Performance Factors Influencing Efficacy and Effectiveness of Iron Fortification Programs of Condiments for Improving Anemia Prevalence and Iron Status in Populations: A Systematic Review

Iron fortification of staple foods is a common practice around the world to reduce the prevalence of iron-deficiency anemia. More recently, fortified condiments, including salts, sauces, and powders, have been tested in various efficacy trials. However, there is limited information on how nutritional, environmental, and experimental factors affect their efficacy and effectiveness. The purpose of the present work was to systematically review performance factors affecting the efficacy of condiment fortification trials. Three databases were searched using a standardized keyword search and included based on four-point inclusion criteria. Studies were evaluated against a quality assessment tool and effect sizes were calculated. Studies were ranked as low or high performing, based on whether or not they significantly improved iron-deficiency outcomes (hemoglobin, anemia prevalence, and ferritin levels). Of the 955 retrieved studies, 23 were included—of which, nine performed poorly, eight performed highly, and six were classified as neither because they did not meet the criteria of assessing the three iron outcomes. Results showed that unsuccessful trials did not consider environmental factors such as parasitic infections, nutritional factors such as micronutrient deficiencies other than iron, consumer acceptability of the product or experimental factors such as monitoring and adherence to the trials. Two common performing factors identified among those studies performing highly vs. those that did not were the control of sensory changes and monitoring of consumption compliance (i.e., dose delivery). The present work can be used as decision-making support for nutrition policy makers when determining the appropriate implementation of condiment fortification programs.


Introduction
Iron deficiency and iron-deficiency anemia are the most prevalent and widespread nutritional conditions around the globe at all socioeconomic levels. The deficiency of iron affects children and women of reproductive age primarily, especially during pregnancy [1]. This condition is linked to poor neurological and cognitive functions in children, to impaired work capacity in adults, and greater mortality in pregnant women [2]. According to the World Health Organization (WHO), the latest prevalence estimates of anemia in 2016 were 41.7% in children and 32.8% in women of reproductive age, being most prevalent in low-and middle-income countries [3]. Practically no progress has been made in the last seven years since 2012 when the United Nations World Health Assembly set a goal for lowering the prevalence of anemia by 50% in

Search Strategy
Experts (n = 2) on novel fortification strategies were initially contacted to collect potentially relevant literature. The Pubmed, Web of Science, and Scopus databases were searched by using the following keyword search: ((fortif *) AND "iron" AND (Condiments [MesH] OR Seasoned OR Seasoning OR Bouillon * OR "Soy sauce *" OR "Fish sauce *" OR Powder * OR "fortified salt *")) NOT ("milk powder *" OR Noodle * OR "home fortification")). A language filter was applied to specify only English and Spanish studies. Other languages were not considered. Once articles were collected, literature was back-searched to identify additional references that met the inclusion criteria. Review articles were used for back-searching references.

Study Selection Criteria
Articles were assessed by title and abstract to include or exclude those based on the following criteria: Include articles that assessed efficacy, effectiveness, anemia prevalence, hemoglobin, or serum ferritin outcomes; studies assessing iron-fortified condiments (sauces, curry powders, bouillon cubes, salts, spices, seasonings, MSG); human studies; randomized or quasi-randomized controlled trials (either individual or cluster-level), cross-over trials, non-randomized controlled trials, clinical trials, community-based studies and prospective observational studies with a control group. Exclude articles not pertaining to the aforementioned fortification strategies; articles published prior to 1999; behavior change only outcome studies; animal studies; survey and consumption data, in vitro studies, or case studies. Articles were collected with publication dates in the last 20 years, January 1999-May 2019, to collect the most relevant and new information. All populations were considered, including children, women, iron-deficient populations, anemic populations, and general populations.

Data Extraction
Data were extracted from included articles into a Microsoft Excel table on the following characteristics: publication year/author/journal, study type, study aim, condiment type, iron source (and preparation of condiment), sample size, duration of trial, population (age, gender, physical status (pregnancy, severity of iron deficiency, infection), diet, geography (country, rural/urban) including setting of trial (school, home, refugee camp, etc.)), performance factors (education framework (if identified; i.e., Bronfenbrenner's ecological framework, the community-based participatory research (CBPR) framework, the social cognitive theory, theory of planned behavior, the health belief model, peer learning theory), communication with participants, behavior change component, policies/legislation, parasite treatment/no treatment, access/coverage, delivery mechanisms, monitoring scheme, adherence to treatment/insufficient or sufficient consumption of fortified condiment, sensory, monitoring compliance, waste, consumption drift, outcomes (anemia, serum ferritin, hemoglobin, hematocrit, mean corpuscular volume, total iron binding capacity, Creactive protein, serum iron, serum soluble transferrin receptor, morbidity outcome: diarrhea, number of days sick, mortality, any reported adverse effects). In the case of reviews or other article types not providing the relevant characteristics to extract, N/A was used as a placeholder.

Study Quality Assessment
The included studies were assessed for quality and risk of bias using the guidelines found in the 'Quality Criteria Checklist: Primary Research' of the Academy of Nutrition and Dietetics Evidence Analysis Manual [15]. This comprehensive tool was designed for food and nutrition interventions [16]. The Quality Criteria Checklist for primary research consists of 10 questions. Depending on the type of design, each question has several sub-questions (n = 2-8) to guide the rating. One author (JEA) assessed the quality of the articles at the study level on clarity of the research question, aspects of selection (random sequence generation and concealment of allocation methods) and comparability of study groups, retention rate (completeness of outcome data), detection (blinding of outcome assessors), performance (blinding of participants) and validity of outcome measures and any intervening factors, quality of statistical analysis and conclusions made with consideration of study limitations and likelihood of sponsorship bias. Any emerging conflicting assessments were discussed with an independent reviewer. Manuscripts were rated as positive (+), neutral (0), or negative (−). The quality assessment was not used to exclude studies from this review, but to support the discussion of performing indicators.

Data Synthesis
Given that the methods used for assessing change in hemoglobin varied across studies (i.e., control or baseline to endpoint), change in hemoglobin between baseline and endpoint were expressed as a standardized mean difference (effect size) to compare all studies based on the same measure. The effect size on hemoglobin levels was calculated using an online software tool based on the difference between the two means at the end of the intervention period, divided by the pooled estimate of standard deviation [17]. This helps calibrate the difference between the experimental and control groups in terms of the standard deviation. Effect sizes are used to compare studies to one another, but not to categorize studies based on effect sizes [18]. This is commonly conducted in meta-analysis reviews and reported as an odds ratio or relative risk [19,20]. The effect sizes can be deemed small (≤0.2), medium (≥0.2-<0.8), or large (≥0.8), based on previous cutoffs [20].

Study Selection
In the initial search of the literature (first phase), 955 studies were identified from Pubmed (n = 260), Web of Science (n = 339), and Scopus (n = 356) databases. After duplicates were removed, 541 studies remained. In the second phase, distillation was conducted by two authors (AWW and LAM) who reviewed the 541 studies by title and abstract and agreed to exclude 514 papers. After the distillation phase, 27 papers were reviewed full-text in the final third phase. After independent full-text review, the reviewers (AWW and LAM) met to discuss the exclusion of four articles under doubtful deliberation. A final agreement was reached based on the following reasons: food matrix not a condiment (n = 1), missing information (n = 1), no iron deficiency indicators as outcomes (n = 1), and duplicate study (n = 1). A total of 23 papers remained for the data extraction, study quality assessment, and data synthesis stages (see Figure 1).

