Determining The Prevalence of Low Bone Health Conditions and Related Factors Using Multisite Quantitative Ultrasound Technique in Young Indian Women: An Epidemiological Community-Based Study

Study settings, design, and sampling

The study was conducted at Guru Nanak Dev University, Amritsar, India. The current study design was population-based and cross-sectional in nature. The study started on October 15^th, 2015, and was completed on July 17^th, 2025. A cluster randomized sampling method was adopted, using the probabilities proportional to size (PPS) criteria to collect samples. The studied sample population comprised both rural and urban women of different socio-economic categories, belonging to the age group of 20–40 years.

Ethical approval, consent, and registration

The study was approved by the Institutional Ethics Committee, Faculty of Sports Medicine and Physiotherapy, Guru Nanak Dev University, Amritsar, India (No-1319/SMP). The protocol was carried out in accordance with the relevant guidelines and regulations from the “Declaration of Helsinki”. A written informed consent explaining the purpose and procedure of the study, translated both in Hindi and Punjabi, was obtained from each participant. The consent for the study was drafted as per the Indian Council of Medical Research (ICMR) guidelines. This study was registered in the Clinical Trial Registry of India (CTRI/2015/10/006240) on October 6^th, 2015.

Sample size

The sample size was calculated to estimate the prevalence of low bone mineral density (BMD) in young Indian women. Based on previous Indian studies indicating a prevalence of approximately 10%, and using a 95% confidence level with a relative precision of ±25%, the initial required sample was 540. Considering that data were collected across nearly 50 clusters, a design effect of 2.0 was applied to account for potential intra-cluster correlation, increasing the target to 1,080 participants. Ultimately, 1,176 women were recruited, exceeding the minimum requirement. This extended recruitment ensured adequate representation across age groups and geographic clusters, enhanced the precision of prevalence estimates, and strengthened the statistical power and generalizability of the findings.

Inclusion, exclusion criteria

Physically able women aged 20–40 years were included in the study. Pregnant and lactating women were excluded from the study. Women with a minimum gap of 6–9 months from the last delivery and lactation period were part of the study. Those who consented to participate were enrolled in the study.

Questionnaires and interview procedure

The study methodology was precisely carried out as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. Physical activity, a 24-hour dietary and self-designed questionnaire encompassing the anthropometric and socio-economic strata were collected. The self-designed questionnaire contained information related to factors that affect bone health.¹¹ The questionnaire contained medical histories which included the questions on whether the participant was currently experiencing the following conditions: hyperparathyroidism, hypothyroidism, renal disease, diabetes mellitus, rheumatoid arthritis, eating disorders, medications that elevate bone density, history of back pain or not, and the question about whether the participant is or was undergoing glucocorticoids and/or thyroid hormone treatment. Another area explored by the interviewer was the obstetric and gynecological history of the participant, which included age at menarche and menopause (assuming it occurred), incidence of amenorrhea, number of offspring and pregnancies, length of breastfeeding, and whether the participant had undergone any surgical events or procedures, including oophorectomy or hysterectomy. There was also an investigation into the use of oral contraceptive pills and/or hormone replacement therapy. The next five questionnaires were used to collect information on socioeconomic status, family occupation, living expenditure, lifelong occupation, and educational attainment. Data on recreational activities and weight-bearing exercises were documented using a validated and translated Punjabi International Physical Activity Questionnaire (IPAQ). Perceived stress was measured using a local translation of a validated and translated Cohen Perceived Stress Scale (PSS-10) using the Punjabi local language. Each respondent was asked to give the time they spent under the sun every day, and a sun-exposure index was calculated according to the approach of Barger-Lux and Heaney (2002).³ The respondents also indicated their eating patterns through the 24-hour dietary recall technique.¹¹

Outcome variable and measures: bone health quantification & quantitative ultrasound (qus) measurement protocol

Bone health was assessed using a portable multisite quantitative ultrasound (QUS) device, the Sunlight Omnisense™ 7000S (BeamMed Ltd., Israel). Speed of Sound (SOS) measurements were taken at two peripheral skeletal sites, the distal one-third radius of the non-dominant arm and the mid-shaft tibia, after procedures to match those described by Shenoy et al. (2017).¹¹ To ensure accuracy, measurements reproducible within a range of −2% of the expected calibration range were accepted daily to calibrate the device. The participants were sitting in a comfortable position as the palpation of the anatomical landmarks was carried out manually. Three successive readings were taken at each site, and the average of the three readings was taken to obtain the value of this mean or the average of the three values as near to the true value as possible. This multisite QUS protocol presented a viable and sound method of characterizing bone qualities in young adult women.¹¹

Operator training and reliability

Two physiotherapists, who had received structured training in the presence of an experienced supervisor, were used to carry out all QUS measurements. Their competence was first put to a pilot sample of 30 people. Inter-operator reliability was high, and the coefficient of variation was 1.8% at the radius and 2.1% at the tibia. Only those operators who passed a training stage and received more than 95 percent agreement in the training phase could carry out measurements during data collection to ensure consistency and accuracy.

Z-Score classification

The device automatically produced Z-scores based on sex- and age-specific reference data based on several populations of Asians. In line with the recommendations of the International Society of Clinical Densitometry (ISCD), a Z-score less than −2.0 was considered as low bone mineral density, whereas a Z-score below −3.0 was regarded as very low bone mineral density. The reason behind the choice of these cut-offs is to enhance the sensitivity of screening in women of the target age group of the present study, i.e., between 20 and 40 years.

Data management and quality control

Data was collected with a pre-established quality-control protocol that included daily checks of calibration, monitoring of signal quality indicators, and verification of the correct positioning of the anatomy during each scan. The entire data was entered in a two-entry procedure to reduce transcription errors. An automated check was used to find outliers, incomplete entries, or inconsistent responses. Records where anthropometric, QUS, or questionnaire data were missing were categorized as missing and were not included in statistical analyses.

Statistical analysis

The descriptive statistics in the study have been evaluated using the exact binomial method. For the multinomial logistic regression analysis, bone mineral density (BMD) was divided into three categories based on z-score values: (1) normal BMD (z-score ≥ −2.0), (2) low BMD (z-score between −2.0 and − 3.0), and (3) very low BMD (z-score ≤ −3.0). These thresholds are aligned with the recommendations of the International Society of Clinical Densitometry (ISCD) for premenopausal women and have been widely used in population-based studies to assess skeletal risk in younger adults. Categorizing BMD in this way allowed us to capture the gradual decline in bone health while ensuring sufficient sample sizes in each group to produce robust and meaningful statistical results. To analyze the relationship of categorical variables (PSS score, obstetrics and gynecological, categories of socio-economic status, physical activity level, nutrition variables, medical history, and anthropometric characteristics) with low bone mineral density, the chi-square test with Bonferroni’s correction was used. Bonferroni’s correction, p-value <0.025, was considered significant. To evaluate the baseline characteristics among different groups, one-way ANOVA with Tukey’s post hoc adjustment was used. Additionally, one-way ANOVA was used to draw a comparative analysis of the average mean of all nutrient variables among three groups. The Levene’s test was applied to check the homogeneity of variances for all variables across groups; when the test was significant, the Welch correction was used. Post-hoc comparisons were conducted using Tukey’s test when variances were equal, and Dunnett T3 when they were unequal. In addition, a multinomial logistic regression analysis using the force-entry method was performed to examine the independent association of various factors with reduced bone mineral density, with the normal BMD group serving as the reference category. To minimize model overfitting, the predictors entered into the model were selected based on biological relevance, bivariate correlations with p < 0.20, and prior evidence from young female cohorts. The final model included age at menarche, gravidity, hypothyroid status, perceived stress, socioeconomic status, daily sun exposure, and dietary calcium intake. Model fit was evaluated using the Hosmer–Lemeshow test, and multicollinearity was assessed through variance inflation factors (VIF < 2.5 considered acceptable). To reduce the likelihood of type I error, multiple comparisons were adjusted using the Bonferroni correction. Adjusted odds ratios (ORs) with 95% confidence intervals were reported, and a p-value <0.05 was considered statistically significant. All analyses were carried out using IBM SPSS Statistics version 26.0.

Prevalence of low bone spectrum

The crude prevalence of low bone mineral density conditions: z score below −2.0) and z score below-3.0 was found to be 22.6% (95% CI, 20.3–25.1), and 8.1% (95% CI, 6.6–9.8), respectively.

Percentage distribution of dietary practices by low z-score categories

Table 3 presents the distribution of bone mineral density (BMD) categories among the 1,176 women who participated in the study. Most women (69.30%) were classified as having normal BMD, indicating generally healthy bone status in this age group. Even so, a considerable number showed reduced bone density. About 22.61% fell into the low BMD range (Z-score − 2 to −3), suggesting early deterioration in bone strength. An additional 8.07% had very low BMD (Z-score < −3), placing them at the greatest risk for osteoporosis-related issues.

Distribution and mean comparison of micro- and macronutrient intake across BMD categories

Table 4 highlights noticeable differences in dietary intake across the three BMD groups. Women with normal BMD were much more likely to fulfil their daily calcium requirements, and this relationship was statistically significant. This suggests that sufficient calcium intake plays an important role in maintaining healthier bone density. A similar trend was seen for energy intake, where women with very low BMD were less likely to meet the recommended calorie levels.

For several other nutrients, such as protein, magnesium, folic acid, iron, and vitamin C the proportions were relatively similar across all groups, indicating that these deficits are widespread in the population rather than specific to any BMD category. On the other hand, nutrients like niacin and riboflavin showed meaningful differences, with adequate intake observed more frequently among women with normal BMD.

In summary, the findings point to insufficient calcium intake as the most influential dietary factor associated with reduced BMD, while shortages of other micronutrients appear to be common across all women, regardless of their bone density status.

Table 5 presents the multinomial regression analysis results after adjusting for the confounding variables BMI, SES, exercise, history of osteoporosis, and stress) showed that more than 3 pregnancies and late menarche age were found to be significant risk factors of low bone mineral density (p < 0.001). The presence of hypothyroidism was positively associated with low bone mineral density conditions. Further, lifestyle variables (high perceived stress, low SES, low sun exposure) were also positively correlated with low z-score categories. In addition, it was found that an increase of every 10 unit increase in physical activity levels was found to be a significant protective factor in delaying the onset of low bone mineral density (p < 0.001).

Nutrient risk factors for low z-score categories

Multiple regression analysis was used to find the dietary risk factors related to low bone mineral density. After adjusting for some confounding variables for low bone mineral density categories, the results revealed that low calcium intake was found to be independently associated with low z-score categories (OR: 7.09, 95% CI: 2.89–17.22; OR: 6.19, 95% CI: 1.52–25.24) (p < 0.01), respectively. All other factors were non-significant (p > 0.05).

Limitations and strengths

The present study has several limitations that should be considered. First, the cross-sectional design precludes any causal inferences between risk factors and bone mineral density. Second, although multi-site quantitative ultrasound (QUS) is a practical and widely used screening tool for assessing bone health in large population studies, it cannot replace dual-energy X-ray absorptiometry (DXA), the gold standard for BMD measurement. Additionally, reference ranges of z-scores for young Indian women using multi-site QUS have not yet been firmly established, which may affect interpretation. Important variables such as serum vitamin D levels, genetic factors, biochemical markers related to bone metabolism, and detailed quantification of physical activity were not assessed. Finally, the study sample was drawn from northern India, limiting the generalizability of the findings to other regions. Despite these limitations, the results emphasize the public health importance of early bone health screening and lifestyle interventions in young Indian women, particularly in settings where access to DXA is limited.

The study has several major strengths. It is the first large-scale Indian study to estimate the prevalence of low bone mineral density in young women using QUS technology. The accuracy and reliability of bone measurements were ensured, as all densitometer assessments were performed by a single trained operator. Nutritional assessments were consistently conducted by a qualified dietician, and all relevant factors related to low bone mineral density were included in the study. Additionally, the questionnaires were translated and validated in Punjabi, and the multisite Omnisense QUS device was validated for use in the Punjabi population for epidemiological purposes. Importantly, the final analytical sample (n = 1,176) exceeded the minimum required size, reflecting higher-than-expected participation and improving statistical precision, without introducing methodological bias, as the same inclusion criteria and ethical protocols were applied throughout.