The effects of wine consumption and lipid profile: A systematic review and meta-analysis of clinical trials

Abstract

Background and objective

The aim of this study was to analyze the effects of wine consumption on the lipid profile, distinguishing between triglycerides, total cholesterol, LDL, HDL and fibrinogen.

Methods

We examined the MEDLINE (via PubMed), Scopus, Cochrane, and Web of Science databases to conduct this systematic review and meta-analysis. PROSPERO has already recorded this study under registration number CRD42023396666.

Results

Thirty-three studies were included in this systematic review, and 29 were included in the meta-analysis. The pooled ES for the effect of red wine consumption on the different lipid profile parameters was significant only for the effect of red wine on the LDL parameter in the prepost studies (−0.29 (95% CI −0.54, −0.05)). The pooled ES for the effect of white wine in prepost studies and clinical trials for the effect of wine consumption on the different parameters did not show any significant results.

Conclusion

This systematic review and meta-analysis revealed that wine consumption has an effect on reducing LDL and has no effect on total cholesterol, HDL, triglycerides, or fibrinogen. This research revealed that the duration of the intervention affects triglyceride and total cholesterol levels, indicating that longer interventions are more effective for these two parameters.

Registration ID

CRD42023396666 (PROSPERO).

URL registration

https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=396666.

1. Introduction

Cardiovascular disease is the leading cause of death, doubling mortality from other diseases worldwide [1,2]. The lipid profile is the principal cardiovascular risk factor, as it is associated with coronary heart disease [3], especially heart stroke [4,5]. This association also affects populations at lower cardiovascular risk [6], such as the young population [7]. Total serum cholesterol and low-density lipoprotein (LDL) are associated with ischemic heart disease [8,9], LDL is the most useful serum marker for predicting the risk of stroke [10], and triglycerides are associated with an increased risk of atherosclerosis and atherothrombosis. Compared with other cardiovascular risk factors, the associations of cardiovascular diseases with lipid profiles have rarely been studied, and these values have tended to increase in recent years in countries such as China and the United States [11]. In Spain, half of the adult population is affected by hypercholesterolemia, and 45% have high LDL cholesterol levels, with no difference found between men and women [12]. In addition, 24% of deaths from cardiovascular disease are related to high cholesterol, which causes 4.4 million deaths per year, accounting for 7.8% of total mortality [13].

Research shows that the treatment of dyslipidemia proportionally decreases the incidence of cardiovascular events [9]. The best treatment for dyslipidemia is pharmacologically based on the use of statins in the case of hypercholesterolaemia [9]. In addition to controlling LDL cholesterol levels, statins improve endothelial function, increase nitric oxide, have an antioxidant effect, and stabilize atheromas [14]. However, this treatment could have negative effects, as it could cause hepatotoxicity, muscle symptomatology, and problems in diabetic patients, increasing insulin resistance [15]. Although the benefits of pharmaceutical drugs are greater than their negative effects are, it is necessary to implement other types of interventions, such as lifestyle modifications, including physical exercise and diet, in the prevention of dyslipidemia [9,16].

Physical activity is associated with a lower incidence of dyslipidemia, an improved HDL cholesterol concentration and a reduced risk for men with low HDL cholesterol [16]. Moreover, diet is associated with the lipid profile, which, as a modifiable factor, has shown that a higher-quality diet benefits HDL and triglyceride concentrations [17]. Wine consumption together with other lifestyle changes, such as physical activity and a Mediterranean diet, also modify the lipid profile of patients taking statins [18]. According to previous evidence, red wine consumption improves HDL and LDL levels in hypercholesterolemic individuals [19]. The benefits of wine on the lipid profile are due to several mechanisms produced by polyphenols, which reduce cholesterol absorption and modify apolipoprotein B secretion rates, very low-density lipoproteins (VLDLs), and triglycerides; therefore, polyphenols could modify the concentration of these parameters at the plasma level [20].

Alcohol consumption has been subject to review over time, as there is controversy. Excessive alcohol consumption is linked to health risks, such as alcohol intoxication, increased risk of injury, impaired judgment and possible long-term consequences for physical and mental health [21]. These risks have led to the implementation of guidelines and consumption limits. The World Health Organization (WHO) recommends avoiding alcohol consumption, although it states that the limit of two drinks per day should not be exceeded for both men and women. More specifically, the WHO suggests not exceeding 20 grams per day for men and 10 grams per day for women [22,23]. The controversy surrounding alcohol consumption is generated especially with wine, owing to the compounds present in wine, which could explain the well-known “French paradox”, where despite high fat and moderate wine consumption, fewer cases of coronary heart disease have been reported [24].

