Vous êtes sur la page 1sur 8

European Journal of Radiology 81 (2012) e344e351

Contents lists available at SciVerse ScienceDirect

European Journal of Radiology


journal homepage: www.elsevier.com/locate/ejrad

Review

A meta-analysis of common risk factors associated with the diagnosis of


developmental dysplasia of the hip in newborns
Clara L. Ortiz-Neira a , Elizabeth Oddone Paolucci b , Tyrone Donnon c,
a
Department of Radiology, Alberta Childrens Hospital, University of Calgary, Faculty of Medicine, 2888 Shaganappi Trail NW, Calgary, Alberta, Canada T3B 6A81
b
Departments of Surgery & Community Health Sciences, Faculty of Medicine, University of Calgary, FMC, North Tower Room #1026, 1403 - 29 Street NW, Calgary, Alberta, Canada
T2N 2T9
c
Department of Community Health Sciences, Faculty of Medicine, University of Calgary, G13 Heritage Medical Research Centre, 3330 Hospital Drive NW, Calgary, Alberta, Canada
T2N 4N1

a r t i c l e i n f o a b s t r a c t

Article history: Background: Although there is no clear consensus about the process of screening for developmental
Received 29 July 2011 dysplasia of the hip (DDH), there are six common risk factors associated with DDH in patients less than
Received in revised form 6 months of age (breech presentation, sex, family history, rst-born, side of hip, and mode of delivery).
24 September 2011
Methods: A meta-analysis of published studies was conducted to identify the relative risk ratio of the six
Accepted 3 November 2011
commonly known risk factors. A total of 31 primary studies consisting of 20,196 DDH patients met the
following inclusion criteria: (1) contained empirical data on at least one common risk factor, (2) were
Keywords:
peer-reviewed from an English language scientic journal, (3) included patients less or equal to 6 months
Developmental dysplasia of the hip
Meta-analysis
of age, and (4) identied method of diagnosis (e.g., ultrasound, radiographs or clinical examination).
Ultrasound Results: Fixed effect and random effects models with 95% condence intervals were calculated for each of
Radiology the six risk factors. Reported relative risk ratio (RR) for each factor in newborns was: breech presentation
Risk factors 3.75 (95% CI: 2.256.24), females 2.54 (95% CI: 2.113.05), left hip side 1.54 (95% CI: 1.251.90), rst born
1.44 (95% CI: 1.121.86), and family history 1.39 (95% CI: 1.231.57). A non-signicant RR value of 1.22
(95% CI: 0.463.23) was found for mode of delivery.
Conclusion: Results suggest that ultrasound and radiology screening methods be used to conrm DDH
in newborns that present with one or a combination of the following common risk factors: breech
presentation, female, left hip affected, rst born and family history of DDH.
2011 Elsevier Ireland Ltd. All rights reserved.

1. Introduction hip, dislocatable hip, and dislocated hip [4], while not excluding the
possibility of the late development of the condition in underdiag-
Developmental dysplasia of the hip (DDH), formerly known as nosed or mild acetabular dysplasia that progresses to instability
congenital dislocation of the hip (CDH) comprises a spectrum of [5].
abnormalities of the hip in infants and children that includes an The preferred screening method for obtaining a prompt diag-
immature hip, acetabular dysplasia with or without dislocation, nosis of DDH remains controversial. Presently, main methods
subluxation of the femoral head as well as being dislocatable and for screening include clinical examination and ultrasound. Pelvic
subluxable [1]. Although some of these changes may resolve spon- radiograph has been used after 4 months of age to assure bone
taneously, if not diagnosed or monitored during infancy, the hip will congruency [6]. Clinical examination is the most commonly used
worsen with time [2]. As some reports [3,4] have documented, dis- method as universal screening for DDH. However, while some
location or dysplasia of the hip can occur after a normal neonatal places in Europe have used ultrasound as a universal screening
screening examination. Thus, the term DDH is preferred because method, others (like in the USA and Canada) use it only for patients
it more accurately refers to a spectrum of abnormalities. These at risk for DDH. The American Academy of Pediatrics (AAP) sub-
abnormalities include: abnormal acetabular shape (i.e., dysplasia, committee on DDH estimated that in about 15% of cases, DDH is not
isolated dysplasia) and/or partial or incomplete displacement of the detectible at birth even by experienced examiners or ultrasonogra-
femoral head from the acetabulum (i.e., subluxed hip, subluxable phers [6,7]. One of the resulting implications of employing different
denitions and diagnostic methods of DDH, is that reported inci-
dence rates of DDH vary widely [811]. More importantly, failure
Corresponding author. Tel.: +1 403 210 9682; fax: +1 403 270 7507. to promptly and accurately diagnose DDH may result in serious
E-mail address: tldonnon@ucalgary.ca (T. Donnon). long-term health care problems such as early degenerative arthritis
1
Tel.: +1 403 955 7985; fax: +1 403 210 7507. [1,6].

