Benefits and Harms of Edible Vegetable Oils and Fats Fortified with Vitamins A and D as a Public Health Intervention in the General Population: A Systematic Review of Interventions

This systematic review aims to assess whether edible vegetable oils and fats fortified with vitamin A and/or D are effective and safe in improving vitamin intake and ameliorating deficiency states in the general population. In November 2022, we systematically searched MEDLINE, Cochrane CENTRAL, Scopus, Global Index Medicus, ClinicalTrials.gov, and WHO ICTRP (International Clinical Trials Registry Platform) for randomized controlled trials (RCT) and non-randomized studies of interventions (NRSI) investigating the fortification of edible vegetable oils and fats with either vitamin A or vitamin D or both as compared to the same vegetable oils and/or fats without vitamin A and D fortification or no interventions, in the general population, without age restriction. We assessed the methodological quality of included RCTs using Cochrane’s risk of bias tool 2.0 and of NRSIs using ROBINS-I tool. We performed random-effects meta-analysis and assessed certainty of evidence using GRADE. We included eight studies. Available evidence showed no significant effect of fortification with vitamin A on serum retinol levels (RCTs: MD 0.35 µmol/L, 95% CI −0.43 to 1.12; two trials; 514 participants; low-certainty evidence; CCTs: MD 0.31 µmol/L, 95% CI −0.18 to 0.80; two trials; 205 participants; very low-certainty evidence) and on subclinical vitamin A deficiency. Low-certainty evidence showed no effect of vitamin D fortification on serum 25-hydroxy vitamin D concentration (MD 6.59 nmol/L, 95% CI −6.89 to 20.07; one trial; 62 participants). In conclusion, vitamin A-fortified vegetable oils and fats may result in little to no difference in serum retinol levels in general populations. The dose of vitamin A used in the trials may be safe but may not be sufficient to reduce subclinical vitamin A deficiency. Further, the evidence suggests that vitamin D fortification results in little to no difference in serum 25-hydroxy vitamin D concentration. Several aspects of providing fortified oils and fats to the general population as a public health intervention should be further investigated, including optimal fortification dose, effects on vitamin D deficiency and its clinical symptoms and potential adverse effects.


