Association of foods enriched in conjugated linoleic acid (CLA) and CLA supplements with lipid profile in human studies: a systematic review and meta-analysis

Abstract

Objective

The present study aimed to review the association of conjugated linoleic acid (CLA) consumption in two forms, foods enriched in CLA and CLA supplements, with serum lipid profile in human studies.

Design

Systematic review and meta-analysis.

Setting

Search process was conducted in PubMed, Cochrane Library, Google Scholar, Scopus and Science Direct. Clinical trials that investigated the association of CLA intakes either in the form of supplements or enriched foods with lipid profile in healthy adults were included. All outcomes were recorded as continuous variables, and the effect size was measured by analysis of the mean and standard deviation before and after the intervention for case and control groups.

Subjects

Healthy adult population.

Results

CLA supplementation was associated with a significant decrease in LDL cholesterol (mean difference=−0·218; 95 % CI −0·358, −0·077; P=0·002), a non-significant decrease in HDL cholesterol (mean difference=−0·051; 95 % CI −0·188, 0·086; P=0·468), a non-significant increase in total cholesterol (mean difference=0·009; 95 % CI −0·128, 0·146; P=0·896) and a non-significant decrease in TAG (mean difference=−0·065; 95 % CI −0·20, 0·07; P=0·344). Foods enriched with CLA were associated with significantly decreased LDL cholesterol (mean difference=−0·231; 95 % CI −0·438, −0·024; P=0·028), non-significantly increased HDL-C (mean difference=0·075; 95 % CI −0·121, 0·270; P=0·455), non-significantly decreased total cholesterol (mean difference=−0·158; 95 % CI −0·349, 0·042; P=0·124) and non-significantly decreased TAG (mean difference=−0·078; 95 % CI −0·274, 0·117; P=0·433).

Conclusions

According to our analysis, consumption of foods enriched with CLA or CLA supplements has favourable effects on LDL cholesterol levels.

Dyslipidaemia consists of different abnormalities in lipid profile and is one of the main risk factors for several diseases such as CVD, diabetes mellitus, hypertension, stroke and acute pancreatitis^{( 1 )}. The prevalence of dyslipidaemia depends on socio-economic status and ethnicity^{( 2 )}. It is increasing in most developed^{( 3 )} and developing countries owing to unhealthy diets and lifestyle changes^{( 4 , 5 )}. The main factors for dyslipidaemia are genetic, diet and lifestyle. According to previous studies, trans-fatty acids (TFA) play an important role in lipid profile disorders^{( 6 )}.

There are two sources of dietary TFA: (i) industrial TFA, which are produced technologically during the partial hydrogenation of vegetable oils; and (ii) ruminant TFA, such as vaccenic acid and conjugated linoleic acid (CLA) that are synthesized by rumen bacteria via the metabolism of MUFA and PUFA^{( 7 – 9 )}. Clinical studies have reported that dietary intake of industrial TFA has a deleterious effect on lipoprotein concentrations; however, ruminant TFA may be less detrimental to blood lipid levels than industrial TFA^{( 10 )}. Two isomers of CLA are cis-9, trans-11 (c9,t11) and trans-10, cis-12 (t10,c12)^{( 11 – 17 )}. The abundance of these isomers is different in foods and industrial supplements^{( 11 – 13 , 18 – 22 )}.

CLA is produced naturally by the rumen bacteria of ruminants^{( 14 , 19 , 23 – 27 )} or by bioconversion of vaccenic acid in the ruminant mammary gland^{( 26 , 28 )}. Moreover, it can even be produced synthetically by partial hydrogenation of linoleic acid^{( 20 , 25 )}. The main dietary sources of CLA are ruminant meats such as beef and lamb, and dairy products such as milk and cheese^{( 11 , 14 – 16 , 19 , 20 , 23 , 24 , 29 )}. The mean CLA intake is estimated at 0·3–2·6 g/d and daily intake of CLA through natural sources is 160 mg/d approximately^{( 22 , 30 )}.

Animal studies have shown that CLA might have various beneficial effects, e.g. prevention of carcinogenesis, decrease body fat, enhancement of lean body mass, empowering the immune system and prevention of diabetes and CVD^{( 14 , 15 , 21 , 31 – 33 )}. However, the findings of human studies are controversial^{( 17 , 25 , 26 , 30 , 34 )}. These differences may be related to the different forms and doses of CLA, study populations and duration of trials^{( 25 )}.

Some human studies have reported that CLA supplementation had no significant effect on plasma lipid concentrations^{( 18 , 21 )}; whereas another study found that CLA supplementation could significantly reduce total cholesterol (TC) and LDL cholesterol (LDL-C) in both genders and HDL cholesterol (HDL-C) only in women^{( 12 )}. Moreover, there are inconsistent findings on foods enriched in CLA. Some studies have claimed that CLA-rich dairy products significantly increased TC and LDL-C and decreased HDL-C; however, they had no significant effect on TAG concentration^{( 34 )}. On the other hand, another study indicated that the consumption of skimmed milk enriched with CLA had no significant effect on plasma lipid variables such as TAG, TC, LDL-C and HDL-C levels^{( 22 )}.

