Early effective intervention can significantly reduce all-cause mortality in prediabetic patients: a systematic review and meta-analysis based on high-quality clinical studies

Background Reducing the occurrence of diabetes is considered a primary criterion for evaluating the effectiveness of interventions for prediabetes. There is existing evidence that early lifestyle-based interventions can significantly decrease the incidence of diabetes. However, whether effective interventions can reduce long-term outcomes in patients, including all-cause mortality, cardiovascular risks, and the occurrence of microvascular complications, which are the most concerning issues for both patients and clinicians, remains a subject of inconsistent research findings. And there is no direct evidence to answer whether effective intervention has long-term benefits for prediabetic patients. Therefore, we conducted a systematic review and meta-analysis to assess the relationship between early effective intervention and macrovascular and microvascular complications in prediabetic patients. Methods PubMed, Embase, and Cochrane Central Register of Controlled Trials were searched for the randomized controlled trials of lifestyle or/and drugs intervention in prediabetes from inception to 2023.9.15. Two investigators independently reviewed the included studies and extracted relevant data. Random or fixed effects model meta-analysis to derive overall relative risk (RR) with 95% CI for all-cause mortality, cardiovascular events, and microvascular complications. Results As of September 15, 2023, a total of 7 effective intervention studies were included, comprising 26 articles out of 25,671 articles. These studies involved 26,389 patients with a total follow-up duration of 178,038.6 person-years. The results indicate that effective intervention can significantly reduce all-cause mortality in prediabetic patients without a history of cardiovascular disease by 17% (RR 0.83, 95% CI 0.70-0.98). Additionally, effective intervention reduced the incidence of retinopathy by 38% (RR 0.62, 95% CI 0.70-0.98). Furthermore, the study results suggest that women and younger individuals have lower all-cause mortality and cardiovascular mortality. Subsequently, we conducted an in-depth analysis of patients without a history of cardiovascular disease. The results revealed that prediabetic patients with a 10-year cardiovascular risk >10% experienced more significant benefits in terms of all-cause mortality (P=0.01). When comparing the results of all-cause mortality and cardiovascular mortality from the Da Qing Diabetes Prevention Outcome Study longitudinally, it was evident that the duration of follow-up is a key factor influencing long-term benefits. In other words, the beneficial effects become more pronounced as the intervention duration reaches a certain threshold. Conclusion Early effective intervention, which significantly reduces the incidence of diabetes, can effectively lower all-cause mortality in prediabetic patients without a history of cardiovascular disease (especially those with a 10-year cardiovascular risk >10%), with women and younger individuals benefiting more significantly. Additionally, the duration of follow-up is a key factor influencing outcomes. The conclusions of this study can provide evidence-based guidance for the clinical treatment of prediabetic patients to prevent cardiovascular and microvascular complications. Systematic review registration https://www.crd.york.ac.uk/prospero, identifier CRD42020160985.


Introduction
Prediabetes is defined as the intermediate metabolic state between normal blood glucose levels and diabetes, including impaired glucose tolerance (IGT) and impaired fasting glucose (IFG).According to the latest estimates of the International Diabetes Federation (IDF) (2017), 352 million adults (7.3%) can be classified as prediabetes (1).Prediabetes is the risk factor for type 2 diabetes and its microvascular complications (2)(3)(4).Multiple large sample meta-analyses have shown that prediabetes is also risk factor for macroangiopathy (5,6) and is associated with increased risk of liver cancer, endometrial cancer, and gastric/colorectal cancer (7).
Therefore, timely identification of prediabetic individuals and effective management are key to preventing the onset of diabetes.American Diabetes Association (ADA) suggests that lifestyle intervention is the basic management for prediabetes.The recommended drug intervention may consider IGT and IFG, age <60 years, body mass index (BMI) >=35kg/m 2 , family history of firstdegree related diabetes, high concentration individuals with triglycerides proceed (8).Studies have shown that early intervention can significantly reduce the incidence of diabetes in patients with prediabetes (9,10).Such as insulin-sensitizing agents and treatment that reduces postprandial hyperglycemia can reduce the risk of progression to T2DM in high-risk prediabetes subjects (11), and lifestyle intervention is the cornerstone for preventing progression to diabetes (12).
However, for prediabetes, just like diabetes, the primary focus of early intervention should be on preventing the occurrence of macrovascular and microvascular complications.So, can interventions that effectively reduce the incidence of diabetes in prediabetic patients also effectively reduce the occurrence of microvascular and macrovascular complications?Based on a comprehensive review of current evidence, the impact of effective intervention on long-term outcome measures, including all-cause mortality and cardiovascular events, remains inconsistent.For example, the Da Qing Diabetes Prevention Outcome Study (DQ) demonstrated that lifestyle intervention could effectively reduce the incidence of diabetes and suggest a reduction in all-cause mortality (13-17).The STOP-Non-insulin dependent diabetes mellitus trial (STOP), which combined lifestyle and acarbose, showed a significant reduction in cardiovascular events (18-21).However, the Acarbose Cardiovascular Evaluation Study (ACE), which applied acarbose in combination with lifestyle intervention, was effective in reducing diabetes incidence but had no effect on cardiovascular events (22,23).Therefore, the current evidence does not provide clear guidance for clinicians regarding the long-term benefits of prediabetic patients.Hence, we aim to assess the impact of effective intervention on the long-term outcomes of prediabetic patients through a systematic review and meta-analysis.