Introduction
Food fortification, where essential micronutrients are added to widely consumed staple foods and condiments during production, either compulsorily or voluntarily, is a strategy that has been used safely and effectively for more than a century to prevent micronutrient deficiencies and related health problems in high-income countries [1].Compared to voluntary food fortification, which is primarily used for marketing purposes, public health fortification campaigns aim to address vitamin and mineral deficiencies at the population level without creating economic inequalities (homogeneous affordability) [1].
Vitamin A is a group of fat-soluble molecules with a similar structure, including retinol, retinal, retinoic acid, and several provitamin A carotenoids (most notably beta-carotene) [2].Vitamin A has diverse functions: it is essential for vision, for embryo development and growth and for maintaining the immune system [3].Therefore, vitamin A deficiency (VAD) can impair the function of neutrophils, macrophages, NK cells, and diminish the Th2 cytokine-production and Th1-mediated immunity [2].VAD is a major nutritional problem in many parts of the world, especially in low-income countries, leading to a number of health problems, including xerophthalmia, increased susceptibility to infections and anemia.VAD is the leading cause of preventable blindness [4], but children with VAD are at increased risk of morbidity and mortality as well [5].The risk factors for the development of VAD are multifactorial, including demographic (mainly men and preschool children), geographical (mainly in Africa and Southeast Asia), childhood (breastfeeding, infections), household (lower socioeconomic status, poor hygiene), and dietary (lower quality and diversity of diet) factors [6].The World Health Organization (WHO) estimates that VAD affected an estimated 190 million pre-school children and 19.1 million pregnant women worldwide between 1995 and 2005, mainly in Africa and Southeast Asia [7].
Vitamin D 3 or cholecalciferol, another fat-soluble vitamin, can be taken up with food, but the main source is the endogenous synthesis in the skin.The active form is obtained after two hydroxylation steps and is called calcitriol (1,25-dihydroxy vitamin D 3 ) [8].Vitamin D plays a central role in calcium homeostasis, and therefore in bone mineralization [9], but it also has immunomodulatory effects in both innate and adaptive immunity, and through the immune cells, in both acute and chronic inflammation as well as in the pathomechanism of several autoimmune processes [8].Dietary sources of vitamin D, including eggs, dairy products, meat, and fish, are limited, so commercially fortified products make a sizeable contribution to daily dietary intake [10,11].Vitamin D deficiency is a global health problem affecting all age groups in almost every country in the world.The global burden of vitamin D deficiency is hard to quantified, as different definitions of deficiency exist, which are all based on serum 25-hydroxyvitamin D levels (25(OH)D) [12].The determinants of lower vitamin D status may vary depending on the location (lower exposure to sunlight, lower consumption of vitamin D-containing foods, urbanization, air pollution, higher body mass index (BMI)) [13].Vitamin D deficiency can primarily cause symptoms in the bones, namely reduced mineralization, leading to nutritional rickets in children and osteoporosis in adults [14], as well as chronic inflammation, autoimmunity, and the increased frequency of infections [8,15].
There are three main strategies which might be effective in the prevention of vitamin deficiencies: increasing diversity, supplementation, and food fortification.Improvements in food diversity are difficult to achieve when limited amounts of food items with high vitamin content are available.Supplements are usually used by a small proportion of the population; therefore, food fortification is the strategy preferred by the WHO in terms of coverage [16].
Edible vegetable oils and fats are one of the most important staple foods worldwide because of their energy density, but they are also natural sources of fat-soluble vitamins (A, D, E, and K) and act as a solvent to enhance the absorption of fat-soluble vitamins.Edible vegetable oils and fats are consumed widely, regardless of wealth.The production of vegetable oils more than doubled between 2000 and 2019 [17].In most countries they are processed centrally by medium and large-scale producers, which facilitates the implementation and monitoring of a potential fortification process [10].
Existing systematic reviews and meta-analyses mainly focus on the health outcomes of vitamin A or vitamin D fortification of all types of staple foods in the general population [18,19] or in children only [20][21][22] and are mainly based on results from clinical trials.A lower number of evidence summaries focus on the fortification of specific vehicles (e.g., bread [23] and yoghurt [24]), but no systematic review has been published on the effects of vitamin A or vitamin D fortification of edible oils and fats in the general population.
The current systematic review aims to synthetize up-to-date data from both interventional and observational trials and provide a systematic assessment of the benefits and harms of edible oils and fat fortified with vitamin A or vitamin D, either alone or in combination to inform policymaking and assist countries in the design and implementation of appropriate food-fortification programs.

Materials and Methods
The methodology and the results are reported according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) reporting guidelines.This study is registered with PROSPERO, CRD42022351689.

Search Strategy
For this systematic review and meta-analysis, we searched the following electronic databases and trial registers from the inception of each database up to 14 November 2022 without restrictions on the language of publication: Ovid MEDLINE, Cochrane Central Register of Controlled Trials (CENTRAL), Global Index Medicus (comprising African Index Medicus (AIM), Index Medicus for the Eastern Mediterranean Region (IMEMR), Index Medicus for the South East Asia Region (IMSEAR), Latin America and the Caribbean Literature on Health Science (LILACS), and Western Pacific Region Index Medicus (WPRO)), Scopus and trial registers (https://clinicaltrials.gov/, WHO ICTRP (International Clinical Trials Registry Platform; apps.who.int/trialsearch)).Details for all search strategies are available in Supplementary File S1.
Using the reference lists of included studies, related systematic reviews, meta-analyses, and health technology assessment reports, we attempted to identify other potentially eligible trials or additional publications.We searched for grey literature, which we defined as searching the Global Index Medicus, as well as trial registers.

Eligibility Criteria
We included randomized controlled trials (RCTs), controlled clinical trials (CCTs), cohort studies, controlled before-after studies, and interrupted time series.For cluster randomized trials, non-randomized cluster trials, and controlled before-after studies, we only included studies with at least two intervention sites and two control sites.We included the general population (including pregnant women), comprising individuals of any age and from any country.Studies of interventions targeted toward participants with a critical illness or severe comorbidities were excluded.
The eligible interventions were edible oils and/or fats (of vegetable origin) for household use fortified with either vitamin A or vitamin D or a combination of vitamin A and D compared to no intervention or the same unfortified oil and/or fat.No restriction was made regarding the type of vegetable oil (extracted from seeds or from other parts of fruits).We excluded studies comparing vitamin A and/or vitamin D oil or fat fortification with other forms of vitamin A and/or vitamin D interventions (i.e., supplementation or dietary diversification) or fortification of other food vehicles (e.g., sugar, flour, milk, and dairy products).