Studies on different forms of CLA, i.e. commercial natural products enriched in CLA or supplement forms, showed various findings; therefore it is necessary to summarize the controversial findings. The present study aimed to review the association of CLA consumption in two forms, foods enriched in CLA or CLA supplements, with serum lipid profile in human studies.

Methods

Literature search

The search was conducted in the following databases: PubMed, Cochrane Library, Google Scholar, Scopus and Science Direct, from 1 June to 23 November 2013. Keywords such as ‘trans-10 cis-12-conjugated linoleic acid’, ‘cis-9 trans-11-conjugated linoleic acid’, ‘CLA fatty acid’, ‘CLA’, ‘conjugated linoleic acid’, ‘trans fatty acid’, ‘TFA’, ‘Triglycerides’, ‘lipoprotein triglyceride’, ‘Lipoproteins, HDL’, ‘Cholesterol, HDL’, ‘Cholesterol, LDL’, ‘Lipoproteins, LDL’, ‘LDL’, ‘HDL’, ‘Total cholesterol’, ‘TG’, ‘triglyceride’, ‘triacylglycerol’, ‘TAG’, ‘lipid profile’, ‘low density lipoprotein’ and ‘high density lipoprotein’ were used. Keywords and medical subject heading (MeSH) terms are presented in Table 1. Age, gender and language were not limited during the search. Clinical trials that investigated the association of CLA intakes either in the form of supplements or enriched foods with lipid profile in healthy adults were included. Animal studies, studies on unhealthy individuals, study designs other than clinical trial, studies that investigated the effect of TFA other than CLA and studies that investigated outcomes other than lipid profile were excluded. Inappropriate forms of CLA, such as CLA plus n-3 fatty acid, CLA plus amino acid, CLA plus chromium picolinate, CLA plus creatine monohydrate or CLA plus exercise were excluded because these forms did not permit us to isolate the precise effect of CLA. Articles without complete data or placebo and articles on participants with metabolic and genetic disorders were excluded. Title and abstract of papers were screened and relevant papers were selected. Then, full texts of relevant papers were read and findings were re-screened. A flowchart of the literature search is shown in Fig. 1.

Table 1.

Search strategy for PubMed, Cochrane Library, Google Scholar, Scopus and Science Direct databases

No.
1	‘trans-10, cis-12-conjugated linoleic acid’ (Supplementary Concept) OR ‘cis-9, trans-11-conjugated linoleic acid’ (Supplementary Concept) OR ‘CLA fatty acid’ (Supplementary Concept) OR ‘CLA’ (tiab) OR ‘conjugated linoleic acid’ (tiab) OR ‘trans fatty acid’ (tiab) OR ‘TFA’ (tiab)
2	‘Triglycerides’ (MeSH) OR ‘lipoprotein triglyceride’ (tiab) OR ‘Lipoproteins, HDL’ (MeSH) OR ‘Cholesterol, HDL’ (MeSH) OR ‘Cholesterol, LDL’ (MeSH) OR ‘Lipoproteins, LDL’ (MeSH) OR ‘LDL’ (tiab) OR ‘HDL’ (tiab) OR ‘Total cholesterol’ (tiab) OR ‘TG’ (tiab) OR ‘triglyceride’ (tiab) OR ‘triacylglycerol’ (tiab) OR ‘TAG’ (tiab) OR ‘lipid profile’ (tiab) OR ‘low density lipoprotein’ (tiab) OR ‘high density lipoprotein’ (tiab)
3	1 AND 2

Fig. 1.

Flowchart of the literature search

Relevant papers were selected according to the title and abstract by three authors (S.-M.D.-R. and M.H.-B., R.K.). Two independent reviewers (S.-M.D.-R. and M.H.-B.) screened papers and read full texts of relevant papers. They assessed full texts for inclusion criteria and extracted data. Statistical analysis was done (M.M.) and cases of disagreement were resolved in consultation with a fourth arbitrating investigator (R.K.). Summaries of the clinical trials that investigated the association of CLA supplementation and foods enriched in CLA with lipid profile in human studies are shown in Tables 2 and 3, respectively.

Table 2.

Summary of clinical trials on the association of conjugated linoleic acid (CLA) supplementation and lipid profile in human studies

