An Objective Look at Intermittent Fasting
By Alan Aragon & Ryan Zielonka
Copyright © By Alan Aragon
August 8th, 2007
Over time, the ancient practice of fasting seems to periodically find its way into modern fitness subcultures. Despite its purported physiological and psychological benefits, scientific data is far from unanimously supporting it. Are there any physiological advantages to increasing or decreasing meal frequency? What are the downsides to intermittent fasting? What are the health advantages shared by fasting, exercise, and caloric restriction? Are there detrimental effects of resistance training in a fasted state? Anecdotal data is fraught with suggestion-driven bias, so this paper will focus on what's been demonstrated objectively in scientific research.
Meal Frequency - Hot Dogma & Cold Facts
Since the infancy of health-nut culture, it's been decreed that one must eat every 2-3 hours, thus minimizing muscle breakdown while keeping the metabolic engine revving. Tupperware and coolers in gyms and cars were (and still are) a common sight, perceived by the public as a sign of diligence and discipline. Alas, the long-held belief in a higher-is-better frequency has been examined with the lens of science, and some interesting data has surfaced.
Effects on Health & Thermogenesis
Research indicates that a haphazard meal frequency, not necessarily a lower frequency, negatively impacts thermogenesis, blood lipids, and insulin sensitivity [1,2]. Contrary to popular belief, a high frequency has no thermodynamic advantage over a low frequency under calorie-controlled conditions (as opposed to ad libitium or free-living conditions) using 24-hr indirect calorimetry [3,4]. So much for the magic of stoking the metabolic furnace with an extreme grazing pattern. It bears mentioning that lower 24-hour insulin levels as well as lower fasting and total LDL-cholesterol levels have been observed with higher meal frequencies [5,6]. However, in discovering this, studies have used unrealistic protocols for the higher frequency treatments, comparing 3 meals to 9 or 17 meals per day.
Effects on Body Composition
With scant exception [7,51], the majority of controlled intervention trials show no improvement in body composition with a higher meal frequency, with treatments ranging from 1 to 9 meals per day [8-12,17]. Unfortunately, a scarcity of research examining meal frequency's effect on body composition under conditions of exercise exists in the literature. To further confound the data, results from studies are mixed. The singular full-length published controlled trial showing a body composition benefit of higher meal frequency (6 versus 3) is limited by its poor study design. In the trial, boxers consumed a 1200 kcal liquid diet over a 2 week period . The 6-a-day group retained more lean mass than the 3-a-day group.
In contrast to the boxer study, a recent abstract presented at the 12th Annual Congress of the European College of Sport Science reported the superiority of 3 meals a day to 6 meals a day for gaining lean mass during a 12 week period involving strength training . Too bad the 3-a-day group also experienced a trend towards fat gain. This raises the possibility that the 3-a-day group simply ate more calories overall. This wouldn't be surprising, considering that comparative research shows an association of greater hunger with meal frequency reduction.
Effects on Appetite
Studies indicating the disappearance or lack of hunger in dieters occurs in either complete starvation, or very low calorie regimes (800 kcal/day or less) [13,14]. Much of this data is irrelevant for most of you reading this. Furthermore, none of those studies systematically measured appetite throughout waking hours. Convenience issues aside, for the purposes of controlling appetite, research indicates the superiority of a higher frequency over a lower one.
In two separate studies led by Speechly, both lean and obese subjects had greater appetite control when pre-test meals were consumed at frequent intervals, in contrast to the same amount of food consumed at a single meal [15,16]. In the ad libitium test meal that followed the pre-test meals, subjects given the single meal consumed on average 26.5% more calories.
Stote's team compared 1 versus 3 meals per day . Among other results, the 1-a-day group reported significantly higher levels of hunger and an increased desire to eat, with the severity of both phenomena increasing throughout the length of the trial. In a recent alternate-day fasting study , Heilbronn’s team concluded, “Overall, these results suggest that a prolonged schedule of fasting and feasting would be marred by aversive subjective states (eg, hunger and irritability), which would likely limit the ability of most individuals to sustain this eating pattern.”
In a recent review, Johnstone suggests that a hierarchy of hunger response exists (at least in obese subjects), whereby hunger directly correlates with the severity of the caloric deficit, citing fasting as the obvious limit of acute deficit . The persistence of hunger was one of the primary reasons he did not recommend fasting as the optimal dieting strategy, despite having been a principal investigator in fasting research. To quote the paper’s conclusion, “There is, however, the problem of elevated hunger during food restriction and this may provide too great a challenge to a ‘faster’ in not breaking compliance to the dieting regime and reaching for the biscuit barrel.”
Skipping Breakfast = Not Too Brilliant
Rampersaud's team examined the results of 47 studies on various effects of breakfast consumption among children and adolescents . Interestingly, while breakfast eaters consumed more daily calories, they were less likely to be overweight. Children who consistently ate breakfast tended to have superior nutritional profiles. This concurs with a cohesive body of data indicating that adults who eat breakfast meet their daily micronutrient needs better than habitual breakfast-skippers [20-22]. While cognitive effects are inconsistent in well-nourished children, breakfast skipping degrades mental performance in malnourished children. Overall, the evidence points to regular breakfast consumption improving cognitive function, test grades, school attendance, memory, and nutrient status. The latter effect pertains to macronutrients and essential vitamins and minerals. The impact of skipping breakfast on the intake of other functional nutrients hasn't been studied.
In a controlled intervention trial on lean subjects, Farshchi's team found that skipping breakfast decreased post-meal insulin sensitivity and increased LDL-cholesterol, despite a high (6-a-day) meal frequency . This data points to the possibility that the body is "metabolically primed" to eat a meal soon after an overnight fast.
Concurring with the above results, noted protein researcher Donald Layman asserted in a recent review that the most critical meal of the day is breakfast after an overnight fast . This is partially due to circadian protein synthesis rates being lowest at this time. He states that the anabolic impact of a meal lasts roughly 5-6 hours based on the rate of post-meal amino acid metabolism, therefore, dietary protein should be provided at approximately 5-hr intervals throughout the day. This recommendation can be challenged by the fact that other studies show longer durations of plasma glucose and amino acid elevations caused by casein or a mixed meal [25,26]. However, the latter research didn't measure the effect of exercise on plasma amino acid flux. In the final analysis, Layman's suggestions are a safe bet without any major convenience impingements.
In three separate controlled experiments, Benton and Parker examined the effect of breakfast versus fasting on cognition . In the first study, fasted subjects took significantly more time than the fed group to complete both the spatial memory task and the word recall. In the second study measuring information processing and short-term memory decay, the fasted group lacked the improvements shown in the breakfast group. In the final trial, memory and intelligence were measured. Although breakfast didn't enhance abstract thought, it was superior to fasting for recalling a story read aloud. The researchers concluded that these trials were in agreement with a substantial body of previous research demonstrating that breakfast benefits memory.
On the observational research front, the National Weight Control Registry (NWCR) is the largest ongoing study of individuals who have successfully maintained substantial weight loss over the long term. To qualify, participants must maintain a weight loss of at least 13.6 kg (30 lb) for at least one year. According to a formal analysis led by Wyatt , 2313 subjects (78%) eat breakfast every day. Only 114 subjects (4%) reported skipping breakfast. This obviously isn’t cause-and-effect data, but it shows the crucial commonalities in the habits of dieters with long-term success. Daily breakfast is clearly one of those habits.
Intermittent Fasting Human Research – Interesting But Inconclusive
Intermittent fasting (IF) can be any number of variations of feed/fast intervals. Alternate-day fasting (ADF) is simply defined by its name. As of this writing, the literature on ADF contains 3 human trials conducted within the last 2 years [18,29,30]. Control groups were absent in all of those studies. As such, no comparative conclusions can be drawn between ADF and linear caloric intake. Animal research has shown promise for the health effects of ADF. However, human research hasn’t quite lived up to the luck of rats. Nevertheless, the data still provides food for thought and further investigation.
Heilbronn's team put non-obese men and women on an ADF for a total of 22 days . Subjects lost an average of 2.1 kg total bodyweight despite instructions to eat double their typical day’s intake every other day. Men maintained normal glucose metabolism and improved insulin response. Impaired glucose tolerance occurred in women by the end of the trial. Although a trend toward increases in resistance to stress occurred in the study, both men and women showed no changes in gene expression involved with fatty acid oxidation.
In another 22-day ADF study led by Heilbronn, non-obese subjects lost an average of 2.5% of
initial bodyweight . Beneficial effects of ADF included a decrease in fasting insulin levels and respiratory quotient, indicating an average fat oxidation increase of roughly 15g per day. Unfortunately, there was an increase in hunger on the first fasting day, a condition that remained elevated for the duration of the trial.
Halberg's team examined the effect of an ADF for a total of 14 days on non-obese young men , and observed an increase in insulin sensitivity. In contrast to the previously discussed study, no change in bodyweight or bodyfat occurred. As a result of the ADF, insulin sensitivity and glucose uptake in muscle increased. However, there was also an increased sensitivity, or uptake-readiness in the fat cells, evidenced by an inhibition of insulin-mediated adipose tissue lipolysis. The next study we'll examine deserves its own section, but not for the reasons you might expect.
A "Controlled" Study?
A recent study led by Stote compared the effects of 1 meal with 3 meals per day, and was the first trial of its kind to control calories between randomly assigned groups . Curiously, the 1-a-day group lost slightly more bodyweight and bodyfat, and gained a small amount of lean mass. Given those results, it's heralded in some fitness circles as the long-awaited shred of research supporting Ramadan-style (12-16 hour daily) fasting for achieving the Holy Grail of body re-composition.
As always, things are never as simple as they seem once the details are exposed. A number of serious design flaws, including common ones such as a small sample size (15 participants completed the trial) and short duration (2 week lead-in, 6 week treatment periods) plague the quality of the conclusions. The withdrawal of 5 subjects was a 28% dropout rate, which raises the huge question of how different the results might have been if the participants hung in there. The authors noted this was anomalously high compared to the typical dropout rate from feeding studies at their facility, which is roughly 6-7%. Statistical outcomes can easily swing either direction due to individual differences in a small sample. Now let’s take a look at how the key limitation of this trial cripples its .
The most crucial short-sight of the investigation was its use of bioelectrical impedance analysis (BIA) for body composition assessment. In a previous study illustrating its inaccuracy for fasting protocols, Faintuch’s team used BIA to measure body composition changes in non-obese subjects undergoing a 42 day fast, consuming only water, vitamins, and electrolytes . BIA registered an unrealistic 32% decrease in bodyfat, and an overall gain in lean mass. The researchers themselves concluded that these impossible readings proved BIA an inappropriate method for this type of protocol. It’s baffling that Stote chose BIA out of all the available methods, given BIA’s woeful track record in fasting research, likely attributable to the aggressive water redistribution found in fasting patients. Since we don't have reliable readings of the study's critical endpoints, its results are basically worthless.
Enter Ramadan - Keep Your Eyes On The Road
Ramadan is considered by practicing Muslims to be the most important period of religious observation in the Islamic calendar. In its strictest version, a complete food and fluid fast is undergone from sunrise to sunset (12-16 hours). This routine is carried out daily for a month. Unsurprisingly, traffic accidents peak during this time, alongside a reduction in working hours . Traffic injuries are the second major cause of death in the United Arab Emirates, with the bulk of the accidents occurring between 8am and 2pm . Research consistently shows a decrease in daytime alertness, mood, and wakefulness during the fasting month of Ramadan [32-35]. Physical performance - speed, agility, and endurance declined in professional soccer players observing Ramadan . Their performance remained low for two weeks post-Ramadan.
Fasting & Exercise - Common Effects
On the Heels of Caloric Restriction
Caloric restriction (CR), defined as a sustained linear calorie reduction without malnutrition, has a substantive body of animal data supporting its benefits on a number of clinical endpoints. Human data is steadily emerging to validate it. [35-37]. Whether or not IF is as effective as CR in humans will be determined by further research, but it appears effective at least in improving HDL levels in women, and insulin sensitivity across the board. Whether IF continues to show detrimental effects on glucose tolerance in women remains to be seen. Improvements in insulin sensitivity, glucose tolerance (in men at least), bodyweight/bodyfat, blood pressure, blood lipids, and heart rate are commonly cited benefits of IF & CR. The question is, can exercise achieve the aforementioned cardiovascular/metabolic benefits without the inherent downsides of periodic food deprivation? The scientifically valid answer is yes [40,41].
One of the highlighted benefits of IF & CR is the ability to prevent aging symptoms of the brain and nervous system. Brain-derived neurotrophic factor (BDNF) is one of a family of brain-based proteins responsible for the survival and growth of neurons involved with memory and learning. Preventing a decline in BDNF can thus prevent and/or lessen the progression of neurodegenerative disorders. IF and CR have both been found to increase BDNF activity . However, few are aware of the fact that exercise has also been demonstrated to elevate BDNF , and the degree of effect appears to be intensity-dependent . In a recent example of this phenomenon, Winter's team found that in comparison to low impact aerobic running and a period of rest, vocabulary learning was 20% faster after high impact anaerobic sprints . Ironically, although fasting can have preventive effects on neurodegeneration, its track record in improving human cognition is bleak.
With all this great data on the common neuro-protective and cardio-protective benefits shared by fasting and exercise, why not combine the two and train in a fasted state? Fasted cardio-respiratory training research has been covered elsewhere . The next section will discuss the effects of fasted resistance training.
Fasted Resistance Training = Not Optimal
Regardless of the inconsistency of performance data on fed versus fasted subjects, the combination of fasting and resistance training has never been a good idea from the standpoint of optimizing protein synthesis and inhibiting protein breakdown.
Recent research by Baty's team showed no resistance training performance benefit of a protein-carb solution taken pre, during, and postworkout . However, two indicators of muscle damage were elevated in the fasted training placebo group. Their myoglobin levels approached significance halfway through the exercise bout, and were significantly elevated 6 hours postexercise. Creatine kinase levels were also significantly elevated 24 hours postexercise.
Tipton's team compared the effect of an immediate pre-resistance training dose of essential amino acids + carbohydrate (EAA+CHO) with the same treatment immediately postworkout . 262% more amino acid uptake was seen in the pre-group compared to the post-group. In a subsequent study, Tipton used a similar protocol with 20g whey protein only, administered either immediately pre- or immediately postworkout . Although no significant differences in protein synthesis were seen, Tipton noted that the study was underpowered to detect differences in such a small sample size. He suggested that a protein-synthetic increase would be seen in the preworkout treatment if there were approximately double the number of subjects. Also of note is that 4 of the 8 subjects in the pre-group had greater amino acid uptake than any of the subjects in post-group. Furthermore, it’s highly likely that more protein synthesis would be seen in the pre-group if carbohydrate was taken with the protein, as was the case in Tipton’s previous study.
Bird's team saw muscle-preserving effects of an EAA+CHO solution ingested during training after a 4-hour fast . The EAA+CHO treatment suppressed any cortisol increase, whereas the fasted group's cortisol levels rose 105% by the end of the training bout. 3-methylhistidine (3MH – an indicator of myofibrillar protein degredation) levels in the fasted group were elevated by 56% 2 days after the exercise bout, whereas 3MH levels the EAA+CHO group were reduced by 27%. Apparently, even a partial fast before resistance training can negatively impact muscle protein status.
§ A haphazard/randomly variable meal frequency, not necessarily a lower frequency, negatively impacts thermogenesis, blood lipids, and insulin sensitivity.
§ Within a day, a higher frequency has no thermodynamic advantage over a lower frequency under controlled conditions.
§ The majority of controlled intervention trials show no improvement in body composition with a higher meal frequency.
§ Studies indicating the disappearance or lack of hunger in dieters occur in either complete starvation, or very low calorie VLCD regimes (800 kcal/day or less).
§ Hunger is a persistent problem with reduced meal frequency in non-starvation and other protocols with calories above VLCD levels.
§ For controlling appetite, the majority of research indicates the superiority of a higher meal frequency.
§ The body appears to be "metabolically primed" to receive calories and nutrients after an overnight fast. Breakfast is a particularly beneficial time to have dietary protein, since muscle protein synthethis rates are typically lowest at this time.
§ Overall, both experimental and observational research points to breakfast improving memory, test grades, school attendance, nutrient status, weight control, and muscle protein synthesis.
§ Animal research has shown a number of positive health effects of ADF and CR.
§ Human ADF research is scarce and less consistent than animal research, showing both benefits (insulin sensitivity is the most consistent outcome) and risks (impaired glucose tolerance in women).
§ So far, control groups are absent in all human ADF studies. Thus, no comparative conclusions can be drawn between ADF and linear caloric intake.
§ The of the single published controlled trial to date (Stote, et al) comparing 1 versus 3 meals is heavily confounded by an exceptionally high dropout rate in the 1-a-day group, and the use of BIA to measure body composition.
§ The 1-a-day group reported increasing hunger levels throughout the length of the trial, echoing the problem of hunger with a reduced meal frequency seen in other similar research.
§ Ramadan fasting (12-16 hours per day, sunrise to sunset) decreases daytime alertness, mood, wakefulness, competitive athletic performance, and increases the incidence of traffic accidents. It's difficult to determine the relative contributions of dehydration and a lack of food to these adverse phenomena.
§ The effects of exercise and meal frequency on body composition is an interesting but largely unexplored area of research.
Fasting & Exercise
§ Improvements in insulin sensitivity, glucose tolerance (except in women undergoing ADF), bodyweight/bodyfat, blood pressure, blood lipids, and heart rate are commonly cited benefits of IF & CR.
§ All of the above benefits can be achieved by exercise, minus the downsides of fasting.
§ IF and CR have both been found to have neuroprotective effects by increasing BDNF levels.
§ A growing body of research shows that exercise can also increase BDNF, and the degree of effect appears to be intensity-dependent.
§ Based on the limited available data, resistance training performance, especially if its not particularly voluminous, might not be enhanced by preworkout EAA+CHO.
§ Despite equivocal performance effects of pre- or midworkout EAA+CHO, it minimizes muscle damage that occurs from fasted resistance training.
§ Immediate preworkout protein and/or EAA+CHO increases protein synthesis more than fasted resistance training with those substrates ingested immediately postworkout.
§ It’s possible that a partial fast (as short as 4 hours) before resistance training can negatively impact muscle protein status.
It's given that personal goals and individual response are the ultimate navigators of any protocol. Therefore, training and meal schedules should be built upon individual preferences & tolerances, which undoubtedly will differ. However, the purpose of this article is to arm the reader with the facts, so that opinions and anecdotes can be judged accordingly. Personal testimony is invariably biased by the powerful placebo effect of suggestion, and sometimes by ulterior agenda. Science is perched on one end of the epistemological spectrum, and hearsay is on the opposite end. As the evidence clearly indicates, IF is not a bed of roses minus the thorns - there are definite pros and cons.
In the world of fitness, recommendations for improving performance and body composition often gain blind acceptance despite a dearth of objective data. This is common in a field where high hopes and obsessive-compulsive tendencies are united with false appeals and incomplete information. In order to be proven effective beyond the mere power of suggestion, supposed truths must be put through the crucible of science. Drawing conclusions from baseless assumptions is a good way to get nowhere - fast.
Rashidi MR. Effects of nibbling and gorging on lipid profiles, blood glucose and insulin levels in healthy subjects.
Saudi Med J. 2003 Sep;24(9):945-8.
Benton D, Pearl Y. Breakfast, blood glucose, and cognition. Am J Clin Nutr. 1998 Apr;67(4):772S-778S.