Reducing the high risk of cardiovascular disease among the Japanese-Brazilian population was the motivation for this behavioral intervention, which could potentially be extended into a broader prevention program throughout the country. In this relatively simple intervention program, most of the participants (71.7%) maintained or improved their glucose tolerance status after the intervention. This study was able to identify baseline characteristics associated with the prevention of the progression of the natural history of glucose metabolism disturbances in two years.
Descriptive analysis identified three baseline characteristics - age ≤ 60 years, presence of glucose intolerance and CRP levels ≤ 0.04 mg/dL - which differed between progressors and non-progressors. Those who improved or maintained glucose tolerance status were younger than those who deteriorated, but age was not independently associated with non-progressor status in logistic regression. This finding might indicate that age is an important factor for deterioration of glucose metabolism, but other metabolic characteristics may minimize the expected effect of aging. According to the multivariate analyses, a lower grade of inflammation (reflected by the CRP levels) could identify those more prone to benefit from a behavioral intervention in terms of non-deterioration of glucose tolerance. Also, individuals with glucose intolerance (IFG or IGT) had a better response to intervention in terms of glucose tolerance outcome. Such finding is in agreement with the DPS, in which individuals with more unfavorable baseline diabetes risk score (FINDRISC) had a higher decrease in diabetes incidence [9
]. This may suggest that the impact of behavioral interventions on cardiometabolic profile is more pronounced in those with any degree of metabolic disturbances. In fact, a higher decrease in plasma glucose in these individuals was found when compared to normal glucose tolerant ones. We speculate that those with abnormalities of glucose metabolism may be more motivated to modify their lifestyle in order to improve their health, when submitted to prevention programs.
This finding of lower CRP level at baseline in the non-progressor group is in the same line of the results from prospective case-control studies, in which elevated levels predicted the development of type 2 diabetes, supporting a possible role for inflammation in diabetogenesis [20
]. In our study, multivariate analysis showed that CRP levels were inversely and independently associated with the non-progression status, even adjusted for age, glucose tolerance status and measurements of adiposity. Among Japanese Americans, Nakanishi et al
also found that CRP was a risk factor for development of type 2 diabetes, independently of either obesity or insulin resistance [2
]. We suggest that the lower levels of CRP might be indicative of a better prognosis in terms of metabolic risk.
The goals of our community-based intervention program were not achieved by the majority of the participants. This could be attributed to our approach - tailored to the reality of a developing country - that was much less intensive than that used in other intervention trials [6
]. Similar proportions of progressors and non-progressors achieved the goals. It means that these goals should be reflecting mainly the effect of the follow-up and reinforces the importance of identifying predictors at baseline. Interestingly, the most commonly achieved goal was the reduction of saturated fat intake. Considering previous findings in the same population, regarding the association of fat intake with metabolic syndrome [5
], such reduction was highly desirable and could have contribute to the improvement in metabolic profile.
Slight, but significant reduction in anthropometry occurred following the intervention in the whole sample. The lack of association of the intervention-induced adiposity reduction on glucose metabolism benefits is in agreement with other studies conducted in Asian populations. The Indian Diabetes Prevention Program showed reduced risk of diabetes without any significant change in anthropometric variables [24
]. Another study based on lifestyle intervention showed a greater reduction in the incidence of diabetes in Japanese men than might be expected simply on the basis of the decrease in BMI [25
], suggesting that while weight loss may be desirable, it does not fully explain the effects of behavioral interventions.
Our study has limitations. The definition of progressors implicates in certain heterogeneity of this group of individuals, including not only 32 incident cases of diabetes, but also those who progressed from normal glucose tolerance to IFG or IGT. Another point to be observed is that we do not have the data of those individuals who were lost to follow-up that could implicate in a selection bias. However, baseline characteristics of the subset that did not complete the intervention program did not differ from those who were evaluated at the end of the intervention. The findings of the present study cannot be extrapolated to other populations.
In summary, our findings suggest that individuals with lower CRP levels and with glucose intolerance at baseline were the ones who will benefit from this relatively simple intervention in terms of glucose metabolism, independent of body adiposity. The homogeneous behavior of the whole sample during the intervention period suggests that the simple fact of being followed-up may result in favorable changes in glucose tolerance, which seems relevant in terms of public health for developing countries.