Keywords hip dislocation, congenital - ultrasonography - diagnosis - newborn
Introduction
There is a lack of studies and evidence available to guide the clinical practice in
the treatment of developmental dysplasia of the hip (DDH). Among the limitations observed,
there is no consensus regarding the classification and diagnosis of DDH, and a wide
variability in decision-making, almost always guided by studies with small samples.[1 ] Most of the existing studies on DDH are retrospective, have small samples (considering
the number of hips and not individuals), and do not correct the results when cases
of bilateral DDH are included.[1 ] Moreover, many studies do not include the entire spectrum of DDH.[1 ]
In the study published by Guarniero et al.[2 ] (1988) based on a meta-analysis and multiple logistic regression protocols, the
estimated prevalence of this condition in Brazil was close to 1.1%, and this data
was obtained through the positivity of the Ortolani maneuver. To improve the evidence
available, it is necessary to use a standardized framework for reports and diagnoses,
with consistency, and to conduct prospective studies with a sound methodological design.[1 ]
Acetabular dysplasia cannot be excluded by a normal physical examination, and ultrasound
can be an important tool for the early diagnosis and treatment of the disease. Articular
instability and hip dislocation can be diagnosed by physical examination using the
Barlow and Ortolani maneuvers respectively.[3 ]
The present study aims to evaluate the prevalence of DDH in newborns aged between
0 and 3 days of life in a public maternity hospital and to correlate the findings
with the main risk factors described in the literature.
Materials and Methods
The present study was approved by the Ethics Committee of both institutions involved,
under numbers 1554 and 2016. The present manuscript was written following the guidelines
of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)
for the communication of observational studies.[4 ]
The present observational, transversal, prospective study was conducted in a high-risk
public maternity hospital in the city of São Paulo, Brazil, in which an average of
7,900 deliveries are performed per year. We performed the exams in the nursery because
this would be the only time when we would have a minimally-acceptable sample to assess
the prevalence among the population of live births, even though spontaneous improvement
regarding the classification has been described in the literature, specifically in
the hips that do not show instability on the dynamic examination. The collection was
performed randomly when the main researcher and her team attended the hospital, from
March to September 2018. The parents or legal guardians were informed about the study
conditions through the Free and Informed Consent Form presented before the ultrasonography
examination.
The population of the present epidemiological study consisted of roomed-innewborns
aged between 0 and 3 days of life for the early detection of the disease. Each participant
was eligible only once, and random sampling was performed. To calculate the sample
size required, we estimated the number of 500 cases to obtain an incidence of up to
5.5% with an estimated error rate of up to 2%. A 95% confidence interval (95%CI) was
established. Since, after the collection, the covariate and risk factor assessments
would be performed using multivariate analysis, with possible data loss, we estimated
an increase of 30% in the size of the sample. Thus, the required sample size for the
present study was estimated as at least 650 cases.
The inclusion criterion was roomed-in newborns in the maternity hospital on the dates
when the research team attended it.
The exclusion criteria were:
1) Extreme preterm birth: due to the need for oxygen and heating in the incubator,
and to avoid excessive heat loss and manipulation of the at-risk newborn, as it is
necessary to use cold gel and change the newborn's position to perform the ultrasonography,
which increases the risk of intracranial hemorrhage;
2) Breathing difficulty: due to the need for oxygen and monitoring, as it is necessary
to manipulate the newborn and change his or her position to perform the ultrasonography;
3) Pathologies and/or conditions that required intensive care: to avoid manipulating
the newborn, as it is necessary to change his or her position to perform the ultrasonography;
4) Congenital anomalies, genetic diseases: because patients with genetic diseases
and/or congenital disorders have a known higher prevalence of DDH, which could be
assessed as a selection bias and increase prevalence, not translating the average
dysplasia in the general population;
5) High-risk pregnancy: due to the need to monitor the newborn;
6) Parents or legal guardians who did not sign the consent form because they did not
agree to participate in the study; and
7) Hips classified as type-IIa according to the Graf Method because of the immaturity
that could normalize in a few weeks.
Upon acceptance to participate in the study, a hip ultrasonography examination was
performed on the patients who had already undergone a clinical examination by a neonatologist,
with the Ortolani and Barlow maneuvers.[5 ] Finally, a standardized questionnaire on the subject of the study, made by the authors,
was applied to record the risk factors for DDH.
Ultrasonography Examination
The method used was developed by Graf[6 ] in 1980; it is considered the reference method, and it is accepted by healthcare
systems in several countries. All examinations in the present study were performed
strictly following the Graf method.
A pediatric sonographer with more than 20 years of experience in pediatric hip ultrasonography,
who was blinded to the clinical examination results, performed the ultrasound examinations.
