8,456 peer-reviewed studies, each with a short note written from its published abstract: what kind of study it was, who took part, what was measured and what the authors reported. Observational studies show associations, not proof.
2014 / Meta-analysis of 6 cohort and 8 case-control studies
A meta-analysis of 14 studies (195,992 individuals, 9,958 hip fracture cases) assessed coffee and tea intake in relation to hip fracture risk. Pooled RRs for the highest versus lowest consumption were 0.94 (95% CI 0.71-1.17) for coffee and 0.84 (95% CI 0.66-1.02) for tea. A nonlinear association was found for tea (p < 0.01): compared with no tea, RRs were 0.72 (95% CI 0.56-0.88) for 1-2 cups/day, 0.63 (95% CI 0.32-0.94) for 2-3 cups/day, and 0.79 (95% CI 0.62-0.96) for 3-4 cups/day.
Shows an association, not proof of cause and effect.
Sheng J, Qu X, Zhang X, et al. Osteoporos Int, 2014. PMID 24196722
2014 / Systematic review and meta-analysis
A meta-analysis of 9 cohort and 6 case-control studies involving 253,514 participants and 12,939 fracture cases examined coffee intake and fracture risk. Highest coffee consumption was linked to greater fracture risk in women (RR 1.14; 95% CI 1.05 to 1.24) but reduced risk in men (RR 0.76; 95% CI 0.62 to 0.94). Daily intake of 2 and 8 cups in women had pooled RRs of 1.02 and 1.54, respectively.
Lee DR, Lee J, Rota M, et al. Bone, 2014. PMID 24576685
2014 / Dose-response meta-analysis of observational studies
A meta-analysis of 9 observational studies including 147,950 individuals evaluated the association between tea drinking and fracture risk. The pooled odds ratio for tea consumption and fracture risk was 0.89 (95% CI, 0.78-1.04). Subgroup analyses showed no significant association in cohort studies (n=3; OR, 0.97; 95% CI, 0.89-1.06) or case-control studies (n=6; OR, 0.91; 95% CI, 0.70-1.19), and no significant association was detected in the dose-response analysis.
Shows an association, not proof of cause and effect.
Chen B, Shi HF, Wu SC Diagn Pathol, 2014. PMID 24588938
2014 / Umbrella review of systematic reviews and meta-analyses
An umbrella review examined 107 systematic reviews, 74 observational meta-analyses, and 87 randomized trial meta-analyses covering 137 outcomes. Concordant observational and intervention findings emerged only for birth weight. Probable links were observed for maternal levels, birth weight, childhood dental caries, and parathyroid hormone in dialysis patients, but evidence failed to demonstrate that solitary vitamin D supplementation reduced fractures or falls in older adults.
Theodoratou E, Tzoulaki I, Zgaga L, et al. BMJ, 2014. PMID 24690624
2014 / Stratified meta-analysis of randomized placebo-controlled trials
A systematic review and stratified meta-analysis of 25 randomized placebo-controlled trials with 3,458 patients assessed glucosamine for osteoarthritis pain. Overall, glucosamine moderately reduced pain (SMD -0.51 [95% CI -0.72, -0.30]) with high inconsistency (I2 = 88%). Brand explained 41% of heterogeneity (P = 0.00032): 12 trials using Rottapharm/Madaus showed significant pain reduction (SMD -1.07 [95% CI -1.47, -0.67]), but a sensitivity analysis of 3 low risk-of-bias trials of this brand showed only a small effect (SMD -0.27 [95% CI -0.43, -0.12]). Thirteen trials using non-Rottapharm/Madaus products failed to show pain reduction (SMD -0.11 [95% CI -0.46, 0.24]).
Eriksen P, Bartels EM, Altman RD, et al. Arthritis Care Res (Hoboken), 2014. PMID 24905534
2013 / Systematic review and dose-response meta-analyses
A systematic review and dose-response meta-analyses examined vitamin B12, folate, and homocysteine status in relation to fractures and BMD across 27 observational studies and one RCT. Meta-analysis of four prospective studies (n = 7,475) linked each 50 pmol/L increase in vitamin B12 to a borderline significant 4% decrease in fracture risk (RR = 0.96, 95% CI = 0.92 to 1.00). Meta-analysis of eight studies (n = 11,511) linked each μmol/L increase in homocysteine to a 4% increase in fracture risk (RR = 1.04, 95% CI = 1.02 to 1.07). Too few studies existed to draw conclusions for folate and fracture risk, and meta-analyses in female populations found no associations between vitamin B12, folate, or homocysteine levels and BMD.
Shows an association, not proof of cause and effect.
van Wijngaarden JP, Doets EL, Szczecińska A, et al. J Nutr Metab, 2013. PMID 23509616
2013 / Meta-analysis of double-blind, placebo-controlled RCTs
A meta-analysis of 25 double-blind, placebo-controlled RCTs evaluated oral DHEA supplementation in 1,353 elderly men with a mean follow-up of 36 weeks. DHEA supplementation was associated with a reduction in fat mass (standardized mean difference -0.35 [-0.65 to -0.05]; P = .02), though this association disappeared after adjusting for increases in total testosterone and estradiol. No effects compared to placebo were observed for lipid and glycemic metabolism, bone health, sexual function, or quality of life.
Corona G, Rastrelli G, Giagulli VA, et al. J Clin Endocrinol Metab, 2013. PMID 23824417
2013 / Systematic review of prospective and retrospective clinical studies
Authors systematically reviewed prospective and retrospective clinical studies from five electronic databases examining the systemic impact of smoking on bone healing in non-spinal orthopaedics. Seventeen studies were evaluated, including nine tibia studies and eight other orthopaedic studies. Of these 17 studies, 13 concluded that smoking negatively influenced bone healing, leading the authors to conclude smoking impacts healing through delayed union, nonunion, and increased complications.
Shows an association, not proof of cause and effect.
Patel RA, Wilson RF, Patel PA, et al. Bone Joint Res, 2013. PMID 23836474
2013 / Systematic review and meta-analysis of controlled trials
Ten randomized and nonrandomized controlled trials evaluated walking interventions on bone mineral density (BMD) in perimenopausal and postmenopausal women. Meta-analysis showed no significant effect on lumbar spine BMD (WMD 0.01 g/cm², 95% CI -0.00 to 0.02; P = 0.05), the radius (WMD -0.01 g/cm, 95% CI -0.06 to 0.04; P = 0.71), or the whole body (WMD 0.04 g/cm², 95% CI -0.00 to 0.08; P = 0.06). While femoral neck BMD showed no significant overall effect across all trials (WMD 0.01 g/cm², 95% CI -0.00 to 0.01; P = 0.07), positive effects were evident with interventions lasting more than 6 months (6 months to 1-2 years).
Ma D, Wu L, He Z Menopause, 2013. PMID 24149921
2012 / Meta-analysis of randomized controlled trials
A meta-analysis of randomized controlled trials examined the effects of soy isoflavones on bone health in women. Isoflavone supplementation significantly increased bone mineral density by 54% and lowered the bone resorption marker urinary deoxypyridinoline by 23% compared to baseline. Effects were greater among postmenopausal women and with daily intakes exceeding 75 mg, whereas no significant changes occurred for serum bone alkaline phosphatase.
Wei P, Liu M, Chen Y, et al. Asian Pac J Trop Med, 2012. PMID 22305793
2012 / Systematic review and meta-analysis of RCTs and quasi-RCTs
A systematic review identified 8 studies evaluating collagen derivatives (six on collagen hydrolysate [CH], two on gelatine, and one on undenatured collagen [UC]) in osteoarthritis patients. The pooled mean difference across three studies for WOMAC pain reduction comparing CH with placebo was -0.49 (95% CI -1.10 to 0.12), with no significant differences for disability. Gelatine was superior to placebo and alternative therapies for pain, whereas UC showed no significant differences compared to glucosamine plus chondroitin. Adverse events were mostly mild to moderate gastrointestinal complaints, and overall evidence quality was moderate to very low.
Van Vijven JP, Luijsterburg PA, Verhagen AP, et al. Osteoarthritis Cartilage, 2012. PMID 22521757
2011 / Meta-analysis of prospective cohorts
A meta-analysis of prospective cohort studies examined daily milk consumption and hip fracture risk. In women, daily milk intake was not associated with hip fracture risk (pooled RR 0.99 per daily glass, 95% CI 0.96-1.02 across 6 studies with 195,102 participants). In men, the pooled RR per daily glass was 0.91 (95% CI 0.81-1.01 across 3 studies with 75,149 participants).
Shows an association, not proof of cause and effect.
Bischoff-Ferrari HA, Dawson-Hughes B, Baron JA, et al. J Bone Miner Res, 2011. PMID 20949604
2010 / Network meta-analysis of randomized controlled trials
A network meta-analysis of 10 large-scale randomized controlled trials evaluated the effects of glucosamine, chondroitin, or their combination in 3,803 patients with hip or knee osteoarthritis. Compared with placebo on a 10 cm visual analogue scale, overall differences in pain intensity were -0.4 cm (95% credible interval -0.7 to -0.1 cm) for glucosamine, -0.3 cm (-0.7 to 0.0 cm) for chondroitin, and -0.5 cm (-0.9 to 0.0 cm) for the combination. None of the 95% credible intervals crossed the prespecified minimal clinically important difference of -0.9 cm. Differences in changes in minimal width of joint space were minute with credible intervals overlapping zero, and industry-independent trials showed smaller effects than commercially funded trials (P=0.02 for interaction).
Wandel S, Jüni P, Tendal B, et al. BMJ, 2010. PMID 20847017
2009 / Meta-analysis
A meta-analysis analyzed 9 studies comprising 2,749 subjects, including 1,880 women and 869 men. Bone mineral density was approximately 4% lower in vegetarians compared with omnivores (95% CI: 2%, 7%) at both the femoral neck and lumbar spine. Vegans showed a 6% lower lumbar spine density (95% CI: 2%, 9%), while lactoovovegetarians showed a 2% reduction (95% CI: 1%, 4%), although the magnitude was deemed clinically insignificant.
Ho-Pham LT, Nguyen ND, Nguyen TV Am J Clin Nutr, 2009. PMID 19571226
2009 / Systematic review and meta-analysis
In cross-sectional studies, protein intake was positively associated with bone mineral density, explaining 1% to 2% of variance. A meta-analysis of randomized placebo-controlled trials showed protein supplementation significantly benefited lumbar spine bone mineral density, although neither soy nor milk basic protein exerted specific effects. Cohort findings indicated protein was not associated with hip fracture risk, showing overall effect sizes that remained modest.
Darling AL, Millward DJ, Torgerson DJ, et al. Am J Clin Nutr, 2009. PMID 19889822
2007 / Meta-analysis of randomized or quasi-randomized controlled trials
A meta-analysis of 20 trials evaluated the effect of chondroitin compared with placebo or no treatment on pain in patients with osteoarthritis of the knee or hip. Overall trials displayed high heterogeneity (I2 = 92%), with larger benefits seen in smaller trials and those with poorer methodology. When restricted to the 3 large trials using intention-to-treat analysis (40% of patients), the effect size was -0.03 (95% CI, -0.13 to 0.07; I2 = 0%), corresponding to a 0.6 mm difference on a 10-cm visual analogue scale. The pooled relative risk for any adverse event across 12 trials was 0.99 (95% CI, 0.76 to 1.31). Limitations noted include generally low trial quality and the need to approximate effect sizes for 9 trials.
Reichenbach S, Sterchi R, Scherer M, et al. Ann Intern Med, 2007. PMID 17438317
2007 / Meta-analysis of randomized, double-blind, placebo-controlled trials
A meta-analysis evaluated 15 randomized, double-blind, placebo-controlled trials of glucosamine for knee or hip osteoarthritis pain to identify factors explaining trial heterogeneity. The overall summary effect size was 0.35 (95% CI 0.14, 0.56) with substantial heterogeneity (I² = 0.80). Summary effect sizes ranged from 0.05 to 0.16 in trials without industry involvement versus 0.47 to 0.55 in trials with industry involvement. By preparation, the effect size was 0.06 for glucosamine hydrochloride and 0.44 for glucosamine sulfate (0.55 for Rottapharm products vs. 0.11 for others). Potential explanations for heterogeneity include different glucosamine preparations, inadequate allocation concealment, and industry bias.
Vlad SC, LaValley MP, McAlindon TE, et al. Arthritis Rheum, 2007. PMID 17599746
2007 / Meta-analysis of randomized controlled trials
A meta-analysis analyzed 29 randomized controlled trials in adults aged 50 years and older. Calcium with or without vitamin D reduced all-type fractures by 12% (risk ratio 0.88, 95% CI 0.83-0.95; 17 trials, n=52,625). Bone loss was also reduced by 0.54% at the hip and 1.19% in the spine (23 trials, n=41,419).
Tang BM, Eslick GD, Nowson C, et al. Lancet, 2007. PMID 17720017
2007 / Meta-analysis of prospective cohorts and randomized controlled trials
Cohort analyses of 170,991 women (pooled RR per 300 mg/d: 1.01; 95% CI: 0.97, 1.05) and 68,606 men (pooled RR per 300 mg/d: 0.92; 95% CI: 0.82, 1.03) showed calcium intake was not significantly associated with hip fracture risk. In 4 clinical trials with 6,504 subjects, the pooled RR for hip fracture between calcium and placebo was 1.64 (95% CI: 1.02, 2.64), leading authors to conclude that trials show no reduction in hip fracture risk and an increased risk is possible. For nonvertebral fractures across 5 trials, calcium supplementation had a neutral effect (RR 0.92; 95% CI: 0.81, 1.05).
Bischoff-Ferrari HA, Dawson-Hughes B, Baron JA, et al. Am J Clin Nutr, 2007. PMID 18065599
2005 / Systematic review and meta-analysis of RCTs
A systematic review and meta-analysis of 20 RCTs evaluated the efficacy and safety of glucosamine in 2,570 osteoarthritis patients. Overall, glucosamine outperformed placebo with a 28% improvement in pain (SMD -0.61, 95% CI -0.95 to -0.28) and a 21% improvement in Lequesne function (SMD -0.51, 95% CI -0.96 to -0.05), but WOMAC pain, function, and stiffness did not reach statistical significance. Analysis restricted to eight studies with adequate allocation concealment failed to show benefit for pain or WOMAC function. Benefits were observed in trials using the Rotta preparation, whereas non-Rotta preparations did not achieve statistically significant improvements in pain or WOMAC function. Glucosamine was as safe as placebo regarding adverse reactions (RR 0.97, 95% CI 0.88 to 1.08).
Towheed TE, Maxwell L, Anastassiades TP, et al. Cochrane Database Syst Rev, 2005. PMID 15846645
2024 / 1-year observational extension of two randomized controlled trials
Following two 16-week randomized trials of the Plants for Joints (PFJ) lifestyle intervention, participants with rheumatoid arthritis (RA) or metabolic syndrome-associated osteoarthritis (MSOA) were followed for 1 year after all received the intervention. Improvements were sustained in the RA group one year post-intervention (mean DAS28 -0.9 points; p<0.001), and WOMAC scores in the MSOA group remained well below baseline (-7.8 points, p<0.001). Sustained improvements were also observed for CRP and waist circumference in both groups, LDL-cholesterol in RA, and weight, HbA1c, and blood pressure in MSOA, alongside net medication decreases.
Shows an association, not proof of cause and effect.
Wagenaar CA, Walrabenstein W, van der Leeden M, et al. RMD Open, 2024. PMID 38413171
2021 / Cross-sectional study
A cross-sectional study compared 63 vegetarian, 52 vegan, and 72 matched omnivore Polish children aged 5-10 years. Compared with omnivores, vegans had lower fat indices, lower LDL (-24 mg/dL), lower HDL (-12.2 mg/dL), and lower high-sensitivity C-reactive protein, but were shorter and had lower bone mineral content (total body minus the head: -3.7%; lumbar spine: -5.6%). Vegans also exhibited lower iron status, higher homocysteine, and higher rates of vitamin B-12 deficiency, iron-deficiency anemia, and low ferritin, though supplementation resolved low B-12 and 25(OH)D concentrations. Vegetarians had lower gluteofemoral adiposity, total cholesterol, and HDL, but higher glucose, VLDL, and triglycerides relative to omnivores.
Shows an association, not proof of cause and effect.
Desmond MA, Sobiecki JG, Jaworski M, et al. Am J Clin Nutr, 2021. PMID 33740036
2020 / Cross-sectional comparative study
Using NHANES (2007-2008 and 2009-2010) data, researchers assessed bone mineral density (BMD) differences between adult vegetarians and nonvegetarians using linear regression models. Initially, vegetarians had significantly lower BMD across sites (P < 0.001), but after adjusting for anthropometrics (BMI and waist circumference), differences became nonsignificant for femoral neck (0.77 vs 0.79 g/cm2, P = 0.10) and total femoral BMD (0.88 vs 0.90 g/cm2, P = 0.12). A small, statistically significant difference remained for total lumbar spine BMD (1.01 vs 1.04 g/cm2, P = 0.005).
Shows an association, not proof of cause and effect.
Karavasiloglou N, Selinger E, Gojda J, et al. J Nutr, 2020. PMID 32055831
2020 / Cross-sectional study
In a cross-sectional study of 73 healthy Australian adults, dietary patterns derived by principal component analysis were evaluated against dual-energy X-ray absorptiometry measures. After adjusting for age, gender, physical activity, and BMI, a 'Plant Based' component correlated positively with BMD (p=0.054, R2=0.439) and T-score (p=0.053, R2=0.221). The 'Meat Based' component showed no association alone, but correlated positively with BMD (p=0.046, R2=0.474) and T-score (p=0.046, R2=0.279) when the Plant Based component was included in the model. No significant associations with BMD or T-scores were observed for a 'Junk Food' component.
Shows an association, not proof of cause and effect.
Berg J, Seyedsadjadi N, Grant R J Nutr Health Aging, 2020. PMID 32242206