Effect of combined exercise and nutrition on bone density in postmenopausal women-a systematic review and meta-analysis

Abstract

Background

Postmenopausal women are subject to hormonal fluctuations, a rapid decline in bone mineral density (BMD) and a significant increase in fracture risk, and both exercise and nutritional interventions have a positive impact on BMD. The aim of this study was to evaluate the effects of combined exercise and nutrition interventions compared to single nutrition or exercise interventions on BMD in postmenopausal women and specific combined strategies to improve BMD.

Methods

A systematic search on PubMed, Embase, Cochrane Library, Web of Science, Google Scholar databases and studies identified until April 4, 2025 were performed following strictly the PRISMA evaluation guidelines.The risk of bias was assessed using the Cochrane Risk of Bias Tool, the quality of evidence was evaluated using the GRADE approach, and data analysis was performed using Stata 17.0.

Results

A total of 24 RCTs involving 2,236 postmenopausal women were included. Meta-analysis results showed that EN vs. N demonstrated a positive effect on BMD at the femoral neck (SMD = 0.17, 95% CI: 0.07–0.27, p = 0.0006), lumbar spine (SMD = 0.20, 95% CI: 0.07–0.33, p = 0.003), and total hip (SMD = 0.16, 95% CI: 0.03–0.28, p = 0.014). However, no statistically significant difference in BMD at the same anatomical site was observed EN vs. N. The robustness of the results of subgroup analysis is limited due to differences in intervention regimens, study populations, and sample sizes across different studies.

Conclusions

Based on current evidence, there is insufficient support for universal evidence-based recommendations regarding specific combined intervention strategies.

Trial registration/protocol registration

PROSPERO (CRD420251053527).

Supplementary Information

The online version contains supplementary material available at 10.1186/s12986-025-01025-9.

Introduction

Osteoporosis (OP) is a disease characterized by low bone mineral density, degeneration of bone tissue, destruction of bone microarchitecture, impaired bone strength, and bone fractures [1]. OP has a high prevalence in postmenopausal women, affecting about one-third of postmenopausal women, and its lifetime risk of fragility fracture due to osteoporosis exceeds even that of breast cancer [2]. Globally, OP affects about 18.3% of the population, with a prevalence as high as 35.3% in older women, according to a 2025 study [3]. Fractures due to OP are becoming increasingly common in women after the age of 55 years, leading to a large number of bone-related complications and increased mortality and healthcare costs [4]. The prevalence of OP increases significantly with age. The global risk of OP-related fractures in people over 50 years of age is expected to double by 2040 [3].

Despite significant advances in drug development and clinical application for the treatment of OP, current treatments still face many challenges, including patient concerns about side effects, insufficient evidence of long-term efficacy, bone loss after drug discontinuation, and limitations in treatment options [5]. These issues result in many patients failing to receive effective treatment, increasing the risk of fracture.

In addition to medication, adequate intake of calcium and vitamin D, avoidance of smoking and excessive alcohol consumption, weight-bearing and resistance training, and fall prevention are among the ways and means of preventing OP [1, 2, 6]. Calcium is a key nutrient for maintaining bone health, and inadequate calcium intake leads to increased bone resorption, while calcium supplements can significantly improve bone mineral density and reduce fracture risk [7]. Vitamin D significantly reduces fracture risk by promoting calcium absorption and regulating bone metabolism, especially in postmenopausal women [8, 9]. Protein is an important nutrient for bone health, and increased protein intake increases insulin-like growth factor-1 (IGF-1) levels, which promotes bone formation and inhibits bone resorption, and high protein intake is beneficial for maintaining bone density and reducing fracture risk [6, 10]. Calcium, vitamin D, and protein are key nutrients for maintaining bone health, and these three nutrients have synergistic effects in promoting bone health. A meta-analysis examined the preventive effect of combined calcium and vitamin D supplementation on fractures and bone loss in people over 50 years of age and found that combined supplementation significantly reduced fracture risk [8, 11, 12]. Adequate calcium, vitamin D, and protein intake during growth is important in preventing fractures in adolescents and older adults [13]. Dairy products are a good source of calcium, vitamin D, and protein, and their intake provides all three nutrients at the same time, which has a significant protective effect on bone health [14–16].

Bones adapt their structure over the course of life to ensure that they are strong enough to withstand the customary levels of loading they are subjected to without accumulating excessive microdamage or fractures [17]. This functional adaptation is achieved through load-induced (re)modeling processes, leading to changes in mass and structure that optimally adjust the skeletal framework to its prevailing mechanical environment [18]. Numerous studies indicate that physical activity and nutritional factors—such as calcium and vitamin D intake—benefit bone health across different life stages [19–21]. Research also suggests that combining exercise with dietary components like calcium, creatine, and protein may produce synergistic effects that positively influence bone health [3, 22–24].

This study integrates fragmented evidence through systematic reviews and meta-analyses to quantify the combined effects of exercise and nutrition interventions versus either nutrition or exercise alone on BMD in postmenopausal women. It identifies optimal combination strategies for enhancing BMD, providing evidence-based guidance for preventing and managing osteoporosis in this population while optimizing non-pharmacological intervention approaches.

Methods

The review was conducted in the International Prospective Registry for Systematic Evaluation (CRD420251053527), following the Cochrane Reporting Handbook for Systematic Evaluation, Meta-analysis and the PRISMA guidelines, in accordance with the PRISMA guidelines [25].

Search strategy

The literature search databases included PubMed, Embase, Cochrane Library, Web of Science, and Google Scholar. The search period spanned from the inception of each database to April 4, 2025. Randomized controlled trials examining the combined effects of exercise and nutrition on BMD in postmenopausal women were retrieved from each database. Key search terms included:“Exercise, aerobic exercise, resistance training, nutrients, dietary supplement, proteins, postmenopausal women, bone mineral density, randomized controlled trial”.The literature was manually searched and included as comprehensively as possible. Only studies published in English were included.

Inclusion and exclusion criteria for studies

Inclusion criteria were determined using the PICOS (Participants, Interventions, Comparisons, Outcomes, and Study Design) methodology. (i) Participants: postmenopausal women diagnosed with reduced or normal bone mass or with osteoporosis. (ii) Intervention: Combination of physical activity and nutritional intake. Exercise includes any form of bone-beneficial exercise (aerobic, resistance, aquatic, whole body vibration, combined exercise, etc.), and nutrition includes all substances that are beneficial to the bones (calcium, vitamin D, proteins, creatine, etc.), and the combination of physical activity and nutritional intake was the main component of the intervention group. Controls: no nutritional intake or placebo in the control group with the same amount of physical activity as in the intervention group or in the control group with the same amount of nutritional intake as in the intervention group with maintenance of daily physical activity but no physical activity. (iii) Outcome measures: femoral neck (FN), lumbar spine (LS), total hip (TH), Ward’s triangle (WT), and whole body (WB) BMD. (iv) Study design: published randomized controlled trials of exercise and nutrition on BMD in postmenopausal women from the construct to April 4, 2025 in each database. Exclusion criteria: (i) Animal experiments; (ii) Published duplicates; (iii) Women with conditions affecting bone metabolism and those taking any medication that affects bone metabolism; (iv) Non-compliance of outcome metrics.

Literature screening and data extraction

Import the retrieved literature into EndNote 21 software and remove duplicate records. Subsequently, two researchers screened the literature and extracted information based on predetermined inclusion and exclusion criteria. Disagreements were resolved through discussion with a third researcher. Characteristics of included studies comprised authors, publication year, country, intervention description (i.e., exercise type, intervention period, frequency, intensity, duration), adherence, sample size, age, exercise status/nutritional intake during the control group trial period, and outcome measures (Tables 1 and 2). BMD at study completion was extracted as the dependent variable, reported in g/cm².

Table 1.