Selection Process
Pairs of review authors (ES, DK, IC, RF, PNN, KI, SL) independently screened the abstract, title, or both, of every record retrieved by the literature searches using COVIDence TM software, https://www.covidence.org/.We obtained the full texts of all potentially relevant records and screened these for eligibility.Any disagreements were resolved through consensus or by recourse to a third review author (SL).Potentially relevant articles written in a language other than English were translated to English prior to full text assessment.Multiple reports of the same study were merged, as each study rather than each report was the unit of interest in this review.All articles excluded after full-text assessment and the reasons for their exclusion are described in the table on characteristics of excluded studies (Supplementary File S2).The trial selection process is presented in a PRISMA flow diagram.

Data Collection
From the full-text publications, we extracted data on study methods, participants, interventions, controls, outcomes, confounders, and funding sources.Data were extracted by one reviewer (IC or RF) and verified for completeness, accuracy, and consistency by a second reviewer (IC or RF).
We included abstracts and conference proceedings but did not use them to extract data, as they did not meet CONSORT requirements.We also extracted data available in the study registers as study results.
The main outcomes, defined by the WHO guideline development group (GDG), were markers of vitamin A and/or D deficiency (measured as serum retinol, serum 25(OH)D, subclinical/clinical VAD, vitamin D deficiency, osteomalacia, nutritional rickets), all-cause morbidity and mortality, and any adverse effects.Additional outcomes were vitamin A status, dietary vitamin A/D intake, iron status, anemia, maternal and infant outcomes, growth, weight change, and any longer-term outcomes.We included outcomes as measured at any given timepoints.
We extracted data on study information, participants, type of intervention, type of outcomes (both primary and secondary outcomes specified and collected, timepoints reported), adjusted and unadjusted outcome measures, confounders, methods used to control confounders, funding, and any notable conflicts of interest of the study authors.Studies reporting outcomes at multiple timepoints, we extracted data for each timepoint.Data extraction was performed by one reviewer and was checked for completeness, accuracy, and consistency by a second independent reviewer.We attempted to obtain missing data from the study investigators.

Risk of Bias Assessment
Two review authors (ÉS and DK) independently assessed the risk of bias of each included trial.Any disagreements were resolved by consensus.Risk of bias in RCTs was assessed using version 2.0 of the Cochrane "Risk of bias" tool (RoB2), while in NRSIs (including quasi-randomized studies, cohort studies, controlled before-and-after studies, and interrupted time series) were assessed using the "Risk of Bias in Non-randomized Studies of Interventions" (ROBINS-I).To illustrate the risk of bias judgements for RCTs and NRSIs, we used the robvis tool to create traffic light plots [25].

Effect Measures
For dichotomous data, we present results as risk ratios (RRs) or odds ratios (ORs) with 95% confidence intervals (CIs).For continuous data, we use mean differences (MDs) with 95% CIs for studies measuring outcomes in the same way and standardized mean differences (SMDs) with 95% CIs for studies measuring outcomes in a variety of ways.

Synthesis Methods
We used RevMan 5 (version 5.4.1) for statistical analyses.As we expected differences between studies in both the population and the intervention, we decided to combine the data using a random effects model, when it was clinically meaningful to do so, to provide an average treatment effect across studies.We used Mantel-Haenszel weighting for dichotomous outcomes and inverse variance for continuous outcomes.In case both individually randomized and cluster-randomized trials were included in a meta-analysis, we planned to use the inverse variance method.
Methodological heterogeneity was assessed by examining risk of bias, while clinical heterogeneity was assessed by examining similarities and differences between studies regarding types of participants, interventions, and outcomes.We considered the size and direction of effect and used a standard χ 2 test with a significance level of α = 0.1 and I 2 statistic, quantifying inconsistency across trials, to assess the impact of heterogeneity on the meta-analysis.We explored heterogeneity by conducting pre-specified subgroup analyses.
We planned to perform subgroup analyses for the following characteristics for both vitamin A and D: age groups, psychological condition, vitamin A/D intake, public health significance of vitamin A/D deficiency in the trial's country, vehicle of intervention, consumption patterns, duration of intervention, amount of added vitamin A/D through fortification, type of vitamin compound, type of fortification intervention, method of cooking, and delivery platform.We planned additional subgroup analyses for vitamin D only for: skin pigmentation, latitude, exposure to environmental pollutants, BMI, exposure to additional vitamin D though other programs, and as a method to stabilize vitamin D.
We planned to conduct sensitivity analyses to examine the potential effects of clustering on the CIs of summary estimates.

Reporting Bias Assessment
We planned to use funnel plots to assess reporting bias (such as publication bias) and to investigate the relationship between effect size and standard error when 10 or more studies were included in a meta-analysis.The degree of funnel plot asymmetry was planned to be quantified using Egger's test.

Certainty Assessment
We followed the GRADE approach to rate the certainty of evidence [26].

Description of Included Studies
We retrieved 5678 unique records through database searching (Figure 1).After removing duplicates, 4532 records were screened based on their titles and abstracts.Most of the references (n = 4441) clearly did not meet the inclusion criteria based on title and abstract review and were excluded.We evaluated 91 full texts or records to determine their eligibility for inclusion in the review.Of these, 28 studies were excluded because they were not RCTs or NRSIs, 2 studies were excluded because the participants were people with a specific disease, 34 studies were excluded because the intervention/exposure was not an oil or fat fortified with either vitamin A or vitamin D or their combination, and 33 studies were excluded because there was no eligible comparator (Supplementary File S2).Eight studies (reported in 24 records) met our inclusion criteria for qualitative, and four studies (reported in 6 records) met the requirements for quantitative synthesis.One of the included studies (with 13 associated records) was a large birth cohort study based on the cancellation of mandatory fortification of margarine in Denmark in 1985, which we will refer to as the "Danish study" (Supplementary File S3).A total of five studies were included for the comparison of vitamin A fortification versus no fortification with vitamin A (Table 1), including two RCTs [27][28][29][30], two CCTs A total of five studies were included for the comparison of vitamin A fortification versus no fortification with vitamin A (Table 1), including two RCTs [27][28][29][30], two CCTs [31,32] and one birth cohort study [33].Participants in all RCTs and CCTs were allocated to groups at the individual level.
Three studies were conducted in a high-income country [33,37], one was conducted in an upper-middle-income country [32], while four studies were conducted in lower-middleincome countries including Indonesia [31], the Philippines [27], Morocco [28], and Iran [35], while no study was conducted in a low income country.
Participant age ranged from 4 to 40 years, while in the birth cohort studies, fetuses or pregnant women were either exposed or not exposed.Sample sizes ranged from 31 [31] to 331,623 [44].Among studies investigating longer-term effects of fortified edible oil consumption, intervention duration lasted between 8 weeks [31] and 6 months [27,28], while in birth cohort studies, there were no detailed information about intervention duration.

Risk of Bias in Included Studies
Overall, two randomized trials (66%) [27][28][29] were rated as having a moderate risk of bias, while one study was evaluated as having a low risk of bias [34,35].In the included non-randomized trials, two studies [32,33] were rated as having a high risk of bias due to the selection of participants, while 60% of the articles had a moderate risk of bias (Supplementary File S4).

Primary Outcomes for Studies on Vitamin A Fortification versus No Fortification with Vitamin A
Two randomized [27,28] and two non-randomized studies [31,32] with intervention durations of 6 months and 2-5 months, respectively, measured serum retinol.Available evidence based on RCTs showed no effect of fortification with vitamin A on serum retinol levels (MD 0.35 µmol/L, 95% CI −0.43 to 1.12; two trials; 514 participants; low-certainty evidence, Table 3), also supported by evidence derived from non-randomized studies (MD 0.31 µmol/L, 95% CI −0.8 to 0.80; two trials; 205 participants; very low-certainty evidence; Supplementary File S5, p. 1).
Similarly, no effect on subclinical vitamin A deficiency, measured as serum retinol ≤ 0.70 µmol/L in one RCT (0/268 vs. 0/144, RR not estimable, one trial, low-certainty evidence, Table 3), supported by evidence derived from the CCT (RR 0.21, 95% CI 0.01 to 4.10; one trial; 31 participants; very low-certainty evidence; Supplementary File S5, p. 6) and no effect on all-cause morbidity (low certainty and very low-certainty evidence, respectively) of fortification with vitamin A was found.
All-cause morbidity was measured in two RCTs [27,28] and one CCT [31]; however, only one RCT [27] and one CCT [31] reported results.Neither the RCTs nor the CCTs described differences between groups in all-cause morbidity (low certainty and very lowcertainty evidence, respectively, Table 3).No studies reported data on clinical vitamin A deficiency, adverse effects, or all-cause mortality.