		Age (years)		Duration	CLA dose and		Placebo dose and form
Reference	Population	Mean	SD	(weeks)	form (g/d)	Isomers	(g/d)	Results
Iwata et al. (2007)( 57 )	Sixty males, healthy overweight and obese	41·5	9·6	12	3·4 g/d, CLA-TAG 6·8 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	10·8 g/d, high-linoleic safflower oil	TAG, HDL-C, LDL-C and TC levels did not change significantly among three groups
Watras et al. (2007)( 51 )	Forty males and females, healthy overweight	33	7·5	24	3·2 g/d, CLA-mix	c9,t11–t10,c12 (39·2:38·5)	4 g/d, safflower oil	No significant changes in TC, LDL-C, HDL-C or TAG concentrations were observed between groups
Gaullier et al. (2004)( 47 )	180 males and females, healthy overweight	45·83	10·3	48	3·4 g/d, CLA-TAG 3·6 g/d, CLA-NEFA	c9,t11–t10,c12 (50:50) c9,t11–t10,c12 (50:50)	4·5 g/d, olive oil	No effect on TC or TAG concentrations; CLA-TAG group had lower HDL-C concentrations and CLA-NEFA group had higher LDL-C concentrations than at baseline of the study
Gaullier et al. (2005)( 32 )	134 males and females, healthy overweight	46·26	9·96	96	3·4 g/d, CLA-TAG 3·4 g/d, CLA-NEFA	c9,t11–t10,c12 (50:50) c9,t11–t10,c12 (50:50)	3·4 g/d, placebo	Plasma TC and LDL-C were reduced, whereas HDL-C and TAG were unchanged
Blankson et al. (2000)( 13 )	Sixty males and females, healthy overweight and obese	44·35	12·95	12	1·7 g/d, CLA-TAG 3·4 g/d, CLA-TAG 5·1 g/d, CLA-TAG 6·8 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	9 g/d, olive oil	No significant differences were observed in blood lipids among the groups
Steck et al. (2007)( 33 )	Forty-eight males and females, healthy obese	34·50	4·85	12	3·2 g/d, CLA-TAG 6·4 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	8 g/d, safflower oil	HDL-C decreased significantly in placebo and 6·4 g CLA/d groups; other clinical laboratory values did not change across all groups
Noone et al. (2002)( 23 )	Fifty-one males and females, healthy normal-weight and overweight	31·37	6·31	8	3 g/d, CLA-TAG 3 g/d, CLA-TAG	c9,t11–t10,c12 (50:50) c9,t11–t10,c12 (80:20)	3 g/d, linoleic acid	Plasma TAG concentrations were significantly decreased in the 50:50 CLA supplement group but not in the 80:20 CLA or control groups; TC had no changes in all supplementation groups; HDL-C concentrations increased non-significantly in the control group; LDL-C concentrations decreased non-significantly in both CLA supplementation groups
Lambert et al. (2007)( 12 )	Sixty-two males and females, healthy regularly exercising non-obese	32	7	12	3·9 g/d, CLA-TAG	c9,t11–t10,c12–other isomers(29·7:30·9:2·9)	3·9 g/d, high-oleic-acid sunflower oil	TC and LDL-C reduced significantly in both genders; HDL-C decreased significantly only in women; TAG did not change significantly
Gaullier et al. (2007)( 31 )	118 males and females, healthy overweight and obese	47·25	9·6	24	3·4 g/d, CLA-TAG	c9,t11–t10,c12 (37·5:38·0)	4·5 g/d, olive oil	HDL-C decreased slightly in the CLA group; other blood lipids were not significantly changed in either group
Berven et al. (2000)( 14 )	Sixty males and females, healthy overweight and obese	47·05	3·9	12	3·4 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	4·5 g/d, olive oil	No significant changes were observed in blood lipid parameters
Mougios et al. (2001)( 15 )	Twenty-four males and females, healthy normal-weight and overweight	22·2	1·5	4–8	(0·7–1·4) g/d, CLA-mix	c9,t11–t10,c12 (50:50)	0·7–1·4 g/d, soyabean oil	HDL-C significantly reduced in all groups of CLA; TAG and TC tended to decrease in the CLA group during the low CLA intake but not during the high CLA intake
Petridou et al. (2003)( 16 )	Sixteen females, healthy sedentary normal-weight and overweight	22·30	1·80	6·5	2·1 g/d, CLA-mix	c9,t11–t10,c12 (50:50)	2·1 g/d, soyabean oil	CLA supplementation had no significant effect on TAG, TC, HDL-C and TC:HDL-C
Kamphuis et al. (2003)( 19 )	Sixty males and females, healthy overweight	35·1	8·35	13	1·8 g/d, CLA-TAG 3·6 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	1·8 g/d, oleic acid 3·6 g/d, oleic acid	CLA supplementation did not have any significant effect on plasma TAG concentrations
Pfeuffer et al. (2011)( 21 )	Eighty-five males, healthy overweight and obese	45–68	–	4	3·4 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	4·5 g/d, safflower oil	CLA decreased TC and LDL-C concentrations not significantly more than safflower oil. HDL-C, fasting and postprandial TAG did not change
