Obesity Classes

Obesity is defined by class of obesity. BMI Calculator. From the NIH page (Classification of Overweight and Obesity by BMI, Waist Circumference, and Associated Disease Risks):

Disease Risk* Relative to Normal Weight and Waist Circumference
Men 102 cm (40 in) or less
Women 88 cm (35 in) or less
Men > 102 cm (40 in)
Women > 88 cm (35 in)
< 18.5
Very High
Very High
Very High
Extreme Obesity
40.0 +
Extremely High
Extremely High


* Disease risk for type 2 diabetes, hypertension, and CVD.
+ Increased waist circumference also can be a marker for increased risk, even in persons of normal weight.

Studying Research – Part 5 – Types of Studies

Here’s a helpful page on different types of studies (Literature Reviews: Types of Clinical Study Designs).

Some types studies are more helpful than others.

Correlational Studies

A good article on the subject (Correlational Studies – A Closer Look at Correlational Research).

The correlational method involves looking at relationships between two or more variables. While researchers can use correlations to see if a relationship exists, the variables themselves are not under the control of the researchers.

Another point is that while correlational research can reveal if a relationship exists between variables, this kind of research cannot prove that changes to one variable lead to changes to another variable. In other words, correlational studies cannot prove cause-and-effect relationships.

This paper (What Is Correlation?) is helpful in understanding the numbers.

Metabolic Ward Studies

Metabolic ward studies confine people to a controlled medical setting. Their intake is controlled and activity levels can be monitored.

Here’s an example of the typical set of controls for this kind of study (Carbohydrate Study Leaves Diet Researchers Divided – As expected, first study from Nutrition Science Initiative makes waves):

The food was prepared in a lab, frozen, then shipped to the study sites, where it was then prepared along with fresh produce. The baseline diet didn’t have a large amount of sugar, so it may have “differed somewhat from the customary diets” of the participants, wrote the authors.

The participants were not allowed to leave or to eat other foods, and were told they had to eat all of the food in front of them. They were allowed to meet with visitors, but only under the supervision of a nurse or research staff. The metabolic chamber measurement times were all at least 23 hours. Blood and urine measurements were also taken.

Food Macro Percentage Studies

Studies which vary the percentage of each macro (Fat, Protein, Carbs) and hold calories constant are also challenging due to the differing thermic effect of food (TEF). The thermic effect of food is the energy required for digestion, absorption, and disposal of ingested nutrients. It’s not possible to hold calories in constant and vary the macro mixture without affected calories out. And TEF even varies within a particular macro.  The magnitude of TEF depends on the composition of the food consumed:

  • Carbohydrates: 5 to 15% of the energy consumed
  • Protein: 20 to 35%
  • Fats: at most 5 to 15 %

Limits of Observational Epidemiology

Here’s a good article (Commentary: The hormone replacement–coronary heart disease conundrum: is this the death of observational epidemiology?
Debbie A Lawlor, George Davey Smith, Shah Ebrahim. International Journal of Epidemiology, Volume 33, Issue 3, 1 June 2004, Pages 464–467) which looks at the limits of these sorts of studies.




Studying Research – Part 4 – Methodological Problems

There are a lot of methodological problems in nutritional research.

Inaccurate Reporting – 24-h recall

Twenty-four hour food recall consists of interviews of people to remember what they ate in the past 24 hours. These sorts of surveys are known to be inaccurate. From (Dietary methods research in the third National Health and Nutrition Examination Survey: underreporting of energy intake. R R Briefel, C T Sempos, M A McDowell, S Chien, K Alaimo. The American Journal of Clinical Nutrition, Volume 65, Issue 4, 1 April 1997, Pages 1203S–1209S) demonstrates the inaccuracy in food logs:

Assessment of diet is a critical component of the third National Health and Nutrition Examination Survey (NHANES III), which was designed to describe the health and nutritional status of the US population. We analyzed data collected with the primary dietary assessment instrument used in NHANES III, the 24-h recall, for 7769 nonpregnant adults aged > or = 20 y to investigate underreporting of total energy intake. Underreporting was addressed by computing a ratio of energy intake (EI) to estimated basal metabolic rate (BMRest). EI:BMRest was 1.47 for men and 1.26 for nonpregnant women; a population level of 1.55 is expected for a sedentary population. About 18% of the men and 28% of the women were classified as underreporters. Underreporting of energy intake was highest in women and persons who were older, overweight, or trying to lose weight. Underreporting varied according to smoking status, level of education, physical activity, and the day of the week the 24-h recall covered. Additionally, underreporting was associated with diets lower in fat (P < 0.01) and alcohol (P < 0.01 in women) when expressed as a percentage of total energy intake.

From the text of the study:

About 10% of men and 8% of women reported that their
intake for the recall day was much less than usual, whereas
3% reported that it was much more.

A unique study provides evidence of the same (Int J Obes (Lond). 2010 Mar;34(3):437-45. doi: 10.1038/ijo.2009.251. Epub 2009 Dec 15. Inaccuracies in food and physical activity diaries of obese subjects: complementary evidence from doubly labeled water and co-twin assessments. Pietiläinen KH1, Korkeila M, Bogl LH, Westerterp KR, Yki-Järvinen H, Kaprio J, Rissanen A.).


Non-obese co-twins consistently reported that their obese twin siblings ate more food overall, consumed less healthy foods and exercised less than the non-obese co-twins do. However, no differences in energy intake (9.6+/-1.0 MJ per day vs 9.8+/-1.1 MJ per day, respectively) in the food diaries or in the mean PA level (1.74+/-0.02 vs 1.79+/-0.04, respectively) in the PA diaries were found between obese and non-obese co-twins. A considerable underreporting of energy intake (3.2+/-1.1 MJ per day, P=0.036) and overreporting of PA (1.8+/-0.8 MJ per day, P=0.049) was observed in the obese, but not in the non-obese co-twins.

Limiting Variables

Good studies limit the variables to the smallest extend possible. Ideally, they’d only change the one subject of interest. For instance, there are studies comparing ad lib low carb to calorie restricted high carb diets. The two variables (low vs high carb) and (ad lib vs calorie restricted) are then confounded in the results. Did the Low Cab diet do better because it was lower carb or because people could eat as much as they want? That particular study showed that that people doing the low carb diet actually ate less calories than the high carb diet people.

Counselling and Food Surveys

Other studies provide counselling and then later ask the people what they ate. Is this testing the effects of a counselling session and people’s compliance? Or is it testing the dietary effect? There’s ways like measuring Respiratory Exchange Rate (RER) to determine the proportion of carbs to fat that the person ate in the last day or so. Without this sort of verification how can the researchers know that the person complied?

These might be a better study of how ineffective a short training session is on getting people to change their dietary habits.

Studying Research – Part 5 – Types of Studies


Studying Research – Part 3 – Study Details

We looked at the summary of a study in Part 2. In this part we will look at the details of this particular study (Annals of Internal Medicine Logo, 18 MAY 2004. A Low-Carbohydrate, Ketogenic Diet versus a Low-Fat Diet To Treat Obesity and Hyperlipidemia: A Randomized, Controlled Trial.
William S. Yancy Jr., MD, MHS; Maren K. Olsen, PhD; John R. Guyton, MD; Ronna P. Bakst, RD; Eric C. Westman, MD, MHS).

The reason for this particular study is further explained.

However, until recently, available data on low-carbohydrate diets came from small studies of short duration, most of which were uncontrolled.

Also our speculation of the alternate Low Fat (LF) diet was based on common recommendations at the time for a diet:

… low-fat, low-cholesterol, reduced-calorie diet commonly used to induce weight loss and decrease serum lipid levels

The participant requirements were people who were fairly messed up but had not been medicated for their health conditions. The participants were selected if they had:

  •  elevated lipid levels (total cholesterol level > 5.17 mmol/L [>200 mg/dL]
  • low-density lipoprotein [LDL] cholesterol level > 3.36 mmol/L [>130 mg/dL]
  • OR triglyceride level > 2.26 mmol/L [200 mg/dL]),

LC Group Diet

Unlike the HG group which was on a caloric restriction, the LC Group was allowed to each as much as they wanted but with certain restrictions:

Participants were permitted unlimited amounts of animal foods (meat, fowl, fish, and shellfish), unlimited eggs, 4 oz of hard cheese, 2 cups of salad vegetables (such as lettuce, spinach, or celery), and 1 cup of low-carbohydrate vegetables (such as broccoli, cauliflower, or squash) daily. Participants were encouraged to drink 6 to 8 glasses of water daily. When participants were halfway to their goal body weight (determined at the week 10 visit with assistance from research personnel), they were advised to add approximately 5 g of carbohydrates to their daily intake each week until they reached a level at which body weight was maintained.

Supplements described

To simulate the practice of the study sponsor, the low-carbohydrate diet group also received daily nutritional supplements (multivitamin, essential oils, diet formulation, and chromium picolinate…

Controlling the number of variables in an experiment is crucial. This list is problematic since it just increases the number of variables. Perhaps one of the items in that group caused the difference? Any why couldn’t most if not all of those items be given to both group?

Completion Rates

The LF group was on a calorie restricted regiment which may explain the poorer completion rate compared to the LC group. The LF…

recommended energy intake was 2.1 to 4.2 MJ (500 to 1000 kcal) less than the participant’s calculated energy intake for weight maintenance (body weight in pounds × 10)

So a typical 200 lb person was put on a 1500 calorie diet for 24 weeks and failed. Not much of a surprise to most of us. Interestingly, the LC group ended up eating less calories a day than the calorie restricted LF group:

The estimated daily energy intake was 6.14 ± 1.37 MJ (1461.0 ± 325.7 kcal) in the low-carbohydrate diet group and 6.31 ± 0.68 MJ (1502.0 ± 162.1 kcal) in the low-fat diet group.

Dietary Macros

The macros weren’t explicitly specified (just LC and LF goals) but the results were:

The low-carbohydrate diet group consumed a mean (±SD) of 29.5 ± 11.1 g of carbohydrates (8% of daily energy intake), 97.9 ± 24.3 g of protein (26% of daily energy intake), and 110.6 ± 27.3 g of fat (68% of daily energy intake) daily.

The low-fat diet group consumed 197.6 ± 34.2 g of carbohydrates (52% of daily energy intake), 70.5 ± 9.7 g of protein (19% of daily energy intake), and 48.9 ± 12.0 g of fat (29% of daily energy intake) daily.

So the LC group was eating quite a bit more protein than the LF group but the LF group was eating the sort of amount of protein that is the recommended amount.

Serum Lipid Levels

The LC results were spectacular. The only “negative” was increased LDL (Cholesterol and Low Carb Diets). In fact, the increase in LDL caused two participants to drop out over concerns about their LDL increase from their doctors. However, the number that really matters is the ratio of triglycerides to HDL which improved on both diets, but dramatically so on LC. Numbers over 3.7 are not good. Only LC reduced the number to a very good number.

Studying Research – Part 4 – Methodological Problems

Studying Research – Part 2 – Examining a Study

In Part 1 of this subject we took a look at using google scholar to find research papers.

Let’s take a closer look at the first result returned by the search. This was the search result:


This looks like a relevant study since it does focus on ketogenic diets.  Ketogenic diets are typically very low in carbohydrates. Low carb can be a much wider definition and in some cases can be much higher amount of carbohydrates.

Focus of the study?

This study compared Low-Carb and Low-Fat diets. That is a pretty specific focus but it’s even more narrow because the purpose was for treatment of obesity and out-of-bound blood test results.

Type of Study?

There are quite a few types of studies. Not all types of studies provide solid evidence. This particular study is a Randomized Controlled Trial (RCT). Another example of a type of study is a correlation study (see this article for the limits of correlation studies).

Looking at the page

The top of the returned page shows:

The publication is listed as Annals of Internal Medicine. The Date of the page is “18 May 2004” so the study is older. That doesn’t mean it is a good or bad study necessarily but there may be newer studies which add more data to the original. The original search engine result showed that there were over 900 citations of this particular study.

Authors List

The list of authors for the study includes quite a few people. One of them, Eric Westman is known in the Low Carb community as an advocate of the diet. It is unknown to me if Eric Westman was a supporter of the diet back in 2004.

The other authors seem unfamiliar but searching for research by them might lead to other related papers.

The body of the page

The rest of the page has a summary of the study as well as the actual study data which followed the summary.

Looking at the Summary Section

Background:Low-carbohydrate diets remain popular despite a paucity of scientific evidence on their effectiveness.

In 2004 that may have been true. It’s not true today so it would be dishonest for someone to claim (today) that there’s no scientific evidence of the effectiveness of a low carb diet.

Objective:To compare the effects of a low-carbohydrate, ketogenic diet program with those of a low-fat, low-cholesterol, reduced-calorie diet.

The objective seems on-point but raises questions which need to be answered by looking into the details. It is good that the diet is defined as a ketogenic diet in contrast to other Low Carb diets since that establishes the limit on the carbohydrate levels.

The comparison diet seems to match the standard recommended diet for weight loss which includes:

  • Low Fat
  • Low Cholesterol
  • Reduced Calories

This also raises the question of the amount of fat in the ketogenic diet, the cholesterol level and whether or not the ketogenic diet was reduced calories. There’s quite a few variables being manipulated on the “standard dietary recommendation” diet but many of them may be dependent (not independent variables).


Randomized, controlled trial.

RCTs are typically good science but it bears some examination to see just how randomized the participants were.


Outpatient research clinic.

An outpatient research clinic sees people regularly to check up on progress but people are allowed to go about their day-to-day lives. That’s often a good thing since it doesn’t cause people to live their lives in an unnatural setting.

However, a setting like a metabolic ward allows for other variables to be controlled since food intake is controlled and outputs (urine, etc) are measured and sampled.


120 overweight, hyperlipidemic volunteers from the community.

The number participants was 120 people. That seems like a decent size. They were volunteers so their motivations were unknown. It is not clear whether this particular group is at all representative of typical overweight individuals. They all did have blood test results which indicate that they had bad test results.


Low-carbohydrate diet (initially, <20 g of carbohydrate daily) plus nutritional supplementation, exercise recommendation, and group meetings, or low-fat diet (<30% energy from fat, <300 mg of cholesterol daily, and deficit of 500 to 1000 kcal/d) plus exercise recommendation and group meetings.

So here’s the real details of the diets.  The summary seems like it is comparing ketogenic levels of carbohydrates (less than 20g a day). Both grouns got exercise recommendations and group meetings. The first group got nutritional supplements and the second group didn’t.

Again there’s a whole lot of variables here as well as unmentioned items like the protein amounts for each group.

Measurements: Body weight, body composition, fasting serum lipid levels, and tolerability.

The metrics seem useful. They seem designed for the purpose of determining the effectiveness of the diet for lipid levels.

Parsing the Results:

Compliance of the LC group was good since more finished the study as a percentage.

A greater proportion of the low-carbohydrate diet group than the low-fat diet group completed the study (76% vs. 57%; P = 0.02).

Surprisingly Weight loss in the calorie controlled LF group was less than the LC group which didn’t mention any calorie restrictions. This begs the question of how well people in the LF group complied with the study. At between 500-1000 calorie deficit per day they LF group should have lost 1-2 lbs per week. However, the summary lists the percent lost not kg lost. The LC group lost nearly twice the weigh (as a percentage) than the LF group lost.

At 24 weeks, weight loss was greater in the low-carbohydrate diet group than in the low-fat diet group (mean change, −12.9% vs. −6.7%; P < 0.001).

Fat loss was also greater (and in proportion to weight loss). Unfortunately, the results here were in kg and not in percent so it’s hard to compare the two.

Patients in both groups lost substantially more fat mass (change, −9.4 kg with the low-carbohydrate diet vs. −4.8 kg with the low-fat diet) than fat-free mass (change, −3.3 kg vs. −2.4 kg, respectively).

The following numbers were the effects on blood lipids which were favorable to the LC diet and of greater magnitude than the weight loss differences. In fact, the LF diet increased the HDL levels which shows a bad result.

Compared with recipients of the low-fat diet, recipients of the low-carbohydrate diet had greater decreases in serum triglyceride levels (change, −0.84 mmol/L vs. −0.31 mmol/L [−74.2 mg/dL vs. −27.9 mg/dL]; P = 0.004) and greater increases in high-density lipoprotein cholesterol levels (0.14 mmol/L vs. −0.04 mmol/L [5.5 mg/dL vs. −1.6 mg/dL]; P < 0.001).

LDL was too close to call but there were some outliers with more in the LC group. This would be higher LDL values in some people which seems common in the LC diet (See our article – Cholesterol and Low Carb Diets) which explains this result in some people (including me) with LC diets.

Changes in low-density lipoprotein cholesterol level did not differ statistically (0.04 mmol/L [1.6 mg/dL] with the low-carbohydrate diet and −0.19 mmol/L [−7.4 mg/dL] with the low-fat diet; P = 0.2). Minor adverse effects were more frequent in the low-carbohydrate diet group.

All in all it looks like LC for the win but if LC came out bad (hasn’t happened yet in any study) we would look at why and report on it.

Studying Research – Part 3 – Study Details

Studying Research – Part 1 – Using google scholar

I use google scholar to locate research papers and studies. When you launch the site you see the familiar google search engine with the difference being that this particular search engine looks through published papers.

As an example, suppose you want to investigate “low carb diet”. Enter that into the search box and select search.

The search engine returns:

Let’s take a look at each section of the results.

How Many Pages?

The search engine indicates how many pages have the particular search term in them. In this case, there’s more than 29,000 pages which have the words low carb diet in them.

That might indicate that looking into a more specific term might be useful. Perhaps add other words like “low carb diet insulin” as an example.

Limiting the Time Range

On the left is a filter section which can be used to select date ranges. Citations can be included or not included. This can help limit to more recent research in the case where there’s a lot of papers.

The Studies

The center lists the title of the particular paper which presented the results of the study.

Looking closer at the first result:

Document Type

The page is an HTML document.

Title of the Study

The title shows the scope of the study. Sometimes the title is misleading. Most of the time the title is helpful.

Type of Study

This is really important. Not all studies are of equal value. The type of this study is a “randomized, controlled trial”. That’s usually a well designed study. However, you will need to look closer to see the study flaws or strengths.


The authors are listed in green along with the date of publication and the journal that the paper was published in.


The abstract of the study is typically listed next. When you click the link it often just shows the abstract but if you search again for the title you may find the full text. I’ve searched for the title and added PDF and had success in some cases at finding the actual paper. The bottom line lists 16 other versions. One of them may have the full paper (if this is only the abstract).

Cited by

For this study it shows that the results are very widely cited. This particular paper was cited in over 900 other papers.

Right Bar Side

The right side shows the document type.

Studying Research – Part 2 – Examining a Study