The Cleft Palate–Craniofacial Journal 00(00) pp. 000–000 Month 0000 Ó Copyright 2015 American Cleft Palate–Craniofacial Association
ORIGINAL ARTICLE Antenatal Ultrasound Detection of Cleft in Western Australia from 2003 to 2012: A Follow-Up Study W. Nicholls, B.Sc., B.A. (Psych), M.H.A., R. Jennings, B.D.Sc., Y. Yeung, M.B.B.S., M. Walters, B.Sc. (Hons), M.Sc., B. Hewitt, M.B.B.S., F.R.A.N.Z.C.O.G., D.D.U., C.O.G.U., D. Gillett, M.B.B.S., F.R.A.C.S. (Plas) Aim: To investigate trends in the rate of antenatal detection of cleft lip and palate (CLP) patients referred to the CLP Unit at Princess Margaret Hospital for Children in Western Australia during the period 2003–2012 and compare data with a previously published report covering the years 1996–2003. Methods: This is a single-center, retrospective survey of antenatal transabdominal ultrasound screenings of mothers of infants born between July 1, 2003 and June 30, 2012 that were referred to the CLP Unit at Princess Margaret Hospital. Results: Detection rates of oral clefts increased significantly when compared with outcomes reported in the same population between 1996 and 2003 (P , .05). An overall detection rate of 71.7% (165/230) was achieved for clefts involving lip and palate. Detection of isolated cleft palate (1/99) and microform (0/8) remained elusive. Most detections (76.5%) were achieved at 15 to 20 weeks of gestational age, corresponding with routine anatomical screening. A further 16.8% were detected post–20 weeks of gestation. Scans were performed by specialist obstetricians, and sonography clinics reported a detection rate of 84.6% (55/65), whereas nonspecialist clinics reported a detection rate of only 67.1% (110/164). Conclusion: The antenatal detection rates of oral clefts involving the lip have improved to the extent that the majority of mothers are now being referred to a cleft unit in Western Australia prior to the births of their children. As a result of this improvement, antenatal counseling is is now a common facet of cleft management. KEY WORDS: oral cleft, antenatal ultrasound detection, cleft lip and palate
In Western Australia (WA), antenatal anatomical screening that includes imaging of the face using transabdominal ultrasound is routinely conducted between 18 and 20 weeks gestation. Many expectant mothers have additional investigations at a range of gestational ages at the discretion of the treating physician. Investigations are performed by obstetricians, medical general practitioners, and sonographers in specialist and nonspecialist settings in WA. If an oral cleft is detected, parents are referred to the Cleft Lip and Palate (CLP) Unit at Princess Margaret Hospital for Children (PMH) for antenatal counseling and treatment discussions. PMH is the only referral center for children born with a CLP condition in WA. WA covers approximately 2.55 million square kilometers and has a population of 2.55 million people with an annual birth rate of 33,600 (Australian Bureau of Statistics, 2013). All children born in WA with a cleft condition are referred to the CLP Unit at PMH for initial assessment and ongoing management. The unit coordinator arranges for assessment by all relevant members of the multidisciplinary cleft care team, including plastic and reconstructive surgery, speech pathology, ear nose and throat, and pediatric dentistry, as well as extended care, including genetics, pediatrics, psychology, and social work as required.
Collectively, oral clefts involving the lip and/or palate (CL/P) represent the most common congenital anomaly of the head and neck region. Antenatal diagnosis of an oral cleft provides expectant parents an opportunity to prepare psychologically and materially for their child’s condition and special needs (Davalbhakta and Hall, 2000). Oral clefts are also indicators of other congenital anomalies associated with a syndromic diagnosis. Transabdominal ultrasound is the most commonly used imaging modality to screen for antenatal anatomical anomalies.
Ms. Nicholls is Clinical Research Associate, Princess Margaret Hospital for Children, Perth, Western Australia. Mr. Jennings is Paediatric Dental Registrar, Princess Margaret Hospital for Children, Perth, Western Australia. Dr. Yeung is General Practitioner, Princess Margaret Hospital for Children, Perth, Western Australia. Mr. Walters is Research Scientist, Princess Margaret Hospital for Children, Perth, Western Australia. Dr. Hewitt is Obstetrician/ Radiologist, Park Ultrasound, Western Australia. Dr. Gillett is Plastic and Reconstructive Surgeon, Princess Margaret Hospital for Children, Perth, Western Australia. Submitted May 2015; Revised June 2015; Accepted June 2015. Address correspondence to: Ms. Wendy Nicholls, Dental Department, Princess Margaret Hospital for Children, Roberts Road, Subiaco, WA 6008, Australia. E-mail [email protected]. gov.au. DOI: 10.1597/15-127 0
0
Cleft Palate–Craniofacial Journal, Month 0000, Vol. 00 No. 00
The sensitivity of detection of oral clefts using transabdominal ultrasound is variable, with reports ranging from 9% and 100% (Maarse et al., 2010). Factors such as the position of the fetus (Hegg et al, 1986), maternal obesity (Hegg et al., 1986), gestational age, and detection of other congenital abnormalities (Christ and Meininger, 1981) influence the rate of antenatal diagnosis. Minor labial clefts and isolated cleft palate is particularly challenging. Studies reporting successful detections are typically cohorts of highrisk groups screened by single-referral centers (Pilu et al., 1986; Hafner et al., 1997; Clementi et al., 2000). The specificity of oral cleft detection by routine transabdominal ultrasound in WA was initially investigated for the years 1996–2003 (Hanikeri et al., 2006). This retrospective survey of 216 parents of children born between January 1996 and June 2003 referred to the CLP Unit at PMH revealed an increase in the detection sensitivity during the study period. In particular, the detection sensitivity of bilateral cleft lip with or without palate (BCL/P) and unilateral cleft lip with or without palate (UCL/P) increased from 22.2% in 1997 to 66.7% in 2003. None of the 95 patients presenting with isolated cleft palate were detected, whereas only one of seven patients with a microform (MF) cleft lip was detected. The aim of this study was to investigate and compare cleft detection sensitivity using routine transabdominal ultrasound in WA for the follow-up period from July 1, 2003 to June 30, 2012. METHODS This study followed a single-center, retrospective survey design. The study population included parents of infants born between July 1, 2003 and June 30, 2012 referred to the CLP Unit at PMH for oral cleft treatment. Parents are routinely asked to complete a written epidemiological survey inclusive of antenatal screening within 12 weeks of the birth of their child. Survey questions were related to the ultrasound history of the pregnancy, including the number and timing of ultrasound scans, details of cleft detection, gestational age, sonographer, and ultrasound clinic they attended. The anatomic presentation of the cleft was accessed from the patient medical record and linked with survey responses. Data were entered into a secure database and de-identified for analysis. Cleft patients with incomplete or ambiguous responses to survey items regarding ultrasound history were excluded as well as those who returned their questionnaires outside of the 12-week return time and those with ultrasound screening conducted outside of WA. For the purposes of analysis and comparison with historic data, oral clefts were subdivided into four groups: UCL/P, BCL/P, microform (MF), and isolated cleft palate (CP). The sample was also stratified according to the imaging clinic attended (major maternity hospital in WA or
rural and remote centers) and obstetrician or imaging technologist. A comparison of the UCL/P and BCL/P detection sensitivities between the study periods was made using Fisher’s exact test of proportions modeled in Microsoft Excel 2010 (Microsoft, Redmond, WA). The Princess Margaret Hospital for Children’s ethics review committee approved this study. RESULTS A total of 561 infants were born with an oral cleft within the study period (2003–2012) and referred to the CLP Unit at PMH. A total of 337 (60%) had sufficient data to include in the study sample. Of the 337 cases, 40% were excluded due to the following reasons: not completing or not sufficiently completing the questionnaire or returning the questionnaire outside of the requested 12-week return period. The decision to only accept questionnaires within this return period was to mitigate recall bias even though it greatly reduced the numbers. The questionnaire response rate for UCL/P and BCL/P combined was 97.1% (230 from 238). The majority of the sample (n ¼ 300) had scans completed in the metropolitan area, and the remaining 37 pregnancies were screened in regional and remote areas of WA. Of the respondents, 75 indicated that their scans were completed in a specialist antenatal imaging center under the supervision and direct care of a medical physician. All respondents indicated that they had attended at least one ultrasound scan. Antenatal screenings were completed at 76 different sites, including four major hospital maternity units (19.6% of participants). The majority of participants were screened at private imaging centers in metropolitan areas, although 10.9% of participants utilized 14 imaging services and public hospitals located in regional or remote areas as defined by the Australian Institute of Health and Welfare Rural, Remote and Metropolitan Areas classification. Approximately one in four participants (n ¼ 96, 28.5%) indicated that their screening was completed by a physician who was either a consultant obstetrician or a general practitioner with training in obstetrics. The mean number of scans during pregnancy was 3.7 (61.6 SD), with a mean interval between scans of 6.6 weeks (62.4 SD). The study population was not a routine population; an anomaly was involved in which many were diagnosed prenatally. When an anomaly has been found, it is not unusual for additional ultrasound scans to be performed outside of the routine antenatal and later in the pregnancy, for example, to gain additional information on the anomaly itself or to double check for other anomalies, including cardiac or growth issues. The majority of mothers (57.6%, n¼194) presented for initial ultrasound examination within the first 15 to 20 weeks of pregnancy. A minority presented post–20 weeks gestation (6.8%) for their initial scan.
Nicholls et al., ANTENATAL ULTRASOUND DETECTION OF CLEFT IN WESTERN AUSTRALIA
FIGURE 1 Antenatal ultrasound first cleft detection sensitivity distributions for the two study periods by gestational age.
Figure 1 shows that the majority of detections in both studies with routine anatomical screening occurred at 15 to 20 weeks of gestational age, which indicates a shift toward earlier detection in the 2003–2012 study. A total of eight oral clefts were detected during ultrasound screening completed during the first 15 weeks gestation (4.8% of all detections). Of the 508 scans performed on undetected pregnancies between 15 and 20 weeks gestation, an additional 127 oral clefts were detected (76.5% of all detections); 61 mothers in this group had been previously scanned with no cleft detected. An additional 20 cases (10.8% of all detections) were detected between 21 and 29 weeks gestation, and 16 of these mothers had previously been scanned without detection. Following 30 weeks gestation, a total of 11 new detections were made (6.6%); all mothers had been examined several times prior to detection. Table 1 shows the antenatal detection of MF (0/8 cases) and isolated CP (1/107 cases) proved to be as challenging in the 2003–2012 study as in the 1996–2003 study (MF 1/7 cases and isolated CP 0/95 cases). The isolated CP case in the 2003–2012 study was detected by a hospital sonographer at a metropolitan hospital at 35 weeks gestational age. The highest detection sensitivity recorded was in the 2003–2012 study period for BCL/P, when 32 of 37 patients were detected (86.5%). Overall, the sensitivity of detection for all cleft types was 49.3% in the 2003–2012 study TABLE 1
compared with 22.2% in the 1996–2003 study. After the exclusion of MF and isolated CP types, the overall detection sensitivity was 71.7% in the 2003–2012 study compared with 41.2% in the 1996–2003 study. Fisher’s exact proportion test indicates this is a statistically significant improvement (P , .05). Figure 2 shows the antenatal detection sensitivity for combined UCL/P and BCL/P per annum across both studies. The detection sensitivity rate improved in the 2003– 2012 study although at a slower rate than that of the 1996– 2003 study, which is anticipated as detection nears saturation. Table 2 shows the antenatal detections for combined BCL/P and UCL/P cleft types for respondents who indicated their screening was completed by an obstetrician (which also includes medical physician) was 84.6% (55/65). This is a significantly higher rate of detection (P , .05) than scans on the same cleft subtypes in which respondents did not identify an obstetrician as their primary imaging consultant (67.1%, 110/164). After classifying oral clefts by subtype, a significant difference in the detection sensitivity was apparent only for unilateral cleft lip (UCL) (P , .05). The regional and remote screening sites (Australian Institute of Health and Welfare, 2004) completed scans for 37 infants born with oral clefts, including 13 with isolated CP. Table 3 shows the antenatal detection for combined BCL/P and UCL/P was 66.7% (16/24). King Edward Memorial Hospital, a major maternity hospital in WA specializing in high-risk births, completed scans for 36 infants, including seven with isolated CP. The combined BCL/P and UCL/P detection sensitivity of 93.1% (27/29) was significantly higher than for the regional and remote screening (P , .05). Significant differences (P , .05) were maintained for the UCL and CL subtypes only. DISCUSSION There is a clear relationship between cleft type and detection sensitivity with antenatal ultrasound screening. Both UCL/P and BCL/P typically present with significant protrusion of the dentoalveolus in the sagittal plane, which
Antenatal Ultrasound Detection Sensitivity Comparison for the Two Study Periods by Cleft Type 1996–2003
Cleft Type*
0
2003–2012
No. of Patients
Detected in Ultrasound
Sensitivity, %
No. of Patients
Detected in Ultrasound
Sensitivity, %
UCL/P BCL/P Subtotal
96 18 114
39 8 47
40.6 44.4 41.2
193 37 230
133 32 165
68.9 86.5 71.7
MF Isolated CP Subtotal
7 95 102
1 0 1
14.3 0.0 1.0
8 99 107
0 1 1
0.0 1.0 0.9
Grand total
216
48
22.2
337
166
49.3
* UCL/P ¼ unilateral cleft lip with or without palate; BCL/P ¼ bilateral cleft lip with or without palate; MF ¼ microform; CP ¼ isolated cleft palate.
0
Cleft Palate–Craniofacial Journal, Month 0000, Vol. 00 No. 00
FIGURE 2 Antenatal ultrasound detection sensitivity for the two study periods by year for combined unilateral cleft lip with or without palate and bilateral cleft lip with or without palate.
may explain the high proportion of detection of these cleft types (68.9% and 86.5%, respectively). Clefts of the lip, either isolated or with notching of the dentoalveolus, present with a lesser degree of facial distortion, which may be identified during imaging in the coronal plane. Cases involving minimal distortion of the lip remain challenging, with no presenting MF patient detected in this study. UCL was the next most challenging case, particularly for less experienced sonographers (52.4%) when compared with 77.4% detection sensitivity by obstetricians. In the 2003– 2013 study, 100% of bilateral cleft lip cases were detected. Detection of a single case in the cohort of 97 isolated palatal clefts was reported. Detection of CP with routine two-dimensional (2D) ultrasound is challenging largely due to the positioning of the tongue into the cleft and shadowing from adjacent anatomical structures. In some practice policies, CP detection is not attempted as part of an anatomy scan list. Detection has been achieved from retrospectively reviewed images from samples known to contain cases of diagnosed CP (Bundy et al., 1986; Goldstein et al., 1999; Wilhelm and Borgers, 2010). Although sonographic measures of dentoalveolar ridge width (Hafner et al., 1997), tongue excursion, and polyhydramnios have been suggested as diagnostic markers of CP, specificity is poor. In 2010, Wilhelm and Borgers reported an effective and efficacious ‘‘equals sign’’ technique using a 2D ultrasound transducer in novel positions to visualize the uvula and soft palate and facilitate detection TABLE 2 Antenatal Ultrasound Detection Sensitivity for the 2003–2012 Study by Obstetrician Versus Nonobstetrician
Cleft Subtype* BCL BCLP UCL UCLP Total
Obstetrician Sensitivity, %
Nonobstetrician Sensitivity, %
100.0 100.0 77.4 85.0 84.6
100.0 71.4 52.4 84.7 67.1
(5/5) (9/9) (24/31) (17/20) (55/65)
(2/2) (15/21) (43/82) (50/59) (110/164)
* BCL ¼ bilateral cleft lip; BCLP ¼ bilateral cleft lip and palate; UCL ¼ unilateral cleft lip; UCLP ¼ unilateral cleft lip and palate.
of CP during routine scans. Incorporation of this technique into scanning protocols and future teaching may facilitate improved detection of CP. The overall antenatal detection sensitivity of oral clefts improved significantly from 41.2% in the 1996–2003 study to 71.7% in the 2003–2012 study. The gestation age at which detection was first made also improved. In the 1996– 2003 study, there were no detections earlier than 15 weeks, and 62.5% were first detected within 15 to 20 weeks gestation. In the 2003–2012 study, 4.8% and 76.5% were detected in the same time periods. These results may be affected by the time at which the initial scans occurred because mothers may have presented at earlier gestational ages. Similarly, this may also have an impact on the results for the number of counts in scans from later presentations. Detection sensitivity across both studies may have been underreported because some scans may not have been anatomical scans but heartbeat scans, for which the data to differentiate are not available. Mothers in WA had ready access to modern sonographic equipment operated by well-trained sonographers for the 1996–2012 period. The improvements in detection may therefore be attributable to improvements in technology, image fidelity, and increased clinician awareness that have occurred since 2003. This includes, to some extent, specific training and inclusion of CLP in anatomy scan checklists. TABLE 3 Antenatal Ultrasound Detection Sensitivity for the 2003–2012 Study by Imaging Site
Cleft Subtype* BCL BCL/P UCL UCL/P Total
King Edward Memorial Hospital Sensitivity, % 100 100 92.3 90.9 93.1
(2/2) (3/3) (12/13) (10/11) (27/29)
Regional and Remote Sensitivity, % 0 100 41.7 83.3 66.7
(0/0) (6/6) (5/12) (6/6) (16/24)
* BCL ¼ bilateral cleft lip; BCLP ¼ bilateral cleft lip and palate; UCL ¼ unilateral cleft lip; UCLP ¼ unilateral cleft lip and palate.