Colakoglu et al. (2006)( 29 )	Forty-four females, healthy exercising normal-weight	21·15	1·85	6	3·6 g/d, CLA-mix	c9,t11–t10,c12	Control	CLA supplementation with or without exercise did not change serum lipid profile (TC, LDL-C, HDL-C, TAG)
Benito et al. (2001)( 58 )	Seventeen females, healthy normal-weight	28·15	6·2	9	3·9 g/d, CLA-TAG	c9,t11–t10,c12–c11,t13–t8,c10–cc–tt (11·4:14·7:15·3:10·8:6·74:5·99)	3·9 g/d, high-linoleic sunflower oil	CLA supplementation did not change the levels of plasma TC, LDL-C, HDL-C and TAG
Tavakoli-Darestani et al.( 59 )	Seventy-six females, healthy menopausal overweight women	55	6·65	12	3·2 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	4 g/d, high-oleic-acid sunflower oil	CLA supplementation had no significant effect on TC, TAG, LDL-C and HDL-C
Risérus et al. (2004)( 11 )	Twenty-five males, healthy overweight and obese	55	5·75	12	3 g/d, CLA-TAG	c9,t11–t10,c12–c9,c11–c10,c12–t9,t11+t10,t12 (83·3:7·3:0·46:0·2:1·4)	3 g/d, olive oil	CLA had no significant effects on lipoprotein or TAG concentrations compared with placebo
Sluijs et al. (2010)( 18 )	401 males and females, healthy overweight and obese	58·4	0·45	24	3·1 g/d, CLA	c9,t11–t10,c12 (80:20)	4 g/d, 80 % palm oil + 20 % soyabean oil	There was no effect of CLA supplementation on concentrations of lipids such as TAG, LDL-C, HDL-C and TC
Whigham et al. (2004)( 20 )	Sixty-four males and females, healthy overweight and obese	42·3	5·35	24	6 g/d, CLA-TAG	c9,t11–t10,c12–tt (37·3:37·6:1·3)	7·5 g/d, high-oleic acid sunflower oil	CLA increased TAG. Other lipids did not change
Song et al. (2005)( 54 )	Twenty-eight males and females, healthy normal-weight	31·35	7·01	12	3 g/d, CLA-TAG	c9,t11–t10,c12 (50:50)	3 g/d, high-oleic-acid sunflower oil	CLA supplementation did not change TC level. HDL-C level decreased significantly after 12 weeks of supplementation. LDL-C did not alter. Plasma TAG levels were increased in the two groups, however; significantly in the CLA group
Taylor et al. (2006)( 52 )	Forty males, healthy overweight and obese	46	7	12	4·5 g/d, CLA-mix	c9,t11–t10,c12 (35:36) c9,c11–c10,c12 (1–2 %) t9,t11–t10,t11 (1·5 %) t8,c10–c11,t13 (<1 %)	4·5 g/d, olive oil	There was no change in TC, TAG, LDL-C and HDL-C
Smedman and Vessby (2001)( 53 )	Fifty-three males and females, healthy	45·2	11·65	12	4·2 g/d, CLA-mix	c9,t11–t10,c12 (50:50)	4·2 g/d, olive oil	TC, LDL-C, HDL-C increased and TAG decreased
Kim (2008)( 48 )	Fifty-one females, healthy overweight Korean women	28·24	20·39	12	2·25 g/d, CLA-NEFA 2·25 g/d, CLA-TAG	c9,t11-CLA–t10,c12-CLA–c9,c11-CLA–t9,t11-CLA (37·95:38·84:0·96:1·35) c9,t11–CLA–t10,c12–CLA–c9,c11-CLA–t9,t11-CLA (37·83:38·55:0·98:1·86)	3 g/d, olive oil	No significant changes were observed within and between treatment groups in blood lipid parameters (TAG, TC, LDL-C or HDL-C)
Tholstrup et al. (2008)( 35 )	Seventy-five females, healthy postmenopausal women	60·16	4·46	16	4·6 g/d, CLA-mix 5·1 g/d, CLA-TAG	c9,t11–t10,c12–other CLA (41·17:39·90:1·79) c9,t11–t10,c12–other CLA (85·03:7·11:0·47)	5·5 g/d, olive oil	CLA mixture decreased HDL-C, increased TC:HDL-C compared with other groups and increased TAG levels compared with control. Plasma LDL-C concentrations did not differ among the three groups
Sahin et al. (2008)( 39 )	Twenty females, healthy overweight or obese premenopausal	22–48 (range)		8	1·8 g/d, CLA-NEFA	c9,t11–t10,c12 (80–84:37–42)	Without placebo	CLA reduced TC, TAG and LDL-C significantly and non-significantly increased HDL-C level
Tricon et al. (2004)( 41 )	Forty-nine males, healthy normal-weight	30·95	1·7	8	0·59, 1·19, 2·38 g/d, CLA-TAG 0·63, 1·26, 2·52 g/d, CLA-TAG	c9,t11–t10,c12 (79·3:7·8) c9,t11–t10,c12 (10·6:84·1)	Without placebo	CLA supplementation had significant effects on TC, LDL-C and no effect on HDL-C in all isomers and doses
von Loeffelholz et al. (2003)( 40 )	Fourteen males and females, healthy bodybuilders	26	4	24	3·78 g/d, CLA-TAG	c9,t11–t10,c12–t8,c10–c11,t13–cc–tt (8·3:7·9:6·0:7·1:4·7:17·7)	Without placebo	CLA supplementation significantly increased LDL-C and TC concentration only in the beginners, not in the advanced athletes, and had no significant effect on TAG or HDL-C levels in either intervention group
Albers et al. (2003)( 36 )	Seventy-one men, healthy overweight or obese	52·33	9	12	1·7 g/d, CLA-NEFA 1·6 g/d, CLA-TAG	c9,t11–t10,c12 (50:50) c9,t11–t10,c12 (80:20)	Sunflower oil fatty acids	CLA supplementation did not affect fasting serum lipids

