You are here

  • Home
  • Online First
  • Cardiorespiratory fitness, body mass index and mortality: a systematic review and meta-analysis

Article Text

Article menu
Systematic review
Cardiorespiratory fitness, body mass index and mortality: a systematic review and meta-analysis
  1. Nathan R Weeldreyer1,
  2. Jeison C De Guzman1,
  3. Craig Paterson2,
  4. Jason D Allen1,
  5. Glenn A Gaesser3,
  6. Siddhartha S Angadi1
  1. 1Department of Kinesiology, School of Education and Human Development, University of Virginia, Charlottesville, Virginia, USA
  2. 2Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK
  3. 3College of Health Solutions, Arizona State University, Phoenix, Arizona, USA
  1. Correspondence to Dr Siddhartha S Angadi; ssa2w{at}virginia.edu

Abstract

Objective The purpose of this review was to assess the joint relationship of cardiorespiratory fitness (CRF) and Body Mass Index (BMI) on both cardiovascular disease (CVD) and all-cause mortality risk.

Design A systematic review and meta-analysis was conducted. Pooled HR and 95% CI were calculated using a three-level restricted maximum likelihood estimation random-effects model with robust variance estimation. The reference group was normal weight-fit and was compared with normal weight-unfit, overweight-unfit and fit, and obese-unfit and fit.

Data sources Electronic databases (PubMed/MEDLINE, Web of Science and SportDiscus) were searched following registration on PROSPERO.

Eligibility criteria Articles meeting the following criteria were included: (1) published between January 1980 and February 2023, (2) prospective cohort study, (3) CRF assessed using a maximal or VO2peak exercise test, (4) BMI reported and directly measured, (5) joint impact of CRF and BMI on all-cause mortality or CVD mortality were analysed, and (6) the reference group was normal weight, fit individuals.

Results 20 articles were included in the analysis resulting in a total of 398 716 observations. Compared with the reference group, overweight-fit (CVD HR (95% CI): 1.50 (0.82–2.76), all-cause HR: 0.96 (0.61–1.50)) and obese-fit (CVD: 1.62 (0.87–3.01), all-cause: 1.11 (0.88–1.40)) did not have a statistically different risk of mortality. Normal weight-unfit (CVD: 2.04 (1.32–3.14), all-cause: 1.92 (1.43–2.57)), overweight-unfit (CVD: 2.58 (1.48–4.52), all-cause: 1.82 (1.47–2.24)) and obese-unfit (CVD: 3.35 (1.17–9.61), all-cause: 2.04 (1.54–2.71)) demonstrated 2–3-fold greater mortality risks.

Conclusions CRF is a strong predictor of CVD and all-cause mortality and attenuates risks associated with overweight and obesity. These data have implications for public health and risk mitigation strategies.

  • Exercise
  • VO2peak
  • Risk factor
  • Cardiovascular Diseases
  • Health promotion

Data availability statement

Data are available upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bjsports-2024-108748

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN

    • Overweight and obesity are associated with an increased risk of all-cause and cardiovascular disease (CVD) mortality.

    • The rates of overweight and obesity have continued to increase globally despite public health guidance to lose weight via dieting.

    WHAT ARE THE NEW FINDINGS

    • Our review included 20 studies with meta-analyses performed on 398 716 observations assessing the joint relationship between Body Mass Index (BMI) and cardiorespiratory fitness (CRF) on mortality risks.

    • Our analyses found that those classified as fit, regardless of BMI status, showed no statistically significant increase in CVD or all-cause mortality risk compared with normal weight-fit individuals. In contrast, all BMI classifications who were unfit showed twofold to threefold increases in risk of CVD and all-cause mortality compared with normal weight-fit individuals.

    • Our review included a greater proportion of females (33%) as well as a more globally diverse cohort compared with previous studies.

    Introduction

    Obesity is associated with an increased risk of all-cause and cardiovascular disease (CVD) mortality rates.1 The prevalence of obesity has increased significantly over the past four decades.2 Today, roughly two in five adults are classified as overweight or obese globally and this is associated with an estimated economic impact of 2.19% of global gross domestic product.3 The public health strategy has largely focused on weight loss; however, this has a recidivism rate of ~100% at 10 years follow-up.4 5 In addition, intentional weight loss alone has not consistently shown improvements in mortality risk in observational studies or randomised controlled trials.6

    Cardiorespiratory fitness (CRF) has been shown to be inversely related to both all-cause and CVD mortality risk.7–10 Due to this, CRF has been proposed as a vital sign11 but is not part of risk management guidelines in overweight and obese individuals.12 In the past three decades, a multitude of prospective studies have investigated the joint associations of CRF and Body Mass Index (BMI; kg/m2) categories on mortality.13–33 These studies consistently show that CRF has a stronger association with mortality risk than does BMI. In the first meta-analysis on the joint association of CRF and BMI on all-cause mortality, Barry et al34 demonstrated that, compared with normal weight (BMI 18.5 kg/m2) fit (generally >20th percentile of age-adjusted CRF) individuals, overweight (BMI 25.0–<30 kg/m2) and obese (BMI >30 kg/m2) fit individuals had no higher risk of all-cause mortality. A subsequent meta-analysis by Barry et al35 demonstrated that CRF attenuated the CVD mortality risk associated with overweight and obesity but did not eliminate it entirely.35 However, these analyses had majority representation from the Aerobics Centre Longitudinal Study/Cooper Centre Longitudinal Study (ACLS/CCLS) cohorts. These reviews by Barry et al did not adequately control for the potential variance dependence when measuring multiple cohorts from the same study. Using a three-level model accounts for this dependency and gives more conservative estimates. In addition, sunset plots provide observed statistical power and true effect sizes which give novel insight not previously provided. Lastly, these reviews primarily included studies with large proportions of male participants (84%34 and 98.6%35 male).

