Article Discussed: Early Time-Restricted Feeding Improves Insulin Sensitivity, Blood Pressure, and Oxidative Stress Even without Weight Loss in Men with Prediabetes
Cell Metabolism, Courtney Peterson et al., Pennington Biomedical Research Center, 2018
The One Rep
A previous study demonstrated that eating in an eight-hour window promotes weight loss by decreasing calories. The study that is the focus of this paper is a follow-up investigation that held calorie intake constant during 5 weeks of 6-hour early time restricted feeding. Subjects demonstrated improved insulin levels, insulin sensitivity, β cell responsiveness, blood pressure, and oxidative stress levels in men with prediabetes even without weight loss. This establishes a benefit intrinsic to eating earlier and less hours per day.
The Warm Up- For Everyone
What is Time Restricted Feeding (TRF)?
TRF is a relatively new form of intermittent fasting (IF), both of which are subtypes of restriction diets (see dietary overview page for differences between types of diets). Time restricted feeding is defined as eating in a 6 to10-hour window with smaller windows being preferred for larger health benefits. There is typically no restriction on calories or elimination of food types in studies done on this diet; only the timing of eating is restricted. However, there is typically an overall decrease in calories on a TRF diet because small snacks generally eaten outside of this window are cut out. This study specifically investigated early TRF (eTRF) with a 6-hour feeding window and dinner before 3pm while holding food intake and calories constant between groups. The purpose of this study was to investigate whether the beneficial effects of TRF are due to weight loss or if they are inherent to the diet itself.
How does the Standard American Diet (SAD) compare to TRF?
Based on a previous TRF study with a control group who ate on their normal schedule :
Bottom 10% participants ate 3.3 times a day, top 10% ate 10.6 times a day
Median time between meals was 3 hours 6 minutes
Only 25% of participants fasted for more than 7 hours
Participants ate 50% of their calories after 3:30PM
In general, calories consumed after 6:30PM were above maintenance calories and therefore contributed to weight gain
Compare these statistics to this study, where the eTRF (intervention group):
Ate 3 meals a day spread over 6 hours
Fasted for 18 hours
Last meal was at 3:00PM
What is the effect of TRF (previous study) and eTRF (this study) on weight and health?
Reduced eating duration by 4 hours 35 minutes
Decrease in calories by 20%
Decrease in weight by 3.3 kg (baseline 94.89kg)
Increased (10 pt scale): morning energy (∆1.5), overall energy (∆0.6), sleep satisfaction (∆1.5)
Decreased (10 pt scale): hunger at bedtime (∆1.25)
Results of this eTRF study, which specifically maintained calorie intake, demonstrated
eTRF decreased fasting insulin by 3.2 mU/L
eTRF increased insulin sensitivity
eTRF decreased morning blood pressure by 11/10 mmHg
Subjectively, eTRF reduced the desire and capacity to eat at night
Not eating for 16 hours seems like a long time. Will I lose muscle mass?
Maybe. A previous n=18 study with men who underwent resistance training on a control diet and TRF diet for eight weeks showed an increase of 2.3kg of muscle with the control diet verses a loss of 0.2 kg in the TRF group . However, there are several confounders to this study. Unlike our current study, this study was not calorie matched between groups; the TRF group consumed 650kcal less than the control group per day. Additionally, the control group ate 1.4 grams of protein per kg of body weight per day while the TRF group ate 1.0 g protein/kg. This is incredibly important as another study showed that a protein intake of 1.3-1.8 g/kg promotes lean mass retention during hypocaloric diets. That study promotes eating 0.25g protein per kg body weight per meal with a total of 1.5g protein per kg body weight per day .
Maximum Effort- For Enthusiasts
Intermittent fasting (IF), of which time restricted feeding is a subtype, has exploded in popularity over the past decade. Hard data regarding these diets began with very promising animal studies and evolved into a handful of human studies. One of the first studies of TRF in humans involved a very particular subtype of TRF known as One Meal A Day (OMAD). This study demonstrated poor results:
In an n=15 study, the intervention group of OMAD with the single meal eaten between 4:00PM to 8:00PM for 8 weeks kept weight stable with a decrease in body fat mass. However, when comparing one meal per day with three (the control), oral glucose insulin sensitivity decreased (403.4 vs 458.8 control) and fasting blood glucose (95.9 vs 85.4 control mg/dL), OGTT glucose, and total, LDL, and HDL cholesterol increased . There was no difference in HOMA-IR, adiponectin, resistin, leptin, ghrelin, or BDNF. This diet had 50% of its calories from carbohydrates.
Other studies fared better, including the previously mentioned TRF study by Dr. Satchin Panda that was not isocaloric between the control and TRF group . As reported above, the TRF intervention group in Dr. Panda’s study reduced eating duration by 4 hours 35 minutes with a concomitant decrease in calories by 20% and 3.3kg of weight loss. The last meal of these participants was between 6PM and 10PM. Other IF and TRF diet studies demonstrated better results with earlier eating times and null results with later eating times. These studies raise two interesting questions that are answered by this study:
What is the optimum eating window for a TRF diet? Answer: An early window that follows the circadian rhythm is optimal.
Is there an intrinsic health benefit to eating within a small window per day or are the health benefits of TRF solely due to weight loss? Answer: There is an intrinsic health benefit to eTRF that does not depend on weight loss.
For the first question, there are a few compelling physiological reasons why an early eating window is optimal:
Diet induced thermogenesis (DIT) is 44% lower in the evening . DIT is defined as “the increase in energy expenditure above basal fasting level” divided by the total energy intake of the day. There are three routes by which energy is expended in humans: DIT, basal metabolic rate, and energy cost of physical activity. DIT typically accounts for 5 to 15% of daily energy expenditure .
Postprandial glucose is 17% higher after the same meal in the evening versus morning. This effect is due to decreased beta cell function in the evening with subsequent 27% lower early-phase insulin .
Insulin sensitivity is impaired by 34% in the evening relative to the morning .
The second question is incredibly important as it changes what is recommended to those seeking a healthy diet:
If the effect of the TRF diet was solely due to weight loss, then any diet that causes weight loss will be just as efficacious and, therefore, just as recommended as TRF
However, if beneficial effects of TRF are seen without weight loss, then there is an intrinsic benefit to fasting 16-18 hours of the day that goes beyond the expected weight loss
One major caveat to IF and TRF is that there are no human IF dietary studies with primary endpoints of all-cause mortality, heart attack, or stroke. These are generally considered the gold standard of research outcomes as they show an actual, non-debatable outcome.
The study that is the subject of this paper was an n=8, 5-week, randomized, crossover, isocaloric controlled feeding trial testing eTRF in men with prediabetes. Subjects were randomized to either the eTRF then control arm (arm 1) or the control then eTRF arm (arm 2). There was a 7-week washout period between arms. The eTRF eating window was 6 hours with dinner before 3:00PM. The control group had an eating window of 12 hours. Food was matched per person between arms; that is, if a person in arm 2 ate a hamburger during the control phase, they were scheduled to eat a hamburger in the eTRF phase as well. All food was prepared in a research kitchen with a rotating menu. Macronutrient ratios were 50% carbohydrates, 35% fat, and 15% protein. Examples of meals include:
Breakfast- bagels with cream cheese; oatmeal with egg whites; waffles with maple syrup; honey nut cheerios with egg whites; or blueberry muffin with a boiled egg
Lunch- Whole wheat ham sandwich; grilled chicken sandwich; whole wheat turkey sandwich; whole wheat chicken salad pita sandwich; or chicken pesto pasta
Dinner- Alfredo pasta with chicken; catfish almondine; lemon sage chicken; spaghetti and meatballs; pork chops with potatoes
Magnitude of Effects and Discussion
Statistically Significant Findings
Many of these outcomes may seem unfamiliar. An in-depth discussion of them is beyond the scope of this paper, but they will be covered in a future review. For now, the parameters with asterisks are related to insulin sensitivity and β cell responsiveness, 8-Isoprostane is a measurement of oxidative stress and PYY is a hormone that promotes satiety. These data show that the eTRF has a considerable beneficial effect on insulin sensitivity, β cell responsiveness, blood pressure and oxidative stress. Triglycerides were significantly elevated in the eTRF group, most likely due to fat mobilization that occurs with fasting.
8-Isoprostane is created from free-radical nonenzymatic peroxidation of fatty acids, most typically arachidonic acid. The structure of 8-isoprostane is similar to prostaglandins (see figure, image credit to ). 8-isoprostane is a sensitive marker of oxidative stress in vivo . Generation of reactive oxygen species have a causal role in atherosclerosis in humans . In this eTRF study, the difference in 8-isoprostane levels were driven by a worsening in the control arm and stable levels in the eTRF arm, thus demonstrating that eTRF prevents oxidative stress that the base diet would have otherwise generated.
The following parameters did not demonstrate a statistically significant difference between the control and eTRF group: weight, fasting glucose, mean glucose, peak glucose, HOMA-IR, Heart rate, augmentation index, pulse wave velocity index, LDL-C, HDL-C, cortisol, hs-CRP, IL-6, active ghrelin, leptin, adiponectin, or GLP-1.
The authors note that the study design was poor for determining eTRF’s impact on blood glucose levels for two reasons: pre-measurement fasting times were not matched between the arms (18 hours of fasting for eTRF vs 12 hours for control) and only a morning blood glucose was measured. These are both unfortunate study design flaws. Even though morning fasting blood glucose was not statistically different, I suspect that the area under the curve of glucose levels throughout the day would be lower in the eTRF group if glucose was checked multiple times as the eTRF reaches the fasting state for longer periods of time than the control diet.
Additionally, remember that calories, not weight, was kept constant for this study. The fact that there was no weight loss difference between the arms could mean one of two things: eTRF does not increase energy expenditure versus a normal diet or this study was not powered to find a difference between the arms. We can calculate the expected change in weight assuming diet induced thermogenesis is 44% less effective in the evening. This calculation is shown in the appendix at the bottom of this article. In short, an n=8 study over 5 weeks is not powered to find this difference.
Biases and Effects
Sequence and Period Effects
Interestingly, subjects who completed the eTRF arm first entered the second arm (the control arm) with >25% lower mean postprandial insulin after the 7-week washout period. This effect is also known as a carryover effect, which, in an AB/ BA designed crossover trial, is synonymous with sequence effect. A carryover or sequence effect in this study can be due to a couple of reasons but I will highlight two: 1.) In the eTRF -> control arm, the washout period was not long enough to fully wash out the beneficial effects, which persisted and were measured in the control group erroneously. Or 2.) Subjects changed their habits in the washout phase after completing eTRF. Additionally, there were period effects for mean postprandial insulin. Period effects means that there was an intrinsic difference between the two periods. For example, both eTRF and control did better or worse in period 1 than in period 2. In general, this could be due to learning, fatigue, or timing factors such as period 1 being in the summer and period 2 being in the winter, which would cause differences in activity levels).
Time restricted feeding is a feasible, acceptable diet for losing weight. The main mechanism of weight loss in TRF is restriction of calories in a way that is easy for patients to understand and implement. The eating window of TRF stops the consumption of late-night snacks, which are notoriously unhealthy to begin with. While calorie counting can be tedious and inexact, eating by the hands of a clock is easily achievable. This study goes beyond these eating guidelines to elucidate the healthy effects of doing TRF with a morning eating window, demonstrating that there are effects on multiple organ systems that goes beyond simple weight loss.
 A Smartphone App Reveals Erratic Diurnal Eating Patterns in Humans that Can be Modulated for Health Benefits. Satchidananda Panda et al., Cell Metabolism Vol 22 Issue 5, 3 November 2015, 789-798
 Time-restricted feeding in young men performing resistance training: A randomized controlled trial. Grant Tinsley et al., European Journal of Sport Science Vol 17, Issue 2, 2017.
 A Brief Review of Higher Dietary Protein Diets in Weight Loss: A Focus on Athletes. Stuart Philips, Sports Medicine, Vol 44, November 2014, 149-153
 Impact of Reduced Meal Frequency Without Caloric Restriction on Glucose Regulation in Healthy, Normal-Weight Middle-Aged Men and Women. Olga Carlson et al., Metabolism. Vol 56, Issue 12, December 2007, 1729-1734
 The Human Circadian System Has a Dominating Role in Causing the Morning/ Evening Difference in Diet-Induced Thermogenesis. Christopher Morris et al., Obesity. Vol 23, Issue 10, 28 September 2015
 Diet Induced Thermogenesis. Klaas Westerterp. Nutr Metab (Lond). 2004;1:5
 Endogenous Circadian System and Circadian Misalignment Impact Glucose Tolerance Via Separate Mechanisms in Humans. Christopher Morris et al., PNAS April 28,2015 112(17)
 Circadian Regulation of Glucose, Lipid, and Energy Metabolism in Humans. Eleonora Poggiogalle et al., Metabolism Vol 84, July 2017, 11-27
 A series of prostaglandin F2-like compounds are produced in vivo in humans by a non-cyclooxygenase, free radical-catalyzed mechanism. Jason Morrow et al., Proc. Natl. Acad. Sci. Vol. 87, pp 9383-9387 Dec 1990
 Role of oxidative stress in atherosclerosis. David Harrison et al., The American Journal of Cardiology, Vol 91, Issue 3, 2003, 7-11
Power Analysis for DIT Weight Loss
Subjects were provided with 3,176 kcal on average given their average age and weight at baseline being 57 and 100kg, respectively, given by the following equation in the appendix
Kcal/day= 2189 +19.6*(weight in kg)- 17.6*(age in years)= 3,176 kcal/day
Assuming that DIT expends 15% of total calories, the eTRF group ate all of their calories in a time period when DIT was maximal (ie, not in the evening), and that only the dinner of the control group was affected by the 44% decrease in effectiveness of DIT, we have the amount of calories burned by DIT in the eTRF group given by
(total kcal/day)*(max DIT)=3,176*0.15= 476 kcal
And the amount of calories burned by DIT in the control group of
(2/3)*(total kcal/day)*(max DIT)+ (1/3)*(total kcal/day)*(max% DIT)*(44% reduction)=
318+80= 398 kcal
Therefore, the difference in energy burned by DIT between the eTRF and control group is 476-398= 78 kcal/day. This equals a total difference of 2,730 kcal over the course of the five weeks of the study. There is approximately 3,500 kcal per pound of fat, which in turn means that this difference in energy consumption would result in 0.78 lbs of fat loss for the eTRF group. Using the baseline weight of 100.7kg ± 18.4 and a power of 0.80, this study would have needed 17,500 subjects to detect this difference. Going further, extrapolating 2,730 kcal over five weeks to one year equates to 8.13 lbs fat loss per year.