The Epidemiology and Demographics of Hip Dysplasia

The etiology of developmental dysplasia of the hip (DDH) is unknown. There are many insights, however, from epidemiologic/demographic information. A systematic medical literature review regarding DDH was performed. There is a predominance of left-sided (64.0%) and unilateral disease (63.4%). The incidence per 1000 live births ranges from 0.06 in Africans in Africa to 76.1 in Native Americans. There is significant variability in incidence within each racial group by geographic location. The incidence of clinical neonatal hip instability at birth ranges from 0.4 in Africans to 61.7 in Polish Caucasians. Predictors of DDH are breech presentation, positive family history, and gender (female). Children born premature, with low birth weights, or to multifetal pregnancies are somewhat protected from DDH. Certain HLA A, B, and D types demonstrate an increase in DDH. Chromosome 17q21 is strongly associated with DDH. Ligamentous laxity and abnormalities in collagen metabolism, estrogen metabolism, and pregnancy-associated pelvic instability are well-described associations with DDH. Many studies demonstrate an increase of DDH in the winter, both in the northern and southern hemispheres. Swaddling is strongly associated with DDH. Amniocentesis, premature labor, and massive radiation exposure may increase the risk of DDH. Associated conditions are congenital muscular torticollis and congenital foot deformities. The opposite hip is frequently abnormal when using rigorous radiographic assessments. The role of acetabular dysplasia and adult hip osteoarthritis is complex. Archeological studies demonstrate that the epidemiology of DDH may be changing.


Introduction
Demography is the study of human populations with reference to size, diversity, growth, age, and other characterizing statistics [1]. Epidemiology is the study of the incidence, distribution, and determinants of disease frequency in groups of individuals who happen to have characteristics in common (e.g., gender, ethnicity, exposure, genetics) [2,3]. Incidence is the proportion of new cases in the population at risk during a specified time interval; prevalence is defined as the proportion of individuals with the disease in the study population of interest. Demographic and epidemiologic studies can determine risk factors for a disease/condition of interest, shed light on its etiology, and guide potential prevention programs.
Developmental dysplasia of the hip (DDH) is an epidemiologic conundrum [4]. DDH encompasses a wide spectrum of pathology ranging from a complete fixed dislocation at birth to asymptomatic acetabular dysplasia in the adult [5][6][7][8][9]. The epidemiologic literature regarding DDH is vast and confusing due to different definitions of hip dysplasia, different methods of diagnosis (e.g., physical exam, plain radiographs, ultrasound), different ages of the population studied (e.g., new born, 1 month old, 3 months old, etc.), clinical experience of the examiner [10], different ethnicities/races in the examined population, and different geographic locations within similar ethnic populations [11,12]. Neonatal hip instability, now even more apparent with hip ultrasonography, must also be addressed [13,14]; the clinical challenge is to separate the neonatal hip instability which resolves spontaneously from that which is significant [15][16][17][18][19][20][21].
The last major review of the epidemiology of hip diseases was in 1977 [22]. The goal of these manuscripts is to update the current knowledge of the epidemiology and demographics of pediatric hip disease which may lead to significant morbidity in later life.

Childhood Hip Dysplasia
Incidence. There are three eras in modern medicine when the incidence of hip DDH has been determined [26]. Period I (1920s to 1950s) was when the incidence was arbitrarily estimated. Period II (1950s to 1980s) was when the incidence was determined based on the detection of unstable hips on neonatal physical exam plus the addition of late-diagnosed patients. Period III (1980s to present) incorporates hip ultrasonography (both diagnostic and screening) (Figure 1). Generally accepted ranges for the incidence of DDH for Period I range from 0-40% (0% Africans and up to 40% in other racial groups), for Period II 0.04%-16.8% (0% in Africans and up to 33% in Native Americans), and for Period III 4.4%-51.8% (4.4% for Africans and 7.15% the lowest for Caucasians) [26]. The wide range of DDH incidence in Period III reflects differences in the definition of DDH. Some studies include any hips with ultrasonographic instability, whereas others include only those types ≥ Graf IIb (Table 1, Figure 1).
Hip ultrasonography creates confusion due to differences between the neonatal physical exam and ultrasound findings [15-17, 27, 28]. The interobserver variability (κ 0.4 to 0.5) in determining hip stability is also poor to moderate [29][30][31]. The wide variability in incidence in the ultrasonographic era is better understood when considering two different groups of neonatal instability. The first is neonatal, sonographic DDH which resolves spontaneously. The second is neonatal instability which if left untreated may progress to true DDH, ranging from acetabular dysplasia to complete dislocation. Hadlow [32] noted that 50% of unstable hips at birth stabilized completely within 5 days; Barlow [33] noted that 90% of unstable hips at birth become normal by 2 months; Abdinejad et al. [34] noted that 97% of neonatally unstable hips resolved spontaneously by 6 months. Many of the sonographic DDH hips likely reflect these clinical findings [20,[32][33][34] and are in the first group.
The incidence/prevalence is quoted as the number per 1000 live births. The data presented here is a best attempt at synthesizing the literature during these different periods; insufficient data was often present making it difficult to calculate demographic variables. Race is classified using the definitions of Eveleth and Tanner: Caucasians, Africans in Africa and of African Ancestry, Asiatics (Amerindians, Hispanics, Indonesian-Malays), Indo-Mediterraneans (inhabitants of the Near East, North Africa, and Indian subcontinent), and Australian Aborigines and Pacific Island peoples [35].