Owing to the importance of lifestyle changes in controlling the lipid profile in the population and the lack of previous evidence of the effect of wine on the lipid profile, there is a lack of information on how it affects the lipid profile according to the type of wine or even the dose‒response relationship. This systematic review and meta-analysis was the first to synthesize previous evidence on the effect of wine consumption on the lipid profile, distinguishing between triglycerides, cholesterol, LDL, HDL and fibrinogen.

2. Methodology

2.1. Search strategy and selection of studies

The guidelines of the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) [25] and the Cochrane Collaboration Handbook [26] were followed in this systematic review and meta-analysis. PROSPERO has already recorded this study under registration number CRD42023396666.

The MEDLINE (via PubMed), Scopus, Cochrane, and Web of Science databases were searched from their inception until March 14, 2023. The PICO structure (population, intervention/exposure, comparison, outcome, and study design) was followed in the search technique, which applied Boolean operators between the following terms: “adults”, “young adults”, “adults population”, “adults subjects”, “older”, “elderly”, “elderly people”, “older people”, “cardiovascular risk markers”, “HDL”, “LDL”, “total cholesterol”, “triglycerides”, “fibrinogen”, “alcohol”, “wine”, “alcohol consumption”, and “wine consumption” (Table S1). In addition, we examined the references from previous systematic reviews and meta-analyses.

2.2. Eligibility

The articles included in this systematic review and meta-analysis were prepost studies and clinical trials measuring the associations between wine consumption and different lipid parameters, such as total cholesterol, triglycerides, HDL, LDL, and fibrinogen. The inclusion criteria were as follows: (i) population over 18 years of age, without restriction as to the presence of pathologies; (ii) exposure: wine consumption; (iii) outcomes: lipid parameters, including triglycerides, HDL, LDL, fibrinogen, and cholesterol, measured with a blood sample; and (iv) study design: prepost studies and clinical trials. In contrast, studies were excluded if they (i) were review studies, ecological studies, editorials, longitudinal studies, or case reports; (ii) were not written in English or Spanish; or (iii) did not report wine consumption separately from other alcoholic beverages.

2.3. Data extraction and quality assessment

The following data were collected from the studies that were included and compiled in an ad hoc table (Table 1): (1) reference: first author and year of publication; (2) country; (3) participant characteristics: sample size, percentage of women, age, and type of population; (4) type of wine: red wine, white wine, or wine; (5) duration of intervention; and (6) outcome: lipid parameters.

Table 1.

Main characteristics of the included studies.

