Egg and Dietary Cholesterol Intake and Risk of All-Cause, Cardiovascular, and Cancer Mortality: A Systematic Review and Dose-Response Meta-Analysis of Prospective Cohort Studies

• Source: Egg and Dietary Cholesterol Intake and Risk of All-Cause, Cardiovascular, and Cancer Mortality: A Systematic Review and Dose-Response Meta-Analysis of Prospective Cohort StudiesPublisher: Frontiers | Author: Manije Darooghegi Mofrad, Sina Naghshi, Keyhan Lotfi, Joseph Beyene, Elina Hypponen, Aliyar Pirouzi, Omid SadeghiArchived: July 1, 2026
• Methods
  • Search Strategy
  • Inclusion Criteria
  • Exclusion Criteria
  • Data Extraction
  • Risk of Bias Assessment
  • Statistical Methods
• Results
  • Characteristics of the Included Studies
  • Findings From the Meta-Analysis on All-Cause Mortality
  • Findings From the Meta-Analysis on CVD Mortality
  • Findings From the Meta-Analysis on Cancer Mortality
  • Subgroup Analyses, Meta-Regression, Sensitivity Analyses, and Publication Bias

§Source: Egg and Dietary Cholesterol Intake and Risk of All-Cause, Cardiovascular, and Cancer Mortality: A Systematic Review and Dose-Response Meta-Analysis of Prospective Cohort Studies
Publisher: Frontiers | Author: Manije Darooghegi Mofrad, Sina Naghshi, Keyhan Lotfi, Joseph Beyene, Elina Hypponen, Aliyar Pirouzi, Omid Sadeghi
Archived: July 1, 2026

§Methods

This systematic review and meta-analysis were conducted in accordance with the Meta-analyses of Observational Studies in Epidemiology (MOOSE) guidelines ().

§Search Strategy

We searched online databases, including PubMed, Scopus, ISI Web of Science, Embase, and Google Scholar, up to April 2021 (Supplementary Table 1). The search was done without applying any filters, including publication date or the language of articles. In addition, the reference lists of the selected articles and recent reviews were cross-checked to identify any articles that may have been missed.

§Inclusion Criteria

Studies were selected if they (1) used a prospective observational design; (2) were conducted on an adult population (≥18 years); and (3) reported relative effect estimates, including risk ratio (RR), hazard ratio (HR), and odds ratio (OR), with 95% confidence intervals (CI), to determine the associations of egg and dietary cholesterol intake with the risk of all-cause, cancer, or CVD mortality. If the results from one study were published in more than one article, we selected the most recent one; otherwise, the one with the greatest number of cases or with the highest quality was included.

§Exclusion Criteria

We excluded studies if they (1) were conducted on subjects who had CVD or cancer at baseline; (2) were letters, abstracts, unpublished studies, reviews, comments, ecological studies, or meta-analyses; and (3) had insufficient data for systematic review and meta-analysis.

§Data Extraction

Two investigators (MDM and SN) screened and extracted data independently, and another author (KL) checked them for accuracy. The following items were extracted from each eligible study: first author’s name, study location, gender, age, sample size, number of cases, study period, categories of egg and dietary cholesterol intake, methods used for dietary evaluation, relative risks, and 95% confidence intervals, and variables adjusted for in the analysis. If an included study reported several effect estimates, the one that was controlled for the most confounding variables was used for the meta-analysis.

§Risk of Bias Assessment

To assess the risk of bias among included studies, we used the risk of bias in non-randomized studies of exposures tool. This tool comprises seven domains through which bias might be introduced. The questions in these domains include bias due to confounding, bias in the selection of participants in the study, bias in the classification of exposures, bias due to departure from intended exposures, bias due to missing data, bias in the measurement of outcomes, and bias in the selection of reported results. Under each domain, we categorized studies as having a low, moderate, serious, or critical risk of bias.

§Statistical Methods

We included the RRs (and 95% CIs) of mortality for the comparison between the highest and lowest intakes of eggs and dietary cholesterol in the meta-analysis. The risk estimates were combined using a random-effects model. If a study reported subgroup risk estimates stratified by gender or any other variables, we first pooled the subgroup estimates using a fixed-effects model, and then the pooled risk estimate was included in the main meta-analysis. In addition, if a study presented the RRs for deaths due to different types of CVDs or cancers, not for overall outcomes, we first pooled the RRs using a fixed-effects model, and then the pooled RR of CVD or cancer mortality was included in the meta-analysis.

Cochran’s Q test and the I2 statistic were applied to evaluate heterogeneity among studies. Subgroup and meta-regression analyses were conducted to detect possible sources of heterogeneity. Publication bias was assessed using Egger’s linear regression test (). In the case of substantial publication bias, the trim-and-fill method was applied to detect the effect of probable missing studies on the overall RR (). To assess the dependency of the overall RR on one study, sensitivity analysis was conducted using a random-effects model.

