Oportunidades

2.3.1 Structured exercise and weight loss

There is considerable ambiguity regarding the effectiveness of exercise for weight management. This is not surprising considering the unhelpful messages portrayed in the media with eye catching headlines such as ‘Why exercise won’t make you thin’

(Time Magazine, 2009) and ‘How exercise can make you pile on the pounds’ (Daily Mail, 2015). The latter headline emanated from the damaging editorial by Malhotra et al. (2015) who refer to ‘the myth of physical inactivity and obesity’. These unhelpful messages from the main stream media and academics alike wrongly reinforce the public’s preference to avoid exercise by suggesting exercise is futile for weight loss.

They condone the largely sedentary lifestyle which is prevalent in most technologically developed countries. This is particularly damaging as there is good evidence that exercise, when carried out over long periods of time, does in fact produce weight loss (Donnelly et al., 2003, Jakicic et al., 2008). There is a dose-response effect; the more exercise carried out, the greater the weight loss. Furthermore, several reviews (Ballor and Keesey, 1991, Catenacci and Wyatt, 2007, Swift et al., 2014), including a

Cochrane review by Shaw et al. (2006), also support the beneficial effect of exercise on weight independent of diet.

2.3.2 Impact of free-living physical activity and sedentary behaviour on adiposity

The relationship between free-living sedentary and active behaviours and weight status has received greater attention particularly since the development of objective PA measurement devices. PA impacts on energy balance through multiple pathways, including increased total EE (Plasqui et al., 2013), improved appetite control (Hopkins and Blundell, 2016, Shook et al., 2015), and there is also evidence to suggest PA has a positive influence on RMR, perhaps due to greater FFM (Speakman and Selman, 2003). On the other hand, a negative association between SB and weight has been reported, however, this relationship is less consistent and questionnaires are often

used to quantify sedentary time (Biddle et al., 2010). Furthermore, TV viewing is often used as a proxy of SB, but TV viewing has been shown to only correlate weakly with overall sedentary time when measured using accelerometers. It has been suggested that SB impacts on weight status by displacing MVPA (Mansoubi et al., 2014) and by altering EI, for example, TV viewing has been associated with increased EI and

snacking (Bowman, 2006). However, this association may not be due to SB per se and could be a result of exposure to food related advertisements (Scully et al., 2009).

Since the development of objective measurement devices, large scale observational and prospective studies have begun to quantify PA and SB using accelerometer based activity monitors. One such study is the National Health and Nutrition Examination Survey (NHANES) in America, which examined the independent and combined associations of PA and SB with obesity. Between 2003 and 2006 Maher et al. (2013) collected PA and SB data for 5,546 adults using accelerometers (ActiGraph 7164) and TV viewing time was assessed with a questionnaire. Stature and weight were

measured by trained health technicians during a physical examination using standardised procedures and BMI was calculated from stature and weight. All

analyses were controlled for potential confounders such as age, ethnicity, EI, alcohol intake and smoking status. Low MVPA was consistently associated with higher risk of obesity regardless of the amount of SB (determined by both accelerometry and TV viewing questionnaire). A similar relationship has been reported when PA and SB were measured using questionnaires (Sugiyama et al., 2008). The relationship between SB and obesity varied depending on the way in which SB was measured. In men, higher TV viewing was associated with greater risk of obesity but there was no relationship in women. A positive association between TV viewing and adiposity (BMI and waist circumference) has previously been reported in another large scale national survey (Heinonen et al., 2013). Accelerometer derived SB was not associated with obesity in men or women. There was a greater risk of obesity when low MVPA was combined with high TV time compared with risk of obesity associated with low MVPA or high TV time alone. Interestingly, Healy et al. (2011b) reported a positive

association between accelerometer measured SB and obesity in the same sample of participants. However, Healy et al. (2012b) used waist circumference (WC) as a measure of adiposity compared to the use of BMI in the study by Maher et al. (2013).

These studies demonstrate the impact that measurement method (for both SB and adiposity) can have on the reported relationship between SB and obesity. In the same sample of participants, two dissimilar conclusions were drawn; one supporting a positive association between SB and obesity and the other showing no relationship.

In another study, 878 participants from two diabetes prevention programmes in the UK had their PA and SB measured objectively (ActiGraph GT3X) (Henson et al., 2013).

There was a positive relationship between total sedentary time and indices of adiposity

and the opposite was true for MVPA and adiposity. After statistically controlling for time spent in MVPA, the relationship between SB and adiposity was no longer significant. However, the relationship between MVPA and adiposity remained after controlling for SB. Similar relationships were reported in breast cancer survivors using data from the NHANES survey (Lynch et al., 2010) and in the International Physical activity and the Environment Network (IPEN) study (Van Dyck et al., 2015).

Interestingly, Healy et al. (2008c) reported the opposite in a sub-sample of the AusDiab 2005 cohort. The relationship between SB and WC was independent of MVPA, but the relationship between MVPA and adiposity was no longer significant after controlling for SB. A possible explanation could be the very low levels of MVPA in the study by Healy et al. compared to the other two studies.

MVPA is consistently beneficially associated with indices of adiposity (Healy et al., 2008c, Lynch et al., 2010, Murabito et al., 2015, Van Dyck et al., 2015). However, the relationship between SB and indices of adiposity is less consistent with some studies reporting a positive association with adiposity (Healy et al., 2008c, Lynch et al., 2010) and others reporting no relationship (McGuire and Ross, 2012, Smith et al., 2014, Van Dyck et al., 2015, Murabito et al., 2015). The inconsistent relationship between

adiposity and SB could be due to the way in which SB is operationally defined and measured. For example, when SB is defined by posture (AP) there is no association with either BMI or total adiposity (Smith et al., 2014). However, when SB is defined by activity intensity (SWA) there is a relationship with both BMI and adiposity (Scheers et al., 2012, Shook et al., 2015). Whether the relationships between SB and adiposity depends on the way SB is defined and measured requires further investigation.

Furthermore, whether the relationships among PA, SB and adiposity remain after statistically controlling for other intensities of activity remains equivocal and requires further examination.

Prospective cohort studies have examined the change in PA, SB and adiposity over time, using statistical models to examine whether change in sedentary and active behaviours predicts change in adiposity and whether change in adiposity predicts change in behaviour. Shook et al. (2015) found that those with low levels of PA at baseline gained the most FM over 12 months. Golubic et al. (2014) showed that MVPA and sedentary time both significantly predicted weight gain over 1 and 7 years. For example, a 1.5 hour reduction in sedentary time and a 16 minute increase in MVPA per day were associated with a 1.4 kg and 0.5 kg reduction in body mass over 1 year, respectively. Furthermore, when MVPA and SB were modelled as the outcome

variable and indices of adiposity the exposure variable there was a three times greater inverse association between adiposity and MVPA compared with when MVPA was the exposure variable. The magnitude of the relationship between adiposity and SB

remained the same as when SB was the exposure. These findings suggest adiposity is

an important determinant of decreased PA as well as increased adiposity being a consequence of decreased MVPA. Interestingly, Ekelund et al. (2008) found that sedentary time was not a significant predictor of indices of adiposity, but rather indices of adiposity predicted sedentary time over a period of 5.6 years.

These cross-sectional and prospective cohort studies suggest in general that SB is associated with higher adiposity and MVPA is associated with lower adiposity, although the former relationship is less consistent. Furthermore, the relationship between MVPA and weight status appears to be independent of SB, whereas controlling for MVPA in the relationship between SB and adiposity often nullifies the association. Prospective cohort studies suggest a bidirectional relationship can account for the link between sedentary and active behaviours and weight status; low levels of PA and high SB will favour weight gain. In turn, greater adiposity will lead to lower MVPA and higher SB. However, the evidence for this bidirectional relationship is speculative and is based on observational studies. Controlled trials provide evidence for the beneficial effects of exercise on weight, however, the same level of evidence does not exist for SB. In order to address the issue of causality, randomised controlled trials would need to be undertaken where SB is manipulated and change in adiposity is measured. However, such trials are unlikely due to the ethical issues related to an enforced increase in SB for long durations.

2.4 Relationship between energy intake and energy