Data Analysis
The selected twenty-three studies , two including subgroups of children and adults [22,42], were then evaluated for the impact of the iron fortification interventions on changes in hemoglobin, the prevalence of anemia and serum ferritin levels. An effort was made to identify negative and positive performing factors that may have influenced the outcomes. As presented in Table 1, the most tested condiment, as a vehicle for efficacy or effectiveness of iron fortification, was salt. This fortification approach included 13 Double-fortified Salt (DFS) [21][22][23][24][25][26][27][28][29][30][31][32][33] and five Multiple Micronutrient Fortified Salt (MMFS) studies [34][35][36][37][38]. Only one study evaluated a seasoning powder [39], two studies evaluated fish sauce [40,41], and two studies evaluated soy sauce [42,43]. The majority of studies evaluated the efficacy of the interventions, and only two [41,42] assessed effectiveness. There were no efficacy or effectiveness studies of iron-fortified bouillon cubes. All studies were conducted in Asia or Africa using as subjects children or women of reproductive age during different time lengths using as fortificants a variety of iron forms that included micronized ground ferric pyrophosphate [21,30,33,38], encapsulated ferrous fumarate [21,31], ferrous fumarate [22], ferrous sulfate [23,28], microencapsulated ferrous fumarate [24], unknown iron source [25][26][27], ferrous sulfate monohydrate chelated with malic acid and sodium hexametaphosphate [29], ferrous sulfate hydrate encapsulated with partially hydrogenated vegetable oil [32], chelated ferrous sulfate [34][35][36][37], H-reduced elemental iron encapsulated with partially hydrogenated vegetable oil [39], ferrous sulfate citrate [40], or NaFeEDTA [40][41][42][43]. Blood hemoglobin was the main hematological indicator in all studies to evaluate the effect of iron fortification. Some studies reported anemia prevalence and only a few analyzed serum ferritin (with or without correction for inflammation). The majority of DFS and MMFS studies assessed urinary iodine excretion (UIE) and verified that iron fortification did not jeopardize iodine fortification. Furthermore, in order to assure the success of the interventions, some studies included as part of the research protocol, the monitoring of potential performing factors such as an education component, assessment of behavioral changes, parasitic treatment, assessment and correction of micronutrient deficiencies, dietary intake, sensory and acceptability evaluation of the product and adherence to the programs. Sensory evaluation and product acceptability were of foremost importance, since some studies reported changes in the color of the DFS and of some foods when DFS was used for cooking [22].  The level of success of iron fortification for increasing hemoglobin, correcting anemia, or improving iron status was not the same for the different studies. Table 2 presents the distribution of responses in changes of hemoglobin, anemia prevalence and ferritin levels in children, women of reproductive age and the whole population at the end of the interventions. Out of 16 children studies, 13 showed increases in the levels of hemoglobin, and there was no change in three of them. In women, six out of eight studies resulted in an increase in hemoglobin and there was no change in one. Of the one study covering the whole population, only one experienced a significant increase in hemoglobin. Anemia in children was reported only by 12 studies, 10 of them decreased the prevalence of this condition and two reported no change. Out of seven studies in women that reported data on anemia, its prevalence decreased only in four studies and there was no change in three. In the one study covering the whole population, anemia was not assessed. Regarding serum ferritin as an indicator of iron deficiency, only 11 out of the 16 studies in children reported data on this parameter showing an increase in nine studies and not change in two of them. In women, four studies showed an increase in ferritin and there was no data in the remaining four. In the whole population, the study did not assess ferritin levels. A general observation in all these studies was that often hemoglobin and/or ferritin levels increased but the prevalence of anemia did not decrease.  1 Children studies include those exclusively assessing children, Chen [42] (disaggregates data based on gender and age), and Asibey-Berko [22] (assesses both women and children). 2 Women studies include those exclusively assessing women, Chen [42] (disaggregates data based on gender and age), and Asibey-Berko [22] (assesses both women and children). * Overall interaction was significant (due to negative changes in the control group) [37].
We then investigated, based on information provided by the authors, why nine of the studies in Table 2 did not experience changes in the right direction-that is, increasing hemoglobin, lowering the prevalence of anemia or increasing serum ferritin, in either children, women or in the whole population. Such studies, associated with potential negative performing factors that may have jeopardized reaching optimal outcomes, are presented in Table 3. Based on the comments made by the authors, there were organoleptic changes in two of the studies, consisting of a change in color of the iron-fortified salt or changes in the color of food when cooking using DFS as an ingredient [22,30]. It was believed that these color changes of the food vehicle decreased the acceptance, and thus the consumption of the fortified product impacting the outcome of the interventions. Two other studies were believed to be impacted negatively in the iron outcome due to an insufficient amount of iron in the fortified product to fulfill iron requirements, especially in women [25,39]. Two studies reported environmental conditions, such as the presence of malaria and parasitism, as factors influencing the outcomes negatively [27,30]. Two other reported deficiencies of micronutrients such as folate and vitamin B12 as possible causes for not attaining a significant reduction in anemia prevalence [24,30]. On the other hand, three of these studies did not reach statistical significance in one or more iron parameters possibly due to statistical issues such as an uneven balance in the prevalence of anemia between experimental and control group at baseline or increases on iron parameters in control groups at the end of the interventions [22,35,37].
There was no apparent association between the type of iron form used and the outcome in iron parameters, except when using FeNa-EDTA in soy and fish sauces, which showed a greater positive response of the intervention. Table 3. Low-performing studies of the efficacy of iron fortification of condiments. Impact on hemoglobin, anemia, and iron status. Notes: DFS-double-fortified salt; FePP-ferric pyrophosphate; FF-ferrous fumarate; EFF-encapsulated FF; FS-ferrous sulfate; Hb-hemoglobin; ID-iron deficiency; mo-month; NA-data or information not available or unknown; NP-non-pregnant; NL-non-lactating; RCT-randomized control trial; wk-week; WRA-women of reproductive age; y-year. * Indicates statistically significant outcome.

Data Synthesis
Effects sizes were estimated for each study, disaggregated by gender, age, iron source, and/or iron concentration when possible (Figure 2). Effect sizes were considered negligible if the standard deviation crossed the 0 x-axis, which was seen in nine studies [22,24,26,27,31,34,36,39,42].
Interestingly, Rajagopalan et al. [26] demonstrated a negligible effect size for men, but a positive effect size for women, indicating that double-fortified salt may be a more effective treatment for women than men. DFS fortified with micronized ferric pyrophosphate showed the highest effect size for this condiment at 1.4 [33], indicating that this iron source may contribute to its greater effect sizes over the other iron sources used for these studies. Seven of the 16 DFS data points (44%) showed negligible effect sizes. Organoleptic changes of the DFS could be associated with weak effect sizes.
Both studies that assessed fish sauce showed positive effect sizes. Interestingly, similar effect sizes were seen for fortified fish sauce using either NaFeEDTA or FeSO4 + citrate, indicating that either iron source can be used for positive effects in fish sauce [40,41].
The fortification of soy sauce showed the highest effect sizes [42,43]; all of the age/gender groups showed considerable effect sizes, with the exception of women 19-30 years of age [42]. Nonetheless, even at a low concentration of NaFeEDTA, the large effect size of 2.5 indicates the strength of this condiment as a fortification vehicle [43].
Seasoning powder demonstrated a negligible effect size (n = 1). However, more studies will need to assess this vehicle as a fortification vehicle before complete conclusions can be drawn [39].
Adherence and/or compliance was addressed by 15 studies, having an average effect size of 0.52 and those that did not address adherence or compliance (n = 7) had an average effect size of 0.67. This relative difference is not large, indicating a negligible effect between study outcomes that did or did not address compliance and adherence.

Discussion
Our review of studies conducted in the last 20 years revealed that during this period, there were 21 studies that assessed the efficacy of iron fortification of condiments in children and adults and only two studies evaluated the effectiveness of this type of intervention on the whole population [41,42]. Most of these studies used salt fortified with iodine and iron or salt fortified with multiple micronutrients. Only a limited number of studies were conducted using other vehicles such as soy and fish sauces. Various forms of iron had been used with the objectives of increasing bioavailability and maintaining the quality and stability of the fortified product, the iron compound and the levels of iodine in salt. Iron forms used included chelated ferrous sulfate, ferrous fumarate, iron pyrophosphate, and in some instances in microencapsulated forms to prevent unwanted interactions between iron and iodine that could jeopardize product quality and organoleptic properties of the fortified product. There has been no evaluation of bouillon cubes and there was only one assessment of a fortified powder when used as a condiment in food preparation. All studies were conducted in Asia or Africa, where the potential of fortifying condiments seems more promising due to culture and tradition. The main iron indicators monitored were the levels of hemoglobin, the prevalence of anemia and in some studies, the levels of serum ferritin.
As shown in Table 2, there was a variety of responses in iron indicators as a result of the iron fortification intervention in children, women, and the whole population. In some cases, the level of hemoglobin did not increase, and in others, the level of hemoglobin increased significantly but the prevalence of anemia did not decrease. Likewise, serum ferritin, as an indicator of iron deficiency, did not always increase. Interventions with children showed the most consistent positive efficacy result. In these studies, 13 out of 16 studies increased the levels of hemoglobin and nine out of 11 studies that measured ferritin showed a significant increase in this indicator. This observation was particularly more evident in children studied and maintained under confined conditions like schools or childcare centers, where monitoring and compliance are easier to control. While it seems that adherence had no effect on the study effect sizes, whether or not the population was iron-deficient at baseline showed an increase in the effect size. Thus, it is hypothesized that a large-scale effectiveness trial using iron-fortified condiments without monitoring adherence and compliance may not see a difference in results than its monitored counterpart. However, further studies are needed to prove this point. Additionally, it should be noted that iron-fortified condiments targeting a specific irondeficient subgroup may see more preferable results than the general population. Policy makers should understand the risks of non-targeted individuals being exposed to iron fortification during targeted trials, specifically those individuals that might possess the thalassemia trait in which iron overload is a risk. We identified nine studies in which one or more iron parameters did not change in the positive direction as a result of the iron interventions, either not increasing hemoglobin significantly, not lowering the prevalence of anemia, or not increasing serum ferritin (Table 3). Upon analysis, several of these studies using salt as a vehicle reported lower acceptability of the fortified product due to organoleptic changes. Those reporting changes, made particularly by women, included a different color appearance of the fortified salt or changes in the color of the food when cooking using fortified salt as an ingredient. Organoleptic changes of DFS, particularly color changes of the fortified product, have been reported by several investigators [22,32], and efforts are being made to correct or at least alleviate this unwanted phenomenon [11]. Visual and flavor modification instead influences the acceptability of the product with a consequential decrease in consumption of the fortified product lowering the iron intake. In addition, two of these nine low-performing studies reported an insufficient amount of iron in the fortified product to meet iron requirements as a cause of not improving iron deficiency. This observation coincides with the potential decrease in the consumption of fortified salt due to organoleptic changes that would also lead to lower iron intake. The need for adding meaningful amounts of fortificants to staple foods and their adequate consumption to fulfill dietary gaps has been demonstrated in successful fortification programs in Latin America, including fortifying foods with iron [44]. In some of these studies in Table 3, negative performing factors reported by the authors were also deficiencies of other micronutrients important in anemia, such as folate and vitamin B12. The lack of folate and vitamin B12 impairs the production of red blood cells and therefore the persistence of anemia despite adequate iron nutrition [45]. Furthermore, as suggested by two of these studies [27,30], environmental factors such as the presence of high parasite burden, particularly from hookworms, as well as malaria, can also contribute to the development of anemia by producing intestinal iron losses and red cell hemolysis, respectively [46]. One systematic review also found that iron supplementation, in addition to deworming of schoolchildren, showed greater changes in hemoglobin and reduction of anemia than studies that just dewormed (n = 8 studies) [47]. An interesting case was seen in one of the included studies, Reddy et al. (2014) [27] where anemia prevalence significantly decreased (−6.3%) among participants consuming DFS and who were dewormed at baseline, but did not significantly decrease (+1.5%) among those who only consumed DFS and were not dewormed at baseline. Nonetheless, this finding is contradictory to a systematic review and meta-analysis conducted by Ramírez-Luzuriaga et al., in which they found deworming at baseline not to be a significant predictor of the risk of anemia and hemoglobin concentration response to DFS interventions [14], and thus requires further investigation.
