Letters to the Editor
Additional analyses in a study on the obesity paradox Dear Sir:
The authors did not declare any conflicts of interest.
Renate M Winkels Zephenia Gomora Moniek van Zutphen Ellen Kampman
1208
REFERENCES 1. Gonzalez MC, Pastore CA, Orlandi SP, Heymsfield SB. Obesity paradox in cancer: new insights provided by body composition. Am J Clin Nutr 2014;99:999–1005. 2. Vrieling A, Kampman E. The role of body mass index, physical activity, and diet in colorectal cancer recurrence and survival: a review of the literature. Am J Clin Nutr 2010;92:471–90. doi: 10.3945/ajcn.114.092742.
Reply to RM Winkels et al Dear Sir: We thank Winkels et al for their interest in our article and their important suggestions. Perhaps we conveyed the wrong message that we ascribe to the obesity paradox concept. On the contrary, we found that when BMI is used to define obesity, the results are misleading and that misperception can be resolved with the use of body composition measurements. As they suggest, the use of the lowest BMI as the reference can give the impression of an obesity paradox. When using as a reference subjects with a BMI (in kg/m2) of 18.5–24.9 we found that the HRs for overweight or obesity compared with normal weight were 0.62 (95% CI: 0.32, 1.22) and 0.74 (95% CI: 0.36, 1.52), respectively. Although decreased in intensity, the effect on risk remained in the same direction: appearance of a protective effect for BMI .25. However, we reiterate that a higher BMI value is not necessarily synonymous with large fat mass. The HR for underweight, as we might expect, becomes an important risk factor with an HR of 6.11 (95% CI: 2.16, 17.28). When we reported that excess fat mass had no protective effect in the presence of low skeletal muscle mass, we based this conclusion on a multivariate Cox regression model. Fat mass index (FMI) was not significantly associated with mortality after controlling for all of the other potential covariates. In fact, the multivariate analysis HR is 1.51 (95% CI: 0.58, 3.92), although this is not statistically significant. As recommended by Winkels et al, the interaction of fat mass and the fat-free mass index (FFMI) was explored by testing the following 4 categories: low FFMI and normal FMI, low FFMI and high FMI, normal FFMI and high FMI, and normal FFMI and normal FMI (reference group). The results showed the importance of a low FFMI as a risk factor, with almost no effect of FMI. Considering normal FFMI and normal FMI as the reference group, a low FFMI
Am J Clin Nutr 2014;100:1208–14. Printed in USA. Ó 2014 American Society for Nutrition
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
In a recent study, ‘‘Obesity paradox in cancer: new insights provided by body composition,’’ Gonzalez et al (1) studied the association of body composition with cancer survival in a prospective cohort study in 175 patients with cancer. The current literature on the role of BMI in colorectal cancer survival is not consistent (2), and part of the inconsistency may be explained by differences in body composition. We therefore highly appreciate the contribution of Gonzalez et al to further study the role of body composition in colorectal cancer survival. Yet, we do have some remarks on the way in which the authors analyzed their data and would like to encourage the authors to comment on the following. First, it is not possible to judge whether there is an obesity paradox from the Cox regression analysis that the authors presented. The authors’ Figure 2A and Table 3 suggest that patients who are overweight or obese according to BMI may have the most beneficial survival rates. Indeed, this may hint toward an ‘‘obesity paradox.’’ Unfortunately, it is not possible to judge from the Cox regression analysis in Table 4 whether those analyses also point toward this obesity paradox, because the authors decided to use the lowest-BMI group (underweight group) as the reference group for BMI. Because of this reference group, the comparison that is made is one of being underweight compared with having normal weight, overweight, or obesity. There would be a greater scientific interest to show the HRs for overweight or obesity compared with being of normal weight (thus, with normal weight as the reference group). Second, the association between fat mass index (FMI) and survival may be different for patients with a low fat-free mass index (FFMI) compared with those with a normal FFMI, but the authors do not fully explore this in their Cox regression. The authors themselves state in their discussion that ‘‘excess FM [fat mass] had no protective effect in the presence of low skeletal muscle mass,’’ yet this is not assessed in the data analysis of the Cox regression model. We would like to encourage the authors to explore the interaction of FMI and FFMI by showing the HRs for the combinations of low FFMI and normal FMI, low FFMI and high FMI, and normal FFMI and high FMI compared with normal FFMI and normal FMI (reference group) to be able to truly assess this possible interaction.
Division of Human Nutrition Wageningen University PO Box 8129 6700 EV Wageningen Netherlands E-mail: [email protected]
1209
LETTERS TO THE EDITOR increases the HR to 48.0 (95% CI: 9.11, 253.1) when associated with a high FMI (sarcopenic obesity) or 3.83 (95% CI: 1.79, 8.18) when the FMI is normal (sarcopenia alone). The HR of obese patients (normal FFMI and high FMI) increased minimally to 1.05 (95% CI: 0.39, 2.82) and shows the absence of an obesity paradox. We thank Winkels et al for their suggestions that have helped guide us in clarifying our conclusions related to the obesity paradox and providing us with the opportunity to explore other combinations of the variables. These efforts confirm that fat-free mass is the component that offers a protective effect for the survival in this sample of patients with cancer and that obesity when classified only by BMI can convey the misimpression of an obesity paradox. The authors had no conflicts of interest to declare.
The author did not declare any conflicts of interest.
Christopher Griffin Obstetrics and Gynecology Clinical Care Unit Carson House 374 Bagot Road Subiaco, Western Australia 6008 Australia E-mail: [email protected]
REFERENCE 1. Lindsay KL, Kennelly M, Culliton M, Maguire OC, Smith T, Shanahan F, Brennan L, McAuliffe FM. Probiotics in obese pregnancy do not reduce maternal fasting glucose: a randomized controlled trial. Am J Clin Nutr 2014;99:1432–9.
Postgraduate Program of Health and Behavior Catholic University of Pelotas Rua Vereador Ariano Requia˜o de Carvalho, 301 Pelotas - Rio Grande do Sul CEP 96055-800 Brazil E-mail: [email protected] Carla A Pastore Silvana P Orlandi Nutrition College of Nutrition Federal University of Pelotas Rio Grande do Sul Brazil Steven B Heymsfield Pennington Biomedical Research Center Baton Rouge, LA
doi: 10.3945/ajcn.114.093054.
Probiotic choices and biological plausibility for metabolic studies in pregnancy Dear Sir: I read with great interest the study published by Lindsay et al (1). I have the following questions: Does Lactobacillus salivarius have appropriate in vitro metabolic research to support its immediate translation into the realm of human clinical trials? What was the biological plausibility for treating only during 24–28 wk and for only 4 wk in total? What was the antibiotic use in each group? Did the ingestion of fermented milk products, live probiotics, or prebiotics feature in the dietary questionnaire? I look forward to the reply.
