Phosphatidic Acid
By Thomas Rowland

In Review: Phosphatidic acid enhances mTOR signalling and resistance exercise induced hypertrophy – Joy et al (2014)


Skeletal muscle mass is dependent on muscle protein balance (MPB), where MPB = muscle protein synthesis (MPS) – muscle protein breakdown (MPB). MPS is regulated in part, by a protein called the mechanistic target of rapamycin (mTOR). Research has shown that mTOR signalling is required for resistance exercise induced increases in MPS.

Phosphatidic acid (PA) is a lipid compound that acts as a signalling molecule, a precursor for other lipids and is a major constituent of cell membranes. Recent research has demonstrated intracellular (inside a cell) PA to increase after mechanical stimuli (i.e. muscle contraction) and this PA increase contributes to the activity of the mTOR pathway.

This study (1) had two phases. The first was a molecular biology based study looking at whether any PA pre-cursors can activate the mTOR pathway. However, because this phase isn’t particularly applicable to strength trainers, only the phase 2 human trial is reported here. Thus the aim of this trial was as follows:

Phase 2: “we performed a double-blind, placebo-controlled study that was designed to assess the effects of orally administered PA on skeletal muscle hypertrophy, strength and power when consumed during a periodized RT program”

The authors provided a good background of previous research and the scientific rational for the current study was made clear. They clearly state the gaps in the research and clearly identify the aims of the current research questions. For NM readers, the research question is interesting because it will help us understand whether PA might be an effective supplement to augment muscle mass gains.


Figure 1 produced by myself shows the basic study protocol. 34 young males who were resistance trained for at least 1 year were recruited. Some dropped out due to busy schedules or failure to comply with the prescribed diet so 24 subjects were used for data analysis.

Baseline body composition (lean body mass:LBM, fat mass:FM and total mass:TM) was measured via a DEXA scan and rectus femoris (the middle quad muscle) cross sectional area (CSA) was measured using ultra-sound. Subjects 1RM was determined in the leg press and bench press exercises and power output was assessed using a cycling test.

Subjects were then matched according to LBM, CSA, and leg press 1RM and divided into two equal groups. One group received 750mg per day of PA supplementation (PA group) and the other group received a placebo (PLA group). The study was double blinded so neither the subjects nor the investigators knew who was in what group. This is important because it minimises any effect of bias.

Figure 1. Flow chart of study design.

Both groups completed an 8 week, 3 day daily undulating periodised resistance training programme with a 2:1 hypertrophy to strength ratio (Table 1). Exercises were changed half way through. All subjects consumed 24g of hydrolysed collagen protein post-exercise, that is low in leucine to minimise any effect of post-exercise nutrition. Those in the PA group consumed 450mg of PA before training and 300mg immediately after. On non-training days, 450mg was consumed with breakfast with 300mg at dinner.

Two weeks before, and throughout the study period, subjects were placed on a diet with individual meal plans by a registered dietician (RD) who specialised in sports nutrition, which consisted of 25% protein, 50% carbohydrates and 25% fat. Weekly logs were kept to ensure compliance and results showed no difference in dietary intake between groups. So pretty good dietary control was used, but the paper did not mention whether subjects were put in a hypo, eu or hyper caloric state. As the harris benedict equation was used I’m assuming it was a eucaloric diet because this equation estimates daily Kcal requirements.

The results of this trial were analysed using a statistical technique called a repeated measures ANOVA and effect sizes were reported. All data is reported as means ± standard deviation.

The study was of an average quality design. In its favour it was a simple, double blinded, placebo controlled, two-group comparison with fairly good dietary control. However, it was not randomised, presumably because of the small sample size and a desire to have matching baseline characteristics. In addition, recent research has shown that not everyone responds to exercise in the same way (2). For a particular measure e.g. VO2max or LBM, some people hyper respond, some don’t respond at all and in some cases a few people even respond negatively (2). In such small studies as the one in review here, the groups have not been balanced for responder status. Thus all it would take is a few extra non-responders in one group for that group to appear inferior. Dietary control was reasonably good, but as is always the case can never be tightly controlled due to bias in self reported intakes. All statistics used were common appropriately prescribed tests and the authors also included effect sizes, which are a measure of how large a response was. Unfortunately there is however no mention of confidence intervals being included.


Body composition:

  • The PA group showed greater increases in rectus femoris CSA (pre 4.5± 1.1cm2, post 5.5±1.3 cm2, ES=0.92) than the PLA group (pre 4.5±1.1 cm2, post 5.1±1.2 cm2, ES=0.52) with a p value of p=0.02.

  • The PA group also showed greater increases in LBM (pre 59.7±6kg, post 62.1±5.5kg, ES=0.42) than the PLA group (pre 59.5±4.7kg, post 60.7±4.7kg, ES=0.26) with a p value of p=0.01.

  • While both groups increased total body mass (time effect, p=0.02) there was not a statistical difference between the groups. The PA group went from 78.1±8.7kg to 78.7±7.9kg and the PLA group went from 75.7±5.8kg to 76.5±6.1kg.

  • Both groups decreased fat mass, with the PA group going from 15.1±4.8kg to 13.8±4.2kg with an ES of -0.28 and the placebo group going from 13±6.5kg to 12.5±6.9kg with an effect size of -0.07. The authors noted a trend for the difference to be greater in the PA group (p=0.068), which is what authors say when they haven’t reached that magical arbitrarily selected threshold of p=0.05.

    Strength and power:

  • The PA group showed greater increases in leg press 1RM (pre 228.7±49.5kg, post 280.6±36.2kg, ES=1.2) than the PLA group (pre 226.3±47.2kg, post 258.7±36.1kg, ES=0.78) with a p value of p<0.05, unfortunately the authors decided not to give the exact p value for this result…hmmm.

  • Both groups increased bench press 1RM (time effect, p<0.01) with the PA group going from 98±13.5kg to 105±12.4kg with an effect size of 0.5 and the PLA group going from 91.4±19.1kg to 96.1±17kg with an effect size of 0.25 but with no statistical difference between groups.

  • The authors also decided to measure total strength, which was just leg press 1RM plus bench press 1RM, which seems like a bit of a silly measure to me. Anyway, they reported that the PA group increased total body strength (pre 327.4±59.9kg, post 386.4±45.8kg, ES=1.1) more than the PLA group (pre 318.3±60.3kg, post 355.5±48.4kg, ES=0.68) with a p value of p<0.05. Well, that’s no surprise really given they noticed a difference in leg press 1RM which contributes more to total body strength than bench press does.

  • Both groups increased peak power (time effect, p<0.01), with PA going from 760.5±166W to 822.8±217.3W and PLA going from 733.5±105.8W to 797.3±122.3W but with no difference between groups (p=0.97).


    Phew! Now all the results and fancy statistics are out the way let’s see what the authors had to say about their results.

    The discussion begins by stating that previous research has suggested endogenous (made by your body) PA might play at least a partial role in regulating the exercise induced increase in MPS and they discuss the molecular mechanisms by which this might occur suggesting similar mechanisms could be at play with supplemental PA.

    The authors note that while LBM and hypertrophy increased in both the PLA and PA groups, this occurred to a greater extent in the PA group. They claim this is a robust effect of PA, which I dispute below in my conclusions. They also note the trend (p=0.068) they found for PA to decrease body fat, suggesting PA might lower body fat by increasing LBM, which is a metabolically active tissue.

    The authors also claim that PA supplementation enhances strength increases (again a claim I dispute below) and suggest the most likely mechanism for this is due to increased muscle CSA rather than any neurological effect. They suggest no effect of PA on power was observed due to the training programme being hypertrophy focussed.

    The authors end by concluding:

    “Oral PA supplementation can directly augment changes in skeletal muscle hypertrophy following a chronic RT stimulus and results in significant increases in strength and LBM over placebo”


    This studies strength was its simple placebo controlled, double-blinded design with good dietary controls. Its downfall however was its small sample, lack of randomisation and no control for exercise responder status. This makes the interpretation of this paper quite difficult and certainly not as conclusive or robust as the authors end up concluding.

    In a bit more detail, the p values aren’t particularly strong (and a bit inconsistently reported sometimes giving the exact value and other times not). In addition, the data concerning body composition is a little confusing. Such that the PA group apparently gained more muscle but didn’t show greater increases in total body mass because they might have lost more body fat. Although plausible, their results aren’t conclusive. In the papers favour, all the effect sizes reported were greater in the PA group compared to placebo so it seems to hint at a possible effect of PA on body composition. This could however simply be because exercise responder status was unbalanced. I would also liked to have seen some confidence intervals. I mentioned why confidence intervals are important in a previous research review ( If CIs for a given statistic in both groups are very large and overlap considerably, then its quite likely if the experiment were repeated that given statistic could end up exactly the same and suggest that PA has no effect.

    Coming back to the exercise responder problem. Assume the placebo group had a few more resistance training LBM non-responders in than the PA group. As subjects didn’t take part in both arms of the trial i.e. take the placebo and PA, it would make it look like the PA group did better when in reality the groups ‘responder’ status just wasn’t balanced. As the individual subjects data points aren’t given, only the means and standard deviations are, this is a possibility we can’t rule out.

    In addition, the study was in part funded by Chemi Nutra (White Bear Lake, MN) and two authors on this paper are paid Chemi Nutra consultants and are named on a pending patent by Chemi Nutra (at the time of publication). These two authors did however have no part in data collection or analysis. Now this is of course not to say the data is false, but this could perhaps be why the results are overstated in favour of PA in the discussion.

    In conclusion, the study didn’t control for exercise responder status, the p values weren’t particularly strong and no confidence intervals were reported, but the effect sizes were at least in favour of PA. I think all we can reasonably conclude is that PA is worth another look in larger, better designed studies.


    (1) Joy JM, Gundermann DM, Lowery RP, Jager R, McCleary SA, Purpura M, Roberts MD, Wilson SMC, Hornberger TA and Wilson JM. Phosphatidic acid enhances mTOR signalling and resistance exercise induced hypertrophy. Nutrition and Metabolism. 11:29, 2014.

    (2) Timmons JA. Variability in training-induced skeletal muscle adaptation. Journal of Applied Physiology. 110(3):846-53, 2011.