Efficacy and Safety of Vonoprazan and Amoxicillin Dual Therapy for Helicobacter pylori Eradication: A Systematic Review and Meta-Analysis

Helicobacter pylori infection is a global concern that affects half of the population worldwide [1]. H. pylori can cause persistent inflammation in the stomach and is associated with many gastrointestinal diseases, including peptic ulcers, mucosa-associated lymphoid tissue lymphoma, and gastric cancer [2]. H. pylori eradication can mitigate the gastritis and reduce the risk of many complications. Furthermore, it is a cost-effective approach to decrease the incidence of gastric cancer [3]. All H. pylori-positive individuals are recommended to receive eradication therapy unless they have competing considerations [4].

However, the cure rate of standard triple therapy is declining over time due to the increasing antibiotic resistance, especially clarithromycin. Treatment success in the regions with a high prevalence of clarithromycin resistance fails to reach even 80% [5]. Currently, standard triple therapy is still the recommended regimen in regions with clarithromycin resistance <15%. For the regions with high clarithromycin resistance, published guidelines recommend bismuth-containing quadruple therapy which contains a proton pump inhibitor (PPI), bismuth, and two antibiotics or concomitant non-bismuth quadruple therapy (PPI + three antibiotics) to achieve satisfactory cure rates [6, 7]. However, the four-drug combination results in higher cost, more adverse events, and lower compliance. Moreover, unsuccessful treatments significantly increase the antibiotic resistance and even multiple drug resistance, leading to an even lower cure rates in rescue treatments [8, 9].

Notably, the prevalence of amoxicillin resistance is low in most regions worldwide and is stable over time even after treatment failure [10]. Dual therapy with PPIs and amoxicillin was first reported in the 1990s for H. pylori eradication, while the cure rate was initially not acceptable [11, 12]. After many years of exploration, the latest evidence revealed that high-dose PPIs and amoxicillin dual therapy (amoxicillin administration ≥ three times a day and PPI administration ≥ twice a day) had similar efficacy as the guideline-recommended therapies and tended to have lower side effects and better compliance [13, 14]. These findings highlight the promising development of dual therapy.

Currently, a novel acid-suppressing agent, vonoprazan, has a stronger effect on the gastric acid suppression than PPIs based on the potassium-competitive acid blocking mechanism and has shown potent efficacy against acid-associated diseases [15, 16]. Furthermore, the efficacy of vonoprazan-based triple therapy has been proven to be superior to that of PPI-based triple therapy for H. pylori eradication [17]. Subsequently, several studies have compared the efficacy of vonoprazan/amoxicillin (VA) dual therapy and standard triple therapies (PPI- or vonoprazan-based triple therapy). Of note, it seems that only 12% of the entire population obtain actual benefit from the addition of clarithromycin in vonoprazan and amoxicillin therapy [18]. In this regard, VA dual therapy is expected to be an appropriate choice for H. pylori eradication. Hence, we performed a systematic review and meta-analysis to investigate the efficacy and safety of VA dual therapy for H. pylori eradication.

Our study was registered on the International Prospective Register of Systematic Reviews (PROSPERO) (ID: CRD42022329712). The whole process of the present study was conducted according to the PRISMA guidelines [19].

We conducted a systematic literature search for relevant studies up to June 2022, using databases including PubMed, Web of Science, EMBASE, and the Cochrane Library. Abstracts from major gastroenterology conferences in the past 7 years were also searched. The following terms were used: (“Helicobacter pylori” OR “HP” OR “H. pylori” OR “Campylobacter pylori”) AND (“vonoprazan” OR “TAK-438” OR “takecab” OR “potassium-competitive acid blocker” OR “P-CAB”) AND (“dual”). The detailed strategies used to search each database are shown in online supplementary material search strategies (for all online suppl. material, see www.karger.com/doi/10.1159/000529622). We also examined the reference list of each study to identify additional eligible studies.

Two researchers (W.L.Z. and B.S.L.) independently checked all articles by reading the titles and abstracts and then excluded those irrelevant to our topic. Second, the full text of each article was assessed for its eligibility. Any disagreements were resolved by mutual agreement with the third reviewer (Y.Y.L.).

The inclusion criteria were listed as follows: (1) patient studies including H. pylori-positive patients, (2) intervention studies using vonoprazan-based dual regimen for H. pylori eradication therapy, (3) comparator studies using PPI- or vonoprazan-based triple therapy as the control group, and (4) outcome-cure rate and adverse events. Studies meeting the following criteria were eventually excluded: (1) animal experiments, (2) studies without a control group, (3) studies published in languages other than English, (4) sub-analysis containing partial data of the included studies, and (5) conference abstracts or studies without full text.

The following data were extracted by two independent investigators (Y.M.D. and Z.X.H.) using an Excel form prepared in advance: name of the first author, the year of publication, country, patients, study design, therapeutic regimens, duration, confirmative testing method after treatment, cure rate, adverse events, and dropout rate (if available). Any disagreement was resolved by consensus in consultation with the third reviewer (Y.Y.L.), and the final decision was made.

The primary outcomes of our research were the cure rates and relative efficacy of VA dual therapy with triple therapies (PPI- or vonoprazan-based triple therapy) in intention-to-treat (ITT) and per-protocol (PP) analyses. Secondary outcomes included the comparison of (1) adverse events, (2) dropout rate, (3) cure rates and relative efficacy based on different acid-suppressing agents in the control group (vonoprazan or PPIs), (4) cure rates and relative efficacy based on clarithromycin susceptibility, (5) pooled cure rates of VA dual therapy based on duration and dosage of amoxicillin, and (6) adverse events according to specific symptoms.

Randomized controlled trials (RCTs) were evaluated by Cochrane Risk of Bias assessment tool, including the methods of generating a randomization schedule and concealing allocations, blindness, management of incomplete outcome data, selective outcome reporting, and other biases [20]. The cohort studies were evaluated by the Newcastle-Ottawa scale (NOS), which consists of three categories: selection, comparability, and outcome [21]. Two authors (W.L.Z. and B.S.L.) evaluated the risk of bias independently. Any disagreements were resolved by discussion.

The pooled cure rates were calculated to assess the efficacy of each regimen. The pooled risk ratio (RR) and 95% confidence interval (CI) were calculated to assess the correlation between the outcomes in the experimental and control groups. To provide a more conservative estimate, the random-effects model was used. Heterogeneity was assessed by Cochrane’s Q test and I² statistics. For Cochrane’s Q test, p > 0.1 indicates no apparent heterogeneity, and for I² statistics, I² <50% indicates an acceptable heterogeneity. Sensitivity analyses were performed to evaluate the stability of the results. We assessed publication bias using Egger’s linear regression and Begg’s rank correlation tests; also, it was assessed qualitatively by inspecting the funnel plots of the logarithmic RR versus its standard errors. Two-tailed p value <0.05 was considered having statistically significant difference. The statistics analysis was conducted by the statistical software Review Manager 5.4 (version 5.4; Cochrane Collaboration, Copenhagen, Denmark) and Stata software (Stata 17, Stata Corporation, College Station, TX, USA).

A flow diagram of the study selection process is shown in Figure 1. In summary, we obtained 161 studies through database searching and one study through manual searching. After excluding the duplicate and irrelevant articles by screening the title and abstracts, seven articles were retained for further full-text review. Two articles that reported the partial data from a multicentral randomized trial were excluded. Ultimately, five articles were included in this analysis.

The characteristics of the included studies are presented in Table 1. These studies were conducted between 2015 and 2021. Of the five studies, three were RCTs [22‒24], one was a prospective cohort study [25], and one was a retrospective cohort study [26]. Three studies explicitly pointed out that the participants were treatment-naive [22, 23, 25], and the remaining two studies did not specify the prior eradication of the participants [24, 26]. Three studies were from Japan and reported 7-day VA dual therapy containing low-dose amoxicillin (1,500 mg/d; 500 mg t.i.d. or 750 mg b.i.d.) [22, 25, 26], while the remaining two studies were from the USA and Pakistan and reported 14-day VA dual therapy containing high-dose amoxicillin (1,000 mg b.i.d. or t.i.d.) [23, 24]. Two studies conducted antibiotic susceptibility tests and analyzed the cure rates according to the amoxicillin or clarithromycin susceptibility [22, 23].

The risk of bias in the RCTs is shown in online supplementary Figure S1. The participants were randomly allocated in all the studies. Zuberi et al. [24] did not describe the blinding method for participants or investigators. Participants and investigators were not blind to the allocation in the study of Suzuki et al. [22]. The VA dual therapy of the study of Chey et al. [23] was open label. All three studies reported the outcomes mentioned in the method section.

The risk of bias in the prospective and retrospective cohort studies is shown in online supplementary Table S1. Gotoda et al. [25] conducted a study among junior high school students where the participants exposed to triple therapy was from 2015 to 2017 and those exposed to VA dual therapy were from 2018. Hence, selection bias may exist.

All five studies reported the cure rates in the ITT analysis, and a total of 1,852 patients were included. The pooled cure rate of VA dual therapy was 85.6% (95% CI: 79.7–91.5%), and moreover, 83.5% (76.3–90.8%) and 89.5% (81.9–97.1%) of RCTs and observational studies, respectively (Fig. 2a). The efficacy of VA dual therapy was not significantly different from that of triple therapy, with a pooled RR of 1.03 (95% CI: 0.97–1.10, p = 0.37). Furthermore, the relative efficacy of VA dual therapy and triple therapy based on the different study types (RCTs and observational studies) appeared to be similar; the pooled RRs were 1.04 (95% CI: 0.93–1.15) and 1.03 (95% CI: 0.94–1.12), respectively. We identified a moderate overall heterogeneity (Cochrane’s Q test, df = 4, p = 0.10, I² = 49%), and the test for subgroup differences showed no heterogeneity (Cochrane’s Q test, df = 1, p = 0.88, I² = 0%, Fig. 3a).

In case of PP analysis, five studies with 1,653 patients were included. The pooled cure rate of VA dual therapy was 88.5% (95% CI: 83.2–93.8%, 87.2% [80.1–94.3%] and 91.1% [83.3–98.8%] of RCTs and observational studies, respectively, Figure 2b). The efficacy of VA dual therapy was not significantly different from that of triple therapy, with a pooled RR of 1.02 (95% CI: 0.98–1.08, p = 0.32). No apparent heterogeneity was identified (Cochrane’s Q test, df = 4, p = 0.24, I² = 27%), and the test for subgroup differences indicated no heterogeneity (Cochrane’s Q test, df = 1, p = 0.90, I² = 0%, Fig. 3b).

Two studies conducted antibiotic susceptibility tests and provided data on the cure rate in the PP analysis according to the antibiotic susceptibility. We calculated the cure rates and compared the efficacy between VA dual and vonoprazan-based triple therapies.

For clarithromycin-resistant strains, the pooled cure rate for VA dual therapy was 86.7% (95% CI: 74.2–99.1%) and for vonoprazan-based triple therapy was 71.4% (95% CI: 62.6–80.2%). The efficacy of VA dual therapy was significantly superior to that of vonoprazan-based triple therapy (RR = 1.20, 95% CI: 1.03–1.39, p = 0.02). No heterogeneity was identified (Cochrane’s Q test, df = 1, p = 0.87, I² = 0%, Fig. 4).

For clarithromycin-sensitive strains, the pooled cure rates for VA dual therapy and vonoprazan-based triple therapy were 83.0% (95% CI: 78.8–87.1%) and 92.8% (95% CI: 88.2–97.3%), respectively. The efficacy of VA dual therapy was significantly inferior to that of vonoprazan-based triple therapy, with a pooled RR of 0.90 (95% CI: 0.85–0.95, p = 0.0002). No heterogeneity was identified (Cochrane’s Q test, df = 1, p = 0.98, I² = 0%, Fig. 4).

In total, four studies compared the efficacy of VA dual therapy with vonoprazan-based triple therapy, and two studies compared it with PPI-based (omeprazole and lansoprazole) triple therapy (Chey et al. [23] compared VA dual therapy with vonoprazan-based or lansoprazole-based triple therapy, so we divided the number of participants in the experimental group into two parts to compare with the control groups separately [27]). The pooled cure rates for vonoprazan-based triple therapy were 85.4% (95% CI: 80.4–90.5%) in ITT analysis and 87.9% (95% CI: 84.3–91.4%) in PP analysis, and for PPI-based (omeprazole and lansoprazole) triple therapy, they were 71.4% (95% CI: 64.2–78.7%) and 76.7% (95% CI: 63.1–90.3%), respectively. The efficacy of VA dual therapy was not inferior to that of vonoprazan-based triple therapy in ITT and PP analysis (RR = 0.98, 95% CI: 0.93–1.03 and RR = 0.98, 95% CI: 0.93–1.03, respectively). Yet, VA dual therapy was significantly superior to the PPI-based (omeprazole and lansoprazole) triple therapy (RR = 1.15, 95% CI: 1.05–1.25, p = 0.001 and RR = 1.13, 95% CI: 1.05–1.23, p = 0.001, respectively). We identified moderate heterogeneity (Cochrane’s Q test, df = 5, p = 0.03, I² = 60%) in both ITT and PP analyses. In addition, the test for subgroup differences revealed apparent heterogeneity (Cochrane’s Q test, df = 1, p = 0.001, I² = 90.1%, online suppl. Fig. S2).

Three studies were from Japan and reported 7-day VA dual therapy containing low-dose amoxicillin (1,500 mg/d; 500 mg t.i.d. or 750 mg b.i.d.). The pooled cure rate of 7-day low-dose VA dual therapy was 87.5% (95% CI: 82.0–93.0%) and 89.6% (95% CI: 84.4–94.7%) in the ITT and PP analyses, respectively.

The remaining two studies were from the USA and Pakistan and reported 14-day VA dual therapy containing high-dose amoxicillin (1,000 mg b.i.d. or t.i.d.). The pooled cure rate of 14-day high-dose VA dual therapy was 83.2% (95% CI: 71.0–95.4%) and 87.3% (95% CI: 75.2–99.4%) in the ITT and PP analyses, respectively.

In terms of adverse events, four studies reported the overall adverse events among the participants, whereas Zuberi et al. [24] only reported common adverse events. Common adverse events included diarrhea, constipation, abdominal pains, nausea, dysgeusia, and skin rashes. Only Chey et al. [23] reported serious adverse events in 1.3% of the patients, and the remaining studies reported no serious adverse events, and most of the symptoms spontaneously ameliorated. The skin rash was mainly due to an allergy to amoxicillin and could be cured by antiallergic agents.

Hence, four studies with 1,707 patients were included to compare the overall adverse events between the two groups. The pooled rates of adverse events for VA dual therapy and triple therapies were 21.2% (95% CI: 11.8–30.7%) and 26.5% (95% CI: 17.8–35.2%), respectively. The adverse effects of VA dual therapy were lower than those of triple therapy in both RCTs (RR = 0.88, 95% CI: 0.74–1.04, p = 0.12) and observational studies (RR = 0.65, 95% CI: 0.37–1.15, p = 0.14); however, the difference was not statistically significant, with a pooled RR of 0.86 (95% CI: 0.73–1.01, p = 0.06). No heterogeneity was identified and no subgroup difference (Cochrane’s Q test, df = 1, p = 0.32, I² = 0%, Fig. 5a).

Additionally, five studies were included to compare specific adverse events. No statistical difference was observed between VA dual therapy and triple therapies regarding skin rash, abdominal pain, nausea, constipation, and dysgeusia. However, VA dual therapy had a significantly lower incidence of diarrhea than triple therapy, with a pooled RR of 0.64 (95% CI: 0.46–0.91, p = 0.01, online suppl. Fig. S3).

Four studies reported the dropout rates due to adverse events. No significant statistical difference was observed between VA dual therapy and triple therapies in both RCTs (RR = 0.67, 95% CI: 0.22–2.08, p = 0.49) and observational studies (RR = 1.35, 95% CI: 0.11–16.09, p = 0.82), with a pooled RR of 0.78 (95% CI: 0.29–2.09, p = 0.62). Specifically, the main reason for discontinuing treatment therapy in the VA dual therapy group was skin rashes, which were proved to be due to amoxicillin allergy. In the triple therapy group, the reasons included skin rashes, abdominal pain, diarrhea, nausea, and others. Furthermore, no apparent heterogeneity was identified, and there was no subgroup difference (Cochrane’s Q test, df = 1, p = 0.62, I² = 0%, Fig. 5b).

We performed a sensitivity analysis by removing one study at a time and analyzing the remaining studies. No statistically significant change was observed in the pooled RR for each outcome. Hence, our results were robust and not markedly affected by a single study. Furthermore, the funnel plot is shown in online supplementary Figure S4. In addition, no significant publication bias was identified based on Begg’s test (p = 0.462) and Egger’s test (p = 0.325).

We systematically summarized available studies and analyzed the efficacy and safety of VA dual therapy for H. pylori eradication. We compared VA dual therapy with PPI- or vonoprazan-based triple therapies on cure rates, adverse effects, dropout rates and performed subgroup analysis based on clarithromycin susceptibility, acid-suppressing agent in the control group, and specific adverse events. Though the pooled cure rate of VA dual therapy did not excess 90%, we found it presented a lower dropout rate and fewer side effects; and in particular, its superiority in clarithromycin-resistance strains.