Pre-post studies
Reference	Country	Characteristics of the participants			Type of wine	Duration of intervention	Dose of wine	Outcome
		N, women (%)	Age (SD)	Type of population
Pikaar et al. 1987	Netherlands	12 (0)	21–29	Healthy male students	Red wine	5 weeks	-0 glasses -Two glasses -Four glasses -Binge drinking	Lipid profile
Lavy et al. 1994	Israel	20 (0)	25–45	Healthy young men	Red wine White wine	2 weeks	400 ml of red wine or 400 ml of white wine/day	Triglycerides Cholesterol LDL
Sharpe et al. 1995	Ireland	20 (45)	31.2(9,39)	Healthy volunteers	Red wine White wine	10 days	200 ml <21 units of alcohol/week in males <14 units/week in females (21 g alcohol each day)	Cholesterol Triglycerides HDL LDL
Nigdikar et al. 1998	United Kingdom	30 (0)	35–65	Healthy men	Red wine White wine	2 weeks	375 mL/day	Cholesterol HDL LDL Triglycerides
Cordain et al. 2000	EE.UU.	20 (100)	NR	Sedentary and overweight middle-aged premenopausal women	Red wine	10 weeks	135 ml of red wine 5 days per week	Cholesterol Triglycerides LDL HDL
Sierksma et al. 2004	Netherlands	18 (100)	57(5)	Postmenopausal women	White wine	3 weeks	250 ml/day	Triglycerides Cholesterol HDL
Flechtner-Mors et al. 2004	Germany	GI: 20 (70) GC: 20 (65)	48.1(11.4)	Obese subjects	GI: White wine GC: Grape juice	12 weeks	GI: 200 ml of white wine/day GC: 200 ml of grape juice/day	Cholesterol HDL Triglycerides Fibrinogen
Retterstol et al. 2005	Norway	87 (65)	50.2(9.6)	Healthy individuals who had not smoked tobacco daily for at least 3 months	Red wine	3 weeks	150 ml/day (15 g alcohol)	Fibrinogen Cholesterol HDL Triglycerides
Avellone et al. 2006	Italy	48 (45)	43.6(10.6)	Healthy subjects	Red wine	8 weeks	250 ml/day	Triglycerides Cholesterol HDL LDL Fibrinogen
Rajdl et al. 2007	Czech Republic	42	41.9(9.7)	Healthy subjects	White wine	4 weeks	375 ml/day	Cholesterol HDL Triglycerides Fibrinogen
Bantle et al. 2008	EE.UU.	18 (61.1)	64	Type 2 diabetic	Red wine White wine	4 weeks	120-240 ml/day	Cholesterol LDL Triglycerides
Joosten et al. 2008	Netherlands	36 (100)	56.5(4.2)	Postmenopausal women	White wine	6 weeks	250 ml/day	HDL LDL
Tozzi et al. 2011	Italy	15 (0)	23.13(0.45)	Male healthy volunteers	Red wine	4 weeks	250 ml/day	Cholesterol Triglycerides HDL LDL Fibrinogen
Estruch et al. 2011	Spain	42 (0)	37.6(7.4)	Healthy adult men	Red wine	4 weeks	160 ml/day	Cholesterol Triglycerides LDL HDL
Tatjána et al. 2012	Hungary	18 (16.7)	Men (28–56) Women (31–67)	Metabolic syndrome patients	White wine	4 weeks	Men: 300 ml (30 g) per day Women: 200 ml (20 g) per day	LDL HDL Triglycerides Cholesterol
Apostolidou et al. 2015	Greece	20	50.4 (14.1)	Healthy male and female volunteers were	Red wine	4 weeks	125 ml/day	Cholesterol Triglycerides HDL LDL
Chiu et al. 2016	Taiwan	10	NR	Healthy hypercholesterolemic volunteers	Red wine	10 weeks	250 ml/day	Triglycerides HDL LDL Fibrinogen Cholesterol
Taborsky et al. 2017	Czech Republic	Total: 146 (58.2) Red wine: 74 (56.8) White wine: 72 (59.7)	NR	Healthy subjects	Red wine White wine	48 weeks	NR	HDL Triglycerides LDL Cholesterol
Roth et al. 2019	Spain	38(0)	55–80	High-risk male volunteers	White wine	3 weeks	255 ml/day	Cholesterol
Roth et al. 2019 (2)	Spain	38(0)	55–80	High-risk male volunteers	White wine	3 weeks	255 ml/day	Cholesterol LDL HDL Fibrinogen

