Association of Common Medications and the Risk of Early-Onset Gastric Cancer: A Population-Based Matched Study

Table 2

Clinical characteristics of EOGC patient and control patient groups.

	EOGC patients ()	Control patients ()	Univariate odds ratio estimate; (95% confidence interval)	Multivariate odds ratio estimate; (95% confidence interval)
PPI use	27 (28.1%)	21 (21.8%)	1.38 (0.72-2.62)	0.58 (0.15-2.18)
H2RA use	8 (8.3%)	2 (2.1%)	4.0 (0.85-18.84)	2.24 (0.427-11.8)
BMI (kg/m^2)	29.2 [25.0, 34.2]	29.5 [26.9, 34.3]	—	—
Smoking status at time of diagnosis			—	—
Never	44 (45.8%)	43 (44.7%)
Former	24 (25.0%)	34 (35.4%)
Current	28 (29.1%)	19 (19.8%)
History of H. pylori infection	7 (7.2%)	2 (2.1%)	3.50 (0.73-16.85)	8.87 (1.03-76.67)
History of GERD	42 (43.8%)	27 (28.1%)	1.94 (1.06-3.54)	2.53 (0.78-8.25)
Prior stomach surgery	6 (6.2%)	0	13.85 (0.612-313.35)	9.35 (0.41-213.22)
Metformin use	4 (4.2%)	9 (9.4%)	0.44 (0.14-1.44)	0.54 (0.15-1.99)
NSAID use	23 (24.0%)	23 (24.0%)	1.00 (0.52-1.92)	—
COX-2 inhibitor use	12 (12.5%)	22 (22.9%)	0.44 (0.19-1.02)	0.39 (0.16-0.94)
Statin use	15 (15.6%)	14 (14.6%)	1.1 (0.47-2.59)	—

Not included in the final parsimonious multivariable model due to value of >0.8 on univariate analysis of maximum likelihood estimates. Odds ratio estimates with 95% confidence intervals that do not cross 1.0 considered significant (bolded). BMI = body mass index, described as median with [interquartile range]. Discrete variables expressed as counts (percent). PPI = proton-pump inhibitor; H2RA = H2-receptor antagonist; GERD = gastroesophageal reflux disease; NSAID = nonsteroidal anti-inflammatory agent; COX-2 inhibitor = cyclooxygenase-2 inhibitor; statin = HMG-CoA reductase inhibitor. EOGC = early-onset gastric cancer, defined as index cancer diagnosis at years old. Control patients were matched 1 : 1 with EOGC patients based on age (within three years), sex, smoking status (current smoker, former smoker, and never smoker), and BMI (obese or not obese, with obesity defined as ). Patient was considered to have a history of use for each medication if they had been prescribed the medication for six months or more prior to index diagnosis of gastric cancer.