Associations between Objectively Determined Physical Activity and Cardiometabolic Health in Adult Women: A Systematic Review and Meta-Analysis

Simple Summary This systematic review and meta-analysis aimed to investigate the association between objectively measured physical activity and cardiometabolic health in adult women. After searching four databases (PubMed, Web of Science, Scopus, and the Cochrane library), 23 eligible studies were included (n = 2105). An accelerometer or pedometer determined physical activities (daily steps, total physical activity, minutes engaged in physical activities at different intensities, and the number of physical activity bouts) and cardiometabolic health indicators (blood pressure, lipids, carbohydrate metabolism, insulin, inflammation markers, and metabolic syndrome) were examined in adult women. Overall, it is compelling that being more physically active has favorable effects on the metabolic syndrome. However, the majority of individual cardiometabolic biomarkers hardly improved following increases in physical activity, with the exception that moderate-intensity physical activity appeared to have a more potent effect on high-density lipoprotein. Although higher-intensity physical activity is more effective for women, it is most important to increase the total volume of physical activity. Meanwhile, strategies to improve body composition and cardiorespiratory fitness are required, since these play an important role in mediating the association between physical activity and cardiometabolic health in women. Abstract The purpose of this systematic review and meta-analysis was to qualitatively synthesize and quantitatively assess the evidence of the relationship between objectively determined volumes of physical activity (PA) and cardiometabolic health in women. Four databases (PubMed, Web of Science, Scopus, and the Cochrane library) were searched and, finally, 24 eligible studies were included, with a total of 2105 women from eight countries. A correlational meta-analysis shows that moderate-to-vigorous intensity physical activity (MVPA) was favorably associated with high-density lipoprotein (r = 0.16; 95% CI: 0.06, 0.25; p = 0.002); however, there was limited evidence for the effects of most of the other cardiometabolic biomarkers recorded from steps, total physical activity, light- and moderate-intensity physical activity and MVPA. It is most compelling and consistent that being more physically active is beneficial to the metabolic syndrome. Overall, PA levels are low in adult women, suggesting that increasing the total volume of PA is more important than emphasizing the intensity and duration of PA. The findings also indicate that, according to the confounding effects of body composition and cardiorespiratory fitness, meeting the minimal level of 150 min of moderate-intensity physical activity recommended is not enough to obtain a significant improvement in cardiometabolic indicators. Nonetheless, the high heterogeneity between studies inhibits robust conclusions.

Four electronic databases, including PubMed, Web of Science, Scopus, and the Cochrane Library, were searched from 1 January 1990 to 31 January 2022 in accordance with the search strategy developed by two researchers with expertise in systematic reviews. Firstly, keywords such as "accelerometer", "pedometer", "objectively", "physical activity", and terms of CVD biomarkers were applied to titles and abstracts since there was no standardized keyword to fully capture the studies, including women-specified associations. Secondly, we conducted a manual search to screen the full text for eligible studies. Thirdly, the reference list from included studies was manually screened to ensure completeness of records. Finally, search results were all imported in Endnote (Endnote 20, Wintertree Software Inc., Beijing, China). The detailed search strategy is provided in Appendix A.

Data Extraction
For each included study, descriptive data, intervention, and correlational findings were independently extracted by two reviewers (Yining Lu and Qiaojun Wang) and inputted into Excel (Microsoft Corp, Redmond, WA, USA). Any disagreements were resolved through discussion and all results were checked by a third reviewer (Shanshan Ying). The relationship between PA and cardiometabolic health outcomes was included if it was measured by t-test/Mann-Whitney U-test (U-test)/Kolmogorov-Smirnov test (K-S test), analysis of variance (ANOVA), correlation, regression, and relative risks.

Risk of Bias and Quality Assessment
The Newcastle-Ottawa Scale (NOS) was used to assess the risk of bias in nonrandomized studies (non-randomized interventions and observational studies) [24]. For the Biology 2022, 11, 925 4 of 35 comparability domain, we considered age to be the most important confounder. The maximum number of stars was seven for the cross-sectional design and nine for the longitudinal design. High quality was defined as four or more stars in cross-sectional designs, and five or more in longitudinal designs. Those below the cut-off point were defined as low quality [25]. The Cochrane collaboration tool was used to assess the risk of bias for random experiments [26].
The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) was used to evaluate the quality of evidence for each category of biomarkers [27].
Each included study was independently rated by two reviewers (Yining Lu and Shanshan Ying). Any discrepancy in rating was resolved by discussion and the results were checked by a third reviewer. Details on the risk of bias and quality assessments are presented in Supplementary Table S1.