Table 3.

Summary of clinical trials on the association of enriched foods with conjugated linoleic acid (CLA) and lipid profile in human studies

		Age (years)		Duration			Placebo form and
Reference	Population	Mean	SD	(weeks)	CLA form and dose	Isomers	dose	Results
Desroches et al. (2005)( 27 )	Sixteen males, healthy overweight and obese	36·6	12·4	8	Butter–CLA (4·22 g CLA/100 g fat)	c9,t11–other isomers (80:20)	Butter (0·38 g CLA/100 g fat)	Butter–CLA diet reduced TC significantly more than control. LDL-C, HDL-C and TAG levels did not change significantly between the two groups
Tricon et al. (2006)( 55 )	Thirty-two males, healthy	45·5	8·7	6	(Butter+cheese+ milk)–CLA (1·421 g CLA/d)	c9,t11	Butter+ cheese+milk (0·151 g CLA/d)	Dairy products enriched with CLA did not significantly affect TAG, TC, LDL-C and HDL-C. They slightly increased LDL-C:HDL-C
Wanders et al. (2010)(34)	Sixty-one males and females, healthy normal weight	30·9	13·7	9	(Margarine+ yoghurt drinks)–CLA (73·7 (sd 0·6) g CLA/100 g fat)	c9,t11–t10,c12 (80:20)	(Margarine+ yoghurt drinks)–oleic acid	TAG level did not change, LDL-C and TC:HDL-C increased, whereas HDL-C decreased in CLA group compared to control
Brown et al. (2011)( 26 )	Eighteen females, healthy normal-weight and overweight	20–40 (range)		8	Beef+dairy (ice cream, cheese, butter)–CLA (1·17 g CLA/d)	c9,t11–other isomers (87·5:12·5)	Beef+dairy (ice cream, cheese, butter) (0·35 g CLA/d)	No significant differences were observed in TC, TAG, LDL-C, HDL-C levels between treatment groups
Sofi et al. (2010)( 24 )	Ten males and females, healthy normal-weight and over weight	51·5		20	Pecorino cheese (1·56 g CLA/100 g lipid)	c9,t11	Placebo cheese (0·19 g CLA/100 g lipid)	TC, TAG, LDL-C and HDL-C did not change during either intervention phases
Raff et al. (2008)( 30 )	Thirty-eight males, healthy normal-weight	25·9	3·9	5	Butter–CLA (4·6 g/d CLA)	c9,t11–t10,c12 (39·4:38·5)	Butter (0·3 g CLA/d)	TC, TAG, LDL-C, HDL-C and TC:HDL-C did not differ during either intervention phase
Chen et al. (2012)( 25 )	Eighty males and females, healthy overweight and obese	32·8	0·8	12	Milk–CLA (1·7 g/d CLA)	c9,t11–t10,c12 (50:50)	Milk	CLA treatment increased levels of TC, TAG and LDL-C, decreased HDL-C concentration. None of these changes were significant
Naumann et al. (2006)( 17 )	Ninety-two males and females, healthy overweight and obese with LDL phenotype B	52·33	7·66	13	Drinkable dairy product–CLA (3 g CLA/d) Drinkable dairy product –CLA (3 g CLA/d)	c9,t11–t10,c12 (>80:<5) t10,c12–c9,t11 (>80:<5)	Drinkable dairy product (3 g high-oleic-acid sunflower oil/d)	LDL-C, HDL-C, TAG, TC:HDL-C, LDL-C:HDL-C did not change in CLA-enriched groups
Laso et al. (2007)( 22 )	Sixty males and females, healthy overweight and obese	53·85	7·73	12	Skimmed milk–CLA (3 g CLA/d)	c9,t11–t10,c12	Skimmed milk	Plasma TAG, TC and LDL-C increased slightly in all CLA groups, however these changes were not significant
Nazare et al. (2007)( 56 )	Forty-four males and females, healthy normal-weight and overweight	28·9	1·14	14	Yoghurt–CLA (3·76 g CLA/d)	c9,t11–t10,c12–tt (35:35:<1)	Yoghurt	CLA-enriched yoghurt did not alter any of the TAG, TC and HDL-C concentrations

Data extraction

Data of thirty-three articles that investigated the effect of CLA intake in either supplement form or enriched foods on lipid profile in healthy adult populations were entered into meta-analysis. Mean and standard deviation for TC, HDL-C, LDL-C and TAG before and after placebo or CLA consumption were extracted. Data from the following studies were not extracted: four studies without complete data for analysis^{( 35 – 38 )}, four studies without a placebo group^{( 32 , 39 – 41 )}, one study that considered special polymorphisms (PPARγ2, Pro12Ala) of healthy adults^{( 42 )}, one study done on adolescents^{( 43 )} and participants of three studies had signs of metabolic syndrome or borderline hyperlipidaemia^{( 44 – 46 )}. Complete information about excluded studies is shown in Fig. 1.