Reply to C Griffin Dear Sir: We thank Griffin for his interest in our study (1). He raises some important questions with regard to the study methodology and probiotic effects to which we are happy to respond. The Lactobacillus salivarius UCC118 probiotic, which was originally isolated from the human ileal-cecal region (2), has undergone several in vitro and in vivo studies to investigate its effects (3). In vitro studies of this strain have predominantly reported antiinflammatory and immune-modulatory effects (4) as well as the probiotic’s successful ability to survive and transit through the gastrointestinal tract (GIT) (5, 6). Randomized controlled trials using L. salivarius UCC118 have also shown its transit through the GIT in both animal (7) and human (8) studies, as well as its ability to modify the gut microflora and engage the immune system in healthy adults (8). However, its in vivo metabolic effects are less well characterized. A recently published trial of L. salivarius UCC118 or vancomycin compared with a control in diet-induced obese mice showed beneficial alteration of gut flora with the use of both treatments (9). However, only vancomycin had metabolic benefits in terms of lower fasting blood glucose, TNF-a, and triglyceride concentrations compared with controls, whereas L. salivarius UCC118 had no such effects (9). Although previous studies of probiotics in pregnancy have used longer intervention periods, the 4-wk intervention period used in our study (1) was decided on the basis of previously conducted research with the use of the L. salivarius UCC118 strain in healthy adults, which showed successful transit through the GIT and modulation of the gut microflora after 21 d (8). Therefore, it was considered that if the previously reported beneficial metabolic effects of probiotics in pregnancy (10) were to be replicated in our study, a 4-wk intervention period should be sufficient while also minimizing burden on participants and increasing the likelihood of compliance. The period of 24–28 wk of gestation was chosen because this interval directly precedes the routine glucose tolerance test performed in our center at 28–29 wk of gestation, thereby coinciding with routine clinical appointments and blood sample collection. Furthermore, glucose intolerance typically first presents in pregnancy from 28 wk onward, and thus the preceding 4 wk was considered an ideal time to trial an intervention for its prevention.
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
doi: 10.3945/ajcn.114.094235.
Maria Cristina Gonzalez
1210
LETTERS TO THE EDITOR
The authors declared no conflicts of interest.
Karen L Lindsay
REFERENCES 1. Lindsay KL, Kennelly M, Culliton M, Maguire OC, Smith T, Shanahan F, Brennan L, McAuliffe FM. Probiotics in obese pregnancy do not reduce maternal fasting glucose: a randomized controlled trial. Am J Clin Nutr 2014;99:1432–9. 2. Thornton GM. Probiotic bacteria. Selection of Lactobacillus and Bifidobacterium strains from the healthy human gastrointestinal tract; characterisation of a novel Lactobacillus derived antibacterial protein. PhD dissertation. National University of Ireland, Dublin, Ireland 1996. 3. Neville BA, O’Toole PW. Probiotic properties of Lactobacillus salivarius and closely related Lactobacillus species. Future Microbiol 2010;5:759–74. 4. McCarthy J, O’Mahony L, O’Callaghan L, Sheil B, Vaughan EE, Fitzsimons N, Fitzgibbon J, O’Sullivan GC, Kiely B, Collins JK, et al. Double blind placebo-controlled trial of two probiotic strains in interleukin 10 knockout mice and mechanistic link with cytokine balance. Gut 2003;52:975–80. 5. Dunne C, Murphy L, Flynn S, O’Mahony L, O’Halloran S, Feeney M, Morrissey D, Thornton G, Fitzgerald G, Daly C, et al. Probiotics: from myth to reality. Demonstration of functionality in animal models of disease and in human clinical trials. Antonie van Leeuwenhoek 1999; 76:279–92. 6. van Pijkeren JP, Canchava C, Ryan KA, Li Y, Claesson MJ, Sheil B, Steidler L, O’Mahony L, Fitzgerald GF, van Sinderen D, et al. Comparative and functional analysis of sortase-dependent proteins in the predicted secretome of Lactobacillus salivarius UCC118. Appl Environ Microbiol 2006;72:4143–53. 7. Murphy L, Dunne C, Kiely B, Shanahan F, O’Sullivan GC, Collins JK. In vivo assessment of potential Lactobacillus salivarius strains: evaluation of their establishment, persistence, and localization in the murine gastrointestinal tract. Microb Ecol Health Dis 1999;11:149–57. 8. Collins JK, Dunne C, Murphy L, Morrissey D, O’Mahony L, O’Sullivan E, Fitzgerald G, Kiely B, O’Sullivan GC, Daly C, et al. A randomized controlled trial of a probiotic Lactobacillus strain in healthy adults: assessment of its delivery, transit, and influence on microbial flora and enteric immunity. Microb Ecol Health Dis 2002;14:81–9. 9. Murphy EF, Cotter PD, Hogan A, O’Sullivan O, Joyce A, Fouhy F, Clarke SF, Marques TM, O’Toole PW, Stanton C, et al. Divergent metabolic outcomes arising from targeted manipulation of the gut microbiota in diet-induced obesity. Gut 2013;62:220–6. 10. Laitinen K, Poussa T, Isolauri E; Nutrition, Allergy, Mucosal Immunology and Intestinal Microbiota Group. Probiotics and dietary counselling contribute to glucose regulation during and after pregnancy: a randomised controlled trial. Br J Nutr 2009;101:1679–87.
UCD Obstetrics and Gynecology School of Medicine and Medical Science University College Dublin National Maternity Hospital Dublin 2 Ireland
doi: 10.3945/ajcn.114.094425.
Protein requirements and aging Lorraine Brennan UCD Institute of Food and Health School of Agriculture and Food Science University College Dublin Belfield Dublin 4 Ireland Fionnuala M McAuliffe UCD Obstetrics and Gynecology School of Medicine and Medical Science University College Dublin National Maternity Hospital Dublin 2 Ireland E-mail: [email protected]
Dear Sir: The recent article involving the use of the indicator amino acid oxidation (IAAO) method to assess protein requirements of octogenarian women (1) represents yet another attempt to show that there is an increase in protein requirements with age, a debate that has existed for decades. One reason for lack of resolution of this debate is that protein and amino acid metabolism is by far the most elaborate of any nutrient. Assuming that the protein requirement is an intake that allows maintenance of an acceptable body composition phenotype and associated normal function, we know that this can occur in population groups exposed to a wide range of habitual protein intakes, through metabolically complex adaptations. Evaluating exactly how adaptations to variation in protein intakes occur, at what cost, if any, and the lower and upper limits of protein intakes at which successful adaptation can occur is extremely challenging. In the absence of functional indicators of protein status of the adult population, all methods to date have been based on some measure of protein
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Antibiotic usage during or around the time of a probiotic intervention would certainly influence any potential effects, and this was therefore considered in our study. Women were asked before and after completion of the capsule intervention if they had taken any antibiotics, and a total of 6 women (2 from the probiotic group, 4 from the placebo group) confirmed that they had. Although details of antibiotic strain or duration of usage were not recorded, a secondary analysis of results that excluded antibiotic users and poor compliers to the capsule intervention showed that no significant differences in any of the metabolic variables or pregnancy outcomes remained (1). As highlighted by Griffin, the ingestion of fermented milk products, live probiotics, or prebiotics by our study participants could also have influenced our results. All participants were asked to avoid fermented milk and probiotic and prebiotic products from recruitment until the end of their pregnancy, allowing a sufficient ‘‘washout’’ period before capsule commencement at 24 wk of gestation. To aid in this, an information sheet outlining the sources of fermented, probiotic, and prebiotic products on the market was developed and explained by the research dietitian to all study participants on recruitment at 12–19 wk of gestation, along with a list of appropriate nonprobiotic yogurts available in local shops. However, we did not directly investigate the consumption of fermented or probiotic products because we gathered information on dietary intakes during the intervention period by using a 3-d food diary rather than a food-frequency questionnaire. We hope we have sufficiently answered the queries raised by Griffin and have instilled confidence in the robustness of our trial’s methodology. As we conclude in our article, further research is warranted into the effects of probiotics in pregnancy on metabolic outcomes, particularly among obese women who may be at higher risk of adverse outcomes. Furthermore, there is a need to establish the optimal species, timing, and dosage of probiotics that may benefit this important patient group.
