Abstract Context Patients with Turner syndrome (TS) are at risk for osteoporotic fractures. Objective The aims of this study were to assess the incidence of clinically important fractures in girls with TS and prospectively describe the development of volumetric bone mineral density (BMD). Design Peripheral quantitative computerized tomography (pQCT) of the radius every other year over the 6 years of observation. Setting Government-funded university referral center. Participants Thirty-two girls with TS, aged 6 to 16 years, were included in the analyses. Fracture incidence was compared with the data in the general population. Bone density and strength were compared with data from 185 healthy girls. Outcomes The main clinical outcome was the fracture occurrence. The secondary outcomes were the changes in Z-scores of the bone parameters. Results Three girls with TS sustained four fractures during 6 years of observation. The fracture rate in TS was not substantially higher than the downward-biased fracture-rate estimate from age-matched, healthy controls (P = 0.48). Whereas the trabecular BMD Z-score decreased with age (β estimate −0.21 ± 0.04, P < 0.001), total bone cross-sectional area correspondingly increased (+0.16 ± 0.04, P < 0.001), which led to normal bone strength. A positive history of incident fractures was not significantly associated with any of the pQCT-derived bone parameters. Conclusions Current pediatric TS patients that are treated with growth hormone and estrogens are not at risk for osteoporotic fractures. Low BMD in TS may be counterweighted by enlarged bone radius, which leads to normal bone strength at the appendicular skeleton. Turner syndrome (TS) is mainly characterized by short stature as a result of the lower-leg growth reduction and by amenorrhea as a result of the lack of ovarian hormone production. Among other symptoms, individuals may be relatively inconspicuous by presenting with recurrent otitis media (1); however, some may develop life-threatening aortic dissection (2). TS affects 1:2000 live-born females (3), and the symptoms are caused by the loss of the short arm of the X chromosome. The first systematic search for associations of different medical conditions with TS was performed in Denmark, based on the national registry for hospitalized patients (4). From this study, it was determined that fractures occur twice as frequently in females with TS as they do in the general population. A higher fracture rate in patients with TS was later confirmed by a different approach consisting of a nationwide questionnaire survey, where the fracture frequency was increased by 25% in TS patients and the mean fracture-free survival was 10 years shorter than that of the general population (5). Increased bone fragility in patients with TS was seemingly and reasonably explained by the lack of estrogen production, which could lead to a low bone mass. Low bone mineral density (BMD), in association with estrogen insufficiency, was indeed reported in patients with TS in several studies (6, 7). However, other studies showed a normal BMD in TS, especially when the short stature of these patients was considered, and estrogens were substituted (8, 9). Parameters other than BMD, such as bone geometry measures, also contribute to whole bone strength. Enlarged width and decreased cortical thickness at the radius were reported in girls and young adult females with TS (10, 11). This skeletal phenotype is estrogen independent and is probably caused by the deletion of the short stature homeobox gene, as this feature is shared among patients with TS and patients with isolated short stature homeobox deficiency (12). Despite all of the previous studies, the age-controlled, time-controlled, and region-controlled prospective fracture incidence among pediatric patients with TS is not known, and the long-term development of volumetric BMD (vBMD) and other geometry-based bone-strength indices has not been previously described in this condition. The aims of this study were the following: (1) to determine whether fracture incidence is increased in girls with TS during their growth, while getting growth hormone (GH) and if needed, estrogens, compared with that in the general female population and (2) to search for associations with the development of BMD and bone strength over a 6-year period. Subjects and Methods Patients This was a prospective, single university hospital referral center study conducted between January 2008 and September 2014. We asked all prepubertal and pubertal patients with TS (n = 47) participating in our previously published cross-sectional peripheral quantitative computerized tomography (pQCT) study (11) to continue in a long-term follow-up. All of the girls (and their parents) consented, but some dropped out as a result of transition to adult care (n = 9), recognition of another disease(s) affecting the bone metabolism (1× Crohn’s disease, 1× primary ciliary dyskinesia with bronchiectasis), or relocation (n = 4). The assessments were made 2, 4, and 6 years after the initial bone scan. Patients were included when bone scans were available for at least three of the four study time points (28 girls) or at the study start and end time points (four girls). Therefore, the final analyses were conducted in 32 patients. The TS diagnosis was confirmed by genetic testing in all participants. Their karyotypes were X monosomy (n = 15), the 45,X/46,XX mosaicism (n = 2), other double or triple cell-line mosaics with at least two aberrant cell lines (n = 10), or there was a structural defect in one X chromosome (two deletions, one isochromosome, and two translocations). At the study start, there were 15 prepubertal (Tanner stage I breast development), 14 pubertal premenarcheal, and three postmenarcheal TS girls. The body height was measured with a stadiometer to the nearest 1 mm. Weight was measured with an electronic scale to the nearest 100 g. Body mass index was calculated as the weight (kilograms) divided by the square of the height (meters). Height, weight, and body mass index Z-scores were calculated using the most recent national reference data (13). The