Fasted or Non Fasted Cardio?
By Thomas Rowland

In Review: Body Composition Changes Associated with Fasted Versus Non-fasted Aerobic Exercise – Schoenfeld et al (2014)


Weight management is dependent on energy balance, whereby energy balance = energy intake – energy expenditure. To lower energy balance we can decrease energy intake and/or increase energy expenditure. Indeed, research suggests doing both these things is the most effective weight loss method.

Nutrient intake can have a dramatic impact on exercise physiology. It has been hypothesized that performing exercise in the overnight fasted state will optimise exercise induced fat loss due to low resting levels of glycogen and circulating insulin thereby causing a fuel utilization shift away from carbohydrate to fat. The authors note that there is some support for this theory in acute studies and although weak evidence, chronic adaptation studies have shown that endurance training in the fasted state causes adaptations that one would suspect to be more favourable toward fat oxidation than training in the fed state.

The authors end by stating that there is a lack of evidence examining fed vs fasted state cardio over time, and none (to the authors knowledge) have examined this with subjects in energy restriction. Thus the purpose of the study:

“was to investigate changes in fat mass and fat-free mass following four weeks of volume-equated fasted versus fed aerobic exercise in young women adhering to a hypocaloric diet”

The authors have clearly stated the gaps in the literature and also clearly stated the primary outcome measures. To natural muscle readers, the interest in finding the answer to this research question is obvious, it will help inform contest prep strategies with the aim to be as shredded on stage as possible.


Figure 1, produced by myself, shows the study method as a flow chart. The authors recruited 20 healthy young females (mean age was 22.4±2.8). Subjects were somewhat aerobically trained. The authors justified the statistical power of the study for detecting a change in body fat percentage based on a prior body composition study. Power analysis is a statistical technique that estimates the sample size needed to detect a difference of some ‘size’ if it does indeed exist.

After baseline anthropometric and body composition measurements, subjects were randomised (randomisation is essential in a well designed study) into one of two groups, FED or FASTED, with 10 subjects in each group, balanced based on initial weight. The FED group performed exercise in the FED state and the FASTED group performed exercise in the fasted state. The intervention was 4 weeks in duration and all subjects performed cardio 3x per week. The exercise session consisted of 1h of steady state treadmill running with a 5 minute warm up at 50% of maximum heart rate (MHR), 50 mins at 70% MHR and a 5 min cool down at 50% MHR, all at a 0% treadmill gradient. The groups differed with subjects in the FED group ingesting a meal replacement shake (250kcal, 20g pro, 40g carb, 0.5g fat) immediately before exercise whereas subjects in the FASTED group ingested it immediately after.

The dietary control in this study was good, at least as good as dietary control realistically can be. Using some calculations to estimate each subject’s energy balance, a caloric target was set for each subject that produced a 500kcal deficit. Macro targets for each subject were 1.8g/kg of protein, fat was 25-30% of total kcal intake and the remaining kcals came from carbohydrates. Dietary adherence was assessed daily by asking subjects to use the myfitnesspal app to track their nutrient intake. That being said, the validity and accuracy of myfitnesspal was not mentioned.

The anthropometrics and body composition measurements (which were performed pre and post intervention) were fairly standard. Body composition was measured using a BodPod, which although not the gold standard method, which is a DEXA scan, is a valid technique.

Figure 1: Flow chart of study design


This section will get a little bit statistical. If you can follow it, great, don’t worry if you can’t, the take home message will be made clear.

There was no difference between the FASTED and FED groups mean macronutrient intake although subjects did eat less than was planned by the dietician so misreporting of true diets is likely and will always be the case in studies that attempt to control diet.

The primary results shown in table 2 were analysed using a linear mixed model. Although lacking certain statistical information such as confidence intervals, they show that time but not condition had an effect on weight loss (time effect, p = 0.0005), BMI reduction (time effect, p = 0.0005) and fat mass loss (time effect, p = 0.02). That is, both groups lost a significant amount of weight and fat mass, and their BMI subsequently lowered, but there was no difference between the FASTED and FED conditions.

The major weakness of the statistical reporting in this paper is that no confidence intervals (CIs) for any statistic were included. Confidence intervals are an important reliability measure because they tell us the likely range in which the mean would fall if the experiment were to be repeated. Thus if confidence intervals are very wide, we can’t be sure the results in the experiment in question are very reliable. However, if the confidence intervals are narrow, we can be more certain of the reliability of the results reported. So, as this paper didn’t include confidence intervals for any statistical estimate I am uncertain of the reliability of some of their estimates.

Looking into the data on weight and fat mass loss in a bit more detail:

  • FASTED lost 1.6kg in weight and FED lost 1kg
  • FASTED lost 1.1kg of fat mass and FED lost 0.7kg

    These small differences were not statistically different so the data do not indicate that fasted cardio is more beneficial.

    The effect sizes (ES; a measure of the size of the effect) were as follows:

  • FASTED weight loss ES was 0.21 and FED 0.18
  • FASTED fat mass loss ES was 0.2 and FED 0.11

    In terms of effect sizes these numbers are very similar.

    Thus the study does seem to suggest that no true difference exists in weight or fat mass loss between the FASTED and FED groups. Was the study underpowered? (i.e. not enough participants to detect a true effect if it does indeed exist), maybe. However, if we consider the actual data, the effect sizes, and the p value results of linear mixed model (which found no difference between groups), as well as the other published literature (see for a brief overview), I think this study adds to the argument that fasted cardio isn’t superior for fat loss than fed state cardio.


    The authors state the study refutes the hypothesis that fasted state cardio will enhance fat loss by pointing out that both groups in the current study lost a significant amount of weight and fat mass but with no difference between conditions.

    They importantly highlight that for a change in body composition, ‘fat burning’ must be considered over the course of days or weeks (like the current study). They further point out that even if fasted cardio does produce greater fat oxidation during exercise this effect might be neutralized by an increased thermic effect of exercise from eating before exercise, that is, eating before exercise allows you to ‘burn’ more fat by working harder.

    The authors point out a number of limitations:

  • Study duration - 4 weeks in aerobically trained women is not a long period to assess changes in body composition.
  • Dietary and activity control - although good in this study, they cannot rule out a confounding effect of misreporting of nutrient intake or exercise that was performed by subjects in their spare time.
  • Only using young women - it is possible differences in menstrual cycle could influence the results obtained. Further, as with any study, the results are only applicable to the population studied so whether the findings are generalizable to say middle aged, highly resistance trained males is not known.
  • Timing of meal replacement shake - The FASTED group ingested the meal replacement shake immediately after exercise thus it is not clear if any effect would be seen by delaying nutrient intake.

    The authors end by concluding that:

    “Our findings indicate that body composition changes associated with aerobic exercise in conjunction with a hypocaloric diet are similar regardless whether or not an individual is fasted prior to training”

    With the practical recommendation that:

    “Those seeking to lose body fat conceivably can choose to train either before or after eating based on preference”

    Reviewers Conclusion

    This small sample study was a well designed randomised controlled trial with good dietary control. Although there are some statistical flaws in not reporting a single confidence interval, the results reported support the author’s conclusion. They certainly didn’t claim to observe an effect that is likely not there, which is always a good thing. Thus the study adds to the literature that suggests fasted cardio is not the holy grail of fat burning.


    Schoenfeld BJ, Aragon AA, Wilborn CD, Krieger JW, Sonmez GT. Body composition associated with fasted versus non-fasted aerobic exercise. Journal of the International Society of Sports Nutrition, 11(15), 2014.