Comparison of quantitative ultrasonography and dual X-ray absorptiometry for bone status assessment in South African children living with HIV.

Children living with HIV (CLHIV) have decreased bone mineral content (BMC) and density (BMD), increasing risk for fracture and future osteoporosis. While DXA is the gold-standard for bone assessments, it lacks availability in resource-constrained settings (RCS). Quantitative ultrasound (QUS) offers an alternative owing to its portability, low cost, ease of handling, and lack of ionizing radiation. While QUS has detected reduced bone quality in CLHIV, the relationship between QUS and DXA in this population remains unexplored. At baseline and 12 months, BMC and BMD of the whole body, lumbar spine, and radius were measured by DXA in a longitudinal cohort of CLHIV in Johannesburg, South Africa. Calcaneal speed of sound (SOS) and broadband ultrasound attenuation (BUA) and radius SOS were obtained by QUS, and calcaneal stiffness index (SI) was calculated. Spearman correlations, with and without HIV stratification, were performed between QUS and DXA measurements at each visit and for absolute difference in measurements between visits. At baseline and 12-months, calcaneal BUA and SI displayed strong positive correlations with DXA, with only modest correlations between radial QUS and DXA at baseline. Longitudinal measures of QUS did not correlate with DXA. At both baseline and 12-months, individuals with DXA whole-body BMD z-score < -1 displayed significantly lower calcaneal BUA and SI. Cross-sectionally, calcaneal QUS correlates strongly with whole body DXA and may represent a viable diagnostic alternative in RCS. Longitudinally, the two methods do not correlate well, possibly reflecting that each method assesses distinct aspects of bone architecture.

Introduction

Improved access to antiretroviral therapy (ART) has allowed many individuals with perinatally-acquired HIV (PHIV) to survive into adulthood. Childhood and adolescence represent a critical period for bone growth and turnover during which 85–90 percent of adult bone mass is acquired; therefore, disruptions of bone growth at this stage have the potential to impact long-term bone health [1]. Compared to uninfected individuals, children and adolescents living with HIV (CLHIV) have consistently been found to have decreased bone mineral content (BMC) and density (BMD), which may place them at risk for future osteoporosis and bone fracture [2].

To date, the majority of studies examining bone health in CLHIV have been conducted using dual-energy X-ray absorptiometry (DXA), generally considered the gold-standard. However, availability of DXA is limited in resource-constrained settings (RCS), due to high equipment cost and the requirement for extensively trained personnel for image acquisition and interpretation [3]. Considering that greater than 95 percent of PHIV instances occur in RCS [4], developing simpler low-cost alternative methods for evaluating bone status is an important consideration for research and clinical care in RCS.

A potential alternative is quantitative ultrasonography (QUS). In contrast to DXA, which utilizes two X-ray beams to generate attenuation coefficients attributable to soft tissue and bone, QUS operates by generating a traveling mechanical vibration whose shape, intensity, and speed are altered by the material properties of the bone through which it passes. Both the velocity and attenuation of the signal as it passes through tissue may be obtained non-invasively and without exposure to ionizing radiation to provide an assessment of bone quality.

Compared to DXA, QUS machines are highly portable, low in cost, require short scan times, and require less technical expertise for operation and interpretation [5]. Additionally, broadband ultrasound attenuation (BUA) measured by QUS has previously been shown to predict fracture risk in certain populations, perhaps with even better predictive power than DXA [6, 7]. As such, QUS offers an attractive option for bone status assessment in RCS, but significant challenges to its application remain, including the variability between the wide variety of QUS devices and inconsistent data supporting their use in specific populations and location of measurement sites.

To support its widespread clinical implementation, comparisons of QUS to the gold standard DXA may be informative. Some previous studies have reported inconsistent correlations between QUS and DXA, owing in part to the heterogeneity of QUS devices and study populations examined [7-9]. Few of these correlation studies between QUS and DXA have been performed in RCS. While we have previously shown that calcaneal QUS measures are reduced in CLHIV [10], these findings have not yet been related to DXA measures of bone quality in CLHIV.

To that end, the present work evaluates the correlations over one year between two different methods of QUS and DXA in young South African CLHIV receiving anti-retroviral therapies. Overall, we identified that calcaneal QUS correlates strongly with DXA in cross-section but not longitudinally, while radial QUS demonstrates a weaker correlation with DXA measurements of bone quality.

Methods

Study population

Subjects included in this analysis were a subset of participants from the Childhood HAART Alterations in Normal Growth, Genes, and aGing Evaluation Study (CHANGES) cohort, a longitudinal study conducted at Rahima Moosa Mother and Child Hospital and Chris Hani Baragwanath Hospital in Johannesburg, South Africa. Details of recruitment have been previously published [11, 12]. Enrollment occurred between March 2014 and May 2016. CLHIV (n = 218) were participants in a prior noninferiority randomized clinical trial (RCT) comparing viral suppression between initially viral-suppressed children who switched to EFV-based therapy and those remaining on LPV/r-based therapy [13]. Controls without HIV (n = 205) were recruited from siblings and household members of CLHIV [11]. Participants included in this study had data collected at baseline and 12-months. The study was approved by the Institutional Review Boards of Columbia University Irving Medical Center (CUIMC; New York, NY, USA) and the University of Witwatersrand (Johannesburg, South Africa). Signed informed consent was obtained from each child’s parent or guardian, and children provided assent if at least 7 years old and deemed able to understand.

