Phalangeal ultrasonography in forearm fracture discrimination

Biomed Pharmacother 56 (2002) 332–338 www.elsevier.com/locate/biopha

Original article

Phalangeal ultrasonography in forearm fracture discrimination R. Giardino a,b,*, R. Rotini c, F. Noia c, C.A. Veronesi c, A. Carpi d, A. Nicolini e, F. de Terlizzi f, M. Fini a, G. Giavaresi a a

Servizio di chirurgia sperimentale, istituto di ricerca Codivilla-Putti, istituti ortopedici Rizzoli, via di Barbiano, 1/10, 40136 Bologna, Italy b Cattedra di fisiopatologia chirurgica, università di Bologna, Bologna, Italy c Modulo dipartimentale Spalla e Gomito, istituto ortopedico Rizzoli, via Pupilli, 1-40136 Bologna, Italy d Dipartimento di medicina della procreazione e dell’Età Evolutiva, università di Pisa, Italy e Dipartimento di medicina interna, università di Pisa, Italy f IGEA Srl, laboratory of clinical biophysics, Carpi–Modena, Italy

Abstract Over the last decade, the use of ultrasounds has been developed into an effective tool for investigating bone tissue and predicting the risk of fracture in osteoporosis. Studies have focused on hip and vertebral fractures while no information is available on the use of phalangeal ultrasonography to identify patients with forearm fractures. Thus, the current authors decided to compare 50 postmenopausal women with low energy forearm fractures (Fractured Group) with a control age-matched group of 94 women (Control Group). Measurements were taken at the distal metaphysis of the proximal phalanxes of the hand of the non-fractured arm using the DBM Sonic Bone Profiler. The reproducibility of the method was assessed by amplitude-dependent speed of sound (AD-SoS) CV% = 0.64 and by Ultrasound Bone Profiler Index (UBPI) CV% = 2.38. In the Control Group, the AD-SoS and UBPI mean values and standard deviations were significantly higher compared to the group with fractures (P < 0.0005). The receiver operating characteristic (ROC) curves were calculated and the areas under the curve (AUC) were 0.78 ± 0.04 for AD-SoS and 0.77 ± 0.05 for UBPI, respectively. Logistic regression analysis adjusted to age revealed that both AD-SoS (78.2%, ORAD-SoS = 12.03, P < 0.0005) and UBPI (76.0%, ORAD-SoS = 7.39, P < 0.0005) parameters discriminated correctly between fractured and non-fractured control women whereas the association of both parameters could not allow better discrimination. The present results showed that ultrasound investigation at the phalanxes is reproducible and efficiently discriminates between subjects with forearm fractures and those in the control subjects. © 2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. Keywords: Forearm fracture; Osteoporosis; Quantitative ultrasound

1. Introduction Osteoporosis is defined as a systemic disease of the skeleton characterised by decreased bone mass and deteriorated architecture of the bone tissue, resulting in increased fragility and fracture risk. The typical fractures induced by osteoporosis can be found at the level of the hip, vertebrae and forearm. Currently, diagnosis of osteoporosis and determination of the postmenopausal female population at the highest risk of fracture are mostly performed by Dual X-ray absorptiometry (DXA) [15,28]. Nevertheless, it has been * Corresponding author. E-mail address: [email protected] (R. Giardino). © 2002 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. PII: S 0 7 5 3 - 3 3 2 2 ( 0 2 ) 0 0 2 4 0 - 8

shown that fracture in osteoporosis has been shown to depend not only, and not exclusively, on bone mineral density (BMD), but also on other factors, such as bone architecture and elasticity. Many studies have indeed demonstrated how these factors are influential in determining a fracture irrespective of BMD g/cm2 [16,21]. These observations have oriented research towards the quantitative ultrasound (QUS) method, which is potentially able to provide information on the elasticity and microarchitecture of bone tissue [12,17,22]. Great varieties of QUS apparatuses have been developed over the last decade to study bone tissue in different skeletal areas: calcaneus (trabecular), tibia (cortical), phalanx (cortical/trabecular), patella (trabecular), and radius (corti-

R. Giardino et al. / Biomed Pharmacother 56 (2002) 332–338

cal) [11]. These types of measurement have been studied exhaustively and have proved to be very accurate. They have good diagnostic sensitivity and are able to assess precisely the risk of fracture in osteoporosis [2,8,10,19,23,27,30]. Several studies have already tested the ability of phalangeal ultrasonography to discriminate between vertebral or hip fractured subjects and controls [1,9,14,23,30], in comparison to BMD measured with DXA at the same sites. In addition, phalangeal ultrasonography seems to be able to monitor bone loss in other metabolic bone diseases, such as secondary osteoporosis, osteomalacia, primary hyperparathyroidism, endo- and exogenous glucorticoid excess [18,20,25] or monitor the effect of therapies for osteoporosis, such as hormonal agents, calcitonin, biphosphonates, and selective oestrogen-receptor modulators [6]. Recently, Montagnani et al. [20] have found that ultrasound (US) signal parameters are differently influenced by bone structural changes, occurring in osteoporosis or primary hyperparathyroidism, pathologies characterised both by bone loss with a different histologic pattern. Nevertheless, information on forearm-fractured patients is lacking. To current authors’ knowledge, only Wu et al. [29] compared phalangeal ultrasonography with DXA and peripheral quantitative computed tomography measurements of the forearm in simulated Colles’ fracture in cadaveric forearms. Their results showed that phalangeal ultrasonography is comparable to forearm densitometry and enhanced the prediction of the strength of the distal radius [29]. In the current study, the authors compared the results of QUS measurements taken on women suffering from osteoporotic fracture of the forearm, with measurements collected in a group of women who responded to a public radio call.

2. Materials and methods The study included 170 postmenopausal women aged over 45 years (mean age 63 ± 7 years) who had undergone phalanx US measurements at the Rizzoli Orthopaedic Institute from September 1998 to June 1999. The ethics committee of Rizzoli Orthopaedic Institute approved the study and women were asked to sign an informed consent to participate in the study. The women were divided into two groups: (a) those with fracture of the forearm (Fractured Group) and (b) those without fracture (Control Group). Women with forearm fractures (n = 50 subjects aged 53–78 years) were recruited from consecutive inpatients with forearm fractures treated at the casualty department of the Institute. Only fractures due to minor trauma, such as falls from a standing position or walking, were considered. The Control Group was made up of 120 women (age range

333

45–82 years) who responded to a public radio call (September 1998). Careful anamnesis by questionnaire enabled precise characterisation of the population of the study: subjects having systemic conditions known to affect bone metabolism (malignancy, paralysis, long-term immobilisation, major rheumatic diseases, chronic renal, hepatic or pulmonary diseases, Paget’s disease of bone) were excluded from the outset. None of the fractured patients was under treatment for more than 6 months with oestrogen, anabolic steroids, calcitonin, bisphosphonates or more than 3 months with glucocorticoids, or antiepileptic drugs. All US measurements were made with the DBM Sonic Bone Profiler (IGEA, Carpi, Italy). The principle of functioning has already been described elsewhere [30]. Briefly, it measures the speed of propagation of a US impulse in the bone tissue, using one transmitting and one receiving probe at the distal metaphyseal level of the first phalanxes of the last four fingers. US measurements were taken in the hand of the non-fractured arm in the Fractured Group, while measurements were taken in the non-dominant hand in the Control Group, as instructed by the manufacturer. The DBM Sonic Bone Profiler was used in the automatic set-up. Indeed, the operator is required only to position the calliper on the distal metaphysis of the phalanx using US contact gel, and scan around the axis of the phalanx by rotating the calliper. Rotation is considered valid by the apparatus provided it does not exceed an angle of 15° in either direction of the horizontal position of the phalanx. The hand is held horizontally during the measurement process by means of a support. While measurements are being taken, the apparatus every 0.5 s automatically records a US signal. For each finger, 24 US signals are recorded and then the apparatus selects the highest amplitude-dependent speed of sound (AD-SoS) measured. In the end, the average AD-SoS of the four fingers is calculated. A quantitative analysis of the pattern of the transmitted US signal was performed and the Ultrasound Bone Profiler Index (UBPI), expressing the probability of the tested subject belongs to the Control Group to have a fracture at the time of the measurement, was calculated according to the following optimum multivariate logistic model [1,4,30]: UBPI = 1/ 关 1 + exp共 − 0.0018 SDy − 0.056 FWA − 1.1467 TF + 3.03 兲 兴 , where SDy (dynamic of the US signal, mV/µs_) is the mean value of the second derivative of amplitude versus time of the first two peaks of the US signal, FWA (fast wave amplitude, mV) is the first highest peak in the first part of the US received and digitised signal, and TF (time frame, µs) is the time interval between the first received signal and the speed value of 1700 m/s.

334

R. Giardino et al. / Biomed Pharmacother 56 (2002) 332–338

Table 1 Baseline biologic and US data of the women examined; values are mean ± SD Parameter a

Unit

Control Group (n = 94)

Fractured Group (n = 50)

P

Age Age at menopause Menopause, length Height Weight BMI AD-SoS UBPI T-score AD-SoS .

years years months cm kg kg/m_ m/s units

63 ± 7 50 ± 4 156 ± 102 161 ± 5 63.6 ± 7.7 24.5 ± 2.6 1885 ± 95 0.34 ± 0.22 –3.42 ± 1.36

64 ± 6 48 ± 5 186 ± 100 159 ± 6 62.2 ± 6.0 24.5 ± 2.1 1790 ± 98 0.18 ± 0.14 –4.76 ± 1.40

n.s. < 0.05 < 0.0005 n.s. n.s. n.s. < 0.0005 < 0.0005 < 0.0005

a

Abbreviations as in Section 2.

The reproducibility of the US measurements was determined by calculating the root mean square: 共 RMS − CV% 兲 =

冉

m

SD/

xi

冊

兺m i=1

100 ,

where xi is the mean value obtained for each patient i, m is the number of subjects 共 i = 1, 2, ..., m 兲 each one measured five times, and SD is

冑

m

2

SDj

兺 m . j=1

2.1. Statistical analysis Statistical analysis was performed using the SPSS v.10.1 software (SPSS/PC, IL). Data are expressed as mean ± SD, at a significant level of P < 0.05. After verifying a normal distribution and homogeneity of variances, groups were compared by Student’s t-test. The linear regression analysis was used to detect the associations between variables yielding correlation coefficients; in particular the distribution and trend of the AD-SoS and UBPI parameters in relation to the age and months since menopause onset were investigated. The logistic regression analysis was also used to calculate the relative risk of fracture with a CI of 95%. Relative risk of fracture was expressed as the odds ratio (OR) per SD decrease of each variable. Finally, the analysis by receiver operating characteristic (ROC) curves was done in order to assess the discrimination ability of different parameters between fractured and non-fractured subjects by calculating the area under ROC curve (AUC).

3. Results The medical history of the women in the Control Group allowed to identify a small number of subjects who, at the time of the examination, had been undergoing therapy for osteoporosis for 6 months. Sixteen women were following hormone replacement therapy (HRT); eight were on treat-

ment with alendronate and two with calcitonin. None of the 50 forearm-fractured subjects was under treatments with osteotrophic drugs. Demographics of women with and without forearm fractures are reported in Table 1. The average age at menopause and the value of months since menopause began in the Fractured Group were significantly (P < 0.05) lower and higher than in the Control Group, respectively. No other differences were observed. Reproducibility of the method tested on five healthy subjects was RMS-CV% = 0.64 for AD-SoS and RMSCV% = 2.38% for UBPI. In both the groups, AD-SoS was significantly correlated with age (P < 0.0005), although to a higher r-value in the Fractured Group (r = –0.568) compared to the Control Group (r = –0.475) (Fig. 1a). A differently significant (P < 0.0005) trend was observed for UBPI: r = –0.551 for the Fractured Group and r = –0.595 for the Control Group (Fig. 1b). Correlations of AD-SoS with the number of months since menopause (P < 0.0005) were r = –0.491 and r = –0.501 for the Fractured and Control groups, respectively (Fig. 2a). For UBPI, the same correlation trend was found: r = –0.472 (P < 0.01) for the Fractured and r = –0.586 (P < 0.0005) for the Control Group (Fig. 2b). Comparison between the two groups by means of the ROC analysis gave the following values: AUC(ADSoS) = 0.783 ± 0.042 and AUC(UBPI) = 0.768 ± 0.045. The values of the odds ratio (OR) were also calculated for AD-SoS and UBPI and were OR(AD-SoS) = 0.989 (95% CI: 0.984–0.994) and OR(UBPI) = 0.939 (95%.CI: 0.909–0.970), respectively. Logistic regression adjusted to age led to the following results in the determination of OR(AD-SoS) = 12.03 (95% CI: 3.06–47.35) and OR(UBPI) = 7.39 (95% CI: 3.14–17.42), classifying correctly all women used in the model in about 75% of the cases. Using the coefficients obtained from the logistic model, the effect of the combination of the two parameters (AD-SoS and UBPI) on the OR and on the AUC was tested: the OR for the combination of AD-SoS and UBPI was 12.44 (95% CI: 2.95–52.48) and the AUC was 0.791 ± 0.042, respectively. Even if these values were all higher than those obtained using separately

R. Giardino et al. / Biomed Pharmacother 56 (2002) 332–338

335

a

b Fig. 1. AD-SoS (a) and UBPI (b) versus age in the Control Group (white circles) and in the Fractured Group (black circles). The black lines represent the linear regression straight lines between the variables involved. Control Group: AD-SoS = –5.80 (age) + 2251.4; UBPI = –0.02 (age) + 1.39. Fractured Group: AD-SoS = –8.04 (age) + 2306.2; UBPI = –0.01 (age) + 0.89.

336

R. Giardino et al. / Biomed Pharmacother 56 (2002) 332–338

a

b Fig. 2. AD-SoS (a) and UBPI (b) versus months since menopause in the Control Group (white circles) and in the Fractured Group (black circles). The black lines represent the linear regression straight lines between the variables involved. Control Group: AD-SoS = –0.47 (months) + 1958.3; UBPI = –0.001 (months) + 0.54. Fractured Group: AD-SoS = –0.42 (months) + 1871.7; UBPI = –0.001 (months) + 0.29.

R. Giardino et al. / Biomed Pharmacother 56 (2002) 332–338

AD-SoS and UBPI, differences were not statistically relevant.

4. Discussion The current authors collected a series of measurements performed on women who responded to a public radio call and on a group of 50 patients treated for recent fractures of the forearm in order to assess the ability of QUS to estimate the risk of osteoporotic fractures in a cross-sectional study. As far as they know, there are no data available in the literature on this type of fracture using the US finger measurement. Furthermore, the whole study was conducted by operating the device in the automatic mode only, which is supposed to be less dependent on operator skilfulness compared to the non-automatic set-up. The reproducibility of the method (0.64% for AD-SoS and 2.38% for UBPI) is comparable to the values reported in the literature and obtained using non-automatic measurements. In the Control Group, all women had the non-dominant hand measured while in the Fractured Group the hand of the non-fractured limb was measured which showed a decrease in phalangeal US velocity following distal forearm fractures [13]. Nevertheless, it should be noted that no significant difference in US velocity was observed at the phalanges when comparing left and right hand [3,26]. With regard to the Control Group, the current authors observed how the two parameters, AD-SoS and UBPI, tend to diminish with increasing age and years after the onset of menopause. The correlation coefficients found for AD-SoS versus age (r = –0.64, P < 0.001) were very similar to those obtained by Wuster et al. [30] (r = –0.58, P <0.0001), Sili Scavalli et al. [24] (r = –0.61, P < 0.0001), Duboeuf et al. [7] (r = –0.71, P < 0.0001). The same correlation, but with lower r-values, was observed in the Fractured Group. Both AD-SoS and UBPI values were significantly lower in the Fractured Group when compared to 94 controls. The ROC analysis showed a satisfactory discriminating power for both AD-SoS and UBPI, as well as significantly elevated odds ratio for AD-SoS and UBPI indicating that these parameters were able, quite in the same manner, to discriminate between Fractured and Control Groups. In addition, these results are consistent with those reported in the literature for hip and vertebral fracture [1,9,14,23,30]. The logistic and ROC analyses were also carried out by combining the effects of the two US parameters and a slight improvement in the results was observed; however, these differences did not reach a statistical relevance, probably due to the lack of a sufficient data. In conclusion, this cross-sectional study has shown for the first time the ability of phalangeal ultrasonography to discriminate subjects with forearm fractures from controls,

337

thus confirming other authors’ findings on vertebral and hip fractures. The current findings seem to indicate that the US technology is easy to use, portable and does not require extensive training of the operator thanks to the automatic acquisition mode. In addition, US technology seems to be reliable, cost-effective, well accepted by the patient and can be used as a first step diagnosis of osteoporosis. This is particularly important in Europe, where the European Commission directive requires that physicians, for diagnostic purposes, shall seek for alternative techniques involving no ionising radiations [5].

Acknowledgements Financial support for this research was partially given by the “Ministry of Health” (Rome, Italy), special strategic project “Fratture osteoporotiche”. The authors have appreciated the technical assistance of Patrizio Di Denia, Claudio Dalfiume, Patrizia Nini, Nicola Corrado and Franca Rambaldi.

References [1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

Alenfeld FE, Wuster C, Funck C, Pereira-Lima JF, Fritz T, Meeder PJ, Ziegler R. Ultrasound measurements at the proximal phalanges in healthy women and patients with hip fractures. Osteoporos Int 1998;8:393–8. Bauer DC, Gluer CC, Cauley JA, Vogt TM, Ensrud KE, Genant HK, Black DM. Broadband ultrasound attenuation predicts fractures strongly and independently of densitometry in older women. A prospective study. Study of Osteoporotic Fractures Research Group. Arch Intern Med 1997;157:629–34. Benitez CL, Schneider DL, Barrett-Connor E, Sartoris DJ. Hand ultrasound for osteoporosis screening in postmenopausal women. Osteoporos Int 2000;11:203–10. Cadossi R, Cané V. Pathways of transmission of ultrasound energy through the distal metaphysis of the second phalanx of pigs: an in vitro study. Osteoporosis Int 1996;6:196–206. Council Directive 97/43 Euratom of 30/6/97 of Health protection of individuals against the danger of ionising radiations. Directive 84/466. Official J Eur Communities 9/7/1997 (No. L.180/24). Drake WM, Brown JP, Banville C, Kendler DL. Use of phalangeal bone mineral density and multi-site speed of sound conduction to monitor therapy with alendronate in postmenopausal women. Osteoporos Int 2002;13:249–56. Duboeuf F, Hans D, Schott AM, Giraud S, Delmas PD, Meunier PJ. Ultrasound velocity measured at the proximal phalanges: precision and age-related changes in normal females. Rev Rhum Engl Ed 1996;63:427–34. Funk C, Wüster C, Alenfeld FE, Pereira-Lima JFS, Fritz T, Meeder PJ, Götz M, Ziegler R. Ultrasound velocity of the tibia in normal German women and hip fracture patients. Calcif Tissue Int 1996;58:390–4. Guglielmi G, Cammisa M, De Serio A, Scillitani A, Chiodini I, Carnevale V, Fusilli S. Phalangeal US velocity discriminates between normal and vertebrally fractured subjects. Eur Radiol 1999;9:1632–7.

338

R. Giardino et al. / Biomed Pharmacother 56 (2002) 332–338

[10] Hans D, Dargent-Molina P, Schott AM, Sebert JL, Cormier C, Kotzki PO, Delmas PD, Pouilles JM, Breart G, Meunier PJ. Ultrasonographic heel measurements to predict hip fracture in elderly women: the EPIDOS prospective study. Lancet 1996;348:511–4. [11] Hans D, Njeh CF, Genant HK, Meunier PJ. Quantitative ultrasound in bone status assessment. Rev Rhum Engl Ed 1998;65:489–98. [12] Hans D, Wu C, Njeh CF, Zhao S, Augat P, Newitt D, Link T, Lu Y, Majumdar S, Genant HK. Ultrasound velocity of trabecular bones reflects mainly bone density and elasticity. Calcif Tissue Int 1999; 64:18–23. [13] Ingle BM, Hay SM, Bottjer HM, Eastell R. Changes in bone mass and bone turnover following distal forearm fracture. Osteoporos Int 1999;10:399–407. [14] Joly J, Westhovens R, Borghs H, Peeters H, Tirry J, Nijs J, Dequeker J. Reference curve and diagnostic sensitivity for a new ultrasound device for the phalanges, the DBM Sonic 1200, in Belgian women. Osteoporos Int 1999;9:284–9. [15] Kanis JA, Melton LJ, Christiansen C, Johnston CC, Khaltaev N. The diagnosis of osteoporosis. J Bone Miner Res 1994;9:1137–41. [16] Kleerekoper M, Villanueva AR, Stanciu J, Rao DS, Parfitt AM. The role of three-dimensional trabecular microstructure in the pathogenesis of vertebral compression fractures. Calcif Tissue Int 1985;37: 594–7. [17] Luisetto G, Camozzi V, De Terlizzi F, Moschini G, Ballanti P. Use of ultrasonography in the diagnosis of osteomalacia: preliminary results on experimental osteomalacia in the rat. J Ultrasound Med 1999;18:225–9. [18] Luisetto G, Camozzi V, De Terlizzi F. Use of quantitative ultrasonography in differentiating osteomalacia from osteoporosis: preliminary study. J Ultrasound Med 2000;19:251–6. [19] Mele R, Masci G, Ventura V, de Aloysio D, Bicocchi M, Cadossi R. Three year longitudinal study with quantitative ultrasound at the hand phalanx in a female population. Osteoporos Int 1997;7:550–7. [20] Montagnani A, Gonnelli S, Cepollaro C, Bruni D, Franci MB, Lucani B, Gennari C. Graphic trace analysis of ultrasound at the phalanges may differentiate between subjects with primary hyperparathyroidism and with osteoporosis: a pilot study. Osteoporos Int 2002;13:222–7.

[21] Mosekilde L, Bentzen SM, Ortoft G, Jorgensen J. The predictive value of quantitative computed tomography for vertebral body compressive strength and ash density. Bone 1989;10:465–70. [22] Njeh CF, Kuo CW, Langton CM, Atrah HI, Boivin CM. Prediction of human femoral bone strength using ultrasound velocity and BMD: an in vitro study. Osteoporos Int 1997;7:471–7. [23] Reginster JY, Dethor M, Pirenne H, Dewe W, Albert A. Reproducibility and diagnostic sensitivity of ultrasonometry of the phalanxes to assess osteoporosis. Int J Gynecol Obstet 1998;63:21–8. [24] Sili Scavalli A, Marini M, Spadaro A, Messineo D, Cremona A, Sensi F, Riccieri V, Taccari E. Ultrasound transmission velocity of the proximal phalanxes of the non-dominant hand in the study of osteoporosis. Clin Rheumatol 1997;16:396–403. [25] Tauchmanova L, Rossi R, Nuzzo V, del Puente A, Esposto-del Puente A, Pizzi C, Fonderico F, Lupoli G, Lombardi G. Bone loss determined by quantitative ultrasonometry correlates inversely with disease activity in patients with endogenous glucorticoid excess due to adrenal mass. Eur J Endocrinol 2001;145:241–7. [26] Ventura V, Mauloni M, Mura M, Paltrinieri F, de Aloysio D. Ultrasound velocity changes at the proximal phalanges of the hand in pre-, peri-, and postmenopausal women. Osteoporos Int 1996;5: 349–53. [27] Wapniarz M, Lehmann R, Banik N, Radwan M, Klein K, Allolio B. Apparent velocity of ultrasound (AVU) at the patella in comparison to bone mineral density at the lumbar spine in normal males and females. Bone Miner 1993;23:243–52. [28] World Health Organisation. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Technical Report Series. Geneva: Switzerland; 1994. [29] Wu C, Hans D, He Y, Fan B, Njeh CF, Augat P, Richards J, Genant HK. Prediction of bone strength of distal forearm using radius bone mineral density and phalangeal speed of sound. Bone 2000;26:529–33. [30] Wuster C, Albanese C, De Aloysio D, Duboeuf F, Gambacciani M, Gonnelli S, Gluer CC, Hans D, Joly J, Reginster JY, De Terlizzi F, Cadossi R. Phalangeal osteosonogrammetry study: age-related changes, diagnostic sensitivity, and discrimination power. The Phalangeal Osteosonogrammetry Study Group. J Bone Miner Res 2000;15:1603–14.