Two structural forms of CLA, i.e. TAG and NEFA, and two isomeric forms, i.e. cis-9, trans-11 isomer (c9,t11) and trans-10, cis-12 isomer (t10,c12), were used as intervention groups^{( 47 , 48 )}. There were different proportions (approximately 50:50 or 80:20; all proportions stated in the paper are by weight) of these isomers and we extracted results of all of them^{( 17 , 23 )}. Two studies reported their results stratified by gender or BMI^{( 12 , 22 )}. We entered their results into meta-analysis separately.

Statistical analysis

All outcomes were recorded as continuous variables, and the effect size was measured by analysis of the mean and standard deviation before and after the intervention for the case and control groups. Pooled meta-analyses were completed on studies that reported the same outcomes. The I ² statistic was used to test for heterogeneity; if there was significant heterogeneity, the random-effects model was used. I ² values of 25 %, 50 % and 75 % were used as evidence of low, moderate and high heterogeneity, respectively. Sensitivity analysis was done by successively removing a particular study that had the highest impact on the heterogeneity test. Comprehensive Meta-Analysis (CMA) software version 2 was used to carry out the data analysis. P values <0·05 were considered statistically significant. All reported P values resulted from two-sided versions of the respective tests. Potential publication bias was evaluated by Egger’s regression test^{( 49 )}. The trim and fill method was used to assess the potential effect of any publication bias on the meta-analysis results^{( 50 )}.

Results

Conjugated linoleic acid supplementation and LDL cholesterol

The summary mean difference and 95 % confidence interval for all fifteen clinical trial studies that investigated the effects of CLA supplementation on LDL-C are shown in Fig. 2. Heterogeneity among studies was significant (I ²=52 %; P _{heterogeneity}=0·040). The clinical trial studies^{( 11 , 13 , 18 , 21 , 29 , 51 – 53 )} contributed most to heterogeneity. In an analysis excluding these studies, CLA supplementation led to a significant decrease in LDL-C level (mean difference=−0·218; 95 % CI −0·358, −0·077; P=0·002); the test for heterogeneity was not statistically significant (I ²=0 %; P _{heterogeneity}=0·934). Publication bias was not significant (Egger’s test P value=0·17).

Fig. 2.

Meta-analysis of the effect of conjugated linoleic acid supplementation on LDL cholesterol in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Conjugated linoleic acid supplementation and HDL cholesterol

The summary mean difference and 95 % confidence interval for all seventeen clinical trial studies that investigated the effects of CLA supplementation on HDL-C are shown in Fig. 3. Heterogeneity among studies was significant (I ²=50 %; P _{heterogeneity}=0·030). The clinical trial studies^{( 11 , 13 , 18 , 21 , 29 , 51 – 53 )} contributed most to heterogeneity. In an analysis excluding these studies, CLA supplementation led to a slight and non-significant decrease in HDL-C level (mean difference=−0·051; 95 % CI −0·188, 0·086; P=0·468); the test for heterogeneity was not statistically significant (I ²=0 %; P _{heterogeneity}=0·649). Publication bias was not significant (Egger’s test P value=0·94).

Fig. 3.

Meta-analysis of the effect of conjugated linoleic acid supplementation on HDL cholesterol in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Conjugated linoleic acid supplementation and total cholesterol

The summary mean difference and 95 % confidence interval for all seventeen clinical trial studies that investigated the effects of CLA supplementation on TC are shown in Fig. 4. Heterogeneity among studies was significant (I ²=55 %; P _{heterogeneity}=0·034). The clinical trial studies^{( 11 , 13 , 18 , 21 , 29 , 51 – 54 )} contributed most to heterogeneity. In an analysis excluding these studies, CLA supplementation led to a slight and non-significant increase in TC level (mean difference=0·009; 95 % CI −0·128, 0·146; P=0·896); the test for heterogeneity was not statistically significant (I ²=0 %; P _{heterogeneity}=0·956). Since publication bias existed, we tried to evaluate the effect of publication bias by the trim and fill method. After eliminating the effect of publication bias, the combined mean difference was 0·0089 (95 % CI −0·125, 0·152), which remained consistent with previous results.

Fig. 4.

Meta-analysis of the effect of conjugated linoleic acid supplementation on total cholesterol in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Conjugated linoleic acid supplementation and TAG

The summary mean difference and 95 % confidence interval for all eighteen clinical trial studies that investigated the effects of CLA supplementation on TAG are shown in Fig. 5. Heterogeneity among studies was significant (I ²=54 %; P _{heterogeneity}=0·041). The clinical trial studies^{( 11 , 13 , 18 , 21 , 29 , 51 – 53 )} contributed most to heterogeneity. In an analysis excluding these studies, CLA supplementation led to a non-significant decrease in TAG level (mean difference=−0·065; 95 % CI −0·200, 0·070; P=0·344) and the test for heterogeneity was not statistically significant (I ²=0 %; P _{heterogeneity}=0·954). Publication bias was not significant (Egger’s test P value=0·08).

Fig. 5.