forearm length Z-scores were calculated based on data from a previously published pQCT study (14). Hormonal therapy All of the girls with TS were treated with recombinant human GH at a median dose of 50 μg/kg/day; the dose has been adjusted according to body weight at each of the subsequent visits at the out-patient clinic (15). Three postmenarcheal girls finished their GH therapy <8 months before the first densitometry assessment. The GH therapy was completed when the near-final height was achieved (i.e., growth velocity <2 cm/year estimated from a period of 6 months or more). Six patients developed spontaneous puberty, whereas 24 and two girls with primary and secondary amenorrhea, respectively, were supplemented with oral estradiol (17-β estradiol) pills to induce/complete their pubertal development. The scheme was based on a low starting dose, which was gradually increased, as described previously in detail (11). All but three girls started menstruating during the 6-year follow-up study. Thyroxine substitution was necessary in eight patients with autoimmune thyroid disease and one patient after total thyroidectomy as a result of a suspect fine-needle aspiration biopsy obtained from a thyroid nodule. All of these patients had been euthyroid over the study duration. Fracture incidence and comparison with the healthy population Fractures were ascertained through a simple questionnaire focused on age, site, and the mechanism of injury. The fracture history was taken at the time of first densitometry (prevalent fractures) and then at every subsequent visit (incident fractures). The fractures were verified through X-rays. Clinically significant fractures, as defined by the International Society for Clinical Densitometry (16), were included in the following analyses. The age-stratified national female population counts and the age-stratified numbers of females hospitalized with a clinically significant fracture (and no other reported diagnosis) were obtained from the Institute of Health Information and Statistics through the National Register of Hospitalised Patients for every year between 2008 and 2014 (www.uzis.cz/en/registers/national-health-registers/nr-hospitalised-patients). Because there was no possibility to gain age-stratified data on outpatient-treated females with fractures at the national level, for reference, we used the age-stratified numbers of females who attended the Emergency Department of Motol university hospital (with 2200 beds as the biggest in the country) because of trauma, resulting in the fracture of the limbs or spine (and no other diagnosis), in every year between 2008 and 2014. We also registered whether they were hospitalized. Control group for pQCT-derived bone parameters The control group consisted of 185 healthy girls who had been included in the Dortmund Nutritional and Anthropometric Longitudinally Designed Study, performed at the Research Institute for Child Nutrition in Dortmund, Germany (17). This was an observational study investigating the inter-relations of nutrition, growth, and metabolism in healthy children. The cohort was recruited in a representative sample of school children of Dortmund that was mostly composed of middle-class families, and all participants were of white race. The age-specific reference values for bone density and geometry parameters in this population were already published (18, 19). For this particular study, the original data from healthy girls were used to create height-specific smooth reference curves for all of the relevant bone parameters. Ethical considerations The current study was approved by the institutional Ethics Committee and conformed to the Declaration of Helsinki. All of the participants included in this study (or a parent if the patient was younger than 18 years) consented to the testing. Bone strength assessment An XCT 2000 scanner (StratecMedizintechnik, Pforzheim, Germany) was used to obtain measurements at the nondominant radius. The dominancy was determined according to handedness. A single tomographic 2.0 mm-thick slice was taken at distances corresponding to the 4% and 65% bone lengths, as measured from the styloid process to the olecranon of the ulna. A voxel size of 0.5 × 0.5 × 2.0 mm was used. The images were processed, and the numerical values were calculated using version 6.20 C of the integrated XCT software. At the distal radius (4% site), the total bone mineral content, total bone cross-sectional area (CSA), and trabecular vBMD were measured through the CBD function, C1/P1 mode, and threshold of 280 mg/cm3. The trabecular bone compartment was assumed to be represented by the inner 45% of the total bone CSA. The polar strength-strain index (SSI polar) was determined using a function SSI, separation mode one, and threshold of 280 mg/cm3. At the proximal radius (65% site), the total bone mineral content and total bone CSA were measured through the CBD function, C1/P1 mode, and threshold of 280 mg/cm3. The cortical bone CSA and cortical vBMD were assessed through the CORTBD function, separation mode one, and threshold of 710 mg/cm3. Cortical vBMD was then adjusted for the partial volume effect by an algorithm suggested by Rittweger et al. (20) and previously described in detail (21). The relative cortical bone CSA was calculated as the ratio between the cortical bone CSA and total bone CSA and expressed in percent (i.e., multiplied by a factor of 100). The cortical thickness was calculated as a difference between the total bone radius and marrow bone radius using the following formula: (total bone CSA/π)0.5 – [(total bone CSA – cortical bone CSA)/π]0.5. The SSI polar was determined by using the SSI function, separation mode one, and threshold of 480 mg/cm3. The precision errors of the pQCT measures were reported to be low at the radius (11). Biochemistry The blood samples were obtained from the patients on the day of the bone assessment and were a part of their regular follow-up. Serum follicle-stimulating hormone (FSH) levels were measured in Motol University Hospital’s accredited laboratory by a chemiluminescence immunoassay; the results are given in international