Introduction

Gastric cancer is the fourth most common cancer worldwide and the fourth leading cause of cancer deaths [1]. Over the last several decades, there has been a steady decline in its incidence and mortality [2]. This burden reduction is mainly attributed to a decrease in the prevalence of Helicobacter pylori (H. pylori) infection [3, 4], and to improvements in food preservation practices, such as an increased use of refrigeration, which enables a greater consumption of fresh fruits and vegetables, and a decreased use of traditional methods of food preservation, such as salting [5].

The most recent World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) Report found strong evidence that consuming foods preserved by salting increases the risk of gastric cancer, while there was limited evidence on the association between total salt or use of table salt and gastric cancer [6]. Most studies examining these associations have been conducted in East Asian countries, and predominantly focus on high-salt foods and salt-preserved foods, including pickled vegetables and salted or dried fish [6]. Additionally, analyses considering cancer anatomical subtype have not been possible and most studies do not take into account H. pylori infection status [6].

The Stomach cancer Pooling (StoP) Project, an international consortium of case–control and nested case–control within cohort studies, which uses an individual participant data approach for the evaluation of the associations between risk factors and gastric cancer [7], allows for some of these limitations to be overcome. Therefore, this study aimed to quantify the association between dietary salt exposure, defined according to different criteria (salt taste preference, use of table salt, total sodium intake, and high-salt and salt-preserved foods intake), and gastric cancer, using an individual participant data meta-analysis of studies participating in the StoP Project.

Methods

For this study, version 3.0 of the StoP Project dataset was used, which includes a total of 12,511 cases of gastric cancer and 29,964 controls from 32 case–control or nested case–control studies [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37]. All data were collected and harmonized according to a pre-specified format at the pooling center. The participating studies were conducted in accordance with applicable laws, regulations and guidelines for the protection of human subjects, and the StoP Project was approved by the University of Milan Review Board (Reference 19/15).

The present analysis used data from 25 studies, corresponding to 10,283 cases and 24,643 controls with information on dietary salt and/or sodium intake: Brazil (two studies) [31, 32], Canada [14], China (four studies) [9, 15, 19, 20], Greece [13], Iran (two studies) [17, 18], Italy (four studies) [8, 10,11,12], Japan [33], Mexico (three studies) [28,29,30], Portugal [22], Russia [16], Spain (two studies) [25, 27], and USA (three studies) [21, 35, 37].

The quality of studies included was assessed using the Newcastle–Ottawa Scale (NOS) [38]. The scale evaluates the quality of studies based on three different categories: selection, comparability, and exposure (case-control studies) or outcome (nested case-control studies). A study can be awarded a maximum of nine stars, which indicates the highest quality.

Dietary salt exposure was defined according to several different criteria, as described in detail in Supplementary Table 1. First, salt taste preference was collected and defined as tasteless, normal, or salty (n = 11 studies). Second, the frequency of use of table salt was assessed and defined as never, sometimes, or always (n = 11 studies). Third, total dietary sodium intake in grams/day, as a computed nutrient, was estimated using food frequency questionnaires (FFQs). Two studies [9, 20] only provided information on the consumption of salt in grams/day as reported by participants. For analysis, total sodium intake in grams/day was categorized into study-specific tertiles according to the distribution in controls (n = 16 studies). Fourth, high-salt and salt-preserved foods intake in grams/day was estimated by adding up the amounts in grams of each single food item or group consumed per day obtained using FFQs. These foods included the main food items or food groups contributing to total dietary sodium intake [39], namely pickles or pickled vegetables, vegetables braised in soy sauce, preserved and/or fermented bean curd, miso soup, salty snacks (e.g., potato chips, corn chips, popcorn, crackers), cheese (e.g., cream cheese, parmesan cheese, sliced cheese), salty condiments (e.g., soy sauce, mayonnaise, ketchup), salted vegetables, fish or meat, smoked, dried or processed fish or meat (e.g., canned tuna, sausages or hot dogs, bacon, ham, cold cuts or lunch meats, croquettes), and grains, cereals and potatoes (e.g., rice, pasta, bread). For analysis, the study-specific tertiles of high-salt and salt-preserved foods intake observed in the controls were used as cut-offs to define groups of exposure (n = 20 studies).