Statistical Analysis
Comparable PA exposures included steps, minutes in LPA, MPA, VPA, MVPA and PA bouts, TPA, meeting/not meeting the guideline. Although different cut-off points and definitions were used, LPA, MPA, VPA, MVPA and TPA were defined as reported in the studies [28]. If studies measured physical activity in metabolic-equivalent tasks (METs), we used the cut-points proposed by Ainsworth et al. (2011) (e.g., 1.6-2.9 METs was defined as LPA, 3-5.9 METs as MPA, and ≥6 METs as VPA) [29]. When more than one statistical analysis was used, the following hierarchy was applied: (1) regression, (2) correlation, (3) ANOVA, (4) t-test/U-test/K-S test [25]. When more than one adjusted model was used, the most-adjusted models were applied [30].
Meta-analysis was planned if more than two studies were eligible for comparable PA measures and biomarkers. Fisher's z transformation and Hedge's g was used for correlational and standardized mean differences meta-analysis, respectively [31]. The effect size was classified as low (r = 0.1/SMD = 0.2), moderate (r = 0.3/SMD = 0.5), or high (r = 0.50/SMD = 0.8) according to Cohen's recommendations [32].
The random-effects model was used due to the diversity of the methodologies. To evaluate the impact of heterogeneity on the meta-analysis, inconsistency was measured using the Higgins' I 2 statistic. Specifically, I 2 = 0 indicated no heterogeneity and low, moderate and high heterogeneity was identified when I 2 < 25%, 25-75%, and >75%, respectively [33]. Publication bias was assessed through Egger's test and funnel plots using at least ten studies [34]. Subgroup analyses were conducted to examine the potential sources of heterogeneity, including age (young (18-39 years old)/middle-age (40-64 years old)/both) [35], BMI (BMI < 25/BMI ≥ 25), menopausal status (postmenopausal/premenopausal/both), country, and ethnicity. All statistical analyses were performed with Review Manager, version 5.4.1 (The Cochrane Collaboration, London, UK, 2020).

Study Selection and Characteristics
A total of 5112 records were yielded from the database and manual search. After screening 126 full texts, 23 eligible studies were finally included in the precent review. The most common reason for exclusion was the unavailability of female-specific data. Figure 1 presents the PRISMA system outlining the study process.
From 14 observational studies, objectively assessed PA included minutes in LPA, MPA, VPA, MVPA and PA bouts, steps, TPA, and met or did not meet the PA guidelines.
Moreover, according to the GRADE framework, very low to moderate quality evidence was reported, with none being upgraded. The small sample size of the intervention studies was the most common reason for the downgrade, and the lower quality for observational studies was mostly due to the inconsistency of findings. Supplementary Table S1C presents the details on the quality of evidence according to study design and the categories of cardiometabolic health outcomes.

Lipid Profile
Both observational and intervention studies reported some favorable relationships with lipid outcomes. One quasi-experimental study found a favorable effect on LDL after a 9-week walking program with 7056 average daily steps [50]. However, from three random experiments, only one study conducted a 24-month moderate-intensity exercise intervention, reporting favorable effects on HDL and TC [57]. The other two random experiments studies showed no effect of increased daily steps on HDL, TG, or TC [51,53].
From nine studies that reported cross-sectional evidence, three studies (75%) reported favorable associations between daily steps and HDL [48] and TG [40,48]. Furthermore, LPA [39] and meeting the recommended 150 min MVPA weekly [47] were reported to be beneficial to TG and TC. Two studies (66.7%) suggested a favorable association between MVPA and HDL [40,45]. Of note, an unfavorable but small relationship with TC was found in overweight Latin women [40]. Similarly, unfavorable relationships with LDL or TC were reported in obese African American women with a lower socioeconomic status [44].
Although two random experiments reported no effects on FPG or HbA1c, which was consistent with findings from observational studies, HOMA-IR was improved as a result of engaging in a walking program [51,52]. These findings were not replicated in quasi-experimental studies as a favorable effect on FPG [53], and PPG [58] was observed following walking programs. In addition, one study used a crossover design to examine the effects of three conditions on carbohydrate metabolism and found that increasing the percentage volume of MVPA had a favorable effect on PPG, while LPA had no effect [55].

Endocrine Regulation
One randomized controlled trial and two quasi-experimental studies consistently indicated no effects on fasting insulin or postprandial insulin after engaging in the walking program [50,52,58]. However, some favorable associations were found in cross-sectional observational studies. One study found favorable associations between daily steps with fasting insulin and postprandial insulin [38]. Furthermore, fasting insulin was shown to be favorably associated with 10-min MVPA bouts [39]. It was worth noting that one study reported an unfavorable association between TPA and fasting insulin [45].