units per liter. The intraindividual coefficient of variation for the kit was 11%. Statistical analyses The statistical computing environment R was used to conduct all of the statistical analyses (22–24). The data are reported as the means [standard deviation (SD)] unless specified otherwise. The height-specific Z-scores for the pQCT-derived bone parameters were established in the same way as we did previously for the muscle functions (25) using Cole’s LMS method (26): in summary, parameters L (the power in the Box-Cox transformation), M (the median), and S (the generalized coefficient of variation) are fitted in the formula to transform the raw data (X) into Z-scores (Z): Z = ln(X/M)/S if L = 0 and Z = (X/M − 1)/S if L = 1. The mean Z-scores were tested for the difference from zero using the one-sample t test. To evaluate the age-dependent changes in individual pQCT-derived bone parameters, a cubic regression model with a random intercept was used for each variable. More precisely, the formula for jth observation of ith girl is Yij=β0+β1(ageij−12)−β2(ageij−12)2+β3(ageij−12)3+ηi+εij;where ηi ∼ N(0, ση2) is the random effect of the ith subject, and εij ∼ N(0, σ2) is the random error. Pointwise confidence intervals for the population means are based on the asymptotic normality of the maximum likelihood estimators. The same model was used to assess the impact of the hormonal treatment and fracture history on the bone parameters by adding the duration of GH treatment and estrogen substitution, respectively, or positive fracture history as the next explanatory variable. The FSH serum concentration at the study end was also added as an independent explanatory variable. Age-stratified national annual fracture incidence in healthy girls and female adolescents aged 0 to 18 years was estimated as a ratio of the incidence of hospitalized healthy girls with a fracture in the country to the percentage of healthy girls with a fracture that were hospitalized within Motol University Hospital. First, we obtained a downward-biased estimate of the age-dependent, 6-year incidence rate for each healthy girl in our dataset by determining the maximum of the corresponding fracture rates. The estimated 6-year fracture incidence in the healthy population of the same age structure (which is representative of our group of girls with TS) was then compared with the observed 6-year fracture incidence in TS, with the P value calculated by the Monte Carlo simulation. Results The mean basic anthropometric measures and the selection of the treatment-related clinical data of the girls with TS are shown in Table 1. Whereas the body height of the patients was decreased compared with the healthy controls, regardless of their pubertal stage, their forearms were even shorter when comparing them with those of healthy girls of the same height. There was no difference in the age of the recombinant human GH treatment initiation [P = 0.156, analysis of variance (ANOVA) test], estrogen substitution initiation (P = 0.366, ANOVA test), or age at menarche among the three groups (P = 0.196, ANOVA test). Table 1. Basic Characteristics of the Girls With TS Prepubertal Girls (n = 15) Pubertal Girls (n = 14) Postmenarcheal Girls (n = 3) Study Start Study End Study Start Study End Study Start Study End Anthropometric data Age, y 10.0 (2.2) 15.9 (2.3) 13.5 (1.5) 18.9 (1.1) 16.1 (0.4) 21.9 (0.5) Height, cm 132.8 (13.4) 155.6 (5.9) 147.7 (8.7) 156.7 (4.5) 157.7 (1.8) 158.3 (2.5) Height Z-score −1.4 (0.9)a −1.4 (0.8)a −1.8 (0.6)a −1.7 (0.7)a −1.4 (0.3)b −1.4 (0.4)b Weight, kg 33.5 (10.4) 54.3 (8.5) 47.3 (10.1) 58.6 (4.2) 58.5 (6.6) 57.9 (3.5) Weight Z-score −0.4 (0.9) −0.2 (0.7) −0.3 (0.6) −0.1 (0.5) 0.1 (0.8) −0.2 (0.4) BMI, kg/m2 18.5 (3.0) 22.4 (3.0) 21.4 (2.7) 23.9 (1.9) 23.5 (2.1) 23.1 (1.5) BMI Z-score 0.3 (0.9) 0.7 (1.0)c 0.7 (0.8)c 1.0 (0.7)c 1.1 (0.8) 0.7 (0.5) Forearm length, mm 190.7 (21.9) 230.6 (10.0) 215.7 (20.9) 228.9 (12.9) 223.3 (10.4) 229.3 (6.0) Forearm length height Z-score −1.8 (1.0)a −1.0 (1.0)c −1.5 (1.2)a −1.5 (1.0)c −2.3 (1.0) −1.7 (0.6)b Clinical data rhGH therapy start, age, y 5.6 (2.6) 7.6 (3.5) 5.0 (1.3) rhGH therapy duration, y 4.4 (3.0) 5.9 (3.7) 11.1 (1.7) Estrogen therapy start, age, y 12.0 (0.9) 12.4 (1.2) 12.9 (0.2) Estrogen therapy duration, y 5.0 (1.4) 7.3 (0.9) 8.9 (0.7) Menarche, age, y 14.2 (1.2) 15.0 (1.4) 15.5 (0.4) Prepubertal Girls (n = 15) Pubertal Girls (n = 14) Postmenarcheal Girls (n = 3) Study Start Study End Study Start Study End Study Start Study End Anthropometric data Age, y 10.0 (2.2) 15.9 (2.3) 13.5 (1.5) 18.9 (1.1) 16.1 (0.4) 21.9 (0.5) Height, cm 132.8 (13.4) 155.6 (5.9) 147.7 (8.7) 156.7 (4.5) 157.7 (1.8) 158.3 (2.5) Height Z-score −1.4 (0.9)a −1.4 (0.8)a −1.8 (0.6)a −1.7 (0.7)a −1.4 (0.3)b −1.4 (0.4)b Weight, kg 33.5 (10.4) 54.3 (8.5) 47.3 (10.1) 58.6 (4.2) 58.5 (6.6) 57.9 (3.5) Weight Z-score −0.4 (0.9) −0.2 (0.7) −0.3 (0.6) −0.1 (0.5) 0.1 (0.8) −0.2 (0.4) BMI, kg/m2 18.5 (3.0) 22.4 (3.0) 21.4 (2.7) 23.9 (1.9) 23.5 (2.1) 23.1 (1.5) BMI Z-score 0.3 (0.9) 0.7 (1.0)c 0.7 (0.8)c 1.0 (0.7)c 1.1 (0.8) 0.7 (0.5) Forearm length, mm 190.7 (21.9) 230.6 (10.0) 215.7 (20.9) 228.9 (12.9) 223.3 (10.4) 229.3 (6.0) Forearm length height Z-score −1.8 (1.0)a −1.0 (1.0)c −1.5 (1.2)a −1.5 (1.0)c −2.3 (1.0) −1.7 (0.6)b Clinical data rhGH therapy start, age, y 5.6 (2.6) 7.6 (3.5) 5.0 (1.3) rhGH therapy duration, y 4.4 (3.0) 5.9 (3.7) 11.1 (1.7) Estrogen therapy start, age, y 12.0 (0.9) 12.4 (1.2) 12.9 (0.2) Estrogen therapy duration, y 5.0 (1.4) 7.3 (0.9) 8.9 (0.7) Menarche, age, y 14.2 (1.2) 15.0 (1.4) 15.5 (0.4) The data are means (SD). The footnotes indicate the difference from zero, tested by one-sample t test. Abbreviations: BMI, body mass index; rhGH, recombinant human GH. a P < 0.001. b P < 0.05. c P < 0.01. View Large Comparison of the fracture rate in girls with TS and healthy girls There were three of 32 girls with TS who had a new, clinically significant fracture during the 6-year observation study. One girl broke her left tibia while falling down the stairs (at age 14.5 years), another girl sustained a left forearm fracture during skating (at age 12.8 years), and the third girl had a fracture of the right humerus (fall on ice, age 11.5 years) and left distal forearm (fall from a bike, age 11.9 years). The humerus and forearm fractures affected the shafts of the bones. On the national level, there were between 43.509 and 66.342 girls (mean ± SD 51.475 ± 6.829 girls) for every 1-year category (0 to 20 years) and every year (2008 to 2014). The number of girls hospitalized with either a limb or spine fracture varied between 1756 and 2122 per year (mean ± SD 1953 ± 130 girls per year). The most frequent diagnosis was a forearm fracture. The national incidence of hospitalized girls with a limb or spine fracture was calculated for every year of observation and then averaged. More than 9000 girls (9263), aged 0 to 18 years, attended Motol University Hospital because of the limb or spine fracture between 2008 and 2014, and only 354 (3.8%) were hospitalized. The ratio of the national incidence of hospitalized girls with a fracture to a single-center proportion of hospitalized girls with a fracture was calculated to estimate the specific fracture rate in “otherwise healthy” girls on the country level. The peak incidence occurred at 11 years and reached 12.5% (Fig. 1). From these data, we estimated the downward-biased, age-dependent, 6-year fracture rate for each healthy girl in our dataset by determining the maximum of the corresponding annual fracture rates. Then, in a subgroup of all healthy girls of the same age as that of our girls with TS, we report the probability of one to 10 girls with fractures (Table 2). As the fracture rate in our girls with TS was not significantly higher compared with the downward-biased fracture rate estimate in healthy girls (P = 0.48), it cannot be higher than the real fracture incidence in healthy girls. Figure 1. View largeDownload slide Country incidence of limb and spine fractures among otherwise healthy girls. Broken lines represent the individual years from 2008 to 2014; the solid line is the average. The vertical lines represent the ages at which the fractures occurred in our girls with TS (followed, on average, from 12.1 to 17.6 years of age). Figure 1. View largeDownload slide Country incidence of limb and spine fractures among otherwise healthy girls. Broken lines represent the individual years from 2008 to 2014; the solid line is the average. The vertical lines represent the ages at which the fractures occurred in our girls with TS (followed, on average, from 12.1 to 17.6 years of age). Table 2. The Likelihood of Fractures in Girls With TS Number of Fractured Girls P value 0 0.06754 1 0.19040 2 0.26114 3 0.23146 4 0.14494 5 0.06866 6 0.02572 7 0.00774 8 0.00182 9 0.00048 10 0.00010 Number of Fractured Girls P value 0 0.06754 1 0.19040 2 0.26114 3 0.23146 4 0.14494 5 0.06866 6 0.02572 7 0.00774 8 0.00182 9 0.00048 10 0.00010 The estimated probabilities of a particular number of TS girls (out of 32) with incident fractures throughout the 6-year observation period based on the Monte Carlo simulation using the national data from control age- and sex-matched population. View Large The development of bone parameters during growth in TS The pQCT-derived bone density and strength parameters at the study beginning and end time points are summarized in Table 3. In addition to the raw data and height-specific Z-scores, we report the 95% confidence intervals for the least significant changes of each pQCT measure. The age-specific, cubic regression model-based TS population means with pointwise confidence intervals for height Z-scores for forearm length and selected pQCT parameters are depicted in Figs. 2 and 3 , respectively. These figures illustrate the estimated development of anthropometric and bone parameters in girls with TS, as calculated from our individual longitudinal data collection. Even though trabecular vBMD was normal before puberty, there was a clear decrease during pubertal development (the Z-score change with age estimate −0.21 ± 0.04, P < 0.001). In contrast, total bone CSA correspondingly increased (0.16 ± 0.04, P < 0.001), which led to normal bone SSI (Table 3; Fig. 3). Partial volume-adjusted cortical vBMD slightly decreased before pubertal induction and increased thereafter (Fig. 3). Therefore, there was no substantial change with age throughout the observation period (0.03 ± 0.05, P = 0.54). Nevertheless, total bone CSA increased with age (0.13 ± 0.04, P < 0.01), which was probably a consequence of the initial decrease in cortical vBMD, which led to a slight increase in bone SSI (0.10 ± 0.03, P < 0.001; Fig. 3). Cortical thickness was normal (Table 3), and no change with age was observed (−0.01 ± 0.03, P = 0.72). Table 3. pQCT Parameters in Girls With TS Prepubertal Girls (n = 15) Pubertal Girls (n = 14) Postmenarcheal Girls (n = 3) Precision Error Least Significant Change Study Start Study End Study Start Study End Study Start Study End (RMS SD) (95% CI) Radial metaphysis (4%) Total bone CSA, mm2 209 (51) 304 (58) 278 (60) 309 (23) 351 (59) 322 (47) 9.6 26.6 Total bone CSA Z-score 0.61 (1.3) 1.3 (1.4)a 1.3 (1.1)b 1.4 (0.76)b 2.2 (1.1) 1.5 (0.97) Trabecular vBMD, mg/cm3 186 (31) 152 (40) 165 (31) 155 (33) 149 (15) 133 (7.0) 2.9 8.0 Trabecular vBMD Z-score −0.11 (0.99) −1.5 (1.7)a −0.88 (1.1)c −1.3 (1.5)c −1.4 (0.65) −2.1 (0.33)a SSI polar, mm3 133 (33) 241 (53) 185 (73) 337 (91) 266 (138) 278 (93) 10 27.8 SSI polar Z-score 0.22 (0.68) 0.28 (0.67) −0.18 (1.6) 1.4 (1.4)c 0.10 (2.0) 0.46 (1.0) Radial diaphysis (65%) Total bone CSA, mm2 107 (27) 137 (29) 114 (19) 124 (18) 111 (27) 135 (24) 1.2 3.3 Total bone CSA Z-score 0.46 (1.2) 1.1 (1.4)a 0.27 (0.93) 0.48 (1.0) −0.48 (1.6) 0.89 (1.2) Cortical vBMD, mg/cm3 1057 (31) 1106 (41) 1084 (37) 1154 (36) 1155 (19) 1138 (9.0) 4.6 12.9 Cortical vBMD Z-score −0.09 (1.2) 0.0 (1.1) −0.15 (0.89) 1.3 (1.0)a 1.3 (0.63) 0.75 (0.39) Cortical thickness, mm 1.3 (0.32) 1.8 (0.28) 1.7 (0.31) 2.1 (0.32) 1.7 (0.22) 1.7 (0.24) 0.068 0.19 Cortical thickness Z-score −0.18 (1.2) −0.09 (0.85) −0.04 (0.90) 0.75 (1.1) −0.54 (0.59) −0.58 (0.59) Relative cortical bone CSA, mm2 39 (10) 49 (9.1) 48 (8) 57 (8.2) 50 (0.42) 46 (4.3) 0.80 2.2 Relative cortical bone CSA Z-score −0.51 (1.5) −0.45 (1.1) −0.16 (1.0) 0.52 (1.1) −0.32 (0.1)a −0.86 (0.48) SSI polar, mm3 140 (40) 248 (50) 185 (47) 243 (38) 199 (70) 250 (64) 7.7 21.3 SSI polar Z-score 0.43 (0.63)c 1.1 (1.0)a 0.36 (0.82) 0.93 (0.83)a −0.30 (1.7) 0.85 (1.3) Prepubertal Girls (n = 15) Pubertal Girls (n = 14) Postmenarcheal Girls (n = 3) Precision Error Least Significant Change Study Start Study End Study Start Study End Study Start Study End (RMS SD) (95% CI) Radial metaphysis (4%) Total bone CSA, mm2 209 (51) 304 (58) 278 (60) 309 (23) 351 (59) 322 (47) 9.6 26.6 Total bone CSA Z-score 0.61 (1.3) 1.3 (1.4)a 1.3 (1.1)b 1.4 (0.76)b 2.2 (1.1) 1.5 (0.97) Trabecular vBMD, mg/cm3 186 (31) 152 (40) 165 (31) 155 (33) 149 (15) 133 (7.0) 2.9 8.0 Trabecular vBMD Z-score −0.11 (0.99) −1.5 (1.7)a −0.88 (1.1)c −1.3 (1.5)c −1.4 (0.65) −2.1 (0.33)a SSI polar, mm3 133 (33) 241 (53) 185 (73) 337 (91) 266 (138) 278 (93) 10 27.8 SSI polar Z-score 0.22 (0.68) 0.28 (0.67) −0.18 (1.6) 1.4 (1.4)c 0.10 (2.0) 0.46 (1.0) Radial diaphysis (65%) Total bone CSA, mm2 107 (27) 137 (29) 114 (19) 124 (18) 111 (27) 135 (24) 1.2 3.3 Total bone CSA Z-score 0.46 (1.2) 1.1 (1.4)a 0.27 (0.93) 0.48 (1.0) −0.48 (1.6) 0.89 (1.2) Cortical vBMD, mg/cm3 1057 (31) 1106 (41) 1084 (37) 1154 (36) 1155 (19) 1138 (9.0) 4.6 12.9 Cortical vBMD Z-score −0.09 (1.2) 0.0 (1.1) −0.15 (0.89) 1.3 (1.0)a 1.3 (0.63) 0.75 (0.39) Cortical thickness, mm 1.3 (0.32) 1.8 (0.28) 1.7 (0.31) 2.1 (0.32) 1.7 (0.22) 1.7 (0.24) 0.068 0.19 Cortical thickness Z-score −0.18 (1.2) −0.09 (0.85) −0.04 (0.90) 0.75 (1.1) −0.54 (0.59) −0.58 (0.59) Relative cortical bone CSA, mm2 39 (10) 49 (9.1) 48 (8) 57 (8.2) 50 (0.42) 46 (4.3) 0.80 2.2 Relative cortical bone CSA Z-score −0.51 (1.5) −0.45 (1.1) −0.16 (1.0) 0.52 (1.1) −0.32 (0.1)a −0.86 (0.48) SSI polar, mm3 140 (40) 248 (50) 185 (47) 243 (38) 199 (70) 250 (64) 7.7 21.3 SSI polar Z-score 0.43 (0.63)c 1.1 (1.0)a 0.36 (0.82) 0.93 (0.83)a −0.30 (1.7) 0.85 (1.3) Data are means (SD). The footnotes indicate the difference from zero tested by one-sample t test. Abbreviations: CI, confidence interval; RMS, root mean square. a P < 0.01. b P < 0.001. c P < 0.05. View Large Figure 2. View largeDownload slide (A) Height and (B) forearm length development in TS patients. Pointwise confidence intervals for the means were estimated from the development of bone parameters in individual girls with TS. Height-specific Z-scores are shown for forearm length. Figure 2. View largeDownload slide (A) Height and (B) forearm length development in TS patients. Pointwise confidence intervals for the means were estimated from the development of bone parameters in individual girls with TS. Height-specific Z-scores are shown for forearm length. Figure 3. View largeDownload slide Bone density and strength parameters’ development in girls with TS. All measures were assessed at the radius by pQCT. (A) Metaphyseal trabecular vBMD, (B) total bone CSA, (C) bone SSI, (D) diaphyseal cortical vBMD, (E) total bone CSA, and (F) bone SSI are shown. Pointwise confidence intervals for the mean were estimated from the development of bone parameters in individual girls with TS. Height-specific Z-scores are shown for all bone parameters. Figure 3. View largeDownload slide Bone density and strength parameters’ development in girls with TS. All measures were assessed at the radius by pQCT. (A) Metaphyseal trabecular vBMD, (B) total bone CSA, (C) bone SSI, (D) diaphyseal cortical vBMD, (E) total bone CSA, and (F) bone SSI are shown. Pointwise confidence intervals for the mean were estimated from the development of bone parameters in individual girls with TS. Height-specific Z-scores are shown for all bone parameters. The effects of hormonal treatment and fracture history on bone development in TS The forearm length Z-score increased with every year of estrogen substitution (β estimate 0.15 ± 0.068, P = 0.029). The cortical vBMD Z-score was higher with longer GH treatment duration (β estimate 0.11 ± 0.051, P = 0.035) and lower with a higher age at GH treatment commencement (β estimate −0.13 ± 0.050, P = 0.016). A positive history of incident fractures was not significantly associated with any of the pQCT-derived bone parameters (data not shown). The FSH serum concentration at study end was negatively associated with SSI Z-score at the metaphysis (β estimate −0.01 ± 0.01, P = 0.040) but not with any other pQCT bone measure. Discussion In this study, we show the controlled clinically significant fracture incidence and the longitudinal development of bone geometry and vBMD at the radius in a pediatric TS population using pQCT. We show that (1) fracture incidence in girls with TS treated with GH and substituted with estrogens is not higher compared with that in the healthy population, (2) trabecular vBMD was normal before puberty and decreased with age, but total bone CSA increased correspondingly and led to normal, estimated bone strength at the radial metaphysis, (3) cortical vBMD slightly decreased before puberty induction but increased thereafter, which together with an increase in total bone CSA, led to slightly increased, calculated bone strength at the radial diaphysis, and (4) a longer GH treatment duration was positively associated with a higher cortical vBMD. The first study mentioning the fracture incidence in TS was published in 1991 (27). The authors calculated that the annual incidence of fractures in girls with TS (aged 4 to 13 years) was not different from that in normal children younger than 15 years (19.9/100 vs 27.8/1000), but there was a substantially higher annual incidence of wrist fracture in TS (9.1/1000 vs 3.5/1000 in normal kids, P < 0.003). It is important to note that only 13/78 girls with TS from the study were treated either with estrogen (n = 13), GH (n = 12), and/or oxandrolone (n = 2) at a maximum duration of 6 months. In contrast, all of our study participants with TS were treated with GH, and all of those without spontaneous puberty were substituted with estrogens, which represents the current standard care for these patients. We found that fractures of the long bones of the limbs (and vertebral fractures) are not more common among girls with TS compared with healthy controls. Therefore, we may speculate that current standard therapy can prevent bone fragility in TS. More importantly, the study by Ross et al. (27) also showed that areal BMD of the wrist (and spine) was not different between the girls with TS and healthy controls after adjusting for height/age and that the BMD was similar between the girls with TS, with and without a wrist fracture. In accordance with this study, we could not find decreased bone strength at the radius in TS in our study or any difference in bone strength between those who had and who had not sustained a fracture. One possible explanation is that common bone densitometry techniques (dual energy X-ray absorptiometry and pQCT) do not generate reliable parameters for fracture prediction, as was described earlier (28, 29). However, factors other than bone quality can cause fractures in TS. In particular, we described low muscle power during a maximum voluntary, single two-leg jump test by mechanography (30), which is a parameter partially influenced by muscle coordination and controlled by the central nervous system. Muscle power is thus a surrogate of visual-spatial cognitive function of the brain that has been shown to be impaired in TS (31). Moreover, conductive hearing loss may be an additional reason for an increased fracture rate in TS (32). All of these studies show that the fracture risk assessment in TS needs a multifactorial evaluation. Normal trabecular vBMD before puberty and a decrease during and after pubertal development have been shown already in our initial cross-sectional study (11). This is in accordance with previous observations of a normal size-adjusted lumbar spine areal BMD before puberty and a lack of adequate pubertal bone mass acquisition in TS girls (27, 33). However, newer studies using pQCT showed normal trabecular vBMD in females with TS (10, 34). We speculated that a different estrogen dose or formula could be the reason for the inconsistent findings (11). Indeed, our girls used (on average) 1.2 mg of 17-β estradiol/day, whereas the participants in the study by Bechtold et al. (10) used estradiol valerate at 2.0 mg/day, and in the study by Holroyd et al. (34), it was ethinyl-estradiol (the daily dose could not be calculated in this case). It seems that there might be an upper-dose limit beyond which there is no additional effect on the skeleton, which was represented by 17-β estradiol, 2 mg/day (35). In contrast, the lower limit has not been studied yet. As regular and continuous estrogen replacement is needed to avoid osteoporosis and reduce increased fracture risk in TS (5, 36), appropriate timing, dosing, and monitoring of estrogen substitution are crucial to improve the quality of life in TS patients. Cortical vBMD showed biphasic development when it decreased before pubertal induction and increased thereafter. Looking over the entire observation period, cortical vBMD was not significantly changed compared with the reference. This contradicts previous studies (10, 34) but is in line with our methodological paper (21). Cortical vBMD is dependent on height (19) and cortical thickness (37). We addressed both of these issues by calculating the height-specific Z-scores and by adjusting cortical vBMD to partial-volume effect [related to thin cortices and described in Rittweger et al. (20)]. These adjustments were not made in previous papers and are thus a possible reason for dissimilar findings. With the search for potential predictors, only GH treatment duration was significantly associated with cortical vBMD. This is in accordance with our initial cross-sectional study and with papers on acromegaly (38) but not with studies of TS (39). Whether this is a real causative relation or a coincidence still needs to be elucidated. Interestingly, the total bone CSA at both radial sites increased, whereas trabecular vBMD (overall) and cortical vBMD (before pubertal induction) decreased, which might be understood as a physiological adaptation of the bone to withstand mechanical loads. Similar findings were shown by Bechtold et al. (10). This process is especially effective during growth (40). Indeed, the calculated bone SSI was normal at the metaphysis and slightly increased at the diaphysis in the girls with TS in our study. Even though trabecular vBMD remained low until postmenarcheal age in TS, cortical vBMD increased from previously low values. This could cause the difference in SSI between the two sites, as a larger bone that has been built already cannot be resorbed on the periosteal site. Normal bone strength in the girls with TS in our study would match with one of our most important findings—that the 6-year fracture incidence in girls with TS was not higher compared with the fracture rate in age-, country-, and time-matched healthy controls. Limitations It would be more appropriate to follow control girls of the same age compared with the girls with TS over the 6-year period to compare directly the fracture incidence between the groups. However, it would only be possible in a limited group of participants with a risk of selection bias. In contrast, our approach was to estimate the 6-year fracture incidence from the national data that were available. As we could not track individual data within the national registry, we estimated the 6-year fracture incidence as a maximum of 6 consecutive years for girls of the same age as the girls with TS. Therefore, it is very probable that the real fracture incidence is even higher in the normal population; thus, it is unlikely that the girls with TS in our study had a higher fracture rate compared with that in healthy girls. Likewise, the development of pQCT-derived bone parameters was estimated from the change of Z-scores calculated from the cross-sectional normative dataset and was not based on parallel parameter tracking in individual controls. Nevertheless, reference data were obtained for the neighboring population with similar ethnicity and health, social, and environmental conditions. Thus, we believe that it is acceptable to assume that the two populations are interchangeable. The results of this study are also limited by the age of the participants. We do not infer that the fracture risk is not increased in adults with TS despite the fact that it has been shown that long-term regular estrogen use prevents low BMD and osteoporosis in TS (36). Whether appropriate continuous estrogen substitution also leads to a reduction of fracture risk in adults with TS still needs to be elucidated. Conclusions This study shows that the 6-year fracture incidence is not increased among girls with TS who were treated with GH during growth and substituted with estrogens (when lacking spontaneous puberty) compared with that in normal girls. Even though trabecular vBMD decreased during puberty, total bone CSA increased and led to normal calculated bone SSI at the metaphysis of the radius, which supports the finding of a “common” fracture rate in TS. Abbreviations: ANOVA analysis of variance BMD bone mineral density CSA cross-sectional area FSH follicle-stimulating hormone GH growth hormone pQCT peripheral quantitative computerized tomography SD standard deviation SSI strength-strain index TS Turner syndrome vBMD volumetric bone mineral density. Acknowledgments We thank Marta Snajderova and Stanislava Kolouskova for recruiting their patients. We also acknowledge all study participants and their families. Financial Support: This work was partially supported by the Ministry of Health, Czech Republic (Project for conceptual development of Research Organization 00064203, Motol University Hospital). Author Contributions: O.S. and Z.S. designed the study and drafted the manuscript. O.S. performed the pQCT measurements, evaluated the results, and established the national fracture incidence data. Z.H. performed the statistics and prepared the figures. J.L. recruited the majority of the TS patients and conducted the clinical care. E.S. and J.W. provided the raw pQCT data of the reference population and calculated additional bone strength indices. All authors reviewed the manuscript. Disclosure Summary: The authors have nothing to disclose. References 1. Davenport ML. Approach to the patient with Turner syndrome. J Clin Endocrinol Metab . 2010; 95( 4): 1487– 1495. Google Scholar CrossRef Search ADS PubMed 2. Schoemaker MJ, Swerdlow AJ, Higgins CD, Wright AF, Jacobs PA; United Kingdom Clinical Cytogenetics Group. Mortality in women with Turner syndrome in Great Britain: a national cohort study. J Clin Endocrinol Metab . 2008; 93( 12): 4735– 4742. Google Scholar CrossRef Search ADS PubMed 3. Stochholm K, Juul S, Juel K, Naeraa RW, Gravholt CH. Prevalence, incidence, diagnostic delay, and mortality in Turner syndrome. J Clin Endocrinol Metab . 2006; 91( 10): 3897– 3902. Google Scholar CrossRef Search ADS PubMed 4. Gravholt CH, Juul S, Naeraa RW, Hansen J. Morbidity in Turner syndrome. J Clin Epidemiol . 1998; 51( 2): 147– 158. Google Scholar CrossRef Search ADS PubMed 5. Gravholt CH, Vestergaard P, Hermann AP, Mosekilde L, Brixen K, Christiansen JS. Increased fracture rates in Turner’s syndrome: a nationwide questionnaire survey. Clin Endocrinol (Oxf) . 2003; 59( 1): 89– 96. Google Scholar CrossRef Search ADS PubMed 6. Costa AM, Lemos-Marini SH, Baptista MT, Morcillo AM, Maciel-Guerra AT, Guerra G, Jr. Bone mineralization in Turner syndrome: a transverse study of the determinant factors in 58 patients. J Bone Miner Metab . 2002; 20( 5): 294– 297. Google Scholar CrossRef Search ADS PubMed 7. Carrascosa A, Gussinyé M, Terradas P, Yeste D, Audí L, Vicens-Calvet E. Spontaneous, but not induced, puberty permits adequate bone mass acquisition in adolescent Turner syndrome patients. J Bone Miner Res . 2000; 15( 10): 2005– 2010. Google Scholar CrossRef Search ADS PubMed 8. Bertelloni S, Cinquanta L, Baroncelli GI, Simi P, Rossi S, Saggese G. Volumetric bone mineral density in young women with Turner’s syndrome treated with estrogens or estrogens plus growth hormone. Horm Res . 2000; 53( 2): 72– 76. Google Scholar PubMed 9. Bakalov VK, Chen ML, Baron J, Hanton LB, Reynolds JC, Stratakis CA, Axelrod LE, Bondy CA. Bone mineral density and fractures in Turner syndrome. Am J Med . 2003; 115( 4): 259– 264. Google Scholar CrossRef Search ADS PubMed 10. Bechtold S, Rauch F, Noelle V, Donhauser S, Neu CM, Schoenau E, Schwarz HP. Musculoskeletal analyses of the forearm in young women with Turner syndrome: a study using peripheral quantitative computed tomography. J Clin Endocrinol Metab . 2001; 86( 12): 5819– 5823. Google Scholar CrossRef Search ADS PubMed 11. Soucek O, Lebl J, Snajderova M, Kolouskova S, Rocek M, Hlavka Z, Cinek O, Rittweger J, Sumnik Z. Bone geometry and volumetric bone mineral density in girls with Turner syndrome of different pubertal stages. Clin Endocrinol (Oxf) . 2011; 74( 4): 445– 452. Google Scholar CrossRef Search ADS PubMed 12. Soucek O, Zapletalova J, Zemkova D, Snajderova M, Novotna D, Hirschfeldova K, Plasilova I, Kolouskova S, Rocek M, Hlavka Z, Lebl J, Sumnik Z. Prepubertal girls with Turner syndrome and children with isolated SHOX deficiency have similar bone geometry at the radius. J Clin Endocrinol Metab . 2013; 98( 7): E1241– E1247. Google Scholar CrossRef Search ADS PubMed 13. Kobzová J, Vignerová J, Bláha P, Krejcovský L, Riedlová J. The 6th nationwide anthropological survey of children and adolescents in the Czech Republic in 2001. Cent Eur J Public Health . 2004; 12( 3): 126– 130. Google Scholar PubMed 14. Neu CM, Rauch F, Manz F, Schoenau E. Modeling of cross-sectional bone size, mass and geometry at the proximal radius: a study of normal bone development using peripheral quantitative computed tomography. Osteoporos Int . 2001; 12( 7): 538– 547. Google Scholar CrossRef Search ADS PubMed 15. Saenger P, Wikland KA, Conway GS, Davenport M, Gravholt CH, Hintz R, Hovatta O, Hultcrantz M, Landin-Wilhelmsen K, Lin A, Lippe B, Pasquino AM, Ranke MB, Rosenfeld R, Silberbach M; Fifth International Symposium on Turner Syndrome. Recommendations for the diagnosis and management of Turner syndrome. J Clin Endocrinol Metab . 2001; 86( 7): 3061– 3069. Google Scholar PubMed 16. Bianchi ML, Leonard MB, Bechtold S, Högler W, Mughal MZ, Schönau E, Sylvester FA, Vogiatzi M, van den Heuvel-Eibrink MM, Ward L; International Society for Clinical Densitometry. Bone health in children and adolescents with chronic diseases that may affect the skeleton: the 2013 ISCD Pediatric Official Positions. J Clin Densitom . 2014; 17( 2): 281– 294. Google Scholar CrossRef Search ADS PubMed 17. Kroke A, Manz F, Kersting M, Remer T, Sichert-Hellert W, Alexy U, Lentze MJ. The DONALD Study. History, current status and future perspectives. Eur J Nutr . 2004; 43( 1): 45– 54. Google Scholar CrossRef Search ADS PubMed 18. Rauch F, Schöenau E. Peripheral quantitative computed tomography of the distal radius in young subjects - new reference data and interpretation of results. J Musculoskelet Neuronal Interact . 2005; 5( 2): 119– 126. Google Scholar PubMed 19. Rauch F, Schoenau E. Peripheral quantitative computed tomography of the proximal radius in young subjects--new reference data and interpretation of results. J Musculoskelet Neuronal Interact . 2008; 8( 3): 217– 226. Google Scholar PubMed 20. Rittweger J, Michaelis I, Giehl M, Wüsecke P, Felsenberg D. Adjusting for the partial volume effect in cortical bone analyses of pQCT images. J Musculoskelet Neuronal Interact . 2004; 4( 4): 436– 441. Google Scholar PubMed 21. Soucek O, Schönau E, Lebl J, Sumnik Z. Artificially low cortical bone mineral density in Turner syndrome is due to the partial volume effect. Osteoporos Int . 2015; 26( 3): 1213– 1218. Google Scholar CrossRef Search ADS PubMed 22. R Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing; 2016. 23. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team. nlme: Linear and Nonlinear Mixed Effects Models. R Package, Version 3.1-128, 2016. Available at: https://CRAN.R-project.org/package=nlme. Accessed January 2016. 24. Bowman AW, Azzalini A. R Package 'sm': Nonparametric Smoothing Methods, Version 2.2-5.4, 2014. Available at: http://www.stats.gla.ac.uk/~adrian/sm, http://azzalini.stat.unipd.it/Book_sm. Accessed January 2016. 25. Sumnik Z, Matyskova J, Hlavka Z, Durdilova L, Soucek O, Zemkova D. Reference data for jumping mechanography in healthy children and adolescents aged 6-18 years. J Musculoskelet Neuronal Interact . 2013; 13( 3): 297– 311. Google Scholar PubMed 26. Cole TJ. The LMS method for constructing normalized growth standards. Eur J Clin Nutr . 1990; 44( 1): 45– 60. Google Scholar PubMed 27. Ross JL, Long LM, Feuillan P, Cassorla F, Cutler GB, Jr. Normal bone density of the wrist and spine and increased wrist fractures in girls with Turner’s syndrome. J Clin Endocrinol Metab . 1991; 73( 2): 355– 359. Google Scholar CrossRef Search ADS PubMed 28. Clark EM, Ness AR, Bishop NJ, Tobias JH. Association between bone mass and fractures in children: a prospective cohort study. J Bone Miner Res . 2006; 21( 9): 1489– 1495. Google Scholar CrossRef Search ADS PubMed 29. Kalkwarf HJ, Laor T, Bean JA. Fracture risk in children with a forearm injury is associated with volumetric bone density and cortical area (by peripheral QCT) and areal bone density (by DXA). Osteoporos Int . 2011; 22( 2): 607– 616. Google Scholar CrossRef Search ADS PubMed 30. Soucek O, Lebl J, Matyskova J, Snajderova M, Kolouskova S, Pruhova S, Hlavka Z, Sumnik Z. Muscle function in Turner syndrome: normal force but decreased power. Clin Endocrinol (Oxf) . 2015; 82( 2): 248– 253. Google Scholar CrossRef Search ADS PubMed 31. Ross JL, Stefanatos GA, Kushner H, Bondy C, Nelson L, Zinn A, Roeltgen D. The effect of genetic differences and ovarian failure: intact cognitive function in adult women with premature ovarian failure versus Turner syndrome. J Clin Endocrinol Metab . 2004; 89( 4): 1817– 1822. Google Scholar CrossRef Search ADS PubMed 32. Han TS, Cadge B, Conway GS. Hearing impairment and low bone mineral density increase the risk of bone fractures in women with Turner’s syndrome. Clin Endocrinol (Oxf) . 2006; 65( 5): 643– 647. Google Scholar CrossRef Search ADS PubMed 33. Shaw NJ, Rehan VK, Husain S, Marshall T, Smith CS. Bone mineral density in Turner’s syndrome--a longitudinal study. Clin Endocrinol (Oxf) . 1997; 47( 3): 367– 370. Google Scholar CrossRef Search ADS PubMed 34. Holroyd CR, Davies JH, Taylor P, Jameson K, Rivett C, Cooper C, Dennison EM. Reduced cortical bone density with normal trabecular bone density in girls with Turner syndrome. Osteoporos Int . 2010; 21( 12): 2093– 2099. Google Scholar CrossRef Search ADS PubMed 35. Cleemann L, Holm K, Kobbernagel H, Kristensen B, Skouby SO, Jensen AK, Gravholt CH. Dosage of estradiol, bone and body composition in Turner syndrome: a 5-year randomized controlled clinical trial. Eur J Endocrinol . 2017; 176( 2): 233– 242. Google Scholar CrossRef Search ADS PubMed 36. Hanton L, Axelrod L, Bakalov V, Bondy CA. The importance of estrogen replacement in young women with Turner syndrome. J Womens Health (Larchmt) . 2003; 12( 10): 971– 977. Google Scholar CrossRef Search ADS PubMed 37. Schoenau E, Neu CM, Rauch F, Manz F. Gender-specific pubertal changes in volumetric cortical bone mineral density at the proximal radius. Bone . 2002; 31( 1): 110– 113. Google Scholar CrossRef Search ADS PubMed 38. Andreassen TT, Oxlund H. The effects of growth hormone on cortical and cancellous bone. J Musculoskelet Neuronal Interact . 2001; 2( 1): 49– 58. Google Scholar PubMed 39. Ari M, Bakalov VK, Hill S, Bondy CA. The effects of growth hormone treatment on bone mineral density and body composition in girls with Turner syndrome. J Clin Endocrinol Metab . 2006; 91( 11): 4302– 4305. Google Scholar CrossRef Search ADS PubMed 40. Ducher G, Bass SL, Saxon L, Daly RM. Effects of repetitive loading on the growth-induced changes in bone mass and cortical bone geometry: a 12-month study in pre/peri- and postmenarcheal tennis players. J Bone Miner Res . 2011; 26( 6): 1321– 1329. Google Scholar CrossRef Search ADS PubMed Copyright © 2018 Endocrine Society
Journal of Clinical Endocrinology and Metabolism – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.
Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.
It’s easy to organize your research with our built-in tools.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera