Physical Activity and Cardiometabolic Characteristics in Overweight Latina Women
- First Online:
- Cite this article as:
- Koniak-Griffin, D., Brecht, M., Takayanagi, S. et al. J Immigrant Minority Health (2014) 16: 856. doi:10.1007/s10903-013-9782-z
- 306 Views
This community-based study examined physical activity (PA) in relation to cardiometabolic risk factors among 223 adult, immigrant Latina women who were overweight or obese. Participants were predominantly of Mexican descent, married and low-income. Data were obtained through accelerometer readings and clinical measures (e.g., BMI, waist circumference, blood pressure, lipid profile, fasting blood sugar). Findings showed that many women were active (mean step count = 8,575 ± 3,191); 27.6 % achieved 10,000 steps per day, and only 11 % were sedentary. They engaged in short bouts of moderate PA rather than long-sustained moderate-to-vigorous activity. Three or more MetS traits were present in 45.3 % of the women, raising concerns about risk for diabetes and cardiovascular disease. More active women had lower BMI, weight, waist circumference, and triglycerides levels. Results point to the importance of targeting maintenance of PA in active women and intervening with culturally tailored programs to promote healthier behavior in those who are sedentary or somewhat active.
The obesity epidemic, a major public health challenge, disproportionately affects Latinos, the fastest growing ethnic minority population in the US. Compared to non-Latino whites (NLW), Latinos have a 21 % greater prevalence of obesity , while Mexican–American (MA) women are 1.5 times as likely to be obese as women in general . Low-income Latina women, particularly those of Mexican descent, face increased risk for cardiovascular disease (CVD) and other chronic illnesses due to overweight/obesity, sedentary lifestyle [2, 3], and other risk factors such as type 2 diabetes (T2D), hypertension , and dyslipidemia . Metabolic syndrome (MetS), the constellation of metabolic abnormalities including dyslipidemia, hypertension, hyperglycemia, and obesity, is associated with increased risk of diabetes  and CVD [7–9]. MA women are 1.5 times as likely to have MetS as NLW women , with abdominal obesity being the most common MetS trait reported in the former group .
MetS and its primary clinical manifestations, CVD and T2D, may be reduced by physical activity (PA) and diet . Moderate-intensity PA for as little as 30 min a day, 5 days a week, reduces risk of CVD in adults . Despite increasing awareness of this preventive health strategy in the general population, low levels of PA continue to prevail, particularly among MA women [2, 3, 13], of whom only 21.9 % meet federal PA recommendations .
Past studies demonstrate the health benefits of PA in increasing cardiovascular fitness, control of hypertension , and glycemic control [16, 17], and decreasing CVD risk . Moreover, the benefits of PA, regardless of weight loss or category of body weight, are supported by large epidemiologic studies showing more active persons have reduced risk of CVD and all-cause mortality [19, 20]. Recent data indicate that engaging in light or moderate/vigorous PA is associated with a decrease in risk of CVD mortality of approximately 30 %, a relationship not notably weakened in the presence of metabolic risk factors such as dyslipidemia and obesity . A Cochrane review noted that exercise resulted in only marginal weight loss in overweight/obese people but significantly improved lipids and blood pressure (BP) . Similarly, the preponderance of evidence in another review indicated that aerobic exercise resulted in positive alterations in lipoproteins in both normolipidemic and dyslipidemic individuals, with most consistent findings demonstrated for increase in HDL-C . Findings from the 2003–2006 National Health and Nutrition Examination Survey (NHANES) showed that moderate-to-vigorous PA was independently associated with HDL-C in predominantly overweight or obese adults . The few studies examining the effects of PA interventions with Latinos reveal mixed findings . Positive outcomes have included improvements in BMI , total cholesterol (Total-C) [26, 27], lower triglycerides (TG), higher fat-free mass, maximum aerobic fitness (VO2max) , and increased activity levels [29–31]. However, a limitation of many prior studies is self-report of PA levels . Comparison of self-reported versus actual measured PA (by accelerometer, pedometer, or indirect methods) shows that women report significantly (183 %) higher PA than objectively measured . Subjective over-reporting of PA has been confirmed in a recent study comparing Latino and NLW populations .
In response to the need for improved measurement of PA and its correlates as well as the limited research involving Latina women, we conducted a community-based investigation using accelerometer readings and objective clinical measures. We report PA levels in relation to MetS factors specified by the American Heart Association  in overweight/obese Latina women, predominantly of Mexican descent.
Data were obtained as part of a randomized controlled trial examining the effects of a 6-month lifestyle behavior intervention involving overweight/obese Latina women. Only baseline data (collected prior to initiation of intervention) are included in the current analyses. All research protocols were approved by the Institutional Review Board of the University of California, Los Angeles (UCLA).
Participants were recruited from December 2009 through July 2010 from community settings such as parent education centers, churches, laundromats, and organizations providing services to children and families (e.g., English-as-a-Second-Language classes, job training). These were located within two adjacent communities of Los Angeles County with similar sociodemographic profiles. The women met the following criteria: self-reported Latina, 35–64 years of age, Spanish- and/or English-speaking, and BMI of ≥25. Those with impaired physical mobility, type 1 diabetes, uncontrolled hypertension, or history of a heart attack or stroke were excluded.
The bilingual, bicultural team of data collectors included community health workers (promotoras) and a registered nurse who performed assessments of lipids, fasting blood sugar (FBS), and BP. Multiple morning data collection sessions were held to accommodate the schedules of participants and the fasting status necessary for blood sample collection. Questionnaires were administered via private one-to-one interviews in Spanish. Participants received a $25 gift card for the evaluation.
The Kenz Lifecorder Plus (Kenz, Nagoya, Japan), an accelerometer with established reliability and validity [36–40], was used to measure PA. This device measures vertical acceleration 10 times each second; data are aggregated over a 4-s interval (epoch), to generate a number of “counts,” providing a “real-time” index (every 2 min) of the amount of body movement. Freedson et al.  demonstrated that activity counts and steady-state oxygen consumption are highly correlated (r = 0.88). The Lifecorder activity counts were converted into METS (1 MET = 3.5 mL/kg min), thus enabling us to classify intensity according to accepted standards as well as calculate step counts. Participants were instructed to wear the monitor for 7 consecutive days during all waking hours except when bathing. After receiving verbal instructions to wear the accelerometer during all hours awake for 7 consecutive days and a demonstration on placement of the device, written protocols were provided with picture illustrations. Total activity level for each participant was averaged for each day and for the entire 7-day period. A minimum of 4 days of data (≥8 h/day) was required for inclusion in the analyses. Research findings on overweight and obese individuals show minimal difference in PA patterns using 8 to 12 h of data from 4 to 6 days of monitoring, with or without weekend days . By recording PA with accelerometers for 21 consecutive days, Matthews and associates  demonstrated that as few as 3–4 days of monitoring was required to achieve 80 % reliability for measuring activity counts and time spent in moderate-to-vigorous activity. A subsequent review of accelerometer-based activity assessments showed 3–5 days of monitoring is sufficient for reliably estimating usual or habitual physical activity .
We examined physiologic measures comprising the MetS in women, as defined by the US National Cholesterol Education Program-Adult Treatment Panel (NCEP-ATP III) ; these include (1) waist circumference > 88 cm; (2) serum TG ≥ 150 mg/dL; (3) high-density lipoprotein (HDL-C) < 50 mg/dL; (4) BP ≥ 130/85 mm Hg; and (5) FBS ≥ 110 mg/dL. In classifying MetS traits, we included women being treated for diabetes or hypertension with medication in accordance with the revised NCEP definition . Systolic and diastolic BP were taken following JNC7  procedural guidelines, using a Welch Allyn and Tycos BP Kit with TR-2 ProCheck Home Aneroid and Stethoscope. This device meets the AAMI accuracy standard of 3 mmHg. Elevated BP readings were rechecked using a mercury sphygmomanometer (average of 3 readings) . The FDA-approved Cholestech LDX System was used to measure Total-C, HDL-C, low-density lipoprotein (LDL-C), TG, and FBS. The Cholestech is Clinical Laboratory Improvement Amendments-waived and meets NCEP guidelines for precision and accuracy. The SECA 769 digital scale was used to measure weight with women wearing light clothing and no shoes. Height was measured using a SECA 220 Hite-Mobile Portable Stadiometer. BMI was calculated as weight in kilograms divided by the square of height in meters. Waist circumference was evaluated with a Gulick tape measure following National Obesity Expert Panel Report guidelines .
Demographic and background information included age, ethnicity, place of birth, education, income, and health history (e.g., hypertension, diabetes, depression). The General Acculturation Index examined cultural orientation and level of involvement in Latino culture (α ≥ 0.80) . The 5 items evaluated linguistic capabilities in speaking and reading English and Spanish, place of childhood residence, ethnic identity of friends, and level of pride in heritage. Item responses and overall scores ranged from Latino (1) to Anglo (5) orientation. Residence index was calculated by subtracting age from years in the US, with smaller negative number closer to zero indicating greater acculturation [50, 51].
Data analyses were performed using SPSS Version 19. Average daily steps, minutes in moderate activity, and minutes in bouts of moderate activity (1 bout ≥10 consecutive minutes in continuous PA at ≥3 METS) were calculated from Lifecorder readings. Sedentary states were indicated by readings of 0.5 METS, whereas 0 METS indicated that the accelerometer was not being worn. Review of the data revealed no evidence of manipulated recordings due to device misuse (e.g., shaking device to increase PA intensity or step count). Thirteen participants were excluded from PA analyses because they did not meet the previously described accelerometer recording criteria. Pearson correlations were computed to examine the relationships of PA with cardiometabolic characteristics. To provide an additional perspective, differences in clinical variables among PA groups (<5,000, 5,000–7,499, 7,500–9,999, and ≥10,000 average daily steps) were evaluated using analysis of variance (ANOVA). A second set of these analyses controlled for age; however, because age-adjusted results did not result in substantive differences in interpretation (based on significance or effect sizes), only the simple (not age-adjusted results) are presented in this paper.
Sociodemographic Background, Cardiometabolic Characteristics, and Physical Activity
Sociodemographic characteristics of participants (N = 223)
Mean ± SD
44.61 ± 7.92
1.48 ± 0.45
Mean years living in USb
18.62 ± 8.26
−25.80 ± 8.13
US (but raised in Mexico)
Other (dominican, Central or South American)
Some HS & high school graduate
Some college/trade school
College degree or higher
Married/living with partner
Uninsured/no health insurance
Diabetes (FBS ≥126 mg/dL or on antidiabetic therapy)
Hypertension (BP ≥ 140/90 mmHg or on antihypertensive medications)
Cardiometabolic characteristics and physical activity of participants (N = 223)
Mean ± SD
Risk classification: frequency (%)
32.62 ± 5.68
Overweight: 92 (41.3 %), obese I: 65 (29.1 %), obese II: 42 (18.8 %), obese III: 24(10.8 %)
175.05 ± 32.60
101.39 ± 11.47
Over 88 cm (35 in): 197 (88.3 %)
Systolic BP (mmHg)
114.08 ± 13.25
SBP < 120: 150 (67.3 %), 120–129: 48 (21.5 %), 130–139: 18 (8.1 %), 140 or above: 7 (3.1 %)
Diastolic BP (mmHg)
75.43 ± 8.98
DBP < 85: 186 (83.4 %), 85–89: 26(11.7 %), 90 or above: 11 (4.9 %)
111.56 ± 29.23
LDL-C < 100: 73 (32.7 %), 100–129: 85 (38.1 %), 130–159: 45 (20.2 %), 160 or above: 11 (4.9 %), missing: 9 (4.0 %)
44.50 ± 13.37
HDL-C 60 or above: 24 (10.8 %), 50–59: 43 (19.3 %), 40–49: 66 (29.6 %), < 40: 90 (40.3 %)
188.44 ± 33.85
TOTAL-C < 200: 150 (67.3), 200–239: 55 (24.7 %): 240 or above: 18 (8.1 %)
164.96 ± 87.87
TG < 150: 120 (53.8 %), 150–199: 48 (21.5 %), 200–499: 53 (23.8 %); 500 or above: 2 (0.9 %)
Fasting blood sugar (mg/dL)
100.43 ± 18.78
FBS 110 or above: 40 (17.9 %)
8,575 ± 3,191
<5,000 (sedentary): 23 (11.0 %), 10,000 or above: 58 (27.6 %)
Daily minutes in moderate PA
22.6 ± 18.9
30 min or more: 53 (25.2 %)
Daily minutes in bouts of moderate PA
16.9 ± 20.5
Levels of PA, as demonstrated by accelerometer step counts, were surprisingly high (8,575 ± 3,191), with 27.6 % of the women achieving 10,000 steps per day. Only 11 % were classified as “sedentary” based on an average daily step count <5,000, and 25.2 % were able to sustain moderate activity (≥3 METS) for 30 min or more daily. Women averaged 22.6 ± 18.9 min of moderate activity per day and 16.9 ± 20.5 min in bouts of moderate activity.
Correlations Between Physical Activity and Cardiometabolic Characteristics
Correlations between physical activity and cardiometabolic characteristics
Average daily steps
Minutes in moderate physical activity
Minutes in bouts of moderate physical activitya
Differences in Cardiometabolic Characteristics by Step Category
Cardiometabolic characteristics by step category
0–4,999 (n = 23)
5,000–7,499 (n = 60)
7,500–9,999 (n = 69)
≥10,000 (n = 58)
Step category based on average # of daily steps
LDL-C (n = 201)
HDL-C (n = 209)
TG (n = 208)
This study examined PA in relation to risk factors for CVD and MetS in Latina adults of predominantly Mexican descent living in an urban area of Southern California. Uniquely, participants were recruited at a grassroots level within the community rather than from clinics or hospitals. Their acculturation levels were low even with lengthy residence and exposure to US culture. Our findings challenge beliefs about sedentary lifestyles among Latina women. Despite overweight/obesity, only 11 % of the women were sedentary, and 28 % achieved a daily step count of 10,000 or higher. The average daily step count indicated that the women were fairly active. The women often engaged in short bouts of moderate PA rather than long-sustained moderate-to-vigorous activity. More active women generally had lower BMI, weight, waist circumference, and TG levels. Additionally, greater minutes in moderate PA were associated with higher HDL-C levels. In interpreting these findings, it is important to consider that aspects of movement reflected in step count differ from those in moderate PA, thus accounting for some of the variations in correlational findings. The absence of a relationship between PA and cardiometabolic indices such as BP and FBS may reflect the nature of the sample (i.e., predominantly normotensive and nondiabetic).
The high prevalence of MetS in our sample raises concern about the women developing T2D and CVD in later years. FBS in the range of 100–125 were evident in 86 (38.6 %) women. The most frequently observed MetS trait was large waist size. Research on correlates of MetS components in overweight/obese Latinos has shown that waist circumference is the only variable that predicts FBS, and BMI does not predict other cardiometabolic variables , suggesting that body weight per se is a poor predictor of metabolic health. Dyslipidemia was fairly prevalent as evidenced by low levels of HDL-C in ~70 % and elevated TG in ~46 % of the participants.
Regarding PA, in comparison to data from NHANES, the women in the present study exhibited higher daily step counts, and slightly fewer minutes in moderate activity as well as less in the sedentary category . This finding about differences in step count was unexpected given the higher energy expenditure of overweight/obese women due to greater body mass and the expectation that they would modify their PA to accommodate their weight. Anecdotally, many of the women engaged in planned, purposeful PA while performing activities of daily living, such as household activities, childcare, and shopping. Furthermore, the differences we observed in BMI across varying PA categories are consistent with NHANES findings [53, 54].
Step counts were not correlated with any of the lipid markers except TG levels; however, greater minutes in moderate PA were associated with higher HDL-C and Total-C levels. The latter finding is not easily explained. In their review of studies examining the association between PA and health, Kokkinos and Meyers  reported the most consistent findings have been demonstrated for increases in HDL-C. Arsenault and colleagues  found no changes in lipid profile following a 6-month exercise intervention in metabolically healthy overweight or obese women with moderately elevated systolic BP.
Our findings suggest that the favorable health behavior of PA may at least partially contribute to the “Latino paradox.” Despite a burden of CVD risk factors similar to non-Latino whites and a higher prevalence of certain risk factors such as T2D, Latinos have lower CVD mortality . Latina women have the highest life expectancy of all ethnic/racial groups at ~83 years, even in the presence of obesity . Creating a physical environment believed to support PA is important in helping women to maintain high PA levels, particularly those who live in communities with limited socioeconomic resources and environmental risk factors (e.g., heavy traffic, crime, lack of park facilities, etc.). Research indicates that perceptions of neighborhood characteristics, such as the availability of PA facilities, impact residents’ levels of PA [58, 59].
Before further considering implications of our findings, we acknowledge a number of limitations. Because the sample was comprised of predominantly women of Mexican descent living in an urban setting, the findings cannot necessarily be generalized to include other US Latino subgroups or Mexican–Americans living in rural settings. In addition, the nature of the PA was not quantified (i.e., leisure-time vs. work-related), and data on daily caloric intake were not collected. Rates of T2D and hypertension were based on our clinical evaluations without medical record verification. The potential effects of antihypertensive drugs on outcomes of the small number of women being treated were not evaluated. Acculturation level and years of residence in the US were not explored as potential modifiers of PA and its correlates. Although this study yields important findings about cardiometabolic factors related to physical activity, causation can not be inferred.
Our findings point to the importance of targeting maintenance of PA in active immigrant women and intervening to promote healthier behaviors in those who are sedentary or somewhat active. Evidence-based guidelines for CVD prevention in women, developed by an American Heart Association expert panel, include regular PA as part of the healthy lifestyle behaviors ranked as Class I (highest recommendation) for CVD prevention . Similarly, the JNC7 found the adoption of healthy lifestyles to be critical for the prevention of high BP and indispensable for the management of HTN . The Department of Health and Human Services Physical Activity Report Guidelines also support that for otherwise healthy sedentary individuals, PA is beneficial, and performing activity ~1 h per week at moderate intensity will provide small increases in cardiorespiratory and muscular fitness . Small changes, such as 5-min walking breaks from sedentary time, can yield beneficial effects on metabolic health variables . Culturally tailored lifestyle interventions that build on the strengths of Latina women while helping them to overcome barriers to PA and gain support for becoming more active are urgently needed. While supporting strategies to promote higher levels of PA, health professionals should recognize that the goal of 10,000 steps or 30 min of moderate activity per day, encouraged in the media and in practice, may not be achievable by all groups. Alternatively, walking fewer steps with higher intensity strides may be considered. Daily step count recommendations vary across countries, and in fact, the US President’s Challenge on Physical Activity and Fitness Awards Program recommends 8,500 steps for adults . Health professionals are encouraged to assess PA status routinely and to identify barriers and facilitators influencing this behavior. Strategies for overcoming barriers and securing needed support should be part of counseling. Family support towards PA among MA women and self-monitoring with pedometers may positively influence activity level [64, 65].
New Contribution to Literature
To our knowledge, this is the first study presenting a comprehensive analysis of accelerometer-derived indicators of PA and related cardiometabolic measures in a sample comprised solely of immigrant Latina women. Inclusion of objective measures more accurately evaluates PA than self-report, thereby strengthening the validity of our findings. Our results point to the need for comprehensive and culturally tailored lifestyle behavior programs for health promotion and disease prevention among overweight/obese, immigrant Latina women. Such interventions have the potential for a positive impact on the health of the fastest growing ethnic minority population within our nation. Future research should examine whether the type of PA (leisure vs. work-related) influences cardiometabolic profile and other CVD risk factors. Large cross-sectional studies with diverse samples of Latina women are needed to compare differences for factors such as country of origin, urban versus rural residence, and level of acculturation. Finally, potential causality between changes in levels of PA activity and CVD benefits may be examined through longitudinal studies.
This research was supported with funds from the National Heart, Lung, and Blood Institute (R01 HL086931) and is part of a registered clinical trial (NCT01333241). We would like to acknowledge the assistance of Antronette Yancey, MD, MPH, James Sallis, Ph.D. and Christian Roberts, Ph.D. for their expert contributions on measurement of physical activity and Carmen Turner for her excellent editorial assistance with manuscript preparation. This study would not have been possible without the cooperation and contributions of the Latina women participants, community partners and research staff involved with the project.