Clinical trials
Reference	Country	Characteristics of the participants			Type of wine	Exposure time	Dose of wine	Outcome
		N, women (%)	Age (SD)	Type of population
Gottrand et al. 1999	France	5 (0)	22.8(3.0)	Nonsmoking male subjects	Red wine	4 weeks	50 g/day	Cholesterol Triglycerides HDL
Van de Gaag et al. 1999	Netherlands	11 (0)	51.7 (5.4)	Nonsmoking male subjects	Red wine	12 weeks	40 g of alcohol	HDL
Hansen et al. 2005	Denmark	Red wine group: 19 (52.63) Placebo: 18 (50)	Red wine group: 50 (2) Placebo: 53 (2)	Healthy subjects	Red wine	4 weeks	Males: 300 ml Female: 200 ml	Cholesterol HDL LDL Fibrinogen
Tsang et al. 2005	United Kingdom	Red wine group: 12 Control group: 8	Red wine group: 35 (8.66) Control group: 29 (7.07)	Healthy nonsmokers subjects	Red wine	2 weeks	375 ml	Cholesterol LDL HDL Triglycerides
Naissides et al. 2005	Australia	Red wine group: 14 (100) Water group: 16 (100)	Red wine group: 58.4 (1.3) Water group: 59.3 (1.4)	Hypercholesterolemic postmenopausal women	Red wine	6 weeks	400 ml/day	Cholesterol Triglycerides
Napoli et al. 2005	Italy	Red wine: 9 (88.9) Control group: 8 (87.5)	Red wine: 58 (2) Control group: 53 (3.7)	Type 2 diabetic	Red wine	2 weeks	360 ml/day	Cholesterol Triglycerides
Banini et al. 2006	EE.UU.	Wine group: 10 (0) Control group: 15 (46.7)	Wine group: 60 (15) Control group: 56 (8)	Type 2 diabetic	Wine	4 weeks	150 ml/day	Cholesterol LDL HDL Triglycerides
Marfella et al. 2006	Italy	Red wine: 57 Control group: 58	Red wine: 36.5 (4.6) Control group: 35.1 (5.1)	Type 2 diabetic	Red wine	48 weeks	118 ml/day	Triglycerides HDL
Shai et al. 2007	Israel	Red wine: 72 Control group: 36 (44.03)	41–74	Type 2 diabetic	Wine	12 weeks	150 ml/day	Triglycerides HDL LDL
Kechagias et al. 2011	Sweden	Red wine group: 20 (65) Total abstention: 21 (85)	Red wine group: 32(8) Total abstention: 34(8)	Healthy subjects	Red wine	12 weeks	150 ml/day women 300 ml/day men	Cholesterol LDL HDL Triglycerides
Droste et al. 2013	Luxembourg	Red wine group: 56 (32) No red wine group: 52 (35)	Red wine group: 64.1 (9.1) No red wine group: 63.0 (9.9)	Patients with the presence of plaques or stenosis without hemodynamic compromise	Red wine	20 weeks	100 ml/day women 200 ml/day men	Cholesterol LDL HDL Triglycerides
Mori et al. 2015	Australia	24 (100)	39.3(7.3)	Healthy premenopausal, nonsmoking women	Red wine	4 weeks	Lower level consumers: <200 g alcohol/wk Higher level consumers: >200 g alcohol/wk	Cholesterol LDL HDL Triglycerides Fibrinogen
Mori et al. 2016	Australia	24 (20.83)	59.3(5.6)	Type 2 diabetic men and postmenopausal women	Red wine	4 weeks	Women: 230 ml/day (24 g alcohol/day) Men: 300 ml/day (31 g alcohol/day)	Cholesterol Triglycerides HDL LDL Fibrinogen

Table with the main data extracted from the studies included in the systematic review.

The Cochrane Collaboration’s tool for measuring risk of bias (Rob2) [27] was used to evaluate the quality of the clinical trials. Six domains are used in this tool: selection bias, performance bias, detection bias, attrition bias, reporting bias, and additional biases. If all the domains are classified as “low risk”, the overall bias is considered “low risk of bias”; if at least one domain is classified as “some concern”; and if more than one domain is evaluated as “some concerns”, “high risk”, or both, it is considered “low risk”.

For prepost studies, the risk of bias in nonrandomized studies of interventions (ROBINS-I) tool [28] was used. This tool evaluates bias in seven domains: confounding, selection of the study participants, measurement of interventions, deviations from intended interventions, missing data, measurement of outcomes, and bias in the selection of reported results. If all domains are evaluated as “low risk”, the overall risk of bias is considered to be “low risk”, “moderate risk” if every domain is considered to be “low risk” or “moderate risk”, “serious risk” if at least one domain is classified as such, “critical risk” if at least one domain has been designated “critical risk”, and “no information” if there is no clear indication that the study is at serious or critical risk of bias and if there is a lack of information in one or more key domains of bias.

Two impartial reviewers (ML-LT and CA-B) carried out the study selection, data extraction, and risk of bias evaluation. Conflicts were resolved by consensus or with the intervention of a third researcher (MC.R.-G.)

2.4. Statistical analysis and data synthesis

A meta-analysis was conducted to analyze the associations between wine consumption and lipid profiles. To highlight some considerations, when two studies included data from the same sample, we included the study with the larger sample size in the meta-analysis. DerSimonian and Laird random effects [29] models were used to calculate a pooled standardized effect size (ES) estimate and the respective 95% confidence intervals (CIs) for each lipid parameter. Taking into account the different types of wine consumed in the intervention, the meta-analysis was performed by type of wine.

Inconsistency, which can range from 0% to 100%, was examined via the Cochrane Handbook recommendations [30]. The I² classified inconsistency as unimportant (0–30%), moderate (≥30–50%), substantial (≥50–75%), or considerable (≥75–100%). The corresponding p values were considered. Furthermore, the τ² test was applied to measure heterogeneity and was defined as low when it was less than 0.04, moderate when it ranged from ≥0.04–0.14, and substantial when it varied from 0.14–0.40 [31].

