Dietary acid load and its association with all-cause mortality, cardiovascular mortality, and events

Abstract

This prospective cohort study examined the association between dietary acid load and cardiovascular disease (CVD) incidence, CVD mortality, and all-cause mortality in an Iranian population. A total of 6,482 adults (59.83% female) aged 35–65 years from the Mashhad Stroke and Heart Atherosclerosis Disorder (MASHAD) study were followed for ten years. Dietary acid load was assessed using net endogenous acid production (NEAP), potential renal acid load (PRAL), and dietary acid load (DAL). Multivariable Cox proportional hazards regression and nonlinear restricted cubic spline models were used. Compared to the first quartile, PRAL was associated with higher CVD mortality in the second (OR: 1.974; 95% CI 1.138–3.424) and third quartiles (OR: 2.323; 95% CI 1.397–3.862) while the highest quartile showed no significant association (OR: 1.335; 95% CI 0.724–2.460). DAL showed similar associations in the second (OR: 1.768; 95% CI 1.046–2.991) and third quartiles (OR: 1.709; 95% CI 1.010–2.894), but not the fourth one (OR: 1.673; 95% CI 0.971–2.883). NEAP was associated with CVD mortality only in the second quartile (OR: 1.802; 95% CI 1.083–2.997). No significant associations were found for all-cause mortality or CVD incidence. Cubic models confirmed a nonlinear relationship for PRAL (P = 0.004) and mid-level associations for DAL (P = 0.009). This nonlinear pattern indicates that the risk of CVD mortality peaks at moderate dietary acid load levels, possibly reflecting adverse metabolic effects of both high acid and high alkaline diets. Overall, these findings suggest that moderate dietary acid load may increase CVD mortality risk, requiring further investigation.

Introduction

According to the World Health Organization (WHO), cardiovascular diseases (CVDs), diabetes, respiratory diseases and some cancers, which are the leading causes of mortality and premature deaths, account for 74% of all deaths globally and affect 41 million people annually^1–5. It is estimated that by 2030, CVDs are responsible for more than 23 million deaths worldwide⁶. In Iran, CVD is the major burden of NCDs⁷. An important modifiable lifestyle risk factor for CVD is an unhealthy diet, and alterations in the diet can reduce the risk of NCDs and global mortality^1,2,4,8.

The individual’s acid–base equilibrium can be affected by dietary intake and the micronutrient profile which is essential to human health^1,4. The food components that are responsible for the release of acid precursors after metabolism include: proteins and phosphates and the alkaline precursors nutrients are potassium, magnesium and calcium⁹. A high consumption of acidogenic foods, that include meat, fish, eggs, whole grains and cheese, can induce endogenous acid production and an elevated dietary acid load^1,9,10. An individual’s dietary acid load can be calculated using net acid production (NEAP), potential renal acid load (PRAL), and dietary acid load (DAL) Eqs. ¹¹. Acid-rich diets, leading to a mild state of metabolic acidosis, have been associated with several health implications. Along with insulin resistance, impaired glucose tolerance, chronic kidney disease, and bone disorders, such diet may also be associated with increased blood pressure due to increased cortisol levels and vasoconstriction¹².

There is very limited evidence regarding the association between dietary acid load and CVD mortality. Initially, the results of Tehran Lipid and Glucose Study were unable to establish an independent association between the incidence of CVD and dietary acid load¹³. In contrast, a cross-sectional study with 11,601 participants concluded that dietary acid load was associated with increased CVD risk independent of obesity and insulin resistance¹⁴. Two other studies investigated the association between dietary acid load and all-cause and CVD mortality. The Golestan Cohort Study, which enrolled 50,045 participants, demonstrated that all-cause and CVD mortality were higher at both the lowest and highest PRAL values, in a roughly U-shaped relationship. The results of this study suggest an adverse association of dietary acidity and alkalinity with increased risk of all-cause and CVD mortality⁴. Similarly, a second study, with a 15-year follow-up (1998–2012), also showed a non-linear U-shaped relationship, indicating higher mortality rates for excess dietary acidity and alkalinity in both sexes¹⁵.