Meta-analysis of the effect of conjugated linoleic acid supplementation on TAG in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Foods enriched in conjugated linoleic acid and LDL cholesterol

The summary mean difference and 95 % confidence interval for all ten clinical trial studies that investigated the effects of foods enriched in CLA on LDL-C are shown in Fig. 6. Heterogeneity among studies was significant (I ²=51 %; P _{heterogeneity}=0·023). One clinical trial study^{( 26 )} contributed most to heterogeneity. In an analysis excluding that study, we found that foods enriched in CLA led to a significant decrease in LDL-C level (mean difference=−0·231; 95 % CI −0·438, −0·024; P=0·028); the test for heterogeneity was not statistically significant (I ²=1 %; P _{heterogeneity}=0·965). Publication bias was not significant (Egger’s test P value=0·18).

Fig. 6.

Meta-analysis of the effect of natural foods enriched with conjugated linoleic acid on LDL cholesterol in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Foods enriched in conjugated linoleic acid and HDL cholesterol

The summary mean difference and 95 % confidence interval for all eleven clinical trial studies that investigated the effects of foods enriched in CLA on HDL-C are shown in Fig. 7. Heterogeneity among studies was significant (I ²=50 %; P _{heterogeneity}=0·045). One clinical trial study^{( 26 )} contributed most to heterogeneity. In an analysis excluding that study, foods enriched in CLA led to a non-significant increase in HDL-C level (mean difference=0·075; 95 % CI −0·121, 0·270; P=0·455) and the test for heterogeneity was not statistically significant (I ²=19 %; P _{heterogeneity}=0·262). Publication bias was not significant (Egger’s test P value=0·07).

Fig. 7.

Meta-analysis of the effect of natural foods enriched with conjugated linoleic acid on HDL cholesterol in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Foods enriched in conjugated linoleic acid and total cholesterol

The summary mean difference and 95 % confidence interval for all eleven clinical trial studies that investigated the effects of foods enriched in CLA on TC are shown in Fig. 8. Heterogeneity among studies was significant (I ²=58 %; P _{heterogeneity}=0·018). One clinical trial study^{( 26 )} contributed most to heterogeneity. In an analysis excluding that study, we found that foods enriched in CLA led to a non-significant decrease in TC level (mean difference=−0·158; 95 % CI −0·349, 0·042; P=0·124) and the test for heterogeneity was not statistically significant (I ²=10 %; P _{heterogeneity}=0·345). Publication bias was not significant (Egger’s test P value=0·84).

Fig. 8.

Meta-analysis of the effect of natural foods enriched with conjugated linoleic acid on total cholesterol in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Foods enriched in conjugated linoleic acid and TAG

The summary mean difference and 95 % confidence interval for all eleven clinical trial studies that investigated the effects of foods enriched in CLA on TAG are shown in Fig. 9. Heterogeneity among studies was significant (I ²=56 %; P _{heterogeneity}=0·033). One clinical trial study^{( 26 )} contributed most to heterogeneity. In an analysis excluding that study, we documented that foods enriched in CLA led to a non-significant decrease in TAG level (mean difference=−0·078; 95 % CI −0·274, 0·117; P=0·433); the test for heterogeneity was not statistically significant (I ²=6 %; P _{heterogeneity}=0·384). Publication bias was not significant (Egger’s test P value=0·71).

Fig. 9.

Meta-analysis of the effect of natural foods enriched with conjugated linoleic acid on TAG in published clinical trials. The study-specific standardized difference (Std diff) in means and 95 % CI are represented by the black square and horizontal line, respectively; the area of the black square is proportional to the specific-study weight to the overall meta-analysis. The centre of the black diamond presents the pooled standardized difference in means and its width represents the pooled 95 % CI

Sensitivity analyses

To identify the source of the heterogeneity between studies, we performed sensitivity analyses by including and excluding some studies. Sensitivity analyses were done sequentially for all of the lipids and all of the studies. In a sensitivity analysis excluding one study at a time, we consistently found statistically the same results. Ranges of summary mean differences were (−0·242, −0·178), (−0·097, −0·016), (0·001, 0·017) and (−0·110, −0·040) for the effect of CLA supplementation on LDL-C, HDL-C, TC and TAG, respectively. Also the sensitivity analysis results based on the effect of enriched foods with CLA for different lipids according to summary mean differences were (−0·280, −0·200), (−0·082, −0·641), (−0·460, −0·222) and (−0·156, −0·047) for LDL-C, HDL-C, TC and TAG, respectively.

Discussion

The present meta-analysis is the first quantitative review of thirty-three randomized controlled clinical studies investigating the effect of CLA supplements and foods enriched in CLA on serum lipids separately. Our meta-analysis showed that intake of foods enriched in CLA decreased LDL-C levels significantly, decreased TC and TAG concentrations non-significantly and increased HDL-C levels non-significantly. CLA supplements decreased LDL-C, HDL-C and TAG levels and increased TC level; however, only the effect on LDL-C level was statistically significant. According to our analysis, consumption of foods enriched in CLA and CLA supplements has favourable effects on LDL-C level.