⊕⊕ Low CRITICAL
All-cause morbidity-non-randomized studies (follow-up: 2 months) 1 [31] randomized trials Serious f not serious j not serious very serious h none One CCT reported morbidity scores (defined as frequency of illness multiplied by duration of illness) and described no significant differences between study groups.CI: confidence interval; MD: mean difference; RR: risk ratio.Explanations: a .Downgraded by one level for RoB since both included studies were rated with some concerns for RoB.b .Downgraded for inconsistency as point estimates varied widely, 95% CI did not overlap between studies, the direction of effect was not consistent, and the magnitude of heterogeneity was high (I2 was 98%, p-value for heterogeneity was <0.0001).Sub-group analyses did not fully explain heterogeneity.c .Not downgraded for imprecision.Although only two studies were included, the magnitude of the median sample size was intermediate (n = 257), and the total sample size was larger than 400 (n = 514).d .Downgraded by one level for RoB since one of the two included studies was rated with a high RoB, and none of the included studies were rated with a low RoB.e .Downgraded by one level for imprecision since the number of included studies was small (n = 2), the magnitude of the median sample size was intermediate (n = 103), and the total sample size was smaller than 400 (n = 205).f .Downgraded by one level for RoB since the included study was rated with some concerns for RoB.g .Downgraded by one level for imprecision.There was only one study included, but the total sample size was larger than 400 (n = 412).h .Downgraded by two levels for imprecision since results are derived from one study, where total sample size was very low (n < 100).i .Downgraded by two levels for RoB, as results were not reported for one out of two studies, and additionally, because none of the included studies was rated with a low RoB.j .This is a single study so inconsistency cannot be judged.

Discussion
To our knowledge, this is the first systematic review summarizing evidence on the consumption of vitamin A and/or D-fortified edible oil or fat compared with the unfortified version of the same oil or fat in a general population.The evidence suggests that vitamin A fortification may result in little to no difference in serum retinol levels in general populations.The dose of vitamin A used in trials may be safe but may not be sufficient to reduce subclinical vitamin A deficiency.Similarly, the consumption of vitamin D-fortified oils and/or fats may result in little to no difference in serum 25(OH)D concentrations.Available evidence suggests that vitamin A-fortified oils/fats might increase dietary vitamin A intake, and therefore vitamin A status; however, there is no current evidence that vitamin D intake is increased by consuming vitamin D-fortified oils/fats.There is no current evidence that gestational vitamin D fortification can influence maternal and neonatal outcomes, but it might be beneficial in growth and weight gain in childhood and might also have some longer-term health effects.
A significant strength of this study is that we used a broad search strategy in both electronic databases and trial registries without applying date or language restrictions.It is unlikely that published trials have been missed; however, unpublished or ongoing trials not registered in clinical trial registries could be missing.Secondly, we aimed to reduce bias wherever possible by having at least two review authors work independently on trial selection, data extraction, and "Risk of bias" assessments.We examined the general population without age restrictions, so that our results can be used widely, not just in certain age groups.Finally, we examined the effects of the fortification of edible oils/fats only, thus reducing the effects of potentially different vitamin absorption from different food types.
However, a major limitation of this systematic review is that several prespecified outcomes were investigated in a small number of trials or that no data were available at all.Due to the low number of studies, we were also not able to explore the potential for publication bias using a funnel plot.
Most existing reviews have addressed the effects of different food vehicles like dairy products, flour, grains, and oils, but most include results mainly from controlled trials and do not considered specific population subgroup analysis.There are only a few metaanalyses discussing the effect of vitamin A [18] or vitamin D [19,23] in the general population; most of them included either only children [20][21][22] or only adults [24,[50][51][52].Almost all meta-analyses found a consistent improvement in vitamin 25(OH)D levels with vitamin D fortification [19,[21][22][23][24]51,52], while the effect of vitamin A fortification is less clear [18,20].By contrast, in this meta-analysis, no significant effect was found on either serum 25(OH)D values or serum retinol, although we were able to only include one RCT for vitamin D fortification and two RCTs as well as two CCTs for vitamin A fortification.Most meta-analyses focused only on serum retinol or serum 25(OH)D levels, while some also investigated the effect of fortified food on clinical/subclinical VAD [18] or other cognitive functions [21].
Although oils and fats are widely consumed staple foods worldwide, providing an ideal solvent for fat-soluble vitamins, there are only a few clinical trials that have investigated the effects of vitamin A/D fortification on edible vegetable oils or fats.Our results suggest that the doses used in the trials so far are safe, but further clinical studies are needed to establish effective doses for the prevention of vitamin A and/or D deficiency.Future research should also clarify the stability of added vitamin A/D in different oils and fats under various conditions and types of usage, as very diverse factors can influence this.
When considering the advantages and disadvantages of the implementation of the fortification of edible oils and fats, it is essential that the worldwide consumption of edible oils and fats; the effectiveness and safety of currently existing fortifying policies; challenges during implementation; and aspects of cost-effectiveness, acceptability, and the potential impact on non-communicable diseases (NCDs) are taken into account.
Currently, 35 countries have mandatory policies and eight countries voluntary policies regarding the fortification of edible oils, mainly in Asian and African countries [11]; however, only about 40% of the population consume fortified vegetable oil and 34% consume adequately fortified oil based on a recent meta-analysis [53].On the other hand, core micronutrient deficiencies, including of iron, zinc, and vitamin A, are still high worldwide, affecting nearly half of pre-school children and non-pregnant women of reproductive age [54].
Fat and oil fortification guidelines should be developed with consideration of the broader nutritional context.Based on guidance formulated by the WHO, total fat should not exceed 30% of total energy intake, the intake of saturated fats should be less than 10%, and that of trans-fats less than 1% of total energy intake [55].Edible oils have different saturated fat contents and fatty acid profiles [56].Currently, palm oil is the most commonly produced oil worldwide, followed by some healthier alternatives, including soybean, rapeseed, and sunflower oils [57].
Based on the regulations of Codex General Principles for the Addition of Essential Nutrients to Foods "fortification should be the responsibility of national authorities since the kinds and amounts of essential nutrients to be added and foods to be fortified will depend upon the particular nutritional problems to be corrected, the characteristics of the target populations, and the food consumption patterns of the area" [58].Although oil consumption should not be promoted in any way, it should be taken into account that in some countries, adequate micronutrient intake through healthy diets is not feasible for large groups of people, so vitamin intake through processed food, which is otherwise consumed regularly, may be a possible solution to prevent vitamin deficiencies.
Although vitamin A and D deficiency is a global health problem and the fortification of oils and fats with vitamin A and D might be a safe strategy that countries could consider making part of their strategy to tackle deficiencies, the results based on the included studies suggest that vitamin A-and D-fortified oils have little or no effect on health; however, more studies are needed as the sample size is presently very low, meaning that the presence of small effects that might be still relevant on the population level, cannot be excluded with a high degree of certainty.
In conclusion, vitamin A and D deficiencies are global health problems, and the fortification of oils and fats with vitamin A and D might be a safe strategy which countries could consider making part of their policies to tackle deficiencies, after assessing local circumstances.In order to be able to formulate recommendations based on higher-certainty evidence, further studies investigating the effectiveness and safety of vitamin A and D fortification are needed.

Figure 1 .
Figure 1.Study selection.CENTRAL: Central Register of Controlled Trials; ICTRP: International Clinical Trials Registry Platform.

Figure 1 .
Figure 1.Study selection.CENTRAL: Central Register of Controlled Trials; ICTRP: International Clinical Trials Registry Platform.

Table 1 .
Key characteristics of included studies with vitamin A fortification as the intervention.

Table 2 .
Key characteristics of included studies with vitamin D fortification as the intervention.

Table 3 .
Vitamin A-fortified oils or fats compared to same oils or fats without vitamin A in the general population as a public health intervention.