A linear dose–response analysis was conducted using the generalized least squares trend estimation method, described by Greenland and Longnecker () and Orsini et al. (). Estimated study-specific slopes were combined using a random-effects model to provide an overall average slope. In the non-linear dose–response analysis, exposures were modeled using restricted cubic splines with three knots at percentiles of 10, 50, and 90% of the distribution. The correlation within each set of provided risk estimates was considered, and the study-specific estimates were combined using a one-stage linear mixed-effects meta-analysis. The significance level for non-linearity was assessed by testing whether or not the coefficient of the second spline was equal to zero. All analyses were done using STATA version 16.0. P < 0.05 was assumed to be statistically significant for all tests.

§Results

In total, we identified 6,503 articles in our initial search. After excluding duplicate articles and studies that did not meet the inclusion criteria, there were 67 potentially relevant publications (Figure 1). After full-text reviews, we excluded two articles owing to enrolling cancer patients (, ). Six publications were excluded because they did not report eligible risk estimates (–). Three articles reported risk estimates for egg consumption substituted for an iso-energetic amount of other foods and thus were excluded (–). One study was excluded because it contained risk estimates for the intake of eggs combined with other protein sources (). We also excluded one abstract without the required data for a meta-analysis (). In addition, we found studies with significant participant overlap, including articles from the National Health and Nutrition Examination Survey (NHANES) (, ), the Health Professionals Follow-Up Study (HPFS) (, ), and the China Health and Nutrition Survey (CHNS) (, ). Since these studies reported risk estimates for similar exposure and outcome variables, we included only the one with higher quality or with the highest number of cases and excluded the duplicate publications (, , ). The articles by Sun et al. () and Chen et al. () used data from the Women’s Health Initiative study, with more complete data presented by Sun et al. However, the study by Chen et al. was included because it presented the RR in relation to both dietary cholesterol and egg consumption. Dehghan et al. () used three different datasets; two with CVD patients and one with the general population. Therefore, we included the risk estimates reported for the general population and excluded those related to patients with vascular diseases. In total, 55 studies (51 publications) remained for the final meta-analysis (, –, –, , –), of which 25 studies examined the association of egg consumption with all-cause mortality (, , , –, –, , , , , , , –, , ), 21 with CVD mortality (, –, , , , , , , , , , , , , , ), and 19 with cancer mortality (, , , , , , , , , , –, , , , –). In terms of dietary cholesterol intake, 11 publications reported risk estimates for all-cause mortality (, , , , , , , , , , ), 10 for CVD mortality (, , , , , , , , , ), and 3 for cancer mortality (, , ).

Figure 1

§Characteristics of the Included Studies

Supplementary Tables 2–4 show the general characteristics of studies included in the current systematic review and meta-analysis. The number of participants in these studies ranged from 162 to 521,120 individuals with an age range between 15 and 103 years. In total, 2,772,486 participants were included in the 55 studies we considered. During the follow-up periods ranging from 5 to 32 years, 228,425 participants died from all causes, 71,745 from CVD, and 67,211 from cancer. Also, nine included only men (, , , , , , , , ), six included only women (, , , , , ), and five publications reported risk estimates for men and women separately (, , , , ). A total of 13 studies were from the United States (, , , , , –, , , , , , ), 12 were from Europe (, , , , , , , , , , , , ), 17 were from Asia (–, , , , , , , , , , , , , , , ), and four publications (13 studies) describe studies that recruited populations from more than one country (, , , ). Dietary intakes were assessed using food frequency questionnaires in 43 publications (, –, , , –, –, , –, –, , –, –, –), 6 applied to food recall or record (, , , , , ), and 3 used dietary history (, , ). Based on the ROBINS-E tool, 24 studies (25%) were rated as having a serious risk of bias and 27 studies (75%) a moderate risk of bias (Supplementary Table 5).

§Findings From the Meta-Analysis on All-Cause Mortality

§Egg Consumption and All-Cause Mortality

A total of 27 studies (22 publications) (, , , –, –, , , , , , , , ) examined the association between egg consumption and risk of all-cause mortality by comparing the highest and lowest egg intakes. In these studies, the median egg intake classified as “high” was 1.07 eggs/day (IQR: 0.73–1.24) and 0.03 egg/day (IQR: 0–0.07) as “low,” with one study not providing quantification (). These studies included a total of 1,153,367 participants and recorded 226,990 deaths from all causes. Combining RRs from these studies did not provide evidence for an association between egg consumption and all-cause mortality (Pooled RR: 1.03, 95% CI: 0.97–1.09, I2 = 85.3%, Pheterogeneity < 0.001) (Table 1; Supplementary Figure 1).