Methods
This study was implemented and reported in accordance with the guidelines for systematic review and meta-analysis of the preferred reporting project (PRISMA) (24).This analysis does not involve personal information, ethical approval or patient consent is exempted, and has been prospectively registered in PROSPERO: CRD42020160985.

Data sources and searches
Pubmed, Embase and Cochrane Central Register of Controlled Trials for randomized controlled studies of prediabetes interventions were searched from inception to 2023.9.15.At the same time, references and related systematic reviews are included in the literature to help complete the search.A manual search of the references of included trials supplemented the electronic search.

Study selection
Studies were included if they met the following PICO(S) (participants, intervention, comparators, outcomes (study designs)) criteria:

Definitions of different indicators/standards
Effective Intervention: In the RCTs we included in the analysis, the intervention that can significantly reduce the occurrence of diabetes is referred to as an effective intervention, and when compared to the control group, it achieves statistical significance (P<0.05).
Cardiovascular History: Patients without a history of cardiovascular disease, meaning individuals included in the study who do not have obvious coronary heart disease.Studies specifically including patients with a history of coronary heart disease are classified as patients with cardiovascular history.
Age Classification: The "Report on the Nutrition and Chronic Diseases Status of Chinese Residents (2020)" indicates that being over 50 is a significant risk factor for the progression from prediabetes to diabetes.Additionally, individuals over the age of 50 tend to exhibit more abnormalities in glucose and lipid metabolism, along with a higher incidence of coronary heart disease.So, we using 50 years as the cutoff, individuals aged 50 or older are categorized as older patients, while others are considered general patients, including those younger than 50 and those older than 50.
Weight Classification: According to the World Health Organization (WHO) classification criteria for obesity, a BMI greater than 25 is considered obese.Therefore, based on a body mass index (BMI) of 25 as the threshold, individuals with a BMI greater than or equal to 25 are classified as overweight patients, while others are considered general patients, including those with a BMI less than 25 and those with a BMI greater than 25.
10-Year Cardiovascular Risk: The 10-year cardiovascular risk of the composite cardiovascular outcome was calculated by multiplying the annualized rate of cardiovascular outcome in the control group by 10 years.High cardiovascular risk is defined as the 10-year cardiovascular risk > = 10%, and low cardiovascular risk is defined as 10-year cardiovascular risk <10%.

Data extraction and quality assessment
Screening out duplicate articles with EndNote X9, two authors (XA and YZ) extracted and checked the data with the Office form separately, included the basic characteristics of the included study (including the country where the study was carried out, the number of patients included, the basic characteristics of the included patients, study design methods, interventions, intervention time, follow-up time, main outcomes, JADAD score) and the basic characteristics of the included patients (age, gender, race, body mass index (BMI), smoking history, hypertension, dyslipidemia), if there is any objection, it would be resolved through consultation.If the result data is reported at multiple follow-up points, the data from the longest follow-up would be selected.
I 2 and Cochran's Q test were used to evaluate the heterogeneity of treatment effects between trials.P <0.05 and I 2 > 50% of the Cochran's Q test indicated significant heterogeneity.For studies with low heterogeneity, fixed-effect model is used.Studies with greater heterogeneity, the source of heterogeneity is first searched, and a random-effects model is used for analysis.For studies with no source of heterogeneity, only descriptive analysis.