0720-048X/$ see front matter 2011 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ejrad.2011.11.003

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351 e345

Risk factors among newborns are associated in 5060% of cases 2.3. Data extraction
with DDH [1217]. We limited our meta-analysis to the idiopathic
risk factors that occur in completely healthy children who do not A standardized protocol used for data extraction on the studies
have obvious dislocations [6]. These risk factors put the patient at to be included in the present meta-analysis was developed by all
risk for occult dysplasia that may or may not be associated with three authors based on a careful review of the existing DDH litera-
dislocation. They include hip side, pre-natal positioning (breech ture. From this extensive review, the common risk factors identied
presentation), genetic factors (family history, sex), and environ- in the literature were selected for investigation (i.e., breech, pre-
mental factors (rst born, mode of delivery). While these main risk sentation, sex, family history, rst born, side of hip and mode of
factors have been identied as increasing the prevalence of DDH delivery). The coding protocol included information on three main
[6], their relative effect size ratios have not been conrmed [18]. domains. The rst involved basic demographics of the studies,
Subsequently, the main goal was to perform a meta-analysis on such as title of the study, number of authors and their names,
the ve most common risk factors associated with DDH in patients year and source of publication, sex and country of afliation of
less than 6 months of age. A secondary purpose was to calculate the rst author, and country the study was undertaken. The second
prevalence of DDH and identify the methods (clinical exam, ultra- domain involved recording information on the quality and strength
sound, radiology) most commonly used to diagnosis DDH based on of each study. There are various systems (some highly intricate and
the studies included in this meta-analysis. complex) used in evidence-based medicine to categorize and rank
different types of clinical evidence according to the strength of their
freedom from various bias. The system we adopted was an ABC
2. Methods
grading scale [19], whereby three levels are identied: (1) Level
A grade is assigned to a randomized controlled trial (RCT), meta-
2.1. Study selection
analysis, or quantitative systematic review; (2) Level B grade refers
to a well-designed nonrandomized clinical trial, non quantitative
To be included in the present meta-analysis, studies had to: (1)
systematic review, lower quality RCT, clinical cohort study, case-
contain empirical data on at least one risk factor (i.e., birth presen-
controlled study, historical uncontrolled study, or epidemiologic
tation, birth order, sex, hip side, family history of DDH, or mode
study; and (3) Level C grade comprises articles based on consen-
of delivery), with sufcient statistical information (e.g., incidence
sus or expert opinion. In addition to this, total sample size was
rates, mean and standard deviation, and F ratio and t-test statis-
recorded for each study, along with the types of statistical analyses
tics) to allow for the calculation of relative risk or effect sizes; (2)
conducted on the study data (i.e., simple frequencies, univariate,
be peer-reviewed from an English-language scientic journal; (3)
multivariate, and/or nonparametric statistics).
include patients less than or equal to 6 months of age, having under-
Finally, information reported in the studies on the specics of
gone a clinical procedure specically designed to diagnose or screen
DDH was also recorded, such as the: (1) denition of DDH if it was
for DDH; (4) describe clinical or imaging diagnostic approaches to
presented in the article; (2) incidence of DDH as it pertained to
DDH; and (5) measure the degree of DDH by the Graf and Harke
the risk factors reported; (3) method of DDH detection (i.e., clin-
rating scales for ultrasound, acetabular index for radiographs, and
ical exam universal, selective clinical exam, universal ultrasound,
Ortolani and Barlow maneuver for clinical examinations.
selective ultrasound, or radiography); and (4) risk factors or moder-
Articles were excluded from the analysis if the: (1) empirical
ators of DDH, such as delivery mode (vaginal or caesarean), delivery
data reported in the study was already published elsewhere; (2)
presentation (breech or non-breech), birth order (rst born or later
article focused exclusively on DDH treatment; and (3) subjects
born), neonate sex (male or female), hip side affected (left, right, or
included premature babies and infants who had neuromuscular
both), and DDH family history (positive or negative).
disturbances, torticollis, scoliosis, postural talipes, foot deformities,
The coding protocol was pilot-tested by having all three authors
postnatal swaddling, oligohydramnios, and associated syndromes,
independently extract data on 10 of the total 31 studies included
or teratologic DDH.
in the meta-analysis. Two of the authors (CON and EOP) proceeded
to code all remaining studies independently. In this initial step the
2.2. Data sources and searches concordance agreement between the initial two raters was 97% fol-
lowed by a 100% concordance agreement when the third rater (TD)
An electronic search for all English-language, empirical stud- was consulted.
ies was conducted on associated risk factors of DDH using The
National Library of Medicines PubMed search engine, the Science 2.4. Data synthesis and statistical analysis
Citation Index, and Cochrane Data base of Controlled Trials (from
1966 to March 2011). Initially, keywords were identied and MeSH Data were extracted from each study based on the standard-
descriptors or subject headings were arranged for the following ized protocol and identied DDH risk factors of interest. While
terms: developmental dysplasia of the hip, DDH, congenital some studies reported data on only one of the risk factors, oth-
dislocation of the hip, CDH, risk factors, diagnosis, screening ers examined multiple risk factors. Effect sizes for each study were
methods, humans, and meta-analysis resulting in the return of computed using random relative risk (RR) ratio calculations along
a total of 466 articles. The authors used the inclusion criteria as a with the 95% condence limits. In addition, combined or over-
guide in reviewing all titles and abstracts, and subsequently elim- all effects across the DDH risk factors were computed using both
inated 276 studies that did not meet the minimum requirements xed and random effects computational models. Although a xed-
(i.e., were not empirical, not peer-reviewed, used duplicate data, effects model assumes that the assessment of the DDH risk factor(s)
or focused on DDH treatments). Copies of full publications for 190 between studies is the same, the random-effects model incorpo-
studies were obtained and a manual hand search of the reference rates the more common effects of heterogeneity in the analysis.
lists was conducted, thereby returning another 47 articles. Over- In order to demonstrate how the point estimate and condence
all, 237 complete studies were collected and reviewed against the intervals differ, both the xed and random effects computational
inclusion and exclusion criteria resulting in a total of 31 articles for models were computed and displayed in Forest plot gures. In
the present meta-analysis. A full list of the initial articles identied analyses where heterogeneity is signicant, a random-effects and
or the completed studies reviewed are available upon request from corresponding 95% CI represents the more conservative estimate
the corresponding author. of the RR ratio risk factor parameter.

