This study has shown that the exercise training performed at Family Health Strategy units was able to reduce the overall cardiovascular risk of previously inactive individuals, by improving blood pressure, body composition, and biochemical blood markers. The FRS remained stable among those who declared to be spontaneously active over 12 months, and increased in physically inactive patients. Therefore, we confirmed the hypothesis that the supervised training applied within the CAP was more effective to reduce the cardiovascular risk over 12 months than spontaneous physical activity.
Similarly to other intervention studies with physical activity applied within the Family Health Strategy (10, 27), a classical randomization of participants in experimental and control groups was not possible in the present study. Firstly, only a small segment of the initially contacted individuals agreed to participate of the supervised exercise program. Moreover, we could not simply dismiss part of volunteers by assigning them to a control group, due to ethical problems – we could lose them, and this was not acceptable. It must noticed that this experiment was carried out with minimum interference in the routine of the Family Health Strategy units. For this reason, there was no change in the procedures usually adopted by the CAP to recruit patients for the supervised exercise program.
The dropout rate among the patients that enrolled in the physical training was of approximately 38%, which is undeniably high. However, this rate was similar to values reported by studies developed in different countries that applied exercise interventions in high social vulnerability contexts (10, 28). Overall, adherence remains a challenge to implement exercise routines within the Family Health Strategy. Despite wide advertising prior to the beginning of the program, the reach was low; of the initially contacted 1,342 patients, only 86 accepted to participate of the supervised training and 53 completed the 12-month intervention. This difficulty has been reported in other studies developed in Brazilian socially vulnerable communities (29).
It is easy to understand that a large proportion of residents studied or worked full-time. Hence, dwellers in those communities seem to be more prone to physical inactivity in comparison with those who inhabit wealthier spaces, due to geographic, social, and economic factors (8). In Brazil, Galvim et al. (10) analyzed the adherence, adhesion, and dropout reasons to a 6-month guided walk program offered to 106 patients attended by Family Health units in the city of São Carlos, SP, Brazil. The dropout rate was of approximately 50% and the main reported reasons were the working hours (28%), health (26%) and personal reasons (22%), or lack of time (11%). Those factors might also explain the high number of dropouts observed in our study, especially in FA. As abovementioned, the rate of employment among dropouts was three times higher than in the group that completed the supervised training. Conversely, the dropout rate in DA was of 10.4%. In this specific group, most individuals remained physically active during the 12-month intervention, perhaps due to the fact they were free to practice the chosen activities at their best convenience of time and local. Further studies are warranted to ratify this premise.
Randomized controlled trials frequently suffer from two major complications, i.e., noncompliance and missing outcomes. In order to avoid bias due to excessive dropouts, a potential solution to this problem is the intention-to-treat concept. The dropout rate within patients that underwent the supervised training was of almost 40%, which was considered high enough to introduce bias in the results. Therefore, we decided to analyze the outcomes using the intention-to-treat approach (30). In short, data from all patients initially assigned to FA were included in the statistical analysis, despite the treatment they actually received, and regardless of withdrawal from the exercise routine. This strategy avoids overoptimistic estimates of the efficacy of a given intervention resulting from the removal of non-compliers, by accepting that noncompliance deviations are likely to occur in actual clinical practice (30).
The intervention program provoked favorable changes in several cardiovascular risk markers, especially blood lipid profile, blood pressure, and type-2 diabetes (31-33). Since those factors integrate the FRS calculation, the lower cardiovascular risk found in DA and FA vs. PA was thereby expected. However, this was particularly true for patients who underwent supervised training (FA). Even including the dropouts in the analysis, the reduction in cardiovascular risk was markedly greater in FA than among patients who declared to be spontaneously active (DA). Actually, 5 of 6 cardiovascular risk markers improved in FA, while only LDL-C reduced in DA (see Table 4).
The association between physical inactivity and increased cardiovascular risk as estimated by the FRS has been demonstrated in previous studies. Galvão et al. (34) reported that a physically inactive lifestyle was more prevalent in men with FRS corresponding to high risk vs. low to moderate risk. Silva et al. (35) revealed that 83.3% of a postmenopausal women sample was classified as active or very active, whereas 87.7% obtained FRS scores corresponding to low cardiovascular risk the physical activity level. In which concerns the ‘Family Health Strategy’, we could find a single trial investigating the effects of a 20-week supervised exercise program on body composition and FRS (36). Although the sample has been restricted to postmenopausal obese women, similarly to our intervention there was a significant improvement in triglycerides and systolic blood pressure, as well as a reduction in FRS score.
Our findings are consistent with the premise that active lifestyles contribute to a lower risk of developing cardiovascular diseases (37). Therefore, this trial contributes to the current knowledge by demonstrating that strategies that contemplate supervised exercise programs within public health units can induce a greater reduction in cardiovascular risk than only stimulating the spontaneous practice of physical activities. It should be highlighted that we recognize the importance of physical activities performed in free time. A prior study from our laboratory has demonstrated that the only presence of physical activities during leisure time can significantly reduce the risk for cardiovascular and metabolic diseases (38).
In the present study, the cardiovascular risk estimated by the FRS increased in PI and remained stable in DA, which means that in terms of the prevention of cardiovascular diseases, spontaneous physical activity would be better than no physical activity at all. The amount of exercise in DA seemed to be enough to prevent the worsening of those factors, which explains the stability of FRS throughout the experiment. In contrast, our results reinforce the idea that supervised physical training, even when simple and inexpensive, might have greater benefits than unstructured activities in which concerns the prevention of cardiovascular diseases. The intention-to-treat approach is conservative in the production of significant effects of experimental interventions. Therefore, our results are promising and deserve consideration within public health promotion policies, particularly in developing countries.
Even though the practice of spontaneous physical activity should be encouraged, our findings suggest that it will not always achieve certain goals. It is feasible to speculate that differences in risk evolution between FA and DA were due to the fact that only FA met the minimum recommendations of health agencies for exercise prescription aiming at health promotion (39). In general, recommendations for physical training to promote health usually include aerobic exercises performed 3- to 5 days a week, with a duration of at least 30 min and moderate- to vigorous intensity. Moreover, recommendations also include a minimum of 1 set of 10 to 15 repetitions for 8 to 10 resistance exercises involving the major muscle groups, as well as flexibility and balance exercises at least 2 days a week (39).
It is usually accepted that dose-response relationships regarding several cardiovascular risk factors increase with exercise intensity and volume (26, 39). Interestingly, 69% of patients assigned to DA classified their physical activities as ‘light’, which reinforces the premise that these patients did not perform a sufficient amount of exercise to induce significant changes of cardiovascular risk factors included in FRS calculation. Our results support the importance of supervised programs to ensure adequate exercise prescription for health promotion. Indeed, there is a concern about the lack of public policies and programs that incorporate this premise. Our data suggest that adequate control of training stimuli within exercise programs developed in public health units is possible, which warrants attention from public health managers.
There is evidence suggesting that interventions conducted within the Family Health Strategy may increase the physical activity levels of participants. Ribeiro et al. (27), for instance, conducted a 12-month non-randomized trial investigating the effect of physical activity and health education interventions on the levels of physical activity of users of Brazilian Unified Health System attended by the Family Health Strategy, in the city of Sao Paulo (SP, southeastern Brazil). Both groups increased their overall time of physical activity (physical exercise, leisure, and transport-related), but individuals assigned to the health education group had a greater tendency to maintain a physically active routine than those who only attended exercise classes.
Unfortunately, we did not assess the physical activity level of the analyzed groups (FA, DA, and PI). The contacts established every 2 months were enough to confirm that no change in patients’ routine occurred. Therefore, those assigned in PI remained physically inactive, while changes in physical activity routine did not occur among in DA. This procedure was considered as enough to exclude or maintain the individuals in our experiment. Prior studies have suggested that sometimes the control groups increase their physical activity level after interviews about this topic (27, 29). Even accepting that this effect might have occurred in DA and PI, this would not change the fact that physical activity would remain spontaneous in DA, while individuals in PI continued to declare that they were physically inactive. Thus, bias due to this possibility was unlikely to have occurred.
The major limitation of this study was that participants were allocated in the experimental groups according to their desire to participate or not of the supervised exercise program in the CAP. As explained above, a classical randomization in experimental and control groups was not possible, since we did not intend to interfere in the routine of the Family Health Strategy unit. Another potential limitation was that the Framingham cohort was based on data from a mostly white population of European origin. Although previous studies have demonstrated the applicability of the FRS in other populations (Mariu et al., 2003; Chiu et al., 2004), we have to acknowledge that this may have interfered with the sensitivity of our results (see http://www.framinghamheartstudy.org). In addition, the sample size was not sufficient to elicit adequate statistical power for comparisons regarding HDL-C. Finally, only a few low-income communities located in the North Zone of Rio de Janeiro City were included in the experiment. Additional trials including health units integrated to the CAP in other regions of the city should be encouraged.