Table 1

Summary risk estimates for the associations between egg and cholesterol intake and risk of all-cause mortality in adults aged ≥18 yearsa.

a

BMI, body mass index; CI, confidence interval; RR, relative risk; DHQ, dietary history questionnaire; FFQ, food frequency questionnaire; US, United States.

b

Number of risk estimates.

c

Obtained from the random-effects model.

d

Inconsistency, the percentage of variation across studies due to heterogeneity.

e

Obtained from the Q-test.

A total of 28 studies (23 publications) (, , , –, –, , , , , , , –, ) were identified for inclusion in the dose–response analysis. Each additional egg per day was associated with a 7% higher risk of all-cause mortality (Pooled RR: 1.07, 95% CI: 1.02–1.12, I2 = 84.8%, Pheterogeneity < 0.001) (Table 1; Supplementary Figure 2). There was evidence for a non-linear association (Pnon-linearity = 0.003); egg consumption from the lowest amount to one egg per day had no association with all-cause mortality, while the consumption of more than 1 egg per day (~1.5 eggs/day) was associated with an increase in all-cause mortality (Figure 2; Supplementary Table 6).

Figure 2

§Dietary Cholesterol Intake and All-Cause Mortality

A total of 14 cohort studies (9 publications) (, , , , , , , , ), including a total of 172,147 deaths among 852,279 participants, investigated the association between high vs. low cholesterol intake and all-cause mortality. In these studies, the median for high cholesterol intake was 419 mg/day (IQR: 330–453) and the median for low cholesterol intake was 114 mg/day (IQR: 80–147). However, two studies provided no quantification (, ). The summary RR was 1.07 (95% confidence interval: 1.02–1.13, I2 = 47.8%, Pheterogeneity = 0.05) (Table 1; Supplementary Figure 3).

In the dose–response meta-analysis, including 14 studies (, , , , , , , , ), we found that a 100 mg/day increase in dietary cholesterol intake was associated with a 6% greater all-cause mortality (pooled RR: 1.06, 95% CI: 1.03–1.08, I2 = 34.5%, Pheterogeneity = 0.16) (Table 1; Supplementary Figure 4). There was evidence of a non-linear association (Pnon-linearity < 0.001) so that the risk of all-cause mortality increased from 450 mg/day dietary cholesterol to the higher amounts (Figure 2; Supplementary Table 7).

§Findings From the Meta-Analysis on CVD Mortality

§Egg Consumption and CVD Mortality

We included 16 cohort studies (, –, , , , , , , , , , ) with 1,479,181 participants and 69,325 cases of deaths in the analysis of the highest vs. lowest egg consumption and CVD mortality. In this analysis, the highest and lowest intakes of eggs were defined as median intakes of 1 egg/day (IQR: 0.7–1.24) and 0.007 egg/day (IQR: 0–0.06), respectively. The summary RR for CVD mortality in relation to egg consumption was 1.01 (95% CI: 0.90–1.13, I2 = 83.1%; Pheterogeneity < 0.001) (Table 2; Supplementary Figure 5).

Table 2

Summary risk estimates for the associations between egg and cholesterol intake and risk of CVD mortality in adults aged ≥18 yearsa.

a

BMI, body mass index; CI, confidence interval; RR, relative risk; DHQ, dietary history questionnaire; FFQ, food frequency questionnaire; US, United States.

b

Number of risk estimates.

c

Obtained from the random-effects model.

d

Inconsistency, the percentage of variation across studies due to heterogeneity.

e

Obtained from the Q-test.

A total of 26 cohort studies (21 publications) (, –, , , , , , , , , , , , , , ) were included in the dose–response analysis. The summary RR for CVD mortality based on 1 egg/day increase was 1.00 (95% CI: 0.92–1.09, I2 = 81.5%, Pheterogeneity < 0.001) (Table 2; Supplementary Figure 6). There was no evidence of a non-linear association (Pnon-linearity = 0.43) (Figure 2; Supplementary Table 6).

§Dietary Cholesterol Intake and CVD Mortality

A total of nine prospective studies (, , , , , , , , ) were included in the analysis of the highest (median: 420 mg/day, IQR: 348–580) vs. lowest (median: 149 mg/day, IQR: 111–187) intake of dietary cholesterol and CVD mortality. These studies included 55,595 deaths among 875,561 participants. Combining data from these studies indicated no significant association between dietary cholesterol intake and CVD mortality (Pooled RR: 1.09, 95% CI: 0.96–1.24, I2 = 64.1%, Pheterogeneity = 0.004) (Table 2; Supplementary Figure 7).

In the dose–response meta-analysis based on 14 studies (nine publications) (, , , , , , , , ), we found no association between a 100 mg/day increase in cholesterol intake and CVD mortality (Pooled RR: 1.04, 95% CI: 0.99–1.10, I2 = 85.9%; Pheterogeneity < 0.001) (Table 2; Supplementary Figure 8). There was statistical evidence of a non-linear association (Pnon-linearity = 0.009) (Figure 2; Supplementary Table 7). Accordingly, the risk of CVD mortality began to increase from 400 mg/day dietary cholesterol to higher amounts. However, the risk among these amounts was not statistically significant.

§Findings From the Meta-Analysis on Cancer Mortality

§Egg Consumption and Cancer Mortality

The association between high vs. low egg intake and cancer mortality was examined in 24 studies (18 publications) (, , , , , , , , , , –, , , , , ) with a total of 1,705,280 participants and 65,261 cancer deaths. In these studies, the median egg intake classified as “high” was 0.81 egg/day (IQR: 0.64–1.00) and 0.04 egg/day (IQR: 0–0.14) as “low,” with one study not providing quantification (). After combining the results of these studies, a significant positive association was observed (Pooled RR: 1.23, 95% CI: 1.05–1.45, I2 = 95%, Pheterogeneity < 0.001) (Table 3; Supplementary Figure 9).

Table 3

Summary risk estimates for the associations between egg and cholesterol intake and risk of cancer mortality in adults aged ≥18 yearsa.

a

BMI, body mass index; CI, confidence interval; RR, relative risk; DHQ, dietary history questionnaire; FFQ, food frequency questionnaire; US, United States.

b

Number of risk estimates.

c

Obtained from the random-effects model.

d

Inconsistency, the percentage of variation across studies due to heterogeneity.

e

Obtained from the Q-test.

A total of 25 cohort studies (18 publications) (, , , , , , , , , –, , , , , , ) were included in the dose–response analysis. The summary RR of CVD mortality based on an increase of 1 egg/day was 1.13 (95% CI: 1.06–1.20, I2 = 54.2%, Pheterogeneity = 0.004) (Table 3; Supplementary Figure 10). There was no evidence of departure from linearity (Pnon-linearity = 0.51) (Figure 2; Supplementary Table 6).

§Dietary Cholesterol Intake and Cancer Mortality

A total of four prospective studies (three publications) (, , ), including 800,622 participants and 53,540 cases of cancer mortality, were identified and included in the analysis of the high (median: 330, IQR 308–480) vs. low (median: 118, IQR: 109–210) intake of dietary cholesterol and cancer mortality. Combining data from these studies indicated a positive association between cholesterol intake and cancer mortality (pooled RR: 1.13, 95% CI: 1.01–1.25, I2 = 69.7%, Pheterogeneity = 0.04) (Table 3; Supplementary Figure 11).

In the dose–response meta-analysis of the four studies, we found that a 100 mg/day increase in cholesterol intake was associated with a 6% higher risk of cancer mortality (Pooled RR: 1.06, 95% CI: 1.05–1.07, I2 = 0%, Pheterogeneity = 0.95) (Table 3; Supplementary Figure 12). There was no evidence of a non-linear association (Pnon-linearity = 0.28) (Figure 2; Supplementary Table 7).

§Subgroup Analyses, Meta-Regression, Sensitivity Analyses, and Publication Bias

Tables 1–3 show findings from different subgroup and meta-regression analyses. A significant positive association was observed between egg consumption and the risk of all-cause and CVD mortality among women, and in studies conducted in the United Studies. For cancer mortality, a significant positive association was seen in studies conducted in the United States, those with a follow-up duration of ≥15 years, those that applied FFQ for dietary assessment, and studies that did not control for energy intake and lipid-lowering medication. According to the meta-regression, study location and dietary assessment methods for all-cause mortality, energy adjustment for CVD mortality, study location, follow-up duration, and energy adjustment for cancer mortality appeared to be the main sources of heterogeneity.

For cholesterol intake and all-cause CVD and cancer mortality, a significant positive association was seen in studies conducted in the United States, those with a follow-up duration of ≥15 years, and those that used FFQ for dietary assessment. Such a significant positive association was also seen for all-cause mortality among studies that controlled their analysis for energy intake and for CVD mortality among studies that did not control this variable. Based on the meta-regression, study location appeared to be the main factor responsible for the heterogeneity observed in the association between dietary cholesterol intake and all-cause mortality.

In the sensitivity analyses that excluded one study at a time from the meta-analysis, the pooled RRs were not substantially altered. Assessment of publication bias using Egger’s linear regression test found no evidence for small-study effects, except for cholesterol intake and all-cause mortality. There, the application of the trim-and-fill method did not change the average effect size \[(1.07, 0.95% CI: 1.02–1.13) vs. (1.07, 0.95% CI: 1.02–1.13)\].