Diet Quality and Risk of SARS-CoV-2 Infection or COVID-19: A Systematic Review of Observational Studies

The COVID-19 pandemic highlighted the importance of healthy diets in the management of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and COVID-19. Evidence suggests the influence of diet and dietary patterns during post–COVID-19, and the impact of the COVID-19 pandemic on dietary habits and quality. However, limited evidence lies on the association between a healthy diet, and risk of SARS-CoV-2 infection or COVID-19. This study aimed to conduct a systematic review of observational studies to examine the association between diet quality, and the risk of SARS-CoV-2 infection or COVID-19 among adult populations. 6158 research articles from Scopus, EMBASE, PubMed, and MEDLINE databases were identified for eligibility. Only observational studies were included. Study quality was assessed using the National Institutes of Health Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. Thirteen studies were included (4 with SARS-CoV-2 infection and 9 with COVID-19 as the outcome); 3 were case-control, 3 were cross-sectional, and 7 were prospective studies. Adherence to Mediterranean diet was examined as exposure in 7 studies, and was associated with decreased risk of SARS-CoV-2 infection in 2 studies, with estimates varying from 12% to 22%, while COVID-19 risk or severity was found to be reduced in 3 studies with odds ratios (ORs) ranging from 36% to 77%. The Dietary Approaches to Stop Hypertension diet was inversely associated with COVID-19 hospitalization (OR: 0.19; 95% CI: 0.07, 0.55), whereas a healthy plant-based diet had an inverse association with both COVID-19 infection (hazard ratio [HR]: 0.91; 95% CI: 0.88, 0.94) and severity (HR: 0.59; 95% CI: 0.47, 0.74). Studies examining individual food groups generally found lower risk of infection or COVID-19 in association with larger dietary intakes of fruits, vegetables, and fiber. The overall findings of the observational studies in this review support the concept that nutritious diets might lower the risk of SARS-CoV-2 infection or COVID-19. This study was registered at PROSPERO as CRD42023397371.


Introduction
During the past 3 y, the world has been dealing with the Coronavirus Disease 19 (COVID-19) pandemic caused by a strain of coronavirus called the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1,2].Inflammation caused by the cytokine storm among COVID-19 infected cases is the result of a weak or overreactive immune system response [2,3].Immunity levels among subjects are influenced by a variety of factors, including stress, physical activity, genetics, vaccination status, comorbidities, such as cardiovascular disease or metabolic syndrome, and diet (nutrition status) [4].Among the mechanisms that are universally attributed to the effects of diet on diseases and health, diet-responsive effectors, specifically, the diet-immune axis, play an important role-for example, low protein can reduce the immunity status due to low antibody production [5,6].Optimal nutrition is required to maintain the inter-relationship between immune system and modulated inflammatory and oxidative stress processes [7].Known dietary and nutrient constituents that have anti-inflammatory and antioxidant properties are vitamin C [8], vitamin A [9], omega-3 fatty acids [10], and dietary fiber [11].Further, diets rich in phytochemicals, such as polyphenols and fiber, are suggested to act as prebiotics, eventually promoting healthy bacterial growth, including Bifidobacterium species, which reduces diarrhea, a common symptom in SARS-CoV-2 infection and COVID-19 cases [12,13].
Among evidence related to diet, few studies have also explored the association between individual foods, and the risk of SARS-CoV-2 infection or COVID-19, including dragonfruit, flax seeds, basil, cinnamon, and ginger [14][15][16].Further, studies have explored an inverse association between micronutrients, including vitamins (B 12 , C, and D) and minerals (iron, zinc, copper, and selenium), and the risk of COVID-19 [17][18][19][20].However, little is known regarding the associations between diet and the risk of SARS-CoV-2 infection or COVID-19 among observational studies among adult populations.
Therefore, the aim of this study was to conduct a systematic review of observational studies to examine the association between diet quality and the risk of SARS-CoV-2 infection or COVID-19 among adult populations.

Protocol registration
This systematic review was performed according to the PRISMA guidelines [21] (see the PRISMA checklist in Supplementary Table 2) and was registered at PROSPERO (reference CRD42023397371).

Information sources and search strategy
A systematic search was conducted in December, 2022 and repeated in July, 2023 among 4 databases (EMBASE, PubMed, Medline, and Scopus), and no date restrictions were applied.Searches were conducted using predefined key words relating to dietary patterns, diets, dietary intakes, COVID-19, and SARS-CoV-2 infection with MeSH terms wherever applicable, such as "Mediterranean diet," AND/OR "Mediterranean dietary pattern," "Diet," AND/OR "Diets," "Dietary intake," AND/OR "Dietary patterns," AND "COVID-19" AND/OR "COVID-19" AND/OR "SARS-CoV-2 infection" (see Supplementary Table 3 for the full search strategy conducted within each database).
The reference lists of all articles were searched for potential studies to be included in the review.The full text versions were stored as PDF files, and all the studies extracted were managed and stored using Rayyan Software (an automated tool for systematic reviews) [22] and Mendeley Reference Manager version 2.79.0.

Eligibility criteria
This systematic review only included observational studies (prospective cohort, case-control, and cross-sectional studies) that explored the association between diet quality (exposure) and risk of SARS-CoV-2 infection or COVID-19 (outcomes).Studies that examined severity and hospitalization because of COVID-19 were also included.Further, studies with adult human participants (>18 y) and examining dietary patterns based on food groups or components as the primary exposure, for example, a Mediterranean diet or "Western" dietary pattern, were included.Finally, studies that examined diets based on energy composition, for example, high-energy, low-fat, or highprotein diets, were also included.
The following studies were excluded: abstract, systematic review articles, editorials, case reports, letters, surveys, literature reviews, conference papers, thesis files, randomized controlled trials, and studies not conducted on human subjects.Also, studies that explored the influence of a single nutrient or food item, for example, vitamin C, iron, or ginger, on the risk of COVID-19 or SARS-CoV-2 infection were excluded.Finally, studies that were published in non-English language were also excluded.

Definition of outcomes
As all the included studies in this systematic review used different methods for ascertainment of SARS-CoV-2 infection and COVID-19, we defined the 2 outcomes for the systematic review according to the NIH COVID-19 Treatment Guidelines (CTG) [23], mainly classified into the following 5 levels of severity: 1) asymptomatic or presymptomatic infection: individuals with a positive test for SARS-CoV-2 using either a nucleic acid amplification test (includes RT-PCR tests along with Abbott ID NOW molecular rapid tests) or an antigen test but with no COVID-19 manifestations; 2) mild illness: patients with any of the COVID-19 symptoms (e.g., loss of taste and smell, fever, headache, malaise, myalgia, nausea, vomiting, diarrhea, cough, and sore throat) but no dyspnea or abnormal chest imaging; 3) moderate illness: individuals with clinical evidence of lower respiratory tract involvement or chest imaging and an oxygen saturation (SpO 2 ) of 94% on room air at sea level; 4) severe illness: individuals with a SpO 2 <94% on room air at sea level, a partial pressure of oxygen/fraction of inspired oxygen ratio <300 mmHg, a respiratory rate >30 breaths/min, or lung infiltrates >50%; and 5) critical illness: individuals suffering from respiratory failure, septic shock, and/or multiple organ dysfu nction.

Selection process
We used Rayyan Software-an automated web tool [22] for the screening process.The automated tool was used to identify duplicate articles from all the databases.Research articles were screened for potential inclusion based on their title and abstract by 1 reviewer (SS) with help of the "automated screening" feature in the Rayyan Software.Following which, an additional screening conducted by the reviewer (SS) as a precautionary measure to lower selection errors.Finally, the full text versions of the studies were retrieved and evaluated for inclusion based on the inclusion criteria stated above, and were assessed by 2 independent reviewers (SS and MB).The screening process is provided using a PRISMA flow chart, version 2020 (Figure 1) [21].

Data extraction and data items
The following data were extracted from the included studies: surname of the first author, publication year, country, period when the study was conducted, type of study design, SARS-CoV-2 infection and/or COVID-19 cases, sample size, study population, dietary assessment used for dietary data, measurement of dietary intake, outcome of interest, assessment of SARS-CoV-2 infection and/or COVID-19, covariates, statistical methods, results, and conclusion (Tables 1 and 2).Two independent reviewers assessed and selected the studies from the data extracted (SS and MB).The data pertaining to the sample, methods, and results from each of the included studies were extracted by 1 author (SS) and was cross-checked by another reviewer (MB).Studies were only included in the review upon mutual agreement of both reviewers (SS and MB) in the data extraction phase.

Statistical analysis
Due to the heterogeneity of the exposure and outcome measurements, and the small number of included studies in this systematic review that have reported the association between diet, and SARS-CoV-2 infection or COVID-19, the data extracted were deemed to be unsuitable for a meta-analysis.Thus, a systematic review was conducted for the same.

Quality assessment process
The National Heart, Lung and Blood Institute (NIH) Quality Assessment Tools [40] for observational cohort and cross-sectional studies (14-item criteria) and case-control studies (12-item criteria) were used to assess the risk of bias of the included studies.This tool does not use a points system to generate an overall quality of assessment score for the included studies.Rather, the categories for methodological quality are based on an overall judgment of the study: "Poor," "Fair," and "Good."Two independent evaluators assessed the quality of the included studies, and any disagreements were resolved through discussing the relevant parts of the paper to check if they had misinterpreted any element.Evaluators (SS and MB) had to select "Yes," "No," or "Not Reported/Not Applicable/Unable to Determine" in the NIH tool.An overall assessment for each study was generated based on the number of times "Yes" was selected under each criterion of the NIH tool; a "Good" study had a maximum of 3 categories that were not rated as a "Yes."Two categories, "validity of outcomes" and "adjustment of confounders," were considered as most important criteria to determine the classification of study quality.
Results of the quality assessments of individual studies for SARS-CoV-2 infection or COVID-19 are summarized in Supplementary Table 1A-C.

Literature selection
Diet and the risk of SARS-CoV-2 infection or COVID-19 A total of 6158 studies were identified from EMBASE (MED-LINE and PubMed) and Scopus databases that explored the relationship between diet and the risk of SARS-CoV-2 infection or COVID-19.An automated tool, Rayyan Software, was used to eliminate the duplicate records.The remaining 4976 records were screened for their eligibility based on title and abstract, of which 4963 studies were excluded for the reasons cited in the PRISMA flow chart in Figure 1.The full text versions of the remaining 13 studies were retrieved and assessed for eligibility.Therefore, a total of 13 observational studies were included in this systematic review.
It should be noted that few early studies [29,31,33] mislabeled the study outcome: SARS-CoV-2 infection as COVID-19; for this, we critically appraised the study methodology and reclassified the study outcome for better clarity (refer to Tables 1  and 2).
Among the cross-sectional studies, 2 were conducted in Iran [25,26] and one in Italy [24].The sample sizes ranged from 250 to 900 participants.The Italian study participants were health care professionals, whereas the Iranian studies included patients hospitalized for COVID-19.The Italian [24] and one Iranian [26] studies explored the Mediterranean diet as dietary exposure [26], whereas the other Iranian study examined individual food groups as the study exposure [25].Further, the Italian study utilized a 36-food item food frequency questionnaire (FFQ) to record dietary data, whereas the Iranian studies used an online validated 168-food item FFQ.The outcomes among these 3 studies differed; the Italian study explored the risk of SARS-CoV-2 infection, but the Iranian studies explored the risk of COVID-19 severity and symptoms.
Among the prospective studies, 3 of them were from United States and United Kingdom [28,33,34], 2 from Spain [29,31], and 1 each from France [36] and Italy [35].The sample size of these 5 cohorts ranged between 1520 and 592,571 participants, mostly recruited from the general population, university graduate students, or health professionals.Dietary data were recorded using varied methods: a couple of studies used concise        versions of FFQ, a 17-food item FFQ [33] and a 27-food item FFQ [28]; whereas one study from France used a 24-h dietary recall [32], and the rest of the studies used a detailed FFQ [29,31,34,35].All 7 studies computed and explored the dietary exposure in varied formats.3 studies assessed adherence to Mediterranean diet through the score proposed by Trichopoulou [27] (ranging 0-9) and its individual food components [29,31,35], and one study used the alternate Mediterranean diet (aMED) adherence score ranging from 0 to 9 [34].Whereas, 1 study analyzed only food components [33], and 3 studies used validated healthy plant-based diet index (hPDI) score [28] and the alternate healthy eating index score (AHEI-2010) [32,34].
Case-control studies explored risk of COVID-19 but examined different exposures: MedDiet score (0-55 points) and a posterioriderived dietary patterns [36], and self-reported dietary patterns [38,39] (refer to Table 1 for further study details).The sample size ranged between 141 and 2884 participants, mostly recruited from online surveys.Dietary data was obtained by using concise versions of FFQ in 2 studies (16-or 47-item FFQs) [36,38], whereas the Iranian study relied on the use of a 147-item FFQ [39].

Results of the included studies
Most studies included in the systematic review reported an inverse association between diet quality and the risk of SARS-CoV-2 infection or COVID-19 risk or severity, regardless of their study design (refer to Table 2 for detailed results).
Among studies that explored only individual food components as the exposure [26,33], it was observed that a higher consumption of fruits, vegetables, and dietary fiber was associated with lower risk of SARS-CoV-2 infection or COVID-19 (refer to Table 2).All studies analyzed data using multivariable-adjusted models, except one [24].Most studies adjusted the models for age, sex, BMI, comorbidities, and physical activity level.

Quality assessment findings
The NIH tool was used to assess the risk of bias of the studies that were included in the review (refer to Supplementary Table 1A-C for a summary of the risk of bias assessment).
Six studies (1 cross-sectional, 1 case-control, and 4 prospective studies) scored "Good" overall, deeming them to be at low risk of bias.However, 7 studies (2 each of case-control and crosssectional studies and 3 prospective studies) scored "Fair" overall, deeming them to have a slightly higher risk of bias due to the following reasons: definition, selection of key confounders, reliability and consistent implementation of measurement of exposures, and assessment duration of exposures.

Discussion
The aim of this systematic review was to examine the evidence from observational studies evaluating the association between diet and the risk of SARS-CoV-2 infection or COVID-19.The overall results, based on 13 studies, indicated that a higher adherence to healthy nutritious diets was inversely associated with the risk of SARS-CoV-2 infection or COVID-19.The studies included in the systematic review underpinned a gap in the literature on the requirement to assess the association of the dietary impact on the risk of SARS-CoV-2 infection or COVID-19 among observational studies.
Previous studies extensively explored the impact of contracting SARS-CoV-2 infection or COVID-19 on diet or the dietary habits/dietary quality during the pandemic and reported a modification in the lifestyle, and diet quality [41,42].Whereas, few studies focused on short-term effects of isolated foods in the form of functional foods, for example, garlic, ginger, flax seeds, and dragonfruit [14][15][16], and single nutrients, for example, vitamin D [43,44], vitamin C, zinc, and selenium [19,20,45], on the risk of SARS-CoV-2 infection or COVID-19.This seemed to indicate an inadequate reductionist approach of exploring the effects of single nutrients or food items on a disease of interest [46][47][48] because foods are mostly consumed in various combinations, making it difficult to attribute the therapeutic effect to a single nutrient or food item of interest [46][47][48].Since the examination of dietary patterns is beneficial in the formation of dietary guidelines [46], our decision to examine the associations between the diet (exposure) in the form of dietary patterns, food components or index scores remained justified.
To the best of our knowledge, this is the first systematic review that investigated the association between diet quality and SARS-CoV-2 infection or COVID-19 among observational studies, thus, aiming to capture an entire SARS-CoV-2 infection and disease spectrum, i.e., from contracting the infection to hospitalization due to COVID-19.A systematic review explored the effect of Mediterranean diet on the inflammatory biomarkers among overweight/obese adults from randomized controlled trials and case-control studies, and observed that a hypocaloric, fiber-dense Mediterranean diet could help lower the inflammatory markers among a high BMI adult population at risk of developing COVID-19 [49].Whereas, our systematic review examined diet quality across all available observational studies (dietary patterns and food component intakes) recorded prior and/or during the COVID-19 pandemic, and explored associations with the risk of SARS-CoV-2 infection or COVID-19.
The results of the included studies were conducted in different countries, i.e., Italy, Spain, United Kingdom, Iran, United States, Lebanon, and France, making the results widely generalizable and replicable.Although the studies measured the dietary data in different formats, for example, Mediterranean diet [24,26,29,31,[34][35][36], AHEI-2010 [28,32], and food groups and individual components [25,33], the results of all the studies consistently suggested that a nutritious diet might be beneficial in lowering the risk of SARS-CoV-2 infection or COVID-19, despite differing in covariate adjustment choices.
Dietary assessment tools and the techniques used to record the dietary data varied among all the studies.Among studies that used FFQ as the dietary assessment tool, the study by Merino et al. [28] used a validated symptom-based algorithm to identify the COVID-19 cases across different populations as their study design included web-based participant recruitment.However, the methodology to compute the diet scores was lesser in detail as they used an FFQ (28-food item Leeds FFQ).This might have caused an underrepresentation of food components and intake important to compute the diet score index.Whereas few studies [26,29,34,35,39] used a detailed FFQ to record the dietary data of the participants in their study.
Further, these studies used the SARS-CoV-2 infection diagnostic results as a confirmation-which might have introduced misdiagnosis and measurement errors in identifying the degree of risk of COVID-19.However, the definition of SARS-CoV-2 infection and COVID-19 was not as detailed as the study by Zargarzadeh et al. [26] and Sharma et al. [35], but, was more robust than in the Merino et al. [28] and the Vu et al. [33] studies, respectively.In contrast, the Zargarzadeh et al. [26] study used the initial measurement of C-reactive protein (CRP) at hospital admission, obtained from the medical records, as well as the 5 levels of severity prescribed by the NIH CTG, whereas Sharma et al. [35] used participant data of anti-nucleocapside protein antibodies for SARS-CoV-2 infection diagnosis in their cohort study.
Further, the study by Vu et al. [33] in the UK Biobank cohort also used an FFQ (17-food item) to record the dietary data similar to the Merino et al. [28] study.This might have affected the quality of the dietary intake assessment, but, nevertheless, the UK Biobank cohort was linked to the national health registry with available COVID-19 data.
In contrast to the studies that used the FFQ as a dietary assessment tool, the Deschasaux-Tanguy et al. [32] study used a 24-h dietary recall to compute the AHEI-2010 score, and utilized a robust methodology to ascertain the SARS-CoV-2 infection but not COVID-19 by using commercial enzyme-linked immunosorbent assay (ELISA).
Although effective hygiene measures, including sanitizers, handwash using soap, K95 masks, and social distancing mandates curbed the overall SARS-CoV-2 infection and COVID-19 rates, studies suggested that infection contraction was related to immunity status, and that COVID-19 was observed among people with low immunity levels [50][51][52][53].Individuals with low grade chronic inflammation have a poor innate immune system, which increases their likelihood of infection [11,54], besides other factors, including genetics, BMI, physical fitness, vaccination status, gut microbiota, stress, illness (cardiovascular disease, diabetes mellitus, cancer, arthritis, obesity, and inflammatory bowel diseases), and diet (nutritional status) [4].Among these, evidence suggested that optimal nutritional status and better diet quality adherence could potentially be associated with lower risk of contracting the SARS-CoV-2 infection and subsequent COVID-19 [28,35,55,56].Studies have suggested that a diet high in olive oil, cereals, fruits and nuts, and vegetables were associated with lower odds of SARS-CoV-2 infection and COVID-19.Underlying biological plausibility explaining these associations include anti-inflammat ory markers, including IL-6, CRP, antioxidants, antithrombotic effects, and adhesion factors that are beneficial in COVID-19 prevention [28,32,35,38].Therefore, promotion of a better diet quality to lower the risk of SARS-CoV-2 infection is vital to decrease the subsequent risk of COVID-19.This is because SARS-CoV-2 infection causes angiotensin-converting enzyme (ACE) and its homolog angiotensin-converting enzyme 2 (ACE/ACE2) balance disruption and renin-angiotensin-aldosterone system (RAAS) activation, which ultimately leads to COVID-19 progression.This is especially seen among individuals with comorbidities, such as diabetes mellitus, hypertension, and cardiovascular disease, which are also preventable through healthy dietary interventions [57][58][59][60].Therefore, a dual approach could potentially be a healthy diet quality adherence impacting overall immunity levels coupled with effective hygiene measures, thus lowering SARS-CoV-2 infection, and eventually COVID-19.

Strengths and limitations
Our study had some strengths and weaknesses to be considered.Despite a relatively low number of observational studies included in the systematic review, NIH risk of bias tool and quality assessment tool yielded an overall score of "Fair."The NIH tool was selected because it was simple yet robust, replicable, and widely used in systematic reviews related to diet and disease [61][62][63][64][65].The NIH tool was user-friendly with a concise set of questions focusing on each criterion that directly impacted overall quality assessment ratings, and did not use a points system for assessment but was based on the judgment of study.It was possible to discriminate and compare the responses for each criterion within each study and the corresponding ratings and between multiple studies (overall quality assessment of each study).Further, the tool helps identify and evaluate potential flaws in study methodologies, including bias (for example, participant selection), measurement and selection of key confounders, inclusion and exclusion criteria definition, assessment of exposure and outcome measurements, and other criteria.Finally, the NIH tool provided brief guidance for each question/criterion helpful for evaluators to efficiently conduct quality assessments-lowering overall risk of bias and judgment error while conducting large systematic reviews.
We included all available observational studies that computed the dietary data in varied measurement formats, but underreporting and bias must be considered as the dietary data was self-reported [66,67].Non-English articles (n ¼ 79) were excluded in the screening phase, but these studies did not examine this review's objective.
This systematic review screened data from multiple databases and used an efficient automated tool, Rayyan Software, to manage the publications, screening, and data extraction process compared to conventional methods, for example, Excel sheets-introducing potential bias and errors in the methodology.
It should also be noted that our systematic review focused on diet quality as the exposure, and not on exposures, such as food habits or ultraprocessed foods.Additionally, this systematic review presents evidence for the entire SARS-CoV-2 infection spectrum potentially contributing toward future research and public health policies.However, few studies, especially those conducted in the early stages of the COVID-19 pandemic, did not adequately define the COVID-19 outcome-because some studies actually analyzed SARS-CoV-2 infection and not COVID-19 per se.

Conclusion
The overall findings of the observational studies included in this review consistently suggested that a nutritious diet might lower the risk of SARS-CoV-2 infection or COVID-19.These results elucidate the importance of a wholesome diet that is beneficial in potentially protecting against SARS-CoV-2 infection or COVID-19.This systematic review highlights the limited number of observational studies, which could be of public health importance and for cautious nutritional advice in clinical settings.

FIGURE 1 .
FIGURE 1.The PRISMA 2020 flow chart shows the study selection process.Identification of studies via databases and registers for diet and risk of SARS-CoV-2 infection or COVID-19.

TABLE 1
Baseline characteristics of included studies exploring the association between diet and the risk of SARS-CoV-2 infection or COVID-19 COVID-19 (i.e., fever, (continued on next page) S. Sharma et al.Advances in Nutrition 14 (2023) 1596-1616

TABLE 2
Results of the included studies exploring the association between diet, and the risk of SARS-CoV-2 infection or COVID-19.
(continued on next page)

TABLE 2
(continued on next page)

TABLE 2
(continued on next page)

TABLE 2
(continued on next page)