Grading of the evidence
The quality of the studies was evaluated by JADAD 7 points (25), including the generation of random sequence, randomization hiding, blind method, withdrawal and exit. 1 to 3 is divided into low-quality study, 4 to 7 is divided into high-quality study.The two authors independently assessed the risk of publication bias in each study, and disputes were resolved through negotiation.

Data analysis
The results were analyzed with RevMan 5.2 provided by Cochrane Collaboration (26) and Prism 9.5.1.The secondary classification results were analyzed by relative risk (RR) and 95% confidence interval (CI).According to the characteristics of the patients, including the history of cardiovascular disease, age, obesity, gender, 10-year cardiovascular risk factors, outcomes were evaluated by subgroup analysis.According to the Cochrane manual, funnel plots were used to assess potential publication bias (27).Heterogeneity was evaluated using I 2 statistics.I 2 <25% represents low heterogeneity, 25-50% represents medium heterogeneity, and> 50% represents high heterogeneity, considering the heterogeneity difference P value, if I 2 <50% and P> 0.05 then Fixed effect model was used, if I 2 > 50% and P <0.05, random effect model was used.For studies with large heterogeneity that cannot find the source of heterogeneity, only descriptive analysis is made.

Results
As shown in Figure 1 in the flowchart, we retrieved a total of 25,671 articles from three databases.First, we excluded duplicate articles, leaving us with 16,539 articles.Next, we reviewed titles and abstracts and excluded 11,322 unrelated articles, including 1,320 reviews, 374 animal experiments, 2,747 registered studies, and 6,881 articles without interventions.After a full-text review, we excluded 5,191 articles that lacked long-term outcome measures, leaving us with 26.No additional articles were added from the reference lists.In the end, a total of 8 studies were included for analysis, and two of these studies utilized a 2x2 factorial design.Simultaneously, we further analyzed seven studies based on the effectiveness of the intervention measures employed in the research (Figures 1, 2).
Included in the studies, the STOP study, DREAM study, NAVIGATOR study, and ACE study were randomized doubleblind placebo-controlled studies, while the DQ study, DPP study, FDPS study were randomized controlled studies.Regarding the intervention measures in different studies, the DQ study and FDPS study solely involved lifestyle interventions.Studies involving drug interventions were all based on lifestyle interventions, including the STOP study (acarbose, 100mg, three times/day), DPP study (metformin, 850mg, twice/day), DREAM study (rosiglitazone 8 mg daily), NAVIGATOR study (dose of valsartan started at 80mg daily and increased to 160mg daily after 2 weeks), and ACE study (acarbose, 50mg, three times/day).We also summarized the baseline characteristics of the included studies, including age, gender ratio, ethnicity, BMI, smoking history, history of hypertension, history of lipid abnormalities, and history of cardiovascular disease (Table 2).
For the studies we included, the JADAD scores were 7 for the STOP, DREAM, NAVIGATOR, and ACE studies, and 5 for the DQ, DPP, and FDPS studies, all indicating high-quality research (Table 3).