Inflammation Markers
Only one random experiment examined the effect on inflammation markers and reported that increasing daily steps had no effect on CRP after a 12-week exercise intervention [51]. However, one observational study reported favorable associations between CRP and MVPA and MVPA bouts [39]. The opposite results were reported by Slater et al. (2021), who found that both TPA and MVPA were detrimental to CRP [45]. There was no relationship between PA and TNF-alpha [37] or IL-6 [39].

Metabolic Syndrome
Findings regarding to MS were unequivocal in intervention and observational studies. One quasi-experimental study reported a positive effect on MS score after participating a 9-week walking program [50]. Cross-sectional studies found the incidence of MS was favorably associated with daily steps [49], the volume of LPA [36] and MVPA [42].

Meta-Analysis
There were six studies (4 cross-sectional, 1 quasi-experimental and 1 RCT) that provided correlational data, which could be pooled to conduct the meta-analysis. Table 4 and Figure 2 illustrate the correlational meta-analysis for the included studies. Table 4. The meta-analysis of the association between physical activity and cardiometabolic health outcomes (all analyses were performed using the random-effect model).

Study Group
No. Studies The results from three studies assessed MVPA could be pooled into a meta-analysis (n = 545). The pooled results showed a significantly favorable but small relationship between MVPA and HDL (r = 0.16; 95% CI: 0.06, 0.25; p < 0.01), with a low heterogeneity between studies (I 2 = 19%, p = 0.29). However, according to subgroup analysis, no significant relationship between MVPA and HDL was detected in young (n = 2) and premenopausal (n = 2) women. Studies conducted in the US (n = 2) and Caucasian (n = 1) women also showed no relationship between MVPA and HDL.
According to a pooled analysis, there were not any significant associations between MVPA and DBP, SBP, LDL, FPG, and TG. Subgroup analysis revealed that age, BMI, menstrual status, country, and ethnicity had no effect on the association.
There was no significant correlation between daily steps and FPG (n = 313) (r = −0.12; 95% CI: −0.24, 0.01; p = 0.06, n = 3). The between-study heterogeneity was low (I 2 = 42%, p = 0.18). The subgroup analysis based on the age, BMI, menstrual status, country, and ethnic did not modify the association. The results from three studies assessed MVPA could be pooled into a meta-analysis (n = 545). The pooled results showed a significantly favorable but small relationship between MVPA and HDL (r = 0.16; 95% CI: 0.06, 0.25; p < 0.01), with a low heterogeneity between studies (I 2 = 19%, p = 0.29). However, according to subgroup analysis, no significant relationship between MVPA and HDL was detected in young (n = 2) and premenopausal (n = 2) women. Studies conducted in the US (n = 2) and Caucasian (n = 1) women also showed no relationship between MVPA and HDL.
According to a pooled analysis, there were not any significant associations between MVPA and DBP, SBP, LDL, FPG, and TG. Subgroup analysis revealed that age, BMI, menstrual status, country, and ethnicity had no effect on the association.
Additionally, we conducted a meta-analysis to examine the effect of meeting PA guidelines on HOMA-IR, since there were three comparable studies (Table 5 and Figure  3). The mean and standard deviation were extracted from those who met the 150 min of MPA and those who did not to calculate the standardized mean differences. The pooled results showed that following the recommended MVPA level had no significant effect on The pooled analysis of four studies (n = 349) revealed that daily steps were not associated with HDL (r = 0.24; 95% CI: −0.07, 0.54; p = 0.13), with a high heterogeneity between studies (I 2 = 84%, p < 0.05). However, subgroup analysis (n = 14) revealed a significantly stronger association in middle-aged women in studies conducted in Japan (r = 0.85; 95% CI: 0.58, 0.95; p < 0.001; n = 1). The between-study heterogeneity was mostly explained by the subgroup analysis of BMI.
Additionally, we conducted a meta-analysis to examine the effect of meeting PA guidelines on HOMA-IR, since there were three comparable studies (Table 5 and Figure 3). The mean and standard deviation were extracted from those who met the 150 min of MPA and those who did not to calculate the standardized mean differences. The pooled results showed that following the recommended MVPA level had no significant effect on HOMA-IR (SMD= −0.22; 95% CI: −0.46, 0.02; p = 0.08), with a low heterogeneity (I 2 = 11%, p = 0.32). According to subgroup analysis, meeting PA guidelines was significantly associated with lower HOMA-IR in studies conducted in the USA (n = 2) and in Caucasian women (n = 1). Table 5. The meta-analysis of the effect of meeting physical activity guideline on HOMA-IR (all analyses were performed using the random-effect model).

Variables
No

Discussion
This systematic review and meta-analysis are the first to synthesize studies that investigated the association between objectively determined PA volume and clinically relevant cardiometabolic biomarkers in adult women across a range of study designs. Although relatively limited by the small number of included studies, the evidence examining the association between objectively assessed PA and cardiometabolic indicators points towards a favorable association between MVPA and HDL. Evidence of a beneficial effect on other cardiometabolic outcomes seems to be limited.

Meta-Analytic Findings
Findings from the meta-analysis revealed that spending more minutes on MVPA were significantly associated with a healthier HDL; however, the pooled effect size was small. No significant associations were observed between MVPA and most cardiometabolic biomarkers, including SBP, DBP, LDL, FPG, and TG.
Subgroup analysis showed significant differences in the association between steps and HDL across various ages, countries and ethnicities. However, it was noteworthy that the number of studies included in the subgroup was small, and caution is required when drawing conclusions from subgroup analyses.

Association between Steps and Cardiometabolic Biomarkers
The observational and experimental evidence examining the associations between daily steps and cardiometabolic health indicate no effects on improving cardiometabolic biomarkers, including BP, lipids, glucose, insulin, and inflammation markers. This was supported by the meta-analytic findings that daily steps were not significantly associated with HDL or FPG.
Even though most studies reported no association between steps and BP, three studies conducted a long-term walking program in obese women and observed decreases in SBP after the intervention [52,53,58]. Participants in these three studies were all obese, with elevated or stage I high SBP at baseline, and moreover, their daily steps doubled to about 10,000 after the intervention. This favorable effect was supported by a systematic

Discussion
This systematic review and meta-analysis are the first to synthesize studies that investigated the association between objectively determined PA volume and clinically relevant cardiometabolic biomarkers in adult women across a range of study designs. Although relatively limited by the small number of included studies, the evidence examining the association between objectively assessed PA and cardiometabolic indicators points towards a favorable association between MVPA and HDL. Evidence of a beneficial effect on other cardiometabolic outcomes seems to be limited.

Meta-Analytic Findings
Findings from the meta-analysis revealed that spending more minutes on MVPA were significantly associated with a healthier HDL; however, the pooled effect size was small. No significant associations were observed between MVPA and most cardiometabolic biomarkers, including SBP, DBP, LDL, FPG, and TG.
Subgroup analysis showed significant differences in the association between steps and HDL across various ages, countries and ethnicities. However, it was noteworthy that the number of studies included in the subgroup was small, and caution is required when drawing conclusions from subgroup analyses.

Association between Steps and Cardiometabolic Biomarkers
The observational and experimental evidence examining the associations between daily steps and cardiometabolic health indicate no effects on improving cardiometabolic biomarkers, including BP, lipids, glucose, insulin, and inflammation markers. This was supported by the meta-analytic findings that daily steps were not significantly associated with HDL or FPG.
Even though most studies reported no association between steps and BP, three studies conducted a long-term walking program in obese women and observed decreases in SBP after the intervention [52,53,58]. Participants in these three studies were all obese, with elevated or stage I high SBP at baseline, and moreover, their daily steps doubled to about 10,000 after the intervention. This favorable effect was supported by a systematic review, in which the decreased SBP was found to be associated with higher baseline values and the magnitude of change in steps per day [59].
Experimental evidence suggests that increasing daily steps after intervention had no effects on HDL, LDL, TG or TC [51,53]. A pedometer-based walking intervention reported no effects on HDL, TG, or TC, while LDL improved at the end of the intervention [50]. The improved LDL was mainly due to weight loss and improved body composition [60]. However, one intervention study reported favorable associations between steps with HDL and TC after the intervention [57]. It was supposed that the improved lipid profiles were mostly due to additional moderate-intensity aerobic training rather than the increased steps, since aerobic exercise interventions had a more consistent and potent effect on improving HDL and associated cardiometabolic health indicators [61,62].
Cross-sectional evidence revealed that there was no association between steps and most blood lipids [38,40,44,48]. This was supported by a cross-sectional study, in which no differences were observed for TC, LDL, HDL, and TG between the group with more than 7500 steps per day and the group with less than 7500 [63]. However, body composition affected this relationship, as steps were significantly associated with HDL and TG after adjusting for fat mass and fat-free mass [48]. Interestingly, the opposite findings were reported by Panton et al. (2007), who found that women who walked at least 5000 steps per day had worse LDL and TC than those walking less than 5000 steps [44]. A potential explanation was that, for obese women, more steps were needed to improve lipid markers.
There was consistent evidence from experimental and observational studies that steps had no effects on carbohydrate metabolism [38,40,46,48,[50][51][52]. Results from other intervention studies also supported the idea that walking programs had no effect on improving FPG [64,65]. Although most studies reported no improvements in FPG, a favorable effect on PPG was observed after increasing daily steps during the 8-week walking program [58]. This finding was consistent with a prospective study that there was a weak favorable correlation between previous daily steps and 2 h-PPG, but no correlation with FPG [65].
Experimental studies reported consistent findings that increased daily steps after intervention had no effects on insulin sensitivity [50,52,58], while a cross-sectional study observed the lower levels of fasting insulin and HOMA-IR in more active women [38].
Despite there being no relationship between daily steps and cardiometabolic health outcomes suggested by most of the studies, consistent beneficial effects were observed for MS score, which was defined as the sum of the number of individual MS indicators [49,50]. Both observational and intervention studies found a favorable association between steps and MS score, which could be supported by longitudinal studies. Huffman et al. (2014) conducted an observational study from NAVIGATOR and found that baseline steps were independently associated with reductions in MS score, which was calculated by summing each standardized MS component [64]. Ponsonby et al. (2011) followed 458 adults with normal glucose and found that a higher level of daily steps was associated with a lower risk of the incidence of abnormal glucose metabolism 5 years later [66].
Walking is incidental to daily life and the accumulated number of steps were mostly at a low intensity. Walking intensity was more important than walking volume in terms of the association with cardiometabolic health, and this might explain why the beneficial effects on cardiometabolic biomarkers were hardly observed [67]. In addition to walking intensity, evidence from experimental studies indicated that the baseline value of biomarkers and magnitude of changes in daily steps affected the relationship between walking and cardiometabolic health outcomes. Likewise, body composition variables suggested by observational cross-sectional evidence could also mediate this relationship. Therefore, more controlled experimental and prospective studies with high-quality experimental designs are needed in the future.

Association between TPA and Cardiometabolic Biomarkers
Evidence from the current review revealed that there was no association between TPA and most cardiometabolic markers, expect fasting insulin and CRP [45]. Among obese women, there was no significant difference in TPA between metabolically healthy and metabolically unhealthy women [36]. However, a cross-sectional finding revealed that TPA displayed stronger associations with cardiometabolic biomarkers, including HDL, TG, FPG, fasting insulin, CRP, and SBP [68]. One potential explanation for the contradictory results might be the discrepancy of TPA between genders, as women engage in less TPA than men. Additionally, the relationship varied according to the ethnicity of the subjects. TPA was associated with fasting insulin in Pacific women but not European women, and the relationship between TPA and CRP was reversed between the two ethnicities. This may be because fasting insulin was nearly two times higher in Pacific women and CRP was positively associated with visceral fat, which was higher in Pacific women.

Association between Volume of PA at Different Intensity and Health Outcomes
Both cross-sectional and intervention evidence supported the idea that there was no relationship between LPA and cardiometabolic health markers [39,43,55]. Our findings were consistent with a previous review [69]. This review summarized the effect of exercise protocols delivered at light intensity and showed little support for the role of LPA in improving cardiometabolic health; moreover, it indicated that the applied dose of LPA was low among the included studies. However, there was emerging evidence that LPA had benefits for health [70][71][72]. In a cross-sectional study, LPA was shown to be significantly associated with TG and TC. These associations were independent of MVPA, but were attenuated by peak oxygen uptake (VO 2peak ) and body composition outcomes, indicating that VO 2peak and body composition might be important contributors to cardiometabolic health [39]. Previous cross-sectional studies showed that VO 2peak was associated with CVD risk factors, with moderate to strong correlations [73]. Likewise, Kodama et al. (2009) conducted a metaanalysis to quantitatively define the relationship between cardiorespiratory fitness and the incidence of CVD. The authors indicated that those with low cardiorespiratory fitness had a risk ratio for CVD events of 1.56 compared to those with high cardiorespiratory fitness [74]. Therefore, the cardiorespiratory fitness appeared to be an important confounder when investigating the relationship between PA and cardiometabolic health. Furthermore, the cardiorespiratory fitness should be taken into consideration when developing exercise protocols aiming to improve cardiometabolic health. Since high-intensity exercises were well-documented to be effective and efficient in improving cardiorespiratory fitness [75][76][77][78], in this regard, PA performed at higher intensity was recommended.
Moreover, a recent systematic review identified 24 cross-sectional and 6 longitudinal studies and found that LPA appears to be independently associated with better WC, TG, fasting insulin, and the presence of MS [79]. Additionally, replacing sitting with LPA was also found to be an effective way to improve health [80]. It was plausible that there is a threshold for PA at which health outcomes improved, and the threshold for LPA would be much higher due to the lower effects accumulated by the low intensities. This statement is supported by findings from a recent systematic review examining the relationship between LPA and cardiometabolic health and mortality in adults. The authors pointed toward the beneficial effects of LPA; however, LPA effects were from two to four times lower than MVPA effects for the same duration [14]. Moreover, the current PA guidelines recommended that at least 150-300 min/week of MPA were required to observe the benefits [3]. Studies that investigated MPA exclusively were sparse. Limited evidence from the present review supported the idea that MPA had no association with HOMA-IR [41,43].
It was generally believed that MVPA appeared to be more potently associated with cardiometabolic biomarkers. However, cross-sectional evidence from the present review suggests that there are no associations between MVPA and cardiometabolic risk indicators [39,40,45,46]. Only one study found a favorable association with the odds of being MS after controlling for age, ethnicity, and smoking [42]. In this study, MS was defined if participants meet three or more of the following criteria: (1) WC ≥ 88 cm; (2) TG ≥ 150 mg/dL or self-reporting on treatment; (3) HDL < 50 mg/dL or self-reporting on treatment; (4) SBP ≥ 130 mm Hg and DBP ≥ 85 mm Hg or self-reporting on treatment); (5) FPG ≥ 100 mg/dL or self-reporting on treatment). Cross-sectional evidence suggested that effects on cardiometabolic health seemed to be limited, while some prospective studies reported beneficial associations. A 10-year longitudinal study investigated the independent association of changes in MVPA and objectively measured cardiometabolic health and concluded that a greater decrease in MVPA was associated with a greater decrease in HDL and increases in clustered cardiometabolic risk score [81]. However, MVPA was self-reported. Mielke et al. (2021) investigated the prospective association between accelerometer-determined MVPA and cardiometabolic health in the transition to adulthood [82]. The authors suggested that young women who increased MVPA from 18 to 22 years old showed improvements in cardiometabolic health at age 22, and moreover, MVPA in 10-min bouts showed a stronger interaction than MVPA in 1 min. Similarly, a previous study conducted by Strath et al. (2008) analyzed data from the 2003-2004 National Health and Nutrition Examination Survey and found that the bouts of MVPA appeared to be a time-efficient strategy [83]. However, evidence from qualitative synthesis in the current review pointed towards there being no differences in beneficial effects between bouts of MVPA that lasted for more than 10 consecutive minutes and no bouts of MVPA [39,40]. One potential explanation for this was that women often engaged in short bouts of MVPA, which were normally less than 10 min [40]. Consistent findings were also reported by recent cross-sectional research showing that the impact of accumulated PA obtained from several short bouts of exercise is the same as the benefits obtained from longer-duration activities [84][85][86][87]. These results were in agreement with findings from prospective studies that short spurts of MVPA could provide protection against the onset of hypertension [88] and all-cause mortality [89]. Although MVPA in 10-min bouts was generally recommended for its health benefits, the accumulated evidence from cross-sectional and prospective studies showed that short-lived MVPA was associated with health outcomes. As such, the move towards recommending MVPA of any duration through the PA guidelines appears be a pragmatic change [3,4].
Research focusing on total volume suggested that there was emerging evidence that the total volume of PA, not the minutes accumulated in bouts, was important in relation to health [68,[86][87][88]. Moreover, PA of a sufficient volume was favorably associated with cardiometabolic health, independent of PA intensity [90]. In a cross-sectional study assessing the relationship between PA and cardiometabolic health in overweight Latina women, minutes of MVPA bouts were shorter than overall minutes of MVPA and moreover, the effect size of the correlation with cardiometabolic indicators was smaller for minutes of MVPA bouts than overall minutes of MVPA [40]. Likewise, Green et al. (2014) found that overall minutes of MVPA was a stronger variable than the bouts of MVPA regarding the association with markers of cardiometabolic health in young women [39]. Although most evidence was from cross-sectional analysis, it was encouraging that the promotion of short bouts of MVPA was more likely to be feasible for most women. From the public health perspective, this has significant implications for inactive individuals, as health benefits could be achieved by simply being more physically active without emphasizing the duration of exercise.
We also examined the effects of meeting the PA recommendation that adults should undertake at least 150 min of MPA a week; overall, meeting PA recommendations had an unclear impact on cardiometabolic health. Few significant differences were found between women who were meeting the recommendations and those who were below the recommended levels [37,41,47]. This was supported by the meta-analytic findings that there was no effect on HOMA-IR when meeting the recommended level. Only TG and TC were found to be improved by meeting the PA recommendations [47]. On the contrary, several previous studies based on large-scale populations showed that following the PA guidelines was strongly associated with a lower risk of cardiometabolic disease [91,92]. Discrepancies among these findings may be explained by the mediating roles of body composition variables on the relationship between PA and insulin resistance [41,93]. Other research also suggested that greater adiposity was associated with higher concentrations of inflammatory markers [94,95].
It is worth noting that most large-scale studies included self-reported PA rather than objectively measured PA, which was believed to attenuate the credibility of the findings. Furthermore, the current guidelines were developed in accordance with reviewed evidence to assess associations between PA and a set of health outcomes; however, most of the evidence was based on subjectively determined PA. Despite the limitations of our relatively low-quality evidence, the results of the current review showed some support of the idea that objectively measured PA was not beneficial to most cardiometabolic outcomes. However, most studies using subjectively determined PA consistently reported a favorable association with health outcomes [96][97][98]. This discrepancy was mainly due to the weak correlation between subjective and objective methods for assessing the intensity and duration of PA [99]. We were unable to judge which was superior because both had several limitations. Therefore, a combination of subjective and objective methods would be expected to further clarify some of the issues revealed by this study.

Strengths, Limitations, and Future Directions
The strengths of the current study include the use of different types of study designs and the inclusion of objectively determined PA volumes. This review was the first to analyze the evidence from different study designs both qualitatively and quantitatively and to explore the association between PA volume and clinical health indicators in adult women.
It was important to note that there were some limitations. First, most of the synthesized evidence ranged from very low to low quality. This was mainly due to the small sample sizes and concerns regarding risk of bias in the results. However, we compared low-quality evidence to high-quality evidence in the discussion, and additional high-quality and wellcontrolled intervention studies with a large sample size will be required to increase the confidence of the findings presented here.
Secondly, most of the included studies were cross-sectional in design, using t-test or ANOVA without controlling for any potential confounders, such as age and body composition. An initiative to address this issue was setting age to be the most import confounder when rating the quality of the included studies. Furthermore, the most-adjusted data were included in the discussion and meta-analysis. In addition, although sedentary behavior was documented to be associated with cardiometabolic health, it was not assessed in the current review. Taken together, the absence of these confounders attenuated the association between PA and cardiometabolic health, and our findings should be interpreted with caution.
Thirdly, our findings must be interpreted with the methodological consideration that PA at different intensities was defined as reported in the studies. Therefore, the heterogeneity in the different definition of PA categories, including different cut-points of counts, METs, and vertical acceleration peaks, was a potential source of inconsistent findings. Furthermore, the use of different epochs might also contribute to overestimation or underestimation of the amount of PA at a particular intensity. For instance, studies using longer epochs (e.g., 10 min) were more likely to underestimate the volume of higherintensity PA than those using 60-s epochs. Finally, accelerometer-determined PA was unable to quantify certain activities, such as yoga, Pilates, and swimming, and unable to precisely calculate the energy expenditure of PA. Likewise, the pedometer was unable to quantify the intensity of walking. To deal with these limitations, standardized cut-points, shorter epochs, and pattern recognition should be applied in future.
Lastly, findings from intervention studies were synthesized with a small sample size. Further large-scale intervention studies were need. Likewise, findings from subgroup analyses were limited and should be considered preliminary due to there being only a few studies including each subgroup category.

Registration
This protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO). The registration name was physical activity and health indicator in women: a systematic review and meta-analysis, and the registration number was CRD42022307774 (https://www.crd.york.ac.uk/prospero/display_record.php?RecordID= 307774). The current systematic review and meta-analysis were conducted with regard to the association between PA and cardiometabolic health.

Implications for Practice and Future Research
Our systematic review and meta-analysis found that accelerometer-and pedometerderived PA were not associated with most individual cardiometabolic health outcomes. These findings were inconsistent with those based on subjectively measured PA. For future improvements in objective measures, the gender-specific cut-points, activity pattern recognition was shown to be more likely to improve our knowledge of the health benefits of PA.
Our review found evidence that walking programs were effective in increasing daily steps among adult women, while significant improvements in cardiometabolic indicators were hardly observed following interventions, except among obesity participants. However, some improvements in SBP were reported among obese women with a higher SBP value at baseline. Furthermore, we found that increasing PA was associated with a higher HDL; however, this favorable association was attenuated among young women. Further research should pay greater attention to potential confounders, such as age, body composition and cardiorespiratory fitness, when investigating the association between PA and cardiometabolic health in adult women.

Conclusions
The findings from the present systematic review and meta-analysis provide evidence that objectively measured PA is not associated with most cardiometabolic health outcomes in healthy adult women. However, it is most compelling that being more physically active is beneficial for MS. For women, it makes more sense to emphasize the volume of PA rather than whether the volume of PA is sporadic or occurs in bouts. Even though lowto-moderate-intensity PA contributes the most to the PA patterns observed in women, PA performed at a higher intensity is more effective in improving cardiometabolic health. The present review also highlights that meeting the recommended 150 min of MVPA each week is not enough to observe significant beneficial effects. However, further high-quality studies with less heterogeneity are needed to yield compelling findings on the association between PA and cardiometabolic health in women.

Acknowledgments:
We would like to thank Yaodong Gu from Faculty of Sport Science at Ningbo University for feedback on the manuscript at all stages of this study.

Conflicts of Interest:
The authors declare no conflict of interest.  "blood pressure" OR "systolic blood pressure" OR "diastolic blood pressure" OR triglyceride OR TG OR "high density lipoprotein" OR HDL OR "low density lipoprotein" OR LDL OR "total cholesterol" OR TC OR insulin OR HOMA OR glucose OR HbA1c OR "glycosylated hemoglobin" OR "glycated hemoglobin" OR CRP OR "C-reactive protein" OR IL-6 OR interleukin-6 OR TNF-alpha OR TNF-α OR "Cardiometabolic Risk ))) AND ("blood pressure" OR "systolic blood pressure" OR "diastolic blood pressure" OR triglyceride OR TG OR "high density lipoprotein" OR HDL OR "low density lipoprotein" OR LDL OR "total cholesterol" OR TC OR insulin OR HOMA OR glucose OR HbA1c OR "glycosylated hemoglobin" OR "glycated hemoglobin" OR CRP OR "C-reactive protein" OR IL-6 OR interleukin-6 OR TNF-alpha OR TNF-α OR "Cardiometabolic Risk #5 "blood pressure" OR "systolic blood pressure" OR "diastolic blood pressure" OR triglyceride OR TG OR "high density lipoprotein" OR HDL OR "low density lipoprotein" OR LDL OR "total cholesterol" OR TC OR insulin OR HOMA OR glucose OR HbA1c OR "glycosylated hemoglobin" OR "glycated hemoglobin" OR CRP OR "C-reactive protein" OR IL-6 OR interleukin-6 OR TNF-alpha OR TNF-α OR "Cardiometabolic Risk Factors" OR "Metabolic Syndrome" #1 OR #2 OR #3 OR #4 AND #5

Appendix A
(TITLE-ABS-KEY(accelerometry OR accelero * OR actigra * OR actigraphy) OR TITLE-ABS-KEY(objectively AND measured AND physical AND activity) OR TITLE-ABS-KEY(objectively AND assessed AND physical AND activity) OR TITLE-ABS-KEY(pedometer)) AND (TITLE-ABS-KEY ("blood pressure" OR "systolic blood pressure" OR "diastolic blood pressure" OR triglyceride OR tg OR "high density lipoprotein" OR hdl OR "low density lipoprotein" OR ldl OR "total cholesterol" OR tc OR insulin OR homa OR glucose OR hba1c OR "glycosylated hemoglobin" OR "glycated hemoglobin" OR crp OR "C-reactive protein" OR il-6 OR interleukin-6 OR TNF-alpha OR TNF-α OR "Cardiometabolic Risk Factors" OR "Metabolic Syndrome")) AND NOT ((child *) OR (old *) OR (eld *) OR (pregnan *) OR (disable *) OR (athlete)) AND (LIMIT-TO (DOCTYPE, "ar")) AND (LIMIT-TO (LANGUAGE, "English")) Database Web of Science (n = 2067) #5 TS = ("blood pressure" OR "systolic blood pressure" OR "diastolic blood pressure" OR triglyceride OR TG OR "high density lipoprotein" OR HDL OR "low density lipoprotein" OR LDL OR "total cholesterol" OR TC OR insulin OR HOMA OR glucose OR HbA1c OR "glycosylated hemoglobin" OR "glycated hemoglobin" OR CRP OR "C-reactive protein" OR IL-6 OR interleukin-6 OR TNF-alpha OR TNF-α OR "Cardiometabolic Risk Factors" OR "Metabolic Syndrome")  #8 "blood pressure" OR "systolic blood pressure" OR "diastolic blood pressure" OR triglyceride OR TG OR "high density lipoprotein" OR HDL OR "low density lipoprotein" OR LDL OR "total cholesterol" OR TC OR insulin OR HOMA OR glucose OR HbA1c OR "glycosylated hemoglobin" OR "glycated hemoglobin" OR CRP OR "C-reactive protein" OR IL-6 OR interleukin-6 OR TNF-alpha OR TNF-α