Anterior Dynamic Ultrasound and Graf s Examination in Neonatal Hip Instability

ORIGINAL ARTICLE ACTA RADIOLOGICA Anterior Dynamic Ultrasound and Graf s Examination in Neonatal Hip Instability T. FINNBOGASON, H.JORULF, E.SÖDERMAN &L.REHNBERG Department of Pediatric Radiology and Department of Pediatric Orthopedics, Astrid Lindgren Children s Hospital, Karolinska University Hospital, Stockholm, Sweden; Department of Woman and Child Health, Karolinska Institute, Stockholm, Sweden Finnbogason T, Jorulf H, Söderman E, Rehnberg L. Anterior dynamic ultrasound and graf s examination in neonatal hip instability. Acta Radiol 2008;49:204 211. Background: Discrepancy between neonatal hip morphology and stability has been reported in the literature. Comparative ultrasound studies on this issue are limited. Purpose: To compare neonatal hip instability, as assessed by dynamic ultrasound and clinical examination, with acetabular morphology, as assessed by Graf s method. Material and Methods: 536 newborn infants with clinical signs of hip instability, ambiguous findings at clinical hip examination, or positive risk factors for DDH were investigated with two ultrasound methods, the Graf method and anterior dynamic ultrasound, at an average age of 12 days. The hips were allocated to three groups according to the Graf result: A, normal (type Ia and b); B, borderline or immature (type IIa); and C, pathologic (type IIc and worse). Graf examination was compared with two diagnostic tests for instability, namely clinical examination by senior pediatric orthopedists and anterior dynamic ultrasound. Results: According to Graf s method, 77% of the hips were normal, 20% borderline/ immature, and 3% pathologic. On clinical hip examination, 82% were stable, 14% unstable, and 4% dislocatable. The dynamic ultrasound outcome was 88% stable hips, 10% unstable, and 2% dislocatable. Of the hips considered unstable or dislocatable on dynamic ultrasound, 21% had normal (type I) and 66% immature acetabular morphology according to the Graf method. Of the hips that were stable on dynamic ultrasound, only one (0.1%) was dysplastic according to the Graf method. Graf s examination showed the smallest number of normal hips, but also the fewest pathologic hips, with many indeterminate results that needed follow-up. Conclusion: Acetabular morphology correlated better to stability as assessed by dynamic ultrasound than to the clinical examination results, with fair to moderate agreement. Graf s examination resulted in a large number of indeterminate results that needed follow-up, but when used as the sole criterion for deciding treatment did not lead to a higher treatment rate than when the decision was based on clinical hip examination. Key words: Comparative studies; hip; pediatrics; skeletal, appendicular; ultrasound Thröstur Finnbogason, Department of Pediatric Radiology, Astrid Lindgren Children s Hospital, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden (tel. +46 8 517 77116 / +46 736259157, fax. +46 8 517 77125, e-mail. throstur.finnbogason@ karolinska.se) Accepted for publication October 4, 2007 Several different ultrasound methods have been introduced as tools for diagnosing developmental dysplasia of the hip (DDH) in need of treatment (1 8). These can be divided into two different ultrasound imaging approaches reflecting the two components of DDH: instability and acetabular development. Dynamic ultrasound examinations are used to detect neonatal hip instability, and with static ultrasound examinations morphologic changes in the acetabulum can be visualized. A study of the relationship between hip stability, as assessed by dynamic ultrasound with a lateral approach, and acetabular morphology, as evaluated by a static ultrasound examination according to the Graf method, has been reported in the literature (9). To our knowledge, no comparisons between static ultrasound by Graf s method and anterior dynamic ultrasound have been published. DOI 10.1080/02841850701775022 # 2008 Informa UK Ltd. (Informa Healthcare, Taylor & Francis AS)

Anterior Dynamic US and Graf s Examination in Neonatal Hip Instability 205 The aim of the present study was to see how the infant acetabular morphology as assessed with Graf s method correlated to the hip stability as judged by use of an anterior dynamic ultrasound method; and, further, to compare the result with the findings on clinical hip examination performed by pediatric orthopedists, and to evaluate the impact of the different methods on the treatment rate. Material and Methods Clinical hip screening was performed by the maternity department pediatricians on all newborn infants born at the maternity unit of this hospital, on the first day of life, according to the standard routine at the hospital. Infants with risk factors for DDH such as breech delivery, foot or neck deformity, or a family history of DDH, and those with a clinically unstable hip or ambiguous findings on clinical examination, were referred for pediatric orthopedic consultation at an age of 10 to 14 days. During a period of 4 years, newborn babies referred for this orthopedic consultation with risk factors or with ambiguous or abnormal clinical findings were included in the study. The study comprised four stages: 1) Clinical hip screening at the maternity ward. Within 24 hours after birth, all infants born at the hospital were examined by a pediatrician with regard to hip instability. The hips were graded into three groups as stable, unstable, or dislocatable/ dislocated on the basis of the Barlow test. All infants with abnormal or ambiguous hips, and all those with risk factors for DDH were referred for the orthopedic consultation. 2) First orthopedic examination. At the orthopedic consultation, a clinical hip examination, including the Barlow and Ortolani tests, was performed. This was done by an experienced pediatric orthopedist who graded the hips into three groups as stable, unstable, or dislocatable. 3) Ultrasound examinations. After the orthopedic examination, the infant was referred to the ultrasound department for static and dynamic ultrasound examinations. In most cases these were done at the same visit to the hospital as for the orthopedic consultation, usually within a few hours. The ultrasound examinations were performed either by an experienced pediatric radiologist or by a sonographer who, prior to the study, had received special training in clinical hip examination, dynamic ultrasound, and the Graf method. A total of six pediatric radiologists with many years of experience in pediatric ultrasound participated in the study. Prior to the study, all ultrasound examiners were trained in the dynamic ultrasound method as well as in the Graf method by one of the authors (T.F.), to ensure that all used the same technique. The findings at the orthopedic examination were not known by the ultrasound examiner. The static ultrasound examination was performed according to Graf s method (10). The examination was carried out in a standard manner, as described in the literature, with the infant in the right and left lateral decubitus positions. At least two images of each hip were acquired in the standard plane. The ultrasound equipment used was an Acuson Sequoia (Siemens Medical Solutions Inc., Mountain View, Calif., USA) with a 5- or 7-MHz linear transducer. The hips were graded according to Graf s classification, and the alpha angle was measured. The examiner noted the findings, and these were included in a report directly after the examination. The dynamic ultrasound method was developed at the department (11), and was a modification of the method described by DAHLSTRÖM and coworkers in 1986 (1). This was an anterior approach that departed from the original method by the use of a special transducer fixing device that allowed one person to perform a hip stress test using Barlow s maneuver and, at the same time, to manage the ultrasound probe. The ultrasound equipment used was an Acuson128XP (Siemens Medical Solutions Inc., Mountain View, Calif., USA) with a 5-MHz sector transducer. The dynamic examination was documented with video recording and directly transferred to the PACS with cineloops. The examiner evaluated the hip stability on the basis of the visual information from the ultrasound together with the tactile perception during the stress test, and the hips were classified as normal, unstable, or dislocatable. These findings were included in a report to the orthopedic surgeon. The dynamic ultrasound classification was subjective. No measurements were used to decide the degree of displacement of the femoral head. A hip that was seen or felt to be displaced out of the acetabulum was classified as dislocatable. A hip that was considered to show significant instability requiring treatment, but not to be dislocatable, was assigned to the unstable group. Hips with minimal (insignificant) instability, together with completely stable hips, were assigned to the stable group. 4) Second orthopedic examination. After the ultrasound examinations, the infant was returned to the orthopedic unit, where a second clinical hip examination was performed by the orthopedist, with the same classification into normal, unstable, and dislocatable. The second orthopedic examination

206 T. Finnbogason et al. Table 1. Distribution of sonographic types according to the Graf classification, and the number (and percentage) of hips that were unstable, dislocatable, or dislocated according to the dynamic ultrasound and the first orthopedic examination Graf group n Unstable hips by dynamic ultrasound Dislocatable or dislocated hips by dynamic ultrasound Unstable hips by first orthopedic examination Dislocatable or dislocated hips by first orthopedic examination Ia 724 20 (2.8%) 1 (0.14%) 89 (12.3%) 5 (0.69%) Ib 97 7 (7.2%) 0 (0.0%) 11 (11.3%) 0 (0.0%) IIa 220 66 (30.0%) 7 (3.2%) 47 (21.4%) 23 (10.4%) IIc 20 17 (85.0%) 2 (10.0%) 6 (30.0%) 5 (25.0%) IId 6 1 (16.7%) 5 (83.3%) 1 (16.7%) 5 (83.3%) IIIa 2 0 (0%) 2 (100%) 0 (0%) 2 (100%) IIIb 3 1 (33.3%) 2 (66.7%) 1 (33.3%)* 1 (33.3%)* IV 0 0 (0%) 0 (0.0%) 0 (0%) 0 (0.0%) Total 1072 112 (10.4%) 19 (1.8%) 155 (14.5%) 41 (3.8%) * One hip that was classified as type IIIb according to Graf s examination (and dislocatable according to dynamic ultrasound) was considered stable at the first orthopedic examination. This hip was revised to dislocatable at the second orthopedic examination. was carried out within hours after the ultrasound examination, even when the sonography was performed on a different day than the first orthopedic examination. On the basis of the clinical findings together with the ultrasound report, the orthopedist made a diagnosis and decided on treatment. All statistical analyses were carried out with SPSS for Windows, version 10.0.1, and Amos version 6.0 (SPSS Inc., Chicago, Ill., USA). Cohen s kappa was used to measure the agreement between the Graf examination and the two tests for stability, namely dynamic ultrasound and the clinical examination. Path analysis was used to study the way in which different diagnostic tests of hip stability influenced the orthopedist s decision regarding treatment. Path analysis is a method for modeling cause and effect based on a covariance or a correlation matrix, and is closely related to multiple regression. A semiexploratory path analysis based on a covariance matrix was used. First, those paths which the study design rendered unfeasible were excluded. The remaining paths were determined by stepwise exclusion of non-significant paths. The direction of the causation was determined by the study design; cause preceded effect. Results During the study period, 18,031 infants were born at the hospital maternity unit. A total of 538 infants were selected for inclusion in the study. Two were excluded from statistical evaluation because of missing data, leaving 536 infants or 1072 hips for the statistical analysis; 64% (342) were girls and 36% (194) boys. The mean gestational age was 39 weeks (SD 1.8 weeks), with a range of 28 to 43 weeks. The mean birth weight was 3375 g (SD 608 g). The mean age at the time of the ultrasound examination was 12.2 days (SD 4.8 days). The static and dynamic ultrasound examinations were performed during the same visit to the radiology department. In most instances, they were done on the same day as the orthopedic examination, but in 14 cases there was an interval exceeding 7 days. The hips were allocated to the standard ultrasonographic types according to Graf s classification. The distribution is illustrated in Table 1, together with the number (and percentage) of hips that were found to be unstable, dislocatable, or dislocated on dynamic ultrasound and at the first orthopedic examination. For statistical purposes, the hips were allocated to three groups based on the Graf classification. Group A comprised hips with normal acetabular morphology, corresponding to type Ia and Ib, with an alpha angle of 60 or more. Group B comprised borderline or immature hips (type IIa), with an alpha angle of 50 59. Group C consisted of dysplastic hips (type IIc and worse), with an alpha angle of below 50 (Fig. 1). Of the 821 hips (77%) with normal morphology (group A) according to the Graf classification, 28 were found to be unstable or dislocatable on dynamic ultrasound and 105 on clinical examination (Tables 1 3). The borderline group, group B, comprised 220 hips. Of these, 32% were judged to be unstable or dislocatable at the clinical orthopedic examination and 33% on dynamic ultrasound. Thirty-one hips were allocated to group C. This group represented pathologic hips that needed treatment; of these, only one was considered stable on the basis of dynamic ultrasound, and the remaining 30 hips (97%) were judged to be unstable or dislocatable. According to the orthopedic examination, 10 hips were stable and 21 (68 %) were unstable or dislocatable. The acetabular morphology of the hips showing instability (Tables 1 3) varied as follows. According to the dynamic ultrasound examination, 131 hips

Anterior Dynamic US and Graf s Examination in Neonatal Hip Instability 207 Fig. 1. Hip sonogram in the standard plane according to Graf, illustrating the different groups. A. Normal hip, type Ia. B. Immature hip, type IIa. C. Dysplastic hip, type IIc. were unstable or dislocatable. Of these, 28 (21%) were normal according to the Graf classification, 73 (56%) were immature, and 30 (23%) were dysplastic. The 196 hips that were considered unstable or dislocatable on clinical hip examination showed Table 2. The results of the dynamic ultrasound compared with those of the Graf examination (n51072 hips) Graf Dynamic ultrasound A B C Total Normal 793 147 1 941 Unstable 27 66 19 112 Dislocatable 1 7 11 19 Total 821 220 31 1072 A: normal hips (type Ia and Ib); B: borderline (type IIa); C: abnormal (type IIc and worse). Kappa50.381. poorer agreement with the acetabular morphology; of these, 105 (53%) were normal, 70 (36%) were immature, and 21 (11%) were dysplastic according to the Graf method. The dynamic ultrasound results were in agreement with those of the first orthopedic examination regarding 867 hips (Table 4). Of the 816 hips that were judged as stable with both methods, only one (0.1%) proved to be dysplastic (type IIc) according to the Graf classification, and 115 (14%) of the hips were immature. One half of the hips that were judged to be unstable with both methods were allocated to group B. Of the 15 hips that were dislocatable with both methods, two-thirds were dysplastic according to the Graf method and none was normal.

208 T. Finnbogason et al. Table 3. The results of the first orthopedic examination compared with those of the Graf examination (n51072 hips) First orthopedic Graf group examination A B C Total Normal 716 150 10 876 Unstable 100 47 8 155 Dislocatable 5 23 13 41 Total 821 220 31 1072 A: normal hips (type Ia and Ib); B: borderline (type IIa); C: abnormal (type IIc and worse). Kappa50.199. Effect on treatment The findings in individual infants were then considered. The hip with the worst score decided treatment. A theoretical treatment rate was calculated on the basis of the outcome of the different diagnostic tests. Regarding the clinical examination and the dynamic ultrasound, all infants with at least one hip considered unstable or dislocatable were assumed to need treatment. The hips in Graf s group A were considered normal and to need no follow-up or treatment. Those in group B were considered indecisive cases that needed a follow-up for decision. Group C comprised pathologic hips that needed treatment without further investigation. In this way, use of the Graf examination to decide treatment led to 132 borderline cases in need of a follow-up examination and 21 that should receive treatment without further investigation. The first orthopedic examination classified 154 infants as having at least one hip unstable or dislocatable, and thus being in need of treatment (Table 5). The dynamic ultrasound suggested that 89 infants should be treated. The use of Graf s method resulted in 132 borderline cases needing a follow-up examination before treatment could be decided, which meant that between 21 and 153 infants should receive treatment. Table 4. The results of the Graf examination in the 867 hips on which the dynamic ultrasound and the first clinical orthopedic examination were in agreement Dynamic ultrasound agrees Graf group with orthopedist s stress test A B C Total Stable 700 115 1 816 Unstable 11 18 7 36 Dislocatable 5 10 15 Total 711 138 18 867 A combination of Graf s examination and dynamic ultrasound was also tested (Table 5), where all infants that met the criteria for treatment with one or both methods were considered in need of treatment. Of the 536 infants, 363 showed normal conditions on both examinations. Eighty-four had stable hips bilaterally but had an immature/borderline acetabulum in at least one hip, and thus needed follow-up. Eighty-nine infants had at least one hip that was unstable or dislocatable, or had acetabular dysplasia type IIc or worse according to the Graf classification, and thus needed treatment. Owing to the large number of indecisive cases resulting from the Graf examination, the calculated treatment rate had a wide range, between 4.9 and 9.6/1000. To address the question how the different estimates of hip instability or pathology influenced the orthopedist s choice of treatment, a path analysis (12) was performed (Fig. 2). The orthopedists were influenced in three steps, as illustrated by the significant regression coefficients in each step. 1) The orthopedist s first estimate, the outcome of the first clinical orthopedic examination, was influenced by the results of the clinical examination in the maternity ward (b50.55). 2) The orthopedist s second estimate, the outcome of the second clinical orthopedic examination, was influenced by the orthopedist s first estimate Table 5. The outcome per patient, based on the hip with the worst score (n5536 infants) Instability/acetabular development based on the hip with worst outcome in test Normal Unstable Borderline Graf Dislocatable Pathologic Graf Number to be treated Treatment rate 95% confidence interval Clinical exam, maternity unit 413 93 30 123 6.8ø 5.7 7.8 First orthopedic exam 382 123 31 154 8.5ø 7.3 9.6 Second orthopedic exam 389 114 33 147 8.2ø 7.0 9.2 Dynamic ultrasound 447 74 15 89 4.9ø 3.9 5.8 Graf s method 383 132 21 21 153* 1.2 8.5ø* 0.6 9.6 Dynamic ultrasound combined with Graf s method 363 84 89 89 173* 4.9 9.6ø* 3.7 10.7 The treatment rate was based on the total number of infants born during the study period, i.e., 18,031. The actual treatment rate during the study period was 8.0ø (144 infants). * The wide range was due to a large number of borderline cases (group B) at the Graf examination that needed a follow-up examination before a decision could be taken on further management.

Anterior Dynamic US and Graf s Examination in Neonatal Hip Instability 209 Fig. 2. Path analysis based on the variance covariance matrix of five estimates of hip pathology (coded as 15normal, 25unstable, 35dislocatable) and treatment (coded as 15no treatment, 25Frejka pillow, 35von Rosen splint) (n51072 hips). Numbers within circles are residual variances not accounted for in the model. Double-ended arrows are covariances between these residuals caused by the study design. Single-ended arrows are significant regression coefficients (P(0.05) that are part of the model. Computations were performed using Amos version 6.0 from SPSS Inc. (b50.83) and by the result of the dynamic ultrasound (b50.23). 3) The choice of treatment was influenced by the orthopedist s second estimate (b50.85), by the dynamic ultrasound (b50.07), by Graf s method (b50.07), and by the clinical examination in the maternity ward (b50.05). In general, the path analysis indicated that, if any of the estimates suggested a need for treatment, the infant was treated. Discussion The literature is overwhelmed with studies/papers dealing with ultrasound as a tool or complement in diagnosing neonatal hip instability and DDH. Since the initial paper by GRAF was published in 1980 (3), several different sonographic methods have been introduced, using either a static or a dynamic approach, or a combination of the two. Reports on studies actually comparing static and dynamic ultrasound for this purpose have been sparse. ROSENDAHL and coworkers (9) compared the Graf examination with dynamic ultrasound using a lateral approach, and found that 49% of the sonographically unstable hips had either normal (Graf type I) or immature (Graf type IIa) acetabular morphology. In our study, using another dynamic ultrasound method, 77% of the sonographically unstable or dislocatable hips were normal or immature according to the Graf classification (Table 2). The dynamic ultrasound methods have varied regarding the direction of visualization of the hip joint, the placement of the ultrasound probe, and the method used for the stress test. We decided to modify an anterior dynamic method first introduced by DAHLSTRÖM and coworkers in 1986 (1), which closely copied the Barlow provocation method (13, 14). By attaching an ultrasound probe holder to a special examination table (11), it was possible to conduct a normal clinical examination of the hips with added ultrasound visualization, complementing the tactile perception of instability with the visual dimension. The goal of this study was to compare this dynamic ultrasound method with Graf s static examination, in a selective population of infants with hip instability or risk factors for DDH. Our results confirm previous observations that sonographically unstable hips can have normal morphology (9, 15). The results show fair to moderate agreement between infant hip instability, as assessed with the dynamic ultrasound approach, and acetabular morphology, as judged by Graf s method (kappa 0.381). The agreement was not as good, poor to fair, regarding the clinical stress test (kappa 0.199), where 105 unstable or dislocatable hips were considered normal according to Graf s examination (Table 3). In the absence of a true gold-standard test for instability, we combined the clinical orthopedic examination with dynamic ultrasound and looked at the hips in which the two tests were in agreement. Our theory was that two independent tests of instability in agreement indicate a more true instability. This combination resulted in better agreement between hip stability and morphology, as follows from Table 4. Only 11 hips with instability were morphologically normal. The large number of indecisive hips belonging to the immature group is a limitation in this study. More than 20% of the hips examined with the Graf

210 T. Finnbogason et al. method were allocated to the immature/borderline group, and would have required a follow-up to make a definite diagnosis and decide the course of management. Hence, the number of hips in this study actually requiring treatment based on the acetabular morphology is not clear, but ranges between 4.9 and 9.6/1000. It is of interest that use of Graf s examination to decide about treatment did not lead to increased treatment as compared to the clinical hip examination. There is no standardization of the hip stress test. The diagnosis of hip instability is subjective. There is a gray zone, with overlapping between normal, unstable, and dislocatable hips. The degree of provokable instability depends on many factors, such as the examination technique, the force applied, and whether the baby is relaxed. These factors are difficult to standardize and rely on the experience of the examiner. This applies both to the anterior dynamic ultrasound examination and the clinical stress test, which use the same technique to test for stability. On the other hand, the classification of hips into subtypes according to Graf s method is highly standardized, but nevertheless significant intra- and interobserver errors have been reported (16, 17), especially regarding abnormal hips, and the value of proper training has been emphasized (18). The prevalence of immature and dysplastic hips in this study was unusually low as compared with several other reports (19 21). One reason for this is that a selected population was studied, and the proportion of borderline or pathologic hips according to the Graf classification among those that were not admitted to the study is not known. The lack of a gold standard regarding the diagnosis of neonatal hip instability and DDH in need of treatment is another limitation. The final diagnosis in this study was decided at a second clinical hip examination that was performed by an orthopedist after the ultrasonography examinations. The orthopedist made his decision on the basis of this second clinical examination, but with access to the result of the prior clinical examinations and the ultrasound examinations. In conclusion, our findings confirm earlier observations that there is a discrepancy between hip stability and acetabular morphology, but that acetabular morphology correlates better with the dynamic ultrasound results than with the findings at the clinical hip examination. Using Graf s examination to decide on treatment will result in a large number of indeterminate results that need followup. When the Graf method is used as the sole criterion for deciding on treatment in a selected population, this does not lead to a higher treatment rate than when the decision is based on the clinical Barlow and Ortolani tests. In view of previous observations that unstable hips with normal morphology probably do not need treatment (9, 15), we believe that static ultrasound can have a role in selective screening of neonatal hip instability and DDH by reducing the treatment rate among hips that are clinically or ultrasonographically unstable. References 1. Dahlström H, Oberg L, Friberg S. Sonography in congenital dislocation of the hip. Acta Orthop Scand 1986;57:402 6. 2. Gomes H, Menanteau B, Motte J, Robilard P. Sonography of the neonatal hip: a dynamic approach. Ann Radiol 1987;30:503 10. 3. Graf R. The diagnosis of congenital hip-joint dislocation by the ultrasonic Combound treatment. Arch Orthop Trauma Surg 1980;97:117 33. 4. Harcke HT, Clarke NMP, Lee MS, Borns PF, MacEwan GD. Examination of the infant hip with real-time ultrasonography. J Ultrasound Med 1984;3: 131 7. 5. Novick G, Ghelman B, Schneider M. Sonography of the neonatal and infant hip. Am J Roentgenol 1983;141:639 45. 6. Saies AD, Foster BK, Lequesne GW. The value of a new ultrasound stress test in assessment and treatment of clinically detected hip instability. J Pediatr Orthop 1988;8:436 41. 7. Suzuki S, Kasahara Y, Futami T, Ushikubo S, Tsuchiya T. Ultrasonography in congenital dislocation of the hip. Simultaneous imaging of both hips from in front. J Bone Joint Surg Br 1991;73:879 83. 8. Terjesen T, Bredland T, Berg V. Ultrasound for hip assessment in the newborn. J Bone Joint Surg Br 1989;71:767 73. 9. Rosendahl K, Markestad T, Lie RT. Ultrasound in the early diagnosis of congenital dislocation of the hip: the significance of hip stability versus acetabular morphology. Pediatr Radiol 1992;22:430 3. 10. Graf R. Hip sonography: diagnosis and management of infant hip dysplasia. Berlin-Heidelberg: Springer, 2006. 11. Finnbogason T, Jorulf H. Dynamic ultrasonography of the infant hip with suspected instability. A new technique. Acta Radiol 1997;38:206 9. 12. Bollen KA. Structural equations with latent variables. New York: Wiley, 1989. 13. Barlow TG. Early diagnosis and treatment of congenital dislocation of the hip. J Bone Joint Surg Br 1962;44B:292 301. 14. Coleman SS. Diagnosis of congenital dysplasia of the hip in the newborn infant. JAMA 1956;162:548 54. 15. Reikeras O, Kristiansen LP, Gunderson R. Ultrasonography of the infant hip: the significance of provokable instability with normal morphology. Orthopedics 2002;25:833 5.

Anterior Dynamic US and Graf s Examination in Neonatal Hip Instability 211 16. Bar-On E, Meyer S, Harari G. Ultrasonography of the hip in developmental hip dysplasia [see comments]. J Bone Joint Surg Br 1998;80:321 4. 17. Omeroglu H, Bicimoglu A, Koparal S, Seber S. Assessment of variations in the measurement of hip ultrasonography by the Graf method in developmental dysplasia of the hip. J Pediatr Orthop B 2001;10:89 95. 18. Roposch A, Graf R, Wright JG. Determining the reliability of the Graf classification for hip dysplasia. Clin Orthop Relat Res 2006;447:119 24. 19. Roovers EA, Boere-Boonekamp MM, Castelein RM, Zielhuis GA, Kerkhoff TH. Effectiveness of ultrasound screening for developmental dysplasia of the hip. Arch Dis Child Fetal Neonatal Ed 2005;90:F25 30. 20. Rosendahl K, Markestad T, Lie RT. Ultrasound screening for developmental dysplasia of the hip in the neonate: the effect on treatment rate and prevalence of late cases. Pediatrics 1994;94:47 52. 21. Tonnis D, Storch K, Ulbrich H. Results of newborn screening for CDH with and without sonography and correlation of risk factors. J Pediatr Orthop 1990;10:145 52.