Human muscle protein turnover—why is it so variable?

Abstract

We undertook a comprehensive review of the literature to unravel the nature of the variability in the reported rate of human muscle protein synthesis. We analyzed the results from studies that report the protein fractional synthesis rate (FSR) in the vastus lateralis in healthy, nonobese, untrained adults ≤50 yr of age in the postabsorptive state at rest by using the primed, constant tracer amino acid infusion method according to experimental design characteristics. We hypothesized that if the variability is methodological (rather than physiological) in nature, systematic clustering of FSR values would be evident, and outliers would become apparent. Overall, as expected, the mixed muscle protein FSR values were significantly (P < 0.001) greater when the muscle vs. the plasma free amino acid enrichment is used as the surrogate precursor pool enrichment, and the average mixed muscle protein FSR values were significantly greater (P = 0.05) than the myofibrillar/myosin heavy chain FSR values. The within-study variability (i.e., population variance) was somewhat smaller in studies that used plasma amino acid/ketoacid enrichments vs. muscle free amino acid enrichment (∼24 vs. ∼31%), but this was not apparent in all circumstances. Furthermore, the between-study consistency of measured FSR values (i.e., interquartile range) was inversely correlated with the average duration between biopsies. Aside from that, the variation in reported FSR values could not be explained by differences in the experimental design and analytical methods, and none of the most commonly used approaches stood out as clearly superior in terms of consistency of results and/or within-study variability. We conclude that the variability in reported values is in part due to 1) differences in experimental design (e.g., choice of precursor pool) and 2) considerable within-subject variability. The summary of the results from our analysis can be used as guidelines for “normal” average basal FSR values at rest in healthy adults.

in the course of life, muscle protein maintenance results from a tightly controlled balance between muscle protein synthesis and muscle protein breakdown. To evaluate the mechanisms responsible for changes in muscle mass that might occur as a result of aging or disease, it is important to accurately measure muscle protein turnover rates. There is considerable debate about the “true” basal, postabsorptive muscle protein turnover rate in human muscle because the values reported in the literature vary by up to approximately eightfold. Part of the variation is most likely due to different approaches used to measure the rate of muscle protein synthesis in vivo in human subjects [e.g., primed, constant tracer amino acid infusion vs. flooding dose tracer amino acid administration to measure the fractional protein synthesis rate (FSR)]; a detailed description and a review of the pros and cons of the various methods can be found elsewhere (22, 37, 68, 89, 92). However, the reported values vary by as much as four- to fivefold even when the comparison is limited to a single experimental approach applied to well-defined study populations and conditions. For example, values reported for the mixed muscle protein FSR in the vastus lateralis portion of the quadriceps femoris of healthy, overnight-fasted young and middle-aged men and women, determined by using the constant tracer amino acid infusion (i.e., the most commonly used approach), range from ∼0.020–0.030%·h⁻¹ (∼2 times total muscle mass per year) (46, 75) in the low range to ∼0.080–0.095%·h⁻¹ (∼8 times total muscle mass per year) (20, 42, 43, 57, 81) in the high range. These differences are not trivial considering that the maximal amino acid-induced increase in FSR is ∼90% above basal values (13) and the biggest reported increase in FSR after a bout of heavy resistance exercise is ∼250% (49). Thus the validity of some of the reported data has been doubted; however, in the absence of a gold standard method and knowledge of the true turnover rate, such criticism may be unfounded.

It is possible that the variability in reported values is simply due to variations in the experimental design, because there are several well-known confounders that affect the final FSR value but do not necessarily invalidate the measurement. For example, it is well established that although theoretically the choice of tracer amino acid should not have an effect on the measured muscle protein synthesis rate, the choice of tracer for studies in vivo in human subjects does in fact, to some small degree, affect the final value (7, 53, 77, 82). The reason(s) for this phenomenon are not entirely clear but may at least in part be related to the fact that instead of measuring the true precursor amino acid pool enrichment (tRNA-bound amino acid enrichment), one relies on the enrichment measured in a surrogate precursor amino acid pool (e.g., plasma or muscle free amino acid pool) in the calculation of the muscle protein FSR [which is derived by using the following equation: FSR = ΔE_protein/E_precursor × 1/t × 100, where t is the time, ΔE_protein is the change in muscle protein labeling (enrichment) over time, and E_precursor is the enrichment of the precursor amino acid pool]. Depending on the tracer amino acid and the surrogate precursor pool, the surrogate more or less closely reflects the true precursor enrichment (6, 50, 77, 88). In fact, Martini et al. (52) demonstrated that in vitro in cultured fibroblasts, the protein FSR is not different when [¹⁵N]glycine, [¹⁵N]proline, or [²H₅]phenylalanine tracers are used in conjunction with the true precursor pool enrichment (i.e., the final “plateau” protein-bound amino acid labeling, which cannot be obtained in vivo in human subjects) for calculation of the FSR. Nevertheless, the variations introduced through the choice of surrogate precursor pool enrichment or tracer are considered too small (approximately <20%) to account for the variability in reported FSR values. However, there are numerous other variations in the experimental design of different studies (e.g., duration between biopsies, methods used for protein enrichment analysis, etc.) that could potentially introduce deviations, and together, several small deviations may add up to relatively big errors in the measured FSR. A systematic review of the literature to find a potential explanation for the apparent inconsistency in the reported values for human muscle protein turnover and the nature of variability is not available.