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
e346 C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351

In order to statistically examine the degree of similarity in the were two non-signicant RRs of less than 1.0 reported by Jones
studies outcomes (i.e., test for heterogeneity across studies), For- [37] (0.59, 95% CI: 0.261.34) and Bache et al. [12] (0.99, 95% CI:
est plots with chi-squared (2 ) tests for heterogeneity of effect sizes 0.891.10). These results contrast with the high RRs ratios found
were performed [20,21]. Heterogeneity was considered to be sig- in the studies of Suzuki et al. [48] (14.87, 95% CI: 9.2023.74) and
nicant, if p values were less than 0.05. These plots demonstrate Paton et al. [43] (13.59, 95% CI: 8.8120.98).
the number of studies included in the analyses across each risk fac-
tor, the precision for each study, the consistency of the effect from 3.4. Hip side
study to study, and the relative weight assigned to each study based
on sample size. Finally, in order to address the issue of publication Empirical data on the prevalence of DDH by hip side was avail-
bias (i.e., statistically signicant studies are more likely to be pub- able in 10 (32%) studies that included 233,358 newborns. Using a
lished than statistically non-signicant ones), the potential impact random effects model, Fig. 3 shows that 3 of the studies had RRs
of this bias was assessed through a sensitivity analysis or funnel less than 1.0 and 7 had RRs greater than 1.0 indicating overall that
plot. a signicantly higher occurrence of DDH is associated with the left
side of the hip. The random RR for DDH was found to be just over
3. Results 1.5 times higher for the left side than the ride side of the hip (1.54,
95% CI: 1.251.90).
3.1. Study characteristics and quality assessment
3.5. Family history
An overview of the 31 primary studies included in the meta-
analysis is provided in Table 1. In addition to identifying the tests From 4 (12%) studies included in the meta-analysis, a family his-
used to diagnosis and screen for DDH, the table includes year of pub- tory of DDH was shown to result in a signicantly higher onset of
lication, study population, country, DDH risk factors investigated, DDH in newborns (Fig. 4). Although 2 (50%) of the studies had RRs
and level of evidence rating for each study. less than 1.0, the other two larger studies showed statistically sig-
The sample size of each study ranged from a minimum of 69 nicant differences, indicating an association of family history with
participants to a maximum of 151,924 participants; with a median an occurrence of DDH. The studies were shown to be homogeneous,
sample size of 6559. The majority of the studies were conducted wherein the use of either a random or xed effects model resulted
in European countries (20; 65%). Of the diagnostic tests used for in a RR of 1.39 (95% CI: 1.231.57).
DDH, most of the studies included data from universal clinical
exams (27; 87%), selective clinical exams (for those infants at risk to 3.6. First born
develop DDH) (24; 77%), universal ultrasound (15; 48%), selective
ultrasound (19; 61%), and radiography (16; 52%). In the 5 (16%) studies that had data available on the prevalence of
Of the 31 studies included in the present meta-analysis, six types DDH with rst born births, 4 had statistically signicant RRs values
of risk factors were examined: 24 studies (77%) investigated sex as less than1.0 (Fig. 5). Using a random effects model analysis resulted
a risk factor, followed by 15 studies (48%) that looked at breech in a statistically signicant RR value of 1.44 with 95% condence
presentation, 10 studies (32%) explored affected hip side, 5 studies intervals from 1.12 to 1.86 (p < .01).
(16%) rst born, 4 (13%) on mode of delivery and 4 (13%) on family
history of DDH (Table 1). Finally, in order to compare methodolog- 3.7. Mode of delivery
ical quality of the studies included in the meta-analysis, we rated
each study with a grade (A, B, or C) for level of evidence. The major- In the 4 (13%) studies that had data available on the prevalence
ity of studies were graded at Level B (23, 74%) followed by Level A of DDH based on mode of delivery (vaginal versus C-section), 2
(8; 28%); there were no studies rated a level of evidence grade of C. studies showed signicant RRs values for vaginal delivery, 1 study
found a signicant RR for C-section, while the fourth study was
3.2. Sex of newborn inconclusive (Fig. 6). The more conservative random effects model
resulted in a non-signicant RR value of 1.22 (95% CI: 0.463.23).
Of the 31 studies included in the meta-analysis, 24 (77%) pro-
vided data on the incidence of DDH in newborns by their sex as 3.8. Heterogeneity and publication bias
either a raw number or percentage of total births from their respec-
tive hospital or region. Using the more conservative random effect Based on the diagnostic test used to conrm DDH and the time
size estimate for interpretation of the combined total, the RR in the tests were conducted on the newborns, subgroup analyses were
newborn females was found to be 2.54 (95% CI: 2.113.05) times limited by inconsistencies found between the studies. For instance,
more prevalent than in males. As illustrated by the Forest plot in a subgroup analysis on the origin of the study (i.e., Asia, Europe, and
Fig. 1, 19 of the 24 studies showed that females were signicantly Australia/USA) failed to explain the heterogeneity found across risk
more likely than males to be diagnosed with DDH. The remaining 5 factors. While the presence of heterogeneity would reect effects
studies were found to be inconclusive as the RR 95% CI crossed 1.0, that are potentially explainable by different confounding variables,
indicating non-signicant results for sex of newborn (Fig. 1). the available data did not allow for further evaluation.
Although the purpose of this meta-analysis would preclude the
3.3. Breech presentation likelihood that we have included only those studies that reported
signicant incidences of DDH specic to any one particular risk fac-
The next most commonly reported data on a risk factor for DDH tor, we plotted the effect size of each trial from all of the studies
was breech presentation (15 or 48% of the studies). As shown in to assess for publication bias. In studies of smaller sample sizes, for
Fig. 2, the majority of studies (13; 87%) revealed a statistically sig- which biases are more often to occur, plotting the relative measures
nicant difference between groups, with a high incidence of DDH of risk ratios against the standard error on a logarithmic scale is
when newborns presented in a breech position during pregnancy standard practice [51]. To emphasize the differences in larger stud-
and at delivery. The random RR was found to be 3.75 (95% CI: ies, however, plotting against precision (1/standard error) yields
2.256.24) times higher than for non-breech births. Of the studies a statistical test that is consistent for comparisons to be made of
that reported the prevalence of DDH in breech presentation, there funnel plot asymmetry.

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351 e347

Table 1
Overview of the 31 studies included in the meta-analysis.

Author Year N Diagnosis test(s)a Risk factors investigated Evidence grade


(moderators) levels (A, B, C)

Akman et al. [22] 2007 403 Turkey US(U) Sex, side B


Anderson et al. [23] 2001 6571 Sweden US(U) Breech B
Ang et al. [24] 1997 19,954 Singapore CE(U,HR), US(HR), R Sex A
Artz et al. [25] 1975 23,720 USA CE(U) Breech, rst born, mode of A
delivery, sex
Bache et al. [12] 2002 29,324 UK CE(HR),US(U,HR) Breech, sex, side B
Bialik et al. [26] 1999 18,060 Israel CE(U,HR),US(U,HR) Side B
Boere-Boonekamp et al. [27] 1998 1968 Netherlands CE(U,HR), US(HR), R Sex B
Castelein et al. [28] 2001 683 Netherlands CE(U,HR), US (U,HR), R Side B
Chan et al. [13] 1997 147,166 Australia CE(U,HR), US(HR) Breech, sex B
Chasiotis-Tourikis et al. [29] 2003 1033 Greece CE(U,HR), US (U,HR) Sex B
Dogruel et al. [30] 2007 3500 Turkey CE(U,HR), US (U,HR) Breech, family history, rst born. B
mode of delivery, sex
Eidelman et al. [31] 2002 768 Israel CE(U,HR),US(U), R Sex, side B
Falliner et al. [32] 1999 6548 Germany CE(U, HR), US (U,HR), R Breech, family history, sex B
Giannakopoulou et al. [33] 2002 6140 Greece CE(U,HR), US (HR), R Sex B
Goss et al. [34] 2002 5166 Australia CE(U,HR), US(HR), R Breech B
Hadlow et al. [35] 1999 10,534 New Zealand CE(U,HR), R Breech, side A
Heikkila et al. [36] 1984 151,924 Finland CE(U,HR) Sex, side B
Jones et al. [37] 1989 426 Swansea CE(U,HR), US(HR), R Breech, family history, mode of A
delivery
Lauge-Pedersen et al. [38] 2006 32,171 Sweden CE(U,HR), US (U,HR), R Sex, side B
Lewis et al. [39] 1999 17,792 Wales CE(U,HR), US (U,HR), R Sex B
Lorente et al. [40] 2002 153 Spain CE(U, HR), US (HR), R Sex, side A
Lotito et al. [41] 2002 69 Italy CE(U,HR), US (HR) Sex A
Miranda et al. [42] 1988 49,937 Spain CE(U,HR), R Breech, rst born, mode of B
delivery, sex
Paton et al. [43] 2005 34,723 UK CE(U,HR), U(U,HR) Breech A
Riboni et al. [44] 2003 8896 Italy CE(U), US(U) Sex B
Rosendahl et al. [45] 1994 3613 Norway CE(U), US (U) Breech, family history, sex, side B
Sahin et al. [46] 2004 5798 Turkey CE(U,HR), US(HR), R Sex B
Stevenson et al. [47] 2009 18,138 USA CE(U) Breech, sex, rst born B
Suzuki et al. [48] 1986 6559 Japan CE(U,HR) Breech, sex B
Terjersen et al. [49] 1989 1000 Norway CE(U,HR),US(U,HR), R Sex B
Yiv et al. [50] 1997 19,622 Australia CE(n/s), US (n/s), R Breech, rst born, sex A
a
CE (U), clinical examination universal; CE (HR), selective clinical examination; US (U), ultrasound universal; US (HR), selective ultrasound; R, radiography; n/s, not
specied.