Depending on the study, FFQs were used to obtain information on the dietary habits of participants for the period of one, two, or five years before diagnosis (for cases from case–control studies), onset of disease or hospital admission (for hospital-based controls) or recruitment (for population-based controls, and participants from the nested case–control study [37]; Supplementary Table 1). Most studies (n = 21) used face-to-face interviews by trained researchers for the administration of FFQs, while the remaining used self-administered FFQs. Fourteen of the included studies reported that the FFQ used was previously validated using 24-h recall interviews and/or dietary records. The FFQs used in the studies included between 19 and 147 individual food and beverage items.

A two-stage modeling approach was used to quantify the association between dietary salt exposure and gastric cancer [40]. First, the study-specific odds ratios (ORs) and corresponding 95% confidence intervals (95% CI) were estimated for the association between each measure of dietary salt exposure and gastric cancer using multivariable unconditional logistic regression models. Models were adjusted for sex, age (5-year age groups: < 40; 40–44; …; 70–74; ≥ 75), socioeconomic status (low, intermediate, or high, as defined in each original study based on education, income or occupation), smoking status (never, former and current smokers of < 10 cigarettes/day [low]; 10 to 20 cigarettes/day [intermediate]; > 20 cigarettes/day [high]), alcohol drinking (never, low: < 13 g of ethanol/day, intermediate: 13 to 47 g of ethanol/day, high: > 47 g of ethanol/day), fruits and vegetables intake (study-specific tertiles), total energy intake (study-specific tertiles), study center (for multicenter studies), and race/ethnicity (White, Black/African American, Asian, Hispanic/Latino, other), when appropriate and available as described in detail in Supplementary Table 2. Second, summary (pooled) effect estimates were computed using random-effects models [41]. Heterogeneity between studies was quantified using I2 (%) statistics [42].

Stratified analyses were also carried out to further explore the effect of dietary salt exposure across strata of sex, age (≤ 55, 56–65, > 65), geographic region, socioeconomic status, smoking status, alcohol drinking, fruits and vegetables intake, and type of study (hospital-based controls, population-based controls, nested case–control study). The difference between groups was assessed through the Q test for heterogeneity [43]. Multinomial logistic regression models were used to estimate the ORs for each cancer anatomical subsite separately (i.e., cardia, non-cardia, not accurately classifiable), and each histological type separately (i.e., intestinal, diffuse, undifferentiated).

Sensitivity analyses were carried out by defining the same categories of exposure for all studies according to the distribution of total sodium intake, and high-salt and salt-preserved foods intake in all controls. Categories of exposure for total sodium intake were also further defined considering the maximum intake amounts recommended by the World Health Organization (WHO), i.e., less than two grams of sodium/day [44]. The cut-offs that describe consumption of less than half of the recommended amount, between half and the recommended amount or more than the recommended amount were used, resulting in three categories (< 1.0, 1.0–1.9, ≥ 2.0 g of sodium/day). Additional sensitivity analyses included comparing the estimates adjusted and not adjusted for total energy intake (n = 16 studies), as well as stratified by seroinfection by H. pylori, among studies with the available information (n = 11 studies). Furthermore, studies that used a self-administered FFQ (n = 4 studies), non-validated FFQs (n = 11), and that scored five or less stars in the NOS (n = 5 studies, lower quality) were removed, and analyses were also restricted to studies evaluating participants more than one year before the gastric cancer diagnosis. Finally, the influence of each study on the pooled estimates was also examined by excluding one study at a time.

Statistical analyses were performed using STATA version 15.1 (STATA Corporation, College Station, Texas, USA). A p-value less than 0.05 was considered statistically significant.

Results

The main sociodemographic characteristics of the cases and controls are described in Supplementary Table 3, and the distribution of dietary salt exposure, defined according to different criteria by case–control status, is shown in Supplementary Tables 4 and 5.

The study-specific and pooled adjusted ORs for dietary salt exposure, defined according to different criteria, and gastric cancer are presented in Fig. 1. A significantly higher risk of gastric cancer was observed for a salty taste preference (OR 1.59, 95% CI 1.25–2.03, I2 = 66.2%), always using table salt (OR 1.33, 95% CI 1.16–1.54, I2 = 13.0%), and a higher high-salt and salt-preserved foods intake (grams/day; OR highest vs. lowest tertile: 1.24, 95% CI 1.01–1.51, I2 = 77.4%). No significant association was observed for a greater total sodium intake (grams/day; OR highest vs. lowest tertile: 1.08, 95% CI 0.82–1.43, I2 = 83.2%). Using the same cut-off for all studies, defined either by the overall distribution in all controls or taking the total sodium intake amounts recommended by the WHO into account, led to estimates of the same magnitude, with slightly lower heterogeneity (Table 1).

Fig. 1
figure 1

Forest plots describing the association between salt taste preference (salty vs. tasteless), use of table salt (always vs. never), total sodium intake (grams/day—study-specific 3rd tertile vs. 1st tertile), and high-salt and salt-preserved foods intake (grams/day—study-specific 3rd tertile vs. 1st tertile), and gastric cancer using estimates from the Stomach cancer Pooling (StoP) Project database. 95% CI 95% confidence interval, aOR adjusted odds ratio, NA not available. aPooled ORs were computed using random-effects models. Study-specific ORs were adjusted, when available and applicable, for sex, age (5-year age groups: < 40; 40–44; …; 70–74; ≥ 75), socioeconomic status (low, intermediate, or high, as defined in each original study based on education, income or occupation), smoking status (never, former and current smokers of < 10 cigarettes/day [low]; 10–20 cigarettes/day [intermediate]; > 20 cigarettes/day [high]), alcohol drinking (never, low: < 13 g of ethanol/day, intermediate: 13 to 47 g of ethanol/day, high: > 47 g of ethanol/day), fruit and vegetables intake (study-specific tertiles), total energy intake (study-specific tertiles), study center (for multicenter studies), and race/ethnicity (White, Black/African American, Asian, Hispanic/Latino, other)

Full size image
Table 1 Pooled odds ratios and 95% confidence intervals for gastric cancer according to salt taste preference (tasteless, normal, salty), use of table salt (no, sometimes, always), total sodium intake (grams/day—study specific, overall distribution and recommended amounts), and high-salt and salt-preserved foods intake (grams/day—study specific, overall distribution)
Full size table

The effect of dietary salt exposure was consistent across most strata of sociodemographic and lifestyle characteristics, and similar when considering each cancer anatomical subsite and histological type (Table 2). Although the difference was not statistically significant, studies with population-based controls presented a higher risk of gastric cancer for salty taste preference (OR 1.83, 95% CI 1.33–2.52, I2 = 72.6%) compared to studies with hospital-based controls (OR 1.27, 95% CI 0.98–1.64, I2 = 0.0%; p for interaction = 0.081). For high-salt and salt-preserved foods intake, there were significant differences according to geographic region (p for interaction = 0.007), with a significant association being observed among studies conducted in Europe (OR 1.62, 95% CI 1.42–1.85, I2 = 3.7%) but not in studies conducted in the Americas (OR 1.22, 95% CI 0.75–1.98, I2 = 80.0%) or Asia (OR 0.86, 95% CI 0.58–1.26, I2 = 69.0%). Differences were also observed considering socioeconomic status (p for interaction = 0.024), though no consistent pattern was observed, a stronger and significant association was found among those with a higher socioeconomic status (OR 1.71, 95% CI 1.32–2.22, I2 = 1.2%). Further sensitivity analyses did not result in major changes in the direction or magnitude of the associations when considering OR estimates adjusted for total energy intake.

Table 2 Pooled odds ratio and 95% confidence interval of gastric cancer for the highest vs. lowest salt taste preference (salt vs. tasteless), use of table salt (always vs. never), total sodium intake (grams/day—study-specific 3rd tertile vs. 1st tertile), and high-salt and salt-preserved foods intake (grams/day—study-specific 3rd tertile vs. 1st tertile) according to strata of selected variables
Full size table

Additional stratified analyses according to study characteristics generally yielded similar and consistent results throughout (Supplementary Table 6). The magnitude of estimates remained essentially unchanged when considering the validity of the FFQ and method of administration. Regarding the period of assessment, the results did not materially differ from those of the main analysis, except for a stronger and statistically significant association between high total sodium intake and gastric cancer being observed when considering studies evaluating dietary intake more than one year before diagnosis (OR 1.60, 95% CI 1.33–1.93, I2 = 0.0%). Finally, applying the NOS to the included studies and removing those with five stars or less (lower quality) also did not change the associations observed in the overall analyses.

