Trimethylamine N-Oxide in Heart Failure: A Meta-Analysis of Prognostic Value.

OBJECTIVE: The present study aimed to explore the prognostic value of trimethylamine N-oxide (TMAO) in heart failure (HF).
METHODS: PubMed, Excerpta Medica Database (EMBASE), Cochrane Library, Web of Science, Wanfang Database, SINOMED, China Science and Technology Journal Database (VIP), and China National Knowledge Infrastructure (CNKI) were searched up to June 1, 2021. Studies recording the major adverse cardiovascular events (MACEs) or all-cause mortality in HF patients and their circulating TMAO concentrations were included. Meta-analysis was performed using Stata 13.0.
RESULTS: Ten articles (12 studies) involving 13,425 participants from 2014 to 2021 were considered. Compared to low-level TMAO, elevated TMAO was correlated with MACEs and all-cause mortality in HF (RR: 1.28, 95% CI: 1.17, 1.39, P < 0.0001, random-effects model and RR: 1.35, 95% CI: 1.28, 1.42, P < 0.0001, random-effects model, respectively). Consistent results were obtained in all examined subgroups as well as in the sensitivity analysis.
CONCLUSION: Elevated TMAO may be an adverse prognostic indicator in patients with HF. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=267208.

Introduction

Heart failure (HF) is an increasing public health problem worldwide (1). Evidence from epidemiological studies has shown that approximately 64.3 million people worldwide suffer from HF (2). Patients with HF usually experience the following malignant disease cycle: “hospitalization-improvement-discharge-rehospitalization” (3). Despite recent advances in the diagnosis and treatment of HF, patient mortality remains high (4, 5), suggesting that many risk factors for HF remain unexplored.

Intestinal microorganisms are involved in the occurrence and progression of many diseases (6). Recent studies have revealed the associations between human gut microbes and HF (7, 8). As one of the metabolic products in intestinal microorganisms, trimethylamine N-oxide (TMAO) has been demonstrated to promote the development of HF and is an important risk factor for patients with HF (9). Schuett et al. (10) followed more than 4,000 patients with HF for 9.7 years; and they reported that after excluding factors, such as body weight, smoking, hypertension, diabetes, and atrial fibrillation, high levels of TMAO are associated with all-cause mortality in patients with HF. TMAO aggravates HF by damaging vascular endothelial function, affecting mitochondrial metabolism, and promoting myocardial fibrosis (11, 12). Owing to the limited sample size, further evaluation is still needed.

Therefore, the present systematic review and meta-analysis aimed to explore the prognostic value of TMAO in HF. To our knowledge, this is the first study to include data from around the world describing the association between TMAO and HF.

Methods

Study Registration

This meta-analysis was performed in accordance with the recommendations of the Preferred Reporting Items for Systematic Reviews (PRISMA) statement (13). The review protocol was registered at PROSPERO as CRD42021267208.

Search Strategy

We searched the following 4 English electronic databases and 3 Chinese literature databases for studies published from inception to June 1, 2021: PubMed, Excerpta Medica Database (EMBASE), Cochrane Library, Web of Science, Wanfang database, SINOMED, VIP database, and China National Knowledge Infrastructure (CNKI). In addition, there was no restriction for language. Taking PubMed as an example, the details of the search strategy are shown in Table 1. This work was completed by two independent reviewers (ZF and JC), and in cases where they disagreed, a third individual (LX) was consulted. Moreover, manual retrieval was performed on Baidu Academic, Google Academic, books, impurities, and conference materials to obtain all the materials related to this study as comprehensively as possible. We also screened the references of the included papers, full texts, and bibliographies of all potential articles, including relevant reviews and meta-analyses, to identify additional eligible studies.

Table 1

Search strategy in the PubMed database.

Number	Search terms
#1	Heart failure [Mesh]
#2	(Cardiac Failure [Title/Abstract]) OR (Heart Decompensation [Title/Abstract])) OR (Decompensation, Heart [Title/Abstract])) OR (Heart Failure, Right-Sided [Title/Abstract])) OR (Heart Failure, Right Sided [Title/Abstract])) OR (Right-Sided Heart Failure [Title/Abstract])) OR (Right Sided Heart Failure [Title/Abstract])) OR (Myocardial Failure [Title/Abstract])) OR (Congestive Heart Failure [Title/Abstract])) OR (Heart Failure, Congestive [Title/Abstract])) OR (Heart Failure, Left-Sided [Title/Abstract])) OR (Heart Failure, Left Sided [Title/Abstract])) OR (Left-Sided Heart Failure [Title/Abstract])) OR (Left Sided Heart Failure [Title/Abstract])
#3	#1OR#2
#4	(“trimethylamine N-oxide” [Mesh])
#5	(trimethyloxamine [Title/Abstract]) OR (trimethylammonium oxide [Title/Abstract]) OR (TMAO [Title/Abstract]) OR (trimethylamine oxide [Title/Abstract])
#6	#4OR#5
#7	#3AND#6

Inclusion and Exclusion Criteria

To prevent bias, the inclusion criteria were prespecified as follows: (a) the subjects were patients with HF; (b) prospective cohort; (c) major adverse cardiac events (MACEs), including cardiovascular mortality, MI, cardiovascular hospitalization or revascularization, in addition to all-cause mortality were reported; and (d) hazard ratio (HR)/relative risk (RR) and 95% CI were reported.

Studies were excluded if any of the following criteria were observed: (a) duplications; (b) lack of data on TMAO levels and heart failure; or (c) case reports, animal trials, review articles, systematic reviews, meta-analyses, commentaries, editorials, or meeting abstracts.

Data Extraction

All data were independently extracted by two reviewers (DL and JL), and disagreements were resolved by discussion. If necessary, a third author (XC) was involved. Study information of the included studies was recorded, including the first author, publication year, participant characteristics, sample size, region, circulating TMAO concentration, duration of follow-up, adjusted risk factors, and outcome assessment.

Quality Assessment

Two authors (XC and LX) independently assessed the methodological quality of the included studies. The Newcastle-Ottawa quality assessment scale (NOS) was used to evaluate the risk of bias based on study group selection, group comparability, and ascertainment of exposure or outcome (14). The maximum score of this scale is 9 points, and studies with a score ≥7 are rated as high quality (15).

Data Synthesis and Statistical Analysis

The data were analyzed with Stata (version 13.0) (16). RRs and 95% CIs were used to estimate the combined effects. The overall effect was calculated by a Z-test, and P < 0.05 (2-tailed) was defined as statistically significant. Potential heterogeneity was evaluated by Cochran Q and I2 statistics (17). A fixed-effect model was employed when heterogeneity was low (P ≥ 0.05, I2 ≤ 50%) (18). However, when high heterogeneity occurred (P < 0.05, I2 > 50%), we further analyzed its potential sources from the following three aspects: clinical heterogeneity, methodological heterogeneity, and statistical heterogeneity. We evaluated clinical heterogeneity first. If there was evident clinical heterogeneity, subgroup analysis was performed. If clinical heterogeneity was evident and subgroup analysis could not be conducted, descriptive analysis was only used. After excluding clinical and methodological heterogeneity, statistical heterogeneity was considered, and a random-effect model was used (19). A funnel chart, Egger's, and Begg's test were constructed to evaluate publication bias (19). Publication bias will be adjusted by the method of trim and filling.

We performed sensitivity analyses to evaluate reliable results. The methods included changing the type of analysis methods (random-effects model or fixed-effects model), eliminating each of the included studies one by one, and then combining the effect quantities.

Ethics

Because patient privacy was not involved in the present study, ethical approval was not needed.

Results

Literature Search and Screening

The search of 8 databases identified 396 articles for further evaluation (129 from CNKI, 23 from WANFANG, 13 from VIP, 36 from SINOMED, 55 from PubMed, 6 from the Cochrane Library, 114 from Embase, and 20 from the Web of Science), of which 190 were removed after review due to duplicate records. After reading the titles and abstracts, 178 were excluded for various reasons. Finally, only 10 studies (10, 20–28) met the inclusion criteria after screening full texts. Figure 1 shows the detailed process of the study selection process.

Figure 1

Flowchart of study selection.

Study Characteristics and Quality Assessment

Ten articles (12 studies) involving 13,425 participants from 2014 to 2021 were included in the present study. Most trials were performed in Europe (3 in the UK, 1 in Norway, 1 in Germany, and 1 in other); however, trials were also performed in the United States (2 trials) and Asia (2 in China). Only three studies (23, 24, 28) included patients with acute HF; all others included patients with chronic HF. Three studies were multicenter, and the others were single-center. The follow-up duration ranged from 1 to 9.7 years with a median follow-up of 3.4 years. The average age of the patients ranged from 57 to 78 years old. Most trials had more males than females. Baseline TMAO levels were reported in all included studies. All the studies reported data on MACEs or all-cause mortality. All observational studies were adjusted for multiple potential confounding factors, and more than half of the studies were rated as “good” quality. The characteristics of the included studies are listed in Table 2.

Table 2

Characteristics of the selected studies and quality assessment.

References	Location	HF type	Scale	Follow-up, years	Participants (male, %)	Age, years	TMAO level (umol/L)	Outcome	Adjusted risk factors	Study quality
Tang et al. (20)	USA	Chronic	Single center	5	720 (59)	66 ± 10	5.0 (3.0, 8.5)	All-cause mortality	Yes	8
Tang et al. (21)	USA	Chronic	Single center	5	112 (75)	57 ± 14	5.8 (3.6, 12.1)	All-cause mortality	Yes	8
Trøseid et al. (22)	Norway	Chronic	Single center	5.2	155 (83)	57 ± 11	NA	All-cause mortality	Yes	7
Suzuki and Heaney (23)	UK	Acute	Single center	1	972 (61)	78 (69–84)	5.6 (3.4, 10.5)	All-cause mortality	Yes	6
Schuett et al. (10)	Germany	Chronic	Single center	9.7	2,490 (NA)	63 ± 10	4.73 (3.4, 6.82)/4.73 (3.22, 6.85)	All-cause mortality	Yes	6
Liu (24)	China	Acute	Single center	1	64 (67)	70 ± 14	6.56 ± 11.83	All-cause mortality	Yes	6
Suzuki et al. (25)	European	Chronic	Multicenter	1	2,234 (74)	70 (61–78)	5.9 (3.6, 10.8)	All-cause mortality; MACE	Yes	8
Suzuki et al. (25)	European	Chronic	Multicenter	2	2,234 (74)	70 (61–78)	5.9 (3.6, 10.8)	All-cause mortality; MACE	Yes	8
Suzuki et al. (25)	European	Chronic	Multicenter	3	2,234 (74)	70 (61–78)	5.9 (3.6, 10.8)	All-cause mortality; MACE	Yes	8
Salzano et al. (26)	UK	Chronic	Single center	5	196 (49)	73 (67–78)	7.0 (4.2, 12.5)	All-cause mortality	Yes	6
Zhou et al. (27)	China	Chronic	Multicenter	1.8	1,208 (689)	73 (64–80)	4.5 (2.83, 7.92)	All-cause mortality, MACE	Yes	7
Israr et al. (28)	UK	Acute	Single center	1	806 (61)	78 (69–84)	10.2 (5.8, 18.7)	All-cause mortality; MACE	Yes	6

USA, the United States; UK, United Kingdom.

Primary Outcome: The Association Between TMAO Level and MACEs in Patients With HF

Five cohort studies with a total of 8,716 participants reported an association of MACEs with TMAO levels (RR: 1.28; 95% CI: 1.17, 1.39; P < 0.0001; I2= 56.1%; P-heterogeneity = 0.058; random-effects model; Figure 2). This result indicated that a high circulating TMAO concentration was associated with a greater risk of MACEs in patients with HF. Due to the high heterogeneity, subgroup analysis was implemented. The subgroup analysis did not affect the results (Figure 3).

Figure 2

Forest plot for the association between trimethylamine N-oxide (TMAO) level and major adverse cardiovascular events (MACEs).

Figure 3

Subgroup analysis for the association between TMAO level and MACEs.

Secondary Outcome: The Association Between TMAO Level and All-Cause Mortality in Patients With HF

The pooled analysis comparing the all-cause mortality in HF patients with high and low circulating TMAO concentrations involved all included studies (RR: 1.35; 95% CI: 1.28, 1.42; P < 0.0001; I2 = 56.5%; P-heterogeneity = 0.008; random-effects model; Figure 4). The results showed that the risk of all-cause mortality was greater with higher circulating TMAO concentrations among patients with HF. Considering significant heterogeneity, we performed subgroup analysis based on study characteristics to explore the potential sources of heterogeneity (Table 3). Seven items, including study scale, location, HF type, sample size, age, follow up, and study quality, were included in the subgroup analysis to evaluate the impact on heterogeneity. In general, the results were not influenced by these factors, indicating that the meta-analysis result was stable.

Figure 4

Forest plot for the association between TMAO level and all-cause mortality.

Table 3

Subgroup analysis for the association between TMAO level and all-cause mortality according to study characteristics.

Subgroups	Studies, n	Effects model	Over effect		Heterogeneity
			RR (95% CI)	P value	I 2, %	P value
All	12	Random	1.35 (1.28, 1.42)	<0.0001	56.5	0.008
Scale
Single-center	8	Fixed	1.32 (1.25, 1.40)	<0.0001	64.8	0.006
Multi-center	4	Random	1.49 (1.29, 1.71)	<0.0001	5.8	0.364
Location
USA	2	Random	1.21 (1.19, 1.34)	<0.0001	52	0.149
European	8	Fixed	1.40 (1.31, 1.49)	<0.0001	47	0.067
Asia	2	Random	1.53 (1.22, 1.90)	<0.0001	69.8	0.069
HF type
Chronic	9	Random	1.34 (1.28, 1.48)	<0.0001	67	0.002
Acute	3	Fixed	1.31 (1.21, 1.42)	<0.0001	0	0.834
Participants
<900	6	Random	1.26 (1.16, 1.36)	<0.0001	55.9	0.045
≥900	6	Fixed	1.43 (1.32, 1.54)	<0.0001	44.1	0.112
Age
<70	5	Random	1.33 (1.22, 1.44)	<0.0001	74.3	0.004
≥70	7	Fixed	1.36 (1.27, 1.46)	<0.0001	36.8	0.148
Follow up
<5	7	Fixed	1.35 (1.26, 1.45)	<0.0001	0	0.449
≥5	5	Random	1.33 (1.22, 1.46)	<0.0001	79.4	0.001
Study quality
Good	8	Random	1.32 (1.24, 1.40)	<0.0001	51.0	0.046
Fair	4	Random	1.43 (1.29, 1.67)	<0.0001	67.6	0.026

Publication Bias Analysis

For the relationship between TMAO and MACEs in the present study, the funnel plot appeared asymmetric (Figure 5A), but further Egger's and Begg's tests demonstrated that there was no publication bias (P = 0.922 for Egger's test and P = 0.462 for Begg's test). In addition, the funnel plot and Egger's test all indicated potential publication bias for the relationship between TMAO and all-cause mortality (P = 0.017 for Egger's test; Figure 5B). Although publication bias existed, the results of trim-and-fill method suggested that the meta-analysis results were robust (RR: 0.251; 95% CI: 0.11, 0.392; P < 0.001 for MACEs; RR: 0.257; 95% CI: 0.153, 0.36; P < 0.001 for all-cause mortality; Figure 6).

Figure 5

Funnel plot for publication bias. (A) Funnel plot for publication bias of the association between TMAO level and MACEs; (B) Funnel plot for publication bias of the association between TMAO level and all-cause mortality.

Figure 6

Trim and fill method on MACEs (A) and all-cause mortality (B).

Sensitivity Analysis

Sensitivity analyses were performed to assess whether the results of this meta-analysis were stable. The results showed that no obvious effect was found after deleting the studies one by one, which suggested that the study results were credible (Figure 7).

Figure 7

The sensitivity analysis of the study. (A) The sensitivity analysis of the association between TMAO level and MACEs; (B) The sensitivity analysis of the association between TMAO level and all-cause mortality.

Discussion

At present, there are few reported meta-analyses on the relationship between TMAO and the poor prognosis of HF; the included studies are mainly from Europe and the United States, and there is a lack of data on Asian HF populations (29). With this background, we performed the first systematic review and meta-analysis involving studies from around the world to explore whether high levels of TMAO are related to the poor prognosis of HF patients. These 10 articles were published from 2014 to 2021, which reflected recent results about the role of circulating TMAO levels in predicting the poor prognosis of patients with HF. This systematic review and meta-analysis provided relatively reliable evidence that elevated TMAO is related to MACEs and all-cause mortality in patients with HF.

Trimethylamine N-oxide (TMAO) is a small odorless molecule (molecular mass of 75.11 g/mol) (30). Dietary choline, betaine, and L-carnitine are metabolized by intestinal microorganisms to generate trimethylamine (TMA), which is further converted to TMAO by hepatic flavin monooxygenase (FMO) (31, 32). Although TMAO can maintain normal physiological activities of the human body, high concentrations of TMAO may cause various cardiovascular diseases (33). Animal studies have shown that both choline supplementation (a precursor to TMAO production) and direct dietary TMAO supplementation increase the level of TMAO, exacerbating myocardial fibrosis and worsening cardiac function (34). Chen et al. (35) proposed that inhibitors of TMA formation reduce circulating TMAO levels in obese mice induced by Western diets, thereby preventing subsequent cardiac dysfunction. TMAO induces HF via multiple pathological mechanisms, including inflammation, mitochondrial dysfunction, oxygen-free radical production, and myocardial fibrosis (Figure 8) (36–40). There are many factors that affect the level of TMAO in the body, including diet, intestinal microbiota, sex, age, kidney function, weight, and drugs (31, 32, 41, 42). In the original studies included in this article, 7 were corrected for age, 5 were corrected for renal function, and 4 were corrected for sex and weight. At the same time, we conducted subgroup analyses of age, scale, location, HF type (chronic or acute), sample size, follow-up, and study quality to explore the potential sources of heterogeneity. Sensitivity analysis was also implemented. Fortunately, subgroup analyses and sensitivity analyses did not change the overall results. We obtained the same result as Li et al. (29), but our original studies were distributed in Europe, the United States, and Asia, which made up for the deficiency of their study. However, it is worth mentioning that we still cannot rule out the influence of diet, gut microbiota, and drugs on TMAO. Red meat, eggs, and sea fish are common sources of TMAO in the diet (31, 41, 43), thereby reminding nutritionists to pay attention to the nutritional benefits of these foods while not ignoring the adverse metabolites (such as TMAO) associated with these foods when formulating nutritional prescriptions for patients with cardiovascular disease.

Figure 8

Mechanism of TMAO exacerbating heart failure (HF) progression.

The rapidly increasing population of patients with HF imposes a heavy social and economic burden. Our results encourage the exploration of interventions to reduce TMAO. At present, methods to reduce TMAO mainly remain in the stage of animal experiments, including 3,3-dimethyl-1-butanol, choline analogs (IMC and FMC), and betaine aldehyde (44–46). The method used to reduce TMAO in clinical is microbial-driven therapy, however, a meta-analysis showed the opposite conclusion (47).

Existing studies have revealed the key role of TMAO in HF. TMAO may participate in the occurrence and development of HF through a variety of mechanisms. However, related scientific research has just started, and only limited studies have revealed the connection between TMAO and HF. There are questions about the relationship between TMAO and HF. First, due to the lack of long-term monitoring of circulating TMAO concentration in patients before HF occurs, it is still difficult to determine whether a high level of TMAO is the triggering factor of HF. The above results provided preliminary evidence that TMAO predicts the poor prognosis of patients with HF. Our findings originated from clinical data, adding new direct evidence, which largely confirms the reliability of the association between high concentrations of TMAO and poor prognosis of HF. Second, TMAO reduction interventions remain in the laboratory stage, and there is no effective treatment in the clinic. Finally, the results of the dose-response relationship provided in the meta-analysis better present the relationship between different dose levels of exposure factors and outcomes, making the results more reliable. However, due to the lack of studies on the relationship between different concentrations of TMAO and the prognosis of HF, the dose-response relationship between TMAO and the poor prognosis of HF remain unknown.

In the future, a large-scale prospective cohort should be further developed to explore the causal relationship between TMAO and HF, starting before the occurrence of HF. At the same time, the relationship between different concentrations of TMAO and the poor prognosis of patients with HF should be discussed. Prospective multicenter randomized controlled trials should be performed to further advance the understanding of TMAO from the laboratory to the clinic to treat patients with HF.

Strengths and Limitations

The present study clarified that TMAO is related to the poor prognosis of patients with HF. Moreover, more than 60 percent of the original studies were of high quality. The present study includes research from around the world, thereby allowing the research results to be widely promoted and overcoming the deficiencies of existing research to a certain extent. The detailed description of baseline characteristics also allowed identification of the source of heterogeneity.

However, the present study had several limitations. First, although we established rigorous inclusion criteria, some results indicated publication bias. Second, there was a lack of studies on the relationship between different concentrations of TMAO and MACE or all-cause death in HF, preventing a dose-response meta-analysis. Finally, we were unable to explore all the variables that influence TMAO levels, such as diet, environment, and medications.

Conclusion

The present study demonstrated that TMAO may be a risk factor for poor prognosis in patients with HF, but the underlying mechanism requires further investigation. Developing therapies that directly target TMAO may constitute a novel approach for preventing and treating HF.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author Contributions

XL and YL: conceptualization and writing original draft. YW and QL: project administration. ZF, JC, and DL: data management and data analysis. JL, XC, and LX: methodology and software application. XL, YL, and QL: writing review and editing and funding acquisition. All authors read and approved the final manuscript, critically reviewed the literature, and contributed to drafting the manuscript.

Funding

This work was supported by the Capital's Funds for Health Improvement and Research (Grant No. 2020-2-4201), National Natural Science Foundation of China (81973622), and Beijing University of Chinese Medicine 1166 Development Program for Junior Scientists (No. 030903010331).

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.