We then wanted to examine the other end of the spectrum, the most successful iron fortification interventions, as detailed in Table 4. That is, those studies that showed changes in all three iron indicators, (increased hemoglobin, decreased anemia prevalence and increased serum ferritin) in the right direction indicating clear efficacy of the iron fortification. We wanted to identify what was done in these particular studies associated with their success, particularly different or contrasting activities to the ones conducted in the less successful studies shown in Table 3. It was revealed that out of eight studies in this category, we found that five were carefully monitored in all aspects of the intervention, particularly the sensory characteristics and acceptability of the fortified product. This was especially important considering the changes in color in iron-fortified salt and when salt is used for cooking. The studies included in the present work using fortified soy and fish sauce did not report any taste acceptability issues. This observation supports the argument that changes in color and/or flavor of the fortified product are critical determinants of acceptability. Due to their distinctive dark brown color and strong flavor, the sauces used in these investigations masked any potential changes in color that could be produced by the addition of iron. In support of this notion, a study conducted in Thailand found good similar sensory acceptability when fortifying soy sauce with NaFeEDTA [48]. Although a metallic taste has been associated in some instances with the fortification of soy sauce with NaFeEDTA, in a systematic review by Huo et al. (2002), only one of 16 studies conducted in China assessing NaFeEDTA-fortified soy sauce reported negative taste effects [49]. In addition, the authors showed a pooled weighted mean difference in hemoglobin of 0.88 g/dL and a decrease in anemia prevalence of 25%, indicating a clear efficacy of these types of interventions. These findings illustrate the significant potential of reducing iron-deficiency anemia prevalence within an appropriate implementation context for condiments.
Five of these studies also monitored and corrected as needed, micronutrient deficiencies, in particular vitamin A. The positive role of vitamin A in iron utilization has been documented since 1998 [50]. Only three of these studies conducted parasitic treatments of the subjects previous to the intervention. While all studies presented in Table 4 demonstrate positive effects in the three iron indicators extracted (hemoglobin, anemia prevalence, and serum ferritin), it should be noted that hemoglobin and serum ferritin are the best indicators of iron deficiency. Anemia prevalence can be an indicator of iron deficiency or other external factors such as other micronutrient deficiencies or parasitic infections. Thus, when analyzing randomized controlled trials, an outcome of lowered anemia prevalence can be attributed to a holistic treatment within the study design.
It needs to be indicated that the remaining six studies that did not conform to the low-or highperforming criteria [23,26,29,34,40,43] showed increases in hemoglobin but did not assess or report anemia, ferritin or both. Therefore, such studies were not classified as high performers based only on the criteria of lacking positive changes in the three parameters. Notes: DFS-double-fortified salt; FePP-ferric pyrophosphate; FF-ferrous fumarate; EFF-encapsulated FF; FS-ferrous sulfate; Hb-hemoglobin; ID-iron deficiency; mo-month; NA-data or information not available or unknown; NP-non-pregnant; NL-non-lactating; RCT-randomized control trial; wk-week; WRA-women of reproductive age; y-year. * Indicates statistically significant outcome.
The purpose of the quality evaluation was to assess the internal validity of the studies, as well as to appraise the quality of the study design. This quality assessment was not used to exclude studies, but to add a layer of internal validity to the already extracted data. The degree of internal validity of study design is used in order to warn the validity of successful trials when it is presented that internal validity is also threatened. For example, Huo et al. showed the highest effect size of all 22 studies included for the effect size calculation (2.88 and 2.49, high and low NaFeEDTA, respectively) [43]. This iron fortificant is known for its higher bioavailability and limited effect on the organoleptic properties of food [51]. Despite these findings, this study [43] was rated as moderate in the quality assessment due to a lack of blinding to prevent bias and inadequate description of statistical analyses. Therefore, its results must be cautiously analyzed, given these reasons for doubting its internal validity. Similarly, Reddy et al. (2016) showed a large positive effect size (0.81) using ferrous sulfate fortified DFS with pregnant women. However, blinding and statistical analyses (e.g., power calculations, ANOVA assumptions) were not used or described in the study design [28]. Blinding is a critical component in randomized, placebo-controlled trials [52]. In the case of those studies using salt that is visibly different from the control, any attempt to blinding will be difficult to accomplish. Thus, it is important that the investigator acknowledges this as a limitation. Many of the studies included in this review did not use blinding, and some that reported using it did not indicate how blinding was operationalized. These examples warrant caution when extrapolating results that might support policy decisions.
It is also important to mention the content of sodium that participants were consuming in the condiment fortification trials. In the salt trials, participants consumed an average of 10 g of salt per day. In the soy sauce and fish sauce trials, participants were also consuming 10 mL of sauce per day. Sodium content in soy sauce ranges between 1.9 and 5 g/100 g [53]. Sodium content in the fish sauce varies, but one study reported 3672 ± 580 mg/100 g [54]. The US Adequate Intake for sodium is 1.5 g/day for healthy individuals aged 9-50 years and 1-1.2 g/day for children 1 to 8 years [55]. The level of salt consumption as found by these evaluated fortification studies may lead to excessive amounts of sodium intake, not aligning with the WHO's non-communicable disease-related target (<5 g salt per day) of reducing salt/sodium intake from current consumption levels by 30% [56]. Nonetheless, no adverse events, related or not to the level of salt intake, were reported by the only one of all reviewed studies who monitored adverse events in the population under investigation [42]. Therefore, the authors recommend the collection and reporting of adverse events within future fortified condiment trials.
Sodium intake continues to be high throughout most of the world, leading to increasingly high rates of hypertension [57]. As such, in populations where the prevalence of hypertension is dangerously high, condiments that are excessive in sodium should be cautioned when used as vehicles in fortification programs. Education and proper communication of benefits and associated risks are critical in program design and implementation.
This review has several limitations that require disclosure. A large proportion of the included studies were on the ability of DFS to address low hemoglobin or anemia, especially in India. A smaller proportion of included studies consisted of other vehicles such as sauces and seasoning powders. Most of the studies included were from India (n = 13) and African countries (Ghana, Morocco, and Cote d'Ivoire), but none from the Americas, Oceania, or Europe; thus, compromising the generalizability of the results to the whole Sub-Saharan Africa and other continents. Most of the studies included in this review were RCTs. Nonetheless, most studies lacked information on program implementation such as randomization, blinding, monitoring, and formative evaluation, which adds to the problem of abstracting and describing poor or good performing factors. Finally, this review included studies only published in English. This is a known limitation for most systematic reviews. However, we considered the findings of articles in other languages as these were included in other systematic reviews, such as in the case of using NaFeEDTA to fortify soy sauce as reported by Huo et al. [49], which included 16 articles, the majority in Chinese.

Conclusions
The present review systematically collected and assessed the efficacy of 23 condiment-based iron-fortification trials conducted in the last 20 years and their associated performance factors. The internal validity of the study designs was assessed via a quality assessment, and effect sizes were calculated. Overall, it was found that effective condiment vehicles for fortification are those that can mask the organoleptic changes due to the physical characteristics of the fortified food, such as fish or soy sauces. These condiments also mask changes in taste due to their inherent strong flavor. The efficacy of fortified salt is well proven and can help to address iron deficiency. However, it is best when using an iron source that does not interfere with other fortificants (e.g., promote oxidation) and does not produce adverse organoleptic changes, as is the case of micronized ferric pyrophosphate.
For best results, programs should rigorously address monitoring compliance and adherence. Furthermore, deworming is necessary to enhance the efficacy of programs implemented in areas with a high parasite burden. Studies are needed to determine the efficacy and effectiveness of bouillon cubes (n = 0 studies), curry powders (n = 0 studies), and seasoning powders (n = 1 study) before conclusions can be drawn for these vehicles. Special attention should be given to those studies that show positive effect sizes as well as a strong internal validity of study design, and caution to those studies that show positive effect sizes but with moderate or weak study design. This review can be used to advise policy makers and those in decision-making positions on best practices and protocols for condiment fortification programs. If implemented correctly, fortifying condiments is a potential nutrition-specific strategy for governments and NGOs to address iron deficiency in at-risk populations.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1. Table S1: Quality assessment of articles included in this review.