In this study, given the limited and inconsistent existing reports on the association between dietary acid load, all-cause mortality, and cardiovascular disease mortality, our aim was to prospectively assess the link between dietary acid load indices (measured by PRAL, NEAP, and DAL) and both CVD incidence and all-cause/CVD mortality. To assess this, we used data from a large population-based cohort study.

Methods

Study design and population

The design of this study is prospective cohort study and the participants were selected from among 9704 adults aged between 35 and 65 years from northeastern Iran that enrolled in Mashhad Stroke and Heart Atherosclerosis Disorder (MASHAD) cohort study. This study started in 2010 with the aim of investigating the association between CVDs and multiple risk factors¹⁶. Demographic, anthropometric data, dietary intake, blood pressure and laboratory measurements were determined at baseline. The individuals excluded from this study were the participants with incomplete dietary intake data, those who over- and under-reported energy intake (< 800 and > 4200 kcal), and women who were pregnant or lactating, leaving 6482 individuals for the final analysis. The study protocol was approved by the Human Research Ethics Committee of Medical School, Mashhad University of Medical Sciences (MUMS) and all of the participants provided written informed consent.

Anthropometric measurements and physical activity level (PAL)

Standard protocols for measuring height, weight, and waist circumference (WC) were performed as previously described¹⁶. To calculate body mass index (BMI), the formula of weight in kilograms divided by the square of height in meters was used. The James and Schofield energy requirement equations were employed to define PAL¹⁷. To assess the physical activity levels, a modified version of the Scottish Heart Health Study (SHHS)/MONICA questionnaire was used. The concerning questions focused on the time spent on activities during work, non-work time and in bed period. Further details have been previously discussed¹⁸.

Biochemical measurements

After a 12- hour overnight fasting period, 10 ml of peripheral venous blood was taken from all individuals and all blood samples were centrifuged at room temperature for 30–45 min. Serum triglyceride (TG), high-density lipoprotein-cholesterol (HDL-C), low-density lipoprotein-cholesterol (LDL-C), total cholesterol and fasting plasma glucose (FPG), were determined using a BT-3000 autoanalyzer (Biothecnica, Italy) and commercial kits supplied by Pars Azmoon (Tehran, Iran). Further applicable details have previously been published¹⁶. Biochemical criteria and clinical parameters were used to define chronic disease. Diabetes mellitus was diagnosed using a FPG ≥ 126 mg/dl, or previous diagnosis of type 2 diabetes. Hypertension was defined as a systolic blood pressure or diastolic blood pressure 140 and 90 or higher, respectively, or treatment of pre-diagnosed hypertension. Dyslipidemia was determined as elevated plasma LDL-C ≥ 130, TG ≥ 150, or a low HDL-C < 40 mg/dl in men, and < 50 mg/dl in women or treatment of previously diagnosed dyslipidemia¹⁶.

Food-frequency questionnaire

In the first phase of the MASHAD study, the participants’ dietary intake was assessed using a previously validated 65-item semi‐quantitative food frequency questionnaire (FFQ) that included five frequency of use, namely per day, week, month, rarely and never and portion size for each food item¹⁶. Portion sizes of foods consumed were reported in household sizes or standard scales of weight and volume of the serving, which were completed by an experienced dietician via face-to-face interview. This FFQ was developed and validated for the Iranian population to represent typical dietary habits and culturally specific food patterns. It covered all major food groups including grains, dairy products, meats and legumes, fruits, vegetables, oils and fats, beverages, and traditional Iranian dishes, providing comprehensive representation of both acid-forming and base-forming foods commonly consumed in Iran. The questionnaire showed acceptable validity and reliability for ranking participants according to their energy and nutrient intakes within this population¹⁹. Dietplan 6 software (Forestfield Software Ltd., Horsham, West Sussex, UK) was used to analyze the Food groups. Food groups and their classifications have been fully described in a previous study¹⁹.

Assessment of dietary acid load indices

NEAP, PRAL and DAL scores were obtained by using estimations of multiple nutrient intakes as follows^14,20–22:

• NEAP score:

NEAP (mEq/d) = [54.5 × protein intake (g/d) ÷ potassium intake (mEq/d)]—10.2

• PRAL score:

PRAL (mEq/d) = 0.4888 × protein intake (g/d) + 0.0366 × phosphorus (mg/d)—0.0205 × potassium (mg/d)—0.0125 × calcium (mg/d)—0.0263× magnesium (mg/d)

• DAL score:

DAL (mEq/day) = PRAL + (body surface area [m2] × 41[mEq/day]/1.73 m²)

• Body surface area:

0.024265 × Height (cm)^0.3964 × Weight (kg).