Measurements and procedures

Digital scale and wall-mounted stadiometer were used to obtain anthropometric measurements [11]. Weight-for-age z score (WAZ), underweight (WAZ ≤ -2), overweight (WAZ > 2) for children aged 5–10 years old, height-for-age z score (HAZ), stunted (HAZ ≤ -2) for children aged 5–19 years old and BMI-for-age z score (BAZ) were determined using WHO norms [14]. Trained physicians assessed stage of pubertal development by using the highest score of breast or public hair development for female individuals and pubic hair for male individuals according to Tanner Staging method [11]. For CLHIV, plasma HIV-RNA levels (copies/ml) were determined with the Abbott RealTime HIV Assay (Abbot Park, Illinois, USA) and CD4+ counts (cells/μl and %) were measured by TruCount Method (BD Biosciences, Germany). Comparisons between ART regimens were performed only with those on a consistent regiment for the full study period.

Dual-energy X-ray absorptiometry (DXA)

Whole body (WB) and lumbar spine (LS) DXA scans were performed by licensed radiographers using a single Hologic Discover Wi bone densitometer at the Department of Radiology of Rahima Moosa Mother and Child Hospital. Scans were analyzed by a single technician, blinded to HIV status, using Apex software version 3.4 (Hologic Inc., Bedford, MA, USA) at the CUIMC Body Composition Unit (New York, NY, USA). Output measures included bone area in cm2, bone mineral content (BMC) in grams, and bone mineral density (BMD) in grams/cm2 [15]. BMD is a two-dimensional parameter and has potentially important limitations in growing children, as it may be strongly influenced by bone size [16]. To address this, BMD and BMC z-scores adjusted for age, sex, race and HAZ were calculated using reference norms from the US Bone Mineral Density in Childhood Study[17], as no South African references were available.

Calcaneal quantitative ultrasonography (QUS)

Calcaneal QUS measures were obtained at each visit using a Lunar Achilles Insight device (GE Healthcare, Madison, WI, USA) by study physicians using the standard operating procedures indicated by the manufacturer’s operating manual [18]. The built-in quality assurance test was performed prior to each use to ensure proper unit temperature, water level, and coupling. All measurements were obtained using manufacturer-provided foot shims to improve positioning. The device’s moveable region of interest (ROI) and the real-time QUS image were used to determine the size and location of the ROI, as well as the proper positioning of the foot [18]. Calcaneal measurements included broadband attenuated ultrasound (BUA) in decibels/megahertz (dB/MHz) and speed of sound (SOS) in meters/second (m/s). Calcaneal stiffness index (SI) was output from these measures using Njeh et al.’s adapted equation: SI = (0.67 * BUA + 0.28 * SOS)– 320 [19].

We excluded participants with SOS values ≥ 1625 m/s at either baseline or 12-months, as we have previously noted that such values represent spurious readings due to small foot size [20].

Radial quantitative ultrasonography

QUS measures were obtained along the distal 1/3 radius using the Sunlight Mini-Omnisense Sonometer (BeamMed, Petah Tikva, Israel). The primary output of this device is a transaxial measurement of SOS in m/s. To take measurements the probe of the device was positioned at the midpoint between the elbow and tip of the middle finger, parallel to the axis of bone. Once aligned, the probe was moved perpendicularly to the bone axis both medially and laterally, as described previously [21].

Statistical analyses

Chi-squared tests were used to compare categorical variables between groups. Student t tests and Wilcoxon rank-sum tests were conducted for normally and non-normally distributed measurements, respectively. Absolute change was calculated as the difference between measurements at 12-months and baseline. For missing data at the 12-month visit, data was imputed from a 6-month study visit using last observation carried forward imputation as a sensitivity analysis. We conducted an additional sensitivity analysis for individuals with missing data in which we performed multiple imputation with the mice package in R [22] with the assumption that data were missing at random.

Spearman correlations were performed to determine the correlation between QUS and DXA measurements. Power calculations were performed for both calcaneal and radial QUS correlations utilizing the pwr package in R [23]. For these calculations, desired power was set at 80 percent, alpha level was 0.05 and sample sizes were input at those present in this study. Correlations were performed at each visit separately, as well as for the absolute change in QUS and DXA measurements between visits. In the main analysis, correlations were pooled across HIV status. Correlation analyses stratified by HIV infection status were conducted. We additionally performed within-subject Spearman correlations for calcaneal BUA and whole body DXA BMD across baseline and follow-up visits, stratified by HIV infection status.

We also assessed if individuals with abnormal BMD as determined by DXA demonstrated differences in Calcaneal QUS measurements. To achieve this, z-scores were calculated for participants at each visit as previously described [24], adjusting for age, sex, race, height-for-age z-score and whole-body weight less head z-score. We established z-score groups by 3 methods: those with z ≤ -2.0, consistent with international consensus guideline definitions for childhood and adolescent osteoporosis [25], as well as those with z ≤ -1.0 or those with a decline in z-score between two visits. We then compared QUS parameters between these groups using Student’s t-tests.

