Higher dietary acid load potentially increases serum triglyceride and obesity prevalence in adults: An updated systematic review and meta-analysis

Background In the current meta-analysis, we aimed to systematically review and summarize the eligible studies evaluating the association between dietary acid load in terms of potential renal acid load (PRAL) and net-endogenous acid production (NEAP) with anthropometric parameters and serum lipids in adult population. Methods In a systematic search from PubMed, Scopus, Web of Sciences and Cochrane electronic databases up to December 2018, relevant studies were included. Cross-sectional, case control or cohort studies evaluating the association between PRAL and NEAP with the mean values of body mass index (BMI), waist circumference (WC), low and high density lipoprotein cholesterol (LDL, HDL), triglyceride (TG), total cholesterol (TC) and the prevalence of obesity were included. Results According to our results, having higher dietary acid load content in terms of high PRAL scores was associated with higher triglyceride concentrations (weighted mean difference (WMD): 3.468; confidence interval (CI): -0.231, 7.166, P = 0.04) and higher obesity prevalence (30% and 27% in highest versus lowest categories). Accordingly, being in the highest category of NEAP was associated with higher prevalence of obesity (25% and 22% in highest versus lowest category). In subgroup analysis, higher PRAL scores was associated with higher BMI in women (WMD: 0.122; CI: -0.001, 0.245; P = 0.049) and higher NEAP in men (WMD: 0.890; CI: 0.430, 1.350; P < 0.001). There was no association between dietary acid load and other studied parameters. Conclusions In the current meta-analysis, high dietary acid load content was associated with higher serum triglyceride concentrations and higher obesity prevalence. Reducing dietary acid load content might be a useful preventive strategy against obesity and metabolic disorders.


Methods
In a systematic search from PubMed, Scopus, Web of Sciences and Cochrane electronic databases up to December 2018, relevant studies were included. Cross-sectional, case control or cohort studies evaluating the association between PRAL and NEAP with the mean values of body mass index (BMI), waist circumference (WC), low and high density lipoprotein cholesterol (LDL, HDL), triglyceride (TG), total cholesterol (TC) and the prevalence of obesity were included.

Results
According to our results, having higher dietary acid load content in terms of high PRAL scores was associated with higher triglyceride concentrations (weighted mean difference (WMD): 3.468; confidence interval (CI): -0.231, 7.166, P = 0.04) and higher obesity prevalence (30% and 27% in highest versus lowest categories). Accordingly, being in the highest category of NEAP was associated with higher prevalence of obesity (25% and 22% in highest versus lowest category). In subgroup analysis, higher PRAL scores was associated with higher BMI in women (WMD: 0.122; CI: -0.001, 0.245; P = 0.049) and higher NEAP in men (WMD: 0.890; CI: 0.430, 1.350; P < 0.001). There was no association between dietary acid load and other studied parameters. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Introduction indices, prevalence and serum lipids status. Therefore, in the current meta-analysis we summarize the results of studies evaluated the association between PRAL of NEAP with general or central obesity indices (e.g. BMI, WC, WHR), obesity prevalence, serum lipids including TG, TC, LDL, HDL in an updated systematic review and meta-analysis.

Search strategy
We performed a systematic search using PubMed, Scopus, Web of Sciences and Cochrane electronic databases to the studies evaluated the association between dietary acid load and general or central obesity, serum lipids and metabolic syndrome up to December 2018. No language restriction was applied. Moreover, hand-searching from reference lists of all relevant papers, previous reviews and meta-analyses was performed to cover all relevant publications. Strategy search was created using a combination of the MeSH (Medical Subject Headings) terms from the PubMed database and free text words were used. For each electronic database, search strategy was adopted. The PICO (patients, intervention, comparator and outcome) for studies' selection is presented in Table 1. We used PICO model because it is one of the most widely used models for formulating clinical questions. The PICO model is one of the frequently used tools for structuring clinical research questions in connection with evidence syntheses. The Cochrane Handbook for Systematic Reviews of Interventions specifies using PICO as a model for developing a review question, thus ensuring that the relevant components of the question are well defined [23,24].

Selection and characteristics of the included studies
Our search obtained 646 potentially relevant articles from PubMed, Scopus, Cochrane and Google Scholar electronic databases. Thereafter 153 manuscripts were remained for full text screening after duplicate remove and exclusion after title and abstract reading. Totally, 121 manuscripts were excluded because of their irrelevant subject, inappropriate design, being reviews including meta-analysis or systematic reviews, conferences and seminars, not relevant age groups, not evaluating the association of studied parameters (dietary acid load, obesity, lipids and metabolic syndrome) or not measuring the routine dietary acid load. Accordingly 32 manuscripts were included in the systematic review. The Flow diagram of study screening and selection process is presented in Fig 1.

Inclusion criteria
In the current systematic review and meta-analysis, observational studies with the design of cross-sectional, case control or cohort evaluating the association between dietary acid load and BMI, WC, WHR, obesity, central obesity, lipid profile and metabolic syndrome were included. According to our set of parameters, we conducted numerous meta-analyses. Dietary acid load- Observational studies with the design of cross-sectional, case control or cohort obesity or metabolic syndrome meta-analysis included the studies evaluated the odds ratio (OR), relative risk (RR) or prevalence of obesity or metabolic syndrome in the lowest versus highest dietary acid load categories. Accordingly, in dietary acid load-body mass index, dietary acid load-waist circumference or dietary acid load-serum lipids meta-analyses, the study must have reported the mean ± SD of body mass index or waist circumference or waist to hip ratio or serum lipids including triglyceride, cholesterol, low-density lipoprotein cholesterol, high density lipoprotein cholesterol in subjects in the highest versus lowest dietary acid load category as the reference group. For the search purpose, we used MESH (Medical Subject Heading) and non-MESH keywords including the following: ("dietary acid load" OR "dietary acid-based load") AND ("body mass index" OR "BMI" OR "obesity" OR "central obesity" OR "serum lipids" OR "lipid profile" OR "triglyceride" OR "cholesterol" OR "LDL-cholesterol "low density lipoprotein cholesterol" OR "HDL" OR "high density lipoprotein cholesterol" OR "hypertension" OR "cardiovascular risk factors" OR "cardiometabolic risk factors" OR "metabolic syndrome" OR "diabetes". The reviewed literatures were inserted into the EndNote  software (version X8, for Windows, Thomson Reuters, Philadelphia, PA, USA). Consequently retrieved citations were merged, duplications were eliminated and the review process has been facilitated. Accordingly, titles and abstracts of all articles had been evaluated independently by three reviewers (MAF, LN, ZN). Articles not meeting the eligibility criteria were excluded. Moreover, the reference lists of relevant review article were also evaluated to include additional studies. Full-texts of relevant articles were retrieved if meeting the eligibility criteria, and wee re-evaluated. Any disagreements were discussed and resolved by consensus.

Quality assessment
The methodological quality assessment of the included papers was performed by a nine-star Newcastle-Ottawa scale (NOS) for quality assessment of the cross-sectional, case-control and cohort studies. The 9-point NOS scale has scoring ranges from 0 to nine and is categorized into selection, comparability, and ascertaining of outcome. Studies with equal or more than 7 stars were categorized as high quality [25].

Data collection and extraction
Data were collected according to a standard data extraction form gathering the information about the study characteristics including information about authors name, publication year, geographical area, study design; information about the population including participants age range, mean age of case and control group, number of case and controls, dietary assessment tool, setting, gender and the sample size and information about the adjusting for possible confounders the main findings and estimates of associations.

Data synthesis and analysis
In the current meta-analysis, three meta-analysis approaches were used: the association between odds of obesity or metabolic syndrome and dietary acid load markers was analyzed by estimating the ORs and 95% confidence intervals by calculating the Ln of ORs and its standard error of mean (s.e.) as the effect size of the meta-analysis. Pooled OR [and 95% confidence interval (CI)] was estimated using a weighted random-effect model (the DerSimonian-Laird approach). The comparison of the continuous variables including BMI, WC, WHR, TG, TC, LDL, HDL between highest versus lowest category of dietary acid load as reference group was performed by measuring the unstandardized mean differences as the effect size calculated by pooled estimate of weighted mean difference (WMD) with 95% confidence interval (CI), and the fixed effects and random effects models.
The prevalence of obesity in highest versus lowest dietary acid load categories was performed by re-calculating the proportions of interest from the relevant numerator and denominator. The overall proportions of interest were derived using meta-analysis techniques by metaprop command in the STATA and presented along with 95% confidence intervals (95% CI) calculated using a normal approximation. Cochran's Q test and I squared test was used to identify between-study heterogeneity; I 2 < 25%, no heterogeneity; I 2 = 25-50%, moderate heterogeneity; I 2 > 50% large heterogeneity [26]. The heterogeneity was considered significant if either the Q statistic had p < 0.1 or I 2 > 50%. Sensitivity analysis was used to explore the extent to which inferences might depend on a particular study or a number of publications. Subgroup analysis was performed to identify possible sources of heterogeneity, if required. Begg's funnel plots was assessed to evaluate the publication bias followed by the Egger's regression asymmetry test and Begg's adjusted rank correlation for formal statistical assessment of funnel plot asymmetry. The data were analyzed using STATA version 13 (STATA Corp, College Station, TX, USA), and P-values less than 0.05 were considered as statistically significant.

Description of the studies reported the dietary acid load as PRAL and NEAP with general and central obesity associations
From all of the systematically reviewed relevant papers (Table 2), totally 29 studies reported the association between PRAL and NEAP with obesity indices including BMI, WC, or the prevalence of general and central obesity [4-7, 11, 14-20, 27-43]. Twelve studies reported higher BMI or WC or the prevalence of general or central obesity in highest versus lowest    [32]. In other study by Kucharska AM et al [35], BMI, WC and the prevalence of overweight or obesity in highest NEAP tertile were higher than the lowest (P and P trend < 0.05). While in PRAL categories, among men, no significant difference in the BMI and WC in the lowest versus highest PRAL tertiles was observed. The prevalence of overweight

Description of the studies reported the dietary acid load as PRAL and NEAP with serum lipids and CVD risk factors associations
From all of the systematically reviewed papers (Table 3) totally, seventeen studies reported the associations between PRAL and NEAP with cardiovascular disease (CVD) and serum lipids [4, 7, 11, 17, 19, 21, 28, 29, 31-37, 43, 44]. Higher concentrations of serum lipids including TG, LDL and lower HDL concentrations and higher prevalence of hypercholesterolemia in the highest versus lowest PRAL or NEAP categories was reported in six studies [7,11,17,32,35,37]. In the study by Kiefte-de Jong JC et al among three cohort of NHS, NHS-II and HPFS, higher prevalence of hypercholesterolemia in the highest versus lowest quintiles of PRAL and NEAP was in both men and women was reported in all of three cohorts [32]. In other study by Kucharska AM [35] examining the association between dietary acid load and cardio-metabolic risk factors among polish adults, only serum TG tended to increase across tertiels of NEAP only among men; while no significant difference among women or across PRAL tertiles were observed [35]. On the other hand, three studies reported lower TC or higher HDL concentrations in highest versus lowest categories of PRAL or NEAP [21,34,35]. Other studies reported no significant difference between lipids across NEAP or PRAL categories [11,28,31,33,35]. The prevalence of CVD was reported among six studies [4,19,29,31,35,36] while one study reporting an inverse association between the prevalence of CVD and NEAP quartile scores [4] and others reported no association. The prevalence of metabolic syndrome or the odds of it has also been reported in three studies [7,11,28]; Bahadoran et al [28] reported no difference in the prevalence of metabolic syndrome across quartiles of PRAL. In a cross-sectional

Findings from meta-analysis of the prevalence of general obesity, central obesity and difference in BMI across lowest and highest dietary acid load categories
The Forest plot of the studies included in the PRAL, NEAP and BMI meta-analysis are presented in Fig 2. No significant association between BMI and dietary acid load indices was Dietary acid load, lipids, obesity   [20] in men led to significance change in the effect size (WMD: 0.230; CI: 0.105, 0.354; P<0.001).
The results of subgroup analysis for the association between BMI with PRAL and NEAP are presented in Tables 4 and 5. Accordingly, for PRAL-BMI associations, country, dietary assessment tool and sample size are the possible source of heterogeneity whereas, in NEAP-BMI associations, country and dietary assessment tool are the source of heterogeneity. Subgroup analysis also revealed gender difference in the associations between PRAL, NEAP and BMI (Table 4). Higher PRAL scores were associated with increased BMI in females (WMD: 0.122; CI: -0.001, 0.245; P = 0.049) and not in men. While, NEAP and BMI association was significant only among men (WMD: 0.890; CI: 0.430, 1.350; P < 0.001). Dietary acid load, lipids, obesity Table 7. Results of subgroup analyses of the association between mean difference in TC and PRAL according to study and participants' characteristics.

Group
No. of studies WMD (95% CI) P within group P between group P heterogeneity I 2 , % Studies eligible for inclusion in the systematic review and meta-analysis https://doi.org/10.1371/journal.pone.0216547.t007 Table 8. Results of subgroup analyses of the association between mean difference in TC and NEAP according to study and participants' characteristics.

Group
No. of studies WMD (95% CI) P within group P between group P heterogeneity I 2 , % Totally, five studies were reported the prevalence of obesity in different PRAL categories [14,17,31,35,38] and the Forest plot is presented in Fig 3. The prevalence of obesity in the highest PRAL category was 30% (CI: 0.29-0.31) and in the lowest category was 27% (CI: 0.26-0.27) with no evidence of heterogeneity. Moreover, three studies were also reported the prevalence of obesity in different NEAP categories [4,14,35] and the Forest plot (Fig 4) indicates that the prevalence of obesity in the highest NEAP category was 25% (CI: 0.24-0.24) and in the lowest category was 22% (CI: 0.22-0.23) with no evidence of heterogeneity. As shown, the higher prevalence of obesity in the highest scores of PRAL and NEAP had been reported. The prevalence of central obesity was identified in only two studies [17,28] and therefore, no meta-analysis was performed. The association between waist circumferences as an indicator of central obesity with PRAL was reported in seven studies and the Forest plot (Fig 5) revealed no association (WMD: -0.021; CI:-1.422, 1.38, P = 0.977) with a great heterogeneity (Heterogeneity chi-squared = 2079.18 (d.f. = 5), P < 0.001; I 2 = 99.8%; Tau 2 = 2.97). In sensitivity analysis Dietary acid load, lipids, obesity excluding the study by Wynn et al [41] revealed a significant association (WMD: 1.036; CI: 0.185, 1.886; P = 0.017). Table 6 presents the possible effects of country and continent on the Table 9. Results of subgroup analyses of the association between mean difference in TG and PRAL according to study and participants' characteristics.

Group
No. of studies WMD (95% CI) P within group P between group P heterogeneity I 2 , % heterogeneity. Among the reviewed studies only two studies reported the WC-NEAP associations and therefore were excluded from the meta-analysis [35,37].

Findings from meta-analysis of mean TC, HDL-C, LDL-C, TG across different categories of dietary acid load
The Forest plot of the effect of dietary acid load on serum TC (Fig 6) shows that no association between PRAL and NEAL with TC is present ( Table 11).

Discussion
In the current meta-analysis, we summarized the results of studies reporting the association between PRAL, NEAP and body mass index, waist circumference, lipid profile and the prevalence of obesity. Accordingly, being in the highest category of PRAL scores was associated with higher TG and higher prevalence of obesity compared with lowest category. No association between BMI, WC and other serum lipids with PRAL or NEAP was observed. In subgroup analysis, increased PRAL scores in women and increased NEAP scores in men were associated Dietary acid load, lipids, obesity with higher BMI. Animal foods including meat, fish, egg, chicken, cheese and also cereals are rich in sulfur containing amino acids, phosphorous and chloride are potentially acid formers; while vegetables and fruits high in malate, citrate and glutamate are potentially base formers therefore, animal based-foods and high contents in western diets are potentially considered as most important acid-producer diets and are associated with higher risk of insulin resistance, high blood pressure and diabetes as established in numerous works [10]. Accordingly, western dietary pattern with high dietary acid load content, is a potent inducer of central obesity and metabolic syndrome; several studies had revealed significant relationships between western dietary pattern and the increased risk of MetS, cholesterol and increased waist circumference and BMI. Accordingly, western dietary pattern with high content of red meat, eggs, and refined grains is associated with increased risk of obesity and increased levels of blood sugar, systolic blood pressure, triglycerides, and reduced levels of HDL [45][46][47]. Higher prevalence of obesity in higher categories of PRAL and NEAP could also be a attributed to the possible adipogenic effects of higher dietary acid load; in the study by Li et al. [48] among 29520 Chinese adults aged 18-70 years and higher prevalence of obesity in higher versus lower quintiles of PRAL was observed. Although we did not observed association between PRAL and NEAP with BMI in total analysis, however, in subgroup analysis of men and women separately, the PRAL-BMI association was significant for women and the NEAP-BMI association was significant for men. These gender-specific results might be due to the difference in the lean body mass; it has been demonstrate that higher dietary acid load reduces lean body mass only among women and not in men and finally leads to higher body fat synthesis [15] and that more alkalinogenic diets are associated with greater skeletal muscle mass among women [40]. Another possible explanation is the difference in sex-hormones affecting acid-base balance [49]. Acidosis leads to loss of muscle mass through reducing protein synthesis and increasing proteolysis and amino acid oxidation, mediated via the ubiquitin proteasome system or via in IGF-1 signaling alterations [50]. This impaired acid-base balance is possibly the reason of reduced calcaneal broadband ultrasound attenuation and bone density in women and not in men [39]. In other word, it will be better to evaluate the association between dietary acid load indices and fat mass or fat free mass as indicators of adiposity instead of BMI to better elucidate the obesity-dietary acid load associations. In the current meta-analysis we also observed a positive association between PRAL and TG. The underlying mechanisms of increased TG concentrations in higher scores of PRAL are not well elucidated; however, several proposed mechanisms might be raised cortisol secretion and reduced insulin sensitivity and secretion and their consequent lipid disorders [35]. As mentioned in the results section, dietary assessment tool and continent could be a source of heterogeneity among observed association. In the current meta-analysis, PRAL and NEAP calculation was based on self-reported data gathered by 24 hours recall method, 24 hours record method and food frequency questionnaire which may be a potential source of bias. Moreover, difference in the items of the FFQ might be a source of heterogeneity; as described previously, the FFQ items ranged from 63 to 168 items and the local foods in the FFQ could also affect the heterogeneity [51], although, almost all of the included studies used valid and reliable FFQs. FFQ covers a wide range of dietary ingredients and is more accurate than 24-hours recall method reflecting usual dietary intake in a short period of time; it has been confirmed that FFQ could be more helpful in evaluating the dietdisease relationships [52]. Another source of heterogeneity, the continent, presents the possible role of geographical distribution, genetic background and cultural factors influencing the association between dietary acid load and metabolic risk factors [53][54][55][56][57]. The current metaanalysis has several limitations and strengths; the current meta-analysis included the results of observational studies with the cross-sectional or cohort design which makes the causal inference impossible; although, the studies were large population-based studies with acceptable quality. Moreover, the PRAL and NEAP were calculated based on self-reported data gathered by 24 hours recall method, food record or food frequency questionnaire which might be potential sources of bias. However, our study, based on our knowledge, is the first meta-analysis evaluating the association between dietary acid load as both PRAL and NEAP scores with a wide range of obesity related parameters including BMI, WC, LDL, HDL, TG, TC and the prevalence of obesity. In conclusion, in the current meta-analysis, we found a positive association between TG and PRAL and a gender-specific associations between PRAL, NEAP and BMI while no association between dietary acid load and other parameters were reported.