LETTERS TO THE EDITOR
FIGURE 1. Concentrations of the ‘‘indicator’’ amino acid (phenylalanine) in meal protein intakes relative to those of a balanced intake at each protein amount. Values shown are calculated from studies in 6 octogenarian women (1) and are the amounts of phenylalanine in each amount of the test meals of protein relative to its content that would have occurred if the test amino acid mixture pattern was that of the reference (egg) amino acid pattern (n). Also shown are the reported rates of expired 13 CO2 measured at the end of each test amount of intake, which reflect phenylalanine oxidation rates (X). In the studies, L-[1-13C]phenylalanine oxidation was measured during the feeding of frequent small meals of the test intake. These test intakes contained the varying amounts of an amino acid mixture patterned on egg protein that was equivalent to the daily protein intakes as shown but with a fixed amount of phenylalanine (30.5 mg kg21 d21) and tyrosine (40.7 mg kg21 d21). This results in a relative excess of phenylalanine at low intakes, limiting amounts at high intakes, and a balanced intake at protein intakes equivalent to 0.56 g kg21 d21. It is argued that this variation in the relative concentration of the phenylalanine ‘‘indicator’’ with protein intake will be the primary influence on the reported shape and breakpoint of the indicator oxidation curve.
in the 4 highest intakes. Because of this, the indicator oxidation rate, shown in Figure 1, reflects the excess or deficiency of the indicator, not the amount of protein intake. The authors refer to the criticism by Millward and Jackson (4) of their approach in their article (1) and argue that ‘‘regardless of the protein intake, the total aromatic amino acid concentration is always 70 mg kg21 d21, which is higher than the aromatic amino acid requirement and thus could not be ‘balanced’ by increasing protein intake. With sufficient tyrosine, the concept of the IAAO approach is to keep the phenylalanine content constant and sufficient at any protein (amino acid mixture) level to reflect the protein oxidation rate.’’ This statement shows a lack of understanding of the postprandial response to varying protein intakes. In fact, with this study design, the protein oxidation rate, which is not measured, will be the opposite of the observed phenylalanine oxidation rate. As pointed out previously (4), on the basis of many published tracer studies of the feeding response, it can be confidently predicted that at low amounts of intake the meal protein will be fully used with low levels of overall amino acid oxidation but with high levels of [13C]phenylalanine oxidation because of its excess. However, as the intake of the amino acid mixture exceeds 0.6 g kg21 d21, overall utilization of the amino acid mixture for net protein synthesis will decrease as it becomes limited by the relative availability of phenylalanine. In consequence, overall amino acid oxidation will increase and [13C]phenylalanine oxidation will decrease. The validity of this argument is easily tested by measuring the response of blood concentrations of phenylalanine (predicted to be high in the excess intake range and low in the deficient range), changes that will be the opposite of other amino acids (eg, leucine). The argument that the total aromatic amino acid concentration always exceeds the aromatic amino acid requirement is irrelevant because in such acute feeding studies the ‘‘requirement’’ for phenylalanine is that which allows efficient utilization of the meal amino acid mixture for net protein deposition. As shown in Figure 1, the meals are phenylalanine deficient at intakes .0.6 g kg21 d21. Thus, these studies tell us nothing about the protein requirement of octogenarian women. It is the case that previous studies by the lead author of this article showed no difference with age in the protein requirements of adults as measured by both nitrogen balance (5) or by [1-13C]leucine balance (6). These 2 reports are separate publications from the same study that, together, comprise the most comprehensive study in the literature on the protein requirements of healthy adults. The study shows quite clearly no effect of age and sex, similar to our own findings (7), with the authors concluding that ‘‘there are no compelling data that the dietary protein needs of old people are different from those of young people when expressed per kg body weight.’’ It is a puzzle, therefore, that in this most recent study (1), the [1-13C]leucine balance article (6) is not quoted at all and the nitrogen balance arm of the study (5) is only briefly mentioned together with a list of reports arguing for an increased protein requirement, none of which include any unequivocal evidence. One would expect experienced investigators to have a consistent message in their published work or at the least explain why they have changed their view. My understanding of the literature in terms of well-conducted nitrogen balance or 13C oxidation studies is that the experimental evidence to date shows that requirement values do not change significantly with advancing age. As indicated in an editorial about this recent article (8), what is really needed are studies that show that incremental increases in protein intake make a difference—ie, that they do affect clinically important outcomes. Sarcopenia has been widely discussed as a potential consequence of inadequate protein intake, although there is very little, if any, unequivocal evidence that the loss of muscle mass and function with age can be influenced by protein intake (9). In the absence of
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
balance, either nitrogen balance or amino acid balance, through measures of amino acid oxidation by using stable isotopes. Concern for inadequacies of the nitrogen balance approach has resulted in investigators adopting acute postprandial studies, such as the IAAO method (1), to evaluate the response to protein or amino acid intakes as a proxy for the ‘‘requirement.’’ Assuming here that such studies can show useful information, it is certainly necessary that investigators adopting postprandial experimental protocols fully understand the metabolic complexity of the response of protein metabolism to protein intake and show that the model assumptions inherent in their studies are correct and that any metabolic response or endpoint does indeed directly relate to the ‘‘requirement.’’ The principal advocates of the IAAO method, Pencharz and Ball (2), have always argued that the change in the oxidation rate, the breakpoint, of a nonlimiting indicator amino acid ([1-13C]phenylalanine) in response to graded intakes of a test amino acid or of protein shows the intake that maximizes protein synthesis and minimizes indicator oxidation: ie, their definition of the ‘‘requirement.’’ In response to a study of the protein requirement of healthy school-aged children determined by the IAAO method (3), Millward and Jackson (4) argued that the use of the IAAO method to assess protein requirements, as opposed to requirements for amino acids, was invalid. This is because, in this specific case, the [13C]phenylalanine indicator does become limiting and this limitation determines the breakpoint. The experimental design of this approach measures [13C]phenylalanine oxidation in response to meals containing increasing amounts of protein (as an amino acid mixture based on egg protein) containing a fixed amount of phenylalanine. This is shown in Figure 1, which plots phenylalanine content of protein meals at each amount of ‘‘protein’’ intake expressed as the content relative to the amount that would have been present if the amino acid mixture was balanced. Thus, intakes of phenylalanine are in excess at the 2 lowest intakes and are deficient
1211
1212
LETTERS TO THE EDITOR
clinical outcomes from well-conducted randomized controlled trials, the identification of a suitable experimental approach that could be adopted by different investigators could allow the requisite much larger numbers of volunteers to be studied and might settle the debate if agreement could be reached on a suitable method. It is quite clear to me that the IAAO method could not serve such a purpose. The author had no conflicts of interest.
D Joe Millward
REFERENCES 1. Tang M, McCabe GP, Elango R, Pencharz PB, Ball RO, Campbell WW. Assessment of protein requirement in octogenarian women with use of the indicator amino acid oxidation technique. Am J Clin Nutr 2014;99: 891–8. 2. Pencharz PB, Ball RO. Different approaches to define individual amino acid requirements. Annu Rev Nutr 2003;23:101–16. 3. Elango R, Humayun MA, Ball RO, Pencharz PB. Protein requirement of healthy school-age children determined by the indicator amino acid oxidation method. Am J Clin Nutr 2011;94:1545–52. 4. Millward DJ, Jackson AA. Protein requirements and the indicator amino acid oxidation method. Am J Clin Nutr 2012;95:1498–501; author reply 1501–2. 5. Campbell WW, Johnson CA, McCabe GP, Carnell NS. Dietary protein requirements of younger and older adults. Am J Clin Nutr 2008;88:1322–9. 6. Conley TB, McCabe GP, Lim E, Yarasheski KE, Johnson CA, Campbell WW. Age and sex affect protein metabolism at protein intakes that span the range of adequacy: comparison of leucine kinetics and nitrogen balance data. J Nutr Biochem 2013;24:693–9. 7. Millward DJ, Fereday A, Gibson N, Pacy PJ. Aging, protein requirements, and protein turnover. Am J Clin Nutr 1997;66:774–86. 8. Fukagawa NK. Protein requirements: methodologic controversy amid a call for change. Am J Clin Nutr 2014;99:761–2. 9. Millward DJ. Nutrition and sarcopenia: evidence for an interaction. Proc Nutr Soc 2012;71:566–75. doi: 10.3945/ajcn.114.089540.
Reply to DJ Millward Dear Sir: With Millward’s letter, he continues to criticize the merits of the indicator amino acid oxidation (IAAO) method to assess human protein requirements. The issues raised are old and repeatedly described and discussed in the literature, as well as carefully considered during a rigorous review of the article before acceptance for publication, and noted in Fukagawa’s editorial published with the article (1). Apparently, Millward’s chief criticism of the IAAO method is that the intake of the indicator amino acid phenylalanine was inadequate at the higher intakes of protein and thus the breakpoint in the response was due to a deficiency of phenylalanine. He used a circular argument with regard to whether phenylalanine intake was adequate or deficient.
The authors had no conflicts of interest to declare.
Minghua Tang Department of Nutrition Science Purdue University West Lafayette, IN George P McCabe Department of Statistics Purdue University West Lafayette, IN
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Department of Nutritional Sciences School of Biosciences and Medicine Faculty of Health and Medical Sciences University of Surrey Guildford GU2 7XH United Kingdom E-mail: [email protected]
We have previously measured the phenylalanine requirement in the presence of an excess of all other amino acids (ie, protein), including tyrosine, and found it to be 13.6 mg kg21 d21 (population upper 95% CI) (2). During the present experiment, we provided phenylalanine at an intake of 30.5 mg kg21 d 21, which is well in excess of the phenylalanine requirement in the presence of excess tyrosine (40 mg kg21 d 21) and other amino acids. Therefore, very clearly, phenylalanine would not be limiting at any intake of protein in the current experiment. In which case, as we discussed in the article, the breakpoint was due to the plateau in protein synthesis that occurred when the intake of protein was adequate. To argue that phenylalanine was deficient because it plateaued in oxidation is therefore incorrect, and ignores the extensive work done to show how the IAAO method works. Two main principles of the IAAO method are 1) that the excess intake of phenylalanine is proportioned between protein synthesis and oxidation and 2) phenylalanine oxidation progressively declines with increasing intake of the limiting amino acid (or of total protein) from deficient to adequate and reaches a steady nadir (breakpoint) when a sufficient amount is consumed. Millward has suggested that the intake of phenylalanine, as the indicator, be allowed to change in proportion to the protein intake. If there was a change in the intake of phenylalanine along with protein intake, then the percentage of dose oxidized would not change with each increment of protein because the degree of excess of phenylalanine would be the same for every protein intake. This means that the percentage of dose oxidized would not vary or vary only very little with intakes between deficient and adequate, and it would increase thereafter because phenylalanine would be in excess. This is the same as using the direct oxidation approach, which has other well-recognized issues (3). This is also the key reason why the indicator amino acid must be controlled to the same intake in all treatments; otherwise, the slope of the response line may be due to changes in intake of the indicator amino acid rather than intake of the test protein or amino acid (3). Millward repeatedly argues that the requirement of every indispensable amino acid varies directly with protein intake, the ‘‘adaptive metabolic demand.’’ However, this theory is untested by direct experimentation. In contrast, there is no evidence currently available that the requirement for any amino acid, other than the single most limiting amino acid in the diet, varies with protein intake. This principle of the limiting amino acid defining the protein requirement is the principle whereby the amino acid score of a protein is derived. If one accepts that amino acid score is a valid concept, as Millward does (4), then ‘‘adaptive metabolic demand’’ cannot also be correct.
LETTERS TO THE EDITOR
Rajavel Elango Department of Pediatrics University of British Columbia Vancouver Canada
1213 Wayne W Campbell
Department of Nutrition Science Purdue University 700 West State Street West Lafayette, IN 47907 E-mail: [email protected] Paul B Pencharz
Department of Pediatrics and Nutrition Science University of Toronto Toronto Canada
REFERENCES
Department of Agricultural, Food, and Nutrition Science University of Alberta Edmonton Canada
doi: 10.3945/ajcn.114.090324.
Erratum Mason C, Xiao L, Imayama I, Duggan C, Wang C-Y, Korde L, McTiernan A. Vitamin D3 supplementation during weight loss: a double-blind randomized controlled trial. Am J Clin Nutr 2014;99:1015–25. The units for C-reactive protein (CRP) are incorrectly stated as ‘‘mg/mL’’ in the following places: the Results section of the Abstract on page 1015, the Results section of the text on page 1020, Table 2 on page 1020, Table 3 on page 1021, and Table 4 on page 1022. The correct unit is ‘‘mg/L.’’ doi: 10.3945/ajcn.114.095448.
Erratum Gale C, Thomas EL, Jeffries S, Durighel G, Logan KM, Parkinson JRC, Uthaya S, Santhakumaran S, Bell JD, Modi N. Adiposity and hepatic lipid in healthy full-term, breastfed, and formula-fed human infants: a prospective short-term longitudinal cohort study. Am J Clin Nutr 2014;99:1034–40. Because of a copyediting error, the abbreviation for magnetic resonance spectroscopy (MRS) incorrectly appears in the fourth paragraph of the Subjects and Methods section. The first sentence of the fourth paragraph [‘‘Whole-body MRS images were acquired on a Phillips 1.5 Tesla system by using a T1-weighted rapid-spin-echo sequence (repetition time of 500 ms, echo time of 17 ms, echo train length of 3) by using a Q body coil.’’] should read as follows: ‘‘Whole body magnetic resonance images were acquired on a Phillips 1.5 Tesla system by using a T1-weighted rapid-spin-echo sequence (repetition time of 500 ms, echo time of 17 ms, echo train length of 3) by using a Q body coil.’’ The fifth sentence of the fourth paragraph [‘‘All MRS images were analyzed independently of the investigators and blind to participant identity and feeding group by VardisGroup (www.vardisgroup.com) by using an image segmentation program (SliceOmatic; Tomovision).’’] should read as follows: ‘‘All magnetic resonance images were analyzed independently of the investigators and blind to participant identity and feeding group by VardisGroup (www.vardisgroup.com) by using an image segmentation program (SliceOmatic; Tomovision).’’ doi: 10.3945/ajcn.114.095620.
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Ronald O Ball
1. Fukagawa NK. Protein requirements: methodologic controversy amid a call for change. Am J Clin Nutr 2014;99:761–2. 2. Zello GA, Pencharz PB, Ball RO. Phenylalanine flux, oxidation and conversion to tyrosine in humans studied with L-(1-13C)-phenylalanine. Am J Physiol 1990;259:E835–43. 3. Pencharz PB, Ball RO. Different approaches to define individual amino acid requirements. Annu Rev Nutr 2003;23:101–16. 4. Millward DJ. Amino acid scoring patterns for protein quality assessment. Br J Nutr 2012;108:S31–43.
1214
LETTERS TO THE EDITOR
Erratum Ryan MF, O’Grada CM, Morris C, Segurado R, Walsh MC, Gibney ER, Brennan L, Roche HM, Gibney MJ. Within-person variation in the postprandial lipemic response of healthy adults. Am J Clin Nutr 2013;97:261–7. An author’s name is spelled incorrectly. In the author list on page 261, ‘‘Colm O Grada’’ should be ‘‘Colm M O’Grada.’’ The initials ‘‘COG’’ in footnote 1 on page 261 and in the Acknowledgments on page 266 should be ‘‘CMO.’’ In the Table of Contents, the third article listed under ‘‘Lipids’’ should read as follows: ‘‘Within-person variation in the postprandial lipemic response of healthy adults. MF Ryan, CM O’Grada, C Morris, R Segurado, MC Walsh, ER Gibney, L Brennan, HM Roche, and MJ Gibney.’’ doi: 10.3945/ajcn.114.095638.
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Additional analyses in a study on the obesity paradox Dear Sir:
The authors did not declare any conflicts of interest.
Renate M Winkels Zephenia Gomora Moniek van Zutphen Ellen Kampman
1208
REFERENCES 1. Gonzalez MC, Pastore CA, Orlandi SP, Heymsfield SB. Obesity paradox in cancer: new insights provided by body composition. Am J Clin Nutr 2014;99:999–1005. 2. Vrieling A, Kampman E. The role of body mass index, physical activity, and diet in colorectal cancer recurrence and survival: a review of the literature. Am J Clin Nutr 2010;92:471–90. doi: 10.3945/ajcn.114.092742.
Reply to RM Winkels et al Dear Sir: We thank Winkels et al for their interest in our article and their important suggestions. Perhaps we conveyed the wrong message that we ascribe to the obesity paradox concept. On the contrary, we found that when BMI is used to define obesity, the results are misleading and that misperception can be resolved with the use of body composition measurements. As they suggest, the use of the lowest BMI as the reference can give the impression of an obesity paradox. When using as a reference subjects with a BMI (in kg/m2) of 18.5–24.9 we found that the HRs for overweight or obesity compared with normal weight were 0.62 (95% CI: 0.32, 1.22) and 0.74 (95% CI: 0.36, 1.52), respectively. Although decreased in intensity, the effect on risk remained in the same direction: appearance of a protective effect for BMI .25. However, we reiterate that a higher BMI value is not necessarily synonymous with large fat mass. The HR for underweight, as we might expect, becomes an important risk factor with an HR of 6.11 (95% CI: 2.16, 17.28). When we reported that excess fat mass had no protective effect in the presence of low skeletal muscle mass, we based this conclusion on a multivariate Cox regression model. Fat mass index (FMI) was not significantly associated with mortality after controlling for all of the other potential covariates. In fact, the multivariate analysis HR is 1.51 (95% CI: 0.58, 3.92), although this is not statistically significant. As recommended by Winkels et al, the interaction of fat mass and the fat-free mass index (FFMI) was explored by testing the following 4 categories: low FFMI and normal FMI, low FFMI and high FMI, normal FFMI and high FMI, and normal FFMI and normal FMI (reference group). The results showed the importance of a low FFMI as a risk factor, with almost no effect of FMI. Considering normal FFMI and normal FMI as the reference group, a low FFMI
Am J Clin Nutr 2014;100:1208–14. Printed in USA. Ó 2014 American Society for Nutrition
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
In a recent study, ‘‘Obesity paradox in cancer: new insights provided by body composition,’’ Gonzalez et al (1) studied the association of body composition with cancer survival in a prospective cohort study in 175 patients with cancer. The current literature on the role of BMI in colorectal cancer survival is not consistent (2), and part of the inconsistency may be explained by differences in body composition. We therefore highly appreciate the contribution of Gonzalez et al to further study the role of body composition in colorectal cancer survival. Yet, we do have some remarks on the way in which the authors analyzed their data and would like to encourage the authors to comment on the following. First, it is not possible to judge whether there is an obesity paradox from the Cox regression analysis that the authors presented. The authors’ Figure 2A and Table 3 suggest that patients who are overweight or obese according to BMI may have the most beneficial survival rates. Indeed, this may hint toward an ‘‘obesity paradox.’’ Unfortunately, it is not possible to judge from the Cox regression analysis in Table 4 whether those analyses also point toward this obesity paradox, because the authors decided to use the lowest-BMI group (underweight group) as the reference group for BMI. Because of this reference group, the comparison that is made is one of being underweight compared with having normal weight, overweight, or obesity. There would be a greater scientific interest to show the HRs for overweight or obesity compared with being of normal weight (thus, with normal weight as the reference group). Second, the association between fat mass index (FMI) and survival may be different for patients with a low fat-free mass index (FFMI) compared with those with a normal FFMI, but the authors do not fully explore this in their Cox regression. The authors themselves state in their discussion that ‘‘excess FM [fat mass] had no protective effect in the presence of low skeletal muscle mass,’’ yet this is not assessed in the data analysis of the Cox regression model. We would like to encourage the authors to explore the interaction of FMI and FFMI by showing the HRs for the combinations of low FFMI and normal FMI, low FFMI and high FMI, and normal FFMI and high FMI compared with normal FFMI and normal FMI (reference group) to be able to truly assess this possible interaction.
Division of Human Nutrition Wageningen University PO Box 8129 6700 EV Wageningen Netherlands E-mail: [email protected]
1209
LETTERS TO THE EDITOR increases the HR to 48.0 (95% CI: 9.11, 253.1) when associated with a high FMI (sarcopenic obesity) or 3.83 (95% CI: 1.79, 8.18) when the FMI is normal (sarcopenia alone). The HR of obese patients (normal FFMI and high FMI) increased minimally to 1.05 (95% CI: 0.39, 2.82) and shows the absence of an obesity paradox. We thank Winkels et al for their suggestions that have helped guide us in clarifying our conclusions related to the obesity paradox and providing us with the opportunity to explore other combinations of the variables. These efforts confirm that fat-free mass is the component that offers a protective effect for the survival in this sample of patients with cancer and that obesity when classified only by BMI can convey the misimpression of an obesity paradox. The authors had no conflicts of interest to declare.
The author did not declare any conflicts of interest.
Christopher Griffin Obstetrics and Gynecology Clinical Care Unit Carson House 374 Bagot Road Subiaco, Western Australia 6008 Australia E-mail: [email protected]
REFERENCE 1. Lindsay KL, Kennelly M, Culliton M, Maguire OC, Smith T, Shanahan F, Brennan L, McAuliffe FM. Probiotics in obese pregnancy do not reduce maternal fasting glucose: a randomized controlled trial. Am J Clin Nutr 2014;99:1432–9.
Postgraduate Program of Health and Behavior Catholic University of Pelotas Rua Vereador Ariano Requia˜o de Carvalho, 301 Pelotas - Rio Grande do Sul CEP 96055-800 Brazil E-mail: [email protected] Carla A Pastore Silvana P Orlandi Nutrition College of Nutrition Federal University of Pelotas Rio Grande do Sul Brazil Steven B Heymsfield Pennington Biomedical Research Center Baton Rouge, LA
doi: 10.3945/ajcn.114.093054.
Probiotic choices and biological plausibility for metabolic studies in pregnancy Dear Sir: I read with great interest the study published by Lindsay et al (1). I have the following questions: Does Lactobacillus salivarius have appropriate in vitro metabolic research to support its immediate translation into the realm of human clinical trials? What was the biological plausibility for treating only during 24–28 wk and for only 4 wk in total? What was the antibiotic use in each group? Did the ingestion of fermented milk products, live probiotics, or prebiotics feature in the dietary questionnaire? I look forward to the reply.
Reply to C Griffin Dear Sir: We thank Griffin for his interest in our study (1). He raises some important questions with regard to the study methodology and probiotic effects to which we are happy to respond. The Lactobacillus salivarius UCC118 probiotic, which was originally isolated from the human ileal-cecal region (2), has undergone several in vitro and in vivo studies to investigate its effects (3). In vitro studies of this strain have predominantly reported antiinflammatory and immune-modulatory effects (4) as well as the probiotic’s successful ability to survive and transit through the gastrointestinal tract (GIT) (5, 6). Randomized controlled trials using L. salivarius UCC118 have also shown its transit through the GIT in both animal (7) and human (8) studies, as well as its ability to modify the gut microflora and engage the immune system in healthy adults (8). However, its in vivo metabolic effects are less well characterized. A recently published trial of L. salivarius UCC118 or vancomycin compared with a control in diet-induced obese mice showed beneficial alteration of gut flora with the use of both treatments (9). However, only vancomycin had metabolic benefits in terms of lower fasting blood glucose, TNF-a, and triglyceride concentrations compared with controls, whereas L. salivarius UCC118 had no such effects (9). Although previous studies of probiotics in pregnancy have used longer intervention periods, the 4-wk intervention period used in our study (1) was decided on the basis of previously conducted research with the use of the L. salivarius UCC118 strain in healthy adults, which showed successful transit through the GIT and modulation of the gut microflora after 21 d (8). Therefore, it was considered that if the previously reported beneficial metabolic effects of probiotics in pregnancy (10) were to be replicated in our study, a 4-wk intervention period should be sufficient while also minimizing burden on participants and increasing the likelihood of compliance. The period of 24–28 wk of gestation was chosen because this interval directly precedes the routine glucose tolerance test performed in our center at 28–29 wk of gestation, thereby coinciding with routine clinical appointments and blood sample collection. Furthermore, glucose intolerance typically first presents in pregnancy from 28 wk onward, and thus the preceding 4 wk was considered an ideal time to trial an intervention for its prevention.
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
doi: 10.3945/ajcn.114.094235.
Maria Cristina Gonzalez
1210
LETTERS TO THE EDITOR
The authors declared no conflicts of interest.
Karen L Lindsay
REFERENCES 1. Lindsay KL, Kennelly M, Culliton M, Maguire OC, Smith T, Shanahan F, Brennan L, McAuliffe FM. Probiotics in obese pregnancy do not reduce maternal fasting glucose: a randomized controlled trial. Am J Clin Nutr 2014;99:1432–9. 2. Thornton GM. Probiotic bacteria. Selection of Lactobacillus and Bifidobacterium strains from the healthy human gastrointestinal tract; characterisation of a novel Lactobacillus derived antibacterial protein. PhD dissertation. National University of Ireland, Dublin, Ireland 1996. 3. Neville BA, O’Toole PW. Probiotic properties of Lactobacillus salivarius and closely related Lactobacillus species. Future Microbiol 2010;5:759–74. 4. McCarthy J, O’Mahony L, O’Callaghan L, Sheil B, Vaughan EE, Fitzsimons N, Fitzgibbon J, O’Sullivan GC, Kiely B, Collins JK, et al. Double blind placebo-controlled trial of two probiotic strains in interleukin 10 knockout mice and mechanistic link with cytokine balance. Gut 2003;52:975–80. 5. Dunne C, Murphy L, Flynn S, O’Mahony L, O’Halloran S, Feeney M, Morrissey D, Thornton G, Fitzgerald G, Daly C, et al. Probiotics: from myth to reality. Demonstration of functionality in animal models of disease and in human clinical trials. Antonie van Leeuwenhoek 1999; 76:279–92. 6. van Pijkeren JP, Canchava C, Ryan KA, Li Y, Claesson MJ, Sheil B, Steidler L, O’Mahony L, Fitzgerald GF, van Sinderen D, et al. Comparative and functional analysis of sortase-dependent proteins in the predicted secretome of Lactobacillus salivarius UCC118. Appl Environ Microbiol 2006;72:4143–53. 7. Murphy L, Dunne C, Kiely B, Shanahan F, O’Sullivan GC, Collins JK. In vivo assessment of potential Lactobacillus salivarius strains: evaluation of their establishment, persistence, and localization in the murine gastrointestinal tract. Microb Ecol Health Dis 1999;11:149–57. 8. Collins JK, Dunne C, Murphy L, Morrissey D, O’Mahony L, O’Sullivan E, Fitzgerald G, Kiely B, O’Sullivan GC, Daly C, et al. A randomized controlled trial of a probiotic Lactobacillus strain in healthy adults: assessment of its delivery, transit, and influence on microbial flora and enteric immunity. Microb Ecol Health Dis 2002;14:81–9. 9. Murphy EF, Cotter PD, Hogan A, O’Sullivan O, Joyce A, Fouhy F, Clarke SF, Marques TM, O’Toole PW, Stanton C, et al. Divergent metabolic outcomes arising from targeted manipulation of the gut microbiota in diet-induced obesity. Gut 2013;62:220–6. 10. Laitinen K, Poussa T, Isolauri E; Nutrition, Allergy, Mucosal Immunology and Intestinal Microbiota Group. Probiotics and dietary counselling contribute to glucose regulation during and after pregnancy: a randomised controlled trial. Br J Nutr 2009;101:1679–87.
UCD Obstetrics and Gynecology School of Medicine and Medical Science University College Dublin National Maternity Hospital Dublin 2 Ireland
doi: 10.3945/ajcn.114.094425.
Protein requirements and aging Lorraine Brennan UCD Institute of Food and Health School of Agriculture and Food Science University College Dublin Belfield Dublin 4 Ireland Fionnuala M McAuliffe UCD Obstetrics and Gynecology School of Medicine and Medical Science University College Dublin National Maternity Hospital Dublin 2 Ireland E-mail: [email protected]
Dear Sir: The recent article involving the use of the indicator amino acid oxidation (IAAO) method to assess protein requirements of octogenarian women (1) represents yet another attempt to show that there is an increase in protein requirements with age, a debate that has existed for decades. One reason for lack of resolution of this debate is that protein and amino acid metabolism is by far the most elaborate of any nutrient. Assuming that the protein requirement is an intake that allows maintenance of an acceptable body composition phenotype and associated normal function, we know that this can occur in population groups exposed to a wide range of habitual protein intakes, through metabolically complex adaptations. Evaluating exactly how adaptations to variation in protein intakes occur, at what cost, if any, and the lower and upper limits of protein intakes at which successful adaptation can occur is extremely challenging. In the absence of functional indicators of protein status of the adult population, all methods to date have been based on some measure of protein
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Antibiotic usage during or around the time of a probiotic intervention would certainly influence any potential effects, and this was therefore considered in our study. Women were asked before and after completion of the capsule intervention if they had taken any antibiotics, and a total of 6 women (2 from the probiotic group, 4 from the placebo group) confirmed that they had. Although details of antibiotic strain or duration of usage were not recorded, a secondary analysis of results that excluded antibiotic users and poor compliers to the capsule intervention showed that no significant differences in any of the metabolic variables or pregnancy outcomes remained (1). As highlighted by Griffin, the ingestion of fermented milk products, live probiotics, or prebiotics by our study participants could also have influenced our results. All participants were asked to avoid fermented milk and probiotic and prebiotic products from recruitment until the end of their pregnancy, allowing a sufficient ‘‘washout’’ period before capsule commencement at 24 wk of gestation. To aid in this, an information sheet outlining the sources of fermented, probiotic, and prebiotic products on the market was developed and explained by the research dietitian to all study participants on recruitment at 12–19 wk of gestation, along with a list of appropriate nonprobiotic yogurts available in local shops. However, we did not directly investigate the consumption of fermented or probiotic products because we gathered information on dietary intakes during the intervention period by using a 3-d food diary rather than a food-frequency questionnaire. We hope we have sufficiently answered the queries raised by Griffin and have instilled confidence in the robustness of our trial’s methodology. As we conclude in our article, further research is warranted into the effects of probiotics in pregnancy on metabolic outcomes, particularly among obese women who may be at higher risk of adverse outcomes. Furthermore, there is a need to establish the optimal species, timing, and dosage of probiotics that may benefit this important patient group.
LETTERS TO THE EDITOR
FIGURE 1. Concentrations of the ‘‘indicator’’ amino acid (phenylalanine) in meal protein intakes relative to those of a balanced intake at each protein amount. Values shown are calculated from studies in 6 octogenarian women (1) and are the amounts of phenylalanine in each amount of the test meals of protein relative to its content that would have occurred if the test amino acid mixture pattern was that of the reference (egg) amino acid pattern (n). Also shown are the reported rates of expired 13 CO2 measured at the end of each test amount of intake, which reflect phenylalanine oxidation rates (X). In the studies, L-[1-13C]phenylalanine oxidation was measured during the feeding of frequent small meals of the test intake. These test intakes contained the varying amounts of an amino acid mixture patterned on egg protein that was equivalent to the daily protein intakes as shown but with a fixed amount of phenylalanine (30.5 mg kg21 d21) and tyrosine (40.7 mg kg21 d21). This results in a relative excess of phenylalanine at low intakes, limiting amounts at high intakes, and a balanced intake at protein intakes equivalent to 0.56 g kg21 d21. It is argued that this variation in the relative concentration of the phenylalanine ‘‘indicator’’ with protein intake will be the primary influence on the reported shape and breakpoint of the indicator oxidation curve.
in the 4 highest intakes. Because of this, the indicator oxidation rate, shown in Figure 1, reflects the excess or deficiency of the indicator, not the amount of protein intake. The authors refer to the criticism by Millward and Jackson (4) of their approach in their article (1) and argue that ‘‘regardless of the protein intake, the total aromatic amino acid concentration is always 70 mg kg21 d21, which is higher than the aromatic amino acid requirement and thus could not be ‘balanced’ by increasing protein intake. With sufficient tyrosine, the concept of the IAAO approach is to keep the phenylalanine content constant and sufficient at any protein (amino acid mixture) level to reflect the protein oxidation rate.’’ This statement shows a lack of understanding of the postprandial response to varying protein intakes. In fact, with this study design, the protein oxidation rate, which is not measured, will be the opposite of the observed phenylalanine oxidation rate. As pointed out previously (4), on the basis of many published tracer studies of the feeding response, it can be confidently predicted that at low amounts of intake the meal protein will be fully used with low levels of overall amino acid oxidation but with high levels of [13C]phenylalanine oxidation because of its excess. However, as the intake of the amino acid mixture exceeds 0.6 g kg21 d21, overall utilization of the amino acid mixture for net protein synthesis will decrease as it becomes limited by the relative availability of phenylalanine. In consequence, overall amino acid oxidation will increase and [13C]phenylalanine oxidation will decrease. The validity of this argument is easily tested by measuring the response of blood concentrations of phenylalanine (predicted to be high in the excess intake range and low in the deficient range), changes that will be the opposite of other amino acids (eg, leucine). The argument that the total aromatic amino acid concentration always exceeds the aromatic amino acid requirement is irrelevant because in such acute feeding studies the ‘‘requirement’’ for phenylalanine is that which allows efficient utilization of the meal amino acid mixture for net protein deposition. As shown in Figure 1, the meals are phenylalanine deficient at intakes .0.6 g kg21 d21. Thus, these studies tell us nothing about the protein requirement of octogenarian women. It is the case that previous studies by the lead author of this article showed no difference with age in the protein requirements of adults as measured by both nitrogen balance (5) or by [1-13C]leucine balance (6). These 2 reports are separate publications from the same study that, together, comprise the most comprehensive study in the literature on the protein requirements of healthy adults. The study shows quite clearly no effect of age and sex, similar to our own findings (7), with the authors concluding that ‘‘there are no compelling data that the dietary protein needs of old people are different from those of young people when expressed per kg body weight.’’ It is a puzzle, therefore, that in this most recent study (1), the [1-13C]leucine balance article (6) is not quoted at all and the nitrogen balance arm of the study (5) is only briefly mentioned together with a list of reports arguing for an increased protein requirement, none of which include any unequivocal evidence. One would expect experienced investigators to have a consistent message in their published work or at the least explain why they have changed their view. My understanding of the literature in terms of well-conducted nitrogen balance or 13C oxidation studies is that the experimental evidence to date shows that requirement values do not change significantly with advancing age. As indicated in an editorial about this recent article (8), what is really needed are studies that show that incremental increases in protein intake make a difference—ie, that they do affect clinically important outcomes. Sarcopenia has been widely discussed as a potential consequence of inadequate protein intake, although there is very little, if any, unequivocal evidence that the loss of muscle mass and function with age can be influenced by protein intake (9). In the absence of
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
balance, either nitrogen balance or amino acid balance, through measures of amino acid oxidation by using stable isotopes. Concern for inadequacies of the nitrogen balance approach has resulted in investigators adopting acute postprandial studies, such as the IAAO method (1), to evaluate the response to protein or amino acid intakes as a proxy for the ‘‘requirement.’’ Assuming here that such studies can show useful information, it is certainly necessary that investigators adopting postprandial experimental protocols fully understand the metabolic complexity of the response of protein metabolism to protein intake and show that the model assumptions inherent in their studies are correct and that any metabolic response or endpoint does indeed directly relate to the ‘‘requirement.’’ The principal advocates of the IAAO method, Pencharz and Ball (2), have always argued that the change in the oxidation rate, the breakpoint, of a nonlimiting indicator amino acid ([1-13C]phenylalanine) in response to graded intakes of a test amino acid or of protein shows the intake that maximizes protein synthesis and minimizes indicator oxidation: ie, their definition of the ‘‘requirement.’’ In response to a study of the protein requirement of healthy school-aged children determined by the IAAO method (3), Millward and Jackson (4) argued that the use of the IAAO method to assess protein requirements, as opposed to requirements for amino acids, was invalid. This is because, in this specific case, the [13C]phenylalanine indicator does become limiting and this limitation determines the breakpoint. The experimental design of this approach measures [13C]phenylalanine oxidation in response to meals containing increasing amounts of protein (as an amino acid mixture based on egg protein) containing a fixed amount of phenylalanine. This is shown in Figure 1, which plots phenylalanine content of protein meals at each amount of ‘‘protein’’ intake expressed as the content relative to the amount that would have been present if the amino acid mixture was balanced. Thus, intakes of phenylalanine are in excess at the 2 lowest intakes and are deficient
1211
1212
LETTERS TO THE EDITOR
clinical outcomes from well-conducted randomized controlled trials, the identification of a suitable experimental approach that could be adopted by different investigators could allow the requisite much larger numbers of volunteers to be studied and might settle the debate if agreement could be reached on a suitable method. It is quite clear to me that the IAAO method could not serve such a purpose. The author had no conflicts of interest.
D Joe Millward
REFERENCES 1. Tang M, McCabe GP, Elango R, Pencharz PB, Ball RO, Campbell WW. Assessment of protein requirement in octogenarian women with use of the indicator amino acid oxidation technique. Am J Clin Nutr 2014;99: 891–8. 2. Pencharz PB, Ball RO. Different approaches to define individual amino acid requirements. Annu Rev Nutr 2003;23:101–16. 3. Elango R, Humayun MA, Ball RO, Pencharz PB. Protein requirement of healthy school-age children determined by the indicator amino acid oxidation method. Am J Clin Nutr 2011;94:1545–52. 4. Millward DJ, Jackson AA. Protein requirements and the indicator amino acid oxidation method. Am J Clin Nutr 2012;95:1498–501; author reply 1501–2. 5. Campbell WW, Johnson CA, McCabe GP, Carnell NS. Dietary protein requirements of younger and older adults. Am J Clin Nutr 2008;88:1322–9. 6. Conley TB, McCabe GP, Lim E, Yarasheski KE, Johnson CA, Campbell WW. Age and sex affect protein metabolism at protein intakes that span the range of adequacy: comparison of leucine kinetics and nitrogen balance data. J Nutr Biochem 2013;24:693–9. 7. Millward DJ, Fereday A, Gibson N, Pacy PJ. Aging, protein requirements, and protein turnover. Am J Clin Nutr 1997;66:774–86. 8. Fukagawa NK. Protein requirements: methodologic controversy amid a call for change. Am J Clin Nutr 2014;99:761–2. 9. Millward DJ. Nutrition and sarcopenia: evidence for an interaction. Proc Nutr Soc 2012;71:566–75. doi: 10.3945/ajcn.114.089540.
Reply to DJ Millward Dear Sir: With Millward’s letter, he continues to criticize the merits of the indicator amino acid oxidation (IAAO) method to assess human protein requirements. The issues raised are old and repeatedly described and discussed in the literature, as well as carefully considered during a rigorous review of the article before acceptance for publication, and noted in Fukagawa’s editorial published with the article (1). Apparently, Millward’s chief criticism of the IAAO method is that the intake of the indicator amino acid phenylalanine was inadequate at the higher intakes of protein and thus the breakpoint in the response was due to a deficiency of phenylalanine. He used a circular argument with regard to whether phenylalanine intake was adequate or deficient.
The authors had no conflicts of interest to declare.
Minghua Tang Department of Nutrition Science Purdue University West Lafayette, IN George P McCabe Department of Statistics Purdue University West Lafayette, IN
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Department of Nutritional Sciences School of Biosciences and Medicine Faculty of Health and Medical Sciences University of Surrey Guildford GU2 7XH United Kingdom E-mail: [email protected]
We have previously measured the phenylalanine requirement in the presence of an excess of all other amino acids (ie, protein), including tyrosine, and found it to be 13.6 mg kg21 d21 (population upper 95% CI) (2). During the present experiment, we provided phenylalanine at an intake of 30.5 mg kg21 d 21, which is well in excess of the phenylalanine requirement in the presence of excess tyrosine (40 mg kg21 d 21) and other amino acids. Therefore, very clearly, phenylalanine would not be limiting at any intake of protein in the current experiment. In which case, as we discussed in the article, the breakpoint was due to the plateau in protein synthesis that occurred when the intake of protein was adequate. To argue that phenylalanine was deficient because it plateaued in oxidation is therefore incorrect, and ignores the extensive work done to show how the IAAO method works. Two main principles of the IAAO method are 1) that the excess intake of phenylalanine is proportioned between protein synthesis and oxidation and 2) phenylalanine oxidation progressively declines with increasing intake of the limiting amino acid (or of total protein) from deficient to adequate and reaches a steady nadir (breakpoint) when a sufficient amount is consumed. Millward has suggested that the intake of phenylalanine, as the indicator, be allowed to change in proportion to the protein intake. If there was a change in the intake of phenylalanine along with protein intake, then the percentage of dose oxidized would not change with each increment of protein because the degree of excess of phenylalanine would be the same for every protein intake. This means that the percentage of dose oxidized would not vary or vary only very little with intakes between deficient and adequate, and it would increase thereafter because phenylalanine would be in excess. This is the same as using the direct oxidation approach, which has other well-recognized issues (3). This is also the key reason why the indicator amino acid must be controlled to the same intake in all treatments; otherwise, the slope of the response line may be due to changes in intake of the indicator amino acid rather than intake of the test protein or amino acid (3). Millward repeatedly argues that the requirement of every indispensable amino acid varies directly with protein intake, the ‘‘adaptive metabolic demand.’’ However, this theory is untested by direct experimentation. In contrast, there is no evidence currently available that the requirement for any amino acid, other than the single most limiting amino acid in the diet, varies with protein intake. This principle of the limiting amino acid defining the protein requirement is the principle whereby the amino acid score of a protein is derived. If one accepts that amino acid score is a valid concept, as Millward does (4), then ‘‘adaptive metabolic demand’’ cannot also be correct.
LETTERS TO THE EDITOR
Rajavel Elango Department of Pediatrics University of British Columbia Vancouver Canada
1213 Wayne W Campbell
Department of Nutrition Science Purdue University 700 West State Street West Lafayette, IN 47907 E-mail: [email protected] Paul B Pencharz
Department of Pediatrics and Nutrition Science University of Toronto Toronto Canada
REFERENCES
Department of Agricultural, Food, and Nutrition Science University of Alberta Edmonton Canada
doi: 10.3945/ajcn.114.090324.
Erratum Mason C, Xiao L, Imayama I, Duggan C, Wang C-Y, Korde L, McTiernan A. Vitamin D3 supplementation during weight loss: a double-blind randomized controlled trial. Am J Clin Nutr 2014;99:1015–25. The units for C-reactive protein (CRP) are incorrectly stated as ‘‘mg/mL’’ in the following places: the Results section of the Abstract on page 1015, the Results section of the text on page 1020, Table 2 on page 1020, Table 3 on page 1021, and Table 4 on page 1022. The correct unit is ‘‘mg/L.’’ doi: 10.3945/ajcn.114.095448.
Erratum Gale C, Thomas EL, Jeffries S, Durighel G, Logan KM, Parkinson JRC, Uthaya S, Santhakumaran S, Bell JD, Modi N. Adiposity and hepatic lipid in healthy full-term, breastfed, and formula-fed human infants: a prospective short-term longitudinal cohort study. Am J Clin Nutr 2014;99:1034–40. Because of a copyediting error, the abbreviation for magnetic resonance spectroscopy (MRS) incorrectly appears in the fourth paragraph of the Subjects and Methods section. The first sentence of the fourth paragraph [‘‘Whole-body MRS images were acquired on a Phillips 1.5 Tesla system by using a T1-weighted rapid-spin-echo sequence (repetition time of 500 ms, echo time of 17 ms, echo train length of 3) by using a Q body coil.’’] should read as follows: ‘‘Whole body magnetic resonance images were acquired on a Phillips 1.5 Tesla system by using a T1-weighted rapid-spin-echo sequence (repetition time of 500 ms, echo time of 17 ms, echo train length of 3) by using a Q body coil.’’ The fifth sentence of the fourth paragraph [‘‘All MRS images were analyzed independently of the investigators and blind to participant identity and feeding group by VardisGroup (www.vardisgroup.com) by using an image segmentation program (SliceOmatic; Tomovision).’’] should read as follows: ‘‘All magnetic resonance images were analyzed independently of the investigators and blind to participant identity and feeding group by VardisGroup (www.vardisgroup.com) by using an image segmentation program (SliceOmatic; Tomovision).’’ doi: 10.3945/ajcn.114.095620.
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
Ronald O Ball
1. Fukagawa NK. Protein requirements: methodologic controversy amid a call for change. Am J Clin Nutr 2014;99:761–2. 2. Zello GA, Pencharz PB, Ball RO. Phenylalanine flux, oxidation and conversion to tyrosine in humans studied with L-(1-13C)-phenylalanine. Am J Physiol 1990;259:E835–43. 3. Pencharz PB, Ball RO. Different approaches to define individual amino acid requirements. Annu Rev Nutr 2003;23:101–16. 4. Millward DJ. Amino acid scoring patterns for protein quality assessment. Br J Nutr 2012;108:S31–43.
1214
LETTERS TO THE EDITOR
Erratum Ryan MF, O’Grada CM, Morris C, Segurado R, Walsh MC, Gibney ER, Brennan L, Roche HM, Gibney MJ. Within-person variation in the postprandial lipemic response of healthy adults. Am J Clin Nutr 2013;97:261–7. An author’s name is spelled incorrectly. In the author list on page 261, ‘‘Colm O Grada’’ should be ‘‘Colm M O’Grada.’’ The initials ‘‘COG’’ in footnote 1 on page 261 and in the Acknowledgments on page 266 should be ‘‘CMO.’’ In the Table of Contents, the third article listed under ‘‘Lipids’’ should read as follows: ‘‘Within-person variation in the postprandial lipemic response of healthy adults. MF Ryan, CM O’Grada, C Morris, R Segurado, MC Walsh, ER Gibney, L Brennan, HM Roche, and MJ Gibney.’’ doi: 10.3945/ajcn.114.095638.
Downloaded from ajcn.nutrition.org at CMU Libraries - library.cmich.edu on December 25, 2014
