Alcohol consumption and the risk of gastric intestinal metaplasia in a U.S. Veterans population.

BACKGROUND: Chronic alcohol use is a risk factor for non-cardia gastric adenocarcinoma. However, it is less well understood whether alcohol use is a risk factor for premalignant mucosal changes, namely gastric intestinal metaplasia. We examined the association between various parameters of alcohol use and risk of gastric intestinal metaplasia.
METHODS: We used data from 2084 participants (including 403 with gastric intestinal metaplasia) recruited between February 2008-August 2013 into a cross-sectional study at the Michael E. DeBakey Veterans Affairs Medical Center in Houston, Texas. All participants underwent a study upper endoscopy with systematic gastric mapping biopsies. Cases had intestinal metaplasia on any non-cardia gastric biopsy. Participants self-reported lifetime history of alcohol consumption, along with other lifestyle risk factors, through a study survey. We calculated odds ratios (OR) and 95% confidence intervals (95% CI) for categories of average alcohol consumption using multivariable logistic regression, and restricted cubic spline regression to explore the potential shape of a dose-response relationship.
RESULTS: Compared to lifelong non-drinkers, individuals who consumed on average ≥28 drinks per week had no elevated risk for gastric intestinal metaplasia (adjusted OR, 1.27; 95% CI, 0.74-2.19). Based on a spline regression curve and its 95% CI, there was also no demonstrable association between cumulative lifetime alcohol consumption and risk of gastric intestinal metaplasia. Similarly, we found no association between beverage type (beer, wine, liquor/spirits) and risk for gastric intestinal metaplasia.
CONCLUSIONS: Neither amount of alcohol consumed nor specific beverage type was associated with risk of gastric intestinal metaplasia.

Introduction

Gastric adenocarcinoma (gastric cancer) is among the leading causes of cancer-related deaths worldwide [1]. Patients diagnosed with gastric cancer have a poor prognosis, especially in advanced stages, with an observed 5 year survival rate of 5–20% [2]. Gastric cancer pathogenesis involves a well-defined pathway from chronic gastritis leading to gastric intestinal metaplasia, which leads to gastric intraepithelial neoplasia (i.e., dysplasia), and ultimately gastric cancer [1-3]. Gastric intestinal metaplasia, the precursor to gastric cancer, is detected on gastric biopsies as mucus secreting goblet cells in the gastric mucosa [4].

Chronic Helicobacter pylori infection has long been established to cause inflammation of the gastric mucosa and subsequently, the development of gastric intestinal metaplasia [4-6]. In populations with high incidence of gastric cancer, there have been multiple studies that investigate environmental risk factors synergistic to H. pylori infections in regard to the development of gastric intestinal metaplasia [6, 7]. However, there are few studies that look primarily at these modifiable factors independent of H. pylori infection, and even fewer in western populations, where H. pylori infection rates are lower, that focus on lifestyle risk factors, such as alcohol use, and their effects on gastric intestinal metaplasia and gastric cancer independent of or synergistically with H. pylori infection.

Alcohol is classified as a class 1 carcinogen by the International Agency for Research on Cancer [8], and heavy alcohol consumption has been demonstrated to independently increase the risk of some but not all subtypes of gastric cancer [9]. A meta-analysis of 44 case-control and 15 cohort studies published through 2010 reported a modest positive relationship between “heavy” alcohol consumption (defined as ≥4 drinks/day) and the risk of non-cardia gastric adenocarcinoma when controlling for other lifestyle factors such as smoking and fruit and vegetable consumption (vs. non-drinkers; relative risk, 1.20; 95% confidence interval [CI], 1.01–1.44) [10]. Conversely, there seems to be no association between alcohol consumption and gastric cardia cancer [7, 11, 12]. Despite the link between heavy alcohol consumption and the development of non-cardia gastric adenocarcinoma, it is unclear whether this association is attributed to the development of gastric intestinal metaplasia, or the progression of gastric intestinal metaplasia to gastric cancer. Few studies have examined the association between alcohol consumption and gastric intestinal metaplasia, with conflicting data. Additionally, those studies have been limited; either to non-U.S. populations or by the absence of data on a dose-dependent relationship between alcohol consumption and gastric intestinal metaplasia. Few studies have examined the relationship with specific beverage type [13-15].

We examined the relationship between alcohol consumption and gastric intestinal metaplasia, and specifically sought to quantify whether increased overall alcohol consumption or consumption of specific beverage types predispose to an increased risk of non-cardia gastric intestinal metaplasia in U.S populations. Because categorization may obscure potentially important differences in risk within and across categories of alcohol consumption, we also modeled the risk of non-cardia gastric intestinal metaplasia associated with alcohol consumption as a continuous measure and examined for linear or nonlinear patterns of risk associated with higher levels of alcohol consumption.

Methods

Study population

We used data from a cross-sectional study conducted among Veterans at the Michael E. DeBakey Veterans Affairs Medical Center (MEDVAMC) in Houston, Texas from February 2008-August 2013 [16, 17]. We recruited participants from two sources: (i) among consecutive eligible patients undergoing an elective esophagogastroduodenoscopy (EGD) for any indication; and (ii) among consecutive patients attending one of seven selected primary care clinics. These two patient groups represent the source population for non-cardia gastric intestinal metaplasia cases at the MEDVAMC. The patients recruited from primary care clinics were eligible for a routine (i.e., average risk) screening colonoscopy and we invited them to participate in the study, which required them to undertake an EGD for the study. None of the primary care patients were primarily referred for EGD and none were approached during a time of a pre-scheduled colonoscopy. Patients recruited from primary care underwent the study EGD and routine screening colonoscopy in the same endoscopic session. Criteria for eligibility included: (1) age 50–80 years (40–80 years for the elective EGD group); (2) no previous gastroesophageal surgery; (3) no previous gastroesophageal cancer; (4) no active lung, liver, colon, breast or stomach cancer; (5) no anticoagulants; (6) no significant liver disease indicated by platelet count below 70 000 ascites, or known gastroesophageal varices; and (7) no history of major stroke or mental condition. The study was approved by the Institutional Review Boards for MEDVAMC and Baylor College of Medicine (H-27828). All participants provided written informed consent to take part in the study. Overall, 70% of patients in the elective EGD group and 43% of eligible patients in the primary care group underwent the study EGD and completed the study survey.

Study procedure

During the study EGD, at least 10 biopsies (2 biopsies at each of the 5–7 biopsy sites according to adoption of the updated Sydney System [18]) were taken from the antrum (both greater and lesser curvature), corpus (proximal greater curvature, proximal lesser curvature, with optional additional biopsies at distal greater curvature and distal lesser curvature), and cardia. We systematically recorded endoscopic findings from the EGD. Biopsy specimens were embedded in paraffin, oriented on edge, sectioned in 5-sections, and stained with hematoxylin and eosin, alcian blue at pH 2.5. A modified silver stain and alcian blue–periodic acid Schiff stain were also used when staining for H. pylori was negative. Two gastrointestinal pathologists independently determined presence and severity of gastric intestinal metaplasia on each specimen. When necessary a third gastrointestinal pathologist was consulted to make the final decision. All pathologists were blinded to endoscopic findings and patient survey responses. Participants with evidence of intestinal metaplasia on ≥1 non-cardia gastric biopsy were classified as gastric intestinal metaplasia cases. Cases were compared to controls, defined as participants without intestinal metaplasia on all non-cardia gastric biopsies.

Patients were considered to have H. pylori infection if H. pylori organisms were isolated on gastric tissue culture or found on histopathology of ≥1 gastric biopsy site (using hematoxylin and eosin, alcian blue at pH 2.5, a modified silver stain, or alcian blue–periodic acid Schiff stain). To process cultures for H. pylori, frozen tissue specimens were thawed, homogenized, and inoculated onto Brain Heart Infusion medium (nutrient rich agar ideal for culturing fastidious microorganisms) and H. pylori Special Peptone Agar plates with 7% horse blood. The plates were incubated at 37°C under micro-aerophilic conditions (5% O2, 10% CO2, and 85% N2) in an Anoxomat jar for up to two weeks. Positive growth was transferred to a fresh, nonselective Brain Heart Infusion blood agar plate and incubated for 48–72 hours. H. pylori were identified when the oxidase, catalase, and urease reactions were positive with a compatible Gram stain. To obtain a pure culture, we selected and subcultured several small round colonies from each patient’s plate 1 or 2 times. Isolated strains were then stored at 80°C in cysteine storage medium containing 20% glycerol.

Data collection

The survey was completed prior to the study EGD with assistance from trained research staff. The survey ascertained information about age, sex, race/ethnicity, education, use of alcohol and smoking, medical history, and use of medications. We calculated body mass index (BMI) from pre-study EGD height and weight measurements. Waist and hip circumferences were measured via flexible tape measure and we calculated waist-to-hip ratio (WHR) as waist circumference divided by hip circumference. A ratio of ≥0.9 and ≥0.85 was used to categorize a high WHR for males and females, respectively.

For alcohol consumption, participants reported whether they currently drank alcohol, were life-long non-drinkers, or had previously drunk alcohol but stopped. Ever drinkers then reported if they had ever consumed alcohol at least monthly for ≥6 months. Participants with a history of monthly alcohol consumption for ≥6 months subsequently reported frequency of consumption for four classes of alcohol (beer, white wine, red wine and liquor/spirits) at ages 20–29, 30–49 and ≥50 years, as applicable. Alcohol consumption (measured as number of all standard drinks [10 g alcohol/drink] summed across all classes of alcohol for all age groups) was divided by duration of drinking (since 20 years of age, in weeks) to calculate total lifetime alcohol consumption. Similar algorithms were used to calculate average lifetime beverage-specific consumptions [19]. In the analysis, participants who reported ever consuming alcohol but who did not consume alcohol at least monthly for ≥6 months were considered non-drinkers and included in the referent group.

Statistical analysis

Our primary aim was to estimate the relative risk of non-cardia gastric intestinal metaplasia associated with average lifetime alcohol consumption, total and beverage specific. We used unconditional multivariate logistic regression to calculate odds ratios (ORs) and associated 95% CIs. Total alcohol consumption was first modeled as a five-level categorical variable (life-long non-drinkers; <7; 7-<14; 14-<28, and ≥28 drinks/week). We used the same 5 categories for beer consumption, but we used fewer categories for wine (life-long non-drinkers; <7; ≥7) and liquor/spirits (life-long non-drinkers; <7; 7-<21; ≥21) due to limited range for these alcohol types. For all analyses, we compared history of alcohol consumption among ever drinkers (current and ex-drinkers) to a referent group of those who were lifelong non-drinkers. The final multivariate models were adjusted for potential confounders, including age (years), sex, race/ethnicity, highest level of education, BMI, smoking status, and H. pylori infection. To test for trend, category of alcohol consumption was included in the multivariate model as an ordinal variable (with category values taking the median of the range observed among the control group) and non-drinkers were excluded. We included participants with missing data for covariates (e.g., BMI, WHR, H. pylori infection, smoking status) in the analyses using an additional category for missing values.

Furthermore, to explore the shape of the dose-response association between alcohol consumption and risk of gastric intestinal metaplasia, we used a logistic regression model with restricted cubic spline for average lifetime alcohol consumption (in drinks/week) as a continuous measure by means of generalized additive logistic models (using the CRAN package mgcv in R software), adjusted for the same covariates. Smoothing splines fixed at 3 degrees of freedom were used to test for significance of nonlinearity against the linear effect [20-22].

Finally, we assessed potential biological interaction between alcohol consumption and (i) smoking status and (ii) H. pylori infection status in relation to the risk of non-cardia gastric intestinal metaplasia by creating new variables that reclassified participants according to their combined exposure and tested for departure from additivity using the Synergy Index (SI). A SI greater than 1 indicates that the joint effect of two risk factors on the risk of gastric intestinal metaplasia is greater than the sum of their independent effects and suggests the presence of biological interaction [23].

All analyses were conducted using Stata 13.0 (StataCorp LP, College Station, TX) and all tests for statistical significance two-sided at α = 0.05.

Results

This study included data from 2084 participants, with 1568 recruited from EGD clinics and 516 from primary care. The mean age of participants was 60.2 years (standard deviation, 8.1 years). Ninety-two percent of participants were male, 57.3% White and 31.3% Black. Most participants were overweight or obese (81.4%) and reported a history of smoking (68.5%).

Among the 2084 included participants, 403 were classified as cases with gastric intestinal metaplasia and 1681 as controls without gastric intestinal metaplasia. The demographic characteristics of cases and controls are shown in Table 1. Compared to controls, cases with non-cardia gastric intestinal metaplasia were significantly more likely to be aged ≥60 years at study enrollment and male, but less likely to be White and have a college degree (Table 1). Cases were also more likely to have H. pylori infection (52.5% vs. 21.9%) and a history of cigarette smoking (80.7% vs. 71.0%).

Table 1

Characteristics of cases and controls.

		Controls	Cases
		N = 1,681	N = 403
		N (%)	N (%)	Pa
Age group, years				0.001
	<60	715 (42.5)	134 (33.3)
	60–69	787 (46.8)	206 (51.1)
	≥70	179 (10.7)	63 (15.6)
Sex				<0.001
	Male	1524 (90.7)	391 (97.0)
	Female	157 (9.3)	12 (3.0)
Race/Ethnicity				<0.001
	White	1028 (61.2)	165 (40.9)
	Black	484 (28.8)	169 (41.9)
	Other	169 (10.0)	69 (17.1)
Highest level of education				<0.001
	Less than high school	94 (5.9)	46 (12.0)
	High school graduate	575 (36.3)	178 (46.2)
	Tech/vocational college	601 (37.9)	104 (27.0)
	College graduate	314 (19.8)	57 (14.8)
	Missing	97	18
BMI, kg/m2				0.06
	<25	300 (17.9)	87 (21.6)
	25–29.9	595 (35.5)	153 (38.0)
	≥30	783 (46.7)	163 (40.4)
	Missing	3	0
WHR				0.94
	Low	241 (14.6)	58 (14.8)
	High	1404 (85.4)	334 (85.2)
	Missing	36	11
H. pylori infection				<0.001
	No	1295 (78.1)	189 (47.5)
	Yes	363 (21.9)	209 (52.5)
	Missing	23	5
Smoking status				<0.001
	Never smoker	456 (29.0)	74 (19.3)
	Ex-smoker	638 (40.5)	166 (43.2)
	Current smoker	480 (30.5)	144 (37.5)
	Missing	107	19
Alcohol drinking status				0.19
	Non-drinker	151 (9.0)	26 (6.4)
	Ex-drinker	639 (38.0)	166 (41.2)
	Current drinker	891 (53.0)	211 (52.4)

BMI, body mass index; WHR, waist-to-hip ratio.

WHR was categorized as high if it was ≥0.9 for males or ≥0.85 for females.

a P-values from chi-square tests for categorical variables; missing category was excluded from statistical tests for differences between controls and cases.

Overall, 91.5% of study participants reported a history of alcohol consumption. Controls were slightly more likely than cases with non-cardia gastric intestinal metaplasia to report being a lifelong non-drinker (9.0% vs. 6.4%). Among controls who reported being ever drinkers, the median lifetime total alcohol consumption was 13 (interquartile range [IQR], 6–42) and 4 (IQR, 2–10) drinks/week for males and females, respectively. Median lifetime total alcohol consumption was numerically higher among male (16 drinks/week; IQR, 6–42) and female (8 drinks/week; IQR, 2–14) cases with non-cardia gastric intestinal metaplasia than controls.

Table 2 shows the unadjusted and adjusted ORs for associations between alcohol consumption and risk of gastric intestinal metaplasia. We found no association between alcohol drinking status and risk of non-cardia gastric intestinal metaplasia (current drinkers vs. lifelong non-drinkers; adjusted OR, 1.03; 95% CI, 0.63–1.68). Likewise, there was no association with average amount of alcohol consumed over the life-course, even for heavy drinkers. Compared to lifelong non-drinkers, individuals who consumed on average ≥28 drinks per week had no elevated risk for non-cardia gastric intestinal metaplasia (adjusted OR, 1.27; 95% CI, 0.74–2.19). Among alcohol drinkers, the risk for gastric intestinal metaplasia did not increase linearly with increasing total average alcohol consumption (p-trend = 0.24). When we analyzed beverage-specific alcohol consumption, there were no associations of beer, wine or liquor consumption with the risk of non-cardia gastric intestinal metaplasia (Table 2).

Table 2

Unadjusted and adjusted odds ratios for associations between alcohol-related variables and risk of gastric intestinal metaplasia.

		Controls	Cases
		N	N	Unadjusted	Adjusted
				OR (95% CI)	ORa (95% CI)
Current alcohol drinking status
	Non-drinkerb	151	26	1.00 (ref)	1.00 (ref)
	Ex-drinker	639	166	1.51 (0.96–2.37)	1.13 (0.69–1.85)
	Current drinker	891	211	1.38 (0.88–2.14)	1.03 (0.63–1.68)
Average lifetime total alcohol consumption (drinks/week)
	<7	432	94	1.30 (0.80–2.11)	1.00 (0.59–1.70)
	7 to <14	222	58	1.56 (0.93–2.62)	1.22 (0.69–2.17)
	14 to <28	196	56	1.70 (1.01–2.88)	1.16 (0.65–2.07)
	≥28	394	119	1.80 (1.12–2.90)	1.27 (0.74–2.19)
	p-trendc			0.05	0.24
Average lifetime beer consumption (drinks/week)
	<7	483	123	1.52 (0.94–2.44)	1.12 (0.66–1.89)
	7 to <14	140	47	2.00 (1.16–3.45)	1.44 (0.78–2.64)
	14 to <28	151	43	1.70 (0.98–2.94)	1.19 (0.65–2.19)
	≥28	311	84	1.61 (0.98–2.64)	1.11 (0.63–1.96)
	p-trendc			0.93	0.79
Average lifetime wine consumption (drinks/week)
	<7	174	54	1.85 (1.09–3.14)	1.47 (0.77–2.82)
	≥7	87	24	1.64 (0.88–3.07)	1.19 (0.55–2.54)
	p-trendc			0.67	0.42
Average lifetime liquor consumption (drinks/week)
	<7	406	102	1.50 (0.92–2.43)	1.08 (0.62–1.87)
	7 to <21	185	45	1.45 (0.84–2.49)	0.95 (0.51–1.77)
	≥21	179	58	1.93 (1.14–3.26)	1.37 (0.75–2.50)
	p-trendc			0.15	0.18

CI, confidence interval; OR, odds ratio.

a Adjusted for age, sex, race/ethnicity, highest level of education, body mass index, smoking status, and H. pylori infection.

b Reference group for all analyses is non-drinkers.

C p for trend excludes non-drinkers.

We also examined the risk of non-cardia gastric intestinal metaplasia associated with average lifetime total alcohol consumption as a continuous measure using both linear and nonlinear regression functions and assigning lifelong non-drinkers a value 0 drinks/week. Fig 1 shows the fitted spline regression curve and its 95% CI, indicating no association with gastric intestinal metaplasia across the full range of average consumption (up to an average of 140 drinks/week). There was no evidence for a nonlinear dose relationship (p = 0.46 testing for departure from linearity).

Fig 1

Dose-response relation of average lifetime total alcohol consumption (drinks/week) on risk of gastric intestinal metaplasia.

(a) solid line = odds ratio; (b) dotted line = 95% confidence interval for odds ratio; (c) dashed line = null hypothesis of no association; (d) extended dashed line = linear association.

As cases were more likely than controls to have a history of cigarette smoking and H. pylori infection, we examined for potential biological interactions between smoking and alcohol consumption and H. pylori and alcohol consumption. The highest risk for gastric intestinal metaplasia was observed among ever smokers with a history of alcohol consumption of ≥28 drinks/week (Table 3). Compared to individuals who were lifelong non-drinkers and also never smoked, individuals who were ever smokers and consumed on average ≥28 drinks/week of alcohol had 2-fold higher risk for gastric intestinal metaplasia (OR, 1.89; 95% CI, 1.00–3.20). However, when formally tested using the SI, we found no evidence for biological interaction of smoking with alcohol suggesting the majority of the increased risk is conferred by smoking, not alcohol (SI, 0.88; 95% CI, 0.13–6.06). Likewise, we found no evidence for biological interaction of H. pylori infection with alcohol consumption on risk of gastric intestinal metaplasia (SI, 1.41; 95% CI 0.52–3.79). H. pylori infection conferred higher risk of gastric intestinal metaplasia irrespective of amount of alcohol consumed (Table 4).

Table 3

Association of combined alcohol and smoking exposure and risk of gastric intestinal metaplasia.

	Controls	Cases
	N	N	Unadjusted	Adjusted
			OR (95% CI)	ORa (95% CI)
Never smoker, non-drinker	91	14	1.00 (ref)	1.00 (ref)
Never smoker, <14 drinks/week	174	29	1.08 (0.55–2.15)	1.02 (0.50–2.11)
Never smoker, 14 to <28 drinks/week	40	5	0.81 (0.27–2.41)	0.61 (0.20–1.90)
Never smoker, ≥28 drinks/week	53	14	1.72 (0.76–3.88)	1.49 (0.63–3.55)
Ever smoker, non-drinker	52	10	1.25 (0.52–3.01)	1.28 (0.50–3.25)
Ever smoker, <14 drinks/week	474	121	1.66 (0.91–3.01)	1.41 (0.75–2.65)
Ever smoker, 14 to <28 drinks/week	154	50	2.11 (1.11–4.03)	1.68 (0.85–3.34)
Ever smoker, ≥28 drinks/week	336	105	2.03 (1.11–3.72)	1.89 (1.00–3.20)
SI (95% CI)				0.88 (0.13–6.06)

CI, confidence interval; OR, odds ratio; SI, synergy index.

a Adjusted for age, sex, race/ethnicity, highest level of education, body mass index, and H. pylori infection.

Table 4

Association of combined alcohol and H. pylori infection exposure and risk of gastric intestinal metaplasia.

	Controls	Cases
	N	N	Unadjusted	Adjusted
			OR (95% CI)	ORa (95% CI)
H. pylori negative, non-drinker	115	11	1.00 (ref)	1.00 (ref)
H. pylori negative, <14 drinks/week	503	72	1.50 (0.77–2.91)	1.17 (0.58–2.33)
H. pylori negative, 14 to <28 drinks/week	143	24	1.75 (0.82–3.73)	1.31 (0.60–2.86)
H. pylori negative, ≥28 drinks/week	304	50	1.72 (0.86–3.42)	1.22 (0.59–2.51)
H. pylori positive, non-drinker	28	13	4.85 (1.97–12.0)	4.04 (1.59–10.2)
H. pylori positive, <14 drinks/week	140	78	5.82 (2.96–11.5)	3.94 (1.94–8.01)
H. pylori positive, 14 to <28 drinks/week	51	30	6.15 (2.86–13.2)	3.81 (1.71–8.51)
H. pylori positive, ≥28 drinks/week	84	68	8.46 (4.22–17.0)	5.59 (2.67–11.7)
SI (95% CI)				1.41 (0.52–3.79)

CI, confidence interval; OR, odds ratio; SI, synergy index.

a Adjusted for age, sex, race/ethnicity, highest level of education, body mass index, and smoking status.

Discussion

In this large cross-sectional study, we examined whether a relationship existed between quantity and type of alcohol consumption and the risk of non-cardia gastric intestinal metaplasia. While there was tentative evidence for a synergistic effect between tobacco smoke and total alcohol consumption on the risk of gastric intestinal metaplasia, total alcohol consumption alone was not associated with increased risk for gastric intestinal metaplasia. Most of the excess risk among individuals who smoked and consumed alcohol was the result of smoking, not alcohol consumption. Likewise, when stratified by specific beverage type, we found no associations between consumption of beer, wine and liquor/sprits and gastric intestinal metaplasia.

The role of alcohol in the pathway to development of gastric adenocarcinoma is still unclear and discrepant amongst particular cancer sub-types [2, 10, 12, 24]. Two meta-analyses conducted by Tramacere et al. concluded that “heavy” alcohol consumption (≥4 drinks per day) was associated with increased risk of non-cardia gastric adenocarcinoma [10] whereas there appears to be no association between any degree of alcohol consumption and the risks of cardia gastric adenocarcinoma or esophageal adenocarcinoma [12]. This discrepancy supports a protocol in which epidemiologic studies evaluating risk factors for gastric cancer specify regionality of the stomach being investigated. Additionally, this framework should be applied when considering risk factors for premalignant conditions, such as gastric intestinal metaplasia. Thus, our study specifically explored alcohol’s relationship, if any, with development of non-cardia gastric intestinal metaplasia.

Alcohol is a known carcinogen [8, 25, 26] and has been associated with many acute pathologic changes in the gut including, but not limited to, hemorrhagic lesions of the gastric mucosa [27] and dose-dependent modulatory effects on gastric acid secretion [28-32]. Given this, and alcohol’s association with non-cardia gastric adenocarcinoma previously mentioned, it is reasonable to inquire whether alcohol is also a risk factor for precursor lesions, such as intestinal metaplasia. Previous literature was sparse, especially in western populations, but predominately indicated no association between alcohol consumption and metaplasia formation [13, 14]. These prior studies alone are insufficient, however, in that neither regarded alcohol consumption as the primary variable of interest and neither stratified alcohol usage based off quantity consumed or beverage type. Given the understanding that effects of alcohol on the gastrointestinal tract are largely dependent on a variety of factors such as dose and lifetime use [33], further studies are warranted to adequately examine for any potential relationship. Our study used numerous stratification schemes to sub-classify alcohol use and search for an association with morphologic changes of the gastric mucosa. No degree of alcohol use correlated to any increased risk for non-cardia gastric intestinal metaplasia development, which provides the strongest evidence to date that there is no independent association between the two.

With our results promoting the absence of an association between alcohol and non-cardia gastric intestinal metaplasia and prior evidence reporting a relationship between heavy alcohol consumption and non-cardia gastric cancer [10], it is plausible to hypothesize that alcohol contributes to the later stages of the non-cardia gastric adenocarcinoma development pathway. Previous studies of upper gastrointestinal tract cancers found that alcohol is a strong risk factor for development of esophageal squamous cell carcinoma [26] and late premalignant conditions such as high grade dysplasia. However, there was no association with early precursor lesions such as low grade dysplasia of the esophagus [34]. This model does offer an interesting framework for how alcohol could plausibly pose a risk in later stages of gastric carcinogenesis. However, further studies are necessary to conclude the presence or absence of any such relationship.

This study has a number of strengths, such as well-defined cases and controls, specific classifications for categorical variables used in the multivariate analysis, and a large sample size. These factors contributed to this study’s ability to fully investigate any potential relationship between varying quantities of alcohol consumption, or consumption of any specific alcoholic beverage and the risk of gastric intestinal metaplasia. The comprehensive pre-study questionnaire provided to subjects allowed for extensive control of potential confounders. This provides assurance that data ascertained from this study directly represents the interplay between alcohol and non-cardia gastric intestinal metaplasia risk. Further, the administration of this questionnaire prior to the study EGD served to limit recall bias in the responses [19], which could otherwise alter the validity of the data. Finally, all EGD procedures and histological classifications of samples were based on established protocols [18]; this coupled with the blinding of the pathologists prevents potential sources of misclassification bias.

Limitations of this study include generalizability of the study population and potential heterogeneity of gastric intestinal metaplasia cases. The VA population predominately consisted of white males, which could hinder this study’s ability to make statements regarding this relationship in other ethnicities or in women. The overall response rate among the primary care group was 43% and this may have biased our results. However, participants who consent are generally healthier than the general population and, if one existed, this would strengthen an association. In regard to positive cases, the parameter set to define a case was a finding of metaplasia on ≥1 non-cardia gastric biopsy. While this aided our study in searching for any preliminary relationship, further studies could look more specifically at extensiveness of metaplasia, specific locations of metaplasia, and subtype of metaplasia. We included in our referent group a fraction of participants who may have ever consumed alcohol but who did not consume alcohol at least monthly for ≥6 months during their lifetime. As a result, we may have biased our risk estimates towards the null. Additionally, the risk estimates may be subject to reporting bias due to self-report alcohol intake. Although we attempted to limit this bias by having participants complete the study surveys prior to the study EGD (and knowledge of case-control study), and the bias in exposure status would likely be non-differential between cases and controls, it may have biased the association toward the null. Diet is an important risk factor for gastric cancer; however, we did not include diet in our study. Nonetheless, lack of adjustment for diet would not explain the null association because, as a confounder, diet would instead have exaggerated an observed association, instead of biasing it toward the null. Finally, as with all cross-sectional investigations, this study cannot demonstrate a causal relationship between two variables–only a relative strength or lack of association. A natural consequence of the cross-sectional design also meant that we could not match controls with cases. Nonetheless, this should not undermine the importance of the observed lack of association.

Conclusions

In conclusion, no specific beverage type or quantity of alcohol consumption posed any increased risk for non-cardia gastric intestinal metaplasia. Further, there was no demonstrable risk between total lifetime alcohol consumption and the same outcome. While there was some evidence of an additive effect between tobacco smoke and alcohol consumption on non-cardia gastric intestinal metaplasia risk, alcohol use to any degree alone may not confer increased risk for gastric intestinal metaplasia.

12 Jul 2021

PONE-D-21-15870

No association between alcohol consumption and the risk of gastric intestinal metaplasia

PLOS ONE

Dear Dr. Thrift,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

The manuscript has ben sent out to two outstanding external referees with a strong research background in epidemiology of alcohol consumption. I would invite the authors to carefully consider their comments attached below. In addition to the reviewers’ comments, this editor had a few points to be addressed:

Estimates were not adjusted for a proxy of socioeconomic status (SES), that is, the main determinants of gastric cancer. This was an important comment from Reviewer #1 that greatly limits your results. Please address this bias.

For what regards the dose-risk relationships, the figure appears to depict a linear trend. In fact, when I go through the manuscript text, I found the author’s sentence “There was no evidence for a nonlinear dose relationship (p=0.46 […]”. Therefore, it does not make sense to report a nonlinear trend when there was no evidence against linearity. Therefore, in figure 1 please report a linear relationship and use a logarithm scale for the y-axis. In fact, in the current figure version, the y axis range is inappropriate and not on a log-scale.

On page 13, the authors wrote “Among alcohol drinkers, the risk for gastric intestinal metaplasia did not increase linearly with increasing total average consumption (p-trend=0.14)”. There is something unclear as it seems that the trend cannot be considered non-linear being the p-value greater than 0.14. Please clarify.

Among the limitations, I would state that controls were not matched to cases, as a “natural” consequence of the cross-sectional design.

Please submit your revised manuscript by Aug 26 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Matteo Rota, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2.  Please include in your Methods section the date ranges over which you recruited participants to this study.

Furthermore, in your Methods section, please provide a justification for the sample size used in your study, including any relevant power calculations (if applicable).

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors found no association between cumulative lifetime alcohol intake and risk of gastric intestinal metaplasia.

Some remarks:

1) The authors stated that “participants who reported ever consuming alcohol but who did not consume alcohol at least monthly for ≥6 months were considered non-drinkers and included in the referent group”. An additional analysis in which these individuals are excluded from the referent group is appropriate. Please, make it available.

2) The authors stated that “The referent group for all analyses was those who were lifelong non-drinkers”. I suppose that ex-drinkers are not included in the referent group. This is correct, but I suggest to clary state this fact.

3) No measures/indexes of social status have been included as potential confounders. Due to the high importance that social status has as potential confounder for the association of alcohol intake with health outcome, authors should make any effort to include it in the multivariable model. On the contrary, the absence of any index of social status in the panel of covariates must be discussed as a strong limitation of the manuscript.

4) Another potentially important variable the authors should adjusted their models for is BMI, also because, beyond formal statistical significance, BMI shows some degree of association with case/control status (Table 1).

5) “Missing category was excluded from statistical tests for differences between controls and cases” (Table 1). However, it is not clear how the authors dealt with missing values in multivariable analyses (adjusted also for smoking and HP infection, that have missings). Please clarify. A case complete approach is not appropriate, and some more robust approach (multiple imputation, for example) for dealing with missing values have to be implemented.

6) The authors repeatedly have used terms as “(statistically) significant” or “non-significant”. It appears that some of their conclusions have been based only on statistically significance. This is not correct. Please, avoid it and base your conclusion more on effect size

7) Authors should provide complete data for biological interaction of H. pylori infection with alcohol consumption on risk of gastric intestinal metaplasia, in the same way they provided in Table 3 data on the association of combined alcohol and smoking exposure

Reviewer #2: This cross-sectional study investigated the relationship between self-reported lifetime alcohol consumption and beverage-specific consumption and gastric intestinal metaplasia in ~2000 US Veterans. This study is clearly written with a well-defined and justified research question (to look at alcohol’s role in precursor of gastric cancer rather than gastric cancer itself, which could add to understanding of alcohol’s role in gastric cancer development), and the statistical methods appeared to be sound. However, the study would benefit from further clarification of the study methods and importantly on the discussions of potential limitations of generalisability and self-reported bias of alcohol consumption. Please see the full list of suggestions below.

Title: I would suggest to avoid stating the conclusion (e.g. no association), but to include the study design and study population in the title.

Abstract:

1. It was slightly confusing to read the number of cases and number of controls separately in the first place, and later realise that it was a cross-sectional study but not a case-control study. I would suggest to first report the overall sample size recruited, and then report the number of cases identified.

2. Please state the full name of Houston VA Medical Center to make it clear the study population was among veterans.

3. Please state the study period (in Abstract and also in the Methods later).

Methods:

1. There was no mentioning of missing data until missing data was shown in Table 1. Please indicate in the Statistical analysis the number of participants with missing data and how missing data was handled in each of the subsequent analyses.

2. Page 7, please move the sentence “Overall, 70% of patients in the elective EGD …” up to the end of the Study population paragraph on page 6.

3. How many participants were recruited from each of the elective EGD group and of the primary care group, respectively?

4. Please specify how much one drink quantifies (e.g., in grams or units).

Results:

1. Page 11, please report the overall sample size and provide a summary of descriptive data of the whole study population in the beginning, before moving on to describe cases and controls separately.

2. Table 1 title: The terms case and controls are self-explanatory, so there is no need to repeat “controls without” and “cases with gastric intestinal metaplasia”. I would suggest to change the title along the lines of “Characteristics of cases and controls” or “Characteristics of US Veterans by case status”

3. Table 1: Please indicate the method used to derive the p-values.

4. Page 13, please report IQR for median of alcohol intake.

5. Table 2: Please be consistent with terminology e.g. non-drinkers (table and main text) vs. never drinkers (footnote).

6. Figure 1 title: Please remove “compared with controls” from the title.

Discussion

1. The authors briefly mentioned generalizability of the study population (predominantly white males) as a limitation. However, other aspect of the study population (Veterans only), as well as the different response rate between the elective EGD group and the primary care group (if the socio-demographics or health seeking behaviours differ between the two groups), may also have implications on the generalizability to the general population and bias. It would be good to see some discussion in the paper on these points.

2. Please discuss the potential bias of self-reported alcohol intake in the limitations.

3. Diet is an important risk factor for stomach cancer but was not measured in the study – this should be discussed in the limitation.

4. Page 20, final sentence “alcohol use to any degree alone proved to be an insufficient predisposing factor”. The word “proved” seems a bit too strongly conclusive, especially in the context of cross-sectional study - a more suggestive tone would be more appropriate.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Augusto Di Castelnuovo

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

9 Sep 2021

July 21, 2021

Matteo Rota, Ph.D.

Academic Editor

PLOS ONE

Dear Dr. Rota,

Thank you for considering our manuscript, and for the helpful comments from the reviewers which accompanied your recent letter. We have taken the opportunity of revising our manuscript in accordance with those comments and herewith submit it to you for further consideration.

Comments from the Editor

Comment #1:

Estimates were not adjusted for a proxy of socioeconomic status (SES), that is, the main determinants of gastric cancer. This was an important comment from Reviewer #1 that greatly limits your results. Please address this bias.

Reply to Comment #1:

We thank the Editor and Reviewer for this important comment. We captured educational background in our study survey, which we now use here as a proxy for social status. As expected, we show in Table 1 that highest level of education is strongly associated with case-control status (cases are more likely to have high school or less education). We have therefore added highest level of education as a covariate in our multivariable models. Additionally adjusting for education level did attenuate the associations with alcohol consumption to the null. We believe adding this data to the paper strengthens our methods and helps to strengthen confidence in our findings.

Comment #2:

For what regards the dose-risk relationships, the figure appears to depict a linear trend. In fact, when I go through the manuscript text, I found the author’s sentence “There was no evidence for a nonlinear dose relationship (p=0.46 […]”. Therefore, it does not make sense to report a nonlinear trend when there was no evidence against linearity. Therefore, in figure 1 please report a linear relationship and use a logarithm scale for the y-axis. In fact, in the current figure version, the y axis range is inappropriate and not on a log-scale.

Reply to Comment #2:

A common approach to the analysis of alcohol consumption in cancer (and other disease) studies has been to categorize participants based on broad measures of alcohol intake. Such an approach has the advantage of deriving an easily interpretable measure of relative risk, and makes no assumption about any underlying trends or patterns of risk. However, categorization may obscure potentially important differences in risk within and across categories of alcohol intake, particularly if the cut-off points for categorization span biologically important thresholds. Modeling risk with alcohol intake as a continuous measure by using standard linear regression models avoids the potential loss of information owing to categorization, but it constrains the dose effect to be linear. Again, such constraints may mask important differences in risk associated with different levels of alcohol intake. As has been well described, alcohol consumption has a J-shaped relationship with all-cause mortality in human beings and there is no reason to discount the possibility that similar nonlinear patterns of risk may be observed for stomach cancer and its precursor, gastric intestinal metaplasia. In the current manuscript, we use a suite of approaches to address this issue. In total, our analyses consistently demonstrate no association with alcohol consumption. The splines demonstrate no hidden/missing non-linear relationship. The lack of linear association across categories is accurate, and there is no ‘hidden’ patterns of risk obscured by the categorization or assumption of linearity.

Comment #3:

On page 13, the authors wrote “Among alcohol drinkers, the risk for gastric intestinal metaplasia did not increase linearly with increasing total average consumption (p-trend=0.14)”. There is something unclear as it seems that the trend cannot be considered non-linear being the p-value greater than 0.14. Please clarify.

Reply to Comment #3:

There is actually no association with alcohol consumption. The associations with categories of alcohol consumption demonstrated a lack of linear association (no increase with increasing category of exposure). The splines show that there is also no non-linear association. Together, these analyses address issues of categorization and assumptions of linearity. They strengthen our conclusion that alcohol consumption does not increase a person’s risk for gastric intestinal metaplasia.

Comment #4:

Among the limitations, I would state that controls were not matched to cases, as a “natural” consequence of the cross-sectional design.

Reply to Comment #4:

We added the following to the Discussion (version with tracked changes; page 24, paragraph 1):

“A natural consequence of the cross-sectional design also meant that we could not match controls with cases.”

Reviewer #1

Comment #1:

The authors stated that “participants who reported ever consuming alcohol but who did not consume alcohol at least monthly for ≥6 months were considered non-drinkers and included in the referent group”. An additional analysis in which these individuals are excluded from the referent group is appropriate. Please, make it available.

Reply to Comment #1:

Unfortunately, the alcohol questions in our study survey were written such that we did not capture any alcohol history among persons who may have ever consumed alcohol but who did not consume alcohol at least monthly for ≥6 months. This was the qualifying question for capturing specific alcohol consumption. We have now noted this as a limitation of our study (page 23, paragraph 2):

“We included in our referent group a fraction of participants who may have ever consumed alcohol but who did not consume alcohol at least monthly for ≥6 months during their lifetime. As a result, we may have biased our risk estimates towards the null.”

Comment #2:

The authors stated that “The referent group for all analyses was those who were lifelong non-drinkers”. I suppose that ex-drinkers are not included in the referent group. This is correct, but I suggest to clary state this fact.

Reply to Comment #2:

The Reviewer is correct. Ex-drinkers were not included in the referent group. We revised the Methods (page 10, paragraph 2) as follows:

“For all analyses, we compared history of alcohol consumption among ever drinkers (current and ex-drinkers) to a referent group of those who were lifelong non-drinkers.”

Comment #3:

No measures/indexes of social status have been included as potential confounders. Due to the high importance that social status has as potential confounder for the association of alcohol intake with health outcome, authors should make any effort to include it in the multivariable model. On the contrary, the absence of any index of social status in the panel of covariates must be discussed as a strong limitation of the manuscript.

Reply to Comment #3:

We thank the Reviewer for this important comment. We captured educational background in our study survey, which we now use here as a proxy for social status. As expected, we show in Table 1 that highest level of education is strongly associated with case-control status (cases are more likely to have high school or less education). We have therefore added highest level of education as a covariate in our multivariable models. Additionally adjusting for education level did attenuate the risk estimates for associations with alcohol consumption to the null.