Follow up and outcomes

The second phase of the MASHAD study was conducted between the years 2018 and 2021. Therefore, all the participants were followed up for 10 years. During the follow-up period, 1111 people were not available any longer for follow-up due to non-response, unwillingness, lack of access, and migration, with 443 people reported dead. Finally, 9404 people participated in phase II of the study. All participants in the second phase were evaluated for probable CVD events, based on clinical history, previous clinical tests and medical examinations such as radioisotope scans, stress echocardiography, computed tomography¹, angiography, coronary angiography, exercise tolerance test⁴ and other relevant medical records available. Three cardiologists and a neurologist judged these findings. In addition, when necessary, a panel of cardiologists decided on the final diagnosis of the participants. The cause of death questionnaire was completed by immediate family members: a spouse or children. The cause of death was confirmed using the health registry of the Iranian Ministry of Health.

Statistical analysis

To check the normal distribution of variables, the Kolmogorov–Smirnov test was carried out. Continuous variables were presented as the mean ± SD, while categorical variables were displayed as counts (percentages). To compare variables between different quartiles of DAL, Chi square test and one-way ANOVA were used for categorical and continuous variables, respectively. A Cox regression model was applied to examine the association between all-cause mortality, CVD mortality and CVD incidence with dietary acid load parameters. Cubic spline analyses examined the connection between the PRAL factor, all-cause mortality and CVD events. Statistical analysis was undertaken using the Statistical Package for the Social Science version 25.0 (SPSS, Chicago, IL) and P < 0.05 was considered statistically significant.

Results

The present study included 6482 participants (59.83% female), aged between 35 and 65 years. Baseline features of participants over quintiles of PRAL are demonstrated in Table 1. Age, gender, job status (P < 0·001), educational level (P = 0.039) and smoking status (P = 0.020) were different among quintiles of PRAL. There were no significant variations in the history of chronic disorders, physical activity level, anthropometric measurements, lipid profiles, and blood sugar levels.

Table 1.

Baseline characteristics, anthropometric, and laboratorial data, according to PRAL quartiles.

Variables	PRAL quartiles				P-value
	Q1 N = 1623	Q2 N = 1620	Q3 N = 1615	Q4 N = 1624
Sociodemographic data
Age (year)	49.24 ± 8.27	48.16 ± 8.09	48.27 ± 8.29	48.08 ± 8.09	< 0.001
Sex, n Male Female	603 (37.22%) 1017 (62.78%)	612 (37.75%) 1009 (62.25%)	653 (40.26%) 969 (59.74%)	736 (45%) 883 (55%)	< 0.001
Educational level, n Diploma or under University educated	1421 (87.72%) 199 (12.28%)	1444 (89.08%) 177 (10.92%)	1472 (90.75%) 150 (9.25%)	1454 (90%) 165 (10%)	0.039
Marriage, n Single Married Divorced Widowed	12 (0.74%) 1506 (92.96%) 18 (1.11%) 84 (5.19%)	13 (0.80%) 1510 (93.15%) 23 (1.42%) 75 (4.63%)	4 (0.25%) 1512 (93.22%) 25 (1.54%) 81 (4.99%)	13 (0.80%) 1500 (93%) 25 (0.2%) 81 (5.00%)	0.590
Job status, n Student Employed Unemployed Retired	9 (0.56%) 581 (35.89%) 839 (51.82%) 190 (11.74%)	1 (0.06%) 580 (35.80%) 867 (53.52%) 172 (10.62%)	6 (0.37%) 594 (36.64%) 876 (54.04%) 145 (8.95%)	3 (0.19%) 671 (41.45%) 763 (47.13%) 182 (11.24%)	< 0.001
Smoking status, n Non-smoking Ex-smoking Current-smoking	1142 (70.49%) 167 (10.31%) 311 (19.20%)	1134 (69.96%) 138 (8.51%) 349 (21.53%)	1129 (69.61%) 168 (10.36%) 325 (20.04%)	1070 (66.09%) 170 (10.50%) 379 (23.41%)	0.020
History of chronic diseases, n Yes  No	1474 (91.78%) 132 (8.22%)	1446 (89.70%) 166 (10.30%)	1454 (89.92%) 163 (10.08%)	1450 (90.40%) 154 (9.60%)	0.182
PAL	1.63 ± 0.30	1.61 ± 0.29	1.60 ± 0.28	1.59 ± 0.30	0.105
Anthropometric measurements
Weight (kg)	72.20 ± 12.64	71.73 ± 12.72	71.40 ± 13.04	71.94 ± 12.88	0.183
Height (m)	1.60 ± 0.09	1.60 ± 0.09	1.60 ± 0.09	1.61 ± 0.09	0.551
BMI (kg/m2)	28.34 ± 4.79	27.98 ± 4.67	27.88 ± 4.79	27.73 ± 4.65	0.268
Laboratory data
FPG (mg/dl)	94.07 ± 39.98	92.78 ± 38.28	94.14 ± 39.59	92.22 ± 36.93	0.505
Triglycerides (mg/dl)	145.47 ± 95.53	142.00 ± 86.87	141.23 ± 87.51	141.48 ± 89.11	0.157
Total cholesterol (mg/dl)	193.71 ± 39.52	191.67 ± 38.81	192.91 ± 39.20	191.76 ± 37.80	0.065
HDL (mg/dl)	43.54 ± 10.14	42.90 ± 10.20	43.13 ± 9.84	43.01 ± 9.88	0.605
LDL (mg/dl)	115.40 ± 35.77	115.62 ± 34.77	117.74 ± 35.69	116.82 ± 35.03	0.862

Continuous variables are presented as the mean ± SD, while categorical variables are displayed as counts (percentages). ANOVA was used to assess continuous variables, and categorical variables were evaluated using the chi-square test. P-value < 0.05 were considered significant. Abbreviations: PRAL: potential renal acid load, PAL: physical activity level, BMI: body mass index, WC: waist circumference, FPG: fasting plasma glucose, LDL: low-density lipoprotein, HDL: high-density lipoprotein, HC: hip circumference, WHR: waist-hip ratio, AVI: adiposity volume index, ABSI: body shape index.

The mean dietary intakes of the participants over PRAL quartiles are presented in Table 2. Participants in the highest PRAL quartile had a greater intake of total carbohydrate, protein, fat, fiber, phosphate, and magnesium in comparison with the participants falling into the lowest quartile. Diets consisting of both higher intakes of rice, meat and lower consumption of vegetables and fruits, were observed to contribute to high acid load (all, P < 0.001).

Table 2.

Nutrient and food group intakes by quartile of PRAL.

Variables	PRAL quartiles				P-value
	Q1 N = 1623	Q2 N = 1620	Q3 N = 1615	Q4 N = 1624
Macronutrients/micronutrients
Energy intake (kcal/d)	1995.64 ± 594.88	1843.31 ± 531.53	1911.58 ± 536.42	2229.91 ± 624.23	0.105
Total carbohydrate intake (g/d)	291.24 ± 98.24	265.96 ± 82.72	278.61 ± 178.27	305.43 ± 93.28	< 0.001
Fat intake (g/d)	65.29 ± 24.63	61.37 ± 23.97	63.22 ± 23.49	76.86 ± 32.22	< 0.001
Protein intake (g/d)	73.22 ± 20.40	68.27 ± 19.14	72.23 ± 19.68	87.85 ± 25.23	< 0.001
Fiber intake (g/d)	28.96 ± 7.09	25.91 ± 6.61	24.97 ± 6.99	24.65 ± 9.13	< 0.001
Cal intake (mg/d)	990.71 ± 352.28	870.96 ± 311.58	904.65 ± 331.08	972.29 ± 330.38	< 0.001
P intake (mg/d)	1337.47 ± 383.65	1209.51 ± 354.94	1249.10 ± 347.48	1448.86 ± 405.89	< 0.001
Mg intake (mg/d)	345.93 ± 99.66	302.86 ± 90.91	304.48 ± 96.67	345.21 ± 112.81	< 0.001
K intake (mg/d)	4096.54 ± 1077.39	3193.23 ± 819.56	2979.19 ± 833.03	3088.39 ± 882.01	< 0.001
Food groups
Fruits intake (g/d)	397.19 ± 280.93	229.23 ± 154.91	169.90 ± 128.49	145.86 ± 114.29	< 0.001
Vegetable intake (g/d)	392.42 ± 174.79	274.87 ± 111.30	231.41 ± 98.78	208.29 ± 102.27	< 0.001
Red meat intake (g/d)	37.23 ± 27.75	36.66 ± 26.98	39.84 ± 29.71	59.07 ± 47.21	< 0.001
White meat intake (g/d)	59.18 ± 30.23	58.35 ± 28.82	61.20 ± 28.51	81.01 ± 55.29	< 0.001
Rice intake (g/d)	127.71 ± 76.83	126.61 ± 73.16	131.64 ± 86.25	160.84 ± 124.93	< 0.001
Coffee intake (mg)	9.73 ± 78.16	3.15 ± 19.35	3.87 ± 34.65	3.44 ± 29.38	< 0.001
Tea intake (mg)	1361.54 ± 1164.72	1129.44 ± 590.96	1024.77 ± 565.43	940.77 ± 509.81	< 0.001
Fish intake (g/d)	8.90 ± 12.58	7.25 ± 9.70	6.64 ± 8.28	7.64 ± 12.31	< 0.001

Data are presented as the mean ± SD. ANOVA was used to compare groups. P value < 0.05 were considered significant. PRAL: potential renal acid load.

The reference group was selected based on data from the first quartile (Q1). No significant variation was observed in the unadjusted model when comparing the ORs for the higher quartiles of the DAL indices (PRAL, NEAP, and DAL) to those in the lowest quartile (P > 0.05). Likewise, following adjustment for smoking status, BMI, sex, the relationship between higher dietary acid load indices and all-cause mortality were still non-significant (P > 0.05) (Table 3). As for PRAL, participants in Q2 and Q3 had a significantly elevated risk of CVD mortality in comparison with Q1 in both models, while Q4 remained non-significant. Similarly, for NEAP, Q2 was associated with a significantly increased risk in both models, whereas Q3 and Q4 showed no significant associations (P > 0.05). As regards DAL, Q2 was linked to a higher risk in both models, while Q3 was significant only in Model 2 (P = 0.046), and Q4 was non-significant in both models (Table 4). Individuals in the highest quartile of dietary acid load (assessed using PRAL, NEAP, DAL) did not exhibit an increased risk of adverse cardiac events compared to those in the lowest quartile (Table 5).

Table 3.

Cox regression model for the association between all-cause mortality and dietary acid load parameters.

Variable	OR	95% CI	P-value
PRAL
Model 1
Q1	Ref	–	–
Q2	1.026	0.744–1.414	0.877
Q3	1.084	0.790–1.489	0.617
Q4	1.051	0.765–1.445	0.757
Model 2
Q1	Ref	–	–
Q2	1.004	0.727–1.386	0.980
Q3	1.061	0.772–1.457	0.715
Q4	0.992	0.721–1.365	0.962
NEAP
Model 1
Q1	Ref	–	–
Q2	1.042	0.758–1.432	0.801
Q3	0.968	0.701–1.336	0.843
Q4	1.076	0.786–1.473	0.609
Model 2
Q1	Ref	–	–
Q2	1.005	0.731–1.383	0.973
Q3	0.924	0.669–1.277	0.633
Q4	1.021	0.745–1.398	0.899
DAL
Model 1
Q1	Ref	–	–
Q2	0.818	0.483–1.385	0.455
Q3	0.703	0.419–1.181	0.183
Q4	0.862	0.510–1.454	0.577
Model 2
Q1	Ref	–	–
Q2	0.816	0.586–1.136	0.228
Q3	0.945	0.687–1.131	0.730
Q4	1.068	0.779–1.465	0.682

Model 1: unadjusted, Model 2: adjusted for BMI, sex, and smoking status. Abbreviations: PRAL: potential renal acid load, NEAP: net endogenous acid production, DAL: dietary acid load.

Table 4.

Cox regression model for the association between CVD mortality and dietary acid load parameters.

Variable	OR	95% CI	P-value
PRAL
Model 1
Q1	Ref	–	–
Q2	1.953	1.154–3.305	0.013
Q3	2.242	1.357–3.702	0.002
Q4	1.238	0.694–2.208	0.470
Model 2
Q1	Ref	–	–
Q2	1.974	1.138–3.424	0.016
Q3	2.323	1.397–3.862	0.001
Q4	1.335	0.724–2.460	0.354
NEAP
Model 1
Q1	Ref	–	–
Q2	1.711	1.044–2.805	0.033
Q3	1.341	0.806–2.232	0.259
Q4	1.317	0.782–2.217	0.300
Model 2
Q1	Ref	–	–
Q2	1.802	1.083–2.997	0.023
Q3	1.388	0.823–2.340	0.219
Q4	1.396	0.813–2.399	0.227
DAL
Model 1
Q1	Ref	–	–
Q2	1.769	1.051–2.977	0.032
Q3	1.613	0.964–2.700	0.069
Q4	1.620	0.965–2.719	0.068
Model 2
Q1	Ref	–	–
Q2	1.768	1.046–2.991	0.033
Q3	1.709	1.010–2.894	0.046
Q4	1.673	0.971–2.883	0.064

Model 1: unadjusted, Model 2: adjusted for BMI, sex, marriage, smoking status, fat intake, and carbohydrate intake. Abbreviations: PRAL: potential renal acid load, NEAP: net endogenous acid production, DAL: dietary acid load.

Significant values are in bold.

Table 5.

Cox regression model for the association between adverse cardiac events and dietary acid load parameters.

Variable	OR	95% CI	P-value
PRAL
Model 1
Q1	Ref	–	–
Q2	1.044	0.832–1.311	0.709
Q3	1.091	0.872–1.366	0.444
Q4	0.959	0.766–1.201	0.718
Model 2
Q1	Ref	–	–
Q2	1.012	0.805–1.272	0.919
Q3	1.071	0.855–1.341	0.552
Q4	0.980	0.779–1.232	0.861
NEAP
Model 1
Q1	Ref	–	–
Q2	1.011	0.803–1.273	0.925
Q3	1.056	0.845–1.319	0.632
Q4	1.006	0.804–1.260	0.956
Model 2
Q1	Ref	–	–
Q2	1.008	0.800–1.269	0.947
Q3	1.070	0.855–1.338	0.555
Q4	1.040	0.828–1.308	0.734
DAL
Model 1
Q1	Ref	–	–
Q2	0.993	0.790–1.248	0.951
Q3	1.023	0.821–1.275	0.837
Q4	0.980	0.783–1.226	0.857
Model 2
Q1	Ref	–	-
Q2	0.958	0.761–1.206	0.714
Q3	0.999	0.800–1.247	0.990
Q4	1.001	0.792–1.265	0.993

Model 1: unadjusted, Model 2: adjusted for BMI, sex, marital status, smoking status, fat intake, and carbohydrate intake. Abbreviations: PRAL: potential renal acid load, NEAP: net endogenous acid production, DAL: dietary acid load.

Figure 1 shows two cubic spline analyses examining the connection between the PRAL factor and three outcomes: all-cause mortality, CVD events, and CVD mortality. In the left spline plot, the x-axis represents the PRAL factor while the y-axis shows the relative risk of all-cause mortality. The blue line represents the estimated relationship, with the gray shaded area indicating the 95% confidence interval (CI). The analysis suggests that there is no significant link between the PRAL factor and all-cause mortality, supported by the flat curve and wide confidence intervals (P = 0.783). Similarly, the spline analysis indicates a non-significant relationship between the PRAL factor and the risk of CVD events, as shown by the predominantly flat curve (P = 0.982). However, PRAL demonstrated a significant association, with CVD mortality increasing sharply with higher PRAL values (P = 0.004).

Fig. 1.

Dose–response relation between PRAL and all-cause mortality, CVD events, and CVD mortality. Abbreviations: PRAL: potential renal acid load, CVD: cardiovascular disorder.

In Fig. 2, the left plot displays the relationship between the DAL factor and all-cause mortality. The analysis suggests a non-linear relationship between the DAL factor and all-cause mortality, with a tendency towards reduced mortality at lower DAL values although this trend is not statistically significant (P = 0.512). On the contrary, the curve representing the association between the DAL factor and CVD events is relatively flat and not statistically significant (P = 0.838). In the analysis of the curve demonstrating the relationship between the DAL factor and CVD mortality, the DAL factor shows a more pronounced increase in CVD mortality with a peak at mid-levels of dietary acid load (P = 0.009).

