Self-reported Resistance Training Is Associated With Better HR-pQCT–derived Bone Microarchitecture in Vegan People

Abstract

Context

A plant-based lifestyle is a global trend; lower bone mineral density and increased fracture risk in vegan people are reported.

Objective

The primary objective was to assess trabecular and cortical bone microarchitecture in vegans and omnivores. Secondary objectives were to evaluate relationships between bone microarchitecture, nutrition parameters, and physical activity.

Methods

This was an observational study at the Medical Department II, St. Vincent Hospital (tertiary referral center for gastrointestinal, metabolic, and bone diseases, and teaching hospital of the Medical University of Vienna), including 43 healthy nonobese female and male subjects on a plant-based diet for at least 5 years, and 45 healthy nonobese female and male subjects on an omnivore diet for at least 5 years. The main outcome measures were the parameters of trabecular and cortical bone microarchitecture (high-resolution peripheral quantitative computed tomography), serum markers of bone turnover, nutrient intake (nutrition protocol), and self-reported resistance training (physical activity questionnaires).

Results

In the vegan group, trabecular and cortical structure were altered compared with omnivores. Vegans not reporting resistance training had diminished bone microarchitecture compared with omnivores not reporting resistance training. In vegans and omnivores reporting resistance training, bone structure was similar. In both vegan subgroups (resistance training and not resistance training), a small number of correlations between nutrient intake and bone microarchitecture were observed without a conclusive pattern.

Conclusion

Bone microarchitecture in vegans differed from matched omnivores but could not be explained solely by nutrient uptake. These differences were attenuated between the subgroups reporting resistance training. In addition to a well-planned diet, progressive resistance training on a regular basis should be part of the vegan lifestyle.

veganism, trabecular microarchitecture, cortical bone, resistance training

A vegan lifestyle is a global trend. In the United States, the percentage of vegans increased from 1% in 2014 to 6% in 2017 (1), and in Europe the number of vegan food products nearly doubled from 2016 to 2019 (2).

Recent data indicate a relationship between a vegan diet and higher susceptibility to fractures as well as low bone mineral density (BMD), as measured by dual-energy X-ray absorptiometry and quantitative ultrasound (3-5). These findings are supported by several meta-analyses. For instance, Li and colleagues (6) reported lower BMD in vegans, and Iguacel et al observed both lower BMD and increased fracture risk (7).

Osteoporosis is defined as low bone strength and increased fragility fracture risk not only due to low BMD but also due to impaired bone microarchitecture (8).

Even though dual-energy X-ray absorptiometry is widely used to assess osteoporosis and an individual fracture risk, this method measures only areal BMD but not 3-dimensional bone microarchitecture. However, microarchitecture is an important component of bone strength (9).

High-resolution peripheral quantitative computed tomography imaging (HR-pQCT) is a noninvasive in vivo assessment of trabecular and cortical bone microarchitecture at the distal radius and tibia. Parameters measured by HR-pQCT strongly correlate with both structural and mechanical properties of the lumbar spine and hip (10).

HR-pQCT will likely provide additional information on potential differences in trabecular and cortical bone structure between vegans and omnivores. To date, there has been no investigations on bone microarchitecture in vegans.

It is well known that resistance training stimulates bone formation (11, 12), whereas other common sports activities, like cycling or swimming, do not exert a relevant effect on BMD (13). Of the studies on vegans cited above, none took the influence of different forms of physical activity on BMD into account.

The aim of this study was to assess bone microarchitecture in vegans and matched omnivores. Furthermore, this study should assess whether self-reported resistance training is associated with differences in trabecular and cortical bone structure.

The primary objective was to evaluate differences in bone structure by state-of-the-art HR-pQCT at the radius and tibia in vegans and omnivores. Secondary objectives were serum levels of bone turnover markers, the analysis of standardized nutrition protocols, as well as physical activity questionnaires and their association with these parameters.

Materials and Methods

This study was conducted at the Medical Department II of the St. Vincent Hospital, a teaching hospital of the Medical University of Vienna and tertiary referral center for gastrointestinal, metabolic, and bone diseases in Vienna, Austria. This study was approved by the St. Vincent Hospital Ethics Committee (EK no. 003-02-2015), and prior to any procedure the subject’s informed consent was obtained. This study was conducted in accordance with the Declaration of Helsinki.

Participants

Healthy nonobese female and male subjects on a long-term plant-based diet were recruited with the friendly support by the Austrian Vegan Society. The inclusion criteria were a vegan diet for at least 5 years, body mass index (BMI) >18.5 and <30 (nonunderweight, nonobese), age >30 and <50 years, and premenopausal status.

Omnivores were recruited via information sheets, email invitations, and social media from hospital staff, rescue organizations, and the Ministry of Health and the Ministry of the Interior. The participants in the omnivore group were subject to the same inclusion criteria, except for consuming an omnivore diet for at least 5 years. A full medical history was obtained from all participants.

The exclusion criteria in both groups were current pregnancy or breastfeeding; oncological or inflammatory diseases (eg, rheumatoid arthritis, spondyloarthritis, systemic lupus erythematosus); illicit drug use, alcohol abuse (>3 International Units per day), or infectious diseases; gastrointestinal diseases (Crohn’s disease, ulcerative colitis, chronic kidney disease, chronic liver disease, diabetes mellitus type I or II, or cardiovascular disease); hypo- or hyperthyroid metabolic status (substitution was not an exclusion criterion); treatment with adrenal or anabolic steroids, glitazones, anticonvulsants, anticoagulants, long-term (>5 years) regular use of proton pump inhibitors, or any osteoporosis-specific treatment (except for calcium and/or vitamin D supplementation).