The robustness of the summary estimates was evaluated via a sensitivity analysis that removed each study individually from the pooled values. To determine whether continuous variables such as the mean age, percentage of women, and period of follow-up might influence the associations between wine intake and each lipid parameter, meta-regression studies were carried out. Egger’s regression asymmetry test [32] was used to evaluate publication bias, and a p value of <0.10 was used to establish whether there was significant publication bias.

All the statistical analyses were conducted with STATA SE software, version 15 (StataCorp, College Station, TX, USA).

3. Results

3.1. Systematic review

The search retrieved 1706 articles. Finally, 33 studies [18,19,[33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64]] were included in this systematic review, and 29 studies [18,19,[34], [35], [36], [37], [38], [39], [40], [41], [42],[44], [45], [46], [47], [48], [49],51,[53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63]] were included in the meta-analysis (Fig. 1). Among the included studies, 20 were prepost studies [19,[33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51]], and 13 were clinical trials [18,[52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63]]. Studies were carried out in 18 countries: four in the Netherlands [33,38,44,53], four in Spain [46,50,51], four in Italy [41,45,57,59], three in Australia [56,62,63], three in EEUU [37,43,58], two in the Czech Republic [42,49], two in the United Kingdom [36,55], two in Israel [34,60], and one each in Denmark [54], Norway [40], France [52], Sweden [61], Greece [19], Taiwan [48], Hungary [47], Germany [39], Ireland [35], and Luxemburg [18]. The studies were published between 1987 and 2019 and included a total of 1283 subjects (718 prepost; 565 in clinical trials), with ages ranging from 21.0–80.0 years. The duration of intervention ranged from 10 days [35] to 48 weeks [49]. Concerning lipid parameters, 27 studies reported triglycerides [18,19,[34], [35], [36], [37], [38], [39], [40], [41], [42], [43],[44], [45], [46], [47], [48], [49],52,[55], [56], [57], [58], [59], [60], [61], [62], [63]], 28 studies reported total cholesterol [18,19,[34], [35], [36], [37], [38], [39], [40], [41], [42], [43],[45], [46], [47], [48], [49], [50], [51], [52],[54], [55], [56], [57], [58],61,62,63], 22 studies reported LDL [18,19,[34], [35], [36], [37],41,[43], [44], [45], [46], [47], [48], [49],51,54,55,58,[60], [61], [62], [63]], 27 studies reported HDL [18,[35], [36], [37], [38], [39], [40], [41], [42],[44], [45], [46], [47], [48], [49],[51], [52], [53], [54], [55],[58], [59], [60], [61], [62], [63]], and ten studies reported fibrinogen [[39], [40], [41], [42],45,48,51,54,62,63]. Table 1 summarizes the characteristics of the included studies. Analyses have been conducted to compare the populations included in prepost studies and clinical trials (Table S2).

Fig. 1.

PRISMA 2020 flow diagram for new systematic reviews that included database searches.

Flowchart of the search strategy, 33 studies were finally included in the systematic review.

3.2. Risk of bias assessment

The overall risk of bias in the clinical trials was high risk in nine studies, low risk in two studies, and some concerns in four studies. In relation to the particular domains, in terms of missing outcome data and in the selection of the reported results, 100.0% of the studies were scored as having low bias; in the randomization process and measurement of the outcome, 69.23% were rated as having some concerns; and in the deviations from the intentional interventions domain, most of the studies had a low risk of bias (61.54%). Finally, in the measurement of the outcome, 30.77% of the studies showed high risk, 15.38% of the studies showed some concerns, and 53.85% of the studies showed low risk (Figure S1, Table S3).

The overall risk of bias from prepost studies revealed some low or moderate risk of bias in the included studies. The risk of bias was evaluated in relation to the particular domains as follows: for confounding, 80.95% of the studies were rated as low; for the selection of study participants, 85.71% were rated as low; for the measurement of the intervention domain, 100% were rated as low; and for deviations from intended interventions, 19% were rated as serious. In the remaining three domains, with respect to bias due to missing data, most studies presented a low risk (66.67%); with respect to bias in the measurement of outcomes, the studies presented a moderate risk (52.38%); and with respect to bias in the selection of reported results, most studies presented a low risk of bias (90.48%) (Figure S2, Table S4).