Fig. 2.

Dose–response relation between DAL and all-cause mortality, CVD events, and CVD mortality. Abbreviations: DAL: dietary acid load, CVD: cardiovascular disorder.

Based on Fig. 3, which specifically shows the relationship between the NEAP factor and three outcomes: all-cause mortality, CVD events and CVD mortality, a weak and non-significant association was observed between the NEAP factor and all-cause mortality (0.651). Similarly, the analysis of the association between the NEAP factor and the risk of CVD events revealed no significant correlation (P = 0.992). However, the analysis of the association between the NEAP factor and the CVD mortality, showed that the link was moderate, peaking around the middle of the factor’s range, with a non-significant trend (P = 0.090).

Fig. 3.

Dose–response relation between NEAP and all-cause mortality, CVD events, and CVD mortality. Abbreviations: NEAP: net endogenous acid production, CVD: cardiovascular disorder.

Discussion

This study shed further light on various aspects of the relationship between dietary acid load indices and health outcomes: all-cause mortality, CVD mortality, and CVD occurrence. Elevated levels of NEAP, PRAL, or DAL did not show a significant association with a higher risk of mortality from all causes or CVD events, while both PRAL and DAL were correlated with higher odds of CVD mortality.

One of the key aspects of the link between DAL and CVD mortality is that higher diet-induced acidosis is tied to metabolic conditions like hypertension²³, the metabolic syndrome²⁴, and insulin resistance²⁵. Nevertheless, there have been relatively few studies on dietary acid load relative to cause-specific mortality^4,15,26,27. One study found that dietary acidity was positively associated with all-cause as well as CVD mortality²⁶, whilst other studies identified a U-shaped association between dietary acidity and mortality rates^4,27. Variations in DAL indices scoring, dietary composition, measurement methods, and study populations may have contributed to these inconsistent results. In line with our study, another research found that dietary acid load was significantly associated with increased CVD mortality¹⁵. However, our study clarified that higher DAL levels is associated with increased CVD mortality, while it has no clear association with mortality from other causes.

High blood pressure is one of the metabolic conditions caused by a high acid load in the diet. By damaging blood vessels and increasing the heart’s workload, high blood pressure raises the risk of cardiovascular diseases such as heart attacks and strokes, while having no direct effect on mortality from other causes²⁸. Earlier research has showed a relationship between high blood pressure and DAL^23,29–31. A combination of physiological and hormonal factors likely explains the mechanism behind the relationship between dietary acidity and elevated blood pressure. One key mechanism involves an increase in hydrogen ion (H⁺) production from acidic food intake, such as animal protein. To maintain pH homeostasis, the kidneys and lungs buffer the excess H⁺, leading to greater sodium reabsorption and decreased potassium excretion, contributing to hypertension²³. Furthermore, diets low in potassium and magnesium, typically found in fruits and vegetables, can lead to vasoconstriction and increased blood pressure^32,33. A higher dietary acid load could also trigger the renin–angiotensin–aldosterone system, raising blood pressure by constricting blood vessels and increasing sodium retention in the kidneys³⁴.

The link between DAL and CVD mortality can be clarified through its relationship with an increased risk of metabolic syndrome and insulin resistance. Previous studies have demonstrated that DAL is significantly associated with higher chances of metabolic syndrome^24,31,35. Consuming a diet rich in acid-forming foods increases insulin secretion, leading to reduced sensitivity of insulin receptors and impaired beta cell function³⁶. Furthermore, high carbohydrate intake and low protein consumption decrease fibroblast growth factor 21 levels, which worsens insulin resistance by boosting fatty acid and cytokine production³⁷. At the molecular level, high acid load increases oxidative stress and chronic inflammation and activates inflammatory pathways such as TNF-α and IL-6, which lead to endothelial damage and atherosclerosis³⁸. These conditions increase the likelihood of heart attack and stroke by reducing blood flow and increasing the risk of clotting.

However, the exact mechanism of the association between dietary acidity and increased probability of mortality from CVD diseases is still not thoroughly understood. Some other possible mechanisms could be explained through DAL impact on bone health, vascular calcification, and muscle function³⁹. Higher acid load diets may cause calcium to be leached from bones to buffer excess acid, leading to weaker bones and an increased risk of vascular calcification, a process in which calcium deposits form in blood vessels, contributing to arterial stiffness and an elevated risk of cardiovascular mortality⁴⁰. Additionally, chronic low-grade acidosis can result in muscle wasting, negatively impacting physical activity levels and overall cardiovascular health⁴¹.

Our results showed that CVD mortality increased at mid-range levels of PRAL and DAL but weakened at the highest levels, suggesting a non-linear, possibly inverted U-shaped association. In our cohort, participants in the middle quartiles had higher intakes of protein and phosphorus, the main acid-generating nutrients, together with lower intakes of potassium, calcium, and magnesium, leading to limited buffering capacity against dietary acid load. This metabolic condition may have raised blood pressure through activation of the renin–angiotensin–aldosterone system, increased sodium retention, and reduced potassium excretion^23,32. A lack of alkaline minerals such as potassium and magnesium could also have contributed to vascular constriction, oxidative stress, and inflammation^32,33,38. In the highest quartile, although the total acid load was greater, slightly higher calcium and magnesium intake might have partially neutralized the excess acidity and reduced its physiological effects, which could explain the observed plateau in risk. A moderate but persistent acid load may have been sufficient to worsen hypertension, insulin resistance, and endothelial dysfunction³¹, conditions that are more strongly related to cardiovascular death than to overall CVD incidence²⁸.

The significantly larger sample size, and the employment of cubic spline analysis to allow for a detailed examination of non-linear relationships—which helped in reconciling inconsistencies seen with previous studies, were what provided for a far more solid ground for the arguments of the present study. Nevertheless, the study was limited by its use of self-reported dietary intake data that may be affected by recall bias and the possibly insufficient adjustment for confounders such as genetic predispositions or environment. In addition, although the 65-item FFQ captures all major food groups, it may still represent a limitation when compared with more detailed versions. However, this FFQ was chosen for its feasibility in large population studies, where shorter questionnaires can reduce respondent burden and improve feasibility within limited time periods. It is also possible that failure to follow-up influenced the representativeness of the cohort, since participants who completed the study may have differed in lifestyle or health status from those who did not. Moreover, dietary intake was assessed only at baseline, so changes in eating patterns over the follow-up period were not recorded. Finally, although the cause-of-death information was verified by the Health Registry of the Iranian Ministry of Health, and is therefore expected to be accurate, the possibility of some misclassification cannot be completely excluded.

Conclusion

In summary, this study indicates that elevated dietary acid load levels are associated with an increased risk of cardiovascular disease mortality. However, other clinical outcomes, like all-cause mortality and cardiovascular events, did not reach statistical significance. Future research is needed to address these gaps, utilizing all available data and appropriately considering confounding factors.

Abbreviations

CVD

Cardiovascular disease

MASHAD

Mashhad stroke and heart atherosclerosis disorder

NEAP

Net acid production

PRAL

Potential renal acid load

DAL

Dietary acid load

BMI

Body mass index

WC

Waist circumference

TG

Triglyceride

HDL-C

High-density lipoprotein-cholesterol

LDL-C

Low-density lipoprotein-cholesterol

FPG

Fasting plasma glucose

Author contributions

Study concept and design: N.S. and M.Gh. data analysis and interpretation of data: F.F. and A.K.; data collection: H.H.; drafting of the manuscript: N.A., N.H., N.S.; supervision and critical revision: N.S., G.A.F., and M.Gh. All authors have approved the final article.

Funding

This study was supported by Mashhad University of Medical Sciences (MUMS). The funders had no involvement in study design; collection, analysis, interpretation of data, the writing of the report, and in the decision to submit the article for publication.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to university data ownership policies, but can be obtained from the corresponding author on reasonable grounds for the request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

All experiments were performed in accordance with the declaration of Helsinki and Mashhad University of Medical Sciences ethical guidelines and regulations. The research protocol was approved by the School of Medicine, Mashhad University of Medical Sciences, Biomedical Research Ethics Committee (IR.MUMS.MEDICAL.REC.1398.228). All participants signed a written informed consent before participating in the study.