Bone Microarchitecture

HR-pQCT (XtremeCT, SCANCO Medical, Brütisellen, Switzerland) examinations were performed after calibration with a standardized phantom (Moehrendorf, Germany) by 2 International Osteoporosis Foundation–certified and well-experienced physicians, according to the manufacturer’s standard in vivo acquisition protocol. The nondominant (except for previous fracture) ultradistal radius and ultradistal tibia were assessed while immobilized in a carbon fiber cast in order to minimize motion artefacts.

The reference line was set manually at the radial and tibial endplate, so that the measurement regions were located 9.5 mm proximal from the reference line at the radius and 22.5 mm at the tibia. The regions consisted of a length of bone of approximately 9 mm, assessed by 110 computed tomography slices. The Scanco standard operation procedure (SOP) scale was used to exclude scans with motion artefacts graded higher than 3. Detailed methodologic information has been reported previously (14). Segmentation of cortical and trabecular bone was performed automatically, according to previous work (15). According to the literature, the precision error is <5% for volumetric BMD as well as trabecular and cortical structure parameters; for cortical porosity, the precision error is <10% (16).

The following parameters of volumetric BMD, and bone microstructure and geometry were assessed: total BMD (Tt.BMD), cortical BMD (Ct.BMD), trabecular BMD (Tb.BMD), trabecular bone volume fraction (Tb.BV/TV), trabecular number (Tb.N/mm), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), cortical thickness (Ct.Th), and cortical porosity (Ct.Po). Due to organizational reasons, the serum markers and the HR-pQCT scans were obtained in 2 different visits.

Serum Markers of Bone Turnover

Serum turnover markers were obtained after overnight fasting (sampling between 8 and 10 am) at the ISO 9001-certified laboratory (Labcon Ltd) at the St. Vincent Hospital (IDS-iSYS microparticle immunoassay system, Immunodiagnostics Systems Ltd, Boldon, UK, Architect ci8200 platform, Abbott Laboratories, Abbott Park, US-IL and ACL TOP 500 CTS, Instrumentation Laboratory, USA).

The following parameters were determined: calcium, 25-hydroxy vitamin D, phosphorus, intact parathyroid hormone (PTH), alkaline phosphatase, and cross-linked C-telopeptide (CTX).

CTX was measured in duplicates using an established manual enzyme-linked immunosorbent assay (Biomedica, Vienna, Austria) with a known intra-assay coefficient of variation of 2.1% to 4.9% (17).

Nutrition

A standardized estimated food diary based on the protocols used for the Austrian Nutrition Study (18) over 4 days (either Wednesday-Saturday or Sunday-Wednesday in order to contain both business days and weekends) was obtained twice during 2 different seasons. Therefore, the protocol represents a period of 8 days and different times of the year. According to previous work (19, 20), an overall period of 7 days is sufficient in order to obtain reliable data while a reporting time of 4 days each reduces the burden to participants and thus increases completion rates ( (21)). The nutrition questionnaires were at least 3 months apart from each other, to assess more than 1 season. As the study was performed over the period of 1 year, all seasons were covered.

The following nutrients relevant for bone health according to the International Osteoporosis Foundation (22) were assessed: protein, vitamin A, vitamin B12, vitamin B6, vitamin D, calcium, magnesium, and zinc. Vitamin A is provided as retinol equivalents, according to previous work (23). In addition, the intakes of energy, phosphate, vitamin B9 (folic acid), and vitamin K were analyzed. Vitamin K was provided as phylloquinone (vitamin K1) and not divided into K1 and K2 due to missing data for vitamin K2 in the food composition database used.

The intake of these nutrients was calculated in the following way: all foods and beverages along with the estimated weights recorded by the participants were entered into the diet analysis program nut.s (dato Denkwerkzeuge, Vienna) using the German food composition database (Bundeslebensmittelschlüssel, Max-Rubner Institute, Germany) and its extension for typical Austrian foods (Österreichische Nährwerttabelle, dato Denkwerkzeuge/Department of Nutritional Sciences, Vienna).

Since supplemental intake for some nutrients (eg, vitamin B12, vitamin D, protein) is quite common in vegans, all participants were asked to report their use of supplements, including brand names and dosage. Supplemental intake of nutrients was calculated, based on the nutrient content provided by the manufacturers. In the case of combination supplements, all components of the product were counted separately toward each nutrient class. Thus, we can provide the overall nutrient intake, including supplements, and nutrient supply only by food (ie, excluding supplements).

Furthermore, several authors estimated dietary acid load (ie, the balance of acid- and base-inducing food) as relevant for bone health and assumed an inverse relationship with BMD (24, 25). Therefore, we calculated net endogenous acid production (NEAP) by using 2 different formulas, according to previous work (24-26): NEAP 1 (mEq/day) = (54.5 × protein g/day/potassium mEq/day) – 10.2; NEAP 2 (mEq/day) = (0.91 × protein g/day) – (0.57 potassium mEq/day) + 21.

Sports Activity

Typical physical activity within the last year was assessed by standardized questionnaires. Specifically, which kind of sports activities had been carried out on a regular basis. Based on their answers, subjects were divided into 2 groups: People performing resistance training on a regular basis, in other words, at least 1 training session per week, were allocated to group 1. People not performing resistance training on a regular basis were allocated to group 2. Previous studies reported that 1 session of resistance training per week is sufficient to exert a positive effect on bone mineral density (27, 28).

Statistical Analysis

Sample size calculations are based on standard deviations from a pilot evaluation of tibial Tb.Th in 8 nonresistance training vegans and 7 nonresistance training omnivores. In this pilot evaluation a high effect size of about 0.5 could be shown. In order to detect a mean difference in tibial Tb.Th of alpha 0.05 with a power of 90%, and in consideration of a dropout rate of about 15% to 20%, 25 subjects had to be included per group. Finally, 23 vegans and 20 omnivores were left for analysis, showing a high effect size of 0.56 with a power of 95% (P < .001) in tibial Tb.Th difference.

Eighty-eight subjects were included in this analysis.

Depending on the distribution of variables, group size and comparison of groups continuous data are either expressed as mean ± SD or mean rank and sum ranks. Categorical variables are expressed in absolute numbers. Normality tests have been measured using the Kolmogorov–Smirnov test and homogeneity of variance was proofed using Levene’s test of homogeneity. Results were categorized as statistically significant with an alpha level set at < .05 and < .01, the reported P values are 2-sided.

First, concerning possible confounders, t-tests were applied to test whether age, BMI, nicotine (packyears), and alcohol (units per day) significantly differ between groups of vegans and omnivores with and without resistance training.

Moreover, multivariate analyses of variance (2-tailed, P < .05) were used for comparing multivariate sample means, to identify significant differences between omnivores and vegans with and without resistance training in BMI-adjusted bone microarchitecture, nutrients including and excluding supplements, serum markers, and acid load.

Th nonparametric Mann–Whitney U (2-tailed, P < .05) test was used for comparing exclusively the aerobic activities group and the no sports group of vegans and omnivores, because of different group size and small groups.

Correlations (Pearson correlation, 2-tailed, P < .05) were used to examine relations between BMI-adjusted bone microarchitecture and nutrient intake. Furthermore, Pearson correlations were calculated to show relations between the duration of vegan diet and bone microarchitecture.

All analyses were performed using SPSS, version 25.0 (IBM Corp., Armonk, NY, USA).

Results

Participants

In total, 105 subjects were screened (Fig. 1). Seventeen subjects had to be excluded due to either menopause, illicit drug consumption, or incompletely filled out questionnaires. Therefore, 88 subjects were included. Mean (SD) age of the study population was 39.2 years (±6.5) with a male:female ratio of 51:49% and a vegan:omnivore ratio of 49:51%.

Figure 1.

Study flow chart.

The vegan group consisted of 22 women and 21 men, the omnivore group of 23 females and 22 males.

Twenty vegans (9 women and 11 men) and 25 omnivores (8 women and 17 men) reported for progressive resistance training on a regular basis: using machines, free weights, or bodyweight resistance exercises at least once a week.

Differences Between the Vegan and Omnivore Groups

Demographics were similar, except for a slightly but significantly lower BMI in the vegan group (Table 1).

Table 1.

Differences in demographics, BMI-adjusted bone microarchitecture, serum markers, and daily nutrient uptake between the vegan and omnivore groups

	Vegans (n = 43)		Omnivores (n = 45)
	Mean	SD	Mean	SD	P value
Demographic variables
Nicotine (packyears)	2.870	6.350	1.000	3.060	.740
Alcohol (units per day)	0.580	1.210	0.640	1.140	.819
Height (cm)	174.25	9.55	173.7	8.48	.763
BMI	22.440	2.650	23.950	2.800	.021
Age (years)	38.060	6.160	40.360	6.610	.072
BMI-adjusted bone microarchitecture
Radius Tt.BMD	324.716	10.005	342.396	9.780	.210
Radius Ct.BMD	884.293	9.019	888.793	8.816	.081
Radius Tb.BMD	166.086	5.681	180.107	5.553	.722
Radius Tb.BV/TV	0.137	0.031	0.151	0.031	.028
Radius Tb.Th	0.072	0.013	0.075	0.014	.188
Radius Ct.Th	0.770	0.208	0.860	0.233	.045
Radius Ct.Po	1.802	1.009	1.756	0.825	.804
Tibia Tt.BMD	282.719	7.617	311.231	7.445	.009
Tibia Ct.BMD	872.998	6.842	884.016	6.6899	.253
Tibia Tb.BMD	167.544	5.680	190.353	5.553	.005
Tibia Tb.BV/TV	0.139	0.032	0.159	0.031	.002
Tibia Tb.Th	0.069	0.011	0.079	0.011	.001
Tibia Ct.Th	1.088	0.209	1.209	0.288	.019
Tibia Ct.Po	5.858	2.393	5.513	2.051	.445
Serum markers
Calcium (mmol/L)	2.329	0.078	2.387	0.109	.009
Phosphate (mmol/L)	1.121	0.204	1.107	0.199	.772
Parathyroid hormone (pg/mL)	57.823	24.324	48.080	18.342	.060
Alkaline phosphatase (U/L)	63.857	17.027	55.879	15.445	.039
25OH vitamin D (ng/mL)	31.733	12.912	28.042	9.185	.169
Crosslaps (CTX, ng/mL)	0.452	0.248	0.382	0.173	.170
Creatinine (mg/dL)	0.777	0.125	0.884	0.141	.001
Daily nutrient uptake (including supplements)
Energy (in kcal) per day (via food)	2248.590	549.178	1964.205	526.150	.011
Protein per kg bodyweight per day	1.024	0.297	1.178	0.409	.037
Vitamin A (retinol equivalent, mg per day)	7.854	11.723	5.151	27.081	.527
Vitamin B12 (cobalamin, µg per day)	120.135	259.133	7.191	15.705	.003
Vitamin B6 (pyridoxine, mg per day)	3.975	8.821	3.903	14.679	.977
Vitamin B9 (folic acid, µg per day)	605.365	412.873	322.408	185.410	<.001
Vitamin D (calciferol, µg per day)	28.867	39.439	6.828	13.076	<.001
Vitamin K (µg per day)	349.631	427.178	105.726	61.556	<.001
Calcium (mg per day)	784.954	286.094	936.531	419.341	.041
Magnesium (mg per day)	626.243	214.895	403.802	218.451	<.001
Phosphate (mg per day)	1267.763	398.453	1361.449	487.328	.305
Zinc (mg per day)	11.603	5.646	12.579	5.912	.410
Acid load (NEAP 1)	23.363	14.536	51.351	15.769	<.001
Acid load (NEAP 2)	17.406	31.138	52.977	28.920	<.001

Abbreviations: BMI, body mass index; Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; NEAP, net endogenous acid production; Tb.BMD, trabecular bone mineral density; Tb.BV/TV, trabecular bone volume fraction; Tb.Th, trabecular thickness; Tt.BMD, total bone mineral density.

BMI-adjusted bone structure was altered in the vegan group at both the radius and tibia: BV/TV and Ct.Th were significantly lower at both the radius and tibia. In addition, at the tibia, significantly lower Tt.BMD, Ct.BMD, Tb.BMD, and Tb.Th were observed.

Intakes of energy, cobalamin, folic acid, vitamin D, vitamin K, and magnesium were significantly higher in the vegan group, whereas intake in protein per kilogram bodyweight and calcium were significantly lower. Dietary acid load was significantly lower in vegans.

Serum markers of bone turnover were within normal values in both groups, with slightly but significantly higher calcium levels in the omnivore group. Nutrient intake excluding supplements is provided elsewhere (Table 1 (29)).

Detailed Subgroup Differences

Demographics

Demographics were similar between the different subgroups, except for significantly higher nicotine consumption in the nonresistance training vegan subgroup and a significantly higher age in the nonresistance training omnivore subgroup (Table 2). There was a trend toward lower body mass index in nonresistance training vegans than in nonresistance training omnivores, but the differences did not reach statistical significance. The duration of the vegan diet did not differ between the subgroups.

Table 2.

Detailed subgroup differences for demographics, BMI-adjusted bone microarchitecture, serum markers and daily nutrient uptake between the vegan and omnivore group, as well as persons with and without self-reported resistance training, respectively

	Vegans (n = 43)					Omnivores (n = 45)					Nonresistance training vegans vs omnivores	Resistance training vegans vs omnivores
	Resistance training (n = 20)		Nonresistance training (n = 23)			Resistance training (n = 25)		Nonresistance training (n = 20)
	Mean	SD	Mean	SD	P value	Mean	SD	Mean	SD	P value	P value	P value
Demographic variables
Nicotine (packyears)	0.050	0.224	5.330	7.976	.004	0.280	1.208	1.810	4.182	.177	.066	.221
Alcohol (units per day)	0.332	0.650	0.814	1.530	.177	0.490	7.390	0.805	1.487	.425	.963	.330
Height (cm)	174.37	10.13	174.69	9.56	.916	176.22	8.11	170.2	8.45	.019	.112	.501
BMI	22.517	2.535	22.454	2.890	.941	23.976	3.129	23.790	2.436	.534	.051	.187
Age (years)	36.300	6.506	39.250	5.605	.071	37.920	6.601	43.273	5.725	.031	.111	.327
Duration of vegan diet (years)	10.900	5.664	9.456	4.200	.344
BMI-adjusted bone microarchitecture
Radius Tt.BMD	335.220	85.877	315.582	41.557	.336	302.71	47.25	265.332	33.935	.823	.159	.685
Radius Ct.BMD	875.065	57.528	892.317	44.674	.275	876.13	44.18	869.39 4	45.62 3	.914	.908	.429
Radius Tb.BMD	179.090	43.340	154.778	30.123	.037	183.48	39.43	153.691	33.201	.413	.048	.678
Radius Tb.BV/TV	0.149	0.036	0.128	0.025	.037	0.157	0.031	0.142	0.027	.411	.049	.686
Radius Tb.Th	0.076	0.013	0.069	0.012	.075	0.078	0.015	0.072	0.011	.573	.249	.977
Radius Ct.Th	0.760	0.277	0.791	0.137	.626	0.922	0.235	0.795	0.202	.213	.689	.068
Radius Ct.Po	1.940	1.281	1.613	0.744	.301	1.892	0.750	1.503	0.838	.144	.688	.805
Tibia Tt.BMD	302.705	47.246	265.339	33.930	.004	323.212	53.332	296.255	54.054	.101	.028	.185
Tibia Ct.BMD	869.39	45.616	876.134	44.179	.625	876.928	41.955	892.875	48.723	.245	.244	.568
Tibia Tb.BMD	183.48	39.427	153.687	33.202	.010	199.076	36.657	179.45	32.042	.066	.014	.177
Tibia Tb.BV/TV	0.153	0.033	0.128	0.027	.011	0.168	0.030	0.146	0.026	.038	.030	.175
Tibia Tb.Th	0.074	0.011	0.065	0.010	.006	0.081	0.010	0.076	0.011	.210	.001	.049
Tibia Ct.Th	1.153	0.223	1.029	0.157	.046	1.286	0.302	1.117	0.242	.100	.123	.139
Tibia Ct.Po	6.231	2.585	5.588	2.385	.478	6.011	2.013	4.772	1.936	.029	.122	.692
Serum markers
Calcium (mmol/L)	2.350	0.080	2.313	0.073	.189	2.392	0.115	2.375	0.113	.729	.023	.390
Phosphate (mmol/L)	1.141	0.248	1.081	0.149	.314	1.127	0.228	1.091	0.172	.505	.961	.831
Parathyroid hormone (pg/mL)	51.108	16.122	62.401	30.713	.169	42.003	10.690	57.031	24.309	.005	.847	.040
Alkaline phosphatase (U/L)	62.176	15.824	63.250	19.303	.712	55.833	16.078	53.154	13.801	.898	.125	.164
25OH vitamin D (ng/mL)	36.319	13.940	28.843	11.869	.098	26.292	10.735	29.992	7.137	.512	.927	.030
Crosslaps (CTX, ng/mL)	0.531	0.326	0.364	0.160	.048	0.417	0.211	0.339	0.112	.786	.456	.115
Daily nutrients uptake (including supplements)
Energy (in kcal) per day (via food)	2337.181	592.641	2156.410	527.141	.296	2130.425	534.550	1742.768	471.226	.015	.010	.226
Protein per kg bodyweight per day	1.050	0.292	1.001	0.326	.618	1.302	0.469	1.020	0.300	.019	.857	.034
Vitamin A (retinol equivalent, mg per day)	6.626	12.657	7.868	10.903	.731	8.930	37.483	1.015	0.384	.351	.008	.794
Vitamin B12 (cobalamin, µg per day)	156.713	278.562	110.097	269.238	.580	9.984	21.525	3.827	1.073	.209	.085	.012
Vitamin B6 (pyridoxine, mg per day)	3.340	2.415	2.363	1.143	.091	6.172	20.262	1.356	0.379	.295	.001	.538
Vitamin B9 (folic acid, µg per day)	621.598	403.144	506.540	174.995	.221	370.550	229.151	255.894	97.465	.043	< .001	.012
Vitamin D (calciferol, µg per day)	30.892	36.418	31.472	45.038	.964	9.518	17.165	3.747	5.406	.156	.009	.013
Vitamin K (µg per day)	419.397	623.064	290.608	145.896	.341	118.393	73.170	92.268	44.643	.169	< .001	.021
Calcium (mg per day)	861.131	340.560	735.482	225.208	.156	1005.414	351.862	876.077	508.711	.320	.237	.173
Magnesium (mg per day)	659.883	253.492	571.403	149.643	.170	481.243	261.828	313.348	117.537	.011	< .001	.015
Phosphate (mg per day)	1348.271	498.105	1226.263	303.211	.331	1568.035	534.826	1106.444	297.402	.001	.200	.165
Zinc (mg per day)	11.371	4.747	10.295	2.858	.668	14.634	7.267	10.125	2.454	.011	.410	.201
Acid load (NEAP 1)	27.611	17.728	19.669	10.064	.074	53.115	18.756	49.146	11.057	.408	< .001	< .001
Acid load (NEAP 2)	11.279	41.254	22.733	17.815	.233	57.182	37.168	47.721	11.814	.281	< .001	< .001

Abbreviations: BMI, body mass index; Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; NEAP, net endogenous acid production; Tb.BMD, trabecular bone mineral density; Tb.BV/TV, trabecular bone volume fraction; Tb.Th, trabecular thickness; Tt.BMD, total bone mineral density.

BMI-adjusted Trabecular and Cortical Bone Microarchitecture

Nonresistance training vegans had significantly diminished bone microarchitecture, compared with nonresistance training omnivores(Fig. 2).

Figure 2.

Selected parameters of trabecular microarchitecture in vegans and omnivores, and in resistance training and nonresistance training persons, respectively. (A) Radial trabecular bone volume fraction. (B) Radial trabecular thickness. (C) Tibial trabecular bone volume fraction. (D) Tibial trabecular thickness. For most parameters (except for radial trabecular thickness), nonresistance training vegans had diminished microarchitecture, compared to resistance training vegans and nonresistance training omnivores. In contrast, resistance training vegans and omnivores had similar bone structure (except for tibial trabecular thickness).

In contrast, in resistance training subjects, there were hardly any differences in bone structure between vegans and omnivores.

Even though bone structure differed between resistance training people and nonresistance training people in both vegans and omnivores, the differences were stronger between the vegan subgroups.

The duration of the vegan diet had no influence on bone microarchitecture (Table 3 (29)).

Vegans performing exclusively aerobic activities (n = 16) and vegans performing no sports activities at all (n = 6) had similar bone microarchitecture. Omnivores performing only aerobic activities (n = 10) and those performing no sports activities at all (n = 10) had similar bone structure as well (Table 4 (29)).

Serum Markers

In all subgroups, serum markers were within normal ranges. In people performing resistance training, vegans had significantly higher PTH and vitamin D levels. In subjects not performing resistance training, vegans had significantly lower calcium levels. Within the vegan group, people performing resistance training had significantly higher CTX levels. Omnivores performing resistance training had significantly lower PTH levels than those who did not.

Nutrient Intake

In nonresistance training people, vegans had significantly higher energy intake, a significantly lower percentage of energy intake from protein, and significantly higher intakes of vitamin A, B6, B9, calciferol, vitamin K, and magnesium.

In resistance training subjects, vegans had a significantly lower protein intake per kilogram bodyweight, significantly lower percentage of energy intake by protein and significantly higher intakes of vitamin B12, B6, calciferol, vitamin K, and magnesium.

Between the 2 vegan subgroups, there were no significant differences in nutrients intake.

Between the 2 omnivore subgroups, those performing resistance training had significantly higher intakes of energy, protein per kg bodyweight, vitamin B9, magnesium, phosphate, and zinc.

Thirty-two vegans and 12 omnivores consumed supplements. Nutrients intake excluding supplements is provided elsewhere (Table 2 (29)).

Correlation Analysis

When testing for relationships between nutrients intake and BMI-adjusted bone microarchitecture in vegans performing resistance training, few correlations were observed: Vitamin D was strongly related to tibia (T)_Tt.BMD, T_Tb.BMD, and T_BV/TV, vitamin A strongly to T_Ct.Th, energy moderately to T_Ct.BMD and strongly T_Ct.Po, and phosphate moderately to T_Ct.Po (Table 3).

Table 3.

Pearson Correlation analysis (correlation coefficients) of daily nutrient uptake and body mass index–adjusted parameters of trabecular and cortical bone microarchitecture. Significant correlations are marked with an asterisk.

Vegans with self-reported resistance training
	Energy (in kcal) per day	Protein per kg bodyweight per day	Vitamin A (retinol equivalent, mg per day)	Vitamin B12 (cobalamin, µg per day)	Vitamin B6 (pyridoxine, mg per day)	Vitamin B9 (folic acid, µg per day)	Vitamin D (calciferol, µg per day)	Vitamin K (µg per day)	Calcium (mg per day)	Magnesium (mg per day)	Phosphate (mg per day)	Zinc (mg per day)
R_Tt.BMD	0.023	0.194	–0.051	0.18	–0.033	0.082	–0.199	0.066	0.043	–0.005	0.007	–0.246
R_Ct.BMD	–0.124	0.142	–0.064	0.175	–0.092	–0.107	0.076	–0.091	–0.134	–0.103	–0.114	–0.067
R_Tb.BMD	0.337	0.255	0.02	0.191	0.159	0.38	–0.416	0.257	0.338	0.248	0.252	–0.222
R_Tb.BV_TV	0.339	0.255	0.017	0.19	0.158	0.379	–0.416	0.256	0.339	0.248	0.252	–0.223
R_Tb_Th	0.169	0.274	–0.121	0.288	0.143	0.308	–0.388	0.195	0.241	0.181	0.205	–0.172
R_Ct_Th	0.219	0.209	–0.081	0.116	0.109	0.144	0.032	0.111	0.046	0.178	0.155	–0.059
R_Ct_po	0.393	0.181	–0.026	–0.201	0.01	0.183	–0.219	0.081	0.251	0.187	0.325	–0.012
T_Tt.BMD	–0.068	0.051	0.329	0.055	–0.12	0.109	–0.511*	0.143	0.14	–0.119	–0.035	–0.347
T_Ct.BMD	–0.462*	–0.148	0.153	0.035	–0.263	–0.252	–0.081	–0.094	–0.267	–0.397	–0.425	–0.356
T_Tb.BMD	0.283	0.274	–0.06	–0.044	–0.002	0.168	–0.513*	0.086	0.241	0.176	0.384	–0.058
T_Tb.BV_TV	0.282	0.273	–0.058	–0.048	–0.006	0.165	–0.516*	0.084	0.239	0.173	0.382	–0.06
T_Tb_Th	0.177	0.198	–0.153	0.031	–0.024	0.075	–0.18	0.034	0.136	0.079	0.191	–0.025
T_Ct_Th	–0.102	–0.117	0.584*	0.243	0.048	0.23	–0.17	0.267	0.184	–0.092	–0.253	–0.324
T_Ct_po	0.566*	0.173	–0.006	–0.076	0.26	0.279	0.093	0.115	0.394	0.427	0.481*	0.328

Vegans without self-reported resistance training
	Energy (kcal)	Protein per kg bodyweight per day	Vitamin A (retinol equivalent, mg per day)	Vitamin B12 (cobalamin, µg per day)	Vitamin B6 (pyridoxine, mg per day)	Vitamin B9 (folic acid, µg per day)	Vitamin D (calciferol, µg per day)	Vitamin K (µg per day)	Calcium (mg per day)	Magnesium (mg per day)	Phosphate (mg per day)	Zinc (mg per day)
R_Tt.BMD	0.014	0.023	0.003	–0.282	–0.013	0.106	0.232	0.409	0.173	0.056	–0.024	–0.009
R_Ct.BMD	–0.043	0.089	0.285	–0.362	–0.076	–0.003	0.007	0.237	0.167	–0.028	–0.028	0.082
R_Tb.BMD	0.166	–0.138	–0.281	0.034	0.115	0.181	0.326	0.212	0.041	0.121	0.002	–0.071
R_Tb.BV_TV	0.164	–0.139	–0.286	0.033	0.115	0.179	0.33	0.21	0.039	0.123	0.001	–0.071
R_Tb_Th	0.146	–0.225	–0.122	0.015	0.004	0.157	0.179	0.043	–0.026	0.051	–0.052	–0.205
R_Ct_Th	0.004	–0.069	0.146	–0.285	–0.055	–0.048	0.127	0.215	0.176	0.027	0.01	0.003
R_Ct_po	0.18	–0.251	–0.234	0.492*	–0.116	–0.27	0.344	–0.117	–0.191	0.126	0.004	–0.025
T_Tt.BMD	0.383	0.194	0.072	–0.339	0.475*	0.462*	0.115	0.23	0.466*	0.435*	0.336	0.192
T_Ct.BMD	–0.029	0.396	0.052	–0.598*	0.253	0.326	–0.378	0.297	0.288	0.165	0.175	0.193
T_Tb.BMD	0.394	–0.102	0.051	0.059	0.173	0.162	0.256	–0.003	0.14	0.228	0.171	0.061
T_Tb.BV_TV	0.392	–0.099	0.047	0.057	0.171	0.16	0.257	–0.006	0.135	0.225	0.168	0.057
T_Tb_Th	0.209	–0.033	–0.089	–0.073	0.036	0.102	0.106	0.052	–0.077	0.109	0.026	–0.051
T_Ct_Th	0.26	0.237	–0.044	–0.573*	0.480*	0.359	–0.055	0.181	0.491*	0.457*	0.367	0.189
T_Ct_po	0.11	–0.374	–0.057	0.487*	–0.051	–0.208	0.545*	–0.169	–0.098	0.01	–0.06	–0.072

Abbreviations: Ct.BMD, cortical bone mineral density; Ct.Po, cortical porosity; Ct.Th, cortical thickness; R, radius; T, tibia; Tb.BMD, trabecular bone mineral density; Tb.BV/TV, trabecular bone volume fraction; Tb.Th, trabecular thickness; Tt.BMD, total bone mineral density.

*represents statistically significant relationships.

Similarly, in vegans not performing resistance training on a regular basis, there was a relatively low number of relations: Vitamin B12 was moderately related to radius (R)_Ct.Po and T_Ct.Po and strongly to T_Ct.BMD and T_Ct.Th, vitamin B6 moderately to T_Tt.BMD and T_Ct.Th, vitamin B 9 moderately to T_Tb.BMD, vitamin D strongly to T_Ct.Po, calcium moderately to T_Tt.BMD and T_Ct.Th, and magnesium to T_Tt.BMD and T_Ct.Th.

Discussion

This study investigated trabecular and cortical bone microarchitecture in people adhering to a long-term purely plant-based diet. We observed significant differences in microarchitectural parameters of both cancellous and cortical bone at the radius and tibia compared with matched omnivores. However, self-reported resistance training was related to better microarchitecture.

When adhering to a plant-based diet, certain risk factors for bone loss need to be considered.

Frequently, protein intake seems to be low in vegans (30-32). Protein consumption promotes calcium absorption, but an adequate protein intake is necessary for this positive effect (33). As protein is an important component of the bone matrix, the structural integrity of the collagenous part of bone also depends on sufficient protein intake (3). In addition, protein is needed for maintaining muscle mass, which in turn influences bone mass, as demonstrated in women with sarcopenia (34).

In addition, as a purely plant-based diet is relatively poor in vitamin B12, in cases of vitamin B12 deficiency homocysteine levels can rise, which is an independent risk factor for high bone turnover (35).

An adequate intake in calcium and vitamin D is needed for the accretion and maintenance of bone mass (36). The intake of calcium might be lower in vegans (32, 37) and it was proposed that the bioavailability of plant-based calcium could be lower (32). In addition, vitamin D intake may be lower in vegans (38). Vitamin D3 seems to more effectively raise and maintain serum 25OH vitamin D levels than vitamin D2 (39, 40). Moreover, vitamin D3 is found predominantly in meat, eggs, and fish, whereas plant-based food mainly contains vitamin D2 (40). These differences highlight the importance of testing serum 25OH vitamin D levels as well as the possible indication for a supplement when adhering to a purely plant-based diet.

In our study, the mean intakes of protein, vitamin B12, calcium, and vitamin D in the vegan group were adequate, without significant differences within the 2 vegan subgroups. However, an adequate intake of vitamin B12 and vitamin D for vegans was mainly due to supplemental intake of these nutrients or due to intake of fortified foods. Both resulted in a relative high intake of vitamin B12 compared with omnivores. Exclusion of supplemental intake of vitamin B12 reduced the mean intake of vitamin B12 to a level slightly below the recommended intake (3.8 µg/day vs 22.0 µg/day). Similarly, vitamin D intake in vegans resulted mainly from supplemental intake leading to an adequate intake of this nutrient compared with the European Food Safety Authority daily reference value (DRV) of 15 µg/day (41) and assuming minimal endogenous synthesis. Without use of supplements, both vegans and omnivores would not reach adequate dietary intakes for vitamin D. However, dietary intake of vitamin D is not essential provided that exposure to sunlight is sufficient for endogenous production of vitamin D. The adequacy of vitamin D status is therefore better reflected by serum concentrations of 25OH vitamin D. No differences for this marker were observed between vegans and omnivores. Both groups had mean values above 20 ng/mL (50 nmol/L), a level considered as adequate (42). The intake of nutrients typically provided by plant foods, such as folates, provitamin A carotenoids, and vitamin K were significantly higher in vegans than in omnivores.

Concerning the relationship between nutrition and bone health, the International Osteoporosis Foundation identified several nutrients that one should ensure sufficient dietary intake to prevent bone loss (22). For vegans, it is already common knowledge that a purely plant-based diet is often low in vitamin B12 and that supplementation may be indicated (43). Moreover, it is well known that sufficient protein intake is necessary for bone health (30, 31).

A plant-based diet may also involve protective factors for bone health. First, dietary acid load, which seems to be indirectly related to bone mass (44), appears to be lower in vegans than in omnivores (24, 30).

Moreover, the overall vitamin K intake may be higher in vegans than in omnivores (30). Vitamin K may influence fracture risk and as well as bone matrix quality and mineralization (45, 46). However, there are several forms of vitamin K and especially K2 seems to be important for bone health (46, 47). Even though the human body can convert vitamin K1 to K2 to some extent (48), overall intake in vitamin K2 may be lower in vegans than in omnivores due to its limited sources (46).

In our vegan subjects, the mean dietary acid load was lower, and vitamin K intake higher, without significant differences among the 2 subgroups. Due to lack of data on the different forms of vitamin K in the food composition database, no information on the intake of vitamin K1 and vitamin K2 can be provided. It is, however, unlikely, that vegans have a low intake of vitamin K2, since it not only occurs in animal food but also in fermented food such as fermented soy beans (49). C-reactive protein values were not available.

Even though the vegans in our study showed significant differences in bone microarchitecture, compared with omnivores, we observed no clear relationships with nutrient intake. Therefore, based on our data, the structural alterations cannot solely be explained by deficits in certain nutrients according to lifestyle. The vegans in this investigation had an adequate intake in nutrients according to the current European Food Safety Authority recommendations (50), with no significant differences between the subgroups, likely due to the intake of supplements and/or fortified foods.

Previous studies reported a lower BMD and an increased fracture risk in vegans (7). A very recent investigation discovered an increased risk of fractures only in women who did not supplement both calcium and vitamin D (51), but others reported increased fracture risk even after correction for BMI, and calcium and protein intake, implicating that other factors also seem to contribute to bone health (5).

While previous studies on bone health in vegans only took BMD, and biochemical and nutritional parameters into account, they did not consider the significant effects of physical activity. By ignoring these effects, important factors influencing bone health are neglected.

Progressive resistance training causes mechanical strain exceeding the threshold at which the skeleton increases bone modeling (52). These mechanical loads lead to biological effects, which are summarized under the term “mechanotransduction.” Osteocytes are the primary mechanosensing organs within the bone. In particular, recent work suggests that mechanical stimulation activates the canonical Wnt/β-catenin pathway, resulting in gene transcription and bone formation, but other pathways also seem to be involved (53).

These findings highlight the impact of progressive resistance training on bone health. Still, the impact of different sports activities on bone varies significantly. While aerobic activity alone appears to exert no osteoanabolic effect, progressive resistance training is known to improve BMD (54, 55). In a study on young adult runners, resistance training once a week led to significantly greater BMD (27). Therefore, we assigned our subjects to the “resistance training” group, if they reported performing resistance training, in other words using machines, free weights, or bodyweight resistance exercises at least once a week.

Vegans performing resistance training at least once a week had a stronger bone microarchitecture, especially at the tibia, than those who did not. Between the omnivore subgroups, however, the differences were smaller. Figure 2 highlights the impact of resistance training between the different subgroups even more: Vegans performing resistance training showed bone structure similar to omnivores, while vegans not performing resistance training had weaker structure parameters. Based on these results, a stronger effect of resistance training on bone microarchitecture in vegans than in omnivores is likely. Thus, resistance training seems to be particularly important to preserve bone health when adhering to a plant-based diet.

Both in the vegan and in the omnivore group, serum markers were within normal ranges. However, total alkaline phosphatase levels were higher in the vegan group, whereas lower calcium levels and a trend toward higher parathyroid hormone levels were observed; vitamin D levels were similar in both groups.

Lower calcium levels could be explained by lower calcium intake in the vegan group.

Parathyroid hormone secretion is exercise dependent. In particular, an exercise threshold needs to be exceeded to increase PTH secretion (56, 57). Within the vegan group, those subjects performing resistance training had slightly (but not significantly) higher PTH levels. However, in the vegan group, the percentage (46%) of the subjects performing resistance training on a regular basis was slightly lower than the omnivore group (55%). Alternatively, higher PTH levels could reflect the attempt to absorb more calcium from the gut due to a relative insufficiency in calcium intake, even though no differences in self-reported calcium intake were observed. This hypothesis would also explain the finding that in vegans performing resistance training, PTH levels were higher than in omnivores performing resistance training.

Resistance training seems to influence bone-specific alkaline phosphatase serum levels (58). In addition, bone-specific alkaline phosphatase serum levels may vary with calcium intake (59). Bone-specific alkaline phosphatase levels were not available in our subjects. However, higher total alkaline phosphatase levels in the vegan group could be partly explained by lower calcium intake.

Our study has some limitations. When taking into account the maximal precision error of first-generation HR-pQCT scanners, some differences seem to be rather small, such as tibial Tb.Th, between the 2 vegan subgroups. However, according to the literature, the precision error for trabecular and cortical structure is even lower than 5% (16). Therefore, the observed differences in bone microarchitecture are considered to be correct.

Even though an observational study cannot imply causal conclusions, the uncertainties are likely to be similar in all subgroups and confounders are considered to be normally distributed to the total sample. Moreover, high authenticity and objectivity can be assumed.

Nutrient intake and resistance training were self-reported, but normal distribution was checked for all subgroups separately and homogeneity of variance had to be given for conduction of parametric analysis. Furthermore, nutrition protocols are a widely used method. Although the accuracy of self-reported physical activity is limited (60, 61) and there is no consensus on the quantification of resistance training (62), we overcame this limitation by allocating subjects into either the resistance training active or inactive group.

Compared with other nutritional studies (63, 64), our sample size was clearly smaller. Therefore, our study may have missed some associations between nutrient uptake and bone microarchitecture.

In the correlation analysis, a relatively high number of comparisons (168) was performed. After Bonferroni correction, the level of significance would change to 0.0003 and none of the correlations would be significant. However, in the correlation analysis, relationships rather than comparison of mean values are analyzed and these relationships do not necessarily refer to causality. Finally, Bonferroni correction is a relatively conservative method with high susceptibility for type 2 error (overlooking significant relationships). Therefore, Bonferroni correction was not performed.

Conclusion

In summary, trabecular and cortical bone microarchitecture in subjects with a long-term vegan diet differed from matched omnivores. We observed significant alterations in both trabecular and cortical microarchitecture, which could not be solely explained by differences in nutrients intake. Resistance training was associated with better bone structure and showed a stronger effect in vegans than omnivores. Therefore, the results of the present study suggest that progressive resistance training on a regular basis should be incorporated in the vegan lifestyle, in addition to a well-planned diet to maintain bone health.

Abbreviations

Abbreviations
 
• BMD

  bone mineral density

 
• BMI

  body mass index

 
• Ct.BMD

  cortical BMD

 
• Ct.Po

  cortical porosity

 
• Ct.Th

  cortical thickness

 
• CTX

  cross-linked C-telopeptide

 
• HR-pQCT

  high-resolution peripheral quantitative computed tomography imaging

 
• NEAP

  net endogenous acid production

 
• PTH

  parathyroid hormone

 
• Tb.BMD

  trabecular BMD

 
• Tb.BV/TV

  trabecular bone volume fraction

 
• Tb.N/mm

  trabecular number

 
• Tb.Sp

  trabecular separation

 
• Tb.Th

  trabecular thickness

 
• Tt.BMD

  total BMD

Financial support

None.

Disclosures

R.W.H. has received research support and/or honoraria from UCB Pharma, Amgen Austria, and gbo Medizintechnik. He has nothing to disclose concerning this manuscript. C.M. has received speaker honoraria from Amgen, Novartis, Servier, Eli Lilly, Nycomed Pharma/Takeda, Kwizda Pharma, Boehringer Ingelheim, Actavis, and Daiichi Sankyo. He has also received educational grants/research support from the Austrian Society for Bone and Mineral Research, Roche Austria, Eli Lilly Austria, Eli Lilly International, and Amgen Austria. He has nothing to disclose concerning this manuscript. P.P. has received research support and/or honoraria from Amgen, Eli Lilly, Fresenius Kabi Austria, Servier Austria, and Shire. He has nothing to disclose concerning this manuscript. All other authors have nothing to disclose.

Data Availability

Some or all datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.