To fill this gap, we performed an extensive literature search for studies that measured the muscle protein FSR, by using stable isotope-labeled tracer techniques, in the quadriceps femoris in young and middle-aged healthy, untrained subjects during basal, postabsorptive conditions at rest and analyzed the results taking into account differences in the experimental design. We limited our investigation to 1) studies that used the primed, constant tracer amino acid infusion FSR approach, because we found it to be the most commonly used approach (see Data retrieval and analysis), and 2) a well-defined subset of subjects to avoid potential confounding of the results due to aging and differences in training status of subjects, among other examples. Theoretically, if the key requirements for the primed, constant tracer amino acid FSR approach (i.e., steady enrichment in the precursor amino acid pool and linear tracer incorporation into muscle protein during the time period of FSR measurement) are met in all studies and sample analyses are robust (i.e., reliable and accurate), then the average FSR values should be consistent (except for systematic deviations due to known confounders such as, e.g., choice of tracer, differences in the surrogate precursor pool) and fall within the range of the population variance, which should be independent of the experimental design. Therefore, we hypothesized that systematic clustering of FSR values according to experimental design characteristics would be evident and that outliers would become apparent. Furthermore, we hypothesized that the measurements would be most robust (and therefore most consistent between studies and with the tightest population variance) with the use of 1) long measurement periods (to maximize the amount of tracer incorporation into muscle protein and therefore the difference in muscle protein labeling in biopsy samples), 2) long lag times between the start of the primed constant tracer amino acid infusion and the start of the measurement period (to minimize the impact of the initial tracer equilibrium and ensure a steady enrichment in the precursor pool enrichment), 3) surrogate precursor pools that allow frequent sampling (e.g., plasma vs. muscle tissue), and 4) isotope ratio mass spectrometry (IRMS) rather than gas chromatography mass spectrometry (GCMS) to measure the muscle protein enrichment (because IRMS is more sensitive than traditional GCMS, and therefore it is thought to be easier to reliably measure small enrichments with IRMS). We expected no difference in the quality of data obtained by using off- vs. online IRMS techniques, because the analytical error of these methods when measuring muscle protein-bound amino acid enrichments has been reported to be the same (3, 93).

METHODS

Data retrieval and analysis.

We searched PubMed Central for articles (published until January 2010) that report muscle protein synthesis rates in human subjects in the basal, postabsorptive state at rest. Search terms were “tracer, muscle protein synthesis,” “leucine, muscle protein synthesis,” “phenylalanine, muscle protein synthesis,” and “flooding dose, muscle protein synthesis.” In addition, we 1) searched for articles published by researchers who are well-known to work in the area of muscle protein metabolism and 2) used the reference lists of all retrieved articles to identify potentially missing sources. We thereby identified 137 articles that report muscle protein synthesis rates in human subjects in the basal, postabsorptive state at rest. We excluded one report (67) because the authors later reported (44) that the FSR in the original report was erroneously high due to methodological issues. Eighty-five (62.5%) of the 136 remaining articles reported muscle protein synthesis rates (mixed, myofibrillar, or myosin heavy chain) in the vastus lateralis portion of the quadriceps femoris in healthy, nonobese, untrained subjects between the ages of 18 and 50 yr; of those, we used only studies that included averaged data that were obtained from at least 5 subjects for our analysis (n = 82 articles). The vast majority of those (71 of the 82 overall and 43 of 44 in the last decade) relied on the primed, constant tracer amino acid infusion technique, and only 11 (13%) used the tracer flooding dose technique; we therefore excluded this technique from our analysis. We included results from both men and women without stratifying by sex, because most studies were conducted in mixed (including men and women) groups of subjects; moreover, we (75) and others (35, 59) have found no evidence for differences in the muscle protein FSR in healthy, untrained young and middle-aged men and women at rest.

Detailed information regarding the results from different studies is presented in Tables 1 and 2. A summary of the findings is presented in Table 3. For data presentation and further analysis, we grouped the results and analyzed them, taking into account known confounders (e.g., choice of tracer and surrogate precursor pool) and other differences in the experimental design (e.g., duration of label incorporation into muscle protein between biopsies) and analytical methods (e.g., instrumentation used for muscle protein enrichment analysis) that could potentially introduce a systematic bias. Within each group we sorted the results by average FSR values in ascending order. In addition to the factors listed, we initially intended to also take into account biopsy sampling site in our analysis. Unfortunately, however, sampling sites are often not reported [only 31 (44%) of the 71 articles we reviewed provided these details]; we therefore omitted this from our analysis. Nevertheless, sampling procedure is likely not a great concern and is unlikely to help explain the variability in reported FSR values, because Nair et al. (55) have shown that there is little variance (2.05 ± 0.63% difference) in the enrichment obtained from repeat biopsies taken from the same site, and we have shown that the difference in mixed muscle protein FSR in different muscles (e.g., soleus vs. triceps) is <15% (54), so we would expect this to be the upper limit of differences in biopsies from the vastus lateralis in different legs. Likewise, we intended to examine the effect of the priming dose-to-constant infusion rate ratio. However, due to considerable homogeneity between studies [e.g., all but one study, which applied a [¹³C]leucine tracer in conjunction with plasma α-ketoisocaproic acid (KIC) as the surrogate precursor pool enrichment and IRMS for protein analysis, and 38 of 44 of the studies in which a phenylalanine tracer was used in combination with muscle free phenylalanine enrichment as the precursor pool enrichment used the same priming dose-to-constant infusion rate ratio; see Tables 1 and 2], it is unlikely that this is a critical factor. We therefore omitted it from further analysis.

Table 1.

Mixed muscle protein FSR in the vastus lateralis portion of the quadriceps femoris muscle in healthy young adults in the postabsorptive state at rest

Reference	Subjects	Tracer	P/I Ratio	Bound Protein Analysis	Biopsy Timing	FSR, % · h−1
KIC tracer with either plasma or average of arterial and muscle free leucine as precursor
Bohe et al. (14)	5 men and 1 woman 33 ± 2 yr 80 ± 12 kg	5,5,5-2H3	1:1	GCMS	0.5 and 3 h	0.076 ± 0.020 (26%)
Chinkes et al. (20)	4 men and 1 woman 73 ± 12 kg	1-13C	1:1	HPLC-carbon nitrogen analyzer-IRMS	1 and 4 h	0.094 ± 0.012 (12%)
Leucine tracer with plasma KIC as precursor
Yarasheski et al. (95)	6 men 30 ± 7 yr 74 ± 7 kg	1-13C	1:1	GC-C-IRMS	1.5 and 14 h	0.037 ± 0.005 (14%)
Yarasheski et al. (93)	9 men Mean age 23 yr	1,2-13C	1:1	GC-C-IRMS	1.5 and 14 h	0.039 ± 0.014 (36%)
Balagopal et al. (4)	6 men and 4 women 30 ± 9 yr 70 ± 9 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.041 ± 0.010 (24%)
Rooyackers et al. (70)	6 men and 6 women 24 ± 3 yr	1-13C	1:1	GC-C-IRMS	5 and 10 h	0.043 ± 0.007 (16%)
Nair et al. (55)	10 men 27 ± 6 yr 73 ± 6 kg	1-13C	1:1	GC-ninhydrin-IRMS	0 and 8 h	0.046 ± 0.009 (20%)
Halliday et al. (44)	18 men 29 ± 6 yr 72 ± 6 kg	1-13C	1:1	GC-ninhydrin-IRMS	2 and 8 h	0.046 ± 0.012 (26%)
Gibson et al. (38)	6 men 23 yr 73 ± 10 kg	1-13C	1:1	GC-ninhydrin -IRMS	7 h with background TTR from previous study	0.046 ± 0.012 (26%)
Charlton et al. (19)	4 men; 2 women Mean age 28 yr Mean BM 70 kg	1-13C	1:1	GC-C-IRMS	2 and 8 h	0.047 ± 0.010 (21%)
Balagopal et al. (5)	8 men and women 23 ± 3 yr 62 ± 12 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.047 ± 0.008 (17%)
Nair et al. (56)	7 young adults	1-13C	0.92:1	GC-ninhydrin -IRMS	0 and 8 h	0.048 ± 0.008 (17%)
Hasten et al. (45)	4 men and 3 women 27 ± 3 yr 63 ± 11 kg	1-13C	1:1	GC/C/IRMS	∼1.5 and 14 h	0.048 ± 0.008 (17%)
Carraro et al. (15)	6 young adults 29 ± 7 yr 70 ± 7 kg	1,2-13C	1:1	HPLC-combustion furnace-IRMS	15 and 19 h	0.048 ± 0.024 (50%)
Gibson et al. (39)	14 men Mean age 26 yr 73 ± 7 kg	1-13C	1:1	GC-ninhydrin-IRMS	0 and 7 h	0.048 ± 0.020 (42%)
	7 men Mean age 48 yr 75 ± 13 kg					0.059 ± 0.012 (20%)
Yarasheski et al. (94)	2 men and 4 women Mean age 24 yr	1-13C or 1,2-13C	1:1	GC-C-IRMS	∼1.5 and 13 h	0.049 ± 0.010 (20%)
Parise et al. (59)	7 young men	1-13C	1:1	GC-C-IRMS	2 and 14 h	0.055 ± 0.020 (36%)
	7 young women					0.057 ± 0.020 (35%)
Gore et al. (41)	5 men 28 ± 2 yr 77 ± 9 kg	1,2-13C	1:1	HPLC-combustion furnace-IRMS	4 h‡	0.056 ± 0.011 (20%)
Leucine tracer with muscle free leucine as precursor
Petersen et al. (61)	8 men 25 ± 3 yr 73 ± 6 kg	15N	0.78:1	GC-C-IRMS	2 and 6 h	0.035 ± 0.015 (44%)
Hasten et al. (45)	4 men and 3 women 27 ± 3 yr 69 ± 11 kg	1-13C	1:1	GC-C-IRMS	∼1.5 and 13 h	0.057 ± 0.006 (28%)
Balagopal et al. (5)	8 men and women 23 ± 3 yr	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.060 ± 0.0014 (23%)
Balagopal et al. (4)	6 men and 4 women 30 ± 9 yr 70 ± 9 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.060 ± 0.006 (32%)
Guillet et al. (42)	8 adults 26 ± 11 yr 79 ± 18 kg	1-13C	1:1	GC-C-IRMS	4 h	0.082 ± 0.028 (34%)
Guillet et al. (43)	6 adults 25 ± 2 y 24 ± 2 kg/m2	1-13C	1:1	GC-C-IRMS	4 h	0.082 ± 0.024 (29%)
Adey et al. (1)	6 adults 41 ± 9 yr 26 ± 3 kg/m2	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.086 ± 0.019* (29%)
Phenylalanine tracer with muscle free phenylalanine as precursor
Smith et al. (75)	8 men 38 ± 6 yr 79 ± 13 kg	ring-2H5	0.58:1	GC-MS	1 and 4 h	0.038 ± 0.0017 (45%)
	8 women 37 ± 7 yr 68 ± 5 kg					0.045 ± 0.014 (31%)
Biolo et al. (8)	6 men 29 ± 12 yr 73 ± 12 kg	ring-13C6	0.67:1	HPLC-carbon-nitrogen analyzer-IRMS	1 and 4 h	0.040 ± 0.017 (43%)
Biolo et al. (10)	5 young men 24 ± 4 yr 72 ± 11 kg	ring-13C6	0.67:1	HPLC-carbon-nitrogen analyzer-IRMS	1 and 4 h	0.043 ± 0.013* (30%)
Drummond et al. (25)	7 men 30 ± 5 yr 89 ± 13 kg	ring-2H5	0.67:1	GC-MS	2 and 3 h	0.044 ± 0.019* (43%)
Volpi et al. (86)	4 men and 3 women 30 ± 5 yr 72 ± 8 kg	ring-2H5	0.67:1	GC-MS	2 and 6 h	0.044 ± 0.010 (23%)
Biolo et al. (12)	5 men 29 ± 11 yr 73 ± 11 kg	ring-13C6	0.67:1	GC-C-IRMS	1 and 4 h	0.048 ± 0.014 (29%)
Fujita et al. (32)	6 men 32 ± 5 yr 84 ± 15 kg	ring-13C6	0.67:1	GC-MS	2 and 4 h	0.053 ± 0.010*† (19%)
						0.057 ± 0.015*† (26%)
Raj et al. (64)	9 men and 1 woman 45 ± 15 yr 74 ± 13 kg	ring-13C6	0.33:1	GC-MS	2 and 5 h	0.054 ± 0.038 (70%)
Biolo et al. (9)	9 men and 1 woman 19-52 yr 74 ± 13 kg	ring-13C6	0.67:1	HPLC-carbon-nitrogen analyzer-IRMS	2 and 5 h	0.054 ± 0.019 (35%)
Bohe et al. (13)	4-6 men and women in each group	ring-2H5	0.87:1	GC-MS	0.5 and 3 h	0.054 ± 0.005 (9%)
						0.057 ± 0.018 (32%)
						0.086 ± 0.020 (23%)
Fujita et al. (33)	7 men 27 ± 5 yr 77 ± 16 kg	ring-2H5	0.67:1	GC-MS	2 and 3 h	0.056 ± 0.021* (38%)
Phillips et al. (62)	4 men and 4 women 23 ± 6 yr 65 ± 11 kg	ring-2H5	0.67:1	GC-MS (CI)	2 and 5 h	0.056 ± 0.014* (25%)
Rasmussen et al. (66)	6 men 32 ± 10 yr 77 ± 15 kg	ring-2H5	0.67:1	GC-MS	2 and 5 h	0.074 ± 0.037 (50%)
Sheffield-Moore et al. (73)	5 men 22 ± 3 yr 76 ± 15 kg	ring-2H5	0.67:1	GC-MS	2 and 5 h	0.057 ± 0.009 (16%)
Sheffield-Moore et al. (72)	6 men 22 ± 3 yr 77 ± 13 kg	ring-2H5	0.67:1	GC-MS (CI)	2 and 5 h	0.057 ± 0.010 (18%)
Volpi et al. (87)	26 men 28 ± 10 yr 76 ± 10 kg	ring-2H5	0.67:1	GC-MS	2 and 6 h	0.058 ± 0.023 (40%)
Drummond et al. (26)	8 men 29 ± 6 yr 74 ± 14 kg	ring-2H5	0.67:1	GC-MS	2.5 and 3.5 h	0.059 ± 0.008* (14%)
Volpi et al. (85)	4 men and 1 woman 30 ± 7 yr 25 ± 1 kg/m2	ring-2H5	0.67:1	GC-MS	2 and 6 h	0.060 ± 0.019 (32%)
Fujita et al. (36)	4 men and 2 women 28 ± 5 yr 74 ± 12 kg	ring-13C6	0.67:1	GC-MS	2 and 4 h	0.060 ± 0.024* (40%)
	4 men and 2 women 29 ± 7 yr 72 ± 17 kg					0.069 ± 0.013* (19%)
	3 men and 4 women 25 ± 5 yr 63 ± 16 kg					0.070 ± 0.024* (34%)
Volpi et al. (83)	5 men and 1 woman 31 ± 7 yr 87 ± 12 kg	ring-2H5	0.67:1	GC-MS	Hourly from 2 to 6 h	0.061 ± 0.032* (52%) (average)
						0.049 ± 0.032* (65%)
						0.057 ± 0.032* (56%)
						0.059 ± 0.039* (66%)
						0.069 ± 0.024* (35%)
Carroll et al. (16)	6 men and 4 women 28 ± 3 yr 70 ± 19 kg	ring-2H5	0.67:1	GC-MS	2 and 5 h	0.063 ± 0.028* (44%)
Rasmussen et al. (65)	3 men and 4 women 25 ± 5 yr 64 ± 16 kg	ring-2H5 or ring-13C6	0.67:1	GC-MS	2 and 5 h	0.063 ± 0.008* (13%)
Dreyer et al. (24)	7 men and 4 women 27 ± 7 yr 71 ± 17 kg	ring-2H5	0.67:1	GC-MS	2 and 3 h	0.063 ± 0.013* (21%)
Paddon-Jones et al. (58)	2 men and 4 women 34 ± 10 yr 6 3 ± 7 kg	ring-2H5	0.67:1	GC-MS	2 and 5 h	0.064 ± 0.017 (27%)
Fujita et al. (34)	7 men and 4 women 27 ± 7 yr 71 ± 17 kg	ring-2H5	0.67:1	GC-MS	3 and 4 h	0.064 ± 0.013* (20%)
Fujita et al. (35)	8 women 26 ± 6 yr 60 ± 14 kg	ring-13C6	0.67:1	GC-MS	1 and 4 h	0.064 ± 0.017* (27%)
	10 men 27 ± 6 yr 75 ± 13 kg					0.066 ± 0.019* (29%)
Dreyer et al. (23)	13 young men	ring-2H5	0.67:1	GC-MS	2 and 3 h	0.065 ± 0.025* (38%)
Ferrando et al. (30)	6 men 28 ± 7 yr 82 ± 17 kg	ring-13C6	0.67:1	HPLC-carbon-nitrogen analyzer-IRMS	2 and 5 h	0.066 ± 0.011 (41%)
Biolo et al. (11)	6 men 29 ± 12 yr 73 ± 12 kg	ring-13C6	0.67:1	HPLC-carbon-nitrogen analyzer-IRMS	1 and 4 h	0.066 ± 0.017 (26%)
Ferrando et al. (31)	7 men 28 ± 5 yr 77 ± 8 kg	ring-2H5	0.67:1	GC-MS (CI)	2 and 5 h	0.067 ± 0.031 (46%)
Sheffield-Moore et al. (71)	6 men 27 ± 7 yr 78 ± 7 kg	ring-2H5	0.67:1	GC-MS	3 h	0.072 ± 0.007 (10%)
Symons et al. (80)	5 men and 5 women 41 ± 8 yr 88 ± 15 kg	ring-13C6	0.42:1	GC-MS	2 and 5 h	0.074 ± 0.016 (22%)
Ferrando et al. (29)	6 men 30 ± 6 yr 65 ± 6 kg	ring-13C6	0.67:1	HPLC-carbon-nitrogen analyzer-IRMS	4 and 7 h	0.074 ± 0.027 (36%)
Pikosky et al. (63)	4 men and 4 women 21 ± 1 yr 76 ± 17 kg	ring-2H5	0.67:1	GC-MS (CI)	2 and 5 h	0.077 ± 0.007 (26%)
Caso et al. (18)	4 men and 4 women 30 ± 8 yr 73 ± 17 kg	ring-2H5	1:1	GC-MS	2 and 5 h	0.081 ± 0.018 (22%)
Symons et al. (81)	3 men and 4 women	ring-13C6	0.42:1	GC-MS	3 and 6 h	0.082 ± 0.024* (29%)
Paddon-Jones et al. (57)	2 men and 4 women	ring-2H5	0.67:1	GC-MS	2 and 5 h	0.083 ± 0.024 (29%)
Phenylalanine tracer with plasma phenylalanine as precursor
Henderson et al. (46)	30 men 20-26 yr 75–88 kg	15N	1:1	GC-C-IRMS	3 and 6 h	0.030 ± 0.005* (17%)
	32 women 20–23 yr 58-73 kg					0.035 ± 0.005* (14%)
Smith et al. (75)	8 women 37 ± 7 yr 68 ± 5 kg	ring-2H5	0.58:1	GC-MS	1 and 4 h	0.031 ± 0.011 (34%)
	8 men 38 ± 6 yr 79 ± 13 kg					0.021 ± 0.008 (39%)
Katsanos et al. (48)	4 men and 4 women 31 ± 6 yr	ring-2H5	0.67:1	GC-MS	1 and 4 h	0.036 ± 0.011 (31%)
	4 men and 4 women 29 ± 8 yr					0.48 ± 0.014 (29%)
Katsanos et al. (47)	5 men and 3 women 31 ± 6 yr 72 ± 13 kg	ring-2H5	0.58:1	GC-MS	1 and 4 h	0.041 ± 0.014* (34%)
Short et al. (74)	20 adults 19–38 yr	15N	1:1	GC-C-IRMS	5 and 10 h	0.046 ± 0.009 (20%)
Caso et al. (18)	4 men and 4 women 30 ± 8 yr 73 ± 17 kg	ring-2H5	1:1	GC-MS	2 and 5 h	0.055 ± 0.008 (15%)

Values are means ± SD with population variance shown in parentheses. Inclusion criteria were basal, postabsorptive muscle protein fractional synthesis rates (FSR), measured at rest, in the vastus lateralis portion of the quadriceps femoris in 5 or more nonobese, healthy, untrained subjects between the ages of 18 and 50 yr. P/I ratio, ratio of priming dose to infusion rate (expressed per hour); GCMS, gas chromatography-mass spectrometry; HPLC, high-performance liquid chromatography; IRMS, isotope ratio mass spectrometry; ninhydrin, reacted with amino acids to produce CO₂; GC-C-IRMS, derivatized amino acids were combusted after being separated by GC before being applied to IRMS; BM, body mass; CI, chemical ionization; carbon-nitrogen analyzer, device used for combustion of amino acids; combustion furnace, device used for combustion of amino acids in the presence of cupric oxide. Subject data include the number, sex, age, and weight of subjects, if reported. Biopsy timing is given relative to the start of tracer infusion.

^*Values were taken from a figure.

^†Subjects were studied on 2 separate occasions.

^‡Plasma protein enrichment preceding infusion was taken as the background enrichment.

Table 2.

Myofibrillar protein and myosin heavy chain FSR in the vastus lateralis portion of the quadriceps femoris in healthy, young adults in the postabsorptive state at rest

Reference	Myo/MHC	Subjects	Tracer	P/I Ratio	Bound Protein Analysis	Biopsy Timing	FSR, % · h−1
Leucine tracer with plasma KIC as precursor
Balagopal et al. (4)	MHC	6 men and 4 women 30 ± 9 yr 70 ± 9 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.030 ± 0.012 (40%)
Mittendorfer et al. (54)	Myo	9 men 30 ± 9 yr 70 ± 9 kg	1,2-13C	1:1	GC-C-IRMS	0 and 3 h	0.031 ± 0.003 (10%)
Cuthbertson et al. (21)	Myo	20 men 28 ± 6 yr 75 ± 10 kg	1,2-13C	0.67:1	GC-C-IRMS	0 and 3 h	0.032 ± 0.013 (41%)
Charlton et al. (19)	MHC	4 men and 2 women Mean age 28 yr Mean BM 70 kg	1-13C	1:1	GC-C-IRMS	2 and 8 h	0.035 ± 0.010 (35%)
Balagopal et al. (5)	MHC	8 men and women 23 ± 3 yr 62 ± 11 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.036 ± 0.003 (8%)
Louis et al. (51)	Myo	6 men 26 ± 7 yr 22 ± 4 kg/m2	1-13C	0.8:1	GC-C-IRMS	0.5 and 3 h	0.036 ± 0.005 (14%)
Hasten et al. (45)	MHC	4 men and 3 women 27 ± 3 yr 69 ± 11 kg	1-13C	1:1	GC-C-IRMS	∼1.5 and 14 h	0.038 ± 0.008 (21%)
Kumar et al. (49)	Myo	5 men	1,2-13C	0.8:1	GC-C-IRMS	0 and 2.5 h	0.039 ± 0.004 (10%)
Welle et al. (90)	Myo	8 men 26 ± 4 yr 81 ± 8 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	2 and 8 h	0.054 ± 0.010 (19%)
Robinson et al. (69)	Myo	6 men 19 ± 2 yr 80 ± 27 kg	1,2-13C	0.8:1	GC-C-IRMS	2 and 6 h	0.059 ± 0.008 (14%)
Welle et al. (91)	Myo	5 men and 4 women 22-31 yr	1-13C	1:1	HPLC-ninhydrin-IRMS	2 and 8 h	0.061 ± 0.012 (20%)
Leucine tracer with muscle free leucine as precursor
Balagopal et al. (4)	MHC	6 men and 4 women 30 ± 9 yr 70 ± 9 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.044 ± 0.018 (41%)
Balagopal et al. (5)	MHC	8 men and women 23 ± 3 yr 62 ± 13 kg	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.047 ± 0.010 (21%)
Hasten et al. (45)	MHC	4 men and 3 women 23 to 32 yr	1-13C	1:1	GC-C-IRMS	∼1.5 and 14 h	0.047 ± 0.010 (21%)
Adey et al. (1)	MHC	6 adults 41 ± 9 yr 26 ± 3 kg/m2	1-13C	1:1	HPLC-ninhydrin-IRMS	5 and 10 h	0.066 ± 0.032* (48%)
Phenylalanine tracer with muscle free phenylalanine as precursor
Glover et al. (40)	Myo	10 men and 2 women 21 ± 3 yr 80 ± 14 kg	ring-13C6	0.67:1	GC-C-IRMS	2.5 h	0.037 ± 0.010 (27%)

Values are means ± SD with population variance shown in parenthesis. Inclusion criteria were basal, postabsorptive muscle protein FSR, measured at rest, in the vastus lateralis portion of the quadriceps femoris in 5 or more nonobese, healthy, untrained subjects between the ages of 18 and 50 yr.

^*Value was taken from a figure. MHC myosin heavy chain; Myo, myofibrillar protein. P/I ratio, priming dose-to-infusion rate (expressed per h) ratio; ninhydrin, amino acids were reacted with ninhydrin to produce CO₂; C, derivatized amino acids were combusted after being separated by GC before being applied to IRMS.

Table 3.

Descriptive statistics for muscle protein FSR in the vastus lateralis portion of the quadriceps femoris in healthy, young adults in the postabsorptive state at rest

Method				FSR, % · h−1		Population Variance, %
Tracer	Precursor/protein analysis	n	Time Between Biopsies, min	Median [quartiles]	Range	Median [quartiles]	Range
Mixed muscle protein
KIC	Plasma or mean of arterial and muscle free leucine/GCMS and IRMS	2	150, 180	0.076, 0.094	0.076-0.094	N/A	12-26
Leucine	Plasma KIC/IRMS	18	483 ± 192a	0.048 [0.046, 0.049]	0.037-0.059	21 [18, 33]a	14-50
Leucine	Muscle free leucine/IRMS	5d	366 ± 183a,b	0.060 [0.057, 0.060]	0.035-0.086	28 [23, 32]	22-44
Phenylalanine	Plasma phenylalanine/GCMS and IRMS	9	193 ± 40b,c	0.036 [0.031, 0.046]	0.021-0.055	29 [17, 31]	14-39
	IRMS only	3	180	0.030, 0.035, 0.046	0.030-0.046	14, 17, 20	14-20
	GCMS only	6	156 ± 53c	0.039 [0.032, 0.046]	0.0210-0.055	31 [30, 33]	15-39
Phenylalanine	Muscle free phenylalanine/GCMS and IRMS	44	160 ± 49c	0.062 [0.056, 0.068]	0.038-0.086	29 [22, 39]	9-70
	IRMS only	7	180	0.054 [0.046, 0.066]	0.040-0.074	35 [30, 39]	26-43
	GCMS only	37	156 ± 53c	0.063 [0.057, 0.069]	0.038-0.086	27 [21, 38]	9-70
Myofibrillar and myosin heavy chain
Leucine	Plasma KIC/IRMS	11	273 ± 155b	0.036 [0.034, 0.047]	0.030-0.061	19 [12, 28]a	8-41
Leucine	Muscle free leucine/IRMS	4	398 ± 195a,b	0.047 [0.046, 0.052]	0.044-0.066	31 [21, 43]	21-48
Phenylalanine	Muscle free phenylalanine/IRMS	1	150	0.037	N/A	27	N/A

Values are means ± SD (time between biopsies) or median values with quartiles given in brackets (FSR and population variance); n = no. of FSR values (individual values are given in columns to the right if n ≤ 3).

^a,b,cP < 0.05, average/median values not sharing the same symbol (within the same column) are significantly different from each other.

^dOnly values obtained from studies with 2 biopsies were included in the analysis. KIC, ketoisocaproic acid.

Comparisons of the results from different experimental approaches (e.g., results from studies that used GCMS vs. IRMS for protein analysis) were evaluated by using either one-way analysis of variance (ANOVA) and Tukey's post hoc procedures (multiple comparisons) or Student's t-test (single comparison) for independent samples to identify significant mean differences. Relationships between parameters were evaluated by using the Pearson product moment or Spearman's correlation coefficients for normally distributed and skewed data sets, respectively. Data are means ± SD or medians [quartiles]. A P value <0.05 was considered statistically significant.

RESULTS

Mixed muscle protein FSR.

In two studies, a KIC tracer (5,5,5-²H₃ or 1-¹³C₁) infusion in conjunction with GCMS or IRMS for protein analysis and intracellular free leucine enrichment as the surrogate precursor pool enrichment were used; the 192 FSR values obtained from these studies were 0.076 and 0.094%·h⁻¹ (Tables 1 and 3).

Sixteen studies (including 18 groups of subjects and therefore 18 average FSR values) were conducted by using a single- or double-labeled [¹³C]leucine tracer in conjunction with IRMS for protein analysis and plasma KIC enrichment as the surrogate precursor pool enrichment; the duration between biopsies ranged from 4 to 12.5 h (Tables 1 and 3). We found fairly good agreement in the mixed muscle protein FSR values obtained from these studies. In fact, this approach, along with the use of a [¹³C]leucine tracer in conjunction with IRMS for protein analysis and muscle free leucine enrichment as the surrogate precursor pool enrichment (see below), yielded the most consistent (i.e., smallest interquartile range of 0.003%·h⁻¹) FSR values (Table 3). Average FSR values ranged from 0.037 to 0.065%·h⁻¹ with the majority (14 of 19) falling between 0.040 and 0.050%·h⁻¹ (Tables 1 and 3). There was, however, no apparent difference between studies that reported high or low values in this group. For example, there was no difference in the FSR values obtained from studies that used a single (1-¹³C₁; n = 14)- or double (1,2-¹³C₂; n = 4)-labeled leucine tracer (0.048 ± 0.006 vs. 0.049 ± 0.008%·h⁻¹, respectively; P = 0.72) or studies that used GC-combustion (C)-IRMS vs. off-line IRMS for protein analysis (0.048 ± 0.008 vs. 0.048 ± 0.005%·h⁻¹, respectively; P = 0.89). There was also no relationship between the reported FSR values and the timing of the initial muscle biopsy (r = −0.16, P = 0.53) or duration between biopsies (r = 0.06, P = 0.83).

Seven studies were conducted by using a leucine tracer (¹⁵N; n = 1 and 1-¹³C₁; n = 6), IRMS for protein analysis, and calculation of the FSR with the intracellular free leucine enrichment as the precursor pool enrichment; the duration between biopsies ranged from 4 to 11.5 h. This approach resulted in values that ranged from 0.035 to 0.086%·h⁻¹ in the five studies that measured the protein labeling from two biopsies (Tables 1 and 2). In two studies, the initial biopsy was omitted and the FSR values were, as expected (76), considerably greater (∼0.082%·h⁻¹) than the average FSR value obtained in studies that included two biopsies (Tables 1 and 3).

The largest number of mixed muscle protein FSR values were obtained by using phenylalanine tracers (37 studies reporting 44 average FSR values) and either IRMS (¹³C; n = 7) or GCMS (¹³C and ²H₅; n = 37) for protein analysis and the intracellular free phenylalanine enrichment as the precursor pool enrichment. [Ring-²H₅]phenylalanine was used for 26 and [ring-¹³C₆]phenylalanine for 17 FSR measurements; one average FSR was obtained by using [ring-²H₅]phenylalanine in some and [ring-¹³C₆]phenylalanine in the other subjects (65). The FSR values in those studies were typically obtained over shorter periods of time (∼3–5 h between biopsies) and ranged from 0.038 to 0.086%·h⁻¹ with no apparent reason for the variability (Tables 1 and 3). For example, there was no relationship between the average FSR and the duration between biopsies (r = 0.01, P = 0.93) or the timing of the initial biopsy (r = 0.22, P = 0.15). There was also no difference in the FSR values obtained from studies that were otherwise similar but used IRMS vs. GCMS for protein analysis (0.056 ± 0.013 vs. 0.063 ± 0.011%·h⁻¹, respectively; P = 0.16) or that used [ring-²H₅] vs. [ring-¹³C₆]phenylalanine (0.062 ± 0.012 vs. 0.061 ± 0.011%·h⁻¹, respectively; P = 0.81). The small sample size (n = 7) for studies that employed IRMS for protein analysis precluded us from examining potential differences between on- and off-line methods within this subgroup of studies in statistical terms, but the data presented in Table 2 do not indicate a systematic bias. Furthermore, the analytical error of these methods when measuring muscle protein bound amino acid enrichments has been reported to be the same (3, 93).

Six studies report a total of 9 FSR values that were obtained by using a phenylalanine tracer (¹⁵N in 3 and ring-²H₅ in the other 6) in conjunction with plasma phenylalanine enrichment as the surrogate precursor pool enrichment and GCMS or IRMS for protein analysis; the duration between biopsies was 3 h in all but one of the studies, in which it was 5 h, and the reported average values ranged from 0.021 to 0.055%·h⁻¹ with no apparent reason for the difference in results (Tables 1 and 3). As expected (17, 77), the mixed muscle protein FSR was significantly (P < 0.001) greater when the muscle free phenylalanine vs. the plasma free phenylalanine enrichment was used as the precursor pool enrichment (Table 3); however, the variability in the average reported FSR values (interquartile range) was not different in these two types of studies.

Myofibrillar protein and myosin heavy chain protein FSR values.

Eight of 11 studies in which a ¹³C-labeled leucine tracer in conjunction with IRMS for protein analysis and plasma KIC enrichment as the precursor pool enrichment were used report myofibrillar FSR values between 0.030 and 0.040%·h⁻¹; the other 3 studies report values ranging from 0.054 to 0.061%·h⁻¹ with no apparent reason for this discrepancy, such as the choice of label (0.041 ± 0.011 vs. 0.040 ± 0.013%·h⁻¹ with single- vs. double-labeled [¹³C]leucine, respectively; P = 0.88), mode of protein analysis (0.039 ± 0.009 vs. 0.045 ± 0.015%·h⁻¹ with on- vs. off-line methods, respectively; P = 0.38), duration between biopsies (r = 0.15, P = 0.65), which ranged from 2.5 to 12.5 h, or timing of the initial biopsy (r = 0.02, P = 0.97) (Tables 2 and 3). The values obtained by using a [¹³C]leucine tracer (all 1-¹³C₁), IRMS for protein analysis, and calculation of the FSR (over measurement periods of 5–12.5 h) with the intracellular free leucine enrichment as the precursor pool enrichment (4 studies) resulted in values of 0.044 and 0.066%·h⁻¹ (Tables 2 and 3). Interestingly, the same two approaches (i.e., leucine tracer, plasma KIC, or muscle free leucine as surrogate precursor pool, IRMS analysis, long duration between biopsies, etc.) yielded the most consistent average mixed muscle protein FSR values with an interquartile range of 0.003%·h⁻¹, whereas the variability in the case of myofibrillar/myosin heavy chain protein (interquartile range of 0.13%·h⁻¹) falls right within the overall variability observed for the other methods analyzed.

In one study, a [¹³C]phenylalanine tracer was used in conjunction with IRMS for protein analysis and calculation of the FSR with the intracellular free phenylalanine enrichment as the precursor pool enrichment, and the FSR was 0.037%·h⁻¹ (Table 2).

Relationship between length of FSR measurement period and variability in reported FSR values.

The between-study consistency of measured FSR values (assessed by the interquartile range of reported FSR values for a given approach) was inversely correlated with the average duration between biopsies (r = −0.86, P = 0.03; Table 3); i.e., the longer the duration between biopsies, the more consistent the between-study average FSR values. There was no relationship between the within-study variability in FSR values (population variance) and the average duration between biopsies.

Population variance.

The median population variance for all studies was 26 [20, 36]%; the smallest reported value was 8% and the largest was 70% (Table 3). There was considerable spread (minimal to maximal) in the variance in all subgroups listed (Table 3), and no one method clearly stood out as producing the tightest data. On average, however, the use of a leucine tracer with plasma KIC enrichment as the surrogate precursor pool enrichment and IRMS for protein analysis (both for mixed muscle protein and myofibrillar/myosin heavy chain protein) resulted in the smallest population variance (∼20%; P = 0.05 compared with all other approaches, which resulted in variances of ∼31% with no difference among them; P = 0.93). This was most likely attributable to the use of plasma KIC enrichment as the surrogate precursor pool enrichment, because overall the population variance was smaller in studies that used plasma amino acid/ketoacid enrichments (vs. the muscle free amino acid enrichment) as the precursor enrichment in the FSR calculation, but the difference, although statistically significant, was small (24 ± 10 vs. 30 ± 12%, respectively; P = 0.01).

No difference in the population variance was observed between studies that were otherwise similar but used IRMS vs. GCMS for protein analysis (e.g., 34 ± 6 vs. 30 ± 13% for mixed muscle protein FSR determined by using a phenylalanine tracer, the muscle free phenylalanine enrichment as the precursor enrichment, and either IRMS vs. GCMS for protein analyses, respectively; P = 0.38) or on- or off-line IRMS methods for protein analysis (e.g., 26 ± 10 vs. 25 ± 11% for mixed muscle protein FSR determined by using a leucine tracer in conjunction with plasma KIC enrichment as the surrogate precursor pool enrichment and either on- or off-line IRMS for protein analyses, respectively; P = 0.79). Furthermore, there was no relationship between the precursor enrichment and the population variance (e.g., r = 0.22, P = 0.30 in studies that used a phenylalanine tracer in combination with GCMS for protein analysis and muscle free phenylalanine enrichment as the precursor pool enrichment and report the muscle free phenylalanine enrichment; n = 25). Variability in the population variance of FSR values within studies that used the same tracer, surrogate precursor pool, and protein analysis technique was also not explained by the duration between biopsies (e.g., r = 0.15, P = 0.39 for studies using a phenylalanine tracer in combination with muscle free phenylalanine as the precursor pool and GCMS protein analysis; r = −0.05, P = 0.85 for studies using a leucine tracer in combination with plasma KIC as the precursor pool and IRMS for protein analysis).

DISCUSSION

Our goal was to elucidate potential reasons for the apparent discrepancy in reported muscle protein FSR values and to find out 1) whether the variability is due to differences in experimental design, 2) which of the many experimental approaches are most robust, and 3) whether “outliers” exist that appear to be operator dependent and thus should be considered unreliable. Overall, as expected (4, 5, 17, 77), the mixed muscle protein FSR values were significantly (P < 0.001) greater when the muscle vs. the plasma free amino acid enrichment was used as the precursor enrichment, and the average mixed muscle protein FSR values were significantly greater (P = 0.05) than the myofibrillar/myosin heavy chain FSR values. The within-study variability (i.e., population variance) was somewhat smaller in studies that used plasma amino acid/ketoacid enrichments vs. the muscle free amino acid enrichment, but this was not apparent in all circumstances. Furthermore, the between-study consistency of measured FSR values (i.e., interquartile range of average FSR values) was inversely correlated with the average duration between biopsies. Aside from that, the variation in reported FSR values could not be explained by differences in the experimental design and analytical methods (e.g., duration between biopsies, GCMS vs. IRMS for protein analysis, etc.), and none of the most commonly used approaches stood out as clearly superior in terms of consistency of results and within-study variability (i.e., population variance). Therefore, we conclude that the variability in reported values is in part due to known experimental bias (e.g., choice of precursor pool) and normal population variance (∼30% on average, which might be influenced by factors such as diet composition or habitual physical activity, etc.) and that whenever possible, one should aim to maximize the difference in protein enrichment (e.g., by extending the time period between biopsies, especially in circumstances when muscle protein synthesis rate is slow, such as during basal, postabsorptive conditions).

The most consistent (smallest interquartile ranges) average FSR values were obtained with mixed muscle protein FSR measured using a [¹³C]leucine tracer, IRMS for protein analysis, and either plasma KIC or muscle free leucine enrichments as the precursor enrichment. The actual range in FSR values (minimal to maximal) obtained using these two approaches, however, was not different from those in other studies and probably reflects the normal overall variability in the population. In fact, the average population variance reported for this approach, although on average somewhat smaller than for other approaches, was still quite variable and comparable in that regard to the variance in other studies. Therefore, we do not consider this approach a lot superior to others. Characteristic of the studies that yielded the most consistent results was the long measurement duration (4–12.5 h between biopsies). This was likely the most important factor for the good agreement in between-study results, because the same approach applied to the measurement of myofibrillar/myosin heavy chain protein FSR, with shorter measurement periods, does not result in similarly consistent results. However, protein separation techniques and purity of the final sample for analyses may also have affected the measured FSR value, because different muscle protein fractions turn over at different rates (54, 69).

Although one might expect that use of the plasma KIC enrichment (multiple samples) as the precursor enrichment in the FSR calculation may also have factored into the robustness of the results with mixed muscle protein FSR measured using a [¹³C]leucine tracer, IRMS for protein analysis, and plasma KIC as the precursor enrichment, it probably played a minor role, because use of the muscle free leucine enrichment (single sample) as the precursor enrichment but an otherwise similar approach (i.e., mixed muscle protein FSR, IRMS for protein enrichment analysis, and duration of measurement) yielded results that were just as consistent with respect to the magnitude of the interquartile range of the mean FSR values. Similarly, there was no difference in the variability of the mean FSR values obtained by using a phenylalanine tracer and either the plasma or muscle free phenylalanine enrichment as precursor enrichment when otherwise the approach was similar, most likely because the variability introduced by the shorter time between biopsies was greater than the reduction in variability by relying on a plasma-derived surrogate precursor pool. In fact, the difference in population variance between studies that used plasma amino acid/ketoacid enrichments vs. the muscle free amino acid enrichment as the precursor enrichment in the FSR calculation was very small. Choosing plasma-based surrogate precursor pools therefore does not seem to provide a major advantage, and the use of the intramuscular free amino acid enrichment as the precursor enrichment seems more appropriate, because the muscle free amino acid enrichment (regardless of tracer amino acid used) better reflects the true precursor enrichment (i.e., aminoacyl-tRNA) (17, 50, 88) and eliminates potential confounding and possibly erroneous results due to differences in plasma/muscle free amino acid equilibria (75).

The use of IRMS for protein analysis, which is considered superior to GCMS by many investigators because it is generally more sensitive and should therefore at least theoretically be better suited for measuring small enrichments, does not appear to provide an advantage either, most likely because investigators in the field are aware of this issue and adjust their analyses accordingly. We (60) have demonstrated that very low enrichments such as those encountered in studies of human muscle protein synthesis can be reliably measured by GCMS when careful attention is paid to the choice of tracer (multiple vs. single label), instrument condition (to achieve a signal-to-noise ratio of ≥100:1 for the tracer), and choice of base ion. In addition, the use of standards is critical to account for day-to-day variability in instrument response, which may vary considerably due to changes in the tuning conditions. For example, although the coefficients of variation for standards of known enrichment in our laboratory are 3.7% for [²H₅]phenylalanine and 4.8% for [²H₃]leucine (tert-butyldimethylsilyl derivatives; Hewlett Packard MSD 5973 System; electron impact mode), we have observed up to 20–30% day-to-day differences (minimal to maximal range; Smith GI, Patterson BW, and Mittendorfer B, unpublished observations). Thus good laboratory practices make it possible to obtain robust results with GCMS.

It has been pointed out to us that leucine tracers may not be suitable for measuring the rate of muscle protein synthesis because they may stimulate muscle protein synthesis (2, 27, 28). However, we are not aware of any studies that would support a stimulatory effect of trace amounts of leucine, or any other essential amino acid, and found no evidence for a relationship between the precursor enrichment (reflecting differences in the tracer infusion rate) and the measured muscle protein FSR value. Nevertheless, the potential stimulatory effect of the tracer amino acid on muscle protein synthesis is a major concern with the flooding dose method (78, 79) and indeed is not limited to leucine but may also include other essential amino acids, such as phenylalanine, because they stimulate muscle protein synthesis when given in large doses (79, 84).

In summary, our review of the literature indicates that there appears to be considerable normal physiological variation in muscle protein turnover during basal, postabsorptive conditions in healthy young and middle-aged adults that needs to be taken into account when planning and interpreting studies (e.g., statistical power). In addition, there are well-established method-dependent biases (e.g., choice of precursor pool affects the FSR value). The variability in reported FSR values appears to simply reflect this. We caution others not to rely on or compare results with “historic controls” (or assumed reference values), at least not without considering in detail the methods employed. The summary in Table 3 provides a guideline for “normal” average basal FSR values at rest in healthy young adults.

GRANTS

This study was supported by National Institutes of Health Grants AR 49869, HD 057796, DK 56341 (Clinical Nutrition Research Unit), and UL1 RR024992 (Washington University Institute of Clinical and Translational Sciences).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).