3.3. Meta-analysis

3.3.1. Prepost studies

3.3.1.1. Red wine

Using the DerSimonian and Laird random effects model, the pooled ES for the effect of red wine consumption on triglycerides was 0.05 (95% CI: −0.10, 0.20), that for total cholesterol was −0.10 (95% CI: −0.25, 0.05), that for LDL was −0.29 (95% CI: −0.54, −0.05), that for HDL was 0.23 (95% CI: −0.02, 0.48), and that for fibrinogen was −0.60 (95% CI: −1.43, 0.24). The heterogeneity of these estimates was low, except in the case of HDL, which was moderate, and in the case of fibrinogen, which was substantial (τ²: 0.0000; τ²: 0.0000; τ²: 0.0575; τ²: 0.0742 and τ²: 0.6062, respectively, according to the meta-analysis) (Fig. 2).

Fig. 2.

Meta-analysis of the associations of wine consumption with lipid profiles in prepost studies.

The meta-analysis only shows significant changes in red wine consumption for LDL cholesterol, for triglycerides, cholesterol, LDL, HDL and fibrinogen no significant results are shown.

3.3.1.2. White wine

Using the DerSimonian and Laird random effects model, the pooled ES for the effect of white wine consumption on triglycerides was −0.08 (95% CI: −0.32, 0.16), that for total cholesterol was −0.14 (95% CI: −0.32, 0.04), that for LDL was 0.00 (95% CI: −0.28, 0.29), that for HDL was −0.04 (95% CI: −0.21, 0.13), and that for fibrinogen was 0.06 (95% CI: −0.51, 0.63). The heterogeneity of these estimates was low for all the parameters except LDL, which was moderate, and for fibrinogen, which was substantial (τ²: 0.0376; τ²: 0.0000; τ²: 0.0624; τ²: 0.000 and τ²: 0.1714, respectively, according to the meta-analysis) (Fig. 2).

3.3.2. Clinical trials

Using the DerSimonian and Laird random effect models, the pooled ES for the effect of wine consumption on triglycerides was −0.15 (95% CI: −0.40, 0.11), that for total cholesterol was −0.11 (95% CI: −0.60, 0.38), that for LDL was −0.13 (95% CI: −0.31, 0.06), that for HDL was 0.43 (95% CI: −0.06, 0.92), and that for fibrinogen could not be determined because the number of studies, but the trend of SEs was less than 0. The heterogeneity of these estimates was low in triglycerides and LDL; in the case of total cholesterol and HDL, it was substantial (τ²: 0.1016; τ²: 0.4980; τ²: 0.0000 and τ²: 0.5811, respectively, according to the meta-analysis) (Fig. 3).

Fig. 3.

Meta-analysis of the associations of wine consumption with lipid profiles in clinical trials.

Meta-analysis show no significant changes in wine consumption for triglycerides, cholesterol, LDL, HDL and fibrinogen.

3.4. Sensitivity analysis and meta-regression models

When data from individual studies were removed from the analysis one at a time, the pooled ES estimations for the associations of wine consumption with lipid parameters did not significantly change (in magnitude or direction) (Tables S3 and S4).

The mean age of the participants, percentage of women and duration of intervention according to random effects meta-regression models for the associations between wine consumption and lipid parameters were only associated with study heterogeneity for the duration of intervention for wine consumption and total cholesterol and triglyceride levels in clinical trials (Table S5 and S6).

3.5. Publication bias

3.5.1. Prepost studies

3.5.1.1. Red wine

Evidence of publication bias was not found through Egger’s test for the association between wine consumption and any lipid parameter (triglycerides p = 0.698; total cholesterol p = 0.558; LDL p = 0.561; HDL p = 0.250; fibrinogen p = 0.486).

3.5.1.2. White wine

Evidence of publication bias was not found through Egger’s test for the association between wine consumption and any lipid parameter (triglycerides p = 0.849; total cholesterol p = 0.877; LDL p = 0.165; HDL p = 0.779; fibrinogen p = 0.244).

3.5.2. Clinical trials

Finally, evidence from publication bias was found through Egger’s test for the associations between wine consumption and total cholesterol (p = 0.096) and LDL (p = 0.053) but not for the associations between wine consumption and HDL (p = 0.913) or triglyceride (p = 0.581) levels.

