Ageing is accompanied by a progressive decline in muscle mass and strength (sarcopenia) and is associated with a lower quality of life due to the reduced ability of an individual to perform daily living activities( Reference Forbes, Little and Candow 1 ). It also predisposes people to the development of chronic metabolic disorders such as diabetes and obesity( Reference Park, Goodpaster and Strotmeyer 2 ). The prevalence of sarcopenia differs by sex and living settings( Reference Cheng, Zhu and Zhang 3 ). For example, age-related muscle loss has been reported to be prevalent in about 68 % of elderly men and 21 % of elderly women living in nursing homes( Reference Masanes, Culla and Navarro-Gonzalez 4 , Reference Landi, Liperoti and Fusco 5 ), but in about 10 % of men and 33 % of women living in the community( Reference Masanes, Culla and Navarro-Gonzalez 4 , Reference Janssen, Shepard and Katzmarzyk 6 ). Sarcopenia results in increased health care costs of approximately $18·5 billion per year in the USA( Reference Janssen, Shepard and Katzmarzyk 6 ).
Age-related muscle loss can result from a variety of modifiable factors including inadequate nutrition, oxidative stress, low physical activity levels, inflammation and reduced hormone concentrations( Reference Candow, Forbes and Little 7 ). Studies have suggested that muscles of the elderly may have a blunted protein synthesis response to food ingestion( Reference Volpi, Mittendorfer and Rasmussen 8 – Reference Cuthbertson, Smith and Babraj 10 ). A number of strategies to increase muscle mass in the elderly have been studied including different nutritional intervention strategies and physical exercise strategies( Reference Kim, Suzuki and Saito 11 – Reference Tieland, Dirks and van der Zwaluw 16 ); however, the findings have been conflicting( Reference Kim, Suzuki and Saito 11 – Reference Koopman, Verdijk and Manders 24 ).
The administration of dietary leucine increases muscle protein synthesis in vivo and in rodents( Reference Leenders and van Loon 25 ). It has been suggested that increasing leucine intake in the elderly may compensate for the blunted muscle protein synthesis response to food ingestion( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Rieu, Balage and Sornet 26 ). Several studies have found that increasing the amount of leucine in meals or in supplemental amino acid mixes increased the muscle protein synthesis response in the elderly( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Rieu, Balage and Sornet 26 ). In addition, essential amino acids (EAA) and leucine supplementation (sometimes given as whey protein) have increased protein synthesis in muscles, and are considered as better strategies for offsetting muscle loss than intact protein( Reference Solerte, Gazzaruso and Bonacasa 12 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 , Reference Leenders and van Loon 25 , Reference Paddon-Jones, Sheffield-Moore and Katsanos 27 , Reference Phillips 28 ). In contrast, other studies did not find any association of increased ingestion of leucine with elevated muscle protein fraction synthetic rate, muscle mass or strength in the elderly( Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Leenders, Verdijk and van der Hoeven 15 , Reference Koopman, Verdijk and Beelen 23 ). A limited number of studies have evaluated the effect of acute and chronic leucine supplementation on lean body mass and/or leg lean mass in elderly populations, and, overall, there have been inconsistent findings of whether leucine supplementation increases these outcomes( Reference Fukagawa 29 ).
Many of the studies evaluating the impact of leucine as a pharmaconutrient on age-related muscle loss have been small. To maximise the biostatistical power of controlled clinical trials, we performed a meta-analysis to assess the ability of leucine supplementation to increase muscle protein fraction synthetic rate, augment lean body mass or leg lean mass in elderly subjects.
Methods
Medline, Cochrane, EMBASE and Google Scholar databases were searched up to 31 December 2013 for clinical trials that investigated the administration of leucine as a nutrient that affects muscle protein metabolism and lean body mass and leg lean mass in elderly subjects. Searches were conducted using the following terms: elderly; elder; older; aging; aged; geriatric; leucine; muscle; muscular; randomized. Randomised controlled trials in which the majority of subjects were elderly (age ≥ 65 years) and that investigated the efficacy of a clearly defined level of leucine were included in the meta-analysis. Included studies were published in English. Excluded studies were non-randomised controlled trials, letters, comments, editorials and case reports. Potential relevant studies were screened by two independent reviewers, and both had to agree on study inclusion. Any disagreement between the reviewers was resolved by a third reviewer.
Data extraction
The following information was extracted from the studies that met the inclusion criteria: the name of the first author; year of publication; study design; demographics; leucine dosing regimen; exercise programme; muscle protein fractional synthetic rate; lean body mass; leg lean mass. Data were extracted by two independent reviewers, and a third reviewer was consulted if there were any uncertainties.
Quality assessment and publication bias
The quality of the studies was evaluated using the Cochrane Risk of Bias Tool to assess the included studies( Reference Higgins and Thompson 30 ). Quality assessment was also performed by two independent reviewers, and a third reviewer was consulted for any ambiguities. Due to the small number of selected studies, it was inappropriate to use the funnel plot for the assessment of publication bias. Therefore, five or fewer studies are not sufficient to detect funnel plot asymmetry( Reference Sutton, Duval and Tweedie 31 ).
Statistical analyses
The primary outcome for the meta-analysis was the protein fractional synthetic rate. Secondary outcomes included lean body mass and leg lean mass. Means and standard deviations or standard errors of means were used to summarise the outcomes before and after the intervention, and the change from baseline was used to evaluate the intervention effect. The standardised difference in mean changes with 95 % CI for subjects treated with leucine supplements (leucine group) compared with those treated with placebo or other nutritional supplements (control group) was calculated for each study. Heterogeneity among the studies was assessed by calculating Cochran's Q and the I 2 statistic( Reference Higgins and Thompson 30 , Reference Lau, Ioannidis and Schmid 32 ). For the Q statistic, P< 0·10 was considered to indicate statistically significant heterogeneity. The I 2 statistic indicates the percentage of the observed between-study variability caused by heterogeneity. Heterogeneity determined using the I 2 statistic was defined as follows: 0–24 %, no heterogeneity; 25–49 %, moderate heterogeneity; 50–74 %, large heterogeneity; 75–100 %, extreme heterogeneity. If heterogeneity existed between studies (a Q statistic with P< 0·1 or an I 2 statistic >50 %), we performed the random-effects model (DerSimonian–Laird method)( Reference DerSimonian and Laird 33 ). Otherwise, the fixed-effects model was used (Mantel–Haenszel method). The pooled standardised difference in mean changes was calculated, and a two-sided P value < 0·05 was considered to indicate statistical significance. Sensitivity analysis was performed for all three outcomes based on the leave-one-out approach. All statistical analyses were performed using the statistical software Comprehensive Meta-Analysis, version 2.0 (Biostat).
Results
Of the 525 studies identified, 492 were excluded and thirty-three underwent a full-text review. Of these, twenty-four were eliminated because they were not randomised controlled trials (n 5), the participants were not elderly (n 13), the amount of leucine administered was not clear (n 3), the outcomes were not the ones being investigated in the present analysis (n 1), there was no placebo control (n 1) or treatment groups were given the same amount of leucine (n 1) (Fig. 1; for the details of the excluded studies, see the online supplementary material). Finally, nine studies met the inclusion criteria( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Ferrando, Paddon-Jones and Hays 14 , Reference Leenders, Verdijk and van der Hoeven 15 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 , Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 , Reference Björkman, Pilvi and Kekkonen 34 , Reference Deutz, Safar and Schutzler 35 ).
Fig. 1 Flow chart for study selection.
Quality assessment
There was a low risk of data bias for the combination of the studies (Fig. 2) and for each individual study (Fig. 3), indicating that the data were of high quality.
Fig. 2 Risk of bias for the included studies. 
, Low risk of bias; 
, unclear risk of bias; 
, high risk of bias. A colour version of this figure can be found online at http://www.journals.cambridge.org/bjn
Fig. 3 Summary of risk of bias for the included studies. Green, low risk of bias; yellow, unclear risk of bias; red, high risk of bias. A colour version of this figure can be found online at http://www.journals.cambridge.org/bjn
Study characteristics
Among the nine included studies, six were randomised controlled trials with parallel treatment arms( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Leenders, Verdijk and van der Hoeven 15 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 , Reference Deutz, Safar and Schutzler 35 ) and three were randomised cross-over trials( Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 , Reference Björkman, Pilvi and Kekkonen 34 ) (Table 1). The total number of patients in the studies ranged from eight to fifty-seven, and the duration of intervention ranged from hours to 6 months (Table 1). Of these studies, four( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 , Reference Deutz, Safar and Schutzler 35 ) investigated the acute effect of leucine and administered leucine only once. The other five studies( Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Leenders, Verdijk and van der Hoeven 15 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 , Reference Björkman, Pilvi and Kekkonen 34 ) administered leucine as a long-term supplement with the length of intervention ranging from 10 d( Reference Ferrando, Paddon-Jones and Hays 14 ) to 6 months( Reference Leenders, Verdijk and van der Hoeven 15 ) (Table 1). Across the studies, the amount of leucine given for long-term supplementation ranged from 2·8 to 16·1 g/d, and for acute administration, it ranged from 2·6 to 17·6 g/d (Table 1). Among these studies, two( Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 ) included exercise as part of the intervention (Table 1). In five of the studies( Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Leenders, Verdijk and van der Hoeven 15 , Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 , Reference Deutz, Safar and Schutzler 35 ), all the subjects were male and five studies( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 , Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 ) included only healthy (or healthy and lean) subjects (Table 1).
Table 1 Characteristics of the studies included in the meta-analysis
RCT, randomised controlled trial; EAA, essential amino acids; CHO, carbohydrate; PRO, protein hydrolysate.
All of the included studies utilised a stable isotope infusion test to assess the muscle protein fractional synthetic rate from the mixed skeletal muscle protein (Table 2) and evaluated body composition by using dual-energy X-ray absorptiometry. The studies used the same measure of muscle protein fractional synthetic rate (%/h) and supplements were administered orally. In two studies that compared EAA with placebo, the muscle protein fractional synthetic rate was approximately 0·07 before long-term supplementation both for placebo and EAA, but decreased for placebo after 10 d to 3 months of treatments (Fig. 2)( Reference Ferrando, Paddon-Jones and Hays 14 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 ). There was little effect of long-term supplementation on lean body mass or leg lean mass (Table 2)( Reference Verhoeven, Vanschoonbeek and Verdijk 13 , Reference Ferrando, Paddon-Jones and Hays 14 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 , Reference Leenders and van Loon 25 ). In the four studies that used acute leucine administration, three studies showed a greater increase in muscle protein fractional synthetic rate from baseline with leucine supplementation than the control (Table 2)( Reference Koopman, Verdijk and Manders 24 , Reference Deutz, Safar and Schutzler 35 , Reference Katsanos, Chinkes and Paddon-Jones 36 ).
Table 2 Primary and secondary outcomes from the studies included in the meta-analysis (Mean values and standard deviations or standard errors)
NA, not available; EAA, essential amino acids; CHO, carbohydrate; PRO, protein hydrolysate.
Muscle protein fractional synthetic rate
Of the studies included in the meta-analysis, four assessed the effect of leucine on the fractional synthetic rate of muscle protein: three reported values before and after intervention( Reference Katsanos, Kobayashi and Sheffield-Moore 9 , Reference Ferrando, Paddon-Jones and Hays 14 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 ) and one reported the fractional synthetic rate as the change from baseline( Reference Deutz, Safar and Schutzler 35 ) (Table 2). After pooling of data, there was no significant heterogeneity found across the studies (heterogeneity test: Q= 4·36, df = 3, P= 0·225; I 2= 31·16 %); therefore, a fixed-effects model of analysis was used. The results indicated that the muscle protein fractional synthetic rate after intervention significantly increased in the leucine group compared with the control group (pooled standardised difference in mean changes 1·08, 95 % CI 0·50, 1·67; P< 0·001; Fig. 4).
Fig. 4 Forest plot showing the results for the meta-analysis of standardised difference (Std diff) in mean changes of muscle protein fractional synthetic rates after leucine v. control intervention.
Lean body mass
For the analysis of lean body mass, we included the four studies that reported lean body mass values both before and after leucine administration( Reference Verhoeven, Vanschoonbeek and Verdijk 13 – Reference Leenders, Verdijk and van der Hoeven 15 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 ). There was no significant heterogeneity found among the studies (heterogeneity test: Q= 2·37, df = 3, P= 0·499; I 2= 0·0 %); therefore, a fixed-effects model of analysis was used. The results showed that the change in lean body mass after intervention did not significantly differ between the leucine group and the control group (pooled standardised difference in mean changes 0·18, 95 % CI − 0·18, 0·54; P= 0·318; Fig. 5).
Fig. 5 Forest plot showing the results for the meta-analysis of standardised difference (Std diff) in mean changes of lean body mass after leucine v. control intervention.
Leg lean mass
Of the studies included in the meta-analysis, three( Reference Verhoeven, Vanschoonbeek and Verdijk 13 – Reference Leenders, Verdijk and van der Hoeven 15 ) reported leg lean mass findings from both before and after intervention. There was no significant heterogeneity found across the studies (heterogeneity test: Q= 0·59, df = 2, P= 0·752; I 2= 0·0 %); consequently, a fixed-effects model was used. There was no significant difference in change in leg lean mass after intervention between the subjects treated with leucine or placebo (pooled standardised difference in mean changes 0·006, 95 % CI − 0·32, 0·44; P= 0·756; Fig. 6).
Fig. 6 Forest plot showing the results for the meta-analysis of standardised difference (Std diff) in mean changes of leg lean mass after intervention: leucine v. control.
Sensitivity analysis
Sensitivity analysis was performed, in which the results were analysed when one study was removed in turn. The direction and magnitude of the pooled estimates for muscle protein fractional synthesis rate (Fig. 7(a)) and lean body mass and leg lean mass (Fig. 7(b) and (c)) did not vary substantially with the removal of any study from the analysis, indicating that one study did not influence the findings.
Fig. 7 Results of sensitivity analysis for the examination of the influence of individual studies on pooled estimates as determined using the leave-one-out approach: (a) muscle protein fractional synthetic rate; (b) lean body mass; (c) leg lean mass.
Discussion
The age-associated loss of skeletal muscle mass is an important factor in the loss of functional performance and the ability to maintain a healthy lifestyle in the elderly( Reference Leenders and van Loon 25 ). Sarcopenia is influenced by a combination of factors including poor diet and sedentary lifestyle( Reference Nair 37 ). In one study( Reference Katsanos, Kobayashi and Sheffield-Moore 9 ), it was found that increasing the leucine content in an amino acid mixture (from 26 to 41 %) could compensate for the blunted response to amino acid ingestion in the elderly, raising the idea that addition of leucine may be an effective strategy to normalise the postprandial response of muscle protein synthesis in the elderly. The present meta-analysis investigated whether leucine is an effective pharmaconutrient that could influence muscle protein fractional synthesis rates, lean body mass and leg lean mass in elderly subjects. We found that the addition of leucine increased the protein fractional synthesis rate compared with the control. Higher levels of leucine did not significantly affect lean body mass or leg lean mass even after long-term supplementation.
Leucine can have an impact on muscle mass in several ways including being a building block for protein synthesis, and also as a nutritional signal that acts via mTOR (mammalian target of rapamycin) in an insulin-dependent and -independent signalling cascade. The mTOR signalling pathway stimulates translation and protein synthesis by the phosphorylation of the translation initiation factor 4E-BP1 (eukaryotic translation initiation factor 4E-binding protein 1)( Reference Leenders and van Loon 25 ). Rodent studies indicated that the EAA leucine may represent an effective pharmaconutrient with the largest anabolic properties( Reference Leenders and van Loon 25 ), and leucine concentrations are thought to be the primary stimulus driving the postprandial response to muscle protein synthesis( Reference Kimball and Jefferson 38 ).
The effect of leucine on muscle protein metabolism is complex, as it influences both muscle protein synthesis and degradation, and this complexity may confound study findings. A prior study showed that intravenous infusion of leucine (0·14 g/kg body weight over a 7 h period) in healthy subjects resulted in decreased protein degradation by about 35 %( Reference Nair, Schwartz and Welle 39 ). Several studies have shown a significant increase (35–50 %) in the rates of muscle protein synthesis in healthy males with intravenous administration of amino acids, with leucine being an important component( Reference Bennet, Connacher and Scrimgeour 40 – Reference Smith, Barua and Watt 42 ). However, other studies have not detected an effect of increased leucine administration on muscle protein synthesis( Reference Koopman, Verdijk and Beelen 23 , Reference Louard, Barrett and Gelfand 43 ). The discrepancy among these studies may be due to the differences in the amount of leucine administered, the time of administration and the population studied. Of the studies included in the present meta-analysis, both the levels of leucine and the duration of dosing differed. For example, five studies had long-term dosing (>10 d) and four had short-term dosing ( ≤ 8 h). However, one of the studies with short dosing time found an improvement in the muscle protein fractional synthesis rate( Reference Katsanos, Kobayashi and Sheffield-Moore 9 ), suggesting that the length of dosing is not the only factor influencing the results. Instead, the study showed that increasing the proportion of leucine in a mixture of EAA may increase the muscle protein fractional synthesis rate( Reference Katsanos, Chinkes and Paddon-Jones 36 ).
There are several reasons for the observed increase in muscle protein synthesis but not in lean muscle mass or leg lean mass with leucine supplementation. One idea is consistent with the ‘anabolic threshold concept’, suggesting that in elderly people, changes in key signalling pathways and changes in catabolic factors and oxidative stress may have negative effects on amino acid or insulin signalling pathways that play a role in the stimulation of muscle anabolism after food intake( Reference Dardevet, Remond and Peyron 44 ). These changes may lead to ‘anabolic resistance’ of muscle, such that there is a requirement for higher anabolic stimuli to promote maximal anabolism and protein retention. Another idea, which is not mutually exclusive, is that the muscle becomes refractory (or ‘full’) when exposed to the persistent levels of amino acid concentrations, independent of the mode of amino acid delivery( Reference Dideriksen, Reitelseder and Holm 22 ). Ingestion of leucine or a protein meal transiently increases myofibrillar protein synthesis after an approximate 45 min delay from intake, for about 45–90 min after which synthesis rapidly declines to pre-intake rates( Reference Atherton, Etheridge and Watt 45 ). The increase in muscle protein synthesis probably results from the activation of processes that regulate mRNA translation. The decline in protein synthesis occurs even in the continued presence of amino acids, suggesting that muscles have a mechanism for regulating the synthesis of new proteins.
The long-term effect of leucine on muscle mass is not clear. Leucine supplementation is known to interact with other key pathways such as insulin signalling and glucose metabolism pathways( Reference Di Camillo, Eduati and Nair 46 ). Long-term treatment with leucine attenuates insulin secretory dysfunction of human diabetic islets via the up-regulation of certain key metabolic genes, and in vivo leucine administration improves glycaemic control in human subjects and rodents with type 2 diabetes( Reference Yang, Chi and Burkhardt 47 ). These findings may have implications for the association between leucine supplementation and type 2 diabetes. In addition, leucine may also attenuate adiposity by increasing fatty oxidation and mitochondrial biogenesis in adipocytes and muscle tissue. In one study, it has been suggested that leucine may be useful in the management of obesity and obesity-related co-morbidities by increasing fat oxidation and reducing oxidative stress and inflammation( Reference Zemel and Bruckbauer 48 ). Muscle contraction also appears to influence muscle protein synthesis; it strongly stimulates muscle protein synthesis and also increases muscle protein degradation, but to a lesser extent, resulting in an improved net muscle protein balance( Reference Leenders and van Loon 25 ). The studies of Koopman et al. ( Reference Koopman, Verdijk and Beelen 23 , Reference Koopman, Verdijk and Manders 24 ) investigated the effect of leucine administration following physical activity in elderly men who received ample amounts of dietary protein on whole-body protein turnover and muscle protein synthetic rate compared with the administration of controls. They found that muscle fraction synthetic rates were not different between the groups. These findings suggest that leucine supplementation and exercise did not further elevate the rates of muscle protein synthesis in elderly men who received ample amounts of protein. The lack of enhancement from additional leucine and exercise may reflect the fact that the subjects were already receiving ample amounts of protein in their diets( Reference Leenders and van Loon 25 ). The influence of protein intake in a subject's diet in the course of a study on whether leucine supplementation does or does not affect muscle protein synthesis is a confounding factor that may, at least in part, explain the discrepancies among the studies. It is possible that long-term leucine supplementation would be more clinically relevant in malnourished elderly or in specific clinical subpopulations( Reference Leenders and van Loon 25 ).
The longest study included in the present meta-analysis was 3 months in duration, raising the issue of the effects of longer-term administration of leucine on muscle protein synthesis and other metabolic processes. In two studies that were not included in the present meta-analysis, the effect of long-term leucine supplementation was evaluated. Zeanandin et al. ( Reference Zeanandin, Balage and Schneider 49 ) evaluated the effects of 6 months of dietary leucine administration on insulin signalling and sensitivity in elderly rats (18 months of age). Rats were fed a 15 % protein diet with or without 4·5 % leucine. They found that the mTOR pathway was not significantly altered in muscle, and glucose tolerance was not changed. No change in skeletal muscle mass was observed, although perirenal adipose tissue mass accumulated in the leucine-supplemented mass. These findings suggest that the effect of leucine is somewhat tissue specific. Guo et al. ( Reference Guo, Yu and Hou 50 ) assessed the metabolic effects of leucine supplementation in an obese/diabetic mouse model, and found that leucine supplementation for 8 months significantly improved glycaemic control, and that the effects of leucine probably acts by multiple mechanisms in different tissues.
All the studies included in the meta-analysis were randomised controlled trials. However, the studies investigated different populations, leucine levels and dosing regimens. This heterogeneity in experimental designs and subject populations could have influenced the findings, and indicates the need for additional studies with more similar experimental designs to address whether leucine supplementation can be used for the normalisation of protein synthesis in the elderly. The pooled results for muscle protein fractional synthetic rates (pooled standardised difference in mean changes 1·08, 95 % CI 0·50, 1·67; P< 0·001) were determined from a combination of acute and long-term supplementation outcomes. Among the four studies comparing the changes in muscle protein fractional synthetic rates between two groups, two used long-term supplementation( Reference Ferrando, Paddon-Jones and Hays 14 , Reference Dillon, Sheffield-Moore and Paddon-Jones 17 ) and two employed acute administration of leucine supplementation( Reference Deutz, Safar and Schutzler 35 , Reference Katsanos, Chinkes and Paddon-Jones 36 ). However, all the four studies reported significant differences in the fractional synthetic rate of muscle protein between the intervention and control groups. Thus, our findings suggest that either long-term or acute leucine supplementation could increase the muscle protein fractional synthetic rate.
In conclusion, we found that ingestion of leucine significantly increased the muscle protein fractional synthetic rate in elderly individuals, and thus may be of benefit to address sarcopenia in this population.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S0007114514002475
Acknowledgements
The present study was supported by the Science and Technology Project of Zhejiang Province (grant no. 2013C33122).
The authors' contributions are as follows: Z.-r. X. contributed to the manuscript preparation and literature search, and was the guarantor of the integrity of the entire study and responsible for the definition of the intellectual content; Z.-j. T. contributed to the definition of the intellectual content, literature research and data acquisition; Q. Z. performed the data analysis and statistical analyses; Q.-f. G. contributed to the manuscript editing; Y.-m. Y. contributed to the study concept, study design and manuscript review.
There are no conflicts of interest.