As shown in the funnel plot for the studies that reported on the with standard diagnostic procedures. This approach would also
incidence of DDH as a function of the sex of the newborn (Fig. 7), produce the highest prevalence estimate.
the trial effects of similar magnitude but opposite directions are Adopting the most basic approach, we found that the total
equidistant from 1.0. The generally inverted funnel and symmet- number of cases with DDH of infants less than 6 months of age
rical plot of the trial effects, with smaller studies scattering more was 20,196 and the total number of individuals in the popula-
widely at the bottom of the graph reecting less precision in the tion of infants less than 6 months of age was 1,065,867. Thus the
estimation of the risk factor effect, supports a non-bias publication prevalence rate of DDH for the studies included in the present
of the incidence of DDH as it pertains to any one particular risk meta-analysis was determined to be 1.9% (20,196/1,065,867).
factor.

4. Discussion
3.9. Prevalence rate of DDH
The present meta-analysis was intended to investigate the risk
A second aim of the present meta-analysis was to determine factors associated with DDH in patients less than 6 months of age,
the prevalence rate of the occurrence of DDH. The prevalence of a to describe current diagnostic and screening practices of DDH, and
disease in a population is dened as the total number of cases of to determine the prevalence rate of the occurrence of DDH. Among
the disease in the population at a given time divided by the number the six risk factors investigated, the present ndings suggest that
of individuals in the population [52]. the most signicant risk factors associated with DDH include: those
As discussed by Scahill and Bearss [53], there are various basic presenting in the breech position during delivery, being female,
approaches used to estimate prevalence, each of which differs having an affected left hip, or having a family history of DDH. Unfor-
in quality of results. The simplest is to count known DDH cases tunately, data was not reported to quantify the effects of combined
and divide by the population. This approach likely underestimates risk factors across studies and the unknown independence or cor-
prevalence because it does not count those who have not yet been relation between variables does not allow us to assume factors are
diagnosed and it may generate skewed statistics because there simply additive. Still, it may be inferred that the combination of
are differences in access to treatment. A second approach, which risk factors (such as a newborn female with a breech presenta-
improves on the rst, would have investigators examine records, tion at birth), would predispose a neonate to be at an even higher
looking for probable cases that have not yet been identied. Finally, risk for DDH. In addition, while the present study did not nd sta-
the third and most reliable approach would be to screen prospec- tistically signicant results for the risk factor of mode of delivery
tively a large sample of an entire community to identify possible (vaginal versus via caesarean-section) we cannot conclusively dis-
cases of DDH and subsequently evaluate each case in greater detail miss this as an unimportant risk factor. One of the reasons we may

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
e348 C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351

Fig. 1. Meta-analysis of random combined RRs incidence of DDH in newborns that are Female.

not have found stronger evidence could be because of the rather of DDH based on screening across several studies for not-at-risk
small number of studies included in the analysis. boys 4.1/1000) and girls (19/1000) resulted in a RR calculation of
Our results are comparable to those published by the American 4.6 (much higher than the 2.54 RR ratio derived from the com-
Academy of Pediatrics (AAP) with the same three main risk fac- bined effect sizes in this meta-analysis). The RR estimate of DDH for
tors identied in a clinical practice guideline for early detection breech presentation was determined to be the highest risk factor
of DDH published in 2000 [54]. The AAP reports that an estimate by the AAP at 6.3 which corresponds to our ndings regarding this

Fig. 2. Meta-analysis of random combined RRs incidence of DDH in newborns that are Breech births.

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351 e349

Fig. 3. Meta-analysis of random combined RRs incidence of DDH in newborns on the Left Side of Hip.

Fig. 4. Meta-analysis of random combined RRs incidence of DDH in newborns with Family History of DDH.

Fig. 5. Meta-analysis of random combined RRs of the incidence of DDH in newborns that are First Born.

Fig. 6. Meta-analysis of random combined RRs incidence of DDH in newborns with Vaginal Delivery.

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
e350 C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351

Fig. 7. Funnel plot of random combined RRs incidence of DDH based on Sex of Newborn.

factor being the greatest risk factor at 3.75, but at a much lower racial distribution of the population, the criteria for a positive ultra-
ratio. Similarly, having a positive family history for DDH was sound, and whether a positive ultrasound with a normal physical
reported as a risk factor with a RR ratio of 1.7 compared to 1.39 in examination is acceptable criteria for DDH.
our study. A corresponding technical report published by Lehmann
et al. [55] in conjunction with the AAP Committee on Quality 4.2. Conclusions
Improvement, Subcommittee on DDH, reported similar odds ratios
for DDH with breech delivery at 5.5, for female sex at 4.1, and fam- As has been argued elsewhere, a well developed screening
ily history at 1.7. Although we were able to derive combined effect program may prevent late presentation of DDH, development of
size RR estimates for other DDH risk factors related to left hip side premature degenerative arthritis in adults, decrease associated
(1.54, p < 0.01), rst born (1.44, p < 0.01) and mode of delivery (1.22, morbidity related to DDH, and reduce costs [1,56]. Alternatively,
p = non-signicant), the AAP did not ag these as potential DDH risk there is a concern that screening leads to over diagnosis of DDH
factors [54,55]. (i.e., false-positives) and most abnormal ndings will completely
In response to the question of diagnostic and screening practices resolve if left untreated. Nevertheless, in a recent decision anal-
of DDH, we can simply comment descriptively based on the prac- ysis model on the utility of screening for DDH it was concluded
tices reported by the researchers of the 31 studies in the present that the optimum strategy associated with the highest probabil-
meta-analysis. We cannot generalize beyond these studies, nor ity of having a non-arthritic hip at the age of sixty years, was to
would it be fair for us to conclude that these studies are in any way screen all neonates for DDH with a physical examination and to
representative of the past or present practices of health care prac- use ultrasonography selectively of infants who are at high risk
titioners. We can say, however, that of the diagnostic tests used [57]. The key to early detection and careful examination of the
for the screening of DDH, clinical examinations remain the most infant in the rst year of life remains important. It is recommended
commonly used (51%), followed by ultrasound (33%), and radiog- that medical school curricula and pediatric and family practice res-
raphy (16%). These results indicate that there is still discrepancy in idency programs emphasize the importance of understanding the
identifying a universally agreed upon screening practice. key risk factors associated with DDH as identied in the present
meta-analysis. Although the determined DDH prevalence rate of
4.1. Limitations 1.9% likely underestimates the true prevalence of DDH, it reinforces
the importance and urgency for a timely diagnosis.
There are several limitations to the present meta-analysis. First, Although various studies have been conducted on DDH, differ-
only studies published in the English language were included. This ences in the operational denition of the construct, and a lack
might reect clinical practice biases specic to English-speaking of agreement on when and how to diagnose the condition, has
countries and potentially underestimate the prevalence of DDH. resulted in unclear and equivocal research results related to the
Second, only studies with infants less than six months of age were associated risk factors of DDH. The present investigation has pro-
included for analysis. This could exclude patients whose diagno- vided the opportunity to assess the phenomenon of DDH from
sis of DDH was made after 6 months of age. Third, an analysis on an empirical perspective. Notwithstanding the limitations inher-
the combined effects of multiple risk factors was not possible. For ent in conducting a meta-analysis and the lack of specic data to
example, although 16 studies reported the number or percentage conduct important moderator analyses, the present meta-analysis
of DDH newborns in relationship to sex and one or more other provides an important summary of the empirical and theoretical
risk factors, only one of these studies [45] stratied the sample by research regarding key risk factors associated with DDH. We sug-
sex (male/female) in connection with another risk factor (i.e., side gest that studies published in the future include the following: (1)
of hip). Fourth, readers should recognize that our calculated preva- a clear denition of the condition of DDH they are adopting; (2)
lence rate is based on 31 studies from various geographic locations. a clear statement of the risk factors associated with the patient
This is important since prevalence in DDH varies depending on diagnosis of DDH; and (3) specication of the method of diagnosis

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.
C.L. Ortiz-Neira et al. / European Journal of Radiology 81 (2012) e344e351 e351

used, as well as if follow-up was performed and by which diag- [28] Castelein RM, Korte J. Limited hip abduction in the infant. J Pediatr Orthop
nostic method. In conclusion, based on the current ndings future 2001;21:66870.
[29] Chasiotis-Tourikis E, Varvarigou A, Yarmenitis S, Vandaros N, Beratis NG. Mater-
research should focus on the development of clinical guidelines nal smoking during pregnancy improves the anatomy of the hip joint in the
related to the screening and diagnosis of DDH in association with female neonate. J Matern Fetal Neonatal Med 2003;14:4550.
the common risk factors identied. [30] Dogruel H, Atalar H, Yavuz OY, Sayli U. Clinical examination versus ultra-
sonography in detecting developmental dysplasia of the hip. Int Orthop
2007;32(3):4159.
Conict of interest [31] Eidelman M, Chezar A, Bialik V. Developmental dysplasia of the hip. Incidence in
Ethiopian Jews revisited: 7 year prospective study. J Pediatr Orthop 2002;11(Pt
B 12):2902.
None declared. [32] Falliner A, Hahne HJ, Hassenpug J. Sonographic hip screening and early man-
agement of developmental dysplasia of the hip. J Pediatr Orthop 1999;8(Pt
References B):1127.
[33] Giannakopoulou C, Aligizakis A, Korakaki E, et al. Neonatal screening for
developmental dysplasia of the hip on the maternity wards in Crete, Greece,
[1] Mahan ST, Katz JN, Kim YJ. To screen or not to screen? A decision analysis of
correlation to risk factors. Clin Exp Obstet Gynecol 2002;29:14852.
the utility of screening for developmental dysplasia of the hip. J Bone Joint Surg
[34] Goss PW. Successful screening for neonatal hip instability in Australia. J Paediatr
Am 2009;91(7):170519.
Child Health 2002;(32):46974.
[2] Babcock DS, Hernandez RJ, Kushner DC, et al. Developmental dysplasia of the
[35] Hadlow V. Neonatal detection of developmental dysplasia of the hip (DDH). J
hip: American College of Radiology ACR appropriateness criteria. Radiology
Bone Joint Surg Br 1999;81:7446.
2000;215(Suppl.):81927.
[36] Heikkila E. Congenital dislocation of the hip in Finland. Acta Orthop Scand
[3] Aronsson DD, Goldberg MJ, Kling Jr TF, Roy DR. Developmental dysplasia of the
1984;55:1259.
hip. Pediatrics 1994;94:2018.
[37] Jones DA. Importance of the clicking hip in screening for congenital dislocation
[4] Shipman SA, Helfand M, Moyer VA, Yawn B. Screening for developmental dys-
of the hip. Lancet 1989;1(8638):599601.
plasia of the hip: a systematic literature review for the US preventive services
[38] Lauge-Pedersen H, Gustafsson J, Hagglund G. 6 weeks with the von Rosen
task force. Pediatrics 2006;117:55776.
splint is sufcient for treatment of neonatal hip instability. Acta Orthop
[5] Kernohan WG, Nugent GE, Haugh PE, Trainor BP, Mollan RA. Sensitivity of man-
2006;77:25761.
ual palpation in testing the neonatal hip. Clin Orthop Relat Res 1993;5:2115.
[39] Lewis K, Jones DA, Powell N. Ultrasound and neonatal hip screening: the
[6] American Academy of Pediatrics Committee on Quality Improvement:
ve-year results of a prospective study in high-risk babies. J Pediatr Orthop
Subcommittee on Developmental Dysplasia of the Hip. Clinical practice
1999;19:7602.
guideline: early detection of developmental dysplasia of the hip. Pediatrics
[40] Lorente Molto FJ, Gregori AM, Casas LM, Perales VM. Three-year prospective
2000;105(4):896905.
study of developmental dysplasia of the hip at birth: should all dislocated or
[7] Hvid I. Neonatal hip instability, developmental dysplasia of the acetabulum,
dislocatable hips be treated? J Pediatr Orthop 2002;22:61321.
and the risk of early osteoarthrosis. Acta Orthop 2008;79(3):3112.
[41] Lotito FM, Rabbaglietti G, Notarantonio M. The ultrasonographic image of the
[8] Klisic PJ. Congenital dislocation of the hip a misleading term: brief report. J
infant hip affected by developmental dysplasia with a positive Ortolanis sign.
Bone Joint Surg Br 1989;71:136.
Pediatr Radiol 2002;32:41822.
[9] Kokavec M, Makai F, Maresch P. Present status of screening and prevention
[42] Miranda L, Palomo JM, Monzonis J, Marti V. Prevention of congenital dislocation
of developmental dysplasia of the hip in the Slovak Republic. J Pediatr Orthop
of the hip in the newborn. J Pediatr Orthop 1988;8:66175.
2003;12(Pt B):1068.
[43] Paton RW, Hinduja K, Thomas CD. The signicance of at-risk factors in
[10] Toma P, Valle M, Rossi U, Brunenghi GM. Paediatric hip-ultrasound screen-
ultrasound surveillance of developmental dysplasia of the hip. A ten-year
ing for developmental dysplasia of the hip: a review. Eur J Ultrasound
prospective study. J Bone Joint Surg Br 2005;87(9):12646.
2001;14:4555.
[44] Riboni G, Bellini A, Serantoni S, Rognoni E, Bisanti L. Ultrasound screening for
[11] Cady RB. Developmental dysplasia of the hip: denition, recognition, and pre-
developmental dysplasia of the hip. Pediatr Radiol 2003;33:47581.
vention of late sequelae. Pediatr Ann 2006;35(2):92101.
[45] Rosendahl K, Markestad T, Lie RT. Ultrasound screening for developmental dys-
[12] Bache C, Clegg J, Herron M. Risk factors for developmental dysplasia of the hip:
plasia of the hip in the neonate: the effect on treatment rate and prevalence of
ultrasonographic ndings in the neonatal period. J Pediatr Orthop 2002;11(3
late cases. Pediatrics 1994;94:4752.
(Pt B)):2128.
[46] Sahin F, Akturk A, Beyazova U, et al. Screening for developmental dyspla-
[13] Chan A, McCaul KA, Cundy PJ, Haan EA, Byron-Scott R. Perinatal risk factors for
sia of the hip; results of a 7-year follow-up study. Pediatr Int 2004;46:
developmental dysplasia of the hip. Arch Dis Child 1997;76:F94100.
1626.
[14] Albinana J, Quesada JA, Cetucha JA. Children at high risk for congenital disloca-
[47] Stevenson D, Mineau G, Kerber R, Viskochil DH, Schaefer C, Roach J. Famil-
tion of the hip: late presentation. J Pediatr Orthop 1993;13:2689.
ial predisposition to developmental dysplasia of the hip. J Pediatr Orthop
[15] Jones DA, Powell N. Ultrasound and neonatal hip screening. J Bone Joint Surg
2009;29(5):4636.
Br 1990;72:4579.
[48] Suzuki S, Yamamuro T. Correlation of fetal posture and congenital dislocation
[16] Tonnis D, Strorch K, Ulbrich H. Results of newborn screening for CDH with
of the hip. Acta Orthop Scand 1986;57:814.
and without sonography and correlation of risk factors. J Pediatr Orthop
[49] Terjesen T, Bredland T, Berg V. Ultrasound for hip assessment in the newborn.
1990;10:14552.
J Bone Joint Surg Br 1989;71(5 Pt B):76773.
[17] Walter RS, Donaldson JS, Davis CL, et al. Ultrasound screening of high risk
[50] Yiv BC, Saidin R, Cundy PJ, et al. Developmental dysplasia of the hip in
infants. Am J Dis Child 1992;146:2304.
South Australia in 1991: prevalence and risk factors. J Paediatr Child Health
[18] Patel H. Preventive health care, 2001 update: screening and management of
1997;33:1516.
developmental dysplasia of the hip in newborns. CMAJ 2001;164:166977.
[51] Egger M, Smith GD, Altam DG. Investigating and dealing with publication and
[19] The Journal of Bone and Joint Surgery: Instructions to Authors. JBJS 2009;91(Pt
other biases. In: Sterne JAC, Egger M, Smith GD, editors. Systematic reviews
A 12):15.
in health care: meta-analysis in context. 2nd ed. Padstow (Cornwall): BMJ
[20] Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in
Publishing Group; 2001. p. 189208.
epidemiology. A proposal for reporting. JAMA 2000;238:200812.
[52] Coggon D, Rose G, Barker DJP. Quantifying disease in populations. In: Epi-
[21] Cochran WG. The combination of estimates from different experiments. Bio-
demiology for the uninitiated. 5th ed. London: BMJ Publishing Group; 2003.
metrics 1954;10(1):10129.
p. 613.
[22] Akman A, Korkmaz A, Aksoy C, Yazici M, Yurdakok M. Evaluation of risk factors
[53] Scahill L, Bearss K. The rise in autism and the mercury myth. J Child Adolesc
in developmental dysplasia of the hip: results of infantile hip ultrasonography.
Psychiatr Nurs 2009;22(1):513.
Turk J Pediatr 2007;49(3):2904.
[54] American Academy of Pediatrics: Committee on Quality Improvement,
[23] Andersson JE, Oden A. The breech presentation and the vertex presentation
Subcommittee on Developmental Dysplasia of the Hip. Clinical practice
following an external version represent risk factors for neonatal hip instability.
guideline: early detection of developmental dysplasia of the hip. Pediatrics
Acta Paediatr 2001;90:8958.
2000;105:896905.
[24] Ang KC, Lee EH, Lee PY, Tan KL. An epidemiological study of developmen-
[55] Lehmann HP, Hinton R, Morello P, Santoli J, In conjunction with the AAP Com-
tal dysplasia of the hip in infants in Singapore. Ann Acad Med Singapore
mittee on Quality Improvement, Subcommittee on Developmental Dysplasia
1997;26:4568.
of the Hip. Developmental dysplasia of the hip practice guideline: technical
[25] Artz TD, Levine DB, Lim WN, Salvati EA, Wilson PD. Neonatal diagnosis,
report. Pediatrics 2006;105:e57.
treatment and related factors of congenital dislocation of the hip. CORR
[56] Engester I, Lie S, Lehmann T, Furnes O, Vollset S, Engester L. Neonatal hip
1975;110:11236.
instability and risk of total hip replacement in young adulthood. Acta Orthop
[26] Bialik V, Bialik GM, Blazer S, Sujov P, Wiener F, Berant M. Developmental dys-
2008;79(3):31320.
plasia of the hip: a new approach to incidence. Pediatrics 1999;103:939.
[57] Mahan ST, Katz JN, Kim Y-J. To screen or not to screen? A decision analysis of
[27] Boere-Boonekamp MM, Kerkhoff TH, Schuil PB, Zielhuis GA. Early detection of
the utility of screening for developmental dysplasia of the hip. J Bone Joint Surg
developmental dysplasia of the hip in The Netherlands: the validity of a stan-
Am 2009;91:170519.
dardized assessment protocol in infants. Am J Public Health 1998;88:2858.

Descargado para Anonymous User (n/a) en ClinicalKey Espanol Colombia, Ecuador & Peru Flood Relief de ClinicalKey.es por Elsevier en agosto 04, 2017.
Para uso personal exclusivamente. No se permiten otros usos sin autorizacin. Copyright 2017. Elsevier Inc. Todos los derechos reservados.