Main outcome
For the main outcome, namely all-cause mortality, a total of 6 studies were included, comprising 25,867 patients with a total follow-up duration of 172,505.4person-years.The results indicated low heterogeneity among the various studies (I 2 = 0%, P=0.40).Using a fixed-effect model for analysis, the results showed that effective intervention could reduce all-cause mortality by 9% (RR 0.91, 95% CI 0.83-1.01),but it was not statistically significant (P=0.07).The funnel plot exhibited a symmetrical shape, suggesting the absence of publication bias (Figure 3).
We further conducted subgroup analyses on patient characteristics, including a history of cardiovascular disease, gender, age, and BMI.The results revealed that early effective intervention significantly reduced all-cause mortality by 18% in prediabetic patients without a history of cardiovascular disease (RR=0.82,95% CI=0.69~0.98,P=0.03,I 2 = 0%), but not in patients with a history of cardiovascular disease (P=0.29).Moreover, women (P=0.01) and younger individuals (P=0.01)exhibited lower all-cause mortality rates.Subgroup analysis based on BMI showed no significant benefit in all-cause mortality for general patients (P=0.08) or overweight patients (P=0.28)(Figure 4).
For patients without a history of cardiovascular disease, we conducted a more in-depth analysis based on a 10-year cardiovascular risk.The results revealed that early intervention effectively reduced all-cause mortality by 21% for patients with a 10year cardiovascular risk >10% (RR=0.79,95% CI=0.66~0.95,P=0.01,I 2 = 0%), while the effect was not significant for studies with a 10-year cardiovascular risk <10% (P=0.67).(Supplementary Appendix).
We conducted subgroup analyses based on patient characteristics, including a history of cardiovascular disease, gender, age, and BMI.The results showed that early effective intervention could significantly reduce cardiovascular mortality by 23% in patients without a history of cardiovascular disease (RR 0.77, 95% CI 0.60-1.00),although it was not statistically significant (P=0.05).Women exhibited lower cardiovascular mortality rates.A 23% reduction in cardiovascular mortality was also observed in younger individuals (RR 0.77, 95% CI 0.60-1.00),but it lacked statistical significance (P=0.05).No significant benefits were demonstrated for overweight and general populations (Figure 5).
For patients without a history of cardiovascular disease, we conducted a deeper analysis based on a 10-year cardiovascular risk.The results showed that effective intervention could reduce cardiovascular mortality by 26% for patients with a 10-year cardiovascular risk >10% (RR=0.74,95% CI=0.57~0.98,P=0.03,I 2 = 0%), while the effect was not significant for patients with a 10year cardiovascular risk <10% (P=0.80)(Supplementary Appendix).
Regarding microvascular complications, the number of included studies was limited.For microvascular events, subgroup analysis was not conducted due to the limited number of studies.The overall results showed that early effective intervention could reduce retinopathy by 38% (RR=0.62,95% CI=0.44~0.86,P=0.005,I 2 = 0%), although there was a trend towards reducing kidney disease by 15%, it was not significant (P =0.05).Peripheral neuropathy did not demonstrate significant benefits (P =0.08) (Figure 3).
During the analysis of the included studies, we noted that the DQ study conducted data summarization and analysis at 6, 20, 23, and 30 years.Although diabetes incidence was reduced with effective intervention at all time points, microvascular events showed benefits early on.However, for cardiovascular events, there were no advantages observed in the first 20 years.Starting at 23 years, the benefits in all-cause mortality (HR) and cardiovascular mortality became significantly prominent.This suggests that besides patient characteristics, follow-up duration is also a crucial factor for effective intervention (Figure 6).

Discussion
To the best of our knowledge, this is the first systematic review and meta-analysis to assess the benefits of effective intervention on microvascular and macrovascular events in prediabetic patients.Through standardized research methods, we found that, for patients without a history of cardiovascular disease, women, and younger individuals, there were lower all-cause mortality rates.In patients without a history of cardiovascular disease, for those with a 10-year cardiovascular risk >10%, there were significant reductions in allcause mortality and cardiovascular mortality.Additionally, through longitudinal comparisons in the DQ study, we found that follow-up duration may be a critical factor in assessing effectiveness.In light of these findings, we provide evidence to guide clinical practitioners in long-term interventions for prediabetic patients.
Combining prior relevant research with our study results, it is evident that prediabetic patients require attention and early management.Some now consider prediabetes as a suboptimal health condition, and prediabetes and diabetes are seen as a continuum of elevated blood glucose and cardiovascular risk.This suggests that the principles applicable to the treatment of type 2 diabetes should also be applied to prediabetes (43,44).Firstly, prediabetes shares the same vascular risk factors as diabetes, such as glucose abnormalities, hypertension, lipid abnormalities, obesity, and insulin resistance (44).These factors may also accelerate the development of atherosclerosis (45)(46)(47), and for prediabetic patients with comorbidities such as atrial fibrillation, there is also an increased likelihood of major cardiovascular and cerebrovascular events (48).Additionally, prediabetic patients may exhibit physiological and pathological abnormalities similar to insulin resistance (44,49), which appears to be a significant factor in the development of cardiovascular diseases (50).For diabetic patients, the risk of cardiovascular events is 2-4 times higher than that of non-diabetic individuals (51, 52), and these risks are nearly equivalent to those of prediabetic patients (53)(54)(55).
Considering the current evidence on prediabetes management, lifestyle intervention forms the foundation.These studies collectively suggest that regardless of medication intervention, lifestyle intervention should be maintained in the long term for individuals with prediabetes.The 'Expert Consensus on the Prevention of Type 2 Diabetes in Chinese Adults' states that when intensive lifestyle intervention persists for more than 6 months and blood glucose control remains unsatisfactory, medication treatment may be considered for young individuals with good economic conditions, higher health needs, and access to medical services.The consensus identifies three classes of medications for prediabetes intervention: metformin, acarbose, and thiazolidinediones (TZDs).However, it emphasizes that lifestyle intervention primarily aids in weight loss, and obesity is a  58).This is because weight loss can reduce or even reverse ectopic deposition, improve chronic inflammation, cardiovascular risk factors, and achieve the delay of disease progression (59).A 4-year follow-up observational longitudinal study indicated that weight loss surgery can significantly increase the remission rate of prediabetes (60).
The progression of prediabetes is reversible, and effective intervention is crucial for the prognosis of individuals with prediabetes (61).Combining diet and physical exercise is more effective in reducing diabetes development than any single strategy, as calorie intake and physical activity are independently associated with reducing diabetes risk, and their combination may produce an additive effect (62)(63)(64).The DQ study indicated that the reduction in diabetes incidence in the intervention group was mainly due to changes in dietary structure, increased physical activity, or improved physical fitness (13).Recent cost-effectiveness analyses have shown that lifestyle intervention is the most cost-effective approach (65).The DPP study demonstrated that intensive lifestyle intervention could reduce cardiovascular risk factors, including high blood pressure, high triglyceride levels, and low high-density lipoprotein levels (66).The effectiveness of lifestyle intervention diminishes over time, indicating the need for long-term adherence to realize significant benefits (62) However, the DQ study showed that the advantages of lifestyle intervention in reducing diabetes incidence persisted for 30 years after intensive lifestyle intervention.
The DQ study provides three potential explanations for the longterm benefits of lifestyle intervention.Firstly, lifestyle intervention may lead to sustained changes in normal behavior beyond the study period.Secondly, these interventions may lead to changes in preventive care and health promotion efforts provided by community clinics, with effects observed beyond the study period.Finally, lifestyle intervention may result in a type of metabolic memory, where motivation for long-term lifestyle changes is crucial for improving intervention compliance (67).In this regard, Penn et al. suggested that an expectation of one year or longer is a necessary condition for establishing long-term behavioral changes (68).For high-risk populations with prediabetes, especially those with additional cardiovascular risk factors, medication treatment is recommended (6).Medication intervention can more rapidly reduce and stabilize risk factors in prediabetic patients.The DPP study showed that metformin did not have a treatment advantage over lifestyle intervention (29), but metformin's effectiveness in reducing the conversion rate of prediabetes still exists (69).Changes in lifestyle may be the most effective, and the addition of pharmacological drugs does not necessarily increase the benefits.However, the STOP study also showed that early intervention with acarbose could significantly reduce cardiovascular outcomes.This could be related to the multiple benefits, including weight reduction, decreased BMI, waist circumference, blood pressure, postprandial 2-hour blood glucose, and triglyceride levels associated with acarbose.Similarly, the ACE study, which also used acarbose, did not demonstrate significant benefits, possibly due to the lower acarbose dosage and the study population's high prevalence of coronary heart disease (22 In the NAVIGATOR study, the most convincing evidence for improved cardiovascular outcomes with higher dosage was seen with valsartan, where patients received twice the daily dose.Apart from lifestyle modifications, medications, and surgical interventions, it is crucial for clinical physicians to develop more rational, precise, and effective comprehensive management strategies for prediabetic patients, especially if high-risk populations, including those at risk for diabetes and cardiovascular diseases, can be accurately identified.Of course, we can refer to currently recognized relevant risk factors, including age, body mass index, family history of diabetes, history of hypertension, and physical activity level.However, more precise prediction models or scoring systems can better quantify the risks associated with prediabetic patients.For example, a recent study in China, based on data from 184,188 prediabetic patients, developed and validated personalized prediabetes prediction charts for Chinese adults, aiding in the identification of high-risk populations (72).Nevertheless, overall, there is currently no widely recommended optimal model available.Therefore, future research should focus on improving the clinical relevance and predictive performance of existing models (73,74).
In our study, due to data limitations, there was relatively less data available for gender, age, and weight.Gender data were only available from the DQ study, and although age was categorized with a cutoff at 50 years, the NAVIGATOR and ACE studies included patients aged 50 and above.However, for the STOP study, DPP, FDPS, and DREAM, as well as NAVIGATOR and ACE, there were significant age differences.Nevertheless, we did not observe additional benefits in older individuals.It is possible that both studies included patients with a history of cardiovascular disease.Regarding gender, subgroup analysis was conducted only in the DQ study, revealing that although a reduction in mortality was primarily seen in female patients, early effective intervention did not significantly benefit males.This could be attributed to a higher number of male smokers, which may have attenuated the benefits of early intervention.Subgroup analysis based on weight revealed that for overweight patients, cardiovascular benefits were not significant.Further analysis based on an average BMI greater than 30 showed that even after excluding the DQ and ACE studies, all-cause mortality remained non-significant.Follow-up time is a key factor to consider for assessing long-term benefits.The DQ study indicated that differences in cardiovascular disease mortality between the intervention and control groups began to appear 12 years after the start of the study and slowly increased to 17% by the 20-year follow-up, reaching statistical significance only after 23 years (15).The WCK study also suggested that at the end of 10 years of treatment, there was no significant impact on mortality (75), but significant differences started to emerge when the follow-up period exceeded 22 years (76).Although the FDPS study did not show a reduced risk of cardiovascular disease over the 10-year follow-up, the incidence of type 2 diabetes in the intervention group remained significantly lower than the control group (40).This was due to a lower complication rate in the intervention group, primarily occurring over a longer time after randomization, possibly indicating a delayed onset of diabetes.This also suggests that follow-up time is a crucial factor for assessing the long-term benefits of intervention.
Our study, based on the synthesis of current high-quality research, indicates that effective intervention, especially lifestyle intervention, significantly benefits individuals with prediabetes, particularly those without a history of coronary heart disease.Subgroup analysis of the data suggests that individuals at high cardiovascular risk for 10 years, as well as women, exhibit more pronounced benefits.At present, the evidence does not clearly define age or BMI as factors influencing the effectiveness of longterm intervention.Lastly, considering the DQ study, follow-up time appears to be a crucial factor for evaluating long-term benefits of intervention.Therefore, patients should adhere to interventions in the long term, and lifestyle intervention is the most cost-effective and easily accepted approach.Given the current data limitations, detailed stratification of patients based on baseline levels such as age, weight, family history, blood pressure, and lipid profiles, which are risk factors for both macrovascular and microvascular complications, cannot be performed.Follow-up time is a critical factor, but the studies included in the analysis had varying followup durations, ranging from 3 to 30 years.Consequently, achieving balanced data among the groups based on current evidence may pose some challenges and potential biases.Therefore, we hope that further research will continue to address the disparity in follow-up times to obtain more reliable evidence.
It is estimated that by 2025, the global population defined as having IFG and/or IGT as prediabetes will reach 472 million (77).Individuals with prediabetes should be aware of their increased risk of future vascular complications (78).However, the vast majority of prediabetic individuals may not even be aware of their risk of developing diabetes (79).Based on the current high-quality evidence, our results indicate that early effective intervention can be used as an effective primary prevention strategy for diabetes, reducing all-cause mortality.This benefit is particularly notable in individuals without a history of cardiovascular disease, those at high cardiovascular risk for 10 years, and women.This provides a basis for clinical guidance in the intervention of individuals with prediabetes.
the abstract or lack of data and cannot obtain full-text articles, unable to extract data for research, and non-English literature.Studies included less than 100 participants.
Comparator: The intervention measures in the control group are either a blank control, a placebo, or lifestyle guidance.4. Outcome: main outcome is all-cause mortality, secondary outcomes include composite cardiovascular outcomes (including cardiovascular mortality, non-fatal myocardial infarction, non-fatal stroke, heart failure hospitalization, arterial revascularization or hospitalization of unstable angina), core cardiovascular outcomes (including cardiovascular mortality, non-fatal myocardial infarction, non-fatal stroke), cardiovascular mortality, microvascular c o m p l i c a t i o n s ( r e t i n o p a t h y , r e n a l d i s e a s e , peripheral neuropathy).5. Study designs: The study type was randomized controlled trial (blind or not); JADAD score between 5 and 7 points.Studies were excluded for: duplicate publications, only

TABLE 1
Basic characteristics of the including studies.

TABLE 2
Basic characteristics of the including patients.
NR, Not reported; T, Treatment group; C, control group.

TABLE 3
Publication bias (JADAD scores) of the including studies.
Subgroup analysis of all-cause mortality.Subgroup analysis of cardiovascular mortality.