Discussion

In this study within the StoP consortium, a higher risk of gastric cancer was observed for a salty taste preference, always using table salt, and a greater intake of high-salt and salt-preserved foods, which was consistent across sociodemographic, lifestyle, and tumor characteristics. No significant association was found for the highest tertile intake of sodium compared to the lowest tertile.

Several systematic reviews reported that excessive salt intake is associated with the risk of gastric cancer [45,46,47] and the WCRF/AICR has classified salt as an important risk factor for gastric cancer [6]. In particular, strong evidence was observed for the association between consuming foods preserved by salting or high-salt foods and a greater risk of gastric cancer [6, 47]. However, the majority of studies examining this association were conducted in Asian countries [6]. In the current study, which includes participants from 11 countries in America, Asia and Europe, we observed an association between high-salt and salt-preserved foods intake and gastric cancer risk. However, a statistically significant association was not found when considering studies conducted in the Americas and Asia. This may have occurred due to the diversity of items included in each study, ranging from pickles or pickled vegetables, salty snacks, salted vegetables or fish to salty condiments, which likely also contributed to the heterogeneity observed. Furthermore, although several systematic reviews found that the risk of gastric cancer was higher among individuals with high salt intake than in those with low levels of consumption [45,46,47], the WCRF/AICR reported limited evidence regarding the association between total salt or added salt and gastric cancer risk [6]. In the current study, we observed an association between always using table salt and gastric cancer, though our findings were less robust when considering total sodium intake.

A concern in assessing total dietary sodium intake is the fact that it is a natural component present in most foods, and it may also be added during the cooking process or at the table in amounts that individuals usually ignore or are unable to accurately report [48]. Although excretion of sodium in urine over a 24-h period is widely regarded as the gold standard method for the assessment of sodium ingested from different sources [49] it cannot be retrospectively used in case–control studies. As such, several different approaches, including salt taste preference, adding salt at the table, total sodium intake, and high-salt and salt-preserved foods intake were considered in the present pooled analyses. In fact, a salty taste preference and the use of salt at the table may result in an increase in overall salt consumption, and may be a more comprehensive measure of dietary salt intake since it considers more than the intrinsic sodium content of foods, which may underestimate intake. In fact, previous studies found that preference for salty foods was moderately associated with daily salt intake [50, 51]. Furthermore, the measurement of total sodium intake in diet is subject to methodological difficulties. In particular, diet records or diet recall often underestimate total sodium intake due to underreporting by participants and due to difficulties in quantifying the concentration of sodium in food items or food groups, as well as discretionary salt intake [49].

High sodium intake is reported to act as a gastric mucosa stimulant, leading to atrophic gastritis, increased DNA synthesis, and cell proliferation, thereby providing the basis for gastric cancer development [47]. Furthermore, high sodium intake may weaken the protective effect of the mucous barrier and promote the carcinogenic effect of H. pylori infection serostatus, which is a known primary risk factor for gastric cancer [52]. This could translate into a stronger association among H. pylori positive individuals; however, we did not observe significant differences in the risk of gastric cancer when stratifying according to H. pylori seroinfection. Nevertheless, these results should be interpreted cautiously since infection status was evaluated using serum samples in case–control studies, which may lead to false negative results in the presence of advanced infection [53]. Additionally, there are potential aetiological differences between gastric cancer anatomical subsites and histological types, with previous studies suggesting a greater influence of lifestyle factors among intestinal type gastric cancers compared to those of the diffuse type [22, 54]. We conducted various analyses considering cancer anatomical subsite and histological type, with consistent results being generally observed.

Low socioeconomic status is a well-recognized risk factor for gastric cancer [55,56,57] partly because of an unfavorable distribution of risk factors including selected dietary and lifestyle habits, which were included in the models as covariates. Individuals with a low socioeconomic status may have a diet high in salt, which may increase the negative effects of other lifestyle risk factors related to a low socioeconomic status, potentially leading to a higher gastric cancer risk. Additionally, in some countries refrigerator use was likely initially restricted to higher socioeconomic status groups, which enabled the consumption of fresh foods including seasonal vegetables and fruits year round, as well as fresh meat, and reduced the need for salting, smoking, curing, and pickling to preserve food [58, 59]. As such, we could hypothesize that lower socioeconomic status groups may have been more exposed to salt-preserved foods due to lack of alternatives rather than a preference for these foods. Indeed, previous studies have found that the consumption of processed or ultra-processed foods increases with education and income level [60, 61], and these foods are major contributors to an individual’s dietary salt intake [62, 63]. In the present individual participant data meta-analysis, although differences according to socioeconomic status were observed for high-salt and salt-preserved foods intake, there was substantial heterogeneity among the studies included and there was an absence of a consistent trend. In general, no convincing evidence was observed for differences in the association between dietary salt intake and gastric cancer according to socioeconomic status, despite several criteria being used to evaluate dietary salt exposure.

Significant geographical differences were observed when considering high-salt and salt-preserved foods, with lower estimates among studies conducted in Asia, and higher ORs for studies from Europe and the Americas. Although not significantly different, higher estimates were observed among studies from Asia for salt taste preference and total sodium intake. The geographical differences observed may reflect the different diets, with higher sodium intakes being generally reported in Asia [64], but also the detail of the FFQs applied regarding the number and types of food items included, which likely contributed to the observed heterogeneity. Nevertheless, the concentration of salt in many processed foods consumed in Europe and North America approaches that of salt-preserved foods [6].

Substantial heterogeneity was observed, particularly when considering salt taste preference, total sodium intake, and high-salt and salt-preserved foods intake. This is mainly due to the different methods used to collect dietary data, particularly the period of dietary assessment, the number and the items included in each FFQ. Within the StoP consortium, most studies used FFQ designed not only to be representative of the diet of each country but also to consider the seasonality of the items included. However, the diversity of food items in each questionnaire, particularly high-salt and salt-preserved foods, likely contributed to the heterogeneity observed. Nevertheless, 12 studies included in the present analysis used previously validated FFQs, and 21 studies collected data using trained interviewers, which have an acceptable validity when compared to reference measures [65, 66]. Overall, our sensitivity analyses considering study characteristics showed no significant differences, providing further support to the robustness of our findings.

Dietary habits were reported by participants that may have led to recall bias, particularly among patients, since changes in lifestyle may occur as cancer develops and becomes symptomatic [67]. Nevertheless, studies recruited incident, histologically confirmed gastric cancer cases, and most obtained dietary information regarding at least the year before diagnosis or the period before changes in dietary habits. We conducted sensitivity analyses considering studies evaluating dietary intake more than one year before diagnosis, and in general, the magnitude and directions of the associations did not change meaningfully, except for the association between high total sodium intake and gastric cancer that became stronger and reached statistical significance among these studies. Furthermore, no significant differences in the results obtained were observed between hospital- or population-based control studies and the prospective cohort study included. Additionally, case–control studies may be prone to selection bias. It is possible that hospital-based controls include individuals with conditions that could potentially be related to dietary salt intake, while population-based controls are more likely to be representative of the study base. Nevertheless, the results of our stratified analysis by type of controls showed that the overall conclusions are not driven by the studies with hospital- vs. population-based controls.

Although substantial heterogeneity was observed, the harmonization of the definition of exposure and control of confounding used in studies of the StoP consortium, contributes to the validity of our estimates. Additionally, the significant effect of salty taste preference, use of table salt, and high-salt and salt-preserved foods intake detected in the main analysis was consistently observed among strata of different sociodemographic, lifestyle, and clinical variables, as well as study characteristics. Sensitivity analyses, either removing one study at a time or considering the same cut-off for all studies, yielded estimates similar to those observed in the main analyses, albeit with less heterogeneity.

In conclusion, our uniquely large individual participant data meta-analysis of studies participating in the StoP Project showed salty taste preference, always using table salt, and a greater intake of high-salt and salt-preserved foods were associated with a risk of gastric cancer. The association was less robust with total sodium intake, which may be due to the high heterogeneity of the food assessment methods used in each study. In particular, this study adds to previous evidence, allowing for analyses considering H. pylori infection status, as well as cancer anatomical subsite and histological type.