Plasma Levels of Branched Chain Amino Acids, Incident Hip Fractures, and Bone Mineral Density of the Hip and Spine

Abstract

Objective

Branched chain amino acids (BCAA) are building blocks for protein, an essential component of bone. However, the association of plasma levels of BCAA with fractures in populations outside of Hong Kong or with hip fractures in particular is not known. The purpose of these analyses was to determine the relationship of BCAA including valine, leucine, and isoleucine and total BCAA (SD of the sum of Z-scores for each BCAA) with incident hip fractures and bone mineral density (BMD) of the hip and lumbar spine in older African American and Caucasian men and women in the Cardiovascular Health Study.

Design

Longitudinal analyses of association of plasma levels of BCAA with incident hip fractures and cross-sectional BMD of the hip and lumbar spine from the Cardiovascular Health Study.

Setting

Community.

Participants

A total of 1850 men (38% of cohort) and women; mean age 73 years.

Main Outcome Measures

Incident hip fractures and cross-sectional BMD of the total hip, femoral neck, and lumbar spine.

Results

In fully adjusted models, over 12 years of follow-up, we observed no significant association between incident hip fracture and plasma values of valine, leucine, isoleucine, or total BCAA per 1 SD higher of each BCAA. Plasma values of leucine but not valine, isoleucine, or total BCAA, were positively and significantly associated with BMD of the total hip (P = .03) and femoral neck (P = .02), but not the lumbar spine (P = .07).

Conclusions

Plasma levels of the BCAA leucine may be associated with higher BMD in older men and women. However, given the lack of significant association with hip fracture risk, further information is needed to determine whether BCAAs would be novel targets for osteoporosis therapies.

amino acids, fractures, bone mineral density

Hip fractures are a significant source of morbidity and mortality in elderly adults (1, 2). There is growing interest in the use of nonpharmacological interventions to reduce hip fracture incidence (3). Protein supplementation is 1 form of nonpharmacological therapy for osteoporosis that might be of potential value for the aging population because dietary protein insufficiency is common in older persons (4). However, studies of the effects of protein supplementation on skeletal health have been conflicting (5, 6, 7, 8, 9, 10, 11, 12). Importantly, these studies did not specifically examine the skeletal effects of individual amino acids, the monomers that compose protein (5, 6, 7, 8, 9, 10, 11, 12). Mechanistically, individual amino acids may differ in their actions on bone. Some, but not all, amino acids stimulate intracellular calcium production (13). Intracellular calcium phosphate in experimental studies in mice increases formation of bone apatite (14). Moreover, some amino acids appear to increase IGF-1, whereas others decrease it (15). Increases in intestinal calcium absorption (16), osteoblast, and type I collagen synthesis with reductions in osteoclast differentiation have also been reported with some amino acids (17). Consistent with these differing mechanisms of actions of amino acids on bone, epidemiological research suggests some amino acids may have favorable effects on bone mineral density (BMD), whereas others do not (18, 19, 20, 21, 22, 23).

Branched chain amino acids (BCAAs), which include valine, leucine, and isoleucine, account for upwards of 40% of the preformed amino acids. These are essential amino acids that must be acquired by dietary means (24). Relative to skeletal muscle tissue, BCAAs are anabolic (25), and their levels decline with age (26). Supplements containing BCAAs are promoted on the internet as “good for bone, muscle strength, and fatigue” and are widely used, particularly by bodybuilders (27). However, few published human studies exist on the associations or effects of BCAAs on skeletal health. Low dietary intakes of leucine have been associated with reduced muscle mass in elderly patients with hip fractures (28). In 1 study including older men and women, isoleucine and valine levels were positively associated with total BMD (21). In another study including Chinese community-dwelling older men and women, serum levels of valine, isoleucine, and leucine were significantly lower in those with osteoporosis compared with those without (29). In men and women aged 65 years and older, circulating levels of BCAAs were positively associated with hip BMD and higher valine intakes were associated with less decline in BMD over 4 years (29). However, in another short-term randomized controlled trial including sixteen women, administration of BCAA increased bone resorption in elderly women at bedrest (30).

To date, only 1 study has examined the association of levels of BCAA to fractures and reported no significant association of serum levels of valine, leucine, or isoleucine with major osteoporotic fracture risk. However, that study was confined to men and women in Hong Kong and did not specifically examine hip fractures (29). No studies have examined the association of BCAAs with incident hip fractures in healthy, older, community-dwelling, African American and Caucasian men and women. Therefore, to better understand the relationship of BCAAs to skeletal health in older adults in the United States, we used the Cardiovascular Health Study (CHS) to measure BCAA plasma levels and determine their prospective associations with incident hip fractures and cross-sectional associations with BMD of the total hip, femoral neck, and lumbar spine.

Methods

Study Participants

The CHS is a longitudinal cohort study that includes men and women aged 65 years and older from 4 US sites drawn from Health Care Finance Administration Medicare eligibility lists (31, 32). The original cohort of 5201 participants was enrolled in 1989 and 1990 and an additional 687 predominantly African American participants were enrolled in 1992 and 1993. All participants gave written informed consent to participate in the study following institutional review board approvals at each study site. From 1989 and 1990 to 1998 and1999, participants were seen in person during annual clinic visits and had telephone contact with information collected at intervening 6-month intervals. Following the 1998 and 1999 visit, participants continued to be contacted biennially to update hospitalizations, incident diagnoses, and medications. Information on incident hip fractures for the whole CHS cohort ended June 30, 2015.

We included all participants who had plasma levels of BCAAs measured. CHS measured plasma levels of BCAAs in 1622 CHS participants from the 1989 and 1990 cohort and 228 African American participants from the 1992 and 1993 cohort in a case-cohort study of cardiovascular disease (33). We assigned probability weights to individuals in this subsample to estimate results as close as possible to those expected from the full CHS population.

Measurements of BCAA

Blood samples were collected after a minimum 8-hour fast and EDTA plasma was immediately separated and frozen for long-term storage at −80 °C at the University of Vermont (34). BCAAs including valine, leucine, and isoleucine were measured using nuclear magnetic resonance spectroscopy, as previously described (35). Samples were tested in 2011, the nuclear magnetic resonance spectra were stored digitally, and the spectra were reanalyzed using the BCAA software algorithm. Frozen stability studies have demonstrated that BCAAs are stable when stored for up to 12 years at −80 °C. Additional studies in 2011 were performed to calculate biological variability and precision. The biological within-subject coefficients of variation were 7.7% for total BCAAs, 9.3% for valine, 9.4% for leucine, and 17.5% for isoleucine. The analytical coefficients of variation were 3.2% to 3.3% for total BCAAs, 2.9% to 3.1% for valine, 4.6% to 5.9% for leucine, and 11.9% to 14.1% for isoleucine (35). These data are comparable to the within-subject variability as reported by The European Federation of Clinical Chemistry and Laboratory Medicine Biological Variation Database, European Federation of Clinical Chemistry and Laboratory Medicine Working Group on Biological Variation, and the Task Group for the Biological Variation Database (36).

Hip Fractures

CHS collected self-report of hip fractures annually, self-report of hospitalizations every 6 months and, to ensure completeness of hospitalization records, checked Medicare claims data to identify any hospitalizations not reported by participants. All incident hip fractures were identified using International Classification of Disease, 9th Revision, code 820.xx review of hospital discharge data. Hip fractures from motor vehicle accidents or other severe injuries were excluded (E810.xx-825.xx). Pathological fractures (International Classification of Disease, 9th Revision, code 773.1x) were also not included in these analyses. Follow-up for hip fractures for these analyses began at CHS visits that occurred during 1989 and 1990 or 1992 and 1993 and was continued to a hip fracture event, death, loss to follow-up, or at 12 years following the 1989 and 1990 or 1992 and 1993 CHS visits for these analyses.

Dual Energy X-ray Absorptiometry Scans

Dual energy x-ray absorptiometry (DXA) measurements of areal BMD were performed at 2 CHS sites on Hologic QDR-2000 densitometers (Hologic, Inc., Waltham, MA) using the array beam mode during years 1995 and 1996. All images were interpreted at the University of California in San Francisco's reading center using Hologic Software, version 7.10. We included the total hip, femoral neck, and lumbar spine measurements in these analyses. The coefficient of variation for the total hip areal BMD was 0.75% (37).

Covariables

Covariable information for each participant was obtained from the same visit at which their plasma BCAA was measured (ie, 1989-1990 or 1992-1993). We a priori selected potential covariables that could affect the association between BCAA and hip fractures or BMD. We included age, body mass index (kg/m²), sex, race, clinic site, self-reported health status (excellent, very good, good vs fair, poor), history of diabetes, highest education level obtained (≥12th or <12th grade), smoking status (current, former, never), current alcohol use (0 to ≤7 drinks/week, >7 drinks/week), renal function from estimated glomerular filtration rate from the combined creatinine-cystatin C equation (25), and baseline fall history (number of falls in the previous year). Medications were determined via direct examination of medication bottles (38). We included medications used to treat or prevent osteoporosis selective estrogen receptor modulators, estrogens, and bisphosphonates, calcium and vitamin D supplements, and medications associated with fracture risk, including oral corticosteroids, loop diuretics, thiazide diuretics, selective serotonin reuptake inhibitors, anticonvulsants, benzodiazepines, sedative/hypnotics, proton pump inhibitors, thiazolidinediones, and thyroid medications.

Diet

We included dietary intakes of valine, leucine, and isoleucine from CHS, which were collected in 1989 and 1990 using a qualitative, picture-sort food frequency questionnaire (FFQ). In a substudy validation of the CHS FFQ including 47 female and 49 males aged 66 to 100 years from the CHS, estimates of mean nutrient intakes from the picture-sort FFQ used in CHS were comparable to estimates based on 24-hour recalls, and correlations with reference data were similar to those reported in the literature for conventionally administered FFQs (39, 40). We excluded those without FFQ data (N = 228) and those with extremes in energy intake (<500 kcal/day and >5000 kcal/) (N = 3) and determined the association between plasma BCAAs and dietary intakes of BCAA in the remaining 1615 CHS participants.

Statistical Analyses

We present descriptive tables of weighted baseline characteristics for the analytic population in 4 ways: compared with the full CHS population, and in the Supplement (41), Table 1, stratified according to median leucine value, presence or absence of subsequent hip fracture, censored without hip fractures, and categories of total hip BMD (normal, osteopenia, osteoporosis).

Table 1.

Distribution of BCAA values in the Cardiovascular Health Study

BCAA	N	Min	Q1	Mean	SD	Median	Q3	99%	Max
Valine (µM)	1846	114	201	229	41	227	254	338	422
Leucine (µM)	1846	15	88	113	35	111	134	215	390
Isoleucine (µM)	1846	7	44	55	16	53	64	100	196

Abbreviation: BCAA, branched chain amino acid.

Weighted multivariable Cox hazards models with robust SEs were used to estimate the hazard ratios and 95% CIs of incident hip fracture associated with an SD higher exposure of valine, leucine, or isoleucine. In addition to summing these to produce simple total BCAA (results in Supplement (41)), we accounted for the different scaling of individual amino acids by standardization to Z-scores with weighted mean zero and weighted SD of 1, then summed their Z-scores and calculated the SD of the sum. SDs for valine, leucine, isoleucine, and simple total BCAA in (μM) were, respectively, 38.93, 33.25, 15.42, and 76.35. We explored the linearity between the BCAA exposures and partial log hazard ratios with splines and found no meaningful departures from linearity.

We censored participants at death, loss of follow up, or after 12 years of follow-up. Cross sectionally, weighted linear regression models with robust SEs were used to estimate the association of BMD of the total hip, femoral neck, and lumbar spine measured in 1995 and 1996 with a 1-SD higher BCAA exposure (from plasma collected in 1989-1990 or 1992-1993). We used nested models adjusting for factors as follows: minimally adjusted: age, sex, race, clinic site; and fully adjusted: age, body mass index, sex, race, clinic site, self-reported health status (excellent, very good, good vs fair or poor), history of diabetes, highest education level obtained (≥12th vs <12th grade), smoking status (current, former, never), current alcohol use (0 to ≤7 drinks/week vs >7 drinks/week), renal function as measured by estimated glomerular filtration rate, baseline fall history, and medication use.

Because of heavy tails in the BCAA values, we conducted a sensitivity analysis in a limited range of exposure values, excluding individuals above the 99th percentile from each exposure, listed in Table 1. This resulted in exclusions of 19 individuals from the valine and total BCAA exposure and 18 individuals from the leucine and isoleucine measurements in the incident hip fracture analysis. In the BMD analyses, limited to 2 CHS sites, these exclusion counts were 5 and 4, respectively, for the BMD measures. We determined Pearson correlation coefficients between dietary intakes of valine, leucine, and isoleucine and plasma levels. All analyses were performed in R (42).

Results

We included 1850 participants with plasma levels of BCAAs. The median, mean, and SD, and minimum and maximum values for valine, leucine, and isoleucine are shown in Table 1.

Demographic and clinical characteristics of the study population and of all CHS participants are shown in Table 2. The mean age of the population used in these analyses was 73 years, 38% were men and 15% were Black. Fourteen percent of these participants had diabetes and the majority were never smokers. Similarly, in the entire CHS population, the mean age was 73 years, 42% were men, 16% Black, 16% had diabetes, and most were never smokers (Table 2). Characteristics according to median values of leucine, later hip fracture, censored without hip fracture, and bone density are available (41).

Table 2.

Weighted demographic and clinical characteristics of analytic cohort and entire CHS population

	Mean summary statistic (analytic cohort)	Mean summary statistic (entire CHS population)
Age (y), mean (SD)	73 (5.4)	73 (5.6)
BMI (kg/m2), mean (SD)	27 (4.8)	27 (4.7)
Female	62%	58%
Race
Black	15%	16%
White	85%	84%
Health status
Excellent/very good/good	80%	75%
Fair/poor	20%	25%
Diabetes (%)	14%	16%
Smoking history
Current	12%	12%
Former	39%	42%
Never	49%	46%
Clinic site
Bowman Gray, NC	26%	26%
Sacramento, CA	30%	26%
Hagerstown, MD	19%	22%
Pittsburgh, PA	25%	26%
Education
<12th grade	45%	44%
≥12th grade	55%	56%
Estimated glomerular filtration rate,a mean (SD)	80.33 (19.2)	77.62 (19.8)
Alcohol use >7 drinks per week	10%	10%
Medication use		1.78%
Medications used to treat osteoporosisb	2.77%	33%
Medication associated with fracture risk usec	29%
Calcium and vitamin D supplements	<1%	<1%
Thyroid medication used	8%	9%
Baseline fall historye
0	87%	84%
1-2	11%	13%
>2	2%	3%

Abbreviations: BMI, body mass index; CHS, Cardiovascular Health Survey.

Estimated glomerular filtration rate calculated from serum creatinine and cystatin C.

Defined as use of selective estrogen receptor modifiers, estrogens, or bisphosphonates.

Defined as use of oral corticosteroids, loop diuretics, thiazide diuretics, selective serotonin reuptake inhibitors, anticonvulsants, benzodiazepines, sedatives/hypnotics, proton pump inhibitors, thiazolidinediones.

Defined as use thyroid medications.

Self-report of number of falls in the previous year.

We documented 141 incident hip fractures, 42 in men and 99 in women, over 12 years of follow-up. The incidence rate of hip fractures per 100 person-years was 0.67 (95% CI, .40-1.12) for the whole population and 0.75 (95% CI, .52-1.07) in women and 0.53 (95% CI, .12-2.37 in men). In fully adjusted Cox proportional hazard models, we observed no significant associations between any BCAA or total BCAAs and incident hip fractures (Table 3 (41)).

Table 3.

Hazard ratios for hip fracture associated with 1-SD higher plasma level of BRAAs

	Minimally adjusted HR (95% CI)a	P	Fully adjustedb HR (95% CI)	P
Predictors
+Valine	.76 (.59-.99)	.04	.86 (.62-1.20)	.37
+Leucine	.78 (.61-1.01)	.06	.80 (.60-1.07)	.13
+Isoleucine	.80 (.61-1.04)	.10	.82 (.59-1.10)	.23
+Total BCAAs Zc	.73 (.55-.97)	.03	.77 (.54-1.11)	.17

Abbreviation: BRAAs, branched chain amino acids.

Adjusted for age, sex, race, and clinic site.

Adjusted for age, body mass index (kg/m²), sex, race, clinic site, self-reported health status (excellent, very good, good vs fair, poor), history of diabetes, smoking status (current, former, never), clinic site, highest education level completed (≥12th or <12th grade), renal function, current alcohol use (0-≤7 drinks/week, >7 drinks/week), fall history, and medication use.

Standardization to Z-scores with weighted mean 0 and weighted SD of 1, with summation of Z-scores and calculation of the SD of the sum.

+ SD: valine 38.93; leucine 33.25; isoleucine 15.42 (μM); total BCAA calculated as the sum of Z-scores for individual BCAA.

Among the 542 participants with DXA scans, each 1-SD higher level of leucine was associated with higher BMD of the total hip (P = .03) and femoral neck (P = .02) but not the lumbar spine (P = .07). There were no significant associations of any other BCAA with total hip, femoral neck, or lumbar spine BMD (Tables 4-6, Supplemental (41) Table 2).

Table 4.

Areal bone mineral density of the total hip by plasma level of branched chain amino acids

	BETA	SE	95% CI	P
+Valine
aMinimally adjusted	2.18	.68	(.84-3.52)	<.01
bFully adjusted	.48	.73	(−.94 to 1.91)	.51
+ Leucine
aMinimally adjusted	2.80	.65	(1.53-4.07)	<.01
bFully adjusted	1.47	.69	(.12-2.82)	.03
+Isoleucine
aMinimally adjusted	.90	.69	(−.46 to 2.26)	.19
b Fully adjusted	.23	.73	(−1.21 to 1.66)	.76
+Total BCAA Zc
aMinimally adjusted	2.37	.68	(1.04-3.7)	<.01
bFully adjusted	.95	.74	(−.5 to 2.41)	.20

Abbreviation: BRAAs, branched chain amino acids.

Adjusted for age, sex, race, and clinic site.

Adjusted for age, body mass index (kg/m²), sex, race, clinic site, self-reported health status (excellent, very good, good vs fair, poor), history of diabetes, smoking status (current, former, never), clinic site, highest education level completed (≥12th or <12th grade), renal function, current alcohol use (0-≤7 drinks/week, >7 drinks/week), and medication use.

Standardization to Z-scores with weighted mean 0 and weighted SD of 1, with summation of Z-scores and calculation of the SD of the sum.

+ SD: valine 38.93; leucine 33.25; isoleucine 15.42; total BCAAs calculated as the sum of Z-scores for individual BCAAs.

Table 5.

Areal bone mineral density of the femoral neck by plasma level of BCAAs

	BETA	SE	95% CI	P
+Valine
aMinimally adjusted	1.63	.61	(.43-2.83)	.01
bFully adjusted	−.19	.65	(−1.46 to 1.07)	.76
+ Leucine
aMinimally adjusted	2.49	.58	(1.35-3.62)	<.01
bFully adjusted	1.48	.61	(.28-2.68)	.02
+Isoleucine
aMinimally adjusted	.60	.62	(−.62 to 1.81)	.34
bFully adjusted	−.22	.65	(−1.50 to 1.06)	.74
+Total BCAA Zc
aMinimally adjusted	1.91	.61	(.72-3.1)	<.01
bFully adjusted	.5	.66	(−.8 to 1.8)	.45

Abbreviation: BRAAs, branched chain amino acids.

Adjusted for age, sex, race, and clinic site.

Adjusted for age, body mass index (kg/m²), sex, race, clinic site, self-reported health status (excellent, very good, good vs fair, poor), history of diabetes, smoking status (current, former, never), clinic site, highest education level completed (≥12th or < 12th grade), renal function, current alcohol use (0-≤7 drinks/week, > 7 drinks/week), and medication use.

Standardization to Z-scores with weighted mean 0 and weighted SD of 1, with summation of Z-scores and calculation of the standard deviation of the sum.

+ SD: valine 38.93; leucine 33.25; isoleucine 15.42; total BCAAs calculated as the sum of Z-scores for individual BCAAs.

Table 6.

Areal bone mineral density of the lumbar spine by plasma level of branched chain amino acids

	BETA	SE	95% CI	P
+Valine
aMinimally adjusted	3.24	1.08	(1.13-5.36)	.01
bFully adjusted	1.33	1.14	(−.91 to 3.57)	.24
+ Leucine
aMinimally adjusted	3.41	1.03	(1.39-5.43)	<.01
bFully adjusted	1.95	1.08	(−.17 to 4.07)	.07
+Isoleucine
aMinimally adjusted	.96	1.10	(−1.19 to 3.12)	.38
bFully adjusted	.30	1.15	(−1.95 to 2.56)	.79
+Total BCAA Zc
aMinimally adjusted	3.07	1.08	(.96-5.18)	.01
bFully adjusted	1.55	1.17	(−.74 to 3.84)	.19

Abbreviation: BRAAs, branched chain amino acids.

Adjusted for age, sex, race, and clinic site.

Adjusted for age, body mass index (kg/m²), sex, race, clinic site, self-reported health status (excellent, very good, good vs fair, poor), history of diabetes, smoking status (current, former, never), clinic site, highest education level completed (≥12th or <12th grade), renal function, current alcohol use (0-≤7 drinks/week, >7 drinks/week), and medication use.

Standardization to Z-scores with weighted mean 0 and weighted SD of 1, with summation of Z-scores and calculation of the SD of the sum.

+ SD: valine 38.93; leucine 33.25; isoleucine 15.42; total BCAAs calculated as the sum of Z-scores for individual BCAAs.

There were no significant interactions at the 0.10 level between valine, leucine, isoleucine, total BCAAs, and age, race, and sex for the outcomes of hip fractures or BMD (data not shown).

Sensitivity analyses excluding BCAA values above the 99th percent produced similar findings (data not shown).

We observed little or no correlation between dietary intakes of valine, leucine, or isoleucine or total BCAA and their respective plasma levels (R = .06, .03, .03, .05, respectively).

Discussion

In this cohort of older community-dwelling men and women, higher levels of the BCAA leucine were significantly associated with higher BMD of the total hip and femoral neck but not with incident hip fracture risk.

To our knowledge, only 1 prior study has examined the association of BCAA with incident fractures (29). In that study conducted in Hong Kong Chinese men and women, serum valine, leucine, and isoleucine were not significantly associated with incident major osteoporotic fractures during up to 10 years of follow-up. Our results extend these earlier findings to a non-Chinese population and to hip fractures in particular.

The association of amino acids with BMD has been examined in a few studies. Similar to our findings of a positive association of leucine with total hip BMD, positive associations of dietary leucine intake with BMD at the spine and forearm have been reported in discordant monozygotic female twins (23). In a 2-sample Mendelian randomization analysis, genetically increased leucine was positively associated with total body BMD; isoleucine and valine levels were also positively associated with BMD (21). In older men and women of Chinese descent, higher serum levels of leucine were associated with a trend for a less decline in BMD of the hip; valine and isoleucine levels also were associated with less BMD decline (29). In a post hoc metabolomics study of older women who had received supplementation with Lactobacillus reuteri or placebo, plasma levels of the BCAA valine were higher in those who received the active supplement and were positively correlated with tibial volumetric BMD (43). The reasons why we did not see a statistically significant association of valine and isoleucine with BMD are not certain but may be related to differences in sample sizes, participant characteristics or covariables considered. In the PROVIDE study, sarcopenic older adults consuming a supplement enriched in vitamin D, calcium, and leucine showed a small positive benefit on total body BMD (44). In contrast with our results, others have reported a potential negative effect of BCAA on bone (30). In a short-term randomized controlled trial including 16 elderly women at bedrest, administration of BCAA increased bone resorption (30). Both studies were of short duration and enrolled few participants (30, 44).

The mechanisms by which BCAA might be favorable for bone were not explored in CHS and remain unclear and deserve further study. In human bone marrow stem cells isoleucine, valine, and leucine (along with proline-proline) at 50 µM increased bone marrow stem cell proliferation, though effects seen with valine and leucine were less robust (45). In vitro studies have reported that all 3 BCAAs, particularly valine, are necessary for osteoclast maturation; however, these associations are concentration dependent because some concentrations of BCAA promote osteoclast differentiation, and others inhibit it (46).

Strengths of these analyses include inclusion of a well-characterized cohort of older US men and women. We had both prospective information on incident hip fractures and a cross-sectional assessment of BMD of the hip and spine. There are also some limitations to this work. The BCAAs were only measured at a single visit. This is a limitation because there may be significant variability in plasma BCAA levels depending on the type of food consumed (47), the speed at which the meal is consumed (48), body weight (49), and gastrointestinal absorptive capacities (50). However, the Cox regression proportional hazard assumption is that the relationship between BCAA levels and our outcome of hip fractures would not change over time, and we observed no violation of this assumption. Furthermore, there is controversy as to the contribution of diet to plasma BCAA concentrations with some studies reporting that higher dietary intakes of BCAA are related to higher plasma BCAA levels (51, 52) and others not (53, 54). In CHS, we found very little correlation between diet and plasma BCAA levels. The biological variability for isoleucine was higher than the other BCAAs, but isoleucine has a low circulating concentration, and all analytical coefficients of variation were in expected ranges. We only had 1 measurement of DXA and only approximately 1/3 of the persons included in the hip fracture analyses had a DXA done. Furthermore, we could not model the association of BCAAs and hip fractures adjusted for BMD because we had too few hip fracture events in the BMD subsample. There were too few hip fracture events over the 12 years of follow-up to reliably assess sex-stratified models, although we did examine the interaction of sex with our outcomes, and it was not significant. We did not have measurements of muscle mass, and it is possible that the benefits of leucine we saw on BMD were at least in part secondary to changes in the muscle. Although we only included those CHS participants who had levels of BCAA measured, we were able to backweight these to be representative of the entire CHS population.

Our results suggest that further attention is warranted to the impact of specific amino acids on skeletal health, rather than just considering protein as a whole. However, given the lack of a significant association with hip fractures, more research is needed as to whether BCAAs at different concentrations would be helpful for skeletal health.

Ethical Approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the Cardiovascular Health Study.

Acknowledgments

This research was supported by contracts HHSN268201200036C, HHSN268200800007C, HHSN268201800001C, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086, and 75N92021D00006, and grants U01HL080295 and U01HL130114 from the National Heart, Lung, and Blood Institute (NHLBI), with additional contribution from the National Institute of Neurological Disorders and Stroke. Additional support was provided by R01AG023629 and K24AG065525 from the National Institute on Aging. A full list of principal Cardiovascular Health Study (CHS) investigators and institutions can be found at CHS-NHLBI.org.

Disclosures

M.A.C. is an employee of Labcorp and owns Labcorp stock. None of the other authors have any competing interests or conflicts of interests with this work. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Data Availability

Original data generated and analyzed during this study are included in this published article or in the data repositories in Cardiovascular Health Study.

Abbreviations

 
• BCAA

  branched chain amino acid

 
• BMD

  bone mineral density

 
• CHS

  Cardiovascular Health Study

 
• DXA

  dual energy x-ray absorptiometry

 
• FFQ

  food frequency questionnaire