Some studies, in agreement with our results, showed that a mixture of CLA isomers decreased LDL-C level significantly in healthy adults^{( 12 , 34 , 39 , 41 )}. Noone et al.^{( 23 )} observed that the daily intake of 3 g CLA supplement (50:50 and 80:20) decreased LDL-C levels non-significantly in CLA groups. However, von Loeffelholz^{( 40 )} claimed that CLA supplementation for 6 months increased LDL-C and TC concentrations significantly. Some studies showed that CLA supplementation^{( 47 , 53 )} or foods enriched in CLA^{( 22 , 25 )} led to a slight, non-significant increase in LDL-C level.

We found that TC level decreased and HDL-C level increased non-significantly after intake of foods enriched in CLA and our findings are in accordance with other studies^{( 17 , 22 , 27 )}. According to our meta-analysis, CLA supplementation led to an adverse non-significant effect on TC or HDL-C level, which is in agreement with some studies on CLA supplements^{( 12 , 15 , 31 , 33 , 35 , 40 , 47 , 53 , 54 )} and is in disagreement with other studies^{( 21 , 39 , 41 )}.

Our analysis showed that TAG level decreased non-significantly after intake of either CLA supplements or CLA-enriched foods, similar to previous studies on either enriched foods or CLA supplements. Some findings suggested that CLA had no significant effect on TAG concentration^{( 11 – 19 , 21 , 22 , 24 – 27 , 29 – 34 , 36 , 40 , 47 – 53 , 55 – 59 )}. However, Chen et al.^{( 25 )} reported that TAG level increased in individuals who consumed foods enriched in CLA. Some trials reported a significant increase in TAG concentration after consuming CLA supplements^{( 20 , 35 , 41 , 54 )}.

The proportion of CLA isomers and their dosage may be important to determine the effect of CLA on lipid profile. Noone et al.^{( 23 )} showed that CLA supplementation with the 50:50 proportions of cis-9, trans-11 and trans-10, cis-12 isomers caused a significant reduction in plasma TAG concentrations; however, this effect disappeared with the 80:20 proportion of CLA isomers. Mougios et al.^{( 15 )} investigated the effect of CLA capsules that included 0·7–1·4 g CLA mixture for 4–8 weeks. They showed that low-dose CLA intake decreased TAG and TC and high CLA intake did not change TAG and TC levels.

Findings from human studies that investigated the effects of CLA mixtures or cis-9, trans-11 and trans-10, cis-12 CLA isomers separately on lipid profile in either enriched foods or supplement forms are controversial. This may be related to differences in the CLA forms (TAG or NEFA), doses of CLA (0·59–6·8 g in supplement forms and 1·17–73·7 g in enriched foods), variation in isomers and their proportions, duration of studies (from 4 weeks to 2 years in supplement forms and from 5 weeks to 5 months in enriched foods), variation in subjects’ body weight and different control groups. As placebo, most of the studies used olive oil or oleic acid extracts; some of them used safflower oil, sunflower oil or linoleic acid extracts; and a few studies used soyabean oil solely or in combination with palm oil. Studies enriched different kinds of dairy products such as cheese, milk, yoghurt, butter and ice cream with CLA. Furthermore, the CLA content of milk and other dairy products ranged from 0·34 % to 1·07 % of total fat, which is influenced by the diet of cows. In European countries, where cows are traditionally pasture grazed, their milk contains higher CLA levels than in countries where cows are mainly fed corn, such as the USA. These can lead to different results in studies^{( 60 )}.

The mechanism of lowering cholesterol level by CLA remains to be determined^{( 28 )}. It was suggested that CLA could decrease LDL-C particles by forbidding the secretion of apo B or by increasing the clearance rate of circulating LDL-C through increasing activity of the LDL receptor^{( 61 , 62 )}. According to evidence, dietary CLA enhances the fecal excretion of total neutral sterols^{( 63 )} and inhibits cholesterol absorption through down-regulation of intestinal acyl-CoA cholesterol acyltransferase^{( 28 )}. CLA can decrease TAG level by inhibiting the expression and activity of hepatic stearoyl-CoA desaturase. This enzyme is involved in the desaturation of substrate for the synthesis of TAG^{( 64 )}.

According to our meta-analysis, foods enriched in CLA and CLA supplements have beneficial effect on LDL-C concentration. CLA did not affect other lipids in the profile. Foods enriched in CLA increased HDL-C and tended to decrease TC non-significantly. Nutrients such as calcium, potassium, vitamin D and vitamin B, or bioactive peptides in dairy products, have been shown to be associated with beneficial outcomes. These nutrients and CLA can influence the lipid profile synergistically^{( 60 )}.

There are some concerns about the potential safety of CLA for human subjects. Studies have shown that supplementation with CLA or trans-10, cis-12 isomer could induce insulin resistance, lipodystrophy in animals, fatty liver, C-reactive protein enhancement and undesirable changes in lipid profile in man^{( 65 , 66 )}.

There is no consensus on the recommended dosage of CLA; however, according to evidence, 3 g/d seems to be most desirable. Consumption of CLA supplements is not recommended in pregnancy^{( 67 , 68 )}.

Conclusion

The present review showed that both CLA supplements and foods enriched in CLA caused a significant reduction in LDL-C level. Foods enriched in CLA, in comparison with CLA supplementation, had a beneficial effect on the whole lipid profile although only the effect on LDL-C level was statistically significant.