Dual therapy is not a novel regimen for H. pylori eradication as it has been proposed for decades. Initially, the cure rate was unacceptable with low-dose amoxicillin (≤2,000 mg/d) or less frequency of PPIs (≤twice daily), which were lower than 60% [11, 12]. Recently, studies exploring modified PPIs and amoxicillin dual therapy highlighted the factors that increase treatment success: [28]: (1) sufficient acid suppression to maintain a neutral intragastric environment; (2) maintaining a higher concentration of amoxicillin than the minimum inhibitory concentration; and (3) adequate treatment duration (up to 14 days). For acid suppression, the effect of traditional PPIs is not stable with the nocturnal acid breakthrough [29]. In addition, PPIs are primarily metabolized by cytochrome P450 (CYP) 2C19 (CYP2C19). Vonoprazan inhibits gastric acid secretion irrespective of the CYP2C19 genotype [16], and its intragastric pH 4 and 5 holding time ratios can reach 100% and 99%, respectively [30]. In our analysis, the efficacy of VA dual therapy was not significantly different from that of the vonoprazan-based triple therapy but superior to that of PPI-based triple therapy. Amoxicillin is a time-dependent antibiotic; hence, increasing the dose or frequency may enhance its efficacy [31]. We found that the cure rates of the 7-day low-dose VA dual therapy were higher than 14-day high-dose VA therapy in both ITT and PP analyses. In addition, a study exploring the optimization of VA dual therapy demonstrated that both 7-day and 10-day (either b.i.d. or t.i.d.) provided satisfying cure rates in China [32]. However, we should note that, in that study, the cure rate of each group in that study was lower than 90%, and therefore, more studies should be conducted to optimize VA dual therapy based on the duration and dosage.

To date, dual therapy has not been widely accepted as the first-line regimen. Only the American College of Gastroenterology Clinical Guideline and the guideline in Japan recommend high-dose PPIs and amoxicillin dual therapy as second-line therapy [33, 34]. In our meta-analysis, more specifically, the pooled cure rates of VA dual therapy were 85.6% and 88.5% in ITT and PP analyses, respectively, which were similar to the pooled cure rates of high-dose PPIs (t.i.d. or q.i.d.) and amoxicillin (750 mg q.i.d. or 1,000 mg t.i.d.) dual therapy [14]. However, the rate still could not meet the criterion as “good therapy” (cure rate >90%) [35]. In this regard, we still should assess the role of VA dual therapy more cautiously until large-scale studies provide further convincing evidence.

Notably, the Toronto consensus indicates that high-dose dual therapy may be an option when multiple antibiotic resistances is suspected and for patients with multiple previous treatment failures [6]. Antibiotic resistance is an important factor influencing eradication success as well as guiding tailored therapy. In fact, clarithromycin resistance is the primary factor accounting for the decrease in the cure rate of standard triple therapy [36]. Our subgroup analysis according to antibiotic susceptibility showed interesting findings. For clarithromycin-resistant strains, the efficacy of VA dual therapy was superior to the vonoprazan-based triple therapy. Two reasons may account for the result. First, H. pylori is sensitive to amoxicillin at a pH of 6–7 [36]. Vonoprazan and clarithromycin are both metabolized by the CYP3A4, and the existence of clarithromycin can increase the maximum plasma concentration of vonoprazan by 1.8-fold [37], which may increase the intragastric pH higher than 7 and thus reduce the potency of amoxicillin. Second, VA dual therapy included fewer drugs and may have better compliance. For clarithromycin-sensitive strains, VA dual therapy was significantly inferior to vonoprazan-based triple therapy. Therefore, VA dual therapy is more suitable for patients with clarithromycin-resistant strains, while not for clarithromycin-sensitive strains. However, because of the data limitation, we were not able to compare the efficacy according to amoxicillin susceptibility. Amoxicillin resistance is low in most regions of the world [5], which may explain why the studies did not specifically emphasize amoxicillin susceptibility.

The safety of the eradication regimens is another important issue. The Chinese fifth guideline specifically points out that the elderly patients should be carefully assessed for benefits/risks before prescribing therapy. The Japanese guideline recommends that patients with decreased liver or renal functions should be treated with a reduced dose of antibacterial agents and with caution regarding side effects [34]. Gao et al. [38] demonstrated a satisfying efficacy of rabeprazole-amoxicillin dual therapy for elderly patients or those with multiple comorbidities, with a cure rate of 90.9%. In fact, the safety of VA dual therapy can be divided into two aspects. On the one hand, vonoprazan is mainly metabolized by CYP3A4; therefore, it does not influence the activity of drugs metabolized by CYP2C19, such as clopidogrel, a drug widely used by patients with coronary atherosclerosis [16]. Furthermore, VA dual therapy contains only one antibiotic, thus having less impact on the liver or renal functions. Additionally, subgroup analysis showed that the incidence of diarrhea in the VA dual therapy group was significantly lower than that in the control group. In this regard, VA dual therapy also had less impact on the gut microbiota [39].

We noted that in VA dual therapy, except for some patients who had skin rashes, most patients who experienced adverse events recovered spontaneously. In particular, skin rash, which was due to amoxicillin allergy, was the main reason for the discontinuation of VA dual therapy. Therefore, a penicillin skin test may be necessary for patients unsure of a penicillin allergy or those who have never used amoxicillin.

The present meta-analysis had several limitations. First, we included both RCTs and observational studies which may induce potential bias. However, we calculated the cure rates, relative efficacy, adverse events, and dropout rates based on different study designs, and no subgroup difference was found. Second, we could not further analyze the efficacy of VA dual therapy in first-line and second-line treatments due to data limitations. The low eradication success for rescue treatment has long been a big challenge due to antibiotic resistance [8], and studies focused on the efficacy of VA dual therapy as second-line treatment are warranted. Third, as we mentioned above, we failed to compare the relative efficacy according to the duration and dosage separately because of the limited data of the included studies; nevertheless, we calculated the pooled cure rates of VA dual therapy (7-day low-dose and 14-day high-dose VA therapies). Previous studies demonstrated that increasing the frequency of PPIs/amoxicillin dual therapy to four times daily could achieve a higher cure rate [40], whereas it is yet to be confirmed in VA dual therapy. Fourth, the PPIs in the included studies were first-generation PPIs (omeprazole and lansoprazole), resulting in low cure rates in the control group, which may overstate the relative efficacy of the VA dual therapy. Fifth, all the included studies compared the efficacy of VA dual therapy with that of triple therapy. According to published guidelines, bismuth-containing quadruple therapy is the first-line empirical therapy in regions with high clarithromycin resistance [7]. The efficacy of VA dual therapy compared with bismuth-containing quadruple therapy remains to be studied, especially in regions with high clarithromycin resistance.

Despite the limitations mentioned above, we elucidated that VA dual therapy showed noninferior efficacy to vonoprazan-based triple therapy and was superior to the omeprazole or lansoprazole-based triple therapy for H. pylori eradication. In particular, VA dual therapy is a good choice for clarithromycin-resistant strains but not appropriate for clarithromycin-sensitive strains. In addition, though its cure rate cannot meet the criterion for first-line therapy, in view of the lower side effects of VA dual therapy, it may be an appropriate choice, especially for elderly patients and those with multiple comorbidities. Notably, despite the superiority of the VA dual therapy, the pooled cure rate of the included studies failed to achieve 90%; therefore, more high-quality studies are warranted to further verify the clinical relevance of VA dual therapy.

The statistical methods of this study were reviewed by Pei-Zhu Wang from Qilu Hospital of Shandong University. We would like to thank Editage (www.editage.cn) for English language editing.

An ethics statement is not applicable because this study is based exclusively on the published literature. Consent to participate statement: Written informed consent was not required because the present article is a meta-analysis of the published studies. The informed consents of the included studies had been obtained when they were conducted.

The authors have no conflicts of interest to declare. All the authors participated in review of the article and approved the final version of the manuscript.

This study was supported by the Natural Science Foundation of Shandong Province (ZR2020LZL003), the Key R&D Program of Shandong Province (2021ZLGX02), and the Clinical Research Center of Shandong University (2020SDUCRCA006). PRISMA 2009 Checklist: The guidelines of the PRISMA 2009 Statement have been adopted.

Wen-Lin Zhang and Bo-Shen Lin performed literature search, selected the relevant studies, and drafted the manuscript. Wen-Lin Zhang, Yu-Ming Ding, and Zhong-Xue Han performed data extraction and the statistical analyses. Wen-Lin Zhang and Bo-Shen Lin evaluated the risk of bias of the included studies. Yue-Yue Li supervised the process, proposed critical comments, and made final decision. Rui Ji reviewed and edited the manuscript.

All data generated or analyzed during this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.