In this cross-sectional analysis of individuals enrolled in the WSHS, greater ambulatory activity was associated with lower waist circumference, and a tendency towards lower DBP and TC/HDL-C ratio in both men and women. Solely among women, a higher level of steps/day was associated with lower body weight, BMI, CRP, triglycerides, total LDL and small LDL particle concentration, but these associations were explained in part by the lower BMIs of the active individuals. Similarly, women taking ≥ 7,500 steps/day had a 62% lower odds of metabolic syndrome and a 43% lower odds of high CRP, that was attenuated after adjustment for BMI, and not present in the men. Although other studies have examined the relationship between physical activity and biological markers of health, to our knowledge this is the first study to examine this relationship specifically in smokers.
Our results regarding the association between ambulatory activity and the various health markers are reasonably consistent with prior work. Physical activity, including ambulatory activity, has been shown to be inversely related to BMI, LDL-C, TC/HDL-C, LDL lipoproteins, triglycerides, and blood pressure.10;29-36
We found very strong relationships between ambulatory activity and metabolic syndrome, consistent with what has been seen in a population of older adults,33
and for physical activity in general.12
The relationship between ambulatory activity and inflammation was also strong, consistent with previous studies.37;38
Overall, the associations we saw in this group of smokers were similar to what has previously been seen in other populations.
The differential associations we noted for women vs. men are somewhat surprising, but may be attributed to measurement issues. Although we objectively measured ambulatory activity, we cannot determine characteristics of that activity such as the intensity or pattern of steps. A previous Australian study noted that men who accumulated ≥ 10,000 steps day were more likely to report vigorous work activity or blue-collar occupations compared to less active men, while in women, the 10,000 steps/day goal was more likely to be met among those participating in ≥150 minutes of leisure-time activity/week.39
Our data are consistent with this finding in that the steps/day were more strongly associated with total moderate-vigorous activity in men than in women. Although correlations with total leisure-time activity were weak overall, they were somewhat stronger in women. Although strong inferences cannot be made as to the intensity of the actual steps recorded, these results suggest that the intensity and/or quality of the steps may have been different, such that a step in the men was not equivalent to a step in the women. Additionally, our men recorded substantially more steps/day than the women such that far fewer men were considered low active than the women. Although these measurement differences may underlie our results, it is also possible that there are biological reasons that explain the differences we saw among male and female smokers, and this appears worthy of further investigation.
After adjustment for BMI, the associations between greater ambulatory activity and many of the biological markers were attenuated; however, this was not unexpected as excess body weight may actually mediate some of these relationships. The interrelationship of activity and body fatness is complex and not always able to be clarified by linear regression modeling. Various studies have found that the relationship between physical activity and inflammation is attenuated after adjustment for fatness.40;41
Increased body size has also been shown to be associated with lipid biomarkers, with higher BMI more strongly associated with higher LDL-C, TC, and lower HDL-C than physical inactivity.40
In our study, greater ambulatory activity was related to a better LDL lipoprotein profile, which was in part related to BMI as was expected. Our results suggest that body fatness is an important mediator of the relationships we examined.
We saw statistically significant associations with many of the biomarkers even though the participants had reasonably healthy biomarker profiles. Participants in the WSHS were recently compared to non-smokers, former smokers, and current smokers who participated in the National Health and Nutrition Examination Survey (NHANES) in 2005-2006.42
Smokers in the WSHS had lower levels of TC (183.8 mg/dL vs. 202.5 mg/dL; p < 0.05), TG (142.2 mg/dL vs. 158.0 mg/dL; p < 0.05), and FG (94.9 ± vs. 102.3; p < 0.05) compared to smokers in NHANES, suggesting that smokers seeking cessation treatment may have a different health profile than current smokers in the general population. Consequently, it is possible that the association between greater ambulatory activity and health status would be stronger in a population of smokers not seeking treatment.
Understanding how much ambulatory activity is related to improved biomarkers of health is an important public health issue. Although limited by the observational nature of our study, we can still examine what level of ambulatory activity is associated with an improved biomarker profile. Waist circumference was significantly lower with ambulatory activity >5,000 steps/day; however, other significant relationships were only seen with higher levels of activity. DBP, FG, and TC/HDL-C were significantly lower in men and women with >10,000 steps/day vs. those with <5,000 steps/day. Among the women only, where we used the 7,500 step/day cutpoint, that level of activity was associated with significantly lower weight, BMI, CRP, TC, LDL particles, and small LDL particles. Additionally, this level of activity was related to substantial and significant reductions in the risk of metabolic syndrome and high-risk levels of CRP. These associations with what is considered to be the “somewhat active” category are consistent with prior studies, although relationships with specific biomarkers have varied.43-45
This suggests that ambulatory activity may not need to be >10,000 steps/day for health benefits to be accrued.
The main limitation of this study is its cross-sectional design, which does not allow us to make strong causal inferences about the observed relations. Also, reactivity to wearing the pedometer, in which subjects will increase their activity when they know they are being monitored, may also be a problem. A recent study suggested that pedometer reactivity was greatest when participants used an unsealed pedometer and recorded their steps/day as our subjects did.46
Therefore, it's possible we overestimated subjects' normal activity levels which would have attenuated predictive relations. However, it's also possible we underestimated activity as pedometers cannot capture non-ambulatory activities like swimming and cycling. As mentioned previously, another limitation is that we could not measure the intensity or duration of the steps taken. Finally, only 69% of the subjects enrolled in the WSHS had valid pedometry data, and therefore our results could be subject to selection bias. However, we did compare those with and without valid pedometry data and the subjects appeared to have similar demographic characteristics. Although it seems low, this 69% figure is similar to the proportion of samples with valid pedometry data seen in prior population-based studies in Colorado (68%)47
and Australia (70%).39
Strengths of this study include the objective measurement of ambulatory activity to assess daily physical activity. Most prior studies have usually measured physical activity through self-report, which is subject to validity issues and recall bias.48
Also, our study measures cumulative ambulatory activity, not just purposeful exercise, which has great translatable potential. With further work, the potential recommendations for a steps/day level has great public health appeal in that walking is a very common activity, people can accumulate steps throughout a day, and pedometers are reasonably cheap and easy to use for personal measurement.