Positioning
The newborn is positioned in lateral decubitus, opposite to the hip to be examined,
with the leg slightly flexed and adducted, and the foot in slight internal rotation
([Fig. 1 ]), using a linear transducer, in the coronal plane of the hip, thus establishing
a standard plane in which it is possible to evaluate the acetabular morphology and
the degree of femoral head coverage.
Fig. 1 Correct position of the newborn during the exam: lateral decubitus, with the lower
limbs flexed and adducted. Credits: art by Vinicius Mustafa.
In the classification of the Graf method, type I is normal, with an α angle > 60°.
Type IIa hips represent immaturity, with an α angle between 50° and 59°, and a β angle > 55°.
Hips with an α angle ≤ 49° are defined as presenting pathological development, and
are classified as types IIc, IId, IIIa, or IIIb. Type-IV hips on the Graf method are
not measurable. Developmental dysplasia of the hip was defined by Graf when the ultrasonography
classification reveals a pattern ≥ IIb; type-IIa hips are classified as immature,
but follow-up and treatment are suggested when they persist after 30 days of life.[7 ]
Statistical Analysis
The data were inserted in an Excel (Microsoft Corp., Redmond, WA, United States) spreadsheet.
For the statistical analysis, the data were exported to the Statistical Package for
the Social Sciences (IBM SPSS Statistics for Macintosh, IBM Corp., Armonk, NY, United
States) software, version 24.0. The descriptive statistics of the categorical data
are presented according to their absolute and relative frequencies. The continuous
data are presented according to their means and respective standard deviations (SDs).
The prevalence was calculated by the ratio between the number of newborns with at
least one hip with a classification ≥ IIc according to the Graf method and the number
of newborns assessed.
For the analysis of the risk factors, binomial logistic regression was performed to
verify the effects of gender (male and female), ethnicity (white and others), parity
(one delivery or multiparity), and the intrauterine presentation of the fetus at the
time of delivery (breech or others) to identify the increased risk of developing DDH.
For the regression analysis, we used independent variables with biological plausibility
that presented an association on the univariate analysis with acceptance in the p ≤ 0.1 model. To be accepted as statistically significant, in the risk factor prediction
equation, the significance value should be p < 0.05.
Results
There were 3,970 deliveries between April and September 2018, of which 733 newborns
were available for examination. However, during that period, 28 newborns were premature,
16 presented transient respiratory distress, and 3 had aspirated meconium, and they
were excluded from the study due to the need for intensive care. Eight newborns whose
parents or legal guardians did not accept to participate in the study were also excluded.
Approximately 8% of the newborns were excluded from the sample according to the stipulated
criteria ([Fig. 2 ]) There were no complications during the examination performed for the research.
Fig. 2 Flowchart of the sample selection.
A total of 678 newborns (1,356 hips) were included in the present study. The prevalence
of DDH was of 5.46% (37 newborns). In total, 24 newborns had unilateral dysplasia:
3, on the right hip (8.1%), and 21, on the left hip (56.7%); and 13 (31,5%) newborns
had bilateral dysplasia. The sample characteristics are described in [Table 1 ]. Among the newborns with DDH examined, the prevalence was assessed in relation to
each type of hip using the Graf method ([Table 2 ]).
Table 1
Characteristic
Children evaluated (N = 678) – n (%)
Newborns with developmental dysplasia of the hip (N = 37; 5,4%) – n (%)
Gender
- Female
356 (52.5%)
32 (86.5%)
- Male
322 (47.5%)
5 (13.5%)
Ethnicity
- White
364 (53.7%)
30 (81%)
- Black
177 (26%)
0
- Brown
137 (20.3%)
7 (19%)
Type of delivery
- Vaginal
449 (66%)
22 (59%)
- Cesarean
229 (34%)
15 (41%)
Presentation
- Pelvic
179 (26.4%)
19 (51%)
- Cephalic
474 (69.9%)
17 (46%)
- Transversal
25 (3.7%)
1 (3%)
Parity
- One delivery
302 (44.5%)
27 (73%)
- Multiparous
376 (55.4%)
10 (27%)
Twinning
4 (0.6%)
0
Family history
19 (2.8%)
1 (2.7%)
Orthopedic pathologies
0
0
Genetic syndromes
0
0
Positive Ortolani maneuver
4 (0.6%)
2 (5.4%)
Table 2
Hip Classification
Right hip – n (%)
Left hip – n (%)
Ia
594 (43.6%)
554 (40.6%)
Ib
8 (0.6%)
23 (2.3%)
IIa
60 (4.4%)
67 (4.9%)
IIc
1 (0.07%)
3 (0.22%)
IId
9 (0.66%)
23 (1.65%)
IIIa
4 (0.3%)
4 (0.3%)
IV
2 (0.14%)
4 (0.3%)
Among the risk factors assessed, the following presented statistical significance
on the univariate analysis:
Female sex (relative risk [RR] = 5.78; 95%CI: 2.28 to 14.67; p = 0.0002);
Breech presentation (RR = 2.94; 95%CI: 1.58 to 5.48; p = 0.0007);
One delivery (RR = 3.36; 95%CI: 1.65 to 6.83; p = 0.0008); and
White ethnicity (RR = 3.69; 95%CI: 1.65 to 8.3; p = 0.001).
