Several cardiovascular and metabolic risk factors were associated with lowered endothelial function, measured by the PAT method, among those at risk of diabetes and cardiovascular disease. HDL cholesterol, current smoking, sex and body mass index explain more than a fourth of the variability in F-RHI.
As in previous studies, lower F-RHI was strongly associated with male sex in the separate linear regression model and the multivariable model [11, 13]. Since vascular function and development of cardiovascular disease differ between the sexes, it would be useful to study men and women separately instead of using sex as a covariate . We performed additional analyses to find possible sex differences in associations between F-RHI and risk factors, but because of the low proportion of male subjects, we were unable to build a reliable model for men alone.
Lower F-RHI was strongly associated with obesity. The strongest single association we found was between F-RHI and BMI. Similar results can be obtained with body fat percentage or WC. In separate sex-adjusted models, all the inter-correlated measures of obesity; body mass index, waist circumference and body fat percentage measured by bioelectrical impedance analysis, were related to lower F-RHI. Of these, BMI was chosen as a candidate for the multivariable model and thus entered the model. This observation was in accordance with earlier studies [11, 13].
Disturbed lipid profile associated with lower F-RHI. In separate sex-adjusted models, high HDL cholesterol was associated with higher F-RHI, and high triglycerides associated with lower F-RHI. HDL cholesterol entered the multivariable model, and was one of the strongest predictors of F-RHI. However, there were no associations between F-RHI and total cholesterol, LDL cholesterol, or lipid-lowering treatment. The lipid values in the study population were better than those of the Finnish working age population in general . One study with dominantly male subjects found a significant association between LDL cholesterol and lower PAT score, but none with HDL cholesterol, triglycerides or hyperlipidemia treatment . Hamburg et al. found significant associations between lower F-RHI and total/HDL cholesterol, triglycerides and lipid-lowering treatment .
The findings concerning diabetes, fasting glucose and current smoking also support earlier studies .
Inflammation plays an important role in the pathogenesis of atherosclerosis . We found no significant association between a marker of systemic inflammation, hs-CRP and F-RHI, although there was a modest non-significant association between higher log-transformed hs-CRP and lower F-RHI in the sex-adjusted model.
Age was not associated with F-RHI in the sex-adjusted model or in the multivariable analysis. It is possible that the range and distribution of age in our study population was not sufficient to show a possible relation. Earlier studies have shown a small but paradoxically positive association between advancing age and higher F-RHI , or no association at all between age and F-RHI .
Associations between F-RHI and blood pressure or resting heart rate were modest and somewhat counterintuitive. We found no association between F-RHI and systolic or diastolic blood pressure. Higher resting heart rate was paradoxically associated with higher F-RHI in the sex-adjusted model, although the beta coefficient was very small. Truschel et al.  found a significant but counterintuitive positive association between F-RHI and systolic blood pressure, but no relationship with resting heart rate . Hamburg et al., conversely found an association between lower F-RHI and higher heart rate .
Blood pressure values in our study population were quite low compared to the Finnish working-age population in general . More hypertensive data might possibly provide better information on the association.
Although the result of the PAT measurement, F-RHI, is mainly nitric oxide dependent, it is likely to be a combination of macro- and micro-vascular reactivity and affected by skin blood flow changes . It is possible that the hyperemic response in the fingertip microvessels differs from other vascular beds . This study confirms some of the associations found earlier between F-RHI and certain risk factors. Many of the established cardio-metabolic risk factors are associated with lower F-RHI as expected, but in this study the relationships with age, blood pressure and heart rate in particular are not parallel with the results of other studies, and need further investigation.
The study’s strengths lie in the relatively large sample size and a good coverage of municipal workers at intermediate cardio-metabolic risk. We were also able to determine a comprehensive set of risk factors and study their associations with a novel measure of endothelial function.
One limitation of this analysis is its cross-sectional design. In addition, the number of male subjects was low, because the proportion of females among Finnish municipal workers is commonly significantly higher. Measurements of vascular reactivity were not carried out in a fasting state due to the study design, although the subjects were instructed to eat only lightly before the measurements. A study using an ultrasound technique for measuring endothelial function suggests that strict requirements for fasting conditions may be unnecessary .
The menstrual cycle phase was not controlled for. The women ranged from peri- to post-menopausal, but information regarding hormonal status or the use of menopausal hormones was not collected. In addition to sex, hormonal status contributes to and modulates the development of cardiovascular disease and should be taken in to account in future study design .
The study sample consisted of Caucasian people of European origin, thus the findings may not apply outside this population.