4. Discussion

This study is the first to synthesize previous evidence on the effects of wine consumption and lipid profiles, distinguishing between triglycerides, total cholesterol, LDL, HDL, and fibrinogen. Our results show only a effect of consuming red wine on LDL, with no effect on the other parameters, even though a protective trend can be observed, that is, an improvement in the values of these parameters for wine consumers. In addition, meta-regressions were performed to observe whether the effects of wine on triglycerides, total cholesterol, LDL, HDL, and fibrinogen could be modified by the percentage of women, duration of the intervention, and age; in the case of clinical trials, the duration of the intervention affects the effects of wine on triglycerides and total cholesterol.

An altered lipid profile, such as high LDL levels, damages the endothelial wall, accelerates the development of atherosclerosis [65], and therefore increases the risk of developing a chronic disease [19], is a cardiovascular risk factor for hyperlipidaemia. Although there are pharmacological treatments, other interventions that are simple and accessible to the whole population, such as dietary modification, can help to control or modify these parameters. In the Mediterranean diet, we recommend moderate wine consumption, which, according to previous evidence, seems to influence these parameters and appears to lower total and LDL cholesterol and increase HDL cholesterol levels [18]. A previous meta-analysis in adults with diabetes revealed that moderate wine consumption reduced total cholesterol and diastolic blood pressure [66] but was not related to other lipid profile parameters. Currently, controversy exists over the effect of wine on the lipid profile, and some studies reporting that the benefits of wine on the lipid profile, such as LDL oxidation, are only observed in regular daily consumption [67]; however, other studies show that daily consumption of wine for three weeks improves fibrinogen levels [68].

Cholesterol is the circulating fatty component in plasma that is most associated with atherogenic processes [69], which come mainly from the intake of animal fats or from some medications [68]. When hyperlipidemia occurs, that is, when the lipid profile is elevated, especially LDL cholesterol, the incidence of cardiovascular disease increases [69]. Because the oxidation of LDL cholesterol takes place in the intima layer of the artery, damaging this region and increasing cardiovascular risk [70], evidence shows that wine reduces this process owing to its antioxidant power and that this protective effect is stronger for red wine than for white wine [71]. Previous studies have shown that wine consumption increases HDL levels [48,61,72], which helps maintain an adequate total cholesterol level and prevents imbalances that increase the risk of cardiovascular disease [73]. Our results show that red wine consumption decreases LDL levels; however, for HDL and total cholesterol, no significant associations were found. The results for total cholesterol are similar between the two types of wine, whereas the results for LDL and HDL are better for red wine. In the meta-regressions for age and percentage of women, the effect of wine consumption was not modified, whereas for time of intervention, the effect of wine on total cholesterol parameters was modified, showing better levels of these parameters for longer interventions.

The analyses of this study revealed no significant effect of wine consumption on the levels of triglycerides or fibrinogen. The above mentioned meta-regressions revealed that only the duration of the intervention affected the effect of wine consumption on triglyceride levels, with better levels being obtained over a longer duration. There is previous evidence on the effect of wine consumption on reducing plasma fibrinogen levels [40,74]. It appears that this reduction is produced by the ethanol present in the wine rather than by the nonalcoholic components of the wine [55,74] because a previous study comparing wine and fermented red grape extract revealed a reduction in fibrinogen only in the group that consumed wine [55]. Moreover, ethanol consumption appears to cause an increase in triglyceride levels [75], and controversy exists concerning the effects of wine consumption and triglycerides. Previous studies have reported lower triglyceride levels in wine drinkers [61,66], whereas other studies have reported an increase in the levels of this parameter [53]. This could be explained by differences in the doses and durations of interventions between studies [66], and this explanation is further supported by the meta-regression conducted in our study.

