Pediatr Radiol (2017) 47:817 821 DOI 10.1007/s00247-017-3802-4 COMMENTARY Hip ultrasound for developmental dysplasia: the 50% rule H. Theodore Harcke 1 & B. Pruszczynski 2 Received: 27 October 2016 /Revised: 5 January 2017 /Accepted: 10 February 2017 /Published online: 27 February 2017 # Springer-Verlag Berlin Heidelberg 2017 For those who perform infant hip ultrasounds, there has always been the issue of measurement as part of the sonographic assessment. The 2013 American Institute of Ultrasound in Medicine (AIUM) practice parameter states: Validation by angle and femoral head coverage measurement is optional [1]. The Graf followers have subscribed to the tables of alpha and beta angles and find publications illustrating their use and results [2, 3]. Those who do not wish to use angles and rely on the Dynamic (Harcke) technique for position and stability [4, 5] have other criteria for acetabular development [6]. Our group directed attention to the coverage of the femoral head by the bony acetabulum [7] as well as to the slope and configuration of the iliac bone. As a consequence, we are referenced as the source of the 50% rule: In a normal hip, the acetabulum should cover 50% (one half or more) of the femoral head. While we have not discouraged the use of this guideline, we have not formally published it as dictum. This review seeks to clarify the rule, present data that support its use, and note pitfalls that can occur with its application. As a caveat, this discussion assumes that when reference is made to a coronal sonogram, whether the hip is extended (neutral) or flexed, the image correctly shows: (1) a straight iliac line, (2) the junction of the iliac bone and tri-radiate cartilage and (3) the echogenic tip of the labrum. * H. Theodore Harcke howard.harcke@gmail.com 1 2 B. Pruszczynski blazejpruszczynski@gmail.com Nemours/Alfred I. dupont Hospital for Children, 1600 Rockland Road, Wilmington, DE 19803, USA Clinics of Orthopaedics and Paediatric Orthopaedics, Medical University of Lodz, Lodz, Poland Early in the development of hip ultrasound, there was a need to establish its validity in comparison with established techniques like the pelvic radiograph. For acetabular development, the radiographic acetabular index (angle) was, and still is, the gold standard. In our quest to validate the coronal flexion sonogram as a reflection of acetabular development, we enlisted Christian Morin, a pediatric orthopedic fellow from France, to compare femoral head coverage as seen by sonogram with a standard pelvic radiograph obtained at the same time. It is important to note the ultrasound technology of the time (1984) was the mechanical sector transducer, which showed a curved configuration of the iliac bone and had less resolution. Comparing bony acetabular depth (d) with the diameter of the cartilaginous femoral head (D) produced a fraction easily converted to a percentage. When comparing percentage of coverage with measure of acetabular angle, we found that 58% or greater coverage always correlated with an accepted normal value. Similarly, coverage of less than 33% always reflected an abnormal radiographic acetabular angle [7]. Many of our sonograms fell in the 33 58% coverage range, where the radiograph measure could be either normal or abnormal. Terjesen et al. [8] also compared radiography and ultrasound using a measure of acetabular coverage called bony roof %. This measure used different landmarks on a coronal neutral view and found some cases with normal ultrasound and dysplasia by radiographic index [8]. When sonography transitioned to linear-array transducers, thereby achieving better resolution and straight iliac lines, Dr. Morin (then with his own program in France) repeated the d/d study with a linear probe. In 1999, he published a new evaluation of d/d that looked at sector and linear probes and compared coverage and alpha angle measurement with radiographic measure [9]. Interestingly,
818 Pediatr Radiol (2017) 47:817 821 there was little difference in d/d correlation with data from sector and linear probes. All acetabula had normal radiographic angles at 56% or more coverage. All hips with less than 40% coverage had abnormal acetabular angles [9]. While the gray zone was reduced from 33 58% to 40 56%, there are still a significant number of cases where there is uncertainty as to radiographic acetabular angle normality. However, it seems likely there is at least a reasonably satisfactory acetabulum with 50% coverage. A variation of the 50% rule occurs when d is measured in a different manner. In the original discussions of d/d, it was always assumed that the d/d determination was being made to compare acetabular depth with head diameter. With a hip that is positioned against the medial wall of the acetabulum, this is done with three parallel lines to determine the d and D values (Figs. 1 and 2). However, when dealing with a subluxated or dislocated hip, it takes four lines to correctly measure whether one is checking the acetabular depth, the original intent of the measure (Figs. 1 and 2) and this detail has not been emphasized in the literature. An alternative interpretation of femoral head coverage can be made with another measurement, which, for clarity in this discussion, we will call d*. Rather than acetabular depth, this measure is the amount of the femoral head covered by the acetabulum and can be compared to the femoral head as d*/d (Figs. 1 and 2). In these figures (b in both), consider d/d and d*/d calculations for a dysplastic acetabulum with hip subluxated. Note that both d/d and d*/d are less than 50%. For a sonogram where the acetabular depth is normal but the femoral head is subluxated, d/d can be greater than 50%, while d*/d will be abnormal (less than 50%). It follows that when a hip is seated (no displacement), d and d* are essentially the same distance and now indicate acetabular depth. Because we know that developmental dysplasia of the hip (DDH) has elements that relate to morphology and stability, we should not be surprised to find that the 50% rule is used to assess both elements depending upon whether d or d* is selected. It is important to recognize which ratio is used and its purpose. Besides relating d/d to radiographic acetabular angle, there is also the question of the relationship of d/d to the alpha angle. Those advocating the use of the alpha angle should be aware that it has limited utility in predicting the radiographic acetabular angle. Morin et al. [9] foundthat it did not correlate as well with acetabular angle as d/d in his 1999 study. More recently, Gunay et al. [10] reported the relationship between alpha angle and d/d coverage in the coronal neutral view. They used an alpha angle of 60 to be the gold standard for a normal hip and found 51% or more coverage always related to an angle of 60 or better. Hence, their d/d of greater than 50% suggests reasonable acetabular development. A lower threshold for adequate development was 39%, with all hips of this coverage or less being pathological. Consequently, between 39% and 51%, there is uncertainty as to alpha measure [10]. We have unpublished data on d/d correlation with alpha angles embedded in our retrospective study of infants with instability published in 2015 [11]. These measurements were made by a single observer on selected views judged to be correctly obtained. Extracting measurements taken as part of this study allows one to examine how d/d and alpha angle measures behaved longitudinally in coronal neutral and coronal flexion views (Table 1). There was a better correlation between d/d and alpha on the second measurements when infants were 3 4 monthsofagethan there was for the initial measurements taken when they were 1 2 months of age. This was attributed to the Fig. 1 Coronal neutral view ultrasound schematic. a Normally seated hip and normal acetabulum with d/d measurement. b Subluxated hip and dysplastic acetabulum with d/d and d*/d measurements. d/d the ratio of acetabular depth to diameter of the femoral head [7, 9], d*/d ratio of femoral head coverage to diameter of the femoral head
Pediatr Radiol (2017) 47:817 821 819 Fig. 2 Coronal flexion view ultrasound schematic. a Normally seated hip and normal acetabulum with d/d measurement. b Subluxated hip and dysplastic acetabulum with d/d and d*/d measurements. d/d the ratio of acetabular depth to diameter of the femoral head [7, 9], d*/d ratio of femoral head coverage to diameter of the femoral head improved delineation of landmarks such as the superior lateral corner of the acetabulum. We also found that serial mean values of alpha reflected improvement in development. While these findings were made in only a small number of infants from a selected population, they provide support for the relationship between the two measures. These data, however, also suggest a slightly better d/d correlation with alpha when measured in the coronal neutral view. This does not surprise us, because the probe position becomes more variable with flexion, affecting the selection of the mid-acetabular plane. From the discussion above, it is evident that leg position introduces another measurement variable. Measurement of d/d and alpha angle is possible in both coronal views (i.e. with the femur extended [neutral] or flexed). Our experience with both measurements leads us to recommend that d/d and d*/d be assessed in the coronal flexion view and alpha angle be assessed in the coronal neutral view. One reason is that coronal flexion is a stress view, and in the presence of instability, all measures of distance can be correlated with application of stress. A second rationale for this is when the literature describing a measurement technique and values is based upon a particular view, application to an alternative view could introduce variation. While the work of Gunay et al. [10] suggesting the use of d/d in the coronal neutral plane is valid, their paper does not address the difference in measure with displaced hips. In a comparable situation, one can measure an alpha angle on the coronal flexion view and find that it does not exactly match the alpha angle measured on the coronal neutral view of the same hip. Such is the effect of subtle differences in probe angle that all sonographers appreciate. Table 1 Comparison of d/d a and alpha angle b sonographic View and statistic 1st measure measurements c (age 1 2 months) 2nd measure (age 3 4 months) Comment Coronal neutral 0.33 0.68 Improved with age d/d vs. alpha angle correlation coefficient d Coronal flexion 0.48 0.61 Improved with age d/d vs. alpha angle correlation coefficient d Coronal neutral mean alpha angle (standard deviation) Coronal flexion mean alpha angle (standard deviation) 60.6 (7.1) 65.1 (6.1) P<0.05 e Improved with age 59.7 (4.5) 64.5 (5.6) P<0.05 e Improved with age a d/d=d/d is the ratio of acetabular depth to diameter of the femoral head [7, 9] b Alpha angle as determined by Graf method [2, 3] c Data derived from Pruszczynski et al. [11] d Correlation coefficient is Cohen kappa; 0.41 to 0.60 moderate agreement; 0.61 to 0.80 substantial agreement e Paired t-tests for mean difference
820 Pediatr Radiol (2017) 47:817 821 Fig. 3 Comparison of radiographic and sonographic measurements in a 4-month-old girl following Pavlik harness treatment for bilateral developmental dysplasia of the hip. a Anteroposterior pelvic radiograph shows increased acetabular angles and superior lateral corner notching (arrows). b Coronal flexion sonogram of the right hip shows d/d is 52% with notched defect at the superior lateral margin (arrow). c, d Coronal neutral (c) andcoronal flexion (d) left hip sonograms show greater than 50% coverage (d/d), but notching defects of the lateral corners (arrows) are a reflection of residual dysplasia. d/d the ratio of acetabular depth to diameter of the femoral head [7, 9] In conclusion, there is a rationale for the 50% rule, although this has not been formally proposed. Its use as an assessment of acetabular development is supported by Morin s data relating d/d and radiographic acetabular index, and as a predictor of the alpha angle it is supported by Gunay s [10] work and our unpublished observations. As an assessment of hip position percentage coverage, d*/ D represents a variation of the 50% rule measure, which decreases with lateral displacement and increases as the femoral head seats and d and d* become the same. We have not endorsed ultrasound measurement as the gold standard for acetabular development this belongs to the pelvic radiograph (Fig. 3). While there is usefulness in having a visual sonographic clue regarding adequacy of acetabular morphology, it must not be forgotten that other acetabular pathology, like rounding of the superior lateral corner and increased echogenicity of the labrum, are not negated by a coverage measurement (Fig. 3). As a rule of thumb there is a logical basis for the 50% rule for acetabular development, as long as the difference between d/d (acetabular development) and d*/d (displacement) is understood and the 50% rule is accepted as a guide and not the golden rule. Compliance with ethical standards Conflicts of interest References None 1. American Institute of Ultrasound in Medicine, Laurel, MD (2013) AIUM practice parameter for the performance of an ultrasound examination for detection and assessment of developmental dysplasia of the hip. J Ultrasound Med 32:1307 1313 2. Graf R (1980) The diagnosis of congenital hip-joint dislocation by the ultrasonic Combound treatment. Arch Orthop Trauma Surg 97: 117 133 3. Graf R (1985) Ultrasonography of the infantile hip. In: Sanders RC, Hill MC (eds) Ultrasound annual. Raven, New York, pp 177 186
Pediatr Radiol (2017) 47:817 821 821 4. Harcke HT, Clarke NM, Lee MS et al (1998) Examination of the infant hip with real-time ultrasonography. J Ultrasound Med 3:131 137 5. Harcke HT, Grissom LE (1990) Performing dynamic sonography of the infant hip. AJR Am J Roentgenol 155:837 844 6. Harcke HT, Grissom LE (1994) Infant hip sonography: current concepts. Semin Ultrasound CT MR 15:256 263 7. Morin C, Harcke HT, MacEwen GD (1985) The infant hip: real-time US assessment of acetabular development. Radiology 157:673 677 8. Terjesen T, Runden TO, Tangerud A (1989) Ultrasonography and radiography of the hip in infants. Acta Orthop Scand 60:651 660 9. Morin C, Zouaoui S, Delvalle-Fayada A et al (1999) Ultrasound assessment of the acetabulum in the infant hip. Acta Orthop Belg 65:261 265 10. Gunay C, Atalar H, Dogruel H et al (2009) Correlation of femoral head coverage and Graf alpha angle in infants being screened for developmental dysplasia of the hip. Int Orthop 33:761 764 11. Pruszczynski B, Harcke HT, Holmes L Jr et al (2014) Natural history of hip instability in infants (without subluxation or dislocation): a three year follow-up. BMC Musculoskelet Disord 15:355