National prevention programs and their effects on gastric cancer Incidence and mortality in East Asian countries

Temporal trends in GC incidence and mortality and related policies

Joinpoint regression was used to compute the annual percent change (APC) and 95% confidence interval (95% CI) of the GC age-standardized incidence (ASIR) and mortality (ASMR) during specific time periods⁶. The most significant decrease in the ASIR and ASMR in South Korea occurred since 2010 [APC_2010–2021: −4.6% (−5.3 to −4.0)] and between 2005 and 2021 [APC_2005–2021: −6.9% (−7.3 to −6.6)], respectively, 2 years subsequent to the screening expansion in 2003 (Figure 1A). The most remarkable decline in the ASIR and ASMR in Japan occurred between 2013 and 2019 [APC_2013–2019: −6.0% (−7.4 to −4.9)] and throughout the nationwide GC screening since 1983 and along with the increased adoption of endoscopy since 2016, respectively. There was a decline in the ASIR [APC: −5.3% (−5.9 to −4.7)] and ASMR [APC: −4.9% (−5.8 to −4.2)] for GC in China between 2010 and 2018. The timing suggests that the decline in the ASIR and ASMR for GC is likely due to the protective effect of local screening programs in part, as summarized in Table S3.

Table S4 shows the changes in ASIR and ASMR between 2004 and 2018, a period for which data were available for the three countries. The ASIR average APC (AAPC) was −2.5% (−2.8 to −2.1) in Japan (from 42.2 to 29.8/100,000) and −2.2% (−2.7 to −1.7) in South Korea (from 41.2 to 30.7/100,000), which is comparable to the ASIR AAPC in China [−2.6% (−2.9 to −2.3)]. However, the ASMR AAPC decreased most rapidly in South Korea [−6.8% (−6.9 to −6.7)], followed by Japan [−3.8% (−3.9 to −3.7)] and China [−2.7% (−3.1 to −2.3)].

Covariation of Hp infection and screening coverage with GC

Increased GC screening was recorded in South Korea between 2010 and 2021 [AAPC: 2.7% (1.6–3.7)], covering 63.8% of the eligible population in 2022 with endoscopy as the major modality (60.7% in 2022; Table S5). The overall coverage of screening in Japan increased from 28.7% in 2007 to 37.2% in 2022 [AAPC: 2.0% (−0.5 to 4.7)]. Radiography was more common but decreased over time, whereas endoscopy was initially applied less often but was on the rise since the nationwide implementation in 2016.

Figure 1B demonstrates the reduced Hp infection rates in parallel to the decreasing incidence and mortality of GC 10–20 years later in South Korea and Japan. The increasing coverage of screening was followed by a declining mortality rate within 3–10 years. Ridge regression or generalized linear mixed models demonstrated a positive association between the Hp infection rate and the incidence in South Korea [β: 1.13 (0.14–2.12)] and with mortality in both South Korea [β: 0.65 (0.04–1.25)] and Japan [β: 0.87 (0.74–1.01)]. An inverse association was also noted between endoscopy coverage and crude mortality in South Korea [β: −1.14 (−1.82 to −0.47)] and Japan [β: −0.77 (−1.02 to −0.53)], but not for radiography coverage (Table S6). The associations did not change substantially in Japan with respect to the sensitivity analysis based on the imputed data (Table S7).

GC attributable to Hp infection and nationwide screening

The population attributable fractions (PAFs) of Hp infection were estimated for endoscopic or radiographic screening based on the level of exposure in the general populations (P) and relative risks (RRs) of factors with GC incidence and death (Table 1). Considering the lag effect of the exposures and the average clinical course of GC over 5 years⁷, the number of cases in 2022 attributable to the factors in 2012 and the number in 2030 to the latest exposures with an approximate 10-year lag time were estimated.

The Hp infection PAF for the GC incidence in Japan decreased from 55% in 1987 to 27% in 2018, while GC-associated deaths decreased from 30% to 11%, leading to 45,538 cases and 6,522 deaths in 2022, and 28,169 cases and 6,667 deaths in 2030 (Figure 1C). The Hp infection PAF was slightly declined in South Korea between 1998 and 2018, resulting in 10,826 cases and 1,895 deaths in 2022, and 10,218 cases and 3,258 deaths in 2030. The number of cases and deaths attributable to Hp infection reached 94,928 and 54,831 in 2022, and is estimated to be 215,382 and 97,877 in 2030 (Table 1).

The absolute values endoscopy PAFs for GC screening increased in Japan and South Korea between 2010 and 2022, while the radiography PAFs decreased slightly in Japan (Figure 1D), resulting in 2,039 deaths prevented by endoscopy in South Korea and, 3,017 by radiography in Japan. Under current strategies, the number of deaths prevented by endoscopy would be 7,467 in South Korea and 4,804 in Japan in 2030 (Table 1). The number of GC deaths prevented by screening in China was assumed to be 0 due to a lack of nationwide screening program.

Sensitivity analyses were performed using the exposure data in 2017 (5 years before) and 2007 (15 years before) to test the robustness of the 2022 estimates. More deaths attributable to Hp infection occurred than deaths prevented by endoscopy and radiography in Japan, but fewer Hp-related deaths occurred than deaths prevented by screening in South Korea (Table S8).

Reduced GC burden due to Hp eradication and screening

The number of additional GC cases and deaths prevented by multiple enhanced strategies, including several ideal scenarios, was predicted. The projected number of GC deaths prevented through full coverage of endoscopy alone in 2030 reached 7,893 in South Korea, 28,559 in Japan, and 353,208 in China, which far exceeded the reduced numbers (4,104, 7,820, and 97,957 in corresponding countries) by cutting down the Hp infection rate to zero alone, both of which are impractical and unattainable in real-world conditions (Figure 1E). Full coverage of endoscopy combined with a zero Hp infection rate is expected to decrease GC deaths to the greatest extent, with 13,513 deaths prevented in South Korea, 40,034 in Japan, and 541,586 in China.

By linking the changes in GC incidence and mortality with the national preventive strategies in Asian countries, that the most rapid decline in mortality was shown to occur within several years of the issuance of related preventive policies in South Korea and Japan. In contrast, GC mortality decreased more slowly in China, a country without nationwide screening programs. The results are suggestive of an early diagnosis of GC due to the interventions and the improvements in GC treatment on secular trends of incidence and mortality.

A major finding in the current study was the superior efficacy of endoscopy in reducing GC deaths compared to radiography. The endoscopy-dominant strategy in South Korea was shown to outperform the radiography-focused approach in Japan, which was also observed in a synthetic control study⁸. The South Korea strategy holds essential reference value for China, a country with a heavy GC burden but without a nationwide cancer screening program. GC deaths attributable to Hp infection were also shown to exceed the deaths prevented by endoscopy in most circumstances. A significant number of GC deaths could be prevented in China by endoscopic screening alone and the number would be significantly increased by jointly implementing Hp eradication, according to the 2030 predictions. Hp eradication reduces GC mortality by decreasing the incidence, which is highly significant in public health as a primary prevention approach. These conventional effective approaches may not be cost-effective in Western countries with a low incidence of GC. The potential harms of endoscopy or radiography also limit extensive applications of the examinations. Novel non-invasive techniques, such as liquid biopsy, may be highly valuable in countries with a low incidence⁹.

This is the first epidemiologic study to quantify and compare the potential impact of nationwide Hp eradication on radiographic and endoscopic screeningon the GC burden. Despite the potential ecologic fallacy, unmeasured confounders, such as changes in healthcare access, dietary patterns, diagnostic improvements, and overestimated PAF by assuming an independent effect of Hp infection and non-screening of endoscopy, the results herein indicated the benefits of nationwide endoscopic screening in reducing GC mortality and the long-term effect of Hp eradication. The findings provide evidence for policy making in GC prevention and control in East Asian countries heavily burdened by GC.

Author contributions

Conceived and designed the analysis: Wanghong Xu, Ruoxin Zhang.

Collected the data: Min Cai, Ruiqi Xia, Ziyang Wang.

Contributed data or analysis tools: Min Cai, Ruiqi Xia, Ziyang Wang.

Wrote the paper: Min Cai.