The benefits of wine are due to its antioxidant components, such as polyphenols. Polyphenols are divided into two groups: flavonoids and nonflavonoids [76]. The most important polyphenols present in wine are anthocyanins and catechins of the flavonoid family and resveratrol and tannins of the nonflavonoid family [71,76]. The polyphenols in red wine work by reducing the absorption of cholesterol in the liver and consequently reducing its concentration in the plasma [77]. Polyphenols also reduce triglyceride levels and LDL levels by increasing lipoprotein lipase activity [19]. Procyanidins, flavonoid polyphenols, work to prevent lipid oxidation in the digestive system, which is the main site of action of nutritional antioxidants [78]. Resveratrol is another of the components present in wine that provides the greatest benefits for the organism because it also inhibits LDL oxidation, suppresses platelet aggregation, has antiatherosclerotic properties, promotes vascular relaxation, provides a protective function for the endothelium and regulates various substances, such as nitric oxide synthase, thioredoxin-1, heme oxygenase-1, vascular endothelial growth factor, manganese superoxide dismutase, and caveolin-1, which induce oxidative stress [78,79]. All the benefits produced by the different polyphenols present in wine could explain the famous French paradox, where there is a low mortality from ischemic heart disease despite high fat consumption, which is attributed to wine consumption in this area [80].

Our study is not free of limitations, which must be mentioned. First, this meta-analysis included studies with different designs, and this variability may introduce heterogeneity in the results. However, including these different methodological approaches may also enrich the understanding of the phenomenon under study, providing a broader view and a more complete synthesis of the evidence currently available. Second, the amount of wine used in the intervention programs and the duration of the interventions differed across the studies, making it difficult to compare the results of the different studies and limiting the generalizability of our results. Future studies may standardize the wine dosage to provide a better overview of the effects of wine consumption on the lipid profile. Third, it has not been possible to analyze the effects of different types of wine in clinical trials since this distinction has not been reported in most studies. Fourth, gray literature was not included, which could impact the findings’ comprehensiveness and validity and represent bias. Fifth, more than half of the clinical trials and prepost studies were rated as having high and severe-moderate risk of bias, respectively, due to the randomization process, measurement of the outcome, and deviations from intended interventions. This may influence the conclusions of this study, although some of them are inherent characteristics of the nature of the available studies. Sixth, our results present publication bias detected by Egger's test and may be due to the small number of studies; for some parameters, studies with larger sample sizes or small effect sizes are needed. Seventh, there are confounding variables such as diet, exercise or different lifestyles that could not be considered in these analyses and that would be important for future studies. Eight, data on the effect of wine consumption on triglycerides and fibrinogen were not significantly associated, and no further analyses to explain the heterogeneity or absence of significant effects on these parameters could be explored through further subgroup analysis or meta-regressions. Finally, there is no safe recommendation for alcohol consumption, and although there are moderate consumption limits established by the WHO, i.e., 20 g/day in men and 10 g/day in women, no global consensus exists, which could be a limitation of our study.

This systematic review and meta-analysis revealed that wine consumption has an effect on reducing LDL cholesterol and that wine consumption has no effect on total cholesterol, HDL, triglycerides or fibrinogen. In addition, this work concluded that the duration of the intervention influences the effects of total cholesterol and triglycerides, demonstrating that longer interventions have greater effects on these two parameters. Taking into account the results of this meta-analysis, it would be important to continue investigating this relationship, with more robust clinical trials to determine the most appropriate type and amount of wine to obtain better results. Considering the results of this study, the inclusion of light wine consumption in the diet could be considered, but specific recommendations cannot be made owing to the health risks of inadequate consumption and the effects on vulnerable people, such as polymedicated, pathological or pregnant women. Therefore, despite the results of this study concerning the effect of wine consumption on LDL cholesterol, other alternatives, such as a healthy diet and physical exercise, may be considered for the improvement of these parameters.

CRediT authorship contribution statement

Conceptualization, M.L.-L.-T. and C.A.-B.; Methodology, M.L.-L.-T., C.Á.-B. and MC.R.-G.; software, MC.R.-G. and C.Á.-B. ; Validation, I.C.-R. and C.D.L.; Formal analysis, M.L.-L.-T. and MC.R.-G.; Investigation, M.L.-L.-T and C.A.-B. ; Resources, M.L.-L.-T, C.P.-M, D.G-G and C.D.L.; Data curation, C.A.-B. and D.G.-G.; Writing—Original draft preparation, M.L.-L.-T. and C.A.-B. ; Writing—Review & Editing, I.C.-R.; Visualization, S.P.-J. and C.D.L.; Supervision, C.A.-B. All of the authors revised and approved the final version of the article.

Declaration of Generative AI and AI-assisted technologies in the writing process

None.

Funding

This research was funded by FEDER funds. MLLT is supported by a grant from the University of Castilla-La Mancha (2022-PROD-20657).

Declaration of competing interest

None.

Appendix A. Supplementary data

The following is Supplementary data to this article: