Characteristics of patients
The characteristics of the 7736 patients were as follows: median age: 61.3 years (range, 15–105 years), gender: 3929 men (50.8 %), and median body mass index (BMI): 23.0 kg/m2 (range, 10.7–37.3). Diabetes mellitus was found in 860 patients (11.1 %). PPIs were taken by 1253 patients (16.2 %). EC was observed in 184 of 7736 patients (2.4 % incidence); this rate was similar to a previous report of regional general hospitals 10. Moreover, 3161 patients had atrophic gastritis (40.9 %) (Table 1).
Predictors for the development of EC
Univariate analysis showed that age over 65 years {odds ratio (OR): 1.95; P < 0.001), diabetes mellitus (OR: 1.85; P = 0.003), PPI use (OR: 2.2; P < 0.001), oral steroid use (OR: 2.96; P = 0.021), inhaled steroid use (OR: 3.62; P = 0.031), atrophic gastritis (OR: 1.63; P = 0.001), advanced gastric cancer (OR: 5.04; P < 0.001), and gastrectomy (OR: 2.48; P = 0.010) were significant risk factors for EC. BMI over 25 kg/m2 (OR: 0.51; P = 0.002) was independently associated with a decreased risk of EC. In contrast, GERD (OR: 0.83; P = 0.737) and esophageal hiatal hernia (OR: 0.55; p = 0.081) were not significant risk factors in this study (Table 2).
Multivariate analysis also revealed that diabetes mellitus {OR: 1.52; 95 % confidence interval (CI), 1.01–2.29, P = 0.043}, PPI use (OR: 1.69; 95 % CI, 1.19–2.41, P = 0.003), atrophic gastritis (OR: 1.60; 95 % CI, 1.17–2.18, P = 0.003), advanced gastric cancer (OR: 4.66; 95 % CI, 2.40–9.05, P < 0.001), and gastrectomy (OR: 2.32; 95 % CI, 1.18–4.57, P = 0.015) were independent risk factors for developing EC. In contrast, obesity (BMI ≥ 25) reduced EC risk (OR: 0.53; 95 % CI, 0.34–0.84, P = 0.006). There were no significant differences between the two groups with regard to steroid use (inhaled: P = 0.069, oral: P = 0.201, respectively) (Fig. 1).
Predictors of EC severity
EC severity was evaluated using the Kodsi classification 12. The severity of EC in the 184 patients examined in our cohort were: 146 patients with Kodsi grade I (79.3 %), 20 with grade II (10.9 %), 16 with grade III (8.7 %), and 2 with grade IV (1.1 %). EC patients were stratified as mild (grade I) or moderate/severe (grade II and higher). Patients with moderate/severe EC were more likely to have advanced gastric cancer (P = 0.003) and gastrectomy (P = 0.006) than patients with mild EC (Table 3). In multivariate analysis, advanced gastric cancer (OR: 17.6; 95 % CI, 3.54–87.9, P < 0.001) and gastrectomy (OR: 23.4; 95 % CI, 4.24–129, P < 0.001) were also identified as significant risk factors for developing sever EC. Atrophic gastritis (P = 0.078) and PPI use (P = 0.072) were not associated with EC severity in the present study (Fig. 2).
Impact of PPI use on the risk of EC in patients with atrophic gastritis
As both PPIs and atrophic gastritis cause reduced acidity, their combination may have a much greater impact on EC development than either alone. Therefore, we next examined whether PPI use may increase the risk for developing EC in patients with atrophic gastritis. The enrolled patients were stratified into the following 4 groups: patients without atrophic gastritis who did not receive PPI (control group, n = 3967), patients with atrophic gastritis who did not receive PPI (gastritis group, n = 2516), patients without atrophic gastritis who received PPI (PPI group, n = 608), and patients with atrophic gastritis who did receive PPI (gastritis plus PPI group, n = 645). The incidence of EC in each group was 1.6 % (control group), 2.6 % (gastritis group), 3.6 % (PPI group), and 5.0 % (gastritis plus PPI group). The OR with respect to the control was 1.59 (95 % CI, 1.12–2.25), 2.25 (95 % CI, 1.38–2.25), and 3.13 (95 % CI, 2.03–4.83) for the gastritis group, PPI group, and gastritis plus PPI group, respectively (Table 4).