.1. Conventional DDH Indigenous Peoples (Table 2(a))
(i) Native Americans. The incidence of DDH is high in Native Americans, likely due to a combination of genetics and swaddling. In the Arizona Fort Apache Indians [43], the incidence was 31, but likely higher, since cases of dysplasia and subluxation were excluded. This particular group of Native Americans represent a very tight gene pool as they have maintained their endogamous marriages (band members only marrying within their own band) after migrating approximately 600 years ago from three different bands in Western Canada. In Navajo children [42] from Fort Defiance, Arizona and Gallup, New Mexico, the incidence was 67. A positive family history was present in 33% of the dysplasia cases but only 6.1% of the nondysplastic cases. In the Navajo from the Arizona Many Farms District, [41] ISRN Orthopedics 3 Abductor muscles * (a) α β (b) α β (c) α β (d)  As described by Graf [23], Roposch et al. [24], and Herring [25].   1.07 * The incidence and either the numerator/denominator were given; appropriate values calculated when possible. @ calculated from the given incidence and total number of births. the overall prevalence of DDH was 37.9 : 33.3 for adults and 40.1 for children. Complete dislocation was more common in adults and simple acetabular dysplasia/subluxation more common in children. The ratio of childhood dysplasia to dislocation was 4.5 to 1 and in adults 0.3 to 1.
In an early study of the Cree-Ojibwa, Island Lake, Northern Manitoba, the incidence was 36 [36]; in a later more detailed study, the incidence of frank dislocation and subluxation was 110 [37]. The prevalence of DDH for all ages was 336 [37] (195 for frank hip dislocation or subluxation, 99 for dysplasia, and 54 for other types). In the Cree in Northern Saskatchewan, the overall prevalence was 13.2 [38]. In Ontario Native Americans [39], the incidence ranged from 12 to 123. Using weighted averages, the average incidence of DDH in Native Americans is 76.1 for all dysplasia (Figure 2(a)).
(iii) Africans. DDH is extremely rare in Africans (Table 2(b)). In Sub-Saharan Africans, 2 cases of typical DDH were described in the Bantu [48]. There were no signs of hip dysplasia at 3 months of age in another study of 16678 Bantu children [47], despite breech presentation in 897 (5.4%). In the Kikuyu Bantu, Kenya, 2 cases of typical DDH are described [154]. In a review of 284 children with congenital orthopaedic malformations in an African teaching hospital (Ibadan, Nigeria), DDH accounted for only 2.2% of all congenital malformations [155].
This immunity of the African infant from DDH may be due to deeper acetabulae [156], genetic factors [157], and the absence of swaddling in African cultures. Carrying the infant in an abducted position straddling the iliac crest is postulated as protective against DDH in the African peoples. However, in the United States, the acetabular indices of Caucasian and African infants showed minimal differences at birth but by 6 to 12 months of age were actually slightly higher (or shallower acetabulae) in Africans [158]. Genetic mixing between Africans and other races with a higher incidence of DDH (e.g., Caucasians in the United States) [48,159] results in a higher but still comparatively low incidence of DDH. Quoted incidences in African Americans are 0.21 [51] and 0.46 [50] compared to 1.5 in American Caucasian infants [50]. Using weighted averages, the incidence of DDH is 0.06 in Africans in Africa and 0.40 in the United States.
(c) Americas. There are few incidence studies in the United States due to its highly mobile population. The incidence is 0.7 in Jefferson County, Alabama [51], 0.8 in Portland, Oregon [150], 1.1 in Iowa [163], and 0.7-6.1 in Utah [164]. In Llanquihue, Chile [149], the incidence is 4. (vii) Mixed Races. In a study of 432778 infants born in Birmingham, England between 1960-1984 [165], the birth prevalence of DDH was 2.77 when both parents were Caucasian, 1.37 when both were South Asian (from India, Pakistan, Bangladesh), and 0.66 when both were Caribbean (primarily African). These numbers are similar to the average weighted incidences in this study (3.6 for the United Kingdom, 2.4 for Indian, and 0.1 for Africans), the value for Caucasians and South Indians slightly lower than ours, while that for the Africans is slightly higher. These values changed with mixed matings; 2.77 to 0.78 when one parent was Caucasian and one South Indian and 0.66 to 1.28 when one parent was Caucasian and one Caribbean/African. This confirms the differences noted in the United States with genetic mixing in Africans. Other incidence figures for mixed or unknown racial groups are shown in Table 2(b). (Table 3) Indigenous Peoples. The incidence of neonatal hip instability in the Maori is less than Caucasians [191], where 16% of the births in one hospital were Maori, but only 7% of the DDH cases were Maori.
(v) Caucasians. In Europe, the incidence of neonatal hip instability is 4.1 in Uxbridge, England [ [188], and 61.7 in Poland [190]. In Australia/New Zealand, it is 3.4 and 8.5 in Auckland, New Zealand [191], 6.6 in Western Australia [192], 6.7 in Sydney, Australia [193], and 19.4 in Victoria, Australia [194]. In North America, it is 1.4 in Iowa City [196], 5.2 in Oklahoma City [167], 8.6 in Salt Lake City [195], 9.2 in Indianapolis, Indiana [198], 5.7 [207] and 9.9 in Vancouver, British Columbia [197], and 15.3 in New York City [168]. The average weighted incidence for all Caucasians is 10.8 (6.0 in Australia/New Zealand to 23.2 in Western Europe). Table 4). For this paper, sonographic DDH is defined as a hip > Graf IIa. North African Black infants (Ethiopian Jews) have a sonographic incidence of 12.4 at birth and 1.5 at 4 to 6 weeks [157]. In Africans living in London, the incidence of sonographic DDH was 0 (0 of 185) [208]. Caucasian infants demonstrate a sonographic incidence at birth from 7.6 [64] to 847 [216], with a weighted average of 131. At 4 to 6 weeks of age, this drops to 12.8-14.8, with a weighted average of 14.1. In Turkey, the incidence of sonographic DDH at 6 weeks is 47.1 [60].

Ultrasonographic and Clinical Screening Period (1980s to Present) (
Composite results (Table 4) denote an average incidence of ultrasonographic DDH in Caucasians at birth of 80.0 (7.6 to 847) and 42.2 (3.8 to 103) for hips > Graf IIa (range).

Gender, Laterality, Family History, Perinatal Factors
Breech presentation/presentation also influences neonatal hip instability. In 6571 live births (257 breech) [244], the odds ratio (OR) of hip instability was 3.42 in all breech babies and 11.1 for those with DDH needing treatment. The incidence of clinical hip instability in breech babies is 44 in Norway [179] (61 specifically in Trondheim, Norway [245]), 71 in New York City [168], 107 in Thailand [236], 131 in Leipzig, Germany [189], and 260 in Malmö, Sweden [176]. In Leipzig [189], the incidence of neonatal hip instability was 131 in breech and 30 in vertex presentations; in Tiachung, Taiwan, these numbers were 8.9 breech and 0.6 vertex [175]. In Dammam, Saudi Arabia [170], breech presentation was present in 13% of newborns with neonatal hip instability and 2.1% without. In Trondