In this study we examined the joint association of physical activity and relative bodyweight with subsequent physical and mental functioning, among middle-aged employees of the City of Helsinki, over a follow-up of 5–7 years. We found that both overweight and physical inactivity jointly contributed to poor physical functioning, although weight tended to dominate the association somewhat. Thus high physical activity at baseline may lead to better physical functioning at follow-up, both among those of normal weight and the overweight, whereas overweight contributes to poor physical functioning even among those who are highly active. The highly active overweight and the inactive normal-weight, were equally associated with poor physical functioning, with those who where inactive and overweight being associated most strongly with poor physical functioning. In mental functioning, physical inactivity tended to dominate the joint association between physical activity and body weight. Adjusting for baseline physical or mental functioning attenuated the association between weight and physical activity with functioning at follow-up, but the associations remained in the overweight inactive group.
There is previous evidence that physical activity and weight maintenance both contribute to better physical functioning
[17–24]. This is in accordance with our study, although we found that weight somewhat dominates the joint association with physical functioning. Most studies
[17–21, 24] show that physical activity is more important than weight maintenance for maintaining good physical functioning. Only two previous studies
[22, 23] suggest that the effect size of weight is of a similar magnitude to that of physical activity. Most of these studies
[17–21] were cross-sectional and as such unable to show the direction of the association. Physical activity may relate to better physical functioning owing to its muscle-strengthening effects
 and improvement of balance control
. Additionally it may prevent various chronic diseases
, which undermine physical functioning. Maintaining normal weight may also prevent several diseases
 and mobility disabilities
 owed to overweight.
The available evidence suggests
[23, 24] that the physically active overweight have better physical functioning than their normal-weight counterparts, but in our study the physical functioning of highly active overweight and inactive normal-weight participants was similar. This could result from the use of distinct measures of physical functioning or different cut-off points between weight and activity groups. To further analyse any potential factors we performed several sensitivity analyses. We applied different cut-off points and classifications, but the results remained similar. A potential bias relates to muscular men, who may be classified as overweight according to their BMI. However, this is less of a problem among men in late middle age.
The association of physical activity and weight with mental functioning is more equivocal. A previous study on the present data
 found that overweight was not associated with mental functioning. Some previous studies
[9, 35] have not considered weight, but they have shown that physical activity can be beneficial for mental functioning. A cross-sectional controlled trial
 found that physical activity was associated with better mood and functioning. Another cross-sectional study
 examining the joint association of physical activity and weight with mental functioning, found that physical activity is more important than weight for mental functioning. It also found that overweight inactive people are at greatest risk of poor mental functioning. This is in accordance with our study, which showed only minor differences in mental functioning between the highly active and the moderately active participants, with the overweight inactive participants being most likely to show poor mental functioning. Another previous study
 suggested that even small amounts of physical activity can improve mental functioning. The study
 also showed that normal weight is associated with better mood and functioning. We also found adverse effects of overweight on mental functioning among the inactive.
Except for baseline functioning, none of our covariates had a substantial effect on the associations between weight, physical activity and subsequent functioning. In addition to health behaviours and socio-demographics, we controlled for limiting longstanding illnesses and common mental disorders. We also conducted sensitivity analyses adjusting for the overall quantity of alcohol used. These covariates had negligible effects on the results (data not shown).
Several further sensitivity analyses were conducted. We used different cut-off points for the physical activity and weight groups. For BMI, 27 kg/m2 and 30 kg/m2 were used as cut-off points. For physical activity, 4 and 14 MET-hours per week were used. We also used different measures for physical functioning, such as the physical functioning (PF) subscale of SF-36, mean values for physical and mental functioning component summaries, and both the lowest quintile and highest quartile of functioning. These analyses did not substantially affect the results reported. Gender stratified analyses were also conducted. However, pooled data were used for the main analyses. Additionally we analysed the effects of vigorous exercise on functioning, as suggested by previous research
[13, 36]. This was done by analysing both the effects of the amount of physical activity and the intensity of the activity, assessed by practising vigorous physical activity or not. There were practically no differences in the effects of a large amount of physical activity compared with practising vigorous activity or not.
The strengths of this study include the prospective design, the large sample of middle-aged women and men originally employed, and the high participation rate at follow-up. Other strengths are the joint assessment of leisure-time physical activity and weight with functioning, the consideration of several covariates and the several sensitivity analyses. Additionally, identical measures for physical and mental functioning, both at baseline and follow-up, were available. The limitations include self-reported measures. These might cause overestimation of physical activity, overestimation of height, and underestimation of weight
. However, it has been shown
 that self-reported and measured weight and height predict health outcomes broadly in a similar way. BMI is not fully accurate as a measure of overweight, because it cannot distinguish between fat and lean mass. However, the accuracy of BMI has been found sufficient for epidemiological studies
. SF-36 is established as a reliable measure of physical and mental functioning