Intermittent fasting (IF for short) is an eating pattern which involves alternating periods of little or no energy intake (i.e., caloric restriction) with intervening periods of normal food intake, on a recurring basis.
One form of intermittent fasting is time restricted feeding (TRF), also called time restricted eating (TRE). With this type of intermittent fasting, one would fast every day for a period of 16–20 hours, and then eat during the remaining 4–8 hours.
In this post, we look at a brand new study (1) on TRF that was published just last week in JAMA Internal Medicine by a team of researchers from the University of California, San Fransisco.
The study was a randomized controlled trial which compared the effects of TRF, relative to a more conventional meal timing pattern (i.e., 3 square meals per day), on body weight and composition, and on several metabolic markers.
Side note: we have also done a review of a study comparing 5:2 intermittent fasting to traditional dieting, if you want to check that out as well.
What the researchers did
Subjects and dietary approaches
The researchers recruited 116 men and women aged between 18 and 64 years old, all with overweight or obesity, and randomly assigned them using block randomization based on gender, age, and BMI to one of two groups:
1. the time restricted feeding (TRF) group: subjects were instructed to eat ad libitum (i.e., as much as they wanted) from 12:00 pm until 8:00 pm, and completely abstain from caloric intake from 8:00 pm until 12:00 pm the following day (i.e., 16 hours fasting and 8 hours eating per day). During fasting times, only non-caloric beverages were permitted.
2. the consistent meal timing (CMT) group: subjects were instructed to eat 3 structured meals per day (i.e., breakfast, lunch, and dinner). Snacking between meals was allowed.
It’s important to note that the study intervention only included recommendations regarding the timing of food intake, with no recommendations given for calorie and macronutrient intakes, food quality, or physical activity.
The total duration of the intervention was 12 weeks.
The primary outcome was the differences between groups in changes in body weight from before to after the study. Each participant recorded their body weight every morning at their home using a Bluetooth scale supplied by the researchers, which would then automatically send the body weight data to the research facility.
Secondary outcomes included changes in body composition (e.g., lean body mass and fat mass assessed using dual-energy X-ray absorptiometry), and other metabolic markers, such as fasting insulin, fasting glucose, HbA1C, HOMA-IR, total energy expenditure, and resting energy expenditure. These were assessed by the researchers in a subset of participants (n=46) who lived within 60 miles of the research facility.
After 12 weeks, both groups lost some weight, with the time restricted feeding group experiencing a statistically significant reduction of 2 pounds (0.94 kg), and the conventional meal timing group experiencing a statistically non-significant reduction of 1.5 pounds (0.68 kg).
Similarly, both groups experienced a reduction in percentage body weight, with the reductions for this outcome reaching statistical significance for both groups.
Importantly, although the time restricted feeding group lost slightly more weight and was the only group for which the reduction in absolute body weight was statistically significant, there were no statistically significant differences between the groups for changes in this outcome.
This is an important point to make, as the primary outcome of interest is the change in body weight with TRF as compared to the control group, not the within-group changes from baseline.
Another thing worth keeping in mind is that the results above (and those of all other outcomes) are group averages. This means that individual responses vary, which you can clearly see in the figure below.
Note how, in both groups, some people lost a lot of weight, others lost less, a few experienced almost no changes at all, and several participants even gained weight.
Neither group experienced significant reductions in fat mass from before to after the intervention, and no statistically significant differences between groups were found.
However, there was a significant decrease in lean mass only in the time restricted feeding group with no statistically significant differences between groups.
Resting and total energy expenditure
There were no significant changes in resting energy expenditure after 12 weeks in either group, and there were no significant differences between groups for this outcome.
In terms of total energy expenditure, statistically significant reductions were detected in both groups (TRF: −177.9 kcal/d; CMT: −127.3 kcal/d) which are generally to be expected with weight loss, with no statistically significant differences between groups.
Blood lipids, glucose, insulin, and other health markers
Although intermittent fasting seems to have become synonymous with health improvements in recent years, this trial did not find any health benefits with time restricted feeding.
More specifically, according to the results, there were no significant improvements over time or differences between the groups in fasting glucose, fasting insulin, HOMA-IR (a marker of insulin resistance), HbA1C (a marker of prediabetes), triglycerides, total cholesterol, LDL, or HDL levels.
The study’s strengths
Overall, the study was well-conducted, with some strong points worth mentioning:
1. It was pre-registered (2): When a study is pre-registered, the researchers explain what they will attempt to investigate and how they will go about it before beginning to collect data. Doing so means that they commit to a pre-specified protocol, so we can be fairly confident that they did not simply “fish” for statistically significant results by doing lots of tests, and that they reported all their findings (even those that didn’t reach statistical significance).
2. It had a control group: Sometimes, weight loss studies test a hypothesis without including a control group. This is problematic in that, often, people lose weight by simply being included in a weight loss study, because they know they are being observed by the researchers. As such, it’s important to include a control group that’s also part of the experiment so that we have something to compare the intervention to.
3. It used block randomization: Block randomization is a randomization method which ensures that, when participants are assigned to the intervention and control groups, they are balanced for critical baseline characteristics that may affect the results (3). In this case, gender, age, and BMI were used.
4. It corrected for multiple comparisons: In trials with multiple outcomes (such as this trial), there is an increased chance of detecting a false positive result due to the fact that more tests are performed (4).
Think of it like this: if you roll a die once, you have a 1 in 6 chance of getting a certain pre-chosen number. However, if you keep rolling that die, the chances of rolling that pre-chosen number increase simply because you are doing more “testing”. It’s the same with clinical trials. More testing equals a higher probability of getting a false positive. In this study, the researchers used a Bonferroni correction, which is a way to control for the issue of multiple comparisons. Most intermittent fasting trials I have read don’t correct for multiple comparisons, which means that some of the results from other trials may have been false positives, especially when those results are from the analyses of secondary outcomes.
5. It was sufficiently powered: Many times, scientific trials may not have enough subjects per group in order to be able to detect a statistically significant effect in the main outcome. In this trial, researchers conducted a power calculation to determine how many participants they had to recruit for the study to be sufficiently powered, and then made sure to recruit enough participants to account for potential dropouts.
6. It was a “real world” intervention: The fact that the study only gave participants recommendations about meal timing, and not about things like food quality, calorie intake, or macronutrient ratios) is important in that it mimics how most people follow time restricted feeding in real life (i.e., by simply restricting their eating window to a specific time period within the day).
The study’s limitations
All studies have limitations, and this study had a few too. These are the main ones:
1. Food intake was not reported: While often misreported (5), it’s still useful to have calorie and macronutrient intakes reported, as these may provide insight into the reasons for some of the findings. For example, the reductions in lean mass in the TRF group may have been the result of a decrease in protein intake. If the researchers had collected food intake data from both groups, we would be able to compare the protein intakes between the groups and see if this may have been the case.
2. Body composition was assessed using 3-compartment DEXA: The DEXA 3-compartment model, although useful and widely used to assess changes in body composition, has its problems. Being a 3-compartment model, it estimates bone mass, lean mass, and fat mass. As such, DEXA cannot differentiate between muscle mass, and other “lean mass” (such as skin, organs, connective tissues, and body water) so the reductions in lean mass in the TRF group don’t necessarily equal a reduction in muscle mass. More likely, the lean mass reductions are a combination of reductions in muscle mass, other tissues, and body water.
3. The study may have been underpowered for secondary outcomes: Although a power calculation was made for the primary outcome (i.e., changes in body weight), power calculations were not conducted for secondary outcomes. This means that the trial may have been underpowered to detect changes in secondary outcomes (especially since these were analyzed in a subgroup of just 46 participants), so some of them may have been false negatives. For example, while no significant differences between groups were detected for changes in lean mass, the test for this outcome approached statistical significance (p=0.09), and may have reached statistical significance if the secondary outcome analyses included more participants.
4. The study used “delayed” time restricted feeding: Two subtypes of time restricted feeding include “early” and “delayed” TRF. The former refers to having your eating window earlier in the day (e.g., between 8 a.m. and 4 p.m.), and the latter refers to having your eating window later in the day (e.g., between 12 p.m. and 9 p.m.). According to some research (6), restricting your your eating window to the earlier part of the day may have a slight advantage in terms of some metabolic health marker improvements.
Summary and take-away points
This is one of the more well-conducted trials I have seen that examined the effects of time restricted feeding on body weight, body composition, and metabolic health markers, so I think we can be fairly confident in its results.
Remember that one well-conducted trial provides more valuable results than several poorly-conducted trials.
With the above said, here are the most important take-away points from this study, in my opinion:
1. In free-living conditions, on average, time restricted feeding works similarly well to a more conventional meal timing schedule of 3 square meals per day.
2. Individual responses vary. Not everyone will lose weight with time restricted feeding (or any other meal timing approach). Some people will even gain weight.
3. Simply being “observed” by researchers tends to cause people to make dietary changes that result in a small degree weight loss.
4. Time restricted feeding may result in lean body mass loss. However, lean mass isn’t necessarily muscle mass, as it includes other things, such as connective tissues, organs, skin, and body water, which 3-compartment DEXA cannot differentiate between.
5. The reduction in lean body mass in the TRF group may have been due to a reduction in protein intake. While this isn’t clear from the current trial as protein intake was not assessed, some research (7) suggests that eating enough protein in a smaller feeding window during the day doesn’t cause lean mass loss.
6. Time restricted feeding doesn’t seem to protect against the reductions in energy expenditure that are generally observed with weight loss (8).
7. While often touted as a way to improve several metabolic health markers (such as insulin resistance and blood lipids), time restricted feeding did not seem to have these effects in this trial. As the results from the available experimental trials in humans have generally been mixed, and due to some important methodological strengths of this trial (e.g., pre-registration and correction for multiple comparisons), I am inclined to call this hypothesis a null for now.
Found this article informative and helpful? If you have, remember to share it with your friends!