Results

Demographic characteristics

Of the 423 individuals with DXA scans at baseline and 12-months in this study, 279 had calcaneal scans at baseline and 12-months. There were few significant differences between the 279 individuals with QUS scans and the 144 individuals without QUS, though a lower proportion of those without scans were CLHIV (S1 Table in S1 File). Of these 279, 109 had an abnormally elevated SOS value (> 1625 m/s) at either baseline or 12-months and were excluded. Differences between those included in analyses and those excluded are presented in S2 Table in S1 File. This resulted in a final sample of 80 CLHIV and 90 controls for calcaneal QUS analyses. At both baseline and 12-month follow-up, CLHIV had lower weight and height and were more often categorized as stunted compared to uninfected controls (Table 1). Both groups were similar in terms of age, sex, categorization as underweight, BMI-for-age z-score, and Tanner stage.

Table 1

Demographic characteristics of participants with insight calcaneal QUS measurements.

	(%)	Baseline (Visit 1)		12-month Follow-up (Visit 2)
	CLHIV (n = 80)	HIV- Control (n = 90)	P-value	CLHIV (n = 80)	HIV- Control (n = 90)	P-value
Age (years)
Mean (SD)	7.14 (1.37)	7.29 (1.55)	0.503	8.13 (1.38)	8.38 (1.53)	0.26
Median (IQR)	7.03 (4.62, 9.45)	7.27 (5.01, 9.92)		7.99 (5.71, 10.3)	8.41 (5.60, 11.2)
Sex, n (%)
Male	40 (50.0%)	51 (56.7%)	0.474	40 (50.0%)	51 (56.7%)	0.474
Female	40 (50.0%)	39 (43.3%)		40 (50.0%)	39 (43.3%)
Weight (kg)
Mean (SD)	21.7 (4.43)	23.7 (5.39)	0.007	24.0 (5.08)	26.9 (6.30)	0.001
Median (IQR)	21.3 (16.2, 26.3)	23.1 (17.7, 28.4)		23.8 (18.1, 29.5)	26.0 (18.6, 33.2)
Weight-for-age z-score
Mean (SD)	-0.579 (1.02)	-1.06 (1.10)	0.004	-0.610 (1.00)	0.0268 (1.13)	<0.001
Median (IQR)	-0.560 (-1.98, 0.86)	-0.230 (-1.63, 1.17)		-6.40 (-2.02, 0.74)	-0.115 (-1.57, 1.34)
Underweight, n (%)
No	73 (91.3%)	88 (97.8%)	0.12	68 (85.0%)	73 (81.1%)	0.121
Yes	7 (8.8%)	2 (2.2%)		6 (7.5%)	1 (1.1%)
Missing				6 (7.5%)	16 (17.8%)
Height (cm)
Mean (SD)	115 (8.37)	119 (8.50)	0.008	121 (7.87)	126 (7.94)	<0.001
Median (IQR)	115 (103, 128)	120 (108, 132)		120 (109, 132)	126 (115, 137)
Height-for-age z-score
Mean (SD)	-1.21 (0.913)	-0.716 (0.792)	<0.001	-1.06 (0.957)	-0.554 (0.822)	<0.001
Median (IQR)	-1.36 (-2.67, -0.05)	-0.665 (-1.85, 0.52)		-1.15 (-2.29, -0.002)	-0.465 (-1.63, 0.70)
Stunted, n (%)
No	62 (77.5%)	84 (93.3%)	0.006	70 (87.5%)	86 (95.6%)	0.104
Yes	18 (22.5%)	6 (6.7%)		10 (12.5%)	4 (4.4%)
BMI
Mean (SD)	16.1 (1.95)	16.6 (2.31)	0.135	16.1 (1.95)	16.9 (2.71)	0.046
Median (IQR)	15.8 (14.1, 17.5)	16.2 (13.6, 18.7)		16.0 (13.6, 18.4)	16.4 (13.4, 19.5)
BMI-for-age z-score
Mean (SD)	0.230 (1.07)	0.429 (1.19)	0.25	0.0489 (0.990)	0.304 (1.30)	0.149
Median (IQR)	0.245 (-0.76, 1.25)	0.305 (-1.18, 1.97)		0.080 (-0.96, 1.12)	0.335 (-1.41, 2.08)
Tanner Stage, n (%)
Tanner 1	78 (97.5%)	88 (97.8%)	1	76 (95.0%)	66 (73.3%)	0.172
Tanner 2	2 (2.5%)	2 (2.2%)		3 (3.8%)	8 (8.9%)
Missing				1 (1.3%)	16 (17.8%)
Viral load (copies/ml) [N (%)]
LDL (≤20 or ≤40)	73 (91.3%)	NA	NA	53 (66.2%)	NA	NA
21–1000	6 (7.5%)			9 (11.2%)
>1000	1 (1.25%)			2 (2.50%)
N.A.	0 (0.0%)			16 (20.0%)
CD4+ count (cells/μl), mean (SD)	1185.5 (373)	NA	NA	1122.7 (393)	NA	NA
ART regimen category [N (%)]
EFV-based	41 (51.2%)	NA	NA	43 (53.7%)	NA	NA
LPV/r-based	39 (48.7%)			37 (46.2%)

Of the original 423 participants, 205 had radial scans at both baseline and the 12-month follow-up. Compared to those with scans, those missing radial QUS scans had lower average height, higher BMI-for-age z-score, and lower CD4 cell count (S3 Table in S1 File). This resulted in a final sample of 101 CLHIV and 104 controls for radial QUS analyses. At baseline and 12-month follow-up, CLHIV had lower weight and height, were more often categorized as underweight, and more often categorized as stunted (Table 2). Both groups were similar in terms of sex, BMI-for-age z-score, and Tanner stage at both baseline and follow-up, while uninfected controls were significantly older at follow-up.

Table 2

Demographic characteristics of participants with Mini-Omnisense QUS measurements.

	Baseline (Visit 1)			12-month Follow-up (Visit 2)
	CLHIV (n = 101)	HIV- Control (n = 104)	P-value	CLHIV (n = 101)	HIV- Control (n = 104)	P-value
Age (years)
Mean (SD)	6.57 (1.32)	6.94 (1.45)	0.059	7.57 (1.33)	7.98 (1.44)	0.035
Median (Min, Max)	6.12 (5.01, 9.77)	6.83 (5.02, 9.99)		7.12 (6.01, 11.1)	7.86 (6.01, 11.0)
Sex, n (%)
Male	50 (49.5%)	55 (52.9%)	0.731	50 (49.5%)	55 (52.9%)	0.731
Female	51 (50.5%)	49 (47.1%)		51 (50.5%)	49 (47.1%)
Weight (kg)
Mean (SD)	19.5 (4.14)	22.2 (5.22)	<0.001	21.8 (4.69)	25.1 (6.32)	<0.001
Median (Min, Max)	18.8 (13.8, 39.9	21.4 (14.6, 48.0)		21.1 (14.8, 46.6)	24.2 (16.7, 54.5)
Weight-adjusted z-score
Mean (SD)	-0.900 (0.902)	-0.281 (0.974)	<0.001	-0.858 (0.891)	-0.179 (1.03)	<0.001
Median (Min, Max)	-0.970 (-2.46, 1.94)	-0.375 (-2.31, 3.64)		-0.810 (-2.60, 1.57)	-0.270 (-2.57, 3.47)
Underweight, n (%)
No	91 (90.1%)	103 (99.0%)	0.011	86 (85.1%)	91 (87.5%)	0.026
Yes	10 (9.9%)	1 (1.0%)		9 (8.9%)	1 (1.0%)
Missing				6 (5.9%)	12 (11.5%)
Height (cm)
Mean (SD)	112 (8.93)	117 (9.27)	<0.001	118 (8.27)	124 (8.90)	<0.001
Median (Min, Max)	110 (94.6, 137)	118 (95.7, 140)		117 (98.5, 142)	124 (104, 147)
Height-adjusted z-score
Mean (SD)	-1.36 (0.925)	-0.675 (0.920)	<0.001	-1.21 (0.897)	-0.472 (0.954)	<0.001
Median (Min, Max)	-1.45 (-3.31, 1.08)	-0.695 (-3.73, 2.11)		-1.28 (-3.36, 1.35)	-0.610 (-3.17, 2.50)
Stunted, n (%)
No	76 (75.2%)	96 (92.3%)	0.002	83 (82.2%)	101 (97.1%)	<0.001
Yes	25 (24.8%)	8 (7.7%)		18 (17.8%)	3 (2.9%)
BMI-for-age z-score
Mean (SD)	-0.0675	0.163 (1.05)	0.101	-0.152 (0.922)	0.0174 (1.13)	0.241
Median (Min, Max)	(0.946)	0.155 (-2.10, 3.31)		-0.160 (-2.61, 2.00)	-0.0300 (-3.20, 3.84)
	-0.0800 (-3.37, 1.87)
Tanner Stage, n (%)
Tanner 1	100 (99.0%)	102 (98.1%)	1	98 (97.0%)	88 (84.6%)	0.428
Tanner 2	1 (1.0%)	2 (1.9%)		3 (3.0%)	6 (5.8%)
Missing					10 (9.6%)
Viral load (copies/ml) [N (%)]
LDL (≤20 or ≤40)	93 (92.1%)	NA	NA	59 (58.4%)	NA	NA
21–1000	5 (4.95%)			12 (11.9%)
>1000	3 (2.97%)			4 (3.96%)
N.A.	0 (0.0%)			26 (25.7%)
CD4+ count (cells/μl), mean (SD)	1281.2 (459)	NA	NA	1185.0 (386)	NA	NA
ART regimen category [N (%)]
EFV-based	54 (53.4%)	NA	NA	57 (56.4%)	NA	NA
LPV/r-based	47 (46.5%)			44 (43.6%)

We additionally conducted sensitivity analyses by imputing missing data for Tanner staging using the last observation carried forward method, in which we utilized Tanner stage from a 6-month follow up for those missing data at 12-months. Results remained similar after imputation, as 15 of 16 individuals with missing data were Tanner Stage 1 at the 6-month visit. Similarly, for those missing viral load data at 12-months, we imputed data from the 6-month visit, which was available for 14 of 16 participants with missing data. Findings were consistent with overall data trends: 11 had viral copies at lower detection limit (LDL) or < 40 copies/ml and 3 had a viral load between 21 and 1000 copies/ml. We performed further sensitivity analyses in which we performed multiple imputation for missing data with the assumption that data was missing at random. By this method, results remained similar: 13 of 16 individuals with missing Tanner staging were Tanner 1 in imputed data, and 12 of 16 individuals with missing viral loads were LDL or < 40 copies/mL after imputation.

Baseline measures

We first compared differences in QUS and DXA measurements between CLHIV and uninfected controls (Table 3). DXA measures of WB BMD and BMC, whole body BMD, were lower in CLHIV than controls, similar to our previous reports [2, 11]. Calcaneal SOS and SI were significantly lower in CLHIV, while radial SOS did not differ between groups (Table 3).

Table 3

Comparison of baseline and follow-up QUS and DXA measures between CLHIV and uninfected controls.

	Calcaneal QUS			Radial QUS
	CLHIV (n = 80)	HIV- Control (n = 90)	p-value	HIV+ (n = 101)	HIV- Control (n = 104)	p-value
Baseline
SOS, mean (SD)	1558.9 (17.6)	1568.8 (19.7)	<0.001	3609.3 (117)	3587.7 (128)	0.213
BUA, mean (SD)	85.15 (11.8)	88.75 (13.1)	0.061	NA	NA	NA
Stiffness Index, mean (SD)	73.26 (10.0)	78.57 (11.5)	0.002	NA	NA	NA
Whole Body BMC, mean (SD)	715.0 (118)	783.8 (133)	<0.001	671.3 (121)	770.7 (141)	<0.001
Whole Body BMD, mean (SD)	0.673 (0.063)	0.708 (0.064)	<0.001	0.652 (0.065)	0.701 (0.067)	<0.001
Lumbar BMC, mean (SD)	17.31 (5.13)	18.52 (5.16)	0.126	15.63 (4.67)	17.29 (5.02)	0.015
Lumbar BMD, mean (SD)	0.600 (0.071)	0.619 (0.066)	0.078	0.586 (0.076)	0.610 (0.068)	0.016
1/3 Distal Radius BMC, mean (SD)	36.99 (9.04)	39.78 (9.53)	0.052	32.95 (8.61)	38.23 (10.0)	<0.001
1/3 Distal Radius BMD, mean (SD)	0.406 (0.056)	0.421 (0.061)	0.099	0.380 (0.058)	0.405 (0.063)	0.003
Follow-up
SOS, mean (SD)	1567.1 (20.9)	1571.4 (20.2)	0.175	3623.9 (105)	3570.6 (134)	0.002
BUA, mean (SD)	88.78 (13.7)	97.88 (13.6)	<0.001	NA	NA	NA
Stiffness Index, mean (SD)	78.09 (12.2)	85.33 (11.0)	<0.001	NA	NA	NA
Whole Body BMC, mean (SD)	813.9 (131)	877.2 (139)	0.003	762.0 (133)	858.7 (147)	<0.001
Whole Body BMD, mean (SD)	0.729 (0.064)	0.754 (0.062)	0.012	0.704 (0.064)	0.743 (0.069)	<0.001
Lumbar BMC, mean (SD)	20.20 (5.29)	21.26 (5.64)	0.212	18.43 (4.94)	20.08 (5.52)	0.026
Lumbar BMD, mean (SD)	0.612 (0.074)	0.632 (0.072)	0.076	0.592 (0.074)	0.626 (0.068)	<0.001
1/3 Distal Radius BMC, mean (SD)	42.52 (9.30)	46.24 (10.3)	0.015	38.22 (8.89)	44.10 (11.1)	<0.001
1/3 Distal Radius BMD, mean (SD)	0.439 (0.060)	0.453 (0.062)	0.150	0.412 (0.057)	0.436 (0.065)	0.006

We performed power calculations to determine the expected effect size given the sample size for both calcaneal and radial QUS. With a predetermined power of 80 percent and alpha level of 0.05, the expected correlation coefficient for calcaneal QUS (n = 170) was 0.213, and the expected coefficient for radial QUS (n = 205) was 0.194 (S1 Fig in S1 File).

In analyses pooled across CLHIV and controls, calcaneal QUS measures were strongly correlated with DXA measures (Fig 1(a) and 1(b)). Both calcaneal BUA and SI were correlated with whole body (r = 0.43 to r = 0.52, p < 0.01), lumbar spine (r = 0.38 to r = 0.43, p < 0.01), and 1/3 distal radius (r = 0.41 to r = 0.52, p < 0.01) BMD and BMC measures. Calcaneal SOS did not correlate with most DXA measures and displayed only modest correlations with whole body BMC (r = 0.15, p < 0.01), and lumbar spine BMC (r = 0.19, p < 0.01).

Fig 1

Correlations between QUS and DXA measures at baseline and follow-up, pooled across HIV groups.

(a) Results of Spearman correlation between baseline calcaneal QUS measures and DXA (n = 170). (b) Results of Spearman correlation between baseline radial QUS measures and DXA (n = 205). Shaded cells indicate p < 0.05, with darker shading indicating more significant values. (c) Results of Spearman correlation between follow-up calcaneal QUS measures and DXA (n = 170). (d) Results of Spearman correlation between follow-up QUS measures and DXA (n = 205). Shaded cells indicate p < 0.05, with darker shading indicating more significant values. Numbers included within individual cells represent Spearman rho values. BMC: bone mineral content; BMD: bone mineral density; BUA: broadband ultrasound attenuation; SOS: speed of sound; DXA: dual-energy x-ray absorptiometry.

Similarly, SOS from the radial QUS correlated with whole body (r = 0.26 to r = 0.31, p < 0.01), lumbar spine (r = 0.19 to r = 0.28, p < 0.01), and 1/3 distal radius (r = 0.29 to r = 0.33, p < 0.01) BMD and BMC measures.

We additionally explored the baseline correlations between calcaneal QUS and DXA in analyses stratified by HIV status. Results of stratified correlations were comparable to the main pooled analysis, with BUA and SI showing strong correlations with BMD and BMC, particularly among CLHIV (S3 and S4 Figs in S1 File).

Follow-up measures

We also determined differences between groups in terms of QUS and DXA measures at the 12-month follow-up visit, and results were largely similar. While significant at baseline, calcaneal SOS did not differ at 12-months. Calcaneal BUA and SI were significantly lower in CLHIV, while radial SOS was significantly higher in CLHIV, at follow-up.

We also performed cross-sectional correlations at 12-months, and calcaneal QUS was strongly correlated with DXA measures (Fig 1(c) and 1(d)). Calcaneal BUA and SI were each correlated with whole body, lumbar spine and 1/3 distal radius BMC and BMD (r = 0.25 to r = 0.45, p < 0.001). Additionally, BUA was correlated with lumbar and 1/3 distal radius (p < 0.05). Calcaneal SI was correlated with 1/3 distal radius BMC and BMD (p < 0.05). Calcaneal SOS displayed modest correlations with whole body BMC and BMD, as well as lumbar BMC (p < 0.05). At follow-up, 1/3 distal radius SOS was modestly correlated with 1/3 distal radius BMD, as well as whole body BMC and BMD (p < 0.05).

We additionally explored the follow-up correlations between calcaneal QUS and DXA in analyses stratified by HIV status. Results of stratified correlations were comparable to the main pooled analysis, with BUA and SI showing strong correlations with DXA, particularly whole-body BMD and BMC (S3 and S4 Figs in S1 File).

Longitudinal measures

Between-group differences were also determined between the absolute change from baseline to follow-up in QUS and DXA measures. Change in whole body BMD differed between groups, with CLHIV demonstrating a greater increase in whole body BMD over the 12-month period. CLHIV also demonstrated a greater increase in calcaneal SOS during the study period. To determine whether changes in calcaneal and radial QUS measures correlated with changes in DXA measures between visits, we performed Spearman correlations between the absolute change between visits of each measurement (S2 Fig in S1 File).

Considering absolute change, calcaneal QUS measures were poorly correlated with DXA measures. Only change in calcaneal SOS had a modest correlation with change whole body BMC and BMD (r = 0.17 to r = 0.22, p < 0.05). Change in radial SOS was also poorly correlated with change in BMC and BMD.

As in the main pooled analyses, correlations between absolute change in calcaneal QUS and DXA were not consistently identified in analyses stratified to CLHIV only or uninfected controls only (S3 and S4 Figs in S1 File).

Inter-visit correlations

We also examined the correlation of calcaneal BUA and whole body DXA BMD between baseline and follow-up visits to determine the degree of within-subject variability for each variable. Considering the correlation between baseline and follow-up for whole body DXA BMD, strong positive correlations were identified for both CLHIV (rho = 0.924, p < 0.001) and control (rho = 0.921, p < 0.001) groups (Fig 2(a) and 2(b)). In contrast, weaker positive correlations were identified for calcaneal BUA between baseline and follow-up visits in CLHIV (rho = 0.318, p = 0.004) and control (rho = 0.508, p < 0.001) groups (Fig 2(c) and 2(d)).

Fig 2

Within-subject Spearman correlations between baseline (Visit 1) and follow-up (Visit 2) visits, stratified by study group.

(a) Correlation between baseline and follow-up whole body DXA BMD within CLHIV group. (b) Correlation between baseline and follow-up whole body DXA BMD within control group. (c) Correlation between baseline and follow-up calcaneal BUA. (d) Correlation between baseline and follow-up calcaneal BUA. Blue lines indicate the line of best fit from Spearman correlation. BMD: bone mineral density; BUA: broadband ultrasound attenuation; DXA: dual-energy x-ray absorptiometry.

QUS comparisons across DXA z-score groups

To determine if differences in QUS measurements were present in those with abnormal BMD, we stratified the cohort into groups according to z-scores at each visit (S5 Table in S1 File). At baseline, those with a DXA whole-body BMD z ≤ -1.0 (n = 61) had significantly lower calcaneal BUA and SI compared to those with z > -1.0 (n = 109), and these findings were recapitulated at 12-months. No difference was identified between calcaneal SOS between these groups. There were no differences in radial QUS parameters between those with whole body DXA BMD z ≤ -2.0 and z > 2.0 at either baseline or follow-up. Similarly, there were no differences in QUS parameters between those who demonstrated a decrease in whole body DXA BMD z-score between baseline and 12-months at either visit.

Discussion

In this longitudinal analysis of the relationships between calcaneal and radial QUS and DXS, we found a strong correlation between calcaneal QUS and DXA at two cross-sectional time points. In particular, both calcaneal BUA and SI were strongly correlated with whole body DXA measures at baseline and 12-month follow-up. However, neither calcaneal nor radial QUS displayed strong correlations with DXA when absolute change between baseline and follow-up was considered. In addition, while DXA BMD displayed strong inter-visit correlations within subjects, calcaneal BUA displayed much weaker, albeit significant, correlations between visits.

Several studies have explored the relationship between QUS and DXA measures in different study populations. Among post-menopausal women, for instance, a strong correlation has consistently been found between calcaneal QUS and DXA at multiple body sites [8, 26]. Similarly, a study of adult patients living with HIV identified more moderate correlations between calcaneal QUS and DXA, particularly whole body DXA measurements, which is consistent with the present study [7]. More variable results are reported in children. Similar to our study, calcaneal SI correlated with whole body BMC and BMD in a cohort of healthy Chinese children and adolescents [9]. Likewise, SOS measured by tibial QUS in a healthy Dutch pediatric cohort correlated strongly with both lumbar and whole body DXA, and a Swedish pediatric study of healthy participants identified strong correlations of calcaneal SI, SOS, and BUA with whole body and lumbar BMD [27, 28]. These studies provide strong evidence that the correspondence between calcaneal QUS and DXA, particularly whole body DXA, is significant and may have clinical utility.

More inconsistent results have been reported with respect to radial QUS, similar to the results we report. In a prior study of healthy Malaysian children, agreement between radial QUS and DXA was weak [29]. This finding was recapitulated in a study comparing radial QUS to DXA among healthy Thai children [30]. A study of previously pre-term infants also failed to identify a correlation between radial SOS and DXA measures [31]. Similarly, in children with hemophilia A, no correlations were identified between radial QUS measures and BMD [32]. As such, the poor agreement we observed between radial QUS and DXA, even at the same measurement site, suggests that the two methods are not comparable for bone status assessment in CLHIV. In our study, this discordance may further be explained by our measurement of only SOS at the radius, a parameter which was not well correlated with DXA at the calcaneus and likely reflects different properties of bone than does BUA.

Despite strong cross-sectional correlations, changes in BUA and SI failed to correlate with changes in DXA measures. This may be due to the fact that QUS outputs such as BUA measure different properties of bone than DXA, and these properties may display different trajectories of growth compared to BMC and BMD measured by DXA. BUA reflects not only the quantity of bone through which the ultrasound beam passes but also the physical arrangement, or micro-architecture, of the bone [33, 34], which may also partially explain why BUA is a better predictor of fracture risk than DXA in some populations [7, 35]. As DXA is unable to distinguish between trabecular and cortical bone, changes in DXA measures may not sensitively reflect some of the micro-architectural changes that occur during periods of growth in childhood and adolescence [36]. Notably, calcaneal BUA displayed a much weaker inter-visit correlation than did DXA BMD, suggesting the measure may not be as stable as DXA over time. Some have suggested that while QUS may adequately measure response to therapy, and correlates well with DXA, it may lack precision in repeat measures longitudinally [37]. Others have described the need for more precise quality control measures for QUS, such as those existing for DXA, that would improve the precision of QUS as a longitudinal measure, particularly noting that QUS devices may be sensitive to temperature changes and require frequent calibration [38-40].

Taken together, our results build upon prior work to suggest the utility of calcaneal QUS for assessment of bone status in pre-pubertal South African CLHIV. We identify that in cross-section, BUA is a strong correlate of DXA, yet differences in which aspects of bone architecture each methodology captures may render longitudinal comparisons less reliable. Furthermore, the finding of poor inter-visit correlation in BUA compared to DXA suggests the ongoing need for improvements in precision and quality control in QUS methodology. Regardless, in RCS where DXA is unlikely to become widely available and in which the majority of CLHIV reside, calcaneal ultrasound may provide a useful tool for initial assessment of bone quality.

(DOCX)

Click here for additional data file.

Supporting dataset.

(XLSX)

Click here for additional data file.

3 Aug 2022

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Robert Daniel Blank, MD, PhD

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf  and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include your full ethics statement in the ‘Methods’ section of your manuscript file. In your statement, please include the full name of the IRB or ethics committee who approved or waived your study, as well as whether or not you obtained informed written or verbal consent. If consent was waived for your study, please include this information in your statement as well.

3. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

4. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The manuscript conducts secondary statistical analysis to assess correlations, both cross-sectional and longitudinal, between two different methods of QUS and DXA on data generated from a previous noninferiority RCT comparing CLHIV versus controls without HIV. There are 2 longitudinal time-points: baseline, and 12-month followup. The study is based on prior science. I have some additional questions.

(a) I understand this is secondary analysis, but a sample size/power justification would assist readers to plan future trials and analysis like this. Authors may simply try to calculate the effect size they wanted to see (at the onset) to assess the differences between the groups.

(b) LOCF is really a sub-optimal method in considering missing data imputation; with several techniques available, the authors may resort to one of the credible ones. A team statistician (within the author list) is recommended, and who can be helpful in using the correct imputation method.

Reviewer #2: This reviewer has a methods concern:

1. The DXA scores were grouped: Z-score groups were z <-1.0, z < -1.5 and those who had declining z-score between visits

Please explain why the data were grouped by z-score less than -1.0 and less than -1.5. Why wasn’t the z-score of <-2.0 used, which is the internationally accepted level of what is considered to be low bone density in children. This reviewer is not convinced that there would be enough difference between the z-score less than -1.0 and less than -1.5 groups to be meaningful, particularly in this sample size.

Doubtful that this will impact the overall findings of the paper, but these groups, based on z-score, seem arbitrary and do not follow any guidelines for what is considered to be low BMD. They do not sensibly stratify the data.

2. Editorial comment: z-scores should be expressed as a decimal (e.g. -1.0). In this paper, z-scores are expressed as -1.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

10 Sep 2022

The authors would like to thank the reviewers for their insightful comments, which we feel have substantially strengthened the manuscript. We have incorporated all suggested edits and have included responses to each individual reviewer’s comments below.

Reviewer #1: The manuscript conducts secondary statistical analysis to assess correlations, both cross-sectional and longitudinal, between two different methods of QUS and DXA on data generated from a previous noninferiority RCT comparing CLHIV versus controls without HIV. There are 2 longitudinal time-points: baseline, and 12-month followup. The study is based on prior science. I have some additional questions.

(a) I understand this is secondary analysis, but a sample size/power justification would assist readers to plan future trials and analysis like this. Authors may simply try to calculate the effect size they wanted to see (at the onset) to assess the differences between the groups.

To address this concern, we have calculated the effect size using the pwr package in R, as described in Lines 205-207. Estimated effect coefficients are included in Lines 277-280. Supplementary Figure 1 has also been added and includes a visual output of power calculations with expected effect size given our sample size and power of 80%.

(b) LOCF is really a sub-optimal method in considering missing data imputation; with several techniques available, the authors may resort to one of the credible ones. A team statistician (within the author list) is recommended, and who can be helpful in using the correct imputation method.

To address this, we have conducted an additional sensitivity analysis in which we performed multiple imputation assuming data was missing at random and have added the details to the Methods section (Lines 200-202). Results of this second sensitivity analysis were consistent with those from the last observation carried forward method, as well as the main analysis (Lines 261-265).

Reviewer #2: This reviewer has a methods concern:

1. The DXA scores were grouped: Z-score groups were z <-1.0, z < -1.5 and those who had declining z-score between visits

Please explain why the data were grouped by z-score less than -1.0 and less than -1.5. Why wasn’t the z-score of <-2.0 used, which is the internationally accepted level of what is considered to be low bone density in children. This reviewer is not convinced that there would be enough difference between the z-score less than -1.0 and less than -1.5 groups to be meaningful, particularly in this sample size.

Doubtful that this will impact the overall findings of the paper, but these groups, based on z-score, seem arbitrary and do not follow any guidelines for what is considered to be low BMD. They do not sensibly stratify the data.

We have amended the methods and results to utilize cutoffs of z <= -2, in accordance with the latest consensus guidelines, and have removed the z <= -1.5 cutoff. As there are a very small number of participants in this study with z <= -2, we have utilized the z <= -1 cutoff to capture a larger number of participants with reduced DXA scores for comparison that did not necessarily meet consensus guidelines for childhood osteoporosis. We have reflected the Methods (Lines 218-219) and Results (Lines 379-380) to reflect these changes.

2. Editorial comment: z-scores should be expressed as a decimal (e.g. -1.0). In this paper, z-scores are expressed as -1.

We have edited the paper to include decimals for z-scores.

Submitted filename: Reviewer Response.docx

Click here for additional data file.

5 Oct 2022

Comparison of Quantitative Ultrasonography and Dual X-ray Absorptiometry for Bone Status Assessment in South African Children Living with HIV

PONE-D-22-15788R1

Dear Dr. Roberts,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Robert Daniel Blank, MD, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: (No Response)

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

**********

7 Oct 2022

PONE-D-22-15788R1

Comparison of Quantitative Ultrasonography and Dual X-ray Absorptiometry for Bone Status Assessment in South African Children Living with HIV

Dear Dr. Roberts:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Robert Daniel Blank

Academic Editor

PLOS ONE