Research characteristics of included research (I)

Athor	Year	Research subject’s bone health status	Country and area	Sample population (n)			Age (Years)
				EN	E	N	EN	E	N
Prince [29]	1995	Bone healtha	Australia	42		42	63 ± 5		62 ± 5
Bemben [30]	2000	Osteopeniaa	America	7		8	51.9 ± 2.3		52.3 ± 1.4
Iwamoto [31]	2001	Osteoporosisa	Japan	30		36	65.3 ± 4.7		64.9 ± 5.7
Kemmler [32]	2002	Osteopeniaa	Germany	59		41	55.4 ± 3.2		55.8 ± 3.3
Going [33]	2003	Bone healtha	America	71		59	55.8 ± 4.7		57.1 ± 5.0
Engelke [34]	2006	Osteopeniaa	Germany	48		30	55.2 ± 3.3		55.5 ± 3.0
Wu [35]	2006	Bone healtha	Japan	31	31	33	54.4 ± 2.9	55.2 ± 2.8	53.8 ± 2.9
Evans [36]	2007	Bone healtha	America	11		10	62 ± 5		62 ± 5
Kemmler [37]	2007	Osteopeniab	Germany	40		28	55.1 ± 3.3		55.5 ± 3.0
Chuin [38]	2009	Bone healtha	Canada	8	11	8	66.3 ± 4	65.4 ± 3.5	65.5 ± 4.5
Cornish [39]	2009	Bone healthc	Canada	25	26		67 ± 7	65 ± 7
Choquette [40]	2011	Bone healthd	Canada	23	25	26	61 ± 3	58 ± 6	58 ± 5
Basat [41]	2013	Osteopeniaa	Turkey	14		14	55.9 ± 4.9		56.2 ± 4
Chilibeck [42]	2013	Bone healthe	Canada	72	77	76	55.8 ± 5	55.3 ± 6.3	56.7 ± 6.5
Daly [43]	2014	Bone healthf	Australia	53	47		73 ± 6.42	73 ± 7.7
Moreira [44]	2014	Bone healtha	Brazil	64		44	58.6 ± 6.71		59.3 ± 6.07
Chilibeck [45]	2015	Bone healthg	Canada	23	24		57 ± 6	57 ± 7
Knobf [46]	2016	Bone healtha	America	76		78	50.6 ± 5.6		53.1 ± 7.2
Mason [47]	2016	Bone healthh	America	93	94		60.3 ± 5.3	59.0 ± 4.7
Yu [48]	2019	Osteopeniaa	China	40		40	61.5 ± 7.5		62.5 ± 6.6
ElDeeb [49]	2020	Osteoporosisi	Egypt	22		21	77.8 ± 3.6		79.2 ± 4.7
Sen [50]	2020	Osteoporosisa	Turkey	15		18	55 ± 4.6		54.5 ± 6
Waltman [51]	2022	Osteopeniaj	America	92		93	54.3 ± 3.3		54.4 ± 3.1
Chilibeck [52]	2023	Osteopeniak	Canada	120	117		59 ± 5.6	59 ± 5.7

EN, combined exercise and nutrition group; E, single exercise group; N, single nutrition group; NR, not record; All values are presented as mean ± SD

a No adverse events

b 1 case of hairline fracture, 1 case of periostitis, 2 cases of pulled muscles

c 3 individuals in a single exercise group reported stomach discomfort, nausea, or diarrhea, all of which were related to corn oil supplements, 1 female accidentally fell outside of training, resulting in patellar fracture (unrelated to intervention)

d Many women in the combined and single exercise groups reported an increase in hot flashes during initial training, but this condition improved after several training sessions

e Serious adverse events have been reported that are unrelated or unlikely to be related to the study intervention, such as hysterectomy, ductal carcinoma, stroke, etc

f 4 individuals reported mild muscle pain and no deterioration of kidney function

g 5 individuals in the combined group reported constipation, diarrhea, heartburn, irritability, and nausea (all mild), 2 individuals reported mild to moderate muscle spasms, and 20 cases reported mild combined pain and strains; 2 individuals in the single exercise group showed mild elevation of liver enzymes or bilirubin levels

h 5 individuals reported non research related sports injuries (such as plantar fasciitis and exacerbation of osteoarthritis)

i 2 individuals in the combined group had migraine; 2 individuals in the single nutrition group stopped taking medication

j Constipation in 7 individuals in the single nutrition group

k 15 individuals in the combined group had mild elevation of creatinine, 7 individuals had elevation of transaminase, 13 individuals had bloating, diarrhea, nausea, 1 individual had a ‘severe’ fall, and multiple individuals experienced mild to moderate falls; 21 individuals in the single exercise group had elevated creatinine levels, 5 individuals had elevated transaminase levels, 14 individuals had bloating, diarrhea, and nausea, and multiple individuals experienced mild to moderate falls

Table 2.

Research characteristics of included research (II)

Athor	Nutritional intervention	Nutritional dosage	Exercise intervention	Intervention cycle	Attendance (%)	Outcome indicators(BMD)units (g/cm2)
Prince [29]	Calcium Calcium lactate gluconate	1 g/day Take before bedtime	MF: aerobic training Supervised weight training: 1 h per session, twice weekly; independent brisk walking or walking: 2 h/week. EF: 3 times/week 1 h each time EI: moderate intensity Heart rate reaches 60% of age-predicted maximum heart rate	2 years	N: 97%~100% EN: 39% Despite low attendance rates, the outcome indicators were favorable.	FN, TH
Bemben [30]	Calcium(Calcium lactate) + Vitamin D	1.5 g/day + 125IU/day oral tablets	MF: resistance training Performed under supervision: Leg press, leg extension, shoulder press, seated row, bicep curl, tricep extension, hip abduction/adduction/flexion, etc. 3 sets of 16 repetitions. EF: 3 times/week EI: moderate intensity 50%1RM	6 months	N: 92.3% EN: 87%	FN, TH, LS, WB, WT
Iwamoto [31]	Calcium(Calcium lactate) + Vitamin D(1α-hydroxyvitamin D₃)	2 g/day + 40IU/day oral tablets	MF: combined training Supervised daily regimen: brisk walking (30% increase in steps) + two sets of calisthenics daily; each set consists of 15 repetitions (straight-leg raises, squats, abdominal exercises, and lower back strengthening exercises). EF: ≥5 times/week EI: moderate intensity 3ཞ4 METs	2 years	N: NR EN: 71%	LS
Kemmler [32]	Calcium(Calcium carbonate) + Vitamin D(cholecalciferol−	1.5 g/day + 500IU/day oral tablets	MF: combined training Supervised warm-up: Jogging + games + low/high-impact cardio; Jumping: Jump rope 3–5 sets × 15–20 reps; Multi-directional jumps (4 variations × 15 reps, including single-leg landings); Strength: Leg press, bench press, rowing, etc. (2–4 sets × 8–20 reps). Dumbbells/resistance bands, squats, bench press, deadlifts, etc.; Stretching (8–10 movements × 30 s). Home Training (2 times/week): Resistance band/isometric exercises + Jump rope (20 reps/set) EF: 4 times/week EI: moderate-high intensity 50–90%1RM	1 year	N: 75% EN: 75%	FN, LS, TH
Going [33]	Calcium (Calcium citrate )	0.8 g/day oral tablets	MF: resistance training Supervised warm-up; strength training (free weights + machines, 70–80% of 1RM); Weighted circuit training (walking/jogging, jumping, stair climbing + weighted vest) EF: 3 times/week EI: moderate-high intensity 70–80%1RM	1 year	N: 92% EN: 75%	FN, LS, WB
Engelke [34]	Calcium(Calcium carbonate) + Vitamin D(cholecalciferol)	1.5 g/day + 500IU/day oral tablets	MF: combined training Supervised group training (twice weekly) Warm-up/endurance; Jump training: Jump rope + 4 sets × 15 reps multi-directional jumps; Resistance training: 13 exercises (major muscle groups); 2–4 sets per exercise, decreasing reps per set (20→8); Bench press, single-arm row, squat/deadlift, 2–4 sets; Stretching: Before and after training, 30 s per muscle group × 1–2 sets Home Training (2 times/week) Jumping rope + resistance bands + isometric exercises + stretching EF: 4 times/week EI: moderate-high intensity 70–92.5.5% 1RM	3 years	N: 100% EN: 100%	LS, TH
Wu [35]	Protein (isoflavone)	0.075 g/day morning oral capsule	MF: aerobic training Supervised warm-up + brisk walking (5–6 km/h) + cool-down EF: 3 times/week 1 h each time EI: moderate intensity 3.5–4.0 METs	1 year	N: 73.5% E: 70.5% EN: 88%	FN, LS, TH, WB, WT
Evans [36]	Protein (protein drink་isoflavone)	25.6 g protein drink + 0.091 g/day ready-to-drink protein drinks	MF: aerobic training conducted under supervision, utilizing various equipment such as indoor tracks, treadmills, rowing machines, and stair climbers. EF: 3 times/week EI: high intensity 75–80% VO₂peak	9 months	NR only report the overall figure of 70.5%	FN, LS, WB
Kemmler [37]	Calcium(Calcium carbonate) + Vitamin D(cholecalciferol)	1.5 g/day + 500IU/day oral tablets	MF: combined training Supervised warm-up/cardio (running + games + high-impact cardio); Jumping (4 types of multi-directional jumps, 15 reps each); Horizontal leg raises, leg curls, bench press, rowing, etc., 3–4 sets per exercise; Bodyweight/Resistance Band Session: 12–15 isometric contractions (2–4 sets) + 3 free weight exercises (Squat/Deadlift, Single-Arm Dumbbell Row, Dumbbell Chest Press) Home Training: Jump rope 3 sets × 20 reps, isometric contractions, resistance band exercises, stretching EF: 4 times/week 1 h each time EI: moderate-high intensity 70–90%1RM	4 years	N: 54.9% EN: 47.7% Reasons for withdrawal: relocation, illness, loss of interest, failure to meet training frequency requirements. Participants training less than twice per week were excluded from analysis.	LS, TH
Chuin [38]	VitaminC + VitaminE	1 g/day + 0.6 g/day oral tablets	MF: resistance training Performed under supervision: Leg press, bench press, leg extension, shoulder press, sit-ups, seated row, triceps extension, biceps curl; 3 sets of 8 repetitions per exercise. EF: 3 times/week 1 h each time EI: high intensity 80%1RM	6 months	NR	FN, LS
Cornish [39]	α-Linolenic acid Flaxseed oil (rich in alpha-linolenic acid, ALA)	14 g/day flaxseed Oil (for Oral Use)	MF: resistance training Performed under supervision, familiarize yourself with the movements: - Muscle building: 2–3 sets of 10–12 reps per set; - Strength enhancement: 3–4 sets of 6–10 reps per set; - Maintenance: 3 sets of 10–12 reps per set. EF: 3 times/week 1 h each time EI: moderate-high intensity 70–85%1RM	3 months	E: 85% EN: 85% EN 1 person withdrew due to an inability to tolerate the taste of the supplement.	LS, TH, WB
Choquette [40]	Protein (isoflavone)	0.07 g/day oral capsules	MF: combined training Supervised training: Leg press, bench press, lat pulldown (4 sets of 4–6 reps) + Power bike, treadmill (4 × 4 min high-intensity intervals + 3 min rest intervals) EF: 3 times/week 1 h each time EI: high intensity 85%1RM	6 months	NR	FN, LS, TH, WB, WT
Basat [41]	Calcium + Vitamin D	1.2 g/day + 800IU/day	MF: resistance training Supervised session: Stationary bike (50 W, 10 min) + Spot walking + Static stretching (10 reps × 1 set per muscle group); Abdominal isometric contraction; Hip extensor, flexor, abductor, and adductor isometric contractions; Knee extensor and flexor isometric contractions; Push-ups EF: 3 times/week 1 h each time EI: high intensity 85%1RM	6 months	E: 85.7% EN: 79%	FN, LS
Chilibeck [42]	Protein (isoflavone) + Calcium(Calcium carbonate) + Vitamin D (cholecalciferol)	0.165 g/day + 1.2 g/day་800IU/day oral tablets	MF: resistance training Supervised brisk walking combined with strength training for major muscle groups: squats and bench presses. Perform 2 sets of 8 repetitions per exercise. EF: 4 times/week 20–30 min/session EI: high intensity 80%1RM	1 year	E: 70% N: 65% EN: 74%	FN, LS, TH, WB, WT
Daly [43]	Protein (Lean Red Meat)	160 g/day Lean red meat	MF: resistance training Supervised warm-up: Squats, lunges, box step-ups, leg extensions, standing leg curls, hip abductions, calf raises, shoulder presses, upright rows, bicep curls, wall push-ups, tricep kickbacks, etc. Using dumbbells, ankle weights, and fitness balls. 3 sets × 8–12 reps; Stretching .EF: 2 times/week 40–60 min/session EI: high intensity based on the RPE, the target is set at 14–16 points.	4 months	N: 72% EN: 75%	FN, LS, TH
Moreira [44]	Calcium(Calcium carbonate) + Vitamin D(Cholecalciferol)	0.5 g/day + 1000IU/day oral tablets	MF: resistance training Supervised warm-up; Strength/Explosive Power Training: Performed using multiple sets with decreasing rest intervals (standing in water up to chest level) Elbow flexion + shoulder abduction, hip abduction/adduction, shoulder horizontal flexion + elbow extension, knee extension (high kicks)/hip extension; Stretching EF: 3 times/week 1 h each time EI: moderate-high intensity 60–90% HRmax	6 months	N: 88.6% EN: ≥85%	FN, LS, WB
Chilibeck [45]	Creatine (creatine monohydrate)	0.1 g∙kg⁻¹/day take with meals	MF: resistance training under supervision, hack squat、hamstrings curl、knee extension、back extension、bench press、lat pulldown、shoulder press、biceps curl、triceps extension、ankle dorsiflexion & plantarflexion Perform 3 sets of each exercise, 10 repetitions per set, aiming for “muscle failure.” EF: 3 times/week EI: high intensity 80%1RM	1 year	E: 77% EN: 75%	FN, LS, TH, WB
Knobf [46]	Calcium + Vitamin D (NR)	1.2 g/day + 400IU/day oral tablets	MF: combined training Performed under supervision, brisk walking on a treadmill、leg press、toe press、leg extension、leg curl、bent-knee sit-ups. 1 set of 8 repetitions per exercise. EF: 3 times/week 30 min/session EI: moderate-high intensity 70%1RM	1 year	E: 97% EN: 97.9%	FN, LS, TH
Mason [47]	Vitamin D (Vitamin D3)	2000IU/day oral tablets	MF: aerobic training Supervised activities: treadmill walking or jogging, stationary cycling, aerobic equipment training, walking/hiking, aerobic exercise classes, cycling EF: 5 times/week ≥ 45 min/session EI: moderate-high intensity	1 year	N: NR EN: 77.4%	FN
Yu [48]	Calcium་Vitamin D (NR)	0.6 g/day + 800IU/day oral tablets	MF: aerobic training Aerobic dance training conducted under supervision, including A-steps, V-steps, marching steps, cross-steps, kicks. EF: 3 times/week 1 h each time EI: high intensity maintain 50%−70% of your individual target heart rate (monitored via POLAR heart rate monitor)	6 months	N: NR EN: 81%	FN
ElDeeb [49]	Calcium + Vitamin D (NR)	1.2 g/day + 800IU/day oral tablets	MF: whole-body vibration Under supervision, the platform vibrates vertically at a frequency of 20–35 Hz, amplitude 2.5–5 mm half squat, wide-stance squat, standard squat, deep squat, single-leg squat, single-leg stance, lunge Hold each position for 30–60 s squatting exercises; repeat 3–9 times EF: 2 times/week EI: moderate intensity 20–35 Hz་2.5–5 mm	6 months	N: NR EN: 88%	FN, LS, WT
Sen [50]	Calcium + Vitamin D (NR)	1.5 g/day + 880IU/day oral tablets	MF: whole-body vibration Under supervision, the platform vibrates vertically at a frequency of 30–40 Hz, amplitude 2–4 mm squats, deep squats, lunges, wide-stance squats, front lunges EF: 3 times/week EI: moderate intensity 20–35 Hz + 2–4 mm	6 months	N: 90% EN: 78.9%	FN, LS, TH
Waltman [51]	Calcium(Calcium citrate) + Vitamin D(cholecalciferol)	1.2 g/day + ≥1000IU/day	MF: combined training Supervised weight-bearing running; strength training for major muscle groups (8–12 repetitions to failure) EF: 3 times/week EI: high intensity 70–85% 1RM	1 year	N: 94.2% EN: 58.9% A few people experience mild muscle soreness	FN, LS, TH
Chilibeck [52]	Creatine	0.14 g∙kg⁻¹/day	MF: combined training Supervised brisk walking training; hack squat、hip abduction/adduction/flexion/extension、bench press、lat pull-down、shoulder press、hamstrings curl、quadriceps extension、biceps curl、triceps extension、back extensio༛2 sets per exercise, 8 reps per set EF: 3–6 times/week 30–60 min/session EI: high intensity ≥80% 1RM	2 years	E: 68% EN: 65%	FN, LS, TH, WB, WT

The intervention group (combined exercise and nutrition group) includes exercise intervention and nutrition intervention; and the control group (single nutrition group and single exercise group); The single nutrition group only received nutrition intervention; and the single exercise group only received exercise intervention; FN Femoral neck; LS Lumbar spine; TH Total hip; WB Whole body; WT Ward’s triangle area; MF Main form of exercise; EF Exercise frequency; EI Exercise intensity, Report exercise intensity according to internationally accepted standards (WHO/ACSM) [53, 54]; RPE Borg Rate of Perceived Exertion scale; 1RM 1-Repetition Maximum; HRmax Maximum Heart Rate; VO₂peak Peak Oxygen Uptake; Hz Hertz; METs Metabolic Equivalents

Quality assessment

The risk of bias for the 24 trials included in this paper was assessed by a self-contained tool in Review Manager 5.4 [26]. Two researchers separately assessed the included literature using the Cochrane Risk of Bias Assessment Tool for seven aspects of quality assessment in terms of randomized sequence generation allocation concealment, blinding of participants and researchers, evaluation of outcome data, data completeness, selective reporting, and other biases Each element was assessed as high risk of bias, low risk of bias, or unknown risk of bias. Disagreements during the assessment process were resolved through discussion.

The quality of evidence for each statistically significant outcome was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) system [27]. Evidence grades will be downgraded by one level based on the following factors: risk of bias, inconsistency, indirectness of evidence, imprecision, and publication bias. Conversely, evidence will be upgraded by one level due to any of the following: large effect size, dose-response relationship, or all plausible biases reducing the apparent treatment effect only. GRADE evidence quality is categorized into four levels: high, moderate, low, or very low [27], S2.

Data synthesis

The missing standard deviation (SD) was calculated using the method detailed in the recent comprehensive meta-analysis by Shojaa et al. [28]. If studies presented confidence intervals (CI) or standard errors (SE), these were converted to SD using a standardization formula [26].

Data analysis

This study employed Stata 17.0 software for meta-effect size analysis and Reviewer Manager 5.4 to construct Cochrane quality assessment plots. The outcome measures in the included studies were continuous variables, with standardized mean difference (SMD) and 95% confidence interval (CI) selected as effect measures for the pooled effect size. Statistical inference was conducted through heterogeneity testing and statistical combination of effect sizes. Heterogeneity was assessed using Chi-square and I² p-value tests. In the heterogeneity test, p > 0.10 indicated negligible heterogeneity among the included studies, while p ≤ 0.10 indicated heterogeneity. When p ≤ 0.10, the assessment of heterogeneity was primarily guided by the I² value. Specifically: 0 ≤ I² ≤ 25% indicates negligible heterogeneity 25% < I² ≤ 50% indicates mild heterogeneity 50% < I² ≤ 75% indicates moderate heterogeneity I² >75% indicates high heterogeneity Given the varying environmental contexts across studies and the resulting differences in intervention effect sizes, we opted to use a random-effects model. Forest plots were used for data presentation. Regression-based funnel plots and Egger’s test were employed to identify and analyze publication bias. Studies with significant bias (p < 0.05) underwent further assessment of potential publication bias using trimming and filling methods. Sensitivity analyses were conducted to examine the robustness of the findings. Meta-regression with robust variance estimation (RVE) was performed using the robumeta package to assess the effects of moderating variables on LS, FN, and TH-BMD in combined exercise and nutritional interventions. Statistical significance was set at p < 0.05 for all outcomes. SMD values of 0.2, 0.5, and 0.8 were interpreted as small, moderate, and large effects, respectively.

Results

Results of literature search

Preliminary screening identified 639 relevant articles, including 168 from PubMed, 115 from the Cochrane Library, 134 from Embase, 109 from Web of Science, and 113 from Google Scholar. EndNote 21 software removed 241 duplicates. After reviewing titles and abstracts, 325 articles were excluded. The remaining 73 full texts were assessed for inclusion. Among these, 5 interventions did not match the target group, 0 controls met the criteria, 16 interventions failed to meet inclusion criteria, 19 outcome measures did not match, and data could not be extracted from 8 studies. Ultimately, 24 articles were included in the meta-analysis [29–52], as shown in Fig. 1.

Fig. 1.

PRISMA research flow chart

Characteristics of the included studies

This study included 24 articles published between 1995 and 2023, involving a total of 2,236 subjects (1,079 in the combined exercise and nutrition group, 452 in the exercise-only group, and 705 in the nutrition-only group). Participants ranged in age from 50 to 80 years. Studies originated from Australia (2 [29, 43]), Japan (2 [31, 35]), Germany (3 [32, 34, 37]), Canada (6 [38–40, 42, 45, 52]), 6 from America [30, 33, 36, 46, 47, 51], 2 from Turkey [41, 50], 1 from Brazil [44], 1 from Egypt [49], and 1 from China [48]. All subjects were postmenopausal women with bone status categorized as bone health, low bone mass, or osteoporosis. 12 studies involved subjects with bone health [29, 33, 35, 36, 38–40, 42–47], 8 with Osteopenia [32, 34, 37, 41, 48, 51, 52], and 3 with osteoporosis [31, 49, 50](Table 1).

Nutritional interventions included having calcium supplements, vitamin D, protein, alpha-linolenic acid and creatine, with 2 articles using calcium supplements [29, 33], at a dose of 0.8–1 g/day, 1 article using vitamin D supplements [47], at a dose of 2000 IU/day, 12 articles using calcium and vitamin D supplements [30–32, 34, 37, 41, 44, 46, 48–51], at 0.5–2 g/day + 40–1000 IU/day,1 article using alpha-linolenic acid [39], at 14 g/day, 1 article using vitamin C and vitamin E supplements [38], at 1 g/day + 0.6 g/day, 2 articles using creatine [45, 52], at 0.1–0.14 g·kg⁻¹/day, 5 articles using protein [35, 36, 40, 42, 43], with isoflavone doses ranging from 0.07 to 0.165 g/day, and animal protein at 160 g/day.

Exercise interventions included aerobic training, resistance training, combined training, and whole-body vibration training. 8 studies employed resistance training [30, 33, 38, 41, 43–45], 5 used aerobic training [29, 35, 47, 48], 9 utilized combined training [31, 32, 34, 37, 40, 42, 46, 51, 52], and 2 studies involved whole-body vibration training [49, 50]. Intervention durations ranged from 3 months to 4 years, with 11 studies lasting < 1 year [30, 36, 38–41, 43, 44, 48–50], 8 studies lasting 1 year [32, 33, 35, 42, 45–47, 51], and 5 had durations of 2–4 years [29, 31, 34, 37, 52]. Weekly intervention frequency ranged from 2 to 6 sessions, with 2 studies at 2 sessions [43, 49], 15 at 3 sessions [29, 30, 33, 35, 36, 38–41, 44–46, 48, 50, 51], and 7 studies had 4–6 sessions per week [31, 32, 34, 37, 42, 47, 52]. Exercise intensity ranged from moderate to moderate-high and high intensity, with 7 studies using moderate intensity [29–31, 35, 41, 49, 50], 8 studies using moderate-high intensity [32–34, 37, 39, 44, 46, 47], and 9 studies involved high-intensity training [36, 38, 40, 42, 43, 45, 48, 51, 52]. The attendance rate for the combined group (EN) ranged from 39% to 97.9%, while the single nutrition group had rates between 54.9% and 100%. The single exercise group reported attendance rates from 70% to 85%. Overall compliance was good. Reasons for attendance rates below 60% are detailed in the table. Measured outcome indicators included BMD for FN, LS, TH, WB, and WT (Table 2).

Discussion

Meta-analysis results indicate that combined nutritional and exercise interventions exert beneficial effects on BMD at the FN, LS, and WT in postmenopausal women; however, the SMDs for all measurements were below 0.2. Both exercise and nutrition can independently improve bone health [12, 28, 55–57], suggesting that combined intervention may yield synergistic effects.This effect may be explained through several molecular mechanisms: calcium and vitamin D promote osteoblast differentiation and activity via the Wnt/β-catenin signaling pathway, and regulate parathyroid hormone (PTH) signaling to maintain the balance between bone formation and resorption [58]. Protein supplementation provides amino acids that stimulate IGF-1 production, enhancing osteoblast function and bone matrix synthesis [59, 60]. Exercise-induced mechanical loading activates mechanotransduction pathways including integrin-mediated signaling and the MAPK/ERK cascade, while suppressing osteopontin expression [58]. Collectively, these pathways promote osteogenesis and inhibit osteoclast-mediated bone resorption. Together, these pathways provide a mechanistic basis for how combined nutritional and exercise interventions improve bone mineral density [61]. the magnitude of this synergistic effect may be limited, resulting in a modest overall effect size. No significant improvement in BMD at the TH and WB was observed with the combined intervention. This may be attributed to the higher proportion of cortical bone in the hip and slower bone remodeling rate, making it difficult to achieve significant changes in the short term even with intervention. Furthermore, while EN vs. N showed a beneficial effect on TH BMD, the effect size remained modest. Skeletal responses to mechanical loading and nutrition exhibit anatomical site specificity; exercise stimulation is more likely to promote bone formation at sites of localized mechanical stress concentration [18, 62], whereas the systemic effects of nutritional interventions may require longer durations or higher doses to manifest.

This study found no significant advantage of EN over E in terms of BMD. The comparison primarily examined the effects of adding nutritional supplementation to exercise, yet no significant differences were observed at the FN, LS, TH, or WT sites. This result may be limited by insufficient sample size and differences in intervention characteristics, aligning with the meta-analysis findings of Fischer et al. [63].Combined interventions may exhibit a “threshold effect,” where additional nutritional supplementation yields only marginal improvements when exercise has already elevated skeletal adaptation to a high level, making further statistically significant gains difficult. Although differences did not reach statistical significance, exercise combined with nutrition may still hold potential value in clinical practice or long-term follow-up. This suggests that exercise may be the primary determinant in skeletal health management, while the additive effects of nutritional supplementation may only become more pronounced under specific conditions.

This study found that EN demonstrated more pronounced improvements in BMD at weight-bearing sites such as the FN, LS, and TH compared to N. This suggests exercise is the core factor in combined interventions. Previous studies have reported that exercise improves BMD at these sites [64], and our findings partially overlap with these reports. While Mohebbi et al. may have focused more on the effects of exercise alone versus control on BMD, this study further validates the combined effect of exercise and nutrition. Subgroup analysis revealed that intervention durations below 1 year and moderate-intensity training were more beneficial for FN and LS-BMD. This may relate to the rapid bone remodeling rate in these sites and their sensitivity to stable mechanical stimulation and nutritional supplementation [65–67]. A frequency of 3 exercise sessions per week positively influenced FN and TH-BMD, while higher frequencies did not show significant effects, potentially due to limited statistical power from small sample sizes. Furthermore, individuals with lower bone mass or at risk of bone loss demonstrated more pronounced improvements in the FN, LS, and TH subgroup analyses, suggesting that bone mass status may influence the efficacy of combined interventions. Those with lower bone mass exhibited greater sensitivity to mechanical loading and nutritional supply, resulting in more significant post-intervention bone formation and mineralization. This indicates that the effectiveness of combined interventions exhibits considerable individual variability.

A meta-analysis demonstrated that calcium and vitamin D fortified dairy products exert a favorable effect on BMD in postmenopausal women. Wolf et al. [3]. investigated the impact of combined protein and exercise interventions on bone health in middle-aged and older adults; however, due to the lack of relevant randomized controlled trials, reliable conclusions could not be drawn. Multiple dietary components may influence osteoclast function, potentially exerting synergistic effects on bone health when combined with physical activity [22]. However, evidence supporting synergistic interactions between physical activity and nutrition to produce benefits is considerably less robust [23]. Subgroup analyses in this study demonstrated positive effects on BMD improvement from supplementation with calcium, vitamin D, or protein.

Further analysis revealed that the combination of calcium, vitamin D, and mixed training significantly improved TH-BMD. Combined training incorporates aerobic and resistance exercises. Aerobic exercise enhances cardiopulmonary function and promotes skeletal blood flow [68], while resistance training stimulates weight-bearing bones through load application. The combination of these two modalities repeatedly applies mechanical stress to the hip during activities like walking, jumping, and squatting, aligning with the hip’s anatomical load-bearing characteristics. In contrast, while calcium and vitamin D combined with whole-body vibration training showed some benefits, only two studies were included, both involving osteoporotic populations, potentially overestimating the intervention’s effectiveness.

Multiple regression analysis revealed that the efficacy of combined interventions was significantly correlated with exercise frequency, duration, type, intensity, population, nutritional supplement type, and exercise-combined-nutrition strategies (all p < 0.05). This suggests that the improvement in BMD from combined interventions is not fixed but depends on the individualized design of exercise prescriptions and nutritional plans. These findings may underscore the multifactorial adaptive nature of bone tissue [69]. In summary, the benefits of combined interventions are not uniform and depend on the specific circumstances of postmenopausal women.

This study employed meta-regression to identify sources of heterogeneity among studies, sensitivity analysis to validate the robustness of results, and GRADE to assess evidence quality. Additionally, several limitations exist: we restricted the search to articles published in English, which may have introduced language bias and led to the exclusion of relevant studies published in other languages. effect sizes for combined interventions were generally small (SMD < 0.2), and some results failed to reach statistical significance, potentially due to insufficient sample sizes in included studies. Intervention protocols exhibited substantial heterogeneity in exercise type, intensity, frequency, duration, and nutritional supplement types, potentially affecting result consistency. Some subgroup studies (e.g., whole-body vibration training combined with nutrition) were extremely limited in number, yielding insufficient evidence robustness and raising concerns about publication bias. Future studies should design larger-sample, long-follow-up randomized controlled trials, considering baseline skeletal status, individualized intervention strategies, and multifactorial synergistic effects to further validate the efficacy of combined interventions.

Conclusion

This meta-analysis indicates that combined exercise and nutritional interventions exert a positive effect on BMD in postmenopausal women, particularly at weight-bearing sites including the femoral neck, lumbar spine, and total hip. However, the efficacy of combined interventions is significantly modulated by intervention characteristics (exercise type, intensity, frequency, duration), nutritional protocols, and baseline bone status in the population, revealing marked individualization and site-specificity. These findings do not support specific evidence-based recommendations.

Overall, exercise remains a core factor in managing bone health among postmenopausal women, with combined nutritional interventions offering synergistic benefits. Future research should focus on personalized intervention design, long-term efficacy, and multifactorial interactions to optimize specific intervention strategies for postmenopausal bone health.

Risk of bias

The overall quality of included studies was high. Risk of bias is summarized in Figs. 2 and 3. 24 studies were assessed using the Cochrane Systematic Review Tool Version 5.1.0 [29–52], yielding the following results: Random sequence generation (low, 19; uncertain, 1; high, 4) Allocation concealment (low, 13; uncertain, 4; high, 7), blinding of participants and personnel (low risk: 11; uncertain risk: 4; high risk: 9), blinding of outcome assessment (low risk: 12; uncertain risk: 10; high risk: 2), incomplete outcome data (low risk: 13; uncertain risk: 9; high risk: 2), selective reporting (low risk: 18; uncertain risk: 3), and other biases (low risk: 21; uncertain risk: 3).

Fig. 2.

Risk of bias in included studies

Fig. 3.

Summary of bias risk of included studies. (√), low risk; (×), high risk; (?), unclear or inadequate description

Meta-analysis results

Effects of combined exercise and nutrition intervention compared to single intervention on femoral neck bone mineral density

A total of 20 studies [29, 30, 32, 33, 35, 36, 38, 40–52] described the effects of EN versus N and E on FN-BMD, including 16 studies [29, 30, 32, 33, 35, 36, 38, 40–42, 44, 46, 48–51] comparing EN versus N and 8 studies [35, 38, 40, 42, 43, 45, 47, 52] comparing EN versus E. Low heterogeneity was observed between studies comparing EN vs. N (I² = 8.6%, p = 0.356), while mild heterogeneity existed between studies comparing EN vs. E (I² = 33.5%, p = 0.161). Using a random-effects model, the results indicated that: EN significantly increased FN-BMD compared to single intervention (E, N)(SMD = 0.17, 95%CI: 0.07–0.27, p = 0.0006).Compared with N, EN significantly increased FN-BMD (SMD = 0.20, 95% CI: 0.09–0.31, p = 0.0007). Compared with E, EN increased FN-BMD (SMD = 0.12, 95% CI: −0.05−0.30, p = 0.17) but without statistical significance (Fig. 4). No significant publication bias was detected by Egger’s test (EN vs. N: p = 0.115; EN vs. E: p = 0.407) or visual interpretation of funnel plots.

Fig. 4.

Forest Plot of the Effect of Combined Exercise and Nutrition Intervention Compared to Single Intervention on Femoral Neck Bone Mineral Density. EN Exercise and Nutrition Group; N Single-nutrient group; E Single-exercise group

FN-EN vs. N subgroup analysis

Since there was no statistically significant difference EN vs. E and the number of included studies was too small, no subgroup analysis was performed. Subgroup analysis was conducted only on the 16 studies included in the EN vs. N comparison (Table 3).

Table 3.

Effects of subgroup analysis on femoral neck BMD

Subgroup	N	SMD 95%CI	I2	Model	P
Duration of exercise program
<1 year	9	0.28(0.08–0.48)	33.1%	Fixed	0.006
= 1 year	6	0.18(0.04–0.33)	0%	Fixed	0.011
>1 year	1	−0.01(−0.44−0.42)	NR	Fixed	0.967
Heterogeneity between groups: p = 0.453
Motion intensity
Moderate intensity	6	0.40(0.01–0.78)	55%	Random	0.044
Medium to high intensity	4	0.10(−0.08−0.28)	0%	Random	0.265
High-intensity	6	0.22(0.04–0.40)	0%	Random	0.015
Frequency of exercise (per week)
2 days	1	0.76(0.14–1.38)	NR	Fixed	0.016
3 days	13	0.20(0.07–0.32)	6.8%	Fixed	0.003
4–6 days	2	0.13(−0.12−0.38)	0%	Fixed	0.318
Heterogeneity between groups: p = 0.175
Type of nutritional supplement
Calcium、vitamin D	11	0.18(0.06–0.31)	29.8%	Fixed	0.005
Protein	4	0.24(0.01–0.48)	0%	Fixed	0.043
Vitamin C + Vitamin E	1	0.54(−0.46−1.54)	NR	Fixed	0.293
Heterogeneity between groups: p = 0.73
Training type
Aerobic training	4	0.15(−0.11−0.50)	0%	Fixed	0.228
Resistance training	5	0.13(−0.13−0.38)	0%	Fixed	0.271
Combined training	5	0.17(0.11–0.45)	0%	Fixed	0.031
Whole body vibration	2	0.76(0.14–1.38)	0%	Fixed	0.0001
Heterogeneity between groups: p = 0.313
Bone state
bone health	9	0.14(0.00–0.28.00.28)	0%	Fixed	0.05
Osteopenia	5	0.19(0.00–0.38.00.38)	0%	Fixed	0.06
Osteoporosis	2	0.95(0.47–1.43)	0%	Fixed	0.0001
Heterogeneity between groups: p = 0.007
Type of nutritional supplement + Training
Calcium、vitamin D + Resistance training	4	0.11(−0.13−0.34)	0%	Fixed	0.877
Calcium、vitamin D + Aerobic training	2	0.02(−0.28−0.33)	0%	Fixed	0.377
Calcium、vitamin D + Combined training	3	0.17(−0.02−0.36)	0%	Fixed	0.073
Calcium、vitamin D + Whole body vibration	2	0.95(0.47–1.43)	0%	Fixed	0.0001
Protein + Aerobic training	2	0.41(−0.02−0.84)	0%	Fixed	0.063
Protein + Combined training	2	0.17(−0.11−0.45)	0%	Fixed	0.231
Vitamin C、E + Resistance training	1	0.54(−0.46−1.54)	NR	Fixed	0.293
Heterogeneity between groups: p = 0.047

N Number of studies; NR Not reported

Duration of training

Training duration of less than 1 year (SMD = 0.28, p = 0.006) and 1 year (SMD = 0.18, p = 0.011) significantly improved FN-BMD. Only one study included training exceeding 1 year, rendering this comparison non-referable. No statistically significant differences were observed between subgroups (p = 0.453).

Training intensity

Moderate-intensity (SMD = 0.40, p = 0.044) or high-intensity (SMD = 0.22, p = 0.015) training significantly increased FN-BMD, while moderate-to-high intensity (SMD = 0.10, p = 0.265) showed no statistically significant difference.

Frequency of training

Training frequency of 3 times per week (SMD = 0.20, p = 0.003) significantly increased FN-BMD, while 4–6 times per week (SMD = 0.13, p = 0.318) showed no statistically significant difference. Training less than 3 times per week could not be referenced as only one study was included. No statistically significant differences were observed between subgroups (p = 0.175).

Nutritional supplement types

Supplementation with calcium and vitamin D (SMD = 0.18, p = 0.005) and protein (SMD = 0.24, p = 0.043) significantly increased FN-BMD. Vitamin C + E supplementation could not be evaluated due to only one included study. No statistically significant differences were observed between subgroups (p = 0.73).

Training types

Aerobic training (SMD = 0.15, p = 0.228) and resistance training (SMD = 0.13, p = 0.271) showed no statistically significant differences in improving FN-BMD. Combined training (SMD = 0.17, p = 0.031), and whole-body vibration (SMD = 0.76, p = 0.0001) significantly increased FN-BMD. No statistically significant differences were observed between subgroups (p = 0.313).

Population

No statistically significant difference in FN-BMD improvement was observed between individuals with bone health (SMD = 0.14, p = 0.05) and osteopenia (SMD = 0.19, p = 0.06). However, individuals with osteoporosis (SMD = 0.95, p = 0.0001) demonstrated a significant increase in FN-BMD. Significant differences existed between subgroups (p = 0.007).

Combined training and nutrition strategies

Calcium, vitamin D + resistance training (SMD = 0.11, p = 0.877); calcium, vitamin D + aerobic training (SMD = 0.02, p = 0.377); calcium, vitamin D + combined training (SMD = 0.17, p = 0.073); Protein + aerobic training (SMD = 0.41, p = 0.063); Protein + combined training (SMD = 0.17, p = 0.231); showed no statistically significant differences in improving FN-BMD. Vitamin C and E supplementation plus resistance training included only one study and thus could not be referenced. Calcium and vitamin D plus whole-body vibration (SMD = 0.95, p = 0.0001) significantly increased FN-BMD. Differences between subgroups were significant (p = 0.047).

Effects of combined exercise and nutrition intervention compared to single intervention on lumbar spine bone mineral density

A total of 21 studies [30–46, 49–52] described the effects of EN compared with N and E on LS-BMD, including 17 studies [30–38, 40–42, 44, 46, 49–51] comparing EN vs. N and 8 studies [35, 38–41, 43, 45, 52] comparing EN vs. E. Low heterogeneity was observed among the included studies comparing EN vs. N (I² = 17%, p = 0.26), while moderate heterogeneity existed between the results of studies comparing EN vs. E (I² = 43%, p = 0.09). Using a random-effects model, the results indicated that EN significantly increased LS-BMD compared to single interventions (E, N) (SMD = 0.15, 95% CI: 0.04–0.27, p = 0.007). Compared with N, EN significantly increased LS-BMD (SMD = 0.20, 95% CI: 0.07–0.33, p = 0.003). Compared with E, EN increased LS-BMD (SMD = 0.05, 95% CI: −0.16−0.27, p = 0.628) but without statistical significance, as shown in Fig. 5. Egger’s test indicated no significant publication bias for EN vs. E (p = 0.922). However, significant publication bias was detected for EN vs. N (p = 0.002) and visually interpreted from the funnel plot.

Fig. 5.

Forest Plot of the Effect of Combined Exercise and Nutrition Intervention Compared to Single Intervention on Lumbar spine Bone Mineral Density. EN Exercise and Nutrition Group; N Single-nutrient group; E Single-exercise group EN vs. N trimming and imputation analysis indicated publication bias, with seven missing studies imputed. After adjusting for these studies, the pooled effect size weakened and became non-significant (SMD = 0.078, 95% CI: −0.024−0.180, p = 0.135), suggesting the initially observed effect may have been partially attributable to publication bias

LS-EN vs. N subgroup analysis

Subgroup analyses were not performed because there were no statistically significant differences and too few studies were included in EN vs. E. Subgroup analyses were performed only on the 17 studies included in EN vs. N (Table 4).

Table 4.

Effects of subgroup analysis on lumbar spine BMD

Subgroup	N	SMD 95%CI	I2	Model	P
Duration of training program
<1 year	8	0.35(0.03–0.66)	40.6%	Random	0.032
= 1 year	6	0.10(−0.05−0.24)	0%	Random	0.180
>1 year	3	0.23(−0.08−0.54)	36.9%	Random	0.203
Motion intensity
Moderate intensity	6	0.51(0.13–0.89)	41.3%	Random	0.008
Medium to high intensity	6	0.11(−0.05−0.27)	0%	Random	0.179
High-intensity	5	0.08(−0.11−0.27)	0%	Random	0.409
Frequency of training (per week)
2 days	1	0.76(0.14–1.39)	NR	Fixed	0.016
3 days	11	0.13(−0.01−0.27)	15.2%	Fixed	0.07
4–6 days	5	0.16(−0.03−0.36)	0%	Fixed	0.103
Heterogeneity between groups: p = 0.147
Type of nutritional supplement
Calcium、vitamin D	12	0.16(0.03–0.29)	39.7%	Fixed	0.016
Protein	4	0.15(−0.09−0.38)	0%	Fixed	0.216
Vitamin C + Vitamin E	1	0.54(−0.46−0.27)	NR	Fixed	0.293
Heterogeneity between groups: p = 0.757
Training type
Aerobic training	2	0.24(−0.19−0.67)	0%	Fixed	0.274
Resistance training	5	0.12(−0.13−0.38)	0%	Fixed	0.28
Combined training	8	0.10(−0.04−0.24)	0%	Fixed	0.171
Whole body vibration	2	0.76(0.14–1.38)	0%	Fixed	0.0001
Heterogeneity between groups: p = 0.019
Bone state
bone health	8	0.11(−0.04−0.26)	0%	Fixed	0.137
Osteopenia	6	0.09(−0.09−0.27)	0%	Fixed	0.367
Osteoporosis	3	0.92(0.50–1.34)	0%	Fixed	0.0001
Heterogeneity between groups: p = 0.001
Type of nutritional supplement་Training
Calcium、vitamin D + Resistance training	4	0.10(−0.13−0.34)	0%	Fixed	0.389
Calcium、vitamin D + Combined training	6	0.09(−0.07−0.25)	0%	Fixed	0.254
Calcium、vitamin D + Whole body vibration	2	0.95(0.47–1.43)	0%	Fixed	0.0001
Protein + Aerobic training	2	0.24(−0.19−0.67)	0%	Fixed	0.274
Protein + Combined training	2	0.11(−0.17−0.39)	0%	Fixed	0.445
Vitamin C、E + Resistance training	1	0.54(−0.46−1.54)	0%	Fixed	0.293
Heterogeneity between groups: p = 0.033

N Number of studies; NR Not reported

Duration of training

Duration below 1 year (SMD = 0.35, p = 0.032) significantly increased LS-BMD, and there was no statistically significant difference in duration of training 1 year (SMD = 0.10, p = 0.18) and above 1 year (SMD = 0.23, p = 0.203).

Training intensity

Moderate-intensity (SMD = 0.51, p = 0.008) training significantly increased LS-BMD, and there was no statistically significant difference between moderate-high intensity (SMD = 0.11, p = 0.179) and high-intensity (SMD = 0.08, p = 0.409).

Training frequency

There was no statistical difference in the number of training sessions per week for 3 sessions per week (SMD = 0.13, p = 0.07), 4–6 sessions per week (SMD = 0.16, p = 0.103). Less than 3 times per week was only included in 1 case study therefore could not be referenced. There was no statistical difference between subgroups (p = 0.147).

Types of nutritional supplements

Supplementation with calcium and vitamin D supplements (SMD = 0.16, p = 0.016) significantly elevated LS-BMD, there was no statistically significant difference in protein (SMD = 0.15, p = 0.216), and supplementation with vitamins C + E was only included in 1 case study and therefore could not be referenced. There was no statistical difference between subgroups (p = 0.757).

Training types

There was no statistically significant difference between performing aerobic (SMD = 0.24,p = 0.274), resistance (SMD = 0.12, p = 0.28) and mixed (SMD = 0.10, p = 0.171) training to improve LS-BMD, and whole-body vibration (SMD = 0.76, p = 0.0001) to significantly improve LS-BMD. the difference was significant between subgroups (p = 0.019).

Population

There was no statistically significant difference between the bone healthy (SMD = 0.11, p = 0.137) and low bone mass (SMD = 0.09, p = 0.367) populations in improving LS-BMD, and the osteoporotic (SMD = 0.92, p = 0.0001) population significantly improved LS-BMD. the difference was significant between subgroups (p = 0.007).

Combined training and nutrition strategies

Calcium, vitamin D + resistance training (SMD = 0.10, p = 0.389); calcium, vitamin D + aerobic training (SMD = 0.09, p = 0.254); calcium, vitamin D + mixed training (SMD = 0.17, p = 0.073); protein + aerobic training (SMD = 0.24, p = 0.274); protein + mixed training (SMD = 0.11, p = 0.445); none of which had a statistically significant difference in LS-BMD improvement. Vitamin C and E supplementation + resistance training was only included in 1 case study therefore cannot be referenced. Calcium, vitamin D + whole body vibration (SMD = 0.95, p = 0.0001) significantly improved LS-BMD. significant difference between subgroups (p = 0.033).

Effects of combined exercise and nutrition intervention compared to single intervention on total hip bone mineral density

A total of 15 studies [29, 30, 32, 34, 35, 37, 39, 40, 42, 43, 45, 46, 50–52] described the effect of EN compared to N and E on TH-BMD, with EN vs. N included in 11 studies [29, 30, 32, 34, 35, 37, 40, 42, 46, 50, 51] and EN vs. E included in 7 studies [35, 39, 40, 42, 43, 45, 52]; low heterogeneity was observed between the results of the EN vs. N inclusion studies (I² = 0%, p = 0.667), whereas moderate heterogeneity was observed between the results of the EN vs. E inclusion studies (I² = 55.9%, p = 0.034), using a random effects model, which showed that EN improved TH-BMD (SMD = 0.11, 95% CI: −0.01−0.23, p = 0.067) compared to a single intervention (E, N), but there was no significant difference. EN (SMD = 0.16, 95% CI: 0.03–0.28, p = 0.014) significantly increased TH-BMD compared to N. EN (SMD = 0.04, 95% CI: −0.2−0.28, p = 0.752) increased TH-BMD compared to E, but there was no statistically significant difference, as shown in Fig. 6. After an Egger ‘s test (EN vs. N: p = 0.428; EN vs. E: p = 0.667) and the visual interpretation of the funnel plot had no significant publication bias.

Fig. 6.

Forest Plot of the Effect of Combined training and Nutrition Intervention Compared to Single Intervention on Total hip Bone Mineral Density. EN Exercise and Nutrition Group; N Single-nutrient group; E Single-exercise group

TH-EN vs. N subgroup analysis

Subgroup analyses were not performed because there were no statistically significant differences and too few studies were included in EN vs. E. Subgroup analyses were performed only on the 11 studies included in EN vs. N (Table 5).

Table 5.

Effects of subgroup analysis on total hip BMD

Subgroup	N	SMD 95%CI	I2	Model	P
Duration of training program
<1 year	3	0.53(−0.16−1.22)	59.8%	Fixed	0.129
= 1 year	5	0.16(0.01–0.31)	30%	Fixed	0.094
>1 year	3	0.08(−0.18−0.34)	0%	Fixed	0.54
Heterogeneity between groups: p = 0.702
Motion intensity
Moderate intensity	4	0.42(−0.08−0.92)	34.6%	Random	0.101
Medium to high intensity	4	0.09(−0.11−0.29)	0%	Random	0.37
High-intensity	3	0.18(−0.12−0.47)	48.5%	Random	0.249
Frequency of training (per week)
3 days	7	0.29(0.06–0.53)	43.1%	Random	0.015
4–6 days	4	0.05(−0.15−0.25)	0%	Random	0.634
Type of nutritional supplement
Including Calcium、vitamin D	8	0.20(0.01–0.38)	30.3%	Random	0.039
Protein	3	0.20(−0.18−0.57)	50.8%	Random	0.300
Training type
Aerobic training	2	0.24(−0.09−0.56)	34.6%	Fixed	0.149
Resistance training	1	−0.07(−1.08−0.95)	NR	Fixed	0.897
Combined training	7	0.13(−0.01−0.27)	0%	Fixed	0.071
Whole body vibration	1	1.22(0.47 − 0.30)	NR	Fixed	0.001
Heterogeneity between groups: p = 0.043
Bone state
Bone health	5	0.09(−0.09−0.27)	7.6%	Fixed	0.324
Osteopenia	5	0.21(0.02–0.40)	0%	Fixed	0.03
Osteoporosis	1	1.22(0.47–1.97)	0%	Fixed	0.001
Heterogeneity between groups: p = 0.014
Type of nutritional supplement + Training
Calcium、vitamin D + Resistance training	1	−0.07(−1.08−0.95)	NR	Fixed	0.897
Calcium、vitamin D + Aerobic training	1	0.06(−0.37−0.49)	NR	Fixed	0.774
Calcium、vitamin D + Combined training	5	0.17(0.00–0.33.00.33)	0%	Fixed	0.045
Calcium、vitamin D + Whole body vibration	1	1.22(0.47–1.97)	NR	Fixed	0.001
Protein + Aerobic training	1	0.48(−0.02−0.97)	NR	Fixed	0.06
Protein + Combined training	2	0.02(−0.26−0.30)	38.6%	Fixed	0.876
Heterogeneity between groups: p = 0.063

N Number of studies; NR Not reported

Duration of training

There was no statistical difference between durations below 1 year (SMD = 0.53, p = 0.129), 1 year of training (SMD = 0.16, p = 0.094) and above 1 year (SMD = 0.08, p = 0.54) in improving TH-BMD. There was no statistical difference between subgroups (p = 0.453).

Training intensity

Moderate-intensity (SMD = 0.42, p = 0.101), moderate-high intensity (SMD = 0.09, p = 0.37), and high-intensity (SMD = 0.18, p = 0.249) training to improve TH-BMD were not statistically different.

Training frequency

Training 3 times per week (SMD = 0.29, p = 0.015) significantly increased TH-BMD, while training four to 6 times per week (SMD = 0.05, p = 0.634) showed no statistically significant difference.

Types of nutritional supplements

Calcium and vitamin D supplementation (SMD = 0.20, p = 0.039) significantly increased TH-BMD, while protein supplementation (SMD = 0.20, p = 0.30) showed no statistically significant difference.

Training types

There was no significant statistical difference in improving TH-BMD between aerobic training (SMD = 0.24, p = 0.149) and combined training (SMD = 0.17, p = 0.031). Resistance training and whole-body vibration could not be compared as only one study each was included. Differences between subgroups were significant (p = 0.043).

Population

There was no statistically significant difference in TH-BMD improvement among individuals with bone health (SMD = 0.09, p = 0.324). Individuals with low bone mass (SMD = 0.21, p = 0.03) demonstrated a significant increase in TH-BMD. Only 1 study included individuals with osteoporosis, rendering this group non-referable. Differences between subgroups were significant (p = 0.014).

Combined training and nutrition strategies

Calcium and vitamin D plus combined training (SMD = 0.17, p = 0.045) significantly increased total hip bone mineral density (TH-BMD); protein plus combined training (SMD = 0.02, p = 0.876) showed no statistically significant difference in TH-BMD improvement. Other combination strategies included only one study each and thus could not be referenced. No statistically significant differences were observed between subgroups (p = 0.063).

Effects of combined exercise and nutrition intervention compared to single intervention on whole body bone mineral density

A total of 10 studies [30, 33, 35, 36, 39, 40, 42, 44, 45, 52] described the effects of EN compared with N and E on WB-BMD, including 7 studies [30, 33, 35, 36, 40, 42, 44] comparing EN vs. N and 6 studies [35, 39, 40, 42, 45, 52] comparing EN vs. E. Low heterogeneity was observed between studies comparing EN vs. N (I² = 0%, p = 0.858), whereas moderate heterogeneity existed between studies comparing EN vs. E (I² = 56.2%, p = 0.044). Therefore, a random-effects model was applied. This model indicated that EN increased WB-BMD compared to single interventions (E, N) (SMD = 0.11, 95% CI: −0.03−0.25, p = 0.115), though the difference was not statistically significant. Compared with N, EN increased WB-BMD (SMD = 0.17, 95% CI: 0.00–0.34.00.34, p = 0.05) but without statistical significance. Compared with E, EN increased WB-BMD (SMD = 0.01, 95% CI: −0.25 to 0.28, p = 0.93) but showed no statistical difference (Fig. 7). Egger’s test (EN vs. N: p = 0.149; EN vs. E: p = 0.376) and visual interpretation of funnel plots indicated no significant publication bias.

Fig. 7.

Forest Plot of the Effect of Combined training and Nutrition Intervention Compared to Single Intervention on Whole body Bone Mineral Density. EN Exercise and Nutrition Group; N Single-nutrient group; E Single-exercise group

Effects of combined exercise and nutrition intervention compared to single intervention on ward’s triangle bone mineral density

A total of 6 studies [30, 35, 40, 42, 49, 52] described the effects of EN compared with N and E on BMD in the Ward’s triangle region, including 5 studies [30, 35, 40, 42, 49] comparing EN vs. N and 4 studies [35, 40, 42, 52] comparing EN vs. E. Low heterogeneity was observed between the results of studies comparing EN vs. N (I² = 3%, p = 0.377), while moderate heterogeneity was present between studies comparing EN vs. E (I² = 62.5%, p = 0.046). Therefore, a random-effects model was used, which showed that EN significantly increased BMD in the Ward’s triangle region compared to single interventions (E, N) (SMD = 0.19, 95% CI: 0.01–0.37, p = 0.037). Compared with N, EN increased BMD at the Ward’s triangle region (SMD = 0.13, 95% CI: −0.09 to 0.35, p = 0.235), but the difference was not statistically significant. Compared with E, EN (SMD = 0.26, 95% CI: −0.06 to 0.58, p = 0.111) increased BMD in the Ward’s triangle region but showed no statistical difference (Fig. 8). Egger’s test (EN vs. N: p = 0.316; EN vs. E: p = 0.197) and visual interpretation of funnel plots revealed no significant publication bias.

Fig. 8.

Forest Plot of the Effect of Combined training and Nutrition Intervention Compared to Single Intervention on Ward’s triangle Bone Mineral Density. EN Exercise and Nutrition Group; N Single-nutrient group; E Single-exercise group

Sensitivity analysis

Sensitivity analysis confirmed the robustness of the effect size for combined exercise and nutritional interventions on BMD at the femoral neck, lumbar spine, total hip, whole body, and Ward’s triangle region. The pooled effect size (95% CI) remained stable and showed no significant changes after sequentially excluding individual studies (S3).

Meta-regression analysis

Meta-regression analysis revealed that the efficacy of combined interventions was significantly associated with exercise frequency, duration, type, intensity, population, supplement type, and exercise-supplement combination strategies (all p < 0.05). These factors collectively constitute important predictors of combined intervention efficacy and partially explain the heterogeneity observed across included studies (S4).

Abbreviations

BMD

Bone mineral density

OP

Osteoporosis

IGF-1

Insulin-like growth factor-1

PRISMA

Preferred reporting items for systematic reviews and meta-analysis

SMD

Standardized mean difference

DXA

Dual energy X-ray absorptiometry

BTMs

Bone turnover marker

BMC

Bone mineral content

SD

Standard deviation

95%CI

95% confidence interval

RCTs

randomized controlled trials

FN

Femoral neck

LS

Lumbar spine

TH

Total hip

WB

Whole body

WT

Ward’s triangle area

MF

Main form of exercise

EF

Exercise frequency

EI

Exercise intensity

RPE

Borg rate of perceived exertion scale

1RM

1-Repetition maximum

HRmax

Maximum heart rate

VO₂peak

Peak oxygen uptake

Hz

Hertz

Author contributions

X.L. Study design. Data collection. Data interpretation. Manuscript review. J.L. Study design. Data collection. Data interpretation. Manuscript writing. Manuscript review. W.H.Z. Data collection.Data analysis. T.T.M. Data interpretation. Manuscript review. X.Q.W. Study design. Manuscript review. All authors read and approved the final Manuscript.

Funding

The authors declare that the research, authorship, and/or publication of this article were financially supported. This study was supported by the Social Science Planning Research Program of Shandong Province in 2021 (21 DTYJ 03).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.