Effect of reducing saturated fat intake on cardiovascular disease in adults: an umbrella review

Introduction Our objective was to explore the effect of the reduction of saturated fat (SAF) intake on cardiovascular disease, mortality and other health-related outcomes in adults. Methods We conducted an umbrella review, searching Medline, Scopus, EMBASE, Cochrane Library, and LILACS databases for systematic reviews from December 1, 2012, to December 1, 2022. We have included meta-analyses of randomized controlled trials (RCTs) and cohort studies. We extracted effect sizes (95%CI), heterogeneity (I2), and evidence quality rating based on the population, intervention, comparator, and outcomes. Results 21 meta-analyses were included (three were from RCTs, and 18 were from cohort studies). Among meta-analyses of RCTs, 15 of the 45 associations were significant. The effect of reduction in SAF intake on combined cardiovascular events (RR 0.79, 95%CI 0.66–0.93) was graded as having moderate certainty of evidence. We found no effect on all-cause mortality, cardiovascular mortality, cancer deaths, and other cardiovascular events. Among meta-analyses of cohort studies, five of the 19 associations were significant. There was an increase in coronary heart disease mortality (HR 1.10, 95% CI 1.01–1.21) and breast cancer mortality (HR 1.51, 95% CI 1.09–2.09) in participants with higher SFA intake compared to reduced SFA. We found no effect on all-cause mortality, cardiovascular mortality, and other cardiovascular events. Conclusion This umbrella review found the reduction in SAF intake probably reduces cardiovascular events and other health outcomes. However, it has little or no effect on cardiovascular mortality and mortality from other causes. More high-quality clinical trials with long-term follow-up are needed. Systematic review registration: CRD42022380859.


Introduction
Cardiovascular disease (CVD) is the leading global cause of death, generating a significant impact on the public health systems of the United States, Europe, and even in low-and middle-income countries (LMCIs), with a secular tendency to increase in recent years.Also, with high annual direct and indirect costs associated with these deaths, including health expenditures and lost productivity (1)(2)(3)(4)(5).The American Heart Association (AHA), in conjunction with the National Institutes of Health (NIH) and other government agencies, provides each year a document named the AHA's Life's Essential 8, which include core health behaviors (smoking, physical activity, diet, and weight) and health factors (cholesterol, blood pressure, and glucose control) that contribute to cardiovascular health (6,7).In this context, improving the nutritional quality of the diet has been recognized as a relevant lifestyle approach to reducing the risk of atherosclerotic cardiovascular disease (ASCVD) (7,8).
On the other hand, most healthcare interventions evaluated in Cochrane Reviews are not supported by high-quality evidence, and harms are under-reported (28).Additionally, some organizations rely on low-quality evidence to formulate recommendations, justifying their decisions as consensus-based guidelines (29).To ensure recommendations based on high-quality evidence, it is essential to develop trustworthy clinical practice guidelines (CPGs) that are informed by a systematic review of evidence and employ a standardized methodology, such as the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach (29,30).
Therefore, this umbrella review aimed to systematically identify meta-analyses of randomized controlled trials (RCTs) and cohort studies investigating the reduction of saturated fat (SAF) intake and its impact on cardiovascular disease, mortality, and other healthrelated outcomes in adults.

Protocol and registration
This study was performed according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (31).The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO), number CRD42022380859.

Search strategy
We searched Medline, Scopus, EMBASE, Cochrane Library, and LILACS database of systematic reviews from December 01, 2012, to December 1, 2022.No language restriction.Our search strategy included Medical Subject Title (MeSH) terms and free-text terms such as "Saturated Fatty Acid, " "Dietary Fats, " "Cardiovascular Diseases, " "Heart Disease Risk Factors, " and "cardiovascular outcomes." We adapted the search algorithms to the requirements of each database.The final search strategy is available as Supplementary Table 1.

Eligibility criteria
Studies were included if they met the following criteria: (1) Population: Systematic reviews that include primary studies in adults (over 18 years), at any risk of cardiovascular disease, with or without cardiovascular disease (but not acutely ill), using or not using lipid-lowering medication; (2) Systematic reviews with metaanalyses of RCTs or observational studies (cohort); (3) Intervention/ comparator: RCTs comparing reduced SFA intake vs. higher SFA intake, and cohort studies comparing categories of low vs. high SFA intake; and (5) Outcomes: studies that reported cardiovascular events and mortality (all-cause mortality, cardiovascular mortality, and cancer deaths) as primary outcomes, and/or other secondary outcomes such as cancer, diabetes, glucose-insulin homeostasis, lipid profile, body weight, blood pressure, and quality of life.We excluded narrative reviews, scoping reviews, meta-analyses of studies with other study designs, comments, editorials, guidelines, and conference abstracts.

Study selection
Duplicate documents were removed with Endnote X20 software.Six independent authors (AA, GD-M, KC-Q, GA-G, CC, and KA-M) selected the articles by titles and abstracts to identify potentially relevant articles.Then, articles were evaluated in full text to assess their eligibility.Any discrepancies were resolved by discussion with the third reviewer (CFL).

Data extraction
Four independent authors (AA, GA-G, KA-M, and CC) extracted the data.Discrepancies were resolved with consensus.We recorded the following variables: author, year of publication, study design, number of participants and included studies, type of intervention/ comparator, outcomes with their effect size with 95% confidence interval (CI), heterogeneity (I 2 ), study follow-up range, and GRADE rating.Another author (GA-G) checked the quality of the data before analysis.
We assessed the overlapping of studies according to the "corrected covered area (CCA)'" for each outcome.CCA >5% was considered as significant overlap (32).In this case, the study result with the highest score was prioritized in a score based on the date of publication, methodological quality, and number of primary studies included.

Quality assessment
Four independent authors (AA, GA-G, CC, and KA-M) independently assessed the quality of the included studies using "A Measurement Tool to Assess Systematic Reviews" (AMSTAR-2), a third author (CFL) settled in case of doubt.This tool consists of 16 items (maximum score: 16. and minimum score: 0).Based on the critical domains, we consider high, moderate, low, and critically low quality in the results (33).

Statistical analysis
We have developed a narrative summary of the data from each included systematic review, including effect estimates with their 95% CI, statistical assessment of heterogeneity (I 2 ), GRADE score (certainty of evidence), and other study characteristics exactly as reported in the included systematic reviews.We performed sensitivity analyses to assess the effect on the outcomes excluded by overlap.

Characteristics of included studies
The three meta-analyses of RCTs (9,34,35) described 45 potential associations of cardiovascular disease and mortality associated with reduction in SAF intake.The number of RCTs were 125, with a sample of 663-56,000 participants and a follow-up duration ranging from 3 to 9 years (additional characteristics in Table 1).
The 18 meta-analyses of cohort studies (17-19, 36-50) described 19 potential associations of cardiovascular disease and mortality associated with reduction in SAF intake.The median number of studies per meta-analysis was 11 (interquartile range, IQR, 7-15), the follow-up duration ranged from 1 to 32 years, and the median sample was 462,268 participants (IQR, 318,747-836,322.5 participants) per meta-analysis (additional characteristics in Supplementary Table 4).

Description and summary of associations 3.4.1 Meta-analyses of RCTs
Fifteen of the 45 associations (33.3%) were statistically significant (p < 0.05) based on random-effects models.The identified associations comprised two types of intervention/comparator, including reduced SFA intake vs. higher SFA intake (51.1%) or replacement of SFA with monounsaturated fatty acids (MUFAs) or polyunsaturated fatty acids (PUFAs) vs. intake without replacement in SFA (48.9%).30 associations analyzed non-repeated outcomes, including nine (30%) cardiovascular and mortality events, six (20%) lipid profile outcomes, nine (30%) glucose-insulin homeostasis outcomes, two (6.7%) blood pressure outcomes, two (6.7%) body weight outcomes, one outcome (3.3%) of quality of life, and another (3.3%) with a diagnosis of cancer.Also, 10 of the 24 associations (41.7%) had heterogeneity (I 2 ) > 50%, and eight (17.8%) of the 45 associations assessed the certainty of evidence using GRADE (three associations were supported by moderate certainty of evidence, others three were supported by very low, and two associations by low certainty of evidence).Summary of all associations in Table 2 and Supplementary Table 6.

Meta-analyses of cohort studies
Five of the 19 associations (26.3%) were statistically significant (p < 0.05) based on random-effects models.The identified associations comprised one type of intervention/comparator: higher vs. reduced SFA intake.18 associations analyzed non-repeated outcomes, including 12 (66.7%)cardiovascular and mortality events, five (27.8%) with a diagnosis of cancer, and one (5.5%)with a diagnosis of diabetes.Additionally, seven of the 17 associations (41.2%) had heterogeneity (I 2 ) > 50%, and two (10.5%) of the 19 associations assessed the strength of evidence using GRADE (both were supported by very low certainty).Summary of all associations in Table 3 and Supplementary Table 7.

Findings of outcomes 3.5.1 Meta-analyses of RCTs
There was a 21% reduction in combined cardiovascular events in people who had reduced SFA compared with those on higher SFA intake (RR 0.79, 95%CI 0.66-0.93,I 2 = 65%, 11 RCTs) (moderate certainty of evidence, GRADE) (9).We found no effect on all-cause mortality, cardiovascular mortality, cancer deaths, and other cardiovascular events such as myocardial infarction, coronary heart disease events, and stroke (moderate, low, and very low certainty of evidence, GRADE) (summary of the studies in Table 2 and Figure 2; and details of the GRADE assessment in Supplementary Table 8).
Among the two associations supported by very low certainty of evidence (GRADE) (38), we found no effect on coronary heart disease (follow-up range: 1-20 years) and ischemic stroke (follow-up range: 7.6-32 years) in participants with higher SFA intake compared to reduced SFA (details of the GRADE assessment in Supplementary Table 9).We also found no effect on all-cause mortality, cardiovascular disease mortality, cancer mortality, and others cardiovascular events as cardiovascular disease, intracerebral hemorrhage, and subarachnoid hemorrhage, with a follow-up range of 1.3-32 years (low and critically low quality, AMSTAR-2) (Table 3).On the other hand, there was a reduction in fatal stroke (RR 0.75, 95% CI 0.59-0.94,I 2 = 0, 4 cohort studies) (critically low quality, AMSTAR-2) (49) and stroke (RR 0.89, 95% CI 0.82-0.96,I 2 = 37.4, 15 cohort studies) (critically low quality, AMSTAR-2) (49) in participants with higher SFA intake compared to reduce SFA (summary of the primary outcomes in Table 3 and Figure 3).About secondary outcomes, there was an increase in liver cancer (RR 1.34, 95% CI 1.06-1.69,I 2 = 16.9, 5 cohort studies) (critically low quality, AMSTAR-2) (36) in participants with higher SFA intake compared to reduce SFA (Supplementary Table 7).
In all significant outcomes, adjustment for confounding variables was performed.Summary of all significant and nonsignificant associations in Table 3 and Supplementary Table 7.

Sensitivity analyses
We have conducted sensitivity analyses, taking into account the potential impact of excluding outcomes due to overlap in cohort studies.We found similar results about mortality (there is no effect on all-cause mortality and cardiovascular disease mortality).In the context of stroke, participants with higher SFA intake experienced a reduction in stroke events compared to those with lower SFA intake.However, differences in the results were observed.While there is no effect on stroke mortality, there was a reduction in events related to stroke subtypes (ischemic and hemorrhagic) (Supplementary Table 10).

Discussion
Our findings indicate that the effect of reduction in SAF intake probably reduces cardiovascular events and other health outcomes.
However, it has little or no effect on cardiovascular mortality and mortality from other causes.Additionally, we observed a reduction in lipid profile (total cholesterol and LDL-C), body weight, BMI, and an improvement in glucose-insulin homeostasis.Moreover, it could enhance the quality of life and reduce the risk of liver cancer.Finally, participants with higher SFA intake, compared to those with reduced SFA intake, may experience a decrease in fatal stroke and stroke events, as suggested by some observational studies.
In our study, we did not observe differences in mortality in RCTs, including both cardiovascular and other causes.However, we found a little effect in observational studies with wide in confidence intervals, such as Cheng et al. (fatal stroke) (49) and Brennan et al. (breast cancer mortality) (50).This suggests the mortality could occur with considerable variability in other countries or contexts.This could be due to the infrequency of the outcome and the small number of studies included, despite having a large study sample.Furthermore, discrepancies among studies could be attributable to the different biological effects produced by various types of saturated fatty acids, influenced by factors such as the food matrix and dietary carbohydrate content.Individual and methodology factors, including age, sex, adiposity levels, and the shorter follow-up time in RCTs, may also contribute to these variations (24-26, 51, 52).Consistent with our findings, the current recommendations from ACC, AHA, and ESC do not justify their decisions based on mortality results but instead aim to reduce the risk of ASCVD (5,10,11).
Regarding the risk of ASCVD, the CPGs recommendations are derived from a combination of observational studies and data from RCTs (5,10,11).However, our findings reveal heterogeneity results.In RCTs, we found significant differences observed when metaanalyzed and when creating a composite outcome that groups various types of cardiovascular events, without differences when metaanalyzed by outcome (9).ACC, AHA, and ESC (5,10,11) have taken   a conservative approach and decided to recommend replacing SAF by PUFAs, principally.This decision was likely made, emphasizing that even a small percentage reduction in cardiovascular-related health outcomes can substantially decrease the number of people developing CVD, both nationally and globally, along with the associated healthcare costs (5).On the other hand, we observed a reduction in lipid profile, body weight, BMI, and an improvement in glucoseinsulin homeostasis.This is compatible with the majority of published data about that (9,34,53,54), and it could be another reason to justify the reduction of SAF intake.Additionally, it could enhance the quality of life and reduce risk of liver cancer, but it is necessary to have more studies to confirm it.Based on these issues, rather than having an universal recommendation, practitioners should give personalized recommendations, taking into account factors such as the habitual dietary patterns of individuals, nutritional status, income level, comorbidities, physical activity, and country-level nutrition data.
It is important to avoid recommendations based on low or very low-quality studies.This is crucial to prevent discordant recommendations, avoid harm to patient care, discourage future RCTs, minimize confusion and frustration among practitioners, and manage health system resources effectively, especially in LMCIs (30, 55).Furthermore, we should assess the benefits of interventions based on critical and important outcomes, avoiding reliance on surrogate measures (56).
In relation to stroke, observational studies suggest a decrease in both fatal stroke and overall stroke events with higher SFA intake.However, there is significant variability among the other studies (13,17,41,45,49), indicating a complex relationship and highlighting the need for further research to fully understand the underlying mechanisms.Another interesting and dual behavior can be observed with highdensity lipoprotein cholesterol (HDL-C) levels.Classically, it is known that HDL-C is inversely associated with CVD risk (11).However, some studies report that very high levels of HDL-C may increase CVD risk and mortality (57,58).Further clarification is needed in future studies.

Limitations and strengths
This study has limitations that are important to mention.First of all, due to the study design (umbrella review, where the unit of searching and data analysis is the systematic review rather than the primary study) (59), our intention was to provide a broad overview of the impact of SAF intake on cardiovascular disease.Our aim was not to evaluate this effect on an individual level, nor did we intend to assess all primary studies included in each meta-analysis.Instead, our focus has been on analyzing the methodology and findings of each systematic review, while acknowledging the inherent limitations in this approach.Secondly, despite conducting a systematic review, we were unable to make recommendations comparable to CPGs.High-quality evidence is the cornerstone of assessing the benefits and harms of an intervention.To maximize the trustworthiness of recommendations within the context of CPGs, they should be rigorously and transparently developed using a standardized methodology.This process should take into account expert opinions, as well as considerations of equity, resource utilization, acceptability, and feasibility (59).Third, we could not re-analyze the outcome data of the systematic reviews, as it was not an objective in our study protocol.Instead, we presented the outcome data exactly as they appear in the included systematic reviews.We believe that this overview format is the most appropriate and a feasible way to address our research question.Fourth, a minority of authors reported evaluations of the certainty of the evidence using the GRADE approach (17.8 and 10.5% of associations in meta-analyses of RCTs and observational studies).However, we assessed the quality of all systematic reviews included using the AMSTAR-2 tool.Finally, the maximum follow-up duration reported in RCTs was 9 years.To address concerns about the potential lack of time to obtain mortality outcomes, it may be necessary for RCTs to have a longer duration.On the other hand, systematic reviews of observational studies reported a maximum follow-up of 32 years.While differences in mortality were found in observational studies, these studies showed a small effect, imprecision, heterogeneity, and a high risk of bias.For all these reasons, our results are exploratory, and should be interpreted with caution.
The strength of this study includes a systematic and exhaustive search of the literature, inclusion of a large body of evidence, and the incorporation of systematic reviews of both RCTs and observational studies.Also, this study stands out as the first umbrella review that focuses on SAF intake's impact on cardiovascular outcomes, considering data from both RCTs and observational studies.

Conclusion
This umbrella review found the reduction in SAF intake probably reduces cardiovascular events and other health outcomes.However, it has little or no effect on cardiovascular mortality and mortality from other causes.A healthy diet and physical activity remain the cornerstones of CVD prevention in all individuals.However, recommendations should be individualized considering factors such as nutritional status, comorbidities, and income level.Additionally, high-quality clinical trials with long-term follow-up are needed to investigate the effects of reduced SAF intake on cardiovascular-related health outcomes.

FIGURE 1 Flow
FIGURE 1Flow chart of study selection.

FIGURE 2
FIGURE 2The effects of reduced intake of saturated fats as reported in meta-analyses of RCTs.CI, Confidence intervals; I 2 , Statistic assessment of heterogeneity; GRADE, Grading of recommendations, assessment, development, and evaluations; AMSTAR-2, A Measurement Tool to Assess Systematic Reviews; RR, Relative risk; MI, Myocardial infarction; CHD, Coronary heart disease; NR, No reported; and SFA, Saturated fatty acids.

FIGURE 3
FIGURE 3The effects of increased intake of saturated fats as reported in cohort meta-analyses.CI, Confidence intervals; I 2 , Statistic assessment of heterogeneity; GRADE, Grading of recommendations, assessment, development, and evaluations; AMSTAR-2, A Measurement Tool to Assess Systematic Reviews; HR, Hazard ratio; RR, Relative risk; CVD, Cardiovascular disease; CHD, Coronary heart disease; ICH, Intracerebral hemorrhage; SAH, Subarachnoid hemorrhage; NR, No reported; and SFA, Saturated fatty acids.

TABLE 1
Characteristics of meta-analyses of randomized clinical trials studying saturated fat intake.
No, Number; AMSTAR-2, A Measurement Tool to Assess Systematic Reviews; SFA, Saturated fatty acids; MUFA, Monounsaturated fat; PUFA, Polyunsaturated fat. a 1 = All-cause mortality, 2 = Cardiovascular mortality, 3 = Coronary heart disease mortality, 4 = Cancer deaths, 5 = Combined cardiovascular events, 6 = Myocardial infarction, 7 = Non-Fatal myocardial infarction, 8 = Coronary heart disease events, 9 = Stroke, 10 = Cancer diagnoses, 11 = Glucose-insulin homeostasis, 12 = Lipidic profile, 13 = Body weight, 14 = Blood Pressure, and 15 = Quality of life.b By suggesting appropriate nutrient based or food-based aims, or which provided a general dietary aim, such as improving heart health or reducing total fat.The intervention had to be dietary advice, supplementation of fats, oils or modified or low-fat foods, or a provided diet.c Which could be a diet high in saturated fat, or a usual diet (not modified in SFA).d Range of mean years in trial.

TABLE 2
Summary of primary findings of meta-analyses of randomized clinical trials studying saturated fat intake.Relative risk.NR, Not reported.a Range of mean years in trial.b Adults (18 years or older, no upper age limit) at any risk of cardiovascular disease, with or without existing cardiovascular disease, using or not using lipid-lower in medication.Participants could be of either gender, but we excluded those who were acutely ill, pregnant or lactating.c By suggesting appropriate nutrient based or food-based aims, or which provided a general dietary aim, such as improving heart health or reducing total fat.The intervention had to be dietary advice, supplementation of fats, oils or modified or low-fat foods, or a provided diet.d Which could be a diet high in saturated fat, or a usual diet (not modified in SFA).
e These included people experiencing any of the following: cardiovascular death, cardiovascular morbidity (non-fatal myocardial infarction, angina, stroke, heart failure, peripheral vascular events, and atrial fibrillation) and unplanned cardiovascular interventions (coronary artery bypass surgery or angioplasty).

TABLE 3
Summary of primary findings of meta-analyses of cohorts studying saturated fat intake.Confidence intervals; I 2 , Statistic assessment of heterogeneity; %, Percentage; AMSTAR-2, A Measurement Tool to Assess Systematic Reviews; GRADE, Grading of recommendations, assessment, development, and evaluations; SFA, Saturated fatty acids; gr, Gramos; NR, No reported; HR, Hazard ratio; RR, Relative risk.aAdjustment for confounding variables: Age, race, sex, education, marital status, poverty to income ratio, physical activity, SBP, smoking, serum cholesterol, alcohol, BMI, ancestry, left ventricular hypertrophy, presence of initial malignant disease, energy intake, serum lipids, systolic blood pressure, glucose intolerance, energy, physical activity, history of hypertension, family history of myocardial infarction < 60 years, profession, dietary fiber, social class, education, frequency of exercise, dietary supplements, diabetes, HDL, LDL, triacylglycerol, hypertension, examination year, % energy from protein, other fatty acids, years of schooling, and waist to hip ratio.bAdjustment for confounding variables: Age, sex, blood pressure, total cholesterol, smoking, diabetes, hypertension, coronary heart disease, Quetelet index, radiation dose, alcohol, mental stress, physical activity, walking, sports, education, total energy intake, animal protein, polyunsaturated fatty acids, vegetables, and fruits.Adjustment for confounding variables: Age at menarche, Quetelet index, total energy intake, age at diagnosis, smoking, body weight, dietary factors, physical activity, BMI, weight change, reproductive factors, treatment, and breast cancer stage at diagnosis.d Adjustment for confounding variables: Age, sex, race, marital status, income, socio-economic status, education, Quetelet index, mental stress, alcohol, smoking, physical activity, walking, sports, menopausal status, hormone use, aspirin use, other medications, multivitamins, radiation dose, blood pressure, total cholesterol, diabetes, hypertension, coronary heart disease, left ventricular hypertrophy, atrial fibrillation, family history of myocardial infarction, total energy intake, fiber, potassium, vitamin E, C and calcium, animal protein, polyunsaturated fatty acids, vegetables, and fruits. c