The risk factors that were not statistically significant were:
Cesarean section (RR = 1.33; 95%CI: 0.70 to 2.52; p = 0.37);
Family history of DDH (RR = 0.96; 95%CI: 0.14 to 6.66; p = 0.97); and
Twins (RR = 1.80; 95%CI: 0.12 to 25.4; p = 0.66).
Regarding family history, there may be information bias due to the difficulty in collecting
data regarding the families, because most of the parents or legal guardians felt insecure
about providing information regarding family history. Twins, also referred to as a
risk factor, did not present statistical significance. It is necessary to consider
a possible selection bias due to prematurity and indication for hospitalization, which
was considered an exclusion factor in the present study.
The sensitivity of the Ortolani maneuver for hips with dysplasia (Graf type-IIc or
higher) was only of 5.41% (95%CI: 0.66 to 18.2%), with a specificity of 99.7% (95%CI:
98.8 to 99.9%). For the diagnosis of dislocated hip (Graf type-IV), the sensitivity
of the Ortolani maneuver was of 50% (95%CI: 6.76% to 93.24%), with a specificity of
99.7% (95%CI: 98.8% to 99.9%). The positive predictive value was of 50% (95%CI: 15.55
to 84.5%), and the negative predictive value was of 99.7% (95%CI: 99.2% to 99.9%).
The binomial logistic regression verified the effects of gender, ethnicity, parity,
and fetal presentation at the time of delivery on the increased risk of developing
DDH. The logistic regression model was statistically significant: χ2 (3) = 44.553; p < 0.001. The model proved to be able to explain 18.4% (Nagelkerke R2 ) of variation in the risk of developing DDH, and correctly classified 94.5% of the
cases. All of the four predictive variables were statistically significant: gender,
ethnicity, parity, and breech presentation. Being white, female, firstborn, and having
breech presentation increased the risk of developing DDH, as shown in [Table 3 ].
Table 3
95% confidence interval for EXP(B)
B
SE
Wald
df
Sig.
Exp(B)
Lower
Upper
Step 1
White ethnicity
0.940
0.443
4.499
1
0.034
2.561
1.074
6.107
Parity
1.240
0.387
10.247
1
0.001
3.455
1.617
7.381
Gender
1.599
0.498
10.293
1
0.001
4.946
1.863
13.134
Breech presentation
0.709
0.359
3.895
1
0.048
2.032
1.005
4.109
Constant
−5.617
0.620
82.012
1
0.000
0.004
The logistic regression analysis showed an increased odds ratio (OR) for the white
ethnicity (OR = 2.561; 95%CI: 1.07 to 6.11); multiparity (OR = 3.50; 95%CI: 1.62 to
7.38); the female gender (OR = 4.95; 95%CI: 1.86 to 13.13); and breech presentation
(OR = 2.03; 95%CI: 1.01 to 4.11).
Discussion
The literature review did not reveal another prospective study estimating the prevalence
of DDH in the city of São Paulo.
Developmental dysplasia of the hip can manifest in three ways: by joint instability,
which can be diagnosed by the Barlow maneuver; by a hip dislocation, which can be
diagnosed by the Ortolani maneuver; and by acetabular dysplasia, which cannot be excluded
after a normal physical examination and requires the help of ultrasound for the diagnosis.
Acetabular dysplasia is one of the most common forms of the disease, represented by
the Graf method as types IIa, IIb, IIc, and D; and this information can be confirmed
in the present study. As the late diagnosis of DDH can cause serious problems in adulthood,
the use of tools such as the ultrasound for identification is justified, especially
in a more recurring fashion.[3 ]
Guarniero et al.[2 ] evaluated the prevalence of DDH in their prospective/retrospective study conducted
in the city of São Paulo through physical examination using the positivity of the
Ortolani maneuver for its statistical calculation, finding a result of 1.1%. A comparison
cannot be made between the present study and the aforementioned one, because the methodologies
are different, and we understand that newborns with acetabular dysplasia who do not
have a dislocated hip may have the disease, and the Ortolani maneuver may be negative.
Barbosa and Albernaz[8 ] were the only authors who tried to estimate the prevalence of DDH in Brazil, in
a retrospective study performed in the School of Medicine at Universidade Católica
de Pelotas. The authors reported many potential selection biases because the analysis
was performed with a small number of cases identified through the patients' medical
records.
In the meta-analysis by Ortiz-Neira et al.,[9 ] the prevalence of DDH was of 1,9% (20,196 cases of DDH among 1,065,867 patients).
In 2019, Zhao et al.[10 ] found a prevalence of 174.9/1,000 in Tibet in a study with design similar to that
of the present study, conducting echographic and clinical evaluations in newborns
for DDH over 1 year, in 10 districts of different altitudes. The prevalence of DDH
showed a significant correlation with high altitudes.[10 ]
“The incidence per 1000 live births ranges from 0.06 in Africans in Africa to 76.1
in Native Americans. There is a significant variability in incidence within each racial
group by geographic location.”[11 ] The role of acetabular dysplasia and adult hip osteoarthritis is complex.[11 ] Archaeological studies demonstrate that the epidemiology of DDH may be changing.[11 ]
In 2010, Pollet et al.[12 ] found a prevalence of 6.6/1,000 live births in a province in Canada, while other
studies described an average incidence of 1 to 2/1,000 based only on clinical screening.
A study[13 ] conducted in the United Kingdom (UK) and Ireland found an overall incidence of 6.7/1,000
in Ireland based on clinical screening with late presentation. In the UK, 0.34/1,000
infants had late DDH (after 3 months) based on an ultrasonography screening performed
in a program involving 107,440 newborns.[13 ] In Norway, Engesæter et al.[14 ] found late presentation (after 1 month) in 0.32/1,000 newborns based on neonatal
ultrasonography screening of a large group of newborns with risk factors and/or clinical
findings. Güler et al.[15 ] reported a prevalence of 9.9% in Turkey.
In the present study, the main risk factors found were: female gender (RR = 5.78;
p = 0.0002), breech presentation (RR = 2.94; p = 0.0007), firstborn (RR = 3.36; p = 0.0008), and white ethnicity (RR = 3.69; p = 0.0015), which were strongly associated with an increased risk of developing DDH.
Black ethnicity is a protective factor; there were no black patients with DDH in the
present study. There were only two pairs of twins eligible for examination in the
first three days of life, who did not present DDH; thus this did not present a considerable
statistical value.
Even though the characteristics family history and being a twin were shown to be significant
risk factors in another study,[16 ] in the present study, they were not adequately assessed due to the selection bias,
a limitation that mainly involves the exclusion factors, such as newborns in intensive
care, prematurity, respiratory distress, hypoglycemia, or any clinical alteration
that requires specific care.[16 ] There were only two pairs of twins eligible for examination in the first three days
of life, who did not present DDH; thus did not present a considerable statistical
value. We could not accurately assess the family history because most parents or legal
guardians did not know about the existence of DDH cases in their families, which may
have been underestimated in the resent study. Other limitations are the fact that
the ultrasounds were performed by a single pediatric sonographer, and there was no
follow-up of the development of the patients after the ultrasound. We also emphasize
that the spresent tudy was performed in a hospital in the city of São Paulo, and we
have reported the prevalence of the disease in this establishment.
Early diagnosis is considered essential for an effective treatment and a good prognosis.[17 ] Considering that there are cases of mild dysplasia, which, according to studies,
resolve spontaneously in ∼ 90% of the cases,[18 ] an ultrasonography assessment of all newborns who do not present risk factors or
clinical signs at 21 days of life could reveal cases of immaturity that would benefit
from conservative treatment. We understand that there is a wide discussion about early
treatment[19 ] in cases of immaturity of the hips. However, we know that the psychological, family
and socioeconomic impact seems to be superior if we compare a treatment performed
with a brace of Pavlik to surgical treatment with plaster-cast immobilization. Carlile
et al.[20 ] suggest that newborns with important risk factors or positive clinical signs should
undergo the examination until the sixth week of life.
Regarding DDH screening, it is recommended that morphology and stability be verified
by ultrasonography, in addition to the clinical examination, which, similar to what
happens in other countries, could be implemented in Brazilian institutions with easy
access to a sonographer with experience to perform this exam. The physical exam does
not always detect early acetabular dislocation, the acetabular index, and subluxation.
The late diagnosis of dislocation is not an evidence that the physical test was not
properly conducted. An increased acetabular index may provide the femoral head to
run laterally out of the acetabulum and develop into a postponed dislocation. The
late diagnosis of subluxated or dislocated hip is not necessarily associated with
the neonatal subluxable or dislocatable hip.[3 ]
There is a high rate of false-negative clinical examinations, even when performed
by experienced physicians.[21 ] To assess whether the result of the present study reflects the prevalence of the
disease among the population of the São Paulo metropolitan area, further studies conducted
in other centers and with larger samples are required.
Conclusion
The present study showed that the prevalence of ultrasound exams showing DDH in the
sample was of 5.45%. The main risk factors associated with a higher risk of developing
DDH were female gender, breech presentation, firstborn, and white ethnicity.