Nicholls et al., ANTENATAL ULTRASOUND DETECTION OF CLEFT IN WESTERN AUSTRALIA
Early detection of the cleft condition facilitates preparation and planning for both parents and the cleft management team. Antenatal counseling is offered to expectant parents by the CLP team at PMH, where an overview of treatment and management for the unborn child is presented. Antenatal counseling may conversely also increase maternal anxiety during the term of pregnancy, which may be resolved following the birth of the child (Davalbhakta and Hall, 2000). A range of techniques have been suggested for antenatal detection with variable sensitivity and specificity. Fetoscopy (Maarse et al., 2010), intrauterine or transvaginal ultrasound (very early scans taken prior to 10 weeks GA), and assay of enzymes drawn from amniocentesis (Raposio et al., 1999; Clementi et al., 2000) have all proven successful in the detection of CLP antenatally. Use of these techniques, however, is limited by their invasiveness, maternal acceptance, and potential fetal morbidity and mortality risks. As definitive facial formation is not complete until 8 to 10 weeks of life, detection before this time is not feasible. Magnetic resonance imaging (MRI) is effective in noninvasively establishing the presence, laterality, and extension of oral clefts and is uniquely suited to antenatal detection of isolated CP (Clementi et al., 2000; Ghi et al., 2003; Maarse et al., 2010; Mailath-Pokorny et al., 2010), but it is also inhibitive because of expense and intensive use of resource. There is no health insurance rebate for fetal MRI in WA, and there is not nearly as much local experience, whereas for further assessment it has the potential to be used more to build up that experience. Fetal reconstructive surgery has been suggested for the management of CLP, advocated for the potential for scarless closure (Lorenz and Longaker, 2003), bone callous formation across the cleft and improved maxillary growth (Clementi et al., 2000; Papaopulous et al., 2005; Maarse et al., 2010). Should current technical and ethical limitations be overcome, successful feto-endoscopic surgery will depend on the timely detection of the malformation to ensure successful management. The current rates of detection would produce substantive cohorts of candidates for these potential procedures. Although this investigation focused on the use of 2D transabdominal ultrasound as a screening tool, many studies have emphasized the benefits of other imaging modalities in antenatal detection and diagnosis of oral clefts. In particular, MRI and three-dimensional (3D) ultrasound techniques have been recommended not only for efficacious detection but also clear diagnosis of cleft morphology of both intraoral and extraoral structures (Platt et al., 2006; Mailath-Pokorny et al., 2010). Although these technologies are available in WA, neither is used routinely during protocol screening of fetal development. In fact, 2D and 3D are extensions of each other rather than distinct modalities at a screening level, and the level of expertise is transferable. The availability and cost-efficacy of 2D ultrasound, coupled
0
with significant improvements in cleft detection and comparable outcomes to international studies, supports its continued use in routine antenatal screening. Alternative imaging techniques will continue to have significant value in WA, in particular the confirmation of findings during routine examination. In this survey, respondents who indicated their routine ultrasounds were completed by a physician (general practitioner or consultant obstetrician) showed a significantly higher rate of cleft detection than those whose screening was completed by a sonographer (P , .05). Dividing by subtype, a statistically significant difference was only maintained for cases of isolated CL. In contrast to unilateral cleft lip and palate and bilateral cleft lip and palate, CL is associated with considerably less volumetric change to 2D sonography and subsequently may be more difficult to detect. It is possible that postgraduate training in ultrasound detection of oral clefts may be a variable in successful antenatal diagnosis, particularly for subtle forms of pathology. REFERENCES Australian Bureau of Statistics. 2013. Available at http://www.abs. gov.au/ausstats/[email protected]/Lookup/1306.5mainþfeatures32014. Accessed June 12, 2015. Australian Institute of Health and Welfare. Rural, regional and remote health: a guide to remoteness classifications. Australian Institute of Health and Welfare catalogue no. PHE 53. Canberra: Australian Institute of Health and Welfare; 2004. Available at http://www.aihw.gov.au/publication-detail/?id¼6442467589. Bundy AL, Saltzmn DH, Emerson D, Fine C, Douilet P, Jones TB. Sonographic features associated with cleft palate. J Clin Ultrasound. 1986;14:486–489. Christ JE, Meininger MG. Ultrasound diagnosis of cleft lip and cleft palate before birth. Plast Reconstr Surg. 1981;68:854–859. Clementi M, Tenconi R, Bianchi F, Stoll C. Evaluation of prenatal diagnosis of cleft lip with or without cleft palate and cleft palate by ultrasound. Experience from 20 European registries. EUROSCAN Study Group. Prenat Diagn. 2000;20:870–875. Davalbhakta A, Hall PN. The impact of antenatal diagnosis and timing of counselling for cleft lip and palate. Br J Plast Surg. 2000;53:298–301. Ghi T, Tani G, Savelli L, Colleoni GG, Pilu G, Bovicelli L. Prenatal imaging of facial clefts by magnetic resonance imaging with emphasis on the posterior palate. Prenat Diagn. 2003;23:970–975. Goldstein I, Jakobi P, Tamir A, Goldstick O. Nomogram of the fetal alveolar ridge: a possible screening tool for the detection of primary cleft palate. Ultrasound Obstet Gynecol. 1999;14:333–337. Hafner E, Sterniste W, Scholler J, Schuchter K, Pjilipp K. Prenatal diagnosis of facial malformations. Prenat Diagn. 1997;17:51–58. Hanikeri M, Savundra J, Gillett D, Walters M, McBain W. Antenatal transabdominal ultrasound detection of cleft lip and palate in Western Australia from 1996 to 2003. Cleft Palate Craniofac J. 2006;43:61–66. Hegge FN, Prescott GH Watson PT. Fetal facial abnormalities identified during obstetric sonography. J Ultrasound Med. 1986;5:679–684. Lorenz HP, Longaker MT. In-utero surgery for cleft lip/palate: minimizing the ‘‘Ripple Effect’’ of scarring. J Craniofac Surg. 2003;14:504–511.
0
Cleft Palate–Craniofacial Journal, Month 0000, Vol. 00 No. 00
Maarse W, Berge´ SJ, Pistorius L, van Barneveld T, Kon M, Breugem C, Mink van der Molen AB. Diagnostic accuracy of transabdominal ultrasound in detecting prenatal cleft lip and palate: a systematic review. Ultrasound Obstet Gynecol. 2010;35:495–502. Maila´th-Pokorny M, Worda C, Krampl-Bettleheim E, Watzinger F, Brugger PC, Prayer D. What does magnetic resonance imaging add to the prenatal ultrasound diagnosis of facial clefts? Ultrasound Obstet Gynecol. 2010;36:445–451. Papadopulous NA, Papadopoulos MA, Kovacs L, Zeilhofer HF, Henke J, Boettcher P, Biemer E. Foetal surgery and cleft lip and palate current status and new perspectives. Br J Plast Surg. 2005;58:593–607.
Pilu G, Reece EA, Romero R, Bovicelli L, Hobbins JC. Prenatal diagnosis of craniofacial malformations with ultrasonography. Am J Obstet Gynecol. 1986;155:45–50. Platt LD, DeVore GR, Pretorius DH. Improving cleft palate/cleft lip antenatal diagnosis by 3-dimensional sonography: the ‘‘flipped face’’ view. J Ultrasound Med. 2006;25:1423–1430. Raposio E, Panarese P, Santi P. Fetal unilateral cleft lip and palate: detection of enzymic anomalies in the amniotic fluid. Last Reconstr Surg. 1999;103:391–394. Wilhelm L, Borgers H. The ‘‘equals sign’’: a novel marker in the diagnosis of fetal isolated cleft palate. Ultrasound Obstet Gynecol. 2010;36:439–444.
ORIGINAL ARTICLE Antenatal Ultrasound Detection of Cleft in Western Australia from 2003 to 2012: A Follow-Up Study W. Nicholls, B.Sc., B.A. (Psych), M.H.A., R. Jennings, B.D.Sc., Y. Yeung, M.B.B.S., M. Walters, B.Sc. (Hons), M.Sc., B. Hewitt, M.B.B.S., F.R.A.N.Z.C.O.G., D.D.U., C.O.G.U., D. Gillett, M.B.B.S., F.R.A.C.S. (Plas) Aim: To investigate trends in the rate of antenatal detection of cleft lip and palate (CLP) patients referred to the CLP Unit at Princess Margaret Hospital for Children in Western Australia during the period 2003–2012 and compare data with a previously published report covering the years 1996–2003. Methods: This is a single-center, retrospective survey of antenatal transabdominal ultrasound screenings of mothers of infants born between July 1, 2003 and June 30, 2012 that were referred to the CLP Unit at Princess Margaret Hospital. Results: Detection rates of oral clefts increased significantly when compared with outcomes reported in the same population between 1996 and 2003 (P , .05). An overall detection rate of 71.7% (165/230) was achieved for clefts involving lip and palate. Detection of isolated cleft palate (1/99) and microform (0/8) remained elusive. Most detections (76.5%) were achieved at 15 to 20 weeks of gestational age, corresponding with routine anatomical screening. A further 16.8% were detected post–20 weeks of gestation. Scans were performed by specialist obstetricians, and sonography clinics reported a detection rate of 84.6% (55/65), whereas nonspecialist clinics reported a detection rate of only 67.1% (110/164). Conclusion: The antenatal detection rates of oral clefts involving the lip have improved to the extent that the majority of mothers are now being referred to a cleft unit in Western Australia prior to the births of their children. As a result of this improvement, antenatal counseling is is now a common facet of cleft management. KEY WORDS: oral cleft, antenatal ultrasound detection, cleft lip and palate
In Western Australia (WA), antenatal anatomical screening that includes imaging of the face using transabdominal ultrasound is routinely conducted between 18 and 20 weeks gestation. Many expectant mothers have additional investigations at a range of gestational ages at the discretion of the treating physician. Investigations are performed by obstetricians, medical general practitioners, and sonographers in specialist and nonspecialist settings in WA. If an oral cleft is detected, parents are referred to the Cleft Lip and Palate (CLP) Unit at Princess Margaret Hospital for Children (PMH) for antenatal counseling and treatment discussions. PMH is the only referral center for children born with a CLP condition in WA. WA covers approximately 2.55 million square kilometers and has a population of 2.55 million people with an annual birth rate of 33,600 (Australian Bureau of Statistics, 2013). All children born in WA with a cleft condition are referred to the CLP Unit at PMH for initial assessment and ongoing management. The unit coordinator arranges for assessment by all relevant members of the multidisciplinary cleft care team, including plastic and reconstructive surgery, speech pathology, ear nose and throat, and pediatric dentistry, as well as extended care, including genetics, pediatrics, psychology, and social work as required.
Collectively, oral clefts involving the lip and/or palate (CL/P) represent the most common congenital anomaly of the head and neck region. Antenatal diagnosis of an oral cleft provides expectant parents an opportunity to prepare psychologically and materially for their child’s condition and special needs (Davalbhakta and Hall, 2000). Oral clefts are also indicators of other congenital anomalies associated with a syndromic diagnosis. Transabdominal ultrasound is the most commonly used imaging modality to screen for antenatal anatomical anomalies.
Ms. Nicholls is Clinical Research Associate, Princess Margaret Hospital for Children, Perth, Western Australia. Mr. Jennings is Paediatric Dental Registrar, Princess Margaret Hospital for Children, Perth, Western Australia. Dr. Yeung is General Practitioner, Princess Margaret Hospital for Children, Perth, Western Australia. Mr. Walters is Research Scientist, Princess Margaret Hospital for Children, Perth, Western Australia. Dr. Hewitt is Obstetrician/ Radiologist, Park Ultrasound, Western Australia. Dr. Gillett is Plastic and Reconstructive Surgeon, Princess Margaret Hospital for Children, Perth, Western Australia. Submitted May 2015; Revised June 2015; Accepted June 2015. Address correspondence to: Ms. Wendy Nicholls, Dental Department, Princess Margaret Hospital for Children, Roberts Road, Subiaco, WA 6008, Australia. E-mail [email protected]. gov.au. DOI: 10.1597/15-127 0
0
Cleft Palate–Craniofacial Journal, Month 0000, Vol. 00 No. 00
The sensitivity of detection of oral clefts using transabdominal ultrasound is variable, with reports ranging from 9% and 100% (Maarse et al., 2010). Factors such as the position of the fetus (Hegg et al, 1986), maternal obesity (Hegg et al., 1986), gestational age, and detection of other congenital abnormalities (Christ and Meininger, 1981) influence the rate of antenatal diagnosis. Minor labial clefts and isolated cleft palate is particularly challenging. Studies reporting successful detections are typically cohorts of highrisk groups screened by single-referral centers (Pilu et al., 1986; Hafner et al., 1997; Clementi et al., 2000). The specificity of oral cleft detection by routine transabdominal ultrasound in WA was initially investigated for the years 1996–2003 (Hanikeri et al., 2006). This retrospective survey of 216 parents of children born between January 1996 and June 2003 referred to the CLP Unit at PMH revealed an increase in the detection sensitivity during the study period. In particular, the detection sensitivity of bilateral cleft lip with or without palate (BCL/P) and unilateral cleft lip with or without palate (UCL/P) increased from 22.2% in 1997 to 66.7% in 2003. None of the 95 patients presenting with isolated cleft palate were detected, whereas only one of seven patients with a microform (MF) cleft lip was detected. The aim of this study was to investigate and compare cleft detection sensitivity using routine transabdominal ultrasound in WA for the follow-up period from July 1, 2003 to June 30, 2012. METHODS This study followed a single-center, retrospective survey design. The study population included parents of infants born between July 1, 2003 and June 30, 2012 referred to the CLP Unit at PMH for oral cleft treatment. Parents are routinely asked to complete a written epidemiological survey inclusive of antenatal screening within 12 weeks of the birth of their child. Survey questions were related to the ultrasound history of the pregnancy, including the number and timing of ultrasound scans, details of cleft detection, gestational age, sonographer, and ultrasound clinic they attended. The anatomic presentation of the cleft was accessed from the patient medical record and linked with survey responses. Data were entered into a secure database and de-identified for analysis. Cleft patients with incomplete or ambiguous responses to survey items regarding ultrasound history were excluded as well as those who returned their questionnaires outside of the 12-week return time and those with ultrasound screening conducted outside of WA. For the purposes of analysis and comparison with historic data, oral clefts were subdivided into four groups: UCL/P, BCL/P, microform (MF), and isolated cleft palate (CP). The sample was also stratified according to the imaging clinic attended (major maternity hospital in WA or
rural and remote centers) and obstetrician or imaging technologist. A comparison of the UCL/P and BCL/P detection sensitivities between the study periods was made using Fisher’s exact test of proportions modeled in Microsoft Excel 2010 (Microsoft, Redmond, WA). The Princess Margaret Hospital for Children’s ethics review committee approved this study. RESULTS A total of 561 infants were born with an oral cleft within the study period (2003–2012) and referred to the CLP Unit at PMH. A total of 337 (60%) had sufficient data to include in the study sample. Of the 337 cases, 40% were excluded due to the following reasons: not completing or not sufficiently completing the questionnaire or returning the questionnaire outside of the requested 12-week return period. The decision to only accept questionnaires within this return period was to mitigate recall bias even though it greatly reduced the numbers. The questionnaire response rate for UCL/P and BCL/P combined was 97.1% (230 from 238). The majority of the sample (n ¼ 300) had scans completed in the metropolitan area, and the remaining 37 pregnancies were screened in regional and remote areas of WA. Of the respondents, 75 indicated that their scans were completed in a specialist antenatal imaging center under the supervision and direct care of a medical physician. All respondents indicated that they had attended at least one ultrasound scan. Antenatal screenings were completed at 76 different sites, including four major hospital maternity units (19.6% of participants). The majority of participants were screened at private imaging centers in metropolitan areas, although 10.9% of participants utilized 14 imaging services and public hospitals located in regional or remote areas as defined by the Australian Institute of Health and Welfare Rural, Remote and Metropolitan Areas classification. Approximately one in four participants (n ¼ 96, 28.5%) indicated that their screening was completed by a physician who was either a consultant obstetrician or a general practitioner with training in obstetrics. The mean number of scans during pregnancy was 3.7 (61.6 SD), with a mean interval between scans of 6.6 weeks (62.4 SD). The study population was not a routine population; an anomaly was involved in which many were diagnosed prenatally. When an anomaly has been found, it is not unusual for additional ultrasound scans to be performed outside of the routine antenatal and later in the pregnancy, for example, to gain additional information on the anomaly itself or to double check for other anomalies, including cardiac or growth issues. The majority of mothers (57.6%, n¼194) presented for initial ultrasound examination within the first 15 to 20 weeks of pregnancy. A minority presented post–20 weeks gestation (6.8%) for their initial scan.
Nicholls et al., ANTENATAL ULTRASOUND DETECTION OF CLEFT IN WESTERN AUSTRALIA
FIGURE 1 Antenatal ultrasound first cleft detection sensitivity distributions for the two study periods by gestational age.
Figure 1 shows that the majority of detections in both studies with routine anatomical screening occurred at 15 to 20 weeks of gestational age, which indicates a shift toward earlier detection in the 2003–2012 study. A total of eight oral clefts were detected during ultrasound screening completed during the first 15 weeks gestation (4.8% of all detections). Of the 508 scans performed on undetected pregnancies between 15 and 20 weeks gestation, an additional 127 oral clefts were detected (76.5% of all detections); 61 mothers in this group had been previously scanned with no cleft detected. An additional 20 cases (10.8% of all detections) were detected between 21 and 29 weeks gestation, and 16 of these mothers had previously been scanned without detection. Following 30 weeks gestation, a total of 11 new detections were made (6.6%); all mothers had been examined several times prior to detection. Table 1 shows the antenatal detection of MF (0/8 cases) and isolated CP (1/107 cases) proved to be as challenging in the 2003–2012 study as in the 1996–2003 study (MF 1/7 cases and isolated CP 0/95 cases). The isolated CP case in the 2003–2012 study was detected by a hospital sonographer at a metropolitan hospital at 35 weeks gestational age. The highest detection sensitivity recorded was in the 2003–2012 study period for BCL/P, when 32 of 37 patients were detected (86.5%). Overall, the sensitivity of detection for all cleft types was 49.3% in the 2003–2012 study TABLE 1
compared with 22.2% in the 1996–2003 study. After the exclusion of MF and isolated CP types, the overall detection sensitivity was 71.7% in the 2003–2012 study compared with 41.2% in the 1996–2003 study. Fisher’s exact proportion test indicates this is a statistically significant improvement (P , .05). Figure 2 shows the antenatal detection sensitivity for combined UCL/P and BCL/P per annum across both studies. The detection sensitivity rate improved in the 2003– 2012 study although at a slower rate than that of the 1996– 2003 study, which is anticipated as detection nears saturation. Table 2 shows the antenatal detections for combined BCL/P and UCL/P cleft types for respondents who indicated their screening was completed by an obstetrician (which also includes medical physician) was 84.6% (55/65). This is a significantly higher rate of detection (P , .05) than scans on the same cleft subtypes in which respondents did not identify an obstetrician as their primary imaging consultant (67.1%, 110/164). After classifying oral clefts by subtype, a significant difference in the detection sensitivity was apparent only for unilateral cleft lip (UCL) (P , .05). The regional and remote screening sites (Australian Institute of Health and Welfare, 2004) completed scans for 37 infants born with oral clefts, including 13 with isolated CP. Table 3 shows the antenatal detection for combined BCL/P and UCL/P was 66.7% (16/24). King Edward Memorial Hospital, a major maternity hospital in WA specializing in high-risk births, completed scans for 36 infants, including seven with isolated CP. The combined BCL/P and UCL/P detection sensitivity of 93.1% (27/29) was significantly higher than for the regional and remote screening (P , .05). Significant differences (P , .05) were maintained for the UCL and CL subtypes only. DISCUSSION There is a clear relationship between cleft type and detection sensitivity with antenatal ultrasound screening. Both UCL/P and BCL/P typically present with significant protrusion of the dentoalveolus in the sagittal plane, which
Antenatal Ultrasound Detection Sensitivity Comparison for the Two Study Periods by Cleft Type 1996–2003
Cleft Type*
0
2003–2012
No. of Patients
Detected in Ultrasound
Sensitivity, %
No. of Patients
Detected in Ultrasound
Sensitivity, %
UCL/P BCL/P Subtotal
96 18 114
39 8 47
40.6 44.4 41.2
193 37 230
133 32 165
68.9 86.5 71.7
MF Isolated CP Subtotal
7 95 102
1 0 1
14.3 0.0 1.0
8 99 107
0 1 1
0.0 1.0 0.9
Grand total
216
48
22.2
337
166
49.3
* UCL/P ¼ unilateral cleft lip with or without palate; BCL/P ¼ bilateral cleft lip with or without palate; MF ¼ microform; CP ¼ isolated cleft palate.
0
Cleft Palate–Craniofacial Journal, Month 0000, Vol. 00 No. 00
FIGURE 2 Antenatal ultrasound detection sensitivity for the two study periods by year for combined unilateral cleft lip with or without palate and bilateral cleft lip with or without palate.
may explain the high proportion of detection of these cleft types (68.9% and 86.5%, respectively). Clefts of the lip, either isolated or with notching of the dentoalveolus, present with a lesser degree of facial distortion, which may be identified during imaging in the coronal plane. Cases involving minimal distortion of the lip remain challenging, with no presenting MF patient detected in this study. UCL was the next most challenging case, particularly for less experienced sonographers (52.4%) when compared with 77.4% detection sensitivity by obstetricians. In the 2003– 2013 study, 100% of bilateral cleft lip cases were detected. Detection of a single case in the cohort of 97 isolated palatal clefts was reported. Detection of CP with routine two-dimensional (2D) ultrasound is challenging largely due to the positioning of the tongue into the cleft and shadowing from adjacent anatomical structures. In some practice policies, CP detection is not attempted as part of an anatomy scan list. Detection has been achieved from retrospectively reviewed images from samples known to contain cases of diagnosed CP (Bundy et al., 1986; Goldstein et al., 1999; Wilhelm and Borgers, 2010). Although sonographic measures of dentoalveolar ridge width (Hafner et al., 1997), tongue excursion, and polyhydramnios have been suggested as diagnostic markers of CP, specificity is poor. In 2010, Wilhelm and Borgers reported an effective and efficacious ‘‘equals sign’’ technique using a 2D ultrasound transducer in novel positions to visualize the uvula and soft palate and facilitate detection TABLE 2 Antenatal Ultrasound Detection Sensitivity for the 2003–2012 Study by Obstetrician Versus Nonobstetrician
Cleft Subtype* BCL BCLP UCL UCLP Total
Obstetrician Sensitivity, %
Nonobstetrician Sensitivity, %
100.0 100.0 77.4 85.0 84.6
100.0 71.4 52.4 84.7 67.1
(5/5) (9/9) (24/31) (17/20) (55/65)
(2/2) (15/21) (43/82) (50/59) (110/164)
* BCL ¼ bilateral cleft lip; BCLP ¼ bilateral cleft lip and palate; UCL ¼ unilateral cleft lip; UCLP ¼ unilateral cleft lip and palate.
of CP during routine scans. Incorporation of this technique into scanning protocols and future teaching may facilitate improved detection of CP. The overall antenatal detection sensitivity of oral clefts improved significantly from 41.2% in the 1996–2003 study to 71.7% in the 2003–2012 study. The gestation age at which detection was first made also improved. In the 1996– 2003 study, there were no detections earlier than 15 weeks, and 62.5% were first detected within 15 to 20 weeks gestation. In the 2003–2012 study, 4.8% and 76.5% were detected in the same time periods. These results may be affected by the time at which the initial scans occurred because mothers may have presented at earlier gestational ages. Similarly, this may also have an impact on the results for the number of counts in scans from later presentations. Detection sensitivity across both studies may have been underreported because some scans may not have been anatomical scans but heartbeat scans, for which the data to differentiate are not available. Mothers in WA had ready access to modern sonographic equipment operated by well-trained sonographers for the 1996–2012 period. The improvements in detection may therefore be attributable to improvements in technology, image fidelity, and increased clinician awareness that have occurred since 2003. This includes, to some extent, specific training and inclusion of CLP in anatomy scan checklists. TABLE 3 Antenatal Ultrasound Detection Sensitivity for the 2003–2012 Study by Imaging Site
Cleft Subtype* BCL BCL/P UCL UCL/P Total
King Edward Memorial Hospital Sensitivity, % 100 100 92.3 90.9 93.1
(2/2) (3/3) (12/13) (10/11) (27/29)
Regional and Remote Sensitivity, % 0 100 41.7 83.3 66.7
(0/0) (6/6) (5/12) (6/6) (16/24)
* BCL ¼ bilateral cleft lip; BCLP ¼ bilateral cleft lip and palate; UCL ¼ unilateral cleft lip; UCLP ¼ unilateral cleft lip and palate.
Nicholls et al., ANTENATAL ULTRASOUND DETECTION OF CLEFT IN WESTERN AUSTRALIA
Early detection of the cleft condition facilitates preparation and planning for both parents and the cleft management team. Antenatal counseling is offered to expectant parents by the CLP team at PMH, where an overview of treatment and management for the unborn child is presented. Antenatal counseling may conversely also increase maternal anxiety during the term of pregnancy, which may be resolved following the birth of the child (Davalbhakta and Hall, 2000). A range of techniques have been suggested for antenatal detection with variable sensitivity and specificity. Fetoscopy (Maarse et al., 2010), intrauterine or transvaginal ultrasound (very early scans taken prior to 10 weeks GA), and assay of enzymes drawn from amniocentesis (Raposio et al., 1999; Clementi et al., 2000) have all proven successful in the detection of CLP antenatally. Use of these techniques, however, is limited by their invasiveness, maternal acceptance, and potential fetal morbidity and mortality risks. As definitive facial formation is not complete until 8 to 10 weeks of life, detection before this time is not feasible. Magnetic resonance imaging (MRI) is effective in noninvasively establishing the presence, laterality, and extension of oral clefts and is uniquely suited to antenatal detection of isolated CP (Clementi et al., 2000; Ghi et al., 2003; Maarse et al., 2010; Mailath-Pokorny et al., 2010), but it is also inhibitive because of expense and intensive use of resource. There is no health insurance rebate for fetal MRI in WA, and there is not nearly as much local experience, whereas for further assessment it has the potential to be used more to build up that experience. Fetal reconstructive surgery has been suggested for the management of CLP, advocated for the potential for scarless closure (Lorenz and Longaker, 2003), bone callous formation across the cleft and improved maxillary growth (Clementi et al., 2000; Papaopulous et al., 2005; Maarse et al., 2010). Should current technical and ethical limitations be overcome, successful feto-endoscopic surgery will depend on the timely detection of the malformation to ensure successful management. The current rates of detection would produce substantive cohorts of candidates for these potential procedures. Although this investigation focused on the use of 2D transabdominal ultrasound as a screening tool, many studies have emphasized the benefits of other imaging modalities in antenatal detection and diagnosis of oral clefts. In particular, MRI and three-dimensional (3D) ultrasound techniques have been recommended not only for efficacious detection but also clear diagnosis of cleft morphology of both intraoral and extraoral structures (Platt et al., 2006; Mailath-Pokorny et al., 2010). Although these technologies are available in WA, neither is used routinely during protocol screening of fetal development. In fact, 2D and 3D are extensions of each other rather than distinct modalities at a screening level, and the level of expertise is transferable. The availability and cost-efficacy of 2D ultrasound, coupled
0
with significant improvements in cleft detection and comparable outcomes to international studies, supports its continued use in routine antenatal screening. Alternative imaging techniques will continue to have significant value in WA, in particular the confirmation of findings during routine examination. In this survey, respondents who indicated their routine ultrasounds were completed by a physician (general practitioner or consultant obstetrician) showed a significantly higher rate of cleft detection than those whose screening was completed by a sonographer (P , .05). Dividing by subtype, a statistically significant difference was only maintained for cases of isolated CL. In contrast to unilateral cleft lip and palate and bilateral cleft lip and palate, CL is associated with considerably less volumetric change to 2D sonography and subsequently may be more difficult to detect. It is possible that postgraduate training in ultrasound detection of oral clefts may be a variable in successful antenatal diagnosis, particularly for subtle forms of pathology. REFERENCES Australian Bureau of Statistics. 2013. Available at http://www.abs. gov.au/ausstats/[email protected]/Lookup/1306.5mainþfeatures32014. Accessed June 12, 2015. Australian Institute of Health and Welfare. Rural, regional and remote health: a guide to remoteness classifications. Australian Institute of Health and Welfare catalogue no. PHE 53. Canberra: Australian Institute of Health and Welfare; 2004. Available at http://www.aihw.gov.au/publication-detail/?id¼6442467589. Bundy AL, Saltzmn DH, Emerson D, Fine C, Douilet P, Jones TB. Sonographic features associated with cleft palate. J Clin Ultrasound. 1986;14:486–489. Christ JE, Meininger MG. Ultrasound diagnosis of cleft lip and cleft palate before birth. Plast Reconstr Surg. 1981;68:854–859. Clementi M, Tenconi R, Bianchi F, Stoll C. Evaluation of prenatal diagnosis of cleft lip with or without cleft palate and cleft palate by ultrasound. Experience from 20 European registries. EUROSCAN Study Group. Prenat Diagn. 2000;20:870–875. Davalbhakta A, Hall PN. The impact of antenatal diagnosis and timing of counselling for cleft lip and palate. Br J Plast Surg. 2000;53:298–301. Ghi T, Tani G, Savelli L, Colleoni GG, Pilu G, Bovicelli L. Prenatal imaging of facial clefts by magnetic resonance imaging with emphasis on the posterior palate. Prenat Diagn. 2003;23:970–975. Goldstein I, Jakobi P, Tamir A, Goldstick O. Nomogram of the fetal alveolar ridge: a possible screening tool for the detection of primary cleft palate. Ultrasound Obstet Gynecol. 1999;14:333–337. Hafner E, Sterniste W, Scholler J, Schuchter K, Pjilipp K. Prenatal diagnosis of facial malformations. Prenat Diagn. 1997;17:51–58. Hanikeri M, Savundra J, Gillett D, Walters M, McBain W. Antenatal transabdominal ultrasound detection of cleft lip and palate in Western Australia from 1996 to 2003. Cleft Palate Craniofac J. 2006;43:61–66. Hegge FN, Prescott GH Watson PT. Fetal facial abnormalities identified during obstetric sonography. J Ultrasound Med. 1986;5:679–684. Lorenz HP, Longaker MT. In-utero surgery for cleft lip/palate: minimizing the ‘‘Ripple Effect’’ of scarring. J Craniofac Surg. 2003;14:504–511.
0
Cleft Palate–Craniofacial Journal, Month 0000, Vol. 00 No. 00
Maarse W, Berge´ SJ, Pistorius L, van Barneveld T, Kon M, Breugem C, Mink van der Molen AB. Diagnostic accuracy of transabdominal ultrasound in detecting prenatal cleft lip and palate: a systematic review. Ultrasound Obstet Gynecol. 2010;35:495–502. Maila´th-Pokorny M, Worda C, Krampl-Bettleheim E, Watzinger F, Brugger PC, Prayer D. What does magnetic resonance imaging add to the prenatal ultrasound diagnosis of facial clefts? Ultrasound Obstet Gynecol. 2010;36:445–451. Papadopulous NA, Papadopoulos MA, Kovacs L, Zeilhofer HF, Henke J, Boettcher P, Biemer E. Foetal surgery and cleft lip and palate current status and new perspectives. Br J Plast Surg. 2005;58:593–607.
Pilu G, Reece EA, Romero R, Bovicelli L, Hobbins JC. Prenatal diagnosis of craniofacial malformations with ultrasonography. Am J Obstet Gynecol. 1986;155:45–50. Platt LD, DeVore GR, Pretorius DH. Improving cleft palate/cleft lip antenatal diagnosis by 3-dimensional sonography: the ‘‘flipped face’’ view. J Ultrasound Med. 2006;25:1423–1430. Raposio E, Panarese P, Santi P. Fetal unilateral cleft lip and palate: detection of enzymic anomalies in the amniotic fluid. Last Reconstr Surg. 1999;103:391–394. Wilhelm L, Borgers H. The ‘‘equals sign’’: a novel marker in the diagnosis of fetal isolated cleft palate. Ultrasound Obstet Gynecol. 2010;36:439–444.