    Since these studies were published, new observational cohort studies with a greater proportion of female participants and greater global representation have been published.18 21 24 28 31 Therefore, the objective of this review was to perform a meta-analysis to update the existing literature and quantify the joint associations of CRF and BMI with all-cause and CVD mortality. We hypothesised that when CRF and BMI were jointly evaluated, CRF would be associated with all-cause and CVD mortality; whereas, BMI would not.

    Methods

    This meta-analysis is reported in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines36 and pre-registered on PROSPERO (CRD42023392979). Covidence (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia) was used for title, abstract and full-text screening.

    Literature review

    The review of literature was performed through PubMed/MEDLINE, Web of Science and SportDiscus search engines using the following terms: ((‘Cardiorespiratory fitness’ OR ‘physical fitness’ OR ‘fitness’ OR ‘maximal oxygen consumption’ OR ‘VO2max’ OR ‘maximal oxygen uptake’ OR ‘stress test’ OR ‘maximal treadmill test’) AND (‘Body composition’ OR ‘BMI’ OR ‘body mass index’ OR ‘obesity’ OR ‘adiposity’) AND (‘mortality’ OR ‘mortalities’ OR ‘death’ OR ‘fatality’ OR ‘fatal’ OR ‘cardiovascular mortality’ OR ‘chronic disease’ OR ‘cardiovascular’ OR ‘metabolic’ OR ‘cardiorespiratory’ OR ‘all-cause mortality’)) between January 1980 and February 2023. In addition, results were limited to those published in the English language.

    Article selection

    First, the titles and abstracts of the articles were screened for eligibility independently by NRW and JCD. The following criteria were determined a priori for article inclusion: (1) published between January 1980 and February 2023, (2) the design was prospective, (3) CRF was assessed using a maximal or VO2peak exercise test, (4) BMI was reported and directly measured in clinic, (5) joint impact of CRF and BMI on all-cause mortality or CVD mortality was analysed, and (6) reference group was normal weight, fit individuals. Included populations were composed of individuals with CVD, diabetes, renal disease, asthma, hormone replacement therapy, smokers and chronic respiratory diseases. Excluded populations included individuals diagnosed with cancer, liver failure/cirrhosis, psychological or psychiatric problems, substance abuse/dependency, eating disorders, neurological degenerative problems and pregnant females. Full texts were reviewed of the remaining articles to determine eligibility. Title and abstract and full-text review were completed independently by two authors (NRW and JCD). Any conflicts were resolved by a third reviewer (SSA). Reasons for study exclusions can be found in online supplemental table 3).

    Data extraction and quality assessment

    Articles meeting inclusion criteria had the following data extracted and systematically organised: (i) title, author and publication year, study design, start date and end date; (ii) continuous variables: sample size, age, mean follow-up years and number of deaths; (iii) categorical variables: sex, disease status, study database, CRF and BMI (kg/m2); and (iv) HRs and 95% CIs.

    CRF was categorised into two groups (ie, fit and unfit). It was determined, a priori, to use the highest CRF group in each article as ‘fit’ and the lowest CRF group as ‘unfit’. BMI was categorised into three groups (ie, normal weight (<25 kg/m2), overweight (25–29.9 kg/m2) and obese (≥30 kg/m2)). CRF and BMI were combined to form five comparison groups (ie, normal weight-unfit, overweight-unfit, obese-unfit, overweight-fit, obese-fit) and a reference group (normal weight-fit).

    Study quality was assessed independently by two authors (NRW and JCD) using the National Institutes of Health quality assessment for observational cohort and cross-sectional studies which included 14 questions on methodology, implementation, sources of bias, confounding, study power and other factors. Each study was scored as either ‘Good’, ‘Fair’ or ‘Poor’ based on the answers to the signalling questions.

    Equity, diversity, and inclusion statement

    Previous reviews conducted in this area34 35 included limited female representation and a mostly US cohort. Since then, new prospective cohort studies have been published including greater female representation and a more globally diverse population. Thus, our goal was to update the literature with a more diverse and generalisable study sample. In addition, we included sex as a potential moderator in our analyses in order to see if sex has an impact on the relationship between BMI, cardiorespiratory fitness and mortality.

    Statistical analysis

    Data were analysed and visualised using the ‘metafor’ and ‘metaviz’ packages in RStudio, respectively (V.4.2.1).37 38 Data are presented as HR with 95% CI. Pooled HRs were estimated using a three-level restricted maximum likelihood estimation random-effects model with robust variance estimation. As several articles reported data from related databases, the data represents a nested structure. The three-level model accounts for the nested structure and the potential variance dependence between study effect estimates arising from the same database/cluster.39 Additionally, robust variance estimation was used to account for the unknown variance dependence within clusters and thus produce more conservative standard errors for each study effect estimate.40 The HRs of mortality between each subgroup and referent group of each trial were inputted into the model to determine the pooled effect.

    Robustness of the pooled results was examined via sunset funnel plots for small study/cluster effects. Sensitivity analysis was performed using Cook’s distance and studentised residuals to identify potentially influential or outlying trials, respectively, using previously described methods.41 If a study or cluster was identified as being potentially influential or outlying, the robustness of the overall analysis was tested by removing the identified trial/cluster. Statistical heterogeneity of the overall model was assessed with using Cochran’s Q and I2, where <25% indicates low heterogeneity, 25%–75% indicates moderate heterogeneity and >75% indicates considerable heterogeneity.42 Sunset funnel plots were used to visualise the statistical power of each trial within each analysis to detect the summary effect of that analysis.43 Separate analyses were performed comparing the risk of all-cause mortality and CVD mortality of normal weight-fit individuals to normal weight-unfit, overweight-fit, overweight-unfit, obese-fit and obese-unfit individuals. To examine the effect of study characteristics on the risk of all-cause mortality and CVD mortality, multiple moderator analyses were performed on possible confounders (mean age (≥50 years or not), sex (males or not) and chronic disease (yes/no), and mean follow-up duration (≥12 years or not)). Alpha was a priori set to be p<0.05.

    Results

    Literature search

    Figure 1 shows the PRISMA diagram where 2279 articles were retrieved. Following screening of titles and abstracts, 2218 studies were excluded leaving 61 studies assessed for full-text eligibility. Following this assessment, 20 articles remained eligible for the current meta-analysis. The study characteristics are shown in table 1. The 20 included studies resulted in 458 784 observations and had a mean age and follow-up ranging from 42.4–64.4 years and 7.7–26 follow-up years, respectively. In addition, 307 385 observations were male (67%) and 151 399 were female (33%). Due to differing CRF criteria used by each study (see online supplemental table 4), a total of 398 716 observations were ultimately included in our analysis.

    Figure 1
    Figure 1

    PRISMA diagram of study selection. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

    View this table:
    Table 1

    Characteristics of studies included in the meta-analysis

    Study quality and statistical power

    Overall, studies were deemed to have good study quality. Visual inspection of the funnel plots revealed some asymmetry; however, the studies identified as being of lower methodological quality had large variance; thus, they did not carry much weight in the analysis. Sunset plots (online supplemental figure S1 and S2) indicate moderate to high statistical power in the analyses comparing fit and unfit groups, and inferences regarding the statistical power of fit vs fit analyses were not possible due to the null result.

    All-cause mortality risk

    The results indicated that compared with normal weight-fit individuals, overweight-fit (HR, (95% CI): 0.96, (0.61–1.50), p=0.8, I2level 2 = 2.71 %, I2level 3 = 88.6%) and obese-fit (HR: 1.11, (0.88–1.40), p=0.25, I2level 2 = 65.4%, I2level 3 = 0%) had no significant increase in risk of all-cause mortality. However, compared with normal weight-fit individuals, an increased risk of all-cause mortality was observed in normal weight-unfit (HR: 1.92, (1.43–2.57), p=0.003, Q=63.4, df=15, p<0.001, I2level 2 = 59.7%, and I2level 3 = 19.5), overweight-unfit (HR: 1.82, (1.47–2.24), p=0.007, Q=64.3, df=12, p<0.001, I2level 2 = 81.6%, I2level 3 <0.001%) and obese-unfit (HR: 2.04, (1.54–2.71), p=0.003, heterogeneity: Q=57.5, df=14, p<0.001, I2level 2 = 67.8%, I2level 3 = 7.9%) individuals. Figure 2 represents the findings from the meta-analysis comparing the joint association of cardiorespiratory fitness and adiposity on all-cause mortality. Data are presented as forest plots with pooled HRs and HRs with 95% CI for the five subgroups (ie, overweight-fit, obese-fit, normal weight-unfit, overweight-unfit and obese-unfit) compared with the reference group (ie, normal weight-fit). In addition, the weight of each study held in the analyses is indicated within each forest plot.

    Figure 2
    Figure 2

    Forest plots for all-cause mortality. (A.) Overweight-fit. (B.) Obese-fit. (C.) Normal weight-unfit. (D.) Overweight-unfit. (E.) Obese-unfit. (US) Stevens et al, 2004, USA-based cohort; (R) Stevens et al, 2004, Russian-based cohort.

    Statistical heterogeneity was found when analysing each of the subgroups. The cluster of studies which included the ACLS/CCLS cohorts15 16 24 27 30 32 33 44 45 was identified as potentially influential as such a large percentage of our population came from these databases. In addition, studies by Church et al15 and Goel et al19 were often flagged due to their large variance seen in the forest plots. However, systematically removing each potentially influential or outlying trials and cluster did not change the significance found when comparing unfit BMI categories to our referent group (online supplemental table 1).

    Visual inspections of the sunset plots (online supplemental figure S1) for fit-unfit analyses showed no evidence of small-study bias. With regard to power, the estimated median statistical power to detect the summary effect of each fit-unfit analysis ranged from 98%–100%. Visual inspections of the sunset plots for fit-fit meta-analyses showed no evidence of small-study bias. As the fit-fit meta-analyses resulted in a null finding, inferences regarding power are inappropriate and thus, these figures should only be interpreted in the same way as traditional funnel plots.

    Moderator analyses

    We performed moderator analyses to examine the effect of sex, age, chronic disease status and length of follow-up. Within each of the subgroups, there were no significant effects of sex, age, chronic disease status or length of follow-up (all p>0.05). This indicates that the effects of fitness appear to show benefit in all-cause mortality risk for each BMI class regardless of these population characteristics.

    Cardiovascular Disease Mortality Risk

    Similar to our results for all-cause mortality, compared with normal weight-fit individuals, we found no statistically greater risk for CVD mortality in overweight-fit (HR: 1.50, (0.82–2.76), p=0.104) or obese-fit (HR: 1.62, (0.87–3.01), p=0.078) individuals. In contrast, compared with normal weight-fit individuals, normal weight-unfit (HR: 2.04, (1.32–3.14), p=0.013, heterogeneity: Q=16.2, df=12, p=0.18, I2level 2 = 3.1%, I2level 3 = 35.3%), overweight-unfit (HR: 2.58, (1.48–4.52), p=0.024, heterogeneity: Q=45.6, df=10, p<0.001, I2level 2= 75.4%, I2level 3 <0.001) and obese-unfit populations (HR: 3.35, (1.17–9.61), p=0.038, heterogeneity: Q=40.8, df=9, p<0.001, I2level 2 = 44.5%, I2level 3 = 36.7%) showed significantly increased CVD risk. Forest plots depicting these HRs for each subgroup can be seen in figure 3.

    Figure 3
    Figure 3

    Forest plots for cardiovascular mortality. (A.) Overweight-fit. (B.) Obese-fit. (C.) Normal weight-unfit. (D.) Overweight-unfit. (E.) Obese-unfit. (US) Stevens et al, 2004, USA-based cohort; (R) Stevens et al, 2004, Russian-based cohort.

    Significant heterogeneity was observed in our obese-unfit vs normal weight-fit model. The study by Stevens et al29 was identified as potentially influential. When removed from the model, our results were no longer significant (HR: 3.99, (0.72–22.2), p=0.06, heterogeneity: Q=22.2, df=7, p=0.002, I2level 2 = 58.0%, I2level 3 = 12.4%). In addition, after removal of the largest ACLS cohort,17 the model was no longer significant (HR: 3.21, (0.98–10.6), p=0.052, heterogeneity: Q=35.5, df=8, p<0.001, I2level 2 = 47.3%, I2level 3 = 32.5%). All other models remained significant after removing all potentially influential clusters (online supplemental table 2).

    Inspection of the sunset plots (online supplemental S2) showed similar results to those for all-cause mortality risk. Fit-unfit analyses showed no evidence of small-study bias. With regard to power, the estimated median statistical power to detect the summary effect of each fit-unfit analysis ranged from 85.4%–99.9%. Visual inspections of the sunset plots for fit-fit meta-analyses showed no evidence of small-study bias. As the fit-fit meta-analyses resulted in a null finding, inferences regarding power are inappropriate and thus, these figures should only be interpreted in the same way as traditional funnel plots.

    Moderator analyses

    We performed moderator analyses to examine the effect of sex, age, chronic disease status and length of follow-up on each subgroup. When examining the overweight-unfit group to normal weight-fit individuals, there was a significant effect of chronic disease status (F=38.8, p=0.008). Compared with normal weight-fit individuals with chronic disease, overweight-unfit individuals without chronic disease had higher CVD mortality (HR: 2.81, (1.29, 6.12)) than did overweight-unfit individuals with chronic disease (HR: 1.84, (1.32, 2.56)). It should be noted that all data points for individuals with chronic disease came from cohort 1 (ACLS/CCLS); whereas, the remaining data points arose from four separate cohorts. Consequently, it is unclear whether this result represents a true effect or is the result of selection bias and as such reflects a limitation of the existing literature in this area. In addition, a significant effect was seen with length of follow-up (F=427.1, p=0.019) such that increased HRs were observed in shorter follow-up studies (HR: 4.49, (2.07, 9.71)) but not in longer follow-up studies (HR: 2.17, (1.28, 3.51)).

    Discussion

    This systematic review and meta-analysis aimed to examine the association of CRF and BMI on all-cause and CVD mortality in males and females using meta-analytical methodology. The major finding of the present meta-analysis was that once CRF was accounted for, there were no significant increases in all-cause or CVD mortality risk for overweight or obese individuals. Importantly, these data extend prior findings to a pooled cohort with greater representation of females (33%).

    Our findings demonstrate that individuals with higher CRF who are overweight or obese are not at a higher risk for all-cause or CVD mortality when compared with normal weight-fit individuals. Importantly, individuals who were unfit had a ~2-fold increase in the risk of all-cause mortality and a ~2–3-fold increase in the risk of CVD mortality. It is important to note that a majority of these studies demonstrated that individuals only needed to exceed the CRF of the study population 20th percentile in order to be considered fit, which suggests that significant reductions in mortality risk may be attained with moderate levels of age-adjusted CRF regardless of BMI status.

    Our findings are in agreement with a previous meta-analysis on all-cause mortality34 but not for CVD mortality.35 Barry et al35 demonstrated statistically significant 25% and 42% higher risks of CVD mortality among overweight-fit and obese-fit, respectively. Although our meta-analysis indicated no statistically significantly higher CVD mortality risk for overweight-fit and obese-fit groups, the HRs were actually greater than those reported by Barry et al. This suggests that CRF may substantially attenuate, but not entirely eliminate, the CVD mortality associated with elevated BMI. The reasons for this are unclear but may be related to the association between obesity and CVD risk factors and type 2 diabetes, which increases the risk of CVD. It is also possible that higher levels of CRF are necessary to further attenuate CVD mortality risk associated with high BMI.

    In addition to the general models, we performed moderator analyses to assess the impact that sex, age, chronic disease status and follow-up duration had on both CVD and all-cause mortality. We found that these moderators had no significant effect on any of our all-cause mortality models. When examining CVD mortality risk, we did find significant effects of both chronic disease status and length of follow-up for the overweight-unfit comparison. However, it is worth noting that all participants with a chronic disease were sampled from the same cohort; whereas, participants without chronic disease were sampled from four separate cohorts. Consequently, it is unclear whether this result represents a true effect or is the result of selection bias. In addition, length of follow-up (≥12-year follow-up or not) showed that those who were overweight-unfit showed increased CVD mortality risks for shorter follow-up duration than those with longer follow-ups. This may indicate that fitness plays a role in short-term outcomes, but given long enough time spans, CRF is less protective due to general ageing.

    While similar results to our review have been previously reported, they are not without significant limitations. In two meta-analyses conducted by Barry et al34 35 looking at all-cause and CVD mortality risk, 8 of 1034 and 7 of 935 of the included studies were from the ACLS/CCLS cohorts. Though they performed sensitivity analyses, they could not account for overlapping data points between studies. Because they did not control for this nested structure with appropriate statistical methods, this may have led to overly confident effect estimates in their analysis. Furthermore, their studies included predominantly males (84%34 and 98.6%35) and since their publication, several large cohort studies have been published that allow us to broaden the pool of participants.

    Our review aimed to address these limitations. First, the present review included the addition of six cohorts outside of ACLS/CCLS, increasing the sample population by ~260 000 individuals. Furthermore, these cohorts had a greater representation of females. As a result, the percentage of males decreased to 67% of the total cohort sample, making our results more generalisable as it better represents sex demographics. To account for the likely dependency between data points coming from the same database, we used a three-level model rather than the traditional two-level model used for meta-analyses. A standard two-level random-effects meta-analysis accounts for only between study and sampling variance; whereas, a three-level model can account for within-study variance as well. Barry et al only removed the two largest ACLS samples during their sensitivity analysis.34 Our data showed that in adjusting for the sample clusters, all models except for the obese-unfit analysis for CVD mortality were not significantly different from the overall effect.

    Our results show that CRF remains a potent predictor of mortality risk independent of BMI. These data are physiologically plausible because exercise training results in weight loss independent improvements in markers associated with mortality risk such as glycaemia,46 47 insulin sensitivity,48–50 cardiovascular function,6 51 inflammation52 and ectopic fat deposition.50 53 In addition, studies from the ACLS/CCLS cohorts (which represent the majority of trials included) used estimated CRF via Bruce protocol treadmill time. When comparing obese and normal weight individuals in this way, their relative VO2peaks (expressed as mL/kg/min) will be the same. However, VO2peak relative to fat-free mass (FFM) (expressed as mL/kg FFM/min) will inherently be greater due to the increased fat mass they carry. This may indicate a more ‘fit’ or metabolically healthy muscle in obese-fit subjects and lead to improved health outcomes via improved muscle signalling.54 In addition, Imboden et al55 recently showed that VO2peakFFM was a stronger predictor of mortality than relative VO2peak. This highlights the need to include measures of body composition when assessing relationships between CRF and mortality risk.55

    Results from this study have implications for public health guidelines. While increased risks of morbidity and mortality are associated with increased BMI,1 weight-centric interventions (interventions primarily concerned with weight loss, typically via calorie restriction) are largely unsuccessful at maintaining long-term weight reduction and thus improved health outcomes.56 Thus, a CRF-centric approach to treating obesity-related health conditions, in which the major focus is on increasing physical activity to improve CRF rather than a specific weight loss target, may improve health outcomes while avoiding pitfalls associated with repeated weight loss attempts.6 57 We do not think weight loss attempts should be discouraged but recognise that this may not be a feasible goal in all adults.58 This is further reinforced by data from the Diabetes Prevention Programme that demonstrated that adults were more likely to achieve physical activity targets as opposed to weight loss targets.59

    This meta-analysis is not without limitations. First, we limited our search to manuscripts published within the English language and this may lead to missed publications in our search strategy. In addition, a limitation of BMI is that it is not a true measure of body composition although, it is strongly correlated to per cent body fat.60 To our knowledge, only Lee et al23 and McAuley et al27 assessed body composition via body fat% obtained by skin callipers and/or hydrostatic weighing. All other studies to date have looked at the joint relationship using BMI as the measure of body fatness. However, similar results to ours held up when Lee et al examined fitness and fat mass, fat-free mass, and waist circumferences.23 To move the field forward, future studies should investigate other measures of adiposity such as non-invasive imaging of visceral adiposity. Another limitation is the dichotomous criteria used when considering CRF levels. It is well known that the reduction in mortality risk with increasing CRF and physical activity levels is not linear.61 62 As such, the varied criteria employed by the included studies may influence the results. However, the largest risk reduction with physical activity and CRF is seen when moving from the least fit category to the moderately fit category, at which point, risk reduction tapers with increased CRF.62 While greater female representation was achieved in our analysis, 67% of the total population was male. This is due to the large representation of ACLS/CCLS cohorts. We are also limited in that studies did not include the full breakdown of females in each BMI/CRF groups. As such, 67% of the 458 784 total sample size were male. We cannot say explicitly how many of the 398 716 included in our analysis were male or female.

    In addition, the majority of these cohorts largely included those from higher socioeconomic statuses, included predominantly Caucasian individuals and were US residents. Future studies should seek to investigate this relationship in disadvantaged populations and different ethnic groups. The lack of data from Africa, the Indian subcontinent and China excludes a significant proportion of the world’s population. In addition, patterns of adiposity differ between ethnic groups63 and may result in different relationships than those found in the current study.

    In conclusion, our results demonstrate a significant reduction in both CVD and all-cause mortality risk in fit compared with unfit individuals regardless of BMI. These results add to a growing body of evidence as to the value of CRF as a predictor of mortality. Given the large reductions in mortality risk with CRF independent of BMI in the present study, future RCTs could disentangle this issue by comparing a CRF-only approach to a CRF+caloric deficit group and an attention-control group.

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    References

    1. ↵
      1. Poirier P,
      2. Cornier M-A,
      3. Mazzone T, et al
      . Bariatric surgery and cardiovascular risk factors: a scientific statement from the American Heart Association. Circulation 2011;123:1683–701. doi:10.1161/CIR.0b013e3182149099
      OpenUrlFREE Full Text
    2. ↵
      1. Craig M H,
      2. Fryar CD,
      3. Ogden CL
      . Prevalence of obesity and severe obesity among adults. CDC.gov2020; 2017. Available: https://www.cdc.gov/nchs/products/databriefs/db360.htm
    3. ↵
      1. Okunogbe A,
      2. Nugent R,
      3. Spencer G, et al
      . Economic impacts of overweight and obesity: current and future estimates for 161 countries. BMJ Glob Health 2022;7:e009773. doi:10.1136/bmjgh-2022-009773
    4. ↵
      1. Grattan BJ,
      2. Connolly-Schoonen J
      . Addressing weight loss recidivism: a clinical focus on metabolic rate and the psychological aspects of obesity. ISRN Obes 2012;2012:567530. doi:10.5402/2012/567530
    5. ↵
      1. Johnson D,
      2. Drenick EJ
      . Therapeutic fasting in morbid obesity. Arch Intern Med 1977;137:1381–2.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Gaesser GA,
      2. Angadi SS
      . Obesity treatment: Weight loss versus increasing fitness and physical activity for reducing health risks. iScience 2021;24:102995. doi:10.1016/j.isci.2021.102995
    7. ↵
      1. Blair SN,
      2. Kohl HW,
      3. Barlow CE, et al
      . Changes in physical fitness and all-cause mortality. A prospective study of healthy and unhealthy men. JAMA 1995;273:1093–8. doi:10.1001/jama.1995.03520380029031
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Kokkinos P,
      2. Myers J,
      3. Kokkinos JP, et al
      . Exercise capacity and mortality in black and white men. Circulation 2008;117:614–22. doi:10.1161/CIRCULATIONAHA.107.734764
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Lang JJ,
      2. Prince SA,
      3. Merucci K, et al
      . Cardiorespiratory fitness is a strong and consistent predictor of morbidity and mortality among adults: an overview of meta-analyses representing over 20.9 million observations from 199 unique cohort studies. Br J Sports Med 2024;58:556–66. doi:10.1136/bjsports-2023-107849
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Mandsager K,
      2. Harb S,
      3. Cremer P, et al
      . Association of Cardiorespiratory Fitness With Long-term Mortality Among Adults Undergoing Exercise Treadmill Testing. JAMA Netw Open 2018;1:e183605. doi:10.1001/jamanetworkopen.2018.3605
    11. ↵
      1. Ross R,
      2. Blair SN,
      3. Arena R, et al
      . Importance of Assessing Cardiorespiratory Fitness in Clinical Practice: A Case for Fitness as a Clinical Vital Sign: A Scientific Statement From the American Heart Association. Circulation 2016;134:e653–99. doi:10.1161/CIR.0000000000000461
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Jensen MD,
      2. Ryan DH,
      3. Apovian CM, et al
      . 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. Circulation 2014;129:S102–38. doi:10.1161/01.cir.0000437739.71477.ee
      OpenUrlFREE Full Text
    13. ↵
      1. Woo J,
      2. Yu R,
      3. Yau F
      . Fitness, fatness and survival in elderly populations. Age (Dordr) 2013;35:973–84. doi:10.1007/s11357-012-9398-6
      OpenUrlCrossRefPubMed
    14. ↵
      1. Church TS,
      2. LaMonte MJ,
      3. Barlow CE, et al
      . Cardiorespiratory fitness and body mass index as predictors of cardiovascular disease mortality among men with diabetes. Arch Intern Med 2005;165:2114–20. doi:10.1001/archinte.165.18.2114
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Church TS,
      2. Cheng YJ,
      3. Earnest CP, et al
      . Exercise capacity and body composition as predictors of mortality among men with diabetes. Diabetes Care 2004;27:83–8. doi:10.2337/diacare.27.1.83
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Farrell SW,
      2. Fitzgerald SJ,
      3. McAuley PA, et al
      . Cardiorespiratory fitness, adiposity, and all-cause mortality in women. Med Sci Sports Exerc 2010;42:2006–12. doi:10.1249/MSS.0b013e3181df12bf
      OpenUrlCrossRefPubMed
    17. ↵
      1. Farrell SW,
      2. Finley CE,
      3. Radford NB, et al
      . Cardiorespiratory fitness, body mass index, and heart failure mortality in men: Cooper Center Longitudinal Study. Circ Heart Fail 2013;6:898–905. doi:10.1161/CIRCHEARTFAILURE.112.000088
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Farrell SW,
      2. Barlow CE,
      3. Willis BL, et al
      . Cardiorespiratory Fitness, Different Measures of Adiposity, and Cardiovascular Disease Mortality Risk in Women. J Womens Health (Larchmt) 2020;29:319–26. doi:10.1089/jwh.2019.7793
      OpenUrlCrossRefPubMed
    19. ↵
      1. Goel K,
      2. Thomas RJ,
      3. Squires RW, et al
      . Combined effect of cardiorespiratory fitness and adiposity on mortality in patients with coronary artery disease. Am Heart J 2011;161:590–7. doi:10.1016/j.ahj.2010.12.012
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Heir T,
      2. Erikssen J,
      3. Sandvik L
      . Overweight as predictor of long-term mortality among healthy, middle-aged men: a prospective cohort study. Prev Med 2011;52:223–6. doi:10.1016/j.ypmed.2011.01.010
      OpenUrlPubMed
    21. ↵
      1. Hemmingsson E,
      2. Väisänen D,
      3. Andersson G, et al
      . Combinations of BMI and cardiorespiratory fitness categories: trends between 1995 and 2020 and associations with CVD incidence and mortality and all-cause mortality in 471 216 adults. Eur J Prev Cardiol 2022;29:959–67. doi:10.1093/eurjpc/zwab169
      OpenUrlCrossRefPubMed
    22. ↵
      1. Jae SY,
      2. Franklin BA,
      3. Kurl S, et al
      . Effect of Cardiorespiratory Fitness on Risk of Sudden Cardiac Death in Overweight/Obese Men Aged 42 to 60 Years. Am J Cardiol 2018;122:775–9. doi:10.1016/j.amjcard.2018.05.017
      OpenUrlCrossRefPubMed
    23. ↵
      1. Lee CD,
      2. Blair SN,
      3. Jackson AS
      . Cardiorespiratory fitness, body composition, and all-cause and cardiovascular disease mortality in men. Am J Clin Nutr 1999;69:373–80. doi:10.1093/ajcn/69.3.373
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. McAuley PA,
      2. Artero EG,
      3. Sui X, et al
      . Fitness, fatness, and survival in adults with prediabetes. Diabetes Care 2014;37:529–36. doi:10.2337/dc13-1347
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. McAuley PA,
      2. Sui X,
      3. Church TS, et al
      . The joint effects of cardiorespiratory fitness and adiposity on mortality risk in men with hypertension. Am J Hypertens 2009;22:1062–9. doi:10.1038/ajh.2009.122
      OpenUrlCrossRefPubMed
    26. ↵
      1. McAuley PA,
      2. Kokkinos PF,
      3. Oliveira RB, et al
      . Obesity paradox and cardiorespiratory fitness in 12,417 male veterans aged 40 to 70 years. Mayo Clin Proc 2010;85:115–21. doi:10.4065/mcp.2009.0562
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. McAuley PA,
      2. Artero EG,
      3. Sui X, et al
      . The obesity paradox, cardiorespiratory fitness, and coronary heart disease. Mayo Clin Proc 2012;87:443–51. doi:10.1016/j.mayocp.2012.01.013
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Rozanski A,
      2. Gransar H,
      3. Hayes SW, et al
      . Synergistic Assessment of Mortality Risk According to Body Mass Index and Exercise Ability and Capacity in Patients Referred for Radionuclide Stress Testing. Mayo Clin Proc 2021;96:3001–11. doi:10.1016/j.mayocp.2021.05.021
      OpenUrlCrossRefPubMed
    29. ↵
      1. Stevens J,
      2. Evenson KR,
      3. Thomas O, et al
      . Associations of fitness and fatness with mortality in Russian and American men in the lipids research clinics study. Int J Obes Relat Metab Disord 2004;28:1463–70. doi:10.1038/sj.ijo.0802770
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Sui X,
      2. LaMonte MJ,
      3. Laditka JN, et al
      . Cardiorespiratory fitness and adiposity as mortality predictors in older adults. JAMA 2007;298:2507–16. doi:10.1001/jama.298.21.2507
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Tarp J,
      2. Grøntved A,
      3. Sanchez-Lastra MA, et al
      . Fitness, Fatness, and Mortality in Men and Women From the UK Biobank: Prospective Cohort Study. J Am Heart Assoc 2021;10:e019605. doi:10.1161/JAHA.120.019605
    32. ↵
      1. Wei M,
      2. Gibbons LW,
      3. Kampert JB, et al
      . Low cardiorespiratory fitness and physical inactivity as predictors of mortality in men with type 2 diabetes. Ann Intern Med 2000;132:605–11. doi:10.7326/0003-4819-132-8-200004180-00002
      OpenUrlPubMedWeb of Science
    33. ↵
      1. Wei M,
      2. Kampert JB,
      3. Barlow CE, et al
      . Relationship between low cardiorespiratory fitness and mortality in normal-weight, overweight, and obese men. JAMA 1999;282:1547–53. doi:10.1001/jama.282.16.1547
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Barry VW,
      2. Baruth M,
      3. Beets MW, et al
      . Fitness vs. fatness on all-cause mortality: a meta-analysis. Prog Cardiovasc Dis 2014;56:382–90. doi:10.1016/j.pcad.2013.09.002
      OpenUrlCrossRefPubMed
    35. ↵
      1. Barry VW,
      2. Caputo JL,
      3. Kang M
      . The Joint Association of Fitness and Fatness on Cardiovascular Disease Mortality: A Meta-Analysis. Prog Cardiovasc Dis 2018;61:136–41. doi:10.1016/j.pcad.2018.07.004
      OpenUrlCrossRefPubMed
    36. ↵
      1. Page MJ,
      2. McKenzie JE,
      3. Bossuyt PM, et al
      . The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:71:n71. doi:10.1136/bmj.n71
      OpenUrlCrossRefPubMed
    37. ↵
      1. Viechtbauer W
      . Conducting Meta-Analyses in R with the metafor Package. J Stat Softw 2010;36:48. doi:10.18637/jss.v036.i03
    38. ↵
      1. Wagner J,
      2. Chelaru F,
      3. Kancherla J, et al
      . Metaviz: interactive statistical and visual analysis of metagenomic data. Nucleic Acids Res 2018;46:2777–87. doi:10.1093/nar/gky136
      OpenUrlCrossRefPubMed
    39. ↵
      1. Assink M,
      2. Wibbelink CJM
      . Fitting three-level meta-analytic models in R: A step-by-step tutorial. TQMP 2016;12:154–74. doi:10.20982/tqmp.12.3.p154
      OpenUrlCrossRef
    40. ↵
      1. Pustejovsky JE,
      2. Tipton E
      . Meta-analysis with Robust Variance Estimation: Expanding the Range of Working Models. Prev Sci 2022;23:425–38. doi:10.1007/s11121-021-01246-3
      OpenUrlCrossRefPubMed
    41. ↵
      1. Viechtbauer W,
      2. Cheung MW-L
      . Outlier and influence diagnostics for meta-analysis. Res Synth Methods 2010;1:112–25. doi:10.1002/jrsm.11
      OpenUrlCrossRefPubMed
    42. ↵
      1. Higgins JPT,
      2. Thompson SG,
      3. Deeks JJ, et al
      . Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. doi:10.1136/bmj.327.7414.557
      OpenUrlFREE Full Text
    43. ↵
      1. Kossmeier M,
      2. Tran US,
      3. Voracek M
      . Power-Enhanced Funnel Plots for Meta-Analysis. Z Psychol 2020;228:43–9. doi:10.1027/2151-2604/a000392
      OpenUrl
    44. ↵
      1. Lee CD,
      2. Jackson AS,
      3. Blair SN
      . US weight guidelines: is it also important to consider cardiorespiratory fitness? Int J Obes Relat Metab Disord 1998;22 Suppl 2:S2–7.
      OpenUrl
    45. ↵
      1. McAuley P,
      2. Pittsley J,
      3. Myers J, et al
      . Fitness and fatness as mortality predictors in healthy older men: the veterans exercise testing study. J Gerontol A Biol Sci Med Sci 2009;64:695–9. doi:10.1093/gerona/gln039
      OpenUrlPubMedWeb of Science
    46. ↵
      1. Boulé NG,
      2. Haddad E,
      3. Kenny GP, et al
      . Effects of exercise on glycemic control and body mass in type 2 diabetes mellitus: a meta-analysis of controlled clinical trials. JAMA 2001;286:1218–27. doi:10.1001/jama.286.10.1218
      OpenUrlCrossRefPubMedWeb of Science
    47. ↵
      1. Snowling NJ,
      2. Hopkins WG
      . Effects of different modes of exercise training on glucose control and risk factors for complications in type 2 diabetic patients: a meta-analysis. Diabetes Care 2006;29:2518–27. doi:10.2337/dc06-1317
      OpenUrlAbstract/FREE Full Text
    48. ↵
      1. Dollet L,
      2. Zierath JR
      . Interplay between diet, exercise and the molecular circadian clock in orchestrating metabolic adaptations of adipose tissue. J Physiol 2019;597:1439–50. doi:10.1113/JP276488
      OpenUrlCrossRefPubMed
    49. ↵
      1. Dewal RS,
      2. Stanford KI
      . Effects of exercise on brown and beige adipocytes. Biochim Biophys Acta Mol Cell Biol Lipids 2019;1864:71–8. doi:10.1016/j.bbalip.2018.04.013
      OpenUrlPubMed
    50. ↵
      1. Fabbrini E,
      2. Magkos F,
      3. Mohammed BS, et al
      . Intrahepatic fat, not visceral fat, is linked with metabolic complications of obesity. Proc Natl Acad Sci U S A 2009;106:15430–5. doi:10.1073/pnas.0904944106
      OpenUrlAbstract/FREE Full Text
    51. ↵
      1. Early KS,
      2. Stewart A,
      3. Johannsen N, et al
      . The Effects of Exercise Training on Brachial Artery Flow-Mediated Dilation. J Cardiopulm Rehabil Prev 2017;37:77–89. doi:10.1097/HCR.0000000000000206
      OpenUrlCrossRefPubMed
    52. ↵
      1. Gaesser GA,
      2. Angadi SS,
      3. Ryan DM, et al
      . Lifestyle Measures to Reduce Inflammation. Am J Lifestyle Med 2012;6:4–13. doi:10.1177/1559827611411646
      OpenUrlCrossRef
    53. ↵
      1. Hens W,
      2. Taeyman J,
      3. Cornelis J, et al
      . The Effect of Lifestyle Interventions on Excess Ectopic Fat Deposition Measured by Noninvasive Techniques in Overweight and Obese Adults: A Systematic Review and Meta-Analysis. J Phys Act Health 2016;13:671–94. doi:10.1123/jpah.2015-0560
      OpenUrlCrossRefPubMed
    54. ↵
      1. Fiuza-Luces C,
      2. Garatachea N,
      3. Berger NA, et al
      . Exercise is the real polypill. Physiol (Bethesda) 2013;28:330–58. doi:10.1152/physiol.00019.2013
      OpenUrlCrossRef
    55. ↵
      1. Imboden MT,
      2. Kaminsky LA,
      3. Peterman JE, et al
      . Cardiorespiratory Fitness Normalized to Fat-Free Mass and Mortality Risk. Med Sci Sports Exerc 2020;52:1532–7. doi:10.1249/MSS.0000000000002289
      OpenUrlCrossRef
    56. ↵
      1. Dansinger ML,
      2. Tatsioni A,
      3. Wong JB, et al
      . Meta-analysis: the effect of dietary counseling for weight loss. Ann Intern Med 2007;147:41–50. doi:10.7326/0003-4819-147-1-200707030-00007
      OpenUrlCrossRefPubMedWeb of Science
    57. ↵
      1. Gaesser GA
      . Type 2 Diabetes Incidence and Mortality: Associations with Physical Activity, Fitness, Weight Loss, and Weight Cycling. Rev Cardiovasc Med 2022;23:364. doi:10.31083/j.rcm2311364
      OpenUrlCrossRefPubMed
    58. ↵
      1. Fildes A,
      2. Charlton J,
      3. Rudisill C, et al
      . Probability of an Obese Person Attaining Normal Body Weight: Cohort Study Using Electronic Health Records. Am J Public Health 2015;105:e54–9. doi:10.2105/AJPH.2015.302773
      OpenUrlCrossRefPubMed
    59. ↵
      1. Knowler WC,
      2. Barrett-Connor E,
      3. Fowler SE, et al
      . Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393–403. doi:10.1056/NEJMoa012512
      OpenUrlCrossRefPubMedWeb of Science
    60. ↵
      1. Garrow JS,
      2. Webster J
      . Quetelet’s index (W/H2) as a measure of fatness. Int J Obes 1985;9:147–53.
      OpenUrlPubMedWeb of Science
    61. ↵
      1. Paluch AE,
      2. Bajpai S,
      3. Bassett DR, et al
      . Daily steps and all-cause mortality: a meta-analysis of 15 international cohorts. Lancet Public Health 2022;7:e219–28. doi:10.1016/S2468-2667(21)00302-9
      OpenUrlCrossRef
    62. ↵
      1. Kokkinos P,
      2. Faselis C,
      3. Samuel IBH, et al
      . Cardiorespiratory Fitness and Mortality Risk Across the Spectra of Age, Race, and Sex. J Am Coll Cardiol 2022;80:598–609. doi:10.1016/j.jacc.2022.05.031
      OpenUrlCrossRefPubMed
    63. ↵
      1. Yaghootkar H,
      2. Whitcher B,
      3. Bell JD, et al
      . Ethnic differences in adiposity and diabetes risk - insights from genetic studies. J Intern Med 2020;288:271–83. doi:10.1111/joim.13082
      OpenUrlCrossRefPubMed

    Supplementary materials

    Footnotes

    • NRW and JCDG contributed equally.

    • Contributors NRW, JCD and SSA conceived the project. NRW and JCD screened and extracted all data. CP conducted the statistical analysis. JDA and GAG provided intellectual input. All authors contributed to the manuscript and approved the final version. SSA is the guarantor.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.