Part II: Systematic Review of Weight Gain Correlates In Literature

Inactivity and Weight Gain

Boston Public Library: Flickr.com

Boston Public Library: Flickr.com

I am addressing the topic of weight gain and obesity in the scientific literature yet again. The first installment in this three part series is available to read here: Part I. This entry in the series will look at inactivity and its correlation to weight gain.

I will remind everyone again of the strange contradiction that increased food intake and lowered activity levels do not correlate with weight gains, when obviously in recovery from an eating disorder, a patient can only restore weight and health if she increases food intake levels and lowers activity levels.

The discrepancy lies in the fact that an energy-balanced individual has no need of any additional energy, whereas an energy-deficient individual has need of additional energy until such time as he or she is energy-balanced. The body handles excess energy very differently from how it handles needed energy. Primarily an energy-balanced body simply burns away excess energy and lowers subsequent energy demand to maintain homeostasis. Carefully note however that that is done unconsciously— managing your weight consciously will harm that energy balance.

We will now look at systematic review data in the research literature regarding inactivity searching for their correlation to either weight gain or obesity prevalence.

Systematic reviews are at the top of the hierarchy of data, as I explain in Orthorexia Nevosa II: Doubt and Certainty. Wherever possible results gleaned from systematic reviews, or systematic reviews specifically provided by the Cochrane Collaboration, give us our best shot at having evidence that can shape the foundation of clinical practice. But, our best shot is not a sure thing, especially when the data don’t match the conclusions offered in those systematic reviews.

Systematic Review: Inactivity and Weight Gain

The first systematic review on the topic looks at whether activity prevents weight gain. The conclusion states that most studies found an inverse association between physical activity and long term weight gain. That means that the more activity a person applied, the less likely she was to gain weight. 1

Yet again we’ll need to trudge through the actual data to see if that conclusion is solidly supported.

The 16 prospective and observational studies included in the review all applied a retrospective questionnaire to assess the habitual (usually for the past year) level of physical activity.

That these studies all depended upon retrospective questionnaires render the results all but useless, but let’s move along to see what they offer and just pretend the results are sound for a moment.

Frustratingly the researchers slice and dice the 16 studies in a way that seems designed to hinder definitive analysis. It appears as though 10 of the 16 studies offered baseline physical activity data and the results were inconclusive.

Three studies showed inverse relationship between baseline physical activity and weight change (meaning a large volume of physical activity resulted in a smaller weight change). One study showed the relationship was present for only men and not women. Another showed work-related physical activity correlated with less weight gain. Two studies showed that vigorous physical activity at baseline was associated with greater weight gain. Three studies found no significance between baseline physical activity, or TV watching, and the magnitude of weight change.

Conversely, four studies providing follow-up physical activity found that a large volume of physical activity or exercise at follow-up was associated with less weight gain. One study found no such association at follow-up.

When reviewing both baseline and follow-up data, seven studies found that an increase in physical activity was associated with less weight gain. Two studies found no association. One study found that increased but not decreased physical activity was associated with weight change. Another study found no association between the cumulative duration of increased physical activity (annual recording) and weight change.

Confused? You should be. Here are the two important takeaways: 1) everyone gains weight no matter the physical activity level or change in level. 2) less than 50% of the studies reviewed actually support the conclusion that an increase in physical activity results in less weight gain.

The researchers then turn their attention to 19 non-randomized weight reduction studies with observational follow-up. 12 of those studies confirm that a large amount of physical activity at follow-up measurement was associated with less weight regain after a weight reduction. And just so we’re clear on that— everyone regains the weight they lost and those who apply a large amount of physical activity gain at a slower rate. And these studies also depended upon retrospective questionnaires or interviews to assess physical activity.

A further eight studies were actual randomized weight reduction interventions with passive follow-up. These studies give a much better sense of what the large amount of physical activity actually does in terms of lessening the rate of weight regain.

The difference between the exercise and non-exercise groups over a 20-month follow-up period after a weight-loss endeavor: 1.2 kg. (2.6 lbs.).

Although this doesn’t appear in the abstract synopsis, I believe the researchers clearly state their bias in their recommendations section:

Despite the fact that we do not find the association between physical activity and weight maintenance to be consistent, we would recommend at least lifestyle activity as one part of a healthy lifestyle.” 2

That’s perhaps equivalent to saying: “Even though we find clear associations between smoking and serious negative health outcomes, we would recommend at least lifestyle smoking as one part of appearing socially acceptable.”

Although this systematic review is addressing activity and weight maintenance, we can see from the data that slightly more than 50% of the 16 prospective and observational studies indicate that inactivity or reduced physical activity have no definitive link to increased weight gain. The non-randomized and randomized data are telling a completely different story altogether because these were all concerted efforts to first lose weight and then maintain that weight loss and that doesn’t really give us any definitive data on activity levels and weight maintenance. The body protects its optimal weight set point and that is why significant weight loss is maintained for approximately 0.003% of those undertaking weight loss, when measured five years out from the point of initial weight loss.

Inactivity and Mortality

But what about inactivity and health outcomes you might ask? Let’s take a look.

The systematic review on this topic extracted 19 high-quality studies for analysis. The researchers clearly identify how they define strong, moderate and insufficient evidence in the review, which I greatly appreciate. However, I would suggest that their marker for strong evidence is not sufficiently rigorous. They indicate strong evidence is present when they identify consistent findings in high-quality data sourced from two or more studies. I would be much more confident in the conclusions drawn if the bar were raised to require that strong evidence rest on consistent findings sourced at or above 50% of the studies providing data on the topic in question.

The researchers confirm what we already know from the previous systematic review mentioned above: “Based on the inconsistent findings among the prospective studies identified, there is insufficient evidence for a longitudinal relationship between sedentary behavior and body weight/BMI gain.” 3

They also found insufficient evidence for sedentary behavior creating a risk for overweight or obesity or waist gain.

They identified two low-quality studies that suggested sedentary behaviors result in an increased risk for type II diabetes mellitus onset. Using their criterion, they classify these studies as offering moderate evidence of correlation.

When I review the actual two studies they reference, the actual risk is as follows:

Study one involved 37,918 men in a prospective study following them for 10 years. In that time frame 1058 developed type II diabetes mellitus. If you were a man in that group, you had a 97.2% chance of not developing type II diabetes mellitus over that 10-year span. Even if you watched greater than 40 hours of TV a week (as per the framework of the study) you had a 92% chance of not developing type II diabetes mellitus over that 10-year span. 4

Study two involved 68, 497 women in a prospective study following them for six years. In that time frame 1515 developed type II diabetes. Participants had a 97.8% chance of not developing type II diabetes in that time frame. Pretty much the same data are present here for women as men when it comes to TV watching as a marker for inactivity and risk of diabetes onset. The researchers additionally estimated approximately 60% of the cases of diabetes onset seen within this study would not have been be preventable had the subjects adopted a relatively active lifestyle. 5

The studies confirm that the huge majority of men and women, regardless of levels of activity, do not develop type II diabetes mellitus.

Getting back to our systematic review, we find the researchers confirm there is insufficient evidence of any correlation for cardiovascular disease onset and inactivity. They also found insufficient evidence of correlation between the onset of endometrial cancer and inactivity.

Then we get to sedentary behavior and mortality for which the researchers suggest there is strong evidence of correlation between sedentary behavior and all cause mortality with the exception of cancer.

Here we go again. So the first low-quality study found no correlation between sedentary work and mortality. But presumably because it is of low quality, it should be overlooked.

Now let’s look at the high-quality study comprising 7278 men and 9735 women who filled out the Canada Fitness survey. The survey was conducted in person in 1981 and mean age at baseline was 42 years. The maximal follow-up period was 12.9 years during which time there were 951 male and 881 female deaths. There were a total of 759 cardiovascular deaths and 547 deaths from cancer.

According to the analysis, a sedentary lifestyle is correlated with all cause mortality with the exception of cancer. Other causes of death categorized within the study included respiratory diseases, 26%; injuries and violence, 24%; mental and nervous system disorders, 13%; digestive system disorders, 13%; and other, 14%. Just as an aside, I am curious as to how sitting correlates with death due to injuries and violence…they don’t expound on this in the published material I’m afraid.

The hazard mortality ratio of being maximally sedentary and dying of cardiovascular disease is 1.54. The hazard mortality ratio of being maximally sedentary and dying of all other causes (with the exception of cancer) is 2.15.

Just as a reminder, standardized mortality ratio is set at 1.00 and it reflects the expected mortality rate for a population at large. A hazard mortality ratio reflects the expected mortality rate within a subsection of the population (in this case, those being observed through the Canada Fitness survey). Therefore those who are most sedentary are expected to die at twice the expected rate.

And we all remember the standardized mortality ratio data for eating disorders— somewhere between 2.75 and 10.00, depending on the data you wish to use.

This high-quality study looks pretty damning when it comes to being sedentary and its impact on risk of dying doesn’t it? Hmm. Except for the fact that the initial baseline survey did not adequately rule out individuals already dealing with the diseases in question. The researchers in this trial used the PAR-Q (physical activity readiness questionnaire) to rule out individuals who might already be suffering from respiratory, cardiovascular, joint or degenerative illnesses. That will not have removed asymptomatic individuals and it depends too heavily upon the respondent accurately recalling and relaying any physical symptoms that might suggest the presence of these conditions.

And that likely means that the hazard mortality ratios are significantly inflated because the data were muddied by those who were already ill and were likely going to become increasingly sedentary as a direct result of their conditions progressing.

When you are sick, you tend to want to be sedentary and that means that correlation may be causative— the patient is highly sedentary because she is sick and not that she is sick because she is highly sedentary.

The final high-quality study referenced for the correlation of inactivity and all-cause mortality involved TV watching and all cause mortality. The study found no correlation between cancer or other non-cardiovascular deaths and TV watching. The hazard mortality ratio for TV watching and cardiovascular disease was 1.46 for TV watching of four or more hours per day. Within the six year study period the subjects had a 99.1% chance of being alive and well at the end of the trial. 6

That’s it folks. Out of 19 prospective studies under scrutiny in the systematic review here’s what the data show:

No correlation between inactivity and weight gain, waist gain, BMI increase or the prevalence of overweightness or obesity.

Questionable and unpersuasive correlation based on only two low-quality studies of inactivity and type II diabetes.

No correlation between inactivity and cardiovascular disease onset.

No correlation between inactivity and endometrial cancer, and all other cancers too.

Correlation of inactivity and all-cause mortality (minus cancer) is almost entirely resting on a study that failed to adequately remove the possibility that those with pre-existing conditions would skew correlation outcomes.

Conclusion

The obesity research arena is so foundationally mired in cultural bias where the presumed objective discipline of scientific research is not even superficially represented in opening statements or valid hypotheses, that the entire field is suspect when it comes to how much influence it has in developing “evidence-based” guidelines that are blatantly not supported by the actual data.

At this point in my three part-series, I have confirmed using systematic review and analysis that food intake does not correlate with overweight and obesity onset, or maintenance, and neither does inactivity.

The final part of this series will address the prevalence of the metabolically healthy obese.


1. Fogelholm, M., and K. KukkonenHarjula. "Does physical activity prevent weight gain–a systematic review." Obesity reviews 1, no. 2 (2000): 95-111.

2. ibid.

3.  Proper, Karin I., Amika S. Singh, Willem Van Mechelen, and Mai JM Chinapaw. "Sedentary behaviors and health outcomes among adults: a systematic review of prospective studies." American journal of preventive medicine 40, no. 2 (2011): 174-182.

4.  Hu, Frank B., Michael F. Leitzmann, Meir J. Stampfer, Graham A. Colditz, Walter C. Willett, and Eric B. Rimm. "Physical activity and television watching in relation to risk for type 2 diabetes mellitus in men." Archives of internal medicine 161, no. 12 (2001): 1542-1548.

5. Hu, Frank B., Tricia Y. Li, Graham A. Colditz, Walter C. Willett, and JoAnn E. Manson. "Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women." Jama 289, no. 14 (2003): 1785-1791.

6. Dunstan, D. W., E. L. M. Barr, G. N. Healy, J. Salmon, J. E. Shaw, Beverley Balkau, D. J. Magliano, A. J. Cameron, P. Z. Zimmet, and N. Owen. "Television viewing time and mortality the australian diabetes, obesity and lifestyle study (AusDiab)." Circulation 121, no. 3 (2010): 384-391.