM) of two measurements. n 14. c n 15. P vs. baseline, * 0.05; ** 0.01.

JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION

45

Fermented Milk Lowers Serum Cholesterol

LDL Cholesterol. Overall serum LDL-cholesterol values after treatments were 2.6% and 1.1% lower than baseline in the L1 FM and ATCC FM group, respectively, but were not significant. However, LDL-cholesterol concentrations in group L1 FM decreased significantly (p 0.01) after receiving yogurt for 2 weeks (data not shown). HDL Cholesterol. Serum HDL-cholesterol values decreased significantly in both groups, i.e., 3.9% in the L1 FM group (p 0.05) and 5.9% in ATCC FM group (p 0.01). Triglycerides. No significant changes of serum triglyceride concentrations were observed in either L1 FM or ATCC FM groups. Intake of placebo FM product was not accompanied by significant changes in serum cholesterol concentrations during either the first or second treatment period (Table 3). LDL-Cholesterol. Changes in serum LDL-cholesterol concentrations followed the pattern observed for total serum cholesterol values (Table 3). In the first treatment period, changes with L1 FM in group II, 4.1%, were significantly greater (p 0.018) than changes with placebo product in group I, 1.8%. Intake of L1 FM product by group II was accompanied by a significant reduction (p 0.031) in values during the first treatment period. Baseline values averaged 4.41 mmol/L, whereas average values after treatment with L1 FM were 4.22 mmol/L. Thus, L1 FM product was accompanied by a 4.1% reduction in LDL-cholesterol concentrations during the first treatment period. However, similar to the results for the total serum cholesterol, the L1 FM product did not significantly affect serum LDL-cholesterol levels during the second treatment period. Intake of placebo FM containing no L. acidophilus cells did not bring about significant changes in LDL-cholesterol values during either the first or second treatment period. HDL Cholesterol. Upon ANOVA testing, no significant differences in HDL-cholesterol values were detected between the changes among the four groups of L1 and placebo FM intake. While intake of L1 FM was accompanied by a significant reduction (p 0.003) in HDL-cholesterol concentrations during the first treatment period, the change scores did not differ significantly from the placebo change scores. Serum Triglycerides. ANOVA analysis did not reveal significant differences in serum triglyceride concentrations between changes among the four groups of L1 FM and placebo FM product intake. Serum triglyceride values averaged 2.07 mmol/L during the first baseline period. Values after 4 weeks of treatment with the L1 FM product averaged 1.81 mmol/

Study II: Serum Lipid Changes

Serum Cholesterol. Table 3 presents serum lipid and body weight responses to daily intake of 200 g of the L1 and placebo FM during a 4-week period in study II. ANOVA showed a significant difference between changes among the four groups of L1 FM and placebo FM intake (p 0.013). In the first treatment period, serum cholesterol changes (treatment minus baseline) with L1 FM product in group II, 3.2%, were significantly greater (p 0.006) than changes with placebo FM in group I, 1.4% (Table 3). Intake of L1 FM was accompanied by a significant decrease in serum cholesterol values during the first treatment period. Serum cholesterol concentrations averaged 6.53 mmol/L during baseline and decreased to 6.31 mmol/L (p 0.003). Values after 4 weeks of treatment averaged 6.25 mmol/L, significantly below (p 0.01) baseline values (see Table 6 presented later). The average treatment values were 3.2% below baseline values and the 4-week values were 3.8% below baseline levels. During the second treatment period, L1 FM did not significantly affect serum cholesterol concentrations (Table 3).

Table 3. Serum Lipida (mmol/L) and Body Weighta (kg) Responses to Fermented Milk Containing L. acidophilus L1 (L1 FM) and Placebo FM Product
Group Ib Total cholesterol LDL cholesterol HDL cholesterol Triglycerides Body weight Group IIc Total cholesterol LDL cholesterol HDL cholesterol Triglycerides Body weight
a b

Baseline 1 6.30 (0.14) 4.30 (0.12) 1.18 (0.07) 1.90 (0.24) 74.0 (3.8) Baseline 1 6.53 (0.17) 4.41 (0.17) 1.25 (0.09) 2.07 (0.24) 77.0 (3.7)

Placebo FM 6.38 (0.14) 4.38 (0.12) 1.17 (0.07) 1.81 (0.18) 73.7 (3.9) L1 FM 6.31 (0.16)** 4.22 (0.12)* 1.21 (0.09)** 2.01 (0.27) 77.4 (3.6)

Change (%) 1.4 (1.4) 1.8 (1.6) 0.8 (0.17) 3.7 (6.3) 0.4 (0.8) Change (%) 3.2 (1.1)** 4.1 (2.0)* 3.2 (0.8) 3.4 (4.8) 0.5 (0.3)

Baseline 2 6.45 (0.16) 4.53 (0.14) 1.18 (0.06) 1.63 (0.14) 74.7 (3.7) Baseline 2 6.42 (0.16) 4.29 (0.16) 1.27 (0.10) 1.91 (0.22) 77.3 (3.6)

L1 FM 6.60 (0.20) 4.52 (0.17) 1.23 (0.08) 1.75 (0.23) 74.7 (3.7) Placebo FM 6.43 (0.17) 4.29 (0.14) 1.28 (0.10) 1.90 (0.19) 77.5 (3.6)

Change (%) 2.3 (0.6) 0.2 (2.0) 2.5 (2.5) 11.7 (10.4) 0.3 (0.3) Change (%) 0.3 (0.0) 0.2 (1.9) 0.8 (1.6) 0.5 (3.7) 0.3 (0.1)

Values represent the mean (SEM) of three (Baseline 1, Placebo FM, L1 FM) or two measurements (Baseline 2). n 19. c n 21. Significant differences are indicated: treatment values are compared to baseline; change scores are compared to placebo; * p 0.05; ** p 0.01.

46

VOL. 18, NO. 1

Fermented Milk Lowers Serum Cholesterol

(data not shown), which was significantly below (p 0.021) baseline data. The 4-week values were 12.4% below baseline values. ATCC 43121 content averaged 1.8 107 CFU/g on day 1 and 6.7 106 CFU/g on day 7. Study II. In the first treatment period, L. acidophilus counts averaged 7.6 107 CFU/g at day 1 and 5.9 107 CFU/g at day 7. In the second period, L. acidophilus counts averaged 5.1 107 CFU/g at day 1 and 4.2 107 CFU/g at day 7. During the second treatment period the average fermentation-time of the L1 FM was significantly longer that in the first treatment period (19.0 1.28 hours vs. 17.2 0.49 hours, respectively).

Study I and II
Body Weight. No significant differences in body weight were found upon ANOVA analysis among all the groups in both studies. Dietary Intake. The average dietary intakes of study I and II are given in Tables 4 and 5, respectively. Both information during baseline periods and FM treatments are presented, including the information of group-I and -II subjects during baseline periods and FM treatments in the second study. There were no significant differences in nutrient intake between all experimental groups and no significant changes are seen. Combined Serum Cholesterol. Since design of the first study and the first wave of the second study were almost identical and both used L. acidophilus L1, we have evaluated the serum cholesterol results combining these two studies. Table 6 summarizes these results. Swine L. acidophilus ATCC 43121 showed a consistent reduction, which was not statistically significant. Placebo FM was associated with no distinct trends. The response to human L. acidophilus L1 showed remarkable similarities in time course for the two studies. ANOVA testing indicated that values differed between the three groups (L1-1st study, L1-2nd study, and placebo) at 2 weeks (p 0.03) and for overall treatment (p 0.011). When compared to baseline values, serum cholesterol values were significantly lower at 2 weeks (p 0.0015), 4 weeks (p 0.02), and overall treatment (p 0.001). Compared to placebo, serum cholesterol values were significantly lower at week 2 (p 0.008), and week 3 (p 0.022), and at overall treatment (p 0.0045). Human L. acidophilus L1 showed a net 4.3% reduction of serum cholesterol values for overall treatment, compared with the placebo FM (p 0.005).

DISCUSSION
In study I we compared a fairly newly isolated strain of L. acidophilus (L1) from human origin with a strain of pig origin (ATCC 43121). Effects on serum cholesterol concentrations of neither strains had been previously studied in human subjects. This first study showed that the swine strain had a very small and statistically insignificant lowering effect on serum cholesterol. However, the human strain had a significant lowering effect on total serum and LDL cholesterol. Study II was a placebo-controlled study to examine effects of human L. acidophilus L1. To increase power of the study a cross-over design was used. While a significant reduction of the serum cholesterol level was observed in the first treatment period, this was not seen in the second treatment period. Since both studies used the same L. acidophilus L1, presented in the same fermented vehicle and prepared in the same manner, we think it is appropriate to look at the combined responses. This also addresses the problem recognized at the outset that these studies did not have adequate numbers of subjects to provide optimal statistical power to address these questions. The combined results indicate that this yogurt product has a potential to lower serum cholesterol by 3 to 4% in hypercholesterolemic individuals. During the second treatment period of study II there were no significant changes in serum lipid concentrations for subjects receiving L1 FM. Several interacting factors may have contributed to the different response of these subjects. It is noteworthy

Bacterial Counts
Study I. L. acidophilus L1 content averaged 3.5 107 CFU/g at day 1 and 3.9 106 CFU/g at day 7. L. acidophilus

Table 4. Daily Dietary Intakesa of Group L1 FM and Group ATCC FM During Baseline and Treatment Periods. Values Were Obtained by Means of 3-day Diet Records
L1 FMb Baseline Energy (kcal/day) Cholesterol (mg/day) Proteins (% energy) Carbohydrates (% energy) Fatty acids (% energy) Alcohol (% energy)
a

ATCC FMc Treatment 1578.5 (121.6) 184.2 (22.6) 18.3 (1.4) 49.9 (2.5) 30.6 (2.3) 1.1 (0.4) Baseline 2084.7 (167.4) 199.8 (31.2) 16.5 (0.9) 51.5 (2.6) 28.6 (1.9) 3.5 (1.3) Treatment 1978.9 (161.3) 228.6 (30.1) 17.3 (0.8) 52.7 (2.1) 27.7 (1.9) 2.3 (0.7)

1845.9 (138.7) 207.8 (27.2) 17.7 (1.2) 50.7 (2.5) 30.5 (2.1) 1.0 (0.6)

Mean (SEM). n 14. c n 15.

b

JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION

47

Fermented Milk Lowers Serum Cholesterol

Table 5. Daily Dietary Intakesa of Group-I and -II Subjects During Cross-over Experiment. Values Were Obtained by Means of 3-day Diet Records (Study II)
Group Ib Energy (kcal) Cholesterol (mg) Saturated fat (g) Proteins Carbohydrates Fatty acids Alcohol Group IIc Energy (kcal) Cholesterol (mg) Saturated fat (g) Proteins Carbohydrates Fatty acids Alcohol
a Mean (SEM). % of the energy intake. b n 19. c n 21.

Baseline 1 2029.5 (140.4) 246.1 (26.9) 24.0 (2.1) 16.9 (0.8) 51.6 (1.3) 30.2 (1.1) 1.2 (0.5) Baseline 1 1865.0 (138.3) 221.1 (23.7) 19.0 (2.4) 16.4 (1.1) 50.8 (2.7) 30.9 (2.7) 1.9 (0.7)

Placebo FM 1776.8 (125.1) 198.9 (23.5) 19.9 (1.7) 18.2 (1.5) 50.1 (1.7) 30.3 (1.4) 1.0 (0.5) L1 FM 1814.4 (128.0) 228.8 (31.6) 21.2 (2.7) 17.8 (1.2) 48.2 (2.4) 30.8 (1.9) 2.8 (1.1)

Baseline 2 2013.2 (165.2) 267.0 (26.9) 27.1 (2.8) 15.9 (0.6) 48.4 (1.5) 34.1 (1.3) 1.6 (0.7) Baseline 2 1801.1 (125.0) 231.2 (22.2) 21.0 (2.1) 18.0 (1.1) 49.9 (1.4) 29.6 (1.1) 2.5 (0.9)

L1 FM 1751.6 (103.4) 200.9 (21.2) 22.2 (1.9) 15.8 (0.6) 50.3 (1.7) 32.1 (1.3) 1.8 (1.3) Placebo FM 1858.5 (157.3) 214.1 (32.4) 21.7 (2.3) 16.9 (1.1) 49.6 (1.5) 31.9 (1.1) 1.5 (0.7)

that group I did show a significant rise during the whole study, both serum cholesterol total and LDL-cholesterol concentrations increasing about 5%. After careful analyses of dietary compliances, adherence to fermented milk products, preparation techniques used for FM and bacterial counts, we can not explain these differences. Since the initial observations of Mann and Spoerry [1] in 1974, a number of investigators have examined effects of intake of FM or yogurt on serum cholesterol concentrations in humans. Harrison and Peat [23] evaluated the effects of introducing L. acidophilus into formula of newborn infants; infants receiving formula containing L. acidophilus had lower serum cholesterol values than infants fed control formula. These early observations suggested that introduction of L. acidophilus into human intestine might have a hypocholesterolemic effect. Since these early observations, at least nine clinical studies have examined the effects of FM or yogurt intake on serum cholesterol concentrations in humans. Early studies prior to 1984 did not control for fat intake on the different diets and are difficult to interpret. Only recent studies have assessed bacterial counts of tested yogurt products. The studies of Payens [7], Rossouw [8] Hepner [3] and Bazzarre [2] are difficult to interpret because fat contents of the various test diets were not controlled. The types of bacteria provided in yogurt are not specified in the studies of Massey [5] and McNamara [6]. Nevertheless, the last five studies [4 6,24,25] appear to be well-controlled studies. Massey and colleagues [5] prepared their yogurt in batches for use over 4 weeks of study. The decrease in cholesterol-lowering effects observed at 4 weeks may have been related to a lower bacterial count at 4 or more weeks after preparation of the yogurt.

Two recent studies [24,25] appear to have been well controlled and both reported a significant reduction in serum cholesterol concentrations with yogurt consumption. Agerbaek et al [24] found a significant 3.5% reduction in serum cholesterol concentration in the experimental group after 3 weeks daily consumption of a FM containing an Enterococcus faecium and two strains of S. thermophilus. After 6 weeks, a 6% reduction was found. In a study of Schaafsma et al [25] a significant 4.5% reduction of serum cholesterol concentration was seen with daily consumption of a yogurt containing a L. acidophilus during 3 weeks. This yogurt also contained 2.5% fructo-oligosaccharides which additionally may have affected serum cholesterol concentrations [26]. The mechanisms responsible for the hypocholesterolemic effects of FM are still under investigation. Gilliland and colleagues [1316,27,28] suggested that bacterial metabolism of cholesterol and bile acids contribute to the hypocholesterolemic effects of L. acidophilus. The early study of Harrison and Peat [23] suggested that introduction of live cultures of L. acidophilus into intestine of humans may have hypocholesterolemic effects. The recent studies of Gilliland and colleagues [16,27,28,33] have shown that some strains of L. acidophilus when growing under anaerobic conditions and in the presence of bile (as would be encountered in the small intestine) assimilate cholesterol, part of which is incorporated into their cellular membrane. Strains of L. acidophilus studied in their laboratories also deconjugate bile acids which can interrupt the enterohepatic circulation of bile acids. Free bile acids can be precipitated in presence of Ca2 , which would lead to a higher elimination of bile salts in the feces. To maintain the steady state the amount of excreted bile salts must be replaced by new

48

VOL. 18, NO. 1

Fermented Milk Lowers Serum Cholesterol

bile acid. In other words if bile salt hydrolase activity increases in the small intestine, there is a greater demand for cholesterol [29]. Another possibility is that free bile acids are absorbed by passive diffussion and have a negative feedback on cholesterol and bile acid synthesis. De Rodas et al [30] have shown in a pig feeding trial, a significant relationship between deconjugation of bile acids in pig and reduction of serum cholesterol concentration caused by a selected strain of L. acidophilus. Also De Smet [31] showed in a pig feeding trial that a bile salt hydrolase (BSH) active Lactobacillus reuteri caused a significant reduction of serum cholesterol. Dietary fibers also may exert their hypocholesterolemic effects through binding of bile acids and decreasing their availability for reabsorption in the distal small intestine [32]. In conclusion, the current studies coupled with previous controlled clinical studies indicate that daily intake of FM products containing selected strains of L. acidophilus has the potential to significantly reduce serum cholesterol concentrations. Reductions of serum cholesterol concentrations of 3 to 4% are clinically meaningful since every 1% reduction in serum cholesterol concentrations leads to a 2 to 3% reduction in estimated risk for coronary heart disease [17]. Thus, regular intake of FM containing an active cholesterol-reducing L. acidophilus could decrease estimated risk for coronary heart disease by 6 to 10%. However further studies are required to better understand the mechanism of action and the magnitude of the effect from FM intake.
3.2 (1.1) 1.4 (1.3) 2.4 (1.1) 0.9 (1.7) 2.9 (0.8) 1.8 (1.0) 0.12 (0.06) 3.6 (1.2) 6.31 (0.14) 0.073 0.008 0.03 6.25 (0.11) 0.02 0.022 0.072 0.27 (0.08) 3.8 (1.3) 6.23 (0.12) 0.001 0.19 (0.05)

% Chg

0.15 (0.07) 0.08 (0.10)

% Chg

Week 4 baseline

Table 6. Combined Study Results for Serum Cholesterol Changes; Values in mmol/L, with Means (SEM)

Week 3 baseline

0.09 (0.11) 0.08 (0.15)

3.2 (1.2) 6.19 (0.24) 1.2 (1.4) 6.30 (0.21)

3.8 (1.7) 6.39 (0.17) 1.9 (1.6) 6.51 (0.15)

Week 3

0.14 (0.08) 0.16 (0.11)

2.0 (1.2) 2.69 (1.7)

% Chg

1.5 (1.6) 0.7 (2.4)

6.25 (0.14) 6.25 (0.19)

Week 4

n.d. n.d.

0.27 (0.11) 0.25 (0.14)

n.d. n.d.

3.8 (1.6) 6.31 (0.16) 1.0 (2.2) 6.38 (0.14)

n.d. n.d.

6.12 (0.20) 6.30 (0.21)

Treatment average

0.22 (0.07) 0.08 (0.09)

Change

0.0045 0.011

ACKNOWLEDGMENTS
We appreciate and thank the following: Dr. C. Hicks for use of the facilities to prepare the yogurt; F. Elbers, Dr. J. Schlatmann and R. Vermin of Campina Melkunie for providing and preparing the yogurt for the study; T. Colson, J. Blake and L. Nichols of the Metabolic Research Group who performed study procedures with subjects; and R. Tijssens for helpful discussions. This study was supported by Campina Melkunie.

% Chg

0.26 (0.10) 0.10 (0.10) 21 6.53 (0.17) 6.27 (0.18) 19 6.30 (0.14) 6.40 (0.13)

Week 2 baseline

0.21 (0.08) 0.08 (0.09)

14 6.27 (0.19) 6.06 (0.17) 15 6.38 (0.23) 6.30 (0.23)

35 6.42 (0.13) 6.18 (0.13) 0.0015

Week 2

0.24 (0.07)

REFERENCES
1. Mann GV, Spoerry A: Studies of a surfactant and cholesterolemia in the Maasai. Am J Clin Nutr 27:464469, 1974. 2. Bazarre TL, Wu SL, Yuhas JA: Total and HDL cholesterol concentrations following yogurt and calcium supplementation. Nutr Rept Int 28:12251232, 1983. 3. Hepner G, Fried R, StJeor S, Fusetti L, Morin R: Hypocholesterolemic effect of yogurt and milk. Am J Clin Nutr 32:1924, 1979. 4. Jaspers DA, Massey LK, Luedecke LO: Effect of consuming yogurt prepared with three culture strains on human serum lipoproteins. J Food Sci 49:11781181, 1984.

No.

Baseline average

1st Study L1 FM ATCC FM 2nd Study L1 FM Placebo Combined L1 FM P vs. baseline P vs. placebo ANOVA

JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION

n.d. no data.

Group

49

Fermented Milk Lowers Serum Cholesterol

5. Massey LK: Effect of changing milk and yogurt consumption on human nutrient intake and serum lipoproteins. J Dairy Sci 55:255 262, 1984. 6. McNamara DJ, Lowell AE, Sabb JE: Effect of yogurt intake on plasma lipid and lipoprotein levels in normolipidemic males. Atherosclerosis 79:167171, 1989. 7. Payens VW, Rethans EJM, de Waard H: The influence of consumption of a large quantity of yoghurt or milk on the serum cholesterol concentration. Milchwissenschaft 30:525530, 1995. 8. Rossouw JE, Burger E, Van Der Vyver P: The effect of skim milk, yoghurt, and full cream milk on human serum lipids. Am J Clin Nutr 34:351356, 1981. 9. Danielson AD, Peo ER, Shahani KM: Anticholesterolemic properties of Lactobacillus acidophilus yogurt fed to mature boars. J Anim Sci 67:966974, 1989. 10. Grunewald KK: Serum cholesterol levels in rats fed skim milk with Lactobacillus acidophilus. J Food Sci 47:20782079, 1992. 11. Kiyosawa H, Sugawara C, Sugawara N, Miyake H: Effect of skim milk and yogurt on serum lipids and development of sudanophilic lesions in cholesterol-fed rabbits. Am J Clin Nutr 40:479484, 1984. 12. Rao DR, Chawan CB, Pulusani SR: Influence of milk and thermophilus milk on plasma cholesterol levels and hepatic cholesterogenesis in rats. J Food Sci 46:13391341, 1981. 13. Buck LM, Gilliland SE: Comparisons of freshly isolated strains of Lactobacillus acidophilus of human intestinal origin for ability to assimilate cholesterol during growth. J Dairy Sci 77:29252933, 1994. 14. Gilliland SE, Walker DK: Factors to consider when selecting a culture of Lactobacillus acidophilus as a dietary adjunct to produce a hypocholesterolemic effect in humans. J Dairy Sci 73:905911, 1990. 15. Walker DK, Gilliland SE: Relationships among bile tolerance, bile salt deconjugation, and assimilation of cholesterol by Lactobacillus acidophilus. J Dairy Sci 76:956961, 1993. 16. Gilliand SE, Nelson CR, Maxwell C: Assimilation of cholesterol by Lactobacillus acidophilus. Appl Environ Microbiol 49:377 381, 1985. 17. Manson JE, Tosteson H, Ridker PM, Satterfield S, Hebert P, OConnor GT: The primary prevention of myocardial infarction. N Engl J Med 326:14061416, 1992. 18. Allain CC, Poon LS, Chan CSG, Richmond W, Fu PC: Enzymatic determination of total serum cholesterol. Clin Chem 20:470475, 1974. 19. Finley PR, Schifman RB, Williams RJ, Lichti DA: Cholesterol in high-density lipoprotein: use of Mg2 /dextran sulfate in its enzymatic measurement. Clin Chem 24:931933, 1978. 20. Bucolo G, David H: Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem 19:476482, 1973. 21. Friedewald WT, Levy RI, Fredrickson DS: Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparative ultracentrifuge. Clin Chem 17:499 502, 1972. 22. Ito M, Kimura M: Influence of lactose on faecal microflora in lactose maldigesters. Microbial Ecolog Health Dis 6:7376, 1993. 23. Harrison VC, Peat G: Serum cholesterol and bowel flora in the newborn. Am J Clin Nutr 28:13511355, 1975. 24. Agerbaek M, Gerdes LU, Richelsen B: Hypocholesterolaemic effect of a new fermented milk product in healthy middle-aged men. Euro J Clin Nutr 49:346352, 1995. 25. Schaafsma G, Meuling WJA, van Dokkum W: Zuivel kan serumcholesterol verlagen. Voeding (Dutch) 57:1214, 1996. 26. Roberfroid M: Dietary fiber, inulin, and oligofructose: a review comparing their physiological effects. Crit Rev Food Sci Nutr 33:103148, 1993. 27. Gilliland SE: Acidophilus milk products: A review of potential benefits to consumers. J Dairy Sci 72:24832494, 1989. 28. Gilliland SE: Health and nutritional benefits from lactic acid bacteria. FEMS Microbiol Rev 87:175188, 1990. 29. De Smet I, Van Voorde N, De Saeyer N, Vande Woestyne M, Verstraete W: In vitro study of Bile Salt Hydrolase (BSH) activity of BSH isogenic Lactobacillus plantarum 80 strains and estimation of cholesterol lowering through enhanced BSH activity. Microbial Ecolog Health Dis 7:315329, 1994. 30. De Rodas BZ, Gilliland SE, Maxwell CV: Hypocholesterolemic action of Lactobacillus acidophilus ATCC 43121 and calcium in swine with hypercholesterolemia induced by diet. J Dairy Sci 79:21212128, 1996. 31. De Smet I: Bile salt hydrolysis by lactobacilli: microbial ecological significance and initial in vivo evaluation of the relation to blood cholesterol. (Thesis). Laboratory of Microbial Ecology, University of Gent (B), June 1995. 32. Anderson JW, Spencer DO, Riddel-Mason S, Floore TL, Dillon DW, Oeltgen PR: Postprandial serum glucose, insulin, and lipoprotein responses to high- and low-fiber diets. Metabolism 44:848 854, 1995. 33. Noh DO, Kim SH, Gilliland SE: Incorporation of cholesterol into cellular membrane of Lactobacillus acidophilus ATCC 43121. J Dairy Sci 80:31073113, 1997.

Received March 1998; revision accepted June 1998.

50

VOL. 18, NO. 1