Summary
Independent of Helicobacter pylori, increased PGE-M in urine is associated with gastric cancer. Higher PGE-M levels may indicate additional risk for gastric cancer even among individuals infected with the more virulent strains of H. pylori in East Asia.
Keywords: Helicobacter pylori, stomach neoplasms, serology, inflammation.
Abstract
Background.
Previous studies suggest that a stable end-product of prostaglandin E2, the urinary metabolite PGE-M, is associated with colorectal cancer, and 1 study of relatively small sample size found an association with gastric cancer among women. In the present study we further investigate the PGE-M, Helicobacter pylori, and gastric cancer association.
Methods.
The present analysis included 359 prospectively ascertained gastric cancer cases and 700 individually matched controls from the Shanghai Women’s and Men’s Health Studies. Urinary PGE-M was measured by a liquid chromatography/tandem mass spectrometric method. Seropositivity to 15 H. pylori recombinantly expressed fusion proteins was detected by H. pylori multiplex serology.
Results.
Adjusting for H. pylori, increasing PGE-M was associated with higher risk of gastric cancer (quartile 4 vs 1: odds ratio [OR], 1.76 [95% confidence interval {CI}, 1.17–2.66], Ptrend = .004). This association remained after excluding those diagnosed within 2 years from sample collection (OR, 1.73 [95% CI, 1.12–2.65], Ptrend = .007). However it was no longer present among individuals with 10 or more years of follow-up (2–4.9 years: OR, 3.15 [95% CI, 1.11–8.91]; 5–9.9 years: OR, 2.23 [95% CI, 1.22–4.06]; ≥10 years: OR, 0.73 [95% CI, .31–1.70]). Compared to H. pylori–negative individuals with below-median PGE-M levels, H. pylori–positive individuals with above-median PGE-M levels had a 5-fold increase in the odds of gastric cancer (OR, 5.08 [95% CI, 2.47–10.43]).
Conclusions.
In China, higher PGE-M levels may indicate an increased risk of gastric cancer independent of the risk conferred by H. pylori infection status, particularly for cancers diagnosed within 10 years of sample collection.
Multiple lines of evidence have indicated an important role for aberrant arachidonic acid metabolism in both inflammation and carcinogenesis [1, 2]. Cyclooxygenase 2 (COX-2), a rate-limiting enzyme, catalyzes the conversion of arachidonic acid to prostaglandin H2 (PGH2) [3, 4]. PGH2 is converted by microsomal prostaglandin E synthase to a key inflammatory mediator, prostaglandin E2 (PGE2) [5]. Upregulation of COX-2 or overproduction of PGE2 is found in various malignancies and is also associated with poor prognosis [6–9]. Catabolism of PGE2 is initiated by 15-hydroxyprostaglandin dehydrogenase and results in a stable end metabolite, PGE-M (11α-hydroxy-9, 15-dioxo-2, 3, 4, 5-tetranorprostane-1, 20-dioic acid), which is excreted in the urine and used as an index of systemic PGE2 production [10, 11].
Levels of PGE-M in healthy individuals are suppressed by both the nonselective COX inhibitor and the COX-2 selective inhibitor [11]. Previously, our findings suggested a linear association between urinary PGE-M levels and gastric cancer risk in women [12]. Importantly, Helicobacter pylori infection is the leading causal factor for gastric cancer, and COX-2 overexpression might be a significant step in H. pylori–associated gastric carcinogenesis. Hence, in the present study, with longer follow-up and more cases in both sexes, we investigated this association with consideration of H. pylori infection in Shanghai, China, an area with documented high rates of gastric cancer and H. pylori infection.
METHODS
Study Design and Study Population
The Shanghai Women’s Health Study (SWHS) and the Shanghai Men’s Health Study (SMHS), 2 population-based prospective cohort studies located in urban Shanghai, China, provided participants for the current nested case-control study. The SWHS recruited 74942 Chinese women aged 40–70 years from 1996 to 2000, while the SMHS recruited 61480 men aged 40–74 years from 2002 to 2006 [13, 14].
Using the International Classification of Diseases for Oncology (including codes C16.0–C16.6, and C16.9), incident gastric cancer was ascertained through registry linkage and active follow-ups. Incidence density sampling was used, with 2 controls chosen at random for each gastric cancer case from the appropriate risk sets consisting of all cohort members alive, free of cancer (excluding nonmelanoma skin cancer), and without history of gastrectomy at the time of diagnosis of the index case. Our final analysis included 359 prospectively ascertained gastric cancer cases with 700 controls matched on sex, age, and date of specimen collection. Written informed consent was provided by all participants and this study was approved by the institutional review boards of Vanderbilt University (Nashville, Tennessee); the German Cancer Research Center (Heidelberg, Germany); and the Shanghai Cancer Institute (Shanghai, China).
Data Collection and Laboratory Methods
At study enrollment, all participants were interviewed in-person by a trained interviewer using a structured questionnaire. The questionnaire consists of the following sections: demographic background, medical history, personal habits, dietary habits, physical activity, family history of cancer, occupational history, weight and height history, reproductive history, and residential history [14, 15].
Urinary PGE-M was measured by a liquid chromatography/tandem mass spectrometric method, as described previously [11, 16]. In brief, 0.75 mL of urine was acidified to pH 3 with HCl and endogenous PGE-M was converted to the O-methyloxime derivative. The methoximated PGE-M was extracted after incubation, applied to a C-18 Sep-Pak, and then eluted with ethyl acetate. With an internal standard of 2H6-O-methyloxime PGE-M, samples were then analyzed by liquid chromatography on a Zorbax Eclipse XDB-C18 column attached to a ThermoFinnigan Surveyor MS Pump. For endogenous PGE-M, the predominant product ion m/z 336 representing [M- (OCH3 + H2O)]− and the analogous ion m/z 339 [M-OC (2H3 + H2O)]− for the deuterated internal standard were monitored in the selected reaction monitoring mode. Quantification of endogenous PGE-M was calculated by comparing the ratio of the mass chromatogram peak areas of the m/z 336 and m/z 339 ions. To account for variations in hydration status, urinary PGE-M levels were standardized using the urinary creatinine levels of each sample and are expressed as nanograms per milligram (ng/mg) creatinine. Urinary creatinine levels were measured using a kit (Sigma). Staff was blinded to case and control status, and duplicate quality-control samples were interspersed among samples. The intra-assay and interassay coefficients of variation for PGE-M levels adjusted for creatinine were 10.8% and 19.9%, respectively.
Based on the glutathione S-transferase capture immunosorbent assay combined with fluorescent bead technology (Luminex, Austin, Texas), human immunoglobulin A, M, and G antibodies to 15 H. pylori recombinantly expressed fusion proteins (UreA, Catalase, GroEL, NapA, CagA, CagM, Cagδ, HP 0231, VacA, HpaA, Cad, HyuA, Omp, HcpC, and HP 0305) were detected by H. pylori multiplex serology simultaneously [17, 18]. Overall, H. pylori seropositivity was defined as ≥4 seropositive results of the 15 H. pylori antigens assessed, in accordance with previous validation applying commercial serological assay classification [17].
Statistical Analyses
The present analysis includes 1059 participants. Quartile cut points were determined using the distribution of PGE-M among controls. Differences of baseline characteristics across quartiles of PGE-M were compared using χ2 tests for categorical variables and Kruskal-Wallis tests for continuous variables.
To determine the associations between the prevalence of H. pylori infection and PGE-M levels, analysis of variance was applied to calculate mean PGE-M levels within groups defined by H. pylori dichotomous status, or Omp and HP 0305 status (comparing Omp– and HP 0305–, Omp– or HP 0305–, and Omp+ and HP 0305+) [19]. As PGE-M levels were not normally distributed, they were log-transformed. Statistical adjustment was made for education, body mass index (BMI), and smoking status with cases and controls matched by sex, age at enrollment, and date of specimen collection.
Conditional logistic regression models were constructed to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for gastric cancer risk across quartiles of PGE-M. In Table 3, model 1 evaluated the odds of gastric cancer incidence by PGE-M level conditioned on matched case-control sets only. Then in model 2 the potential confounders of education, BMI, and smoking were added. Model 3 adjusted for all variables in model 2 together with H. pylori infection status categorized dichotomously. Model 4 then adjusted for H. pylori based on seropositivity of combined Omp and HP 0305 status instead of H. pylori dichotomously, as these H. pylori proteins were previously found to be better markers than H. pylori seropositivity of gastric cancer risk in our population [19]. Tests for linear trend were performed by entering the categorical variables as continuous parameters. With the same analysis strategy as in Table 3, stratified analyses were conducted by follow-up period (2–4.9 years, 5–9.9 years, ≥10 years) from sample collection to diagnosis. Again with the same covariates as used in model 2, the combined effect of PGE-M and H. pylori was evaluated. We also examined effect modification by H. pylori status using a likelihood ratio test to compare models with and without interaction terms. Sensitivity analyses were performed excluding cases diagnosed within 2 years after sample collection, current smokers, or ever drinkers and their matched controls, which were suggested to be associated with PGE-M levels and gastric cancer [20–22]. Statistical analyses were conducted with SAS statistical software version 9.4 (SAS Institute, Cary, North Carolina). P values of <.05 (2-sided probability) were interpreted as being statistically significant for all analyses.
Table 3.
Risk of Gastric Cancer by Urinary PGE-M Levels in the Helicobacter pylori Biomarker Cohort Consortium
| PGE-M Level (Quartiles, ng/mg Creatinine) | Cases/Controls, No. | OR (95% CI)a | OR (95% CI)b | OR (95% CI)c | OR (95% CI)d |
|---|---|---|---|---|---|
| Overall | |||||
| Q1 | 69/176 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 75/174 | 1.15 (.77–1.72) | 1.06 (.70–1.60) | 1.05 (.69–1.60) | 0.98 (.64–1.51) |
| Q3 | 92/175 | 1.49 (1.00–2.23) | 1.40 (.93–2.10) | 1.26 (.83–1.91) | 1.22 (.80–1.86) |
| Q4 | 123/175 | 2.03 (1.36–3.01) | 1.92 (1.28–2.88) | 1.76 (1.17–2.66) | 1.72 (1.13–2.62) |
| Ptrend | … | <.001 | <.001 | .004 | .004 |
| 2–4.9 y of follow-up from sample collection to diagnosis | |||||
| Q1 | 11/35 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 14/25 | 1.73 (.63–4.70) | 1.86 (.65–5.30) | 1.90 (.62–5.76) | 1.91 (.57–6.42) |
| Q3 | 15/35 | 1.48 (.55–3.98) | 1.54 (.57–4.19) | 1.54 (.55–4.32) | 1.73 (.56–5.39) |
| Q4 | 24/29 | 2.87 (1.10–7.49) | 3.29 (1.21–8.94) | 3.15 (1.11–8.91) | 2.95 (.98–8.86) |
| Ptrend | … | .044 | .031 | .046 | .071 |
| 5–9.9 y of follow-up from sample collection to diagnosis | |||||
| Q1 | 33/96 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 34/80 | 1.30 (.73–2.29) | 1.16 (.64–2.11) | 1.10 (.60–2.02) | 1.05 (.57–1.97) |
| Q3 | 41/80 | 1.69 (.95–3.01) | 1.59 (.88–2.89) | 1.42 (.78–2.62) | 1.41 (.75–2.63) |
| Q4 | 64/81 | 2.69 (1.53–4.73) | 2.47 (1.37–4.45) | 2.23 (1.22–4.06) | 2.20 (1.19–4.10) |
| Ptrend | … | <.001 | .002 | .006 | .007 |
| ≥10 y of follow-up from sample collection to diagnosis | |||||
| Q1 | 21/37 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
| Q2 | 22/54 | 0.73 (.34–1.57) | 0.63 (.28–1.42) | 0.66 (.29–1.53) | 0.54 (.23–1.29) |
| Q3 | 29/42 | 1.34 (.62–2.87) | 1.22 (.55–2.72) | 1.08 (.48–2.44) | 0.98 (.42–2.27) |
| Q4 | 22/49 | 0.82 (.37–1.83) | 0.77 (.33–1.76) | 0.73 (.31–1.70) | 0.69 (.29–1.66) |
| Ptrend | … | .950 | .989 | .775 | .806 |
Abbreviations: CI, confidence interval; OR, odds ratio; PGE-M, prostaglandin E2 metabolite; Q, quartile.
aConditional logistic regression with cases and controls matched on sex, age, and date of biological collection.
bConditional logistic regression with cases and controls matched on sex, age, and date of biological collection, adjusted for education, body mass index (BMI), and smoking.
cConditional logistic regression with cases and controls matched on sex, age, and date of biological collection, adjusted for education, BMI, smoking, and Helicobacter pylori infection (positive or negative).
dConditional logistic regression with cases and controls matched on sex, age and date of biological collection, adjusted for education, BMI, smoking, and Omp and HP 0305 status (Omp– and HP 0305–, OMP+ or HP 0305+, and Omp+ and HP 0305+).
RESULTS
Study Cohorts’ Characteristics
Among 1059 participants, individuals with higher PGE-M quartiles were more likely to be older, current smokers, less educated, and seropositive to H. pylori overall and seropositive to both H. pylori proteins Omp and HP 0305 specifically (Table 1). The differences among quartiles for PGE-M in follow-up time, sex, ever drinkers, BMI, nonsteroidal anti-inflammatory drugs (NSAIDs) or cold medication use, and history of gastritis were not statistically significant.
Table 1.
Distribution of Selected Demographic Characteristics Across Quartiles of Urinary PGE-M Levels at Baseline in the Helicobacter pylori Biomarker Cohort Consortium
| Characteristic | PGE-M (ng/mg Creatinine) | P Value | |||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | ||
| Age, y, median (IQR) | 55.6 (48.1–64.0) | 60.0 (48.6–65.5) | 61.9 (54.1–66.4) | 63.3 (55.9–67.3) | <.001 |
| Follow-up time, y, median (IQR) | 7.8 (5.1–9.9) | 7.8 (5.1–10.6) | 7.7 (4.1–10.2) | 7.0 (4.7–9.6) | .356 |
| Sex | .930 | ||||
| Female | 198 (80.8) | 206 (82.7) | 219 (82.0) | 241 (80.9) | |
| Male | 47 (19.2) | 43 (17.3) | 48 (18.0) | 57 (19.1) | |
| Smoking | |||||
| Female | .058 | ||||
| Noncurrent smoker | 193 (97.5) | 203 (98.5) | 208 (95.0) | 227 (94.2) | |
| Current smoker | 5 (2.5) | 3 (1.5) | 11 (5.0) | 14 (5.8) | |
| Male | .028 | ||||
| Noncurrent smoker | 29 (61.7) | 30 (69.8) | 22 (45.8) | 25 (43.9) | |
| Current smoker | 18 (38.3) | 13 (30.2) | 26 (54.2) | 32 (56.1) | |
| Drinking | |||||
| Female | .288 | ||||
| Never drinker | 195 (98.5) | 201 (97.6) | 212 (96.8) | 239 (99.2) | |
| Ever drinker | 3 (1.5) | 5 (2.4) | 7 (3.2) | 2 (0.8) | |
| Male | .277 | ||||
| Never drinker | 30 (63.8) | 27 (62.8) | 35 (72.9) | 31 (54.4) | |
| Ever drinker | 17 (36.2) | 16 (37.2) | 13 (27.1) | 26 (45.6) | |
| Educationa | <.001 | ||||
| Elementary school or less | 54 (22.0) | 76 (30.5) | 107 (40.1) | 146 (49.0) | |
| Junior high school | 81 (33.1) | 95 (38.2) | 79 (29.6) | 74 (24.8) | |
| High school | 66 (26.9) | 62 (24.9) | 45 (16.9) | 45 (15.1) | |
| Professional education or above | 43 (17.6) | 16 (6.4) | 35 (13.0) | 31 (10.4) | |
| Missing | 1 (0.4) | 0 (0) | 1 (0.4) | 2 (0.7) | |
| BMI, kg/m2 | .217 | ||||
| <23.0 | 88 (35.9) | 95 (38.1) | 87 (32.6) | 86 (28.9) | |
| 23.0–24.9 | 66 (26.9) | 49 (19.7) | 56 (21.0) | 71 (23.8) | |
| 25.0–27.4 | 53 (21.7) | 56 (22.5) | 67 (25.0) | 75 (25.2) | |
| ≥27.5 | 38 (15.5) | 49 (19.7) | 57 (21.4) | 66 (22.1) | |
| NSAID/cold medication use, within past day | .612 | ||||
| No | 225 (91.8) | 236 (94.8) | 250 (93.6) | 277 (93.0) | |
| Yes | 20 (8.2) | 13 (5.2) | 17 (6.4) | 21 (7.0) | |
| History of gastritis | .837 | ||||
| No | 187 (76.3) | 191 (76.7) | 212 (79.4) | 232 (77.9) | |
| Yes | 58 (23.7) | 58 (23.3) | 55 (20.6) | 66 (22.1) | |
| H. pylori infection status | .008 | ||||
| Negative | 44 (18.0) | 34 (13.7) | 28 (10.5) | 26 (8.7) | |
| Positive | 201 (82.0) | 215 (86.3) | 239 (89.5) | 272 (91.3) | |
| Omp and HP 0305 status | .024 | ||||
| Omp– and HP 0305– | 62 (25.3) | 42 (16.9) | 40 (15.0) | 50 (16.8) | |
| Omp+ or HP 0305+ | 75 (30.6) | 81 (32.5) | 88 (33.0) | 81 (27.2) | |
| Omp+ and HP 0305+ | 108 (44.1) | 126 (50.6) | 139 (52.0) | 167 (56.0) | |
Abbreviations: BMI, body mass index; IQR, interquartile range; NSAID, nonsteroidal anti-inflammatory drug; PGE-M, prostaglandin E2 metabolite; Q, quartile.
aThe χ2 test was done within those nonmissing subjects.
Helicobacter pylori or Omp and HP0305 Status Associated With PGE-M Levels
Overall, H. pylori–seropositive individuals had an average PGE-M level that was 19.4% higher than H. pylori–seronegative individuals (P < .001) (Table 2). While this trend was apparent separately among both cases and controls, the difference reached significance only among the controls, who had overall lower PGE-M levels than the cases. Comparing PGE-M levels by combined status of seropositivity to Omp and HP 0305 instead of by H. pylori dichotomously, individuals in the least virulent group (Omp– and HP 0305–) had an average PGE-M level that was 15.4% lower than those in the most virulent group (Omp+ and HP 0305+) (P = .004). While there was an overall trend of increasing PGE-M levels with increasing H. pylori virulence based on Omp and HP 0305 status, similarly there were no significant differences in mean PGE-M levels among cases.
Table 2.
Association of Helicobacter pylori or Omp and HP0305 Status Seroprevalence With PGE-M Levels in the H. pylori Biomarker Cohort Consortium
| H. pylori Status | PGE-M Levels (ng/mg Creatinine) | ||
|---|---|---|---|
| Entire Population (N = 1059) | Cases (n = 359) |
Controls (n = 700) |
|
| H. pylori –, geometric mean level (95% CI)a | 6.2 (5.6–6.8) | 7.2 (5.5–9.4) | 6.0 (5.4–6.6) |
| H. pylori +, geometric mean level (95% CI)a | 7.4 (7.1–7.6) | 7.9 (7.4–8.4) | 7.1 (6.8–7.4) |
| Difference between H. pylori– and H. pylori+ | |||
| P valueb | <.001 | .497 | .003 |
| Percentage increasec | 19.4% | 9.7% | 18.3% |
| Omp– and HP 0305–, geometric mean level (95% CI)a | 6.5 (6.0–7.1) | 7.3 (5.8–9.2) | 6.4 (5.8–6.9) |
| Omp– or HP 0305–, geometric mean level (95% CI)a | 7.2 (6.8–7.7) | 7.7 (6.9–8.7) | 7.0 (6.4–7.5) |
| Omp+ and HP 0305+, geometric mean level (95% CI)a | 7.5 (7.1–7.9) | 8.0 (7.4–8.6) | 7.1 (6.7–7.6) |
| Difference between (Omp– and HP 0305–) and (Omp– or HP 0305–) | |||
| P valueb | .047 | .662 | .038 |
| Percentage increasec | 10.8% | 5.5% | 9.4% |
| Difference between (Omp– and HP 0305–) and (Omp+ and HP 0305+) | |||
| P valueb | .004 | .461 | .038 |
| Percentage increasec | 15.4% | 9.6% | 10.9% |
| Difference between (Omp– or HP 0305–) and (Omp+ and HP 0305+) | |||
| P valueb | .380 | .631 | .615 |
| Percentage increasec | 4.2% | 3.9% | 1.4% |
Abbreviation: CI, confidence interval; PGE-M, prostaglandin E2 metabolite.
aThe geometric mean and its 95% CI.
bAnalysis of variance (ANOVA) with cases and controls matched on sex, age, and date of biological collection, adjusted for education, body mass index, and smoking.
cFor relative difference.
Gastric Cancer Risk by Urinary PGE-M Levels
Increasing level of PGE-M was associated with increasing risk of gastric cancer in all 4 models in this study (Ptrend < .01) (Table 3). After adjustment for H. pylori status along with the other risk factors of education, BMI, and smoking status, the odds of gastric cancer for individuals in the highest quartile of PGE-M levels was 72%–76% greater than those in the lowest quartile of PGE-M levels (adjusting for H. pylori dichotomously: OR, 1.76 [95% CI, 1.17–2.66], Ptrend = .004; adjusting for H. pylori using combined Omp and HP 0305 status: OR, 1.72 [95% CI, 1.13–2.62], Ptrend = .004). If cases diagnosed within 2 years of blood draw together with their matched controls were excluded, similar increasing risks of gastric cancer by PGE-M level were found (adjusting for H. pylori dichotomously: OR, 1.73 [95% CI, 1.12–2.65], Ptrend = .007; adjusting for H. pylori by Omp and HP 0305 status: OR, 1.70 [95% CI, 1.10–2.63], Ptrend = .007). When analyses were stratified by sex, an increased trend of gastric cancer risk was observed among both men and women, although the strength of the association in men was weaker than that of women, potentially due to the small sample size of male cases (data not shown).
When analyses were stratified by follow-up period from sample collection to diagnosis, the risk of gastric cancer associated with increasing PGE-M levels decreased over time (Table 3). Adjusting for H. pylori infection, individuals in the highest PGE-M quartile diagnosed within 2–4.9 years from sample collection had an approximate 3-fold increase in the odds of gastric cancer (adjusting for H. pylori dichotomously: OR, 3.15 [95% CI, 1.11–8.91], Ptrend = .046; adjusting for H. pylori by Omp and HP 0305 status: OR, 2.95 [95% CI, .98–8.86], Ptrend = .071), while for the group with 5–9.9 years of follow-up, the increased risk associated with high PGE-M levels was only 2-fold (adjusting for H. pylori dichotomously: OR, 2.23 [95% CI, 1.22–4.06], Ptrend = .006; adjusting for H. pylori by Omp and HP 0305 status: OR, 2.20 [95% CI, 1.19–4.10], Ptrend = .007). For those with ≥10 years of follow-up, no association of PGE-M level and gastric cancer risk was found (adjusting for H. pylori dichotomously: OR, 0.73 [95% CI, .31–1.70], Ptrend = .775; adjusting for H. pylori by Omp and HP 0305 status: OR, 0.69 [95% CI, .29–1.66], Ptrend = .806).
Combined Effect by PGE-M Levels and H. pylori
Grouping individuals into categories by both H. pylori status and PGE-M level, H. pylori–positive participants with above-median PGE-M levels had a 5-fold increased odds for gastric cancer compared to H. pylori–negative participants with below-median PGE-M levels (OR, 5.08 [95% CI, 2.47–10.43]) (Table 4). Similarly, when we used Omp and HP 0305 to categorize H. pylori status instead, individuals categorized into the highest-risk H. pylori group (Omp+ and HP 0305+) and with above-median PGE-M levels had a >6-fold increased odds of gastric cancer compared to low-risk H. pylori individuals (Omp- and HP 0305-) with below-median PGE-M levels (OR, 6.54 [95% CI, 3.40–12.60]).
Table 4.
Association of Gastric Cancer Risk According to Urinary PGE-M Levels, Helicobacter pylori, and Omp and HP 0305 in the H. pylori Biomarker Cohort Consortium
| Helicobacter pylori Status | PGE-M Levels (ng/mg Creatinine) | Cases/Controls, No. | OR (95% CI) |
|---|---|---|---|
| H. pylori – | ≤50th percentile | 9/69 | 1.00 (reference) |
| >50th percentile | 9/45 | 1.61 (.59–4.44) | |
| H. pylori + | ≤50th percentile | 135/281 | 3.52 (1.70–7.29) |
| >50th percentile | 206/305 | 5.08 (2.47–10.43) | |
| Omp– and HP 0305– | ≤50th percentile | 12/92 | 1.00 (reference) |
| >50th percentile | 14/76 | 1.38 (.59–3.24) | |
| Omp+ or HP 0305+ | ≤50th percentile | 45/111 | 3.09 (1.55–6.17) |
| >50th percentile | 59/110 | 4.41 (2.20–8.81) | |
| Omp+ and HP 0305+ | ≤50th percentile | 87/147 | 4.33 (2.25–8.34) |
| >50th percentile | 142/164 | 6.54 (3.40–12.60) |
Conditional logistic regression with cases and controls matched on sex, age, and date of biological collection, adjusted for education, body mass index, and smoking.
Abbreviations: –, negative; +, positive; CI, confidence interval; OR, odds ratio; PGE-M, prostaglandin E2 metabolite.
No effect modification of the PGE-M and gastric cancer association was found by H. pylori infection, combined Omp and HP 0305 status, sex, or follow-up time (data not shown). Secondary analyses that examined the PGE-M association with gastric cancer, excluding current smokers, ever drinkers, or cases included in our previously published article, did not find differing results (data not shown) [12]. Additionally, when stratifying by site, a trend of increasing risk with increasing PGE-M level was observed for both cardia and noncardia gastric cancer, although the association was weaker and did not reach statistical significance for noncardia gastric cancer (highest vs lowest quartile of PGE-M levels: OR, 1.44 [95% CI, .94–2.22], Ptrend = .057).
DISCUSSION
Utilizing prospectively ascertained data in a region with high gastric cancer incidence, we found that after adjustment for H. pylori infection, high urinary levels of PGE-M, a stable end metabolite of PGE2, were associated with the subsequent risk of gastric cancer. Consistent with prior studies, the timing of sample collection in relation to cancer development was suggested to be a factor affecting the association, which was strongest for individuals diagnosed within 5 or 10 years of blood draw, leading to a significant over 2-fold increased risk of gastric cancer for individuals with PGE-M levels in the highest quartile [12, 16]. Combined with harboring antibodies to H. pylori, an individual with above-median levels of PGE-M was at a 5-fold increase in odds of gastric cancer diagnosis, and this rose to a >6-fold increase in odds for individuals also seropositive to the virulent H. pylori proteins Omp and HP 0305. These results suggest an additive predictive value to H. pylori status of PGE-M measurement. Previous studies have found enhanced expression of COX-2 in tumor tissues and excessive prostaglandin production in gastric carcinogenesis [23, 24]. In addition, overexpression of COX-2 in metaplastic and adenomatous cells suggests that COX-2 may contribute to the early event in the formation of gastric cancer [23, 24].
Previously a study within SWHS found a strong association of PGE-M levels and colorectal cancer risk, with relative risks of 2.5, 4.5, and 5.6 with increasing quartiles of PGE-M [16]. For breast cancer overall, no association with PGE-M in the SWHS was found, although among normal-weight postmenopausal women only, there was a linear trend of increasing PGE-M quartile and increasing risk (P = .005) [25]. A different study in Shanghai found a positive association with pancreatic cancer risk comparing the third tertile of PGE-M levels to the first (OR, 1.63 [95% CI, 1.01–2.63]) [26]. However, only 1 publication has examined the association between PGE-M and gastric cancer risk [12]. This study from our group observed a significant linear association, but was hampered by small numbers and the inability to consider the heterogeneity of H. pylori infection.
Several studies have shown that H. pylori induces COX-2 expression in human gastric mucosa, and COX-2 expression is decreased after anti–H. pylori treatment [24, 27, 28]. Hence, in H. pylori–associated gastric carcinogenesis, COX-2 overexpression might be an important step [29]. Consistent with previous findings, our study found that more virulent H. pylori strains were associated with higher PGE-M levels in all groups of the population. And the highest gastric cancer risk was observed in the more virulent H. pylori strains together with higher quartiles of PGE-M. Therefore, urinary PGE-M may be a promising noninvasive biomarker that, taken together with H. pylori infection, could aid in the risk assessment of gastric cancer.
Epidemiological studies suggest that the long-term use of NSAIDs decreases the incidence of and mortality from gastrointestinal cancers [30, 31]. The principal pharmacological effects of NSAIDs are due to their ability to inhibit COX-2 and prostaglandin synthesis [32, 33]. However, in our study population, only 4.1% participants took NSAIDs regularly. And COX-dependent mechanisms may not completely explain the chemopreventive effect of aspirin on cancer prevention [34].
Consistent with the prior literature, we observed increasing levels of urinary PGE-M were related to several lifestyle factors, including age, smoking, and obesity. Elevated levels of multiple proinflammatory mediators, known inducers of COX-2 and PGE2 synthesis, occur during aging and may contribute to this finding [35]. It has also been suggested that reduced levels of estrogen, with known anti-inflammatory properties, may contribute to aging-related increases in urinary PGE-M levels [36]. Alternatively, the age-related increase in PGE-M may simply reflect the presence of atherosclerosis or other age- related inflammatory conditions [37]. In the current study, we similarly observed a direct association between age and higher PGE-M levels. Smoking was also an independent determinant of elevated urinary PGE-M. Previously, smokers were shown to have high urinary PGE-M levels due to increased COX-2 activity, ultimately leading to increased PGE2 production [20, 38]. In our study, we also found current smokers to be overrepresented in the higher PGE-M level groups. Additionally, obesity is recognized as an inflammatory condition [39]. Increasing adiposity leads to the recruitment of macrophages that produce and release proinflammatory cytokines, thus upregulating COX-2 expression, and elevated PGE2 levels have been found in inflamed white adipose tissue [39]. Consistent with a previous study, we found that obesity was suggested to be associated with elevated levels of urinary PGE-M [25, 36].
Certain limitations of the present study should be considered when interpreting the results. While individuals who self-reported a history of cancer at baseline were excluded from these analyses, both the SMHS and SWHS participants were not screened for undiagnosed cancer, which is typical in large prospective cohort studies. Thus, it is possible that undiagnosed asymptomatic cancer could explain part of the association observed in the early years of follow-up. However the association persisted, even after excluding cases diagnosed in the first 2 years of follow-up in our study, suggesting that this finding could not be entirely explained by cancer-related PGE-M production [12, 16]. Also, premalignant gastric lesions were not accessible, and further exploration is warranted to explain the role of PGE-M in the progress of gastric cancer at differing stages.
Our study’s strengths include its prospective design, large study size, urine samples provided by most participants, long follow-up period, and adjustment for most gastric cancer risk factors, especially H. pylori infection. To our knowledge, it comprises the largest number of prospectively ascertained gastric cancer cases to examine the association of PGE-M with gastric cancer risk with adjustment of H. pylori in the high-risk region of East Asia.
In conclusion, this large prospective study in China provides the insight that, independent of H. pylori, the major causal factor for gastric cancer, increased PGE-M in urine is associated with gastric cancer in this population at high baseline risk. Also, higher PGE-M levels may indicate additional risk for gastric cancer even among individuals infected with the more virulent strains of H. pylori in East Asia. Hence, urinary PGE-M, together with H. pylori protein–specific antibodies, may be promising building blocks for a model to assess risk of gastric cancer in East Asia.
Notes
Financial support. This study was supported by the National Cancer Institute, National Institutes of Health (grant numbers R01 CA174853, UM1 CA182910, UM1 CA173640) and the China Scholarship Council.
Potential conflicts of interest. All authors: No reported conflicts of interest. The authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1. Chen X, Sood S, Yang CS, Li N, Sun Z. Five-lipoxygenase pathway of arachidonic acid metabolism in carcino-genesis and cancer chemoprevention. Curr Cancer Drug Targets 2006; 6:613–22. [DOI] [PubMed] [Google Scholar]
- 2. Chen X, Li N, Wang S, et al. Aberrant arachidonic acid metabolism in esophageal adenocarcinogenesis, and the effects of sulindac, nordihydroguaiaretic acid, and alpha-difluoromethylornithine on tumorigenesis in a rat surgical model. Carcinogenesis 2002; 23:2095–102. [DOI] [PubMed] [Google Scholar]
- 3. DuBois RN, Awad J, Morrow J, Roberts LJ, 2nd, Bishop PR. Regulation of eicosanoid production and mitogenesis in rat intestinal epithelial cells by transforming growth factor-alpha and phorbol ester. J Clin Invest 1994; 93:493–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Smith WL, DeWitt DL, Garavito RM. Cyclooxygenases: structural, cellular, and molecular biology. Annu Rev Biochem 2000; 69:145–82. [DOI] [PubMed] [Google Scholar]
- 5. Dannenberg AJ, Subbaramaiah K. Targeting cyclooxygenase-2 in human neoplasia: rationale and promise. Cancer Cell 2003; 4:431–6. [DOI] [PubMed] [Google Scholar]
- 6. Rigas B, Goldman IS, Levine L. Altered eicosanoid levels in human colon cancer. J Lab Clin Med 1993; 122:518–23. [PubMed] [Google Scholar]
- 7. Wang D, Dubois RN. Cyclooxygenase-2: a potential target in breast cancer. Semin Oncol 2004; 31:64–73. [DOI] [PubMed] [Google Scholar]
- 8. McLemore TL, Hubbard WC, Litterst CL, et al. Profiles of prostaglandin biosynthesis in normal lung and tumor tissue from lung cancer patients. Cancer Res 1988; 48:3140–7. [PubMed] [Google Scholar]
- 9. Hambek M, Baghi M, Wagenblast J, Schmitt J, Baumann H, Knecht R. Inverse correlation between serum PGE2 and T classification in head and neck cancer. Head Neck 2007; 29:244–8. [DOI] [PubMed] [Google Scholar]
- 10. Hamberg M, Samuelsson B. On the metabolism of prostaglandins E 1 and E 2 in man. J Biol Chem 1971; 246:6713–21. [PubMed] [Google Scholar]
- 11. Murphey LJ, Williams MK, Sanchez SC, et al. Quantification of the major urinary metabolite of PGE2 by a liquid chromatographic/mass spectrometric assay: determination of cyclooxygenase-specific PGE2 synthesis in healthy humans and those with lung cancer. Anal Biochem 2004; 334:266–75. [DOI] [PubMed] [Google Scholar]
- 12. Dong LM, Shu XO, Gao YT, et al. Urinary prostaglandin E2 metabolite and gastric cancer risk in the Shanghai women’s health study. Cancer Epidemiol Biomarkers Prev 2009; 18:3075–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Shu XO, Li H, Yang G, et al. Cohort profile: the Shanghai men’s health study. Int J Epidemiol 2015; 44:810–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zheng W, Chow WH, Yang G, et al. The Shanghai women’s health study: rationale, study design, and baseline characteristics. Am J Epidemiol 2005; 162:1123–31. [DOI] [PubMed] [Google Scholar]
- 15. Shu XO, Li H, Yang G, et al. Cohort profile: the Shanghai men’s health study. Int J Epidemiol 2015; 44:810–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Cai Q, Gao YT, Chow WH, et al. Prospective study of urinary prostaglandin E2 metabolite and colorectal cancer risk. J Clin Oncol 2006; 24:5010–6. [DOI] [PubMed] [Google Scholar]
- 17. Michel A, Waterboer T, Kist M, Pawlita M. Helicobacter pylori multiplex serology. Helicobacter 2009; 14:525–35. [DOI] [PubMed] [Google Scholar]
- 18. Waterboer T, Sehr P, Michael KM, et al. Multiplex human papillomavirus serology based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem 2005; 51:1845–53. [DOI] [PubMed] [Google Scholar]
- 19. Cai H, Ye F, Michel A, et al. Helicobacter pylori blood biomarker for gastric cancer risk in East Asia. Int J Epidemiol 2016; 45:774–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Moraitis D, Du B, De Lorenzo MS, et al. Levels of cyclooxygenase-2 are increased in the oral mucosa of smokers: evidence for the role of epidermal growth factor receptor and its ligands. Cancer Res 2005; 65:664–70. [PubMed] [Google Scholar]
- 21. Hansson LE, Baron J, Nyrén O, Bergström R, Wolk A, Adami HO. Tobacco, alcohol and the risk of gastric cancer. A population-based case-control study in Sweden. Int J Cancer 1994; 57:26–31. [DOI] [PubMed] [Google Scholar]
- 22. Zaridze D, Borisova E, Maximovitch D, Chkhikvadze V. Alcohol consumption, smoking and risk of gastric cancer: case-control study from Moscow, Russia. Cancer Causes Control 2000; 11:363–71. [DOI] [PubMed] [Google Scholar]
- 23. Ristimäki A, Honkanen N, Jänkälä H, Sipponen P, Härkönen M. Expression of cyclooxygenase-2 in human gastric carcinoma. Cancer Res 1997; 57:1276–80. [PubMed] [Google Scholar]
- 24. Zhang Y, Pan KF, Zhang L, et al. Helicobacter pylori, cyclooxygenase-2 and evolution of gastric lesions: results from an intervention trial in China. Carcinogenesis 2015; 36:1572–9. [DOI] [PubMed] [Google Scholar]
- 25. Cui Y, Shu XO, Gao YT, et al. Urinary prostaglandin E2 metabolite and breast cancer risk. Cancer Epidemiol Biomarkers Prev 2014; 23:2866–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Zhao J, Wang J, Du J, et al. Urinary prostaglandin E2 metabolite and pancreatic cancer risk: case-control study in urban Shanghai. PLoS One 2015; 10:e0118004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. McCarthy CJ, Crofford LJ, Greenson J, Scheiman JM. Cyclooxygenase-2 expression in gastric antral mucosa before and after eradication of Helicobacter pylori infection. Am J Gastroenterol 1999; 94:1218–23. [DOI] [PubMed] [Google Scholar]
- 28. Wong BC, Zhang L, Ma JL, et al. Effects of selective COX-2 inhibitor and Helicobacter pylori eradication on precancerous gastric lesions. Gut 2012; 61:812–8. [DOI] [PubMed] [Google Scholar]
- 29. Sung JJ, Leung WK, Go MY, et al. Cyclooxygenase-2 expression in Helicobacter pylori-associated premalignant and malignant gastric lesions. Am J Pathol 2000; 157:729–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Yang P, Zhou Y, Chen B, et al. Aspirin use and the risk of gastric cancer: a meta-analysis. Dig Dis Sci 2010; 55:1533–9. [DOI] [PubMed] [Google Scholar]
- 31. Akre K, Ekström AM, Signorello LB, Hansson LE, Nyrén O. Aspirin and risk for gastric cancer: a population-based case-control study in Sweden. Br J Cancer 2001; 84:965–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Brown JR, DuBois RN. COX-2: a molecular target for colorectal cancer prevention. J Clin Oncol 2005; 23:2840–55. [DOI] [PubMed] [Google Scholar]
- 33. Gupta RA, Dubois RN. Colorectal cancer prevention and treatment by inhibition of cyclooxygenase-2. Nat Rev Cancer 2001; 1:11–21. [DOI] [PubMed] [Google Scholar]
- 34. Fedirko V, Bradshaw PT, Figueiredo JC, et al. Urinary metabolites of prostanoids and risk of recurrent colorectal adenomas in the Aspirin/Folate Polyp Prevention Study (AFPPS). Cancer Prev Res (Phila) 2015; 8:1061–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Bruunsgaard H, Pedersen M, Pedersen BK. Aging and proinflammatory cytokines. Curr Opin Hematol 2001; 8:131–6. [DOI] [PubMed] [Google Scholar]
- 36. Morris PG, Zhou XK, Milne GL, et al. Increased levels of urinary PGE-M, a biomarker of inflammation, occur in association with obesity, aging, and lung metastases in patients with breast cancer. Cancer Prev Res (Phila) 2013; 6:428–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Wang M, Zukas AM, Hui Y, Ricciotti E, Puré E, FitzGerald GA. Deletion of microsomal prostaglandin E synthase-1 augments prostacyclin and retards atherogenesis. Proc Natl Acad Sci U S A 2006; 103:14507–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Gross ND, Boyle JO, Morrow JD, et al. Levels of prostaglandin E metabolite, the major urinary metabolite of prostaglandin E2, are increased in smokers. Clin Cancer Res 2005; 11:6087–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Das UN. Is obesity an inflammatory condition? Nutrition 2001; 17:953–66. [DOI] [PubMed] [Google Scholar]