Comment #4:

Another potentially important variable the authors should adjusted their models for is BMI, also because, beyond formal statistical significance, BMI shows some degree of association with case/control status (Table 1).

Reply to Comment #4:

We agree with the Reviewer and have now adjusted our models for BMI.

Comment #5:

“Missing category was excluded from statistical tests for differences between controls and cases” (Table 1). However, it is not clear how the authors dealt with missing values in multivariable analyses (adjusted also for smoking and HP infection, that have missings). Please clarify. A case complete approach is not appropriate, and some more robust approach (multiple imputation, for example) for dealing with missing values have to be implemented.

Reply to Comment #5:

We apologize for excluding this information from the Methods. As shown in Table 1, we had missing data on BMI, WHR, H. pylori infection, and smoking status among a minority of study participants. We added the following to the Methods (page 10, paragraph 2):

“We included participants with missing data for covariates (e.g., BMI, WHR, H. pylori infection, smoking status) in the analyses using an additional category for missing values.”

Comment #6:

The authors repeatedly have used terms as “(statistically) significant” or “non-significant”. It appears that some of their conclusions have been based only on statistically significance. This is not correct. Please, avoid it and base your conclusion more on effect size.

Reply to Comment #6:

We have revised our manuscript throughout to remove emphasis on “statistical significance”.

Comment #7:

Authors should provide complete data for biological interaction of H. pylori infection with alcohol consumption on risk of gastric intestinal metaplasia, in the same way they provided in Table 3 data on the association of combined alcohol and smoking exposure.

Reply to Comment #7:

We have now added this data in Table 4 (page 20).

Reviewer #2

Title: I would suggest to avoid stating the conclusion (e.g. no association), but to include the study design and study population in the title.

Reply: We have changed our title to “Alcohol consumption and the risk of gastric intestinal metaplasia in a U.S. Veterans population”.

Abstract:

1. It was slightly confusing to read the number of cases and number of controls separately in the first place, and later realise that it was a cross-sectional study but not a case-control study. I would suggest to first report the overall sample size recruited, and then report the number of cases identified.

Reply: We revised the Abstract (page 3) as follows:

“We used data from 2084 participants (including 403 with gastric intestinal metaplasia) recruited between February 2008-August 2013 into a cross-sectional study at the Michael E. DeBakey Veterans Affairs Medical Center in Houston, Texas.”

2. Please state the full name of Houston VA Medical Center to make it clear the study population was among veterans.

Reply: We replaced “Houston VA Medical Center” with “Michael E. DeBakey Veterans Affairs Medical Center in Houston, Texas”.

3. Please state the study period (in Abstract and also in the Methods later).

Reply: We added the study period to both the Abstract (page 3) and Methods (page 7).

Methods:

1. There was no mentioning of missing data until missing data was shown in Table 1. Please indicate in the Statistical analysis the number of participants with missing data and how missing data was handled in each of the subsequent analyses.

Reply: We apologize for excluding this information from the Methods. As shown in Table 1, we had missing data on BMI, WHR, H. pylori infection, and smoking status among a minority of study participants. We added the following to the Methods (page 10, paragraph 2):

“We included participants with missing data for covariates (e.g., BMI, WHR, H. pylori infection, smoking status) in the analyses using an additional category for missing values.”

2. Page 7, please move the sentence “Overall, 70% of patients in the elective EGD …” up to the end of the Study population paragraph on page 6.

Reply: We moved this sentence as suggested.

3. How many participants were recruited from each of the elective EGD group and of the primary care group, respectively?

Reply: We added this information to the Results (page 12, paragraph 1):

“This study included data from 2084 participants, with 1568 recruited from EGD clinics and 516 from primary care.”

4. Please specify how much one drink quantifies (e.g., in grams or units).

Reply: We now noted in the Methods that one standard drink was 10g of alcohol.

Results:

1. Page 11, please report the overall sample size and provide a summary of descriptive data of the whole study population in the beginning, before moving on to describe cases and controls separately.

Reply: We revised the Results (page 12, paragraph 1) as follows:

“This study included data from 2084 participants, with 1568 recruited from EGD clinics and 516 from primary care. The mean age of participants was 60.2 years (standard deviation, 8.1 years). Ninety-two percent of participants were male, 57.3% White and 31.3% Black. Most participants were overweight or obese (81.4%) and reported a history of smoking (68.5%).”

2. Table 1 title: The terms case and controls are self-explanatory, so there is no need to repeat “controls without” and “cases with gastric intestinal metaplasia”. I would suggest to change the title along the lines of “Characteristics of cases and controls” or “Characteristics of US Veterans by case status”

Reply: We changed the title to “Characteristics of cases and controls.

3. Table 1: Please indicate the method used to derive the p-values.

Reply: We added this to the footnote of Table 1.

4. Page 13, please report IQR for median of alcohol intake.

Reply: We revised the results (page 15, paragraph 1) as follows:

“Among controls who reported being ever drinkers, the median lifetime total alcohol consumption was 13 (interquartile range [IQR], 6-42) and 4 (IQR, 2-10) drinks/week for males and females, respectively. Median lifetime total alcohol consumption was numerically higher among male (16 drinks/week; IQR, 6-42) and female (8 drinks/week; IQR, 2-14) cases with non-cardia gastric intestinal metaplasia than controls.”

5. Table 2: Please be consistent with terminology e.g. non-drinkers (table and main text) vs. never drinkers (footnote).

Reply: We placed “never drinkers” with “non-drinkers” in the footnote (page 17).

6. Figure 1 title: Please remove “compared with controls” from the title.

Reply: We removed “compared with controls” from the figure title.

Discussion

1. The authors briefly mentioned generalizability of the study population (predominantly white males) as a limitation. However, other aspect of the study population (Veterans only), as well as the different response rate between the elective EGD group and the primary care group (if the socio-demographics or health seeking behaviours differ between the two groups), may also have implications on the generalizability to the general population and bias. It would be good to see some discussion in the paper on these points.

Reply: We added the following to the Discussion (page 23, paragraph 2):

“The overall response rate among the primary care group was 43% and this may have biased our results. However, participants who consent are generally healthier than the general population and, if one existed, this would strengthen an association.”

2. Please discuss the potential bias of self-reported alcohol intake in the limitations.

Reply: We added the following to the Discussion (page 24, paragraph 1):

“Additionally, the risk estimates may be subject to reporting bias due to self-report alcohol intake. We attempted to limit this bias by having participants complete the study surveys prior to the study EGD (and knowledge of case-control study), increasing the likelihood that any bias would be non-differential between cases and controls.”

3. Diet is an important risk factor for stomach cancer but was not measured in the study – this should be discussed in the limitation.

Reply: We added the following to the Discussion (page 24, paragraph 1):

“Diet is an important risk factor for gastric cancer; however, we did not include diet in our study, which may have confounded our associations with alcohol consumption.”

4. Page 20, final sentence “alcohol use to any degree alone proved to be an insufficient predisposing factor”. The word “proved” seems a bit too strongly conclusive, especially in the context of cross-sectional study - a more suggestive tone would be more appropriate.

Reply: We revised the sentence (page 24, paragraph 2) as follows:

“While there was some evidence of an additive effect between tobacco smoke and alcohol consumption on non-cardia gastric intestinal metaplasia risk, alcohol use to any degree alone may not confer increased risk for gastric intestinal metaplasia.”

We look forward to your consideration of our revised manuscript, and trust that these revisions meet with your approval.

Yours sincerely,

Aaron P. Thrift, PhD

On behalf of all authors

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

7 Oct 2021

PONE-D-21-15870R1​Alcohol consumption and the risk of gastric intestinal metaplasia in a U.S. Veterans populationPLOS ONE

Dear Dr. Thrift,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

There are still two minor comments that need to be better addressed before taking a positive decision on the manuscript. See below. This is a minor revision. In addition to these reviewer #2 comments, I would ask to the authors to better set the y-axis limits of Figure 1, now ranging from 0.1 to 10. I suggest to set a range from 0.25 to 4, using a logarithmic scale, as common for plots of dose-risk relationships.

I also want to point out that my previous comment #2 was partly misunderstood by the authors. The assessment of dose-risk relationship shape trough restricted cubic spline (rcs) showed no evidence of a non-linear association. This is fine, but from a statistical point of view, this means that there is no reason to consider a less parsimonious model (as the rcs is) when no evidence of a better fit emerged as compared to the linear line. However, I still agree with the author choice to show the rcs dose-risk relationship, but I also suggest to report the linear line in a supplementary figure, if not possible within Figure 1.

Please submit your revised manuscript by Nov 21 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Matteo Rota, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments (if provided):

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors provided a satisfactory revision

Reviewer #2: The authors have generally done well in revising the manuscript and responding to the reviewers’ comments. However, I still have a few minor comments:

1. Limitation paragraph, discussion on self-report bias in alcohol consumption: even if the bias in exposure status is non-differential between cases and controls, non-differential misclassification of the exposure can still bias the association toward the null.

2. Limitation paragraph, discussion on diet: The first half of the sentence “Diet is an important risk factor for gastric cancer; however, we did not include diet in our study…” sounds fine as poor diet/nutrition is associated with heavy drinking, and is a possible confounder in any observed association between alcohol consumption and gastric cancer-related outcome. However, the second part of the sentence “… which may have confounded our associations with alcohol consumption” is a bit be confusing in this context since there is no observed association between alcohol consumption and risk of gastric intestinal metaplasia in this study. I would suggest the authors to revise the second part of sentence to discuss this issue more specifically in the context of this study. For example, the lack of diet info in this study is unlikely to have major impact on your conclusions (since as a confounder diet would have exaggerated the observed associations, instead of biasing it toward the null).

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Augusto Di Castelnuovo

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

22 Oct 2021

October 20, 2021

Matteo Rota, Ph.D.

Academic Editor

PLOS ONE

Dear Dr. Rota,

Thank you for considering our revised manuscript, and for the additional comments in your recent letter. We have taken the opportunity of revising our manuscript in accordance with those comments and herewith submit it to you for further consideration.

Comments from the Editor

Comment #1:

There are still two minor comments that need to be better addressed before taking a positive decision on the manuscript. See below. This is a minor revision. In addition to these reviewer #2 comments, I would ask to the authors to better set the y-axis limits of Figure 1, now ranging from 0.1 to 10. I suggest to set a range from 0.25 to 4, using a logarithmic scale, as common for plots of dose-risk relationships.

I also want to point out that my previous comment #2 was partly misunderstood by the authors. The assessment of dose-risk relationship shape trough restricted cubic spline (rcs) showed no evidence of a non-linear association. This is fine, but from a statistical point of view, this means that there is no reason to consider a less parsimonious model (as the rcs is) when no evidence of a better fit emerged as compared to the linear line. However, I still agree with the author choice to show the rcs dose-risk relationship, but I also suggest to report the linear line in a supplementary figure, if not possible within Figure 1.

Reply to Comment #1:

We apologize for our misunderstanding your prior comment. We have revised the Figure to set the y-axis limits as suggested and added a linear line. We hope that this new Figure better reflects the findings from our study.

Reviewer #2

The authors have generally done well in revising the manuscript and responding to the reviewers’ comments. However, I still have a few minor comments:

Comment #1:

Limitation paragraph, discussion on self-report bias in alcohol consumption: even if the bias in exposure status is non-differential between cases and controls, non-differential misclassification of the exposure can still bias the association toward the null.

Reply to Comment #1:

Thank you for this important comment. We have revised the Discussion (page 22, paragraph 1) as follows:

“Additionally, the risk estimates may be subject to reporting bias due to self-report alcohol intake. Although we attempted to limit this bias by having participants complete the study surveys prior to the study EGD (and knowledge of case-control study), and the bias in exposure status would likely be non-differential between cases and controls, it may have biased the association toward the null.”

Comment #2:

Limitation paragraph, discussion on diet: The first half of the sentence “Diet is an important risk factor for gastric cancer; however, we did not include diet in our study…” sounds fine as poor diet/nutrition is associated with heavy drinking, and is a possible confounder in any observed association between alcohol consumption and gastric cancer-related outcome. However, the second part of the sentence “… which may have confounded our associations with alcohol consumption” is a bit be confusing in this context since there is no observed association between alcohol consumption and risk of gastric intestinal metaplasia in this study. I would suggest the authors to revise the second part of sentence to discuss this issue more specifically in the context of this study. For example, the lack of diet info in this study is unlikely to have major impact on your conclusions (since as a confounder diet would have exaggerated the observed associations, instead of biasing it toward the null).

Reply to Comment #2:

We have revised the sentence (page 22, paragraph 1) as follows:

“Diet is an important risk factor for gastric cancer; however, we did not include diet in our study. Nonetheless, lack of adjustment for diet would not explain the null association because, as a confounder, diet would instead have exaggerated an observed association, instead of biasing it toward the null.”

We look forward to your consideration of our revised manuscript, and trust that these revisions meet with your approval.

Yours sincerely,

Aaron P. Thrift, PhD

On behalf of all authors

Associate Professor

Section of Epidemiology and Population Sciences

Department of Medicine

Baylor College of Medicine

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

2 Nov 2021

​Alcohol consumption and the risk of gastric intestinal metaplasia in a U.S. Veterans population

PONE-D-21-15870R2

Dear Dr. Thrift,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Matteo Rota, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: (No Response)

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: (No Response)

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: (No Response)

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: (No Response)

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

5 Nov 2021

PONE-D-21-15870R2

Alcohol consumption and the risk of gastric intestinal metaplasia in a U.S. Veterans population

Dear Dr. Thrift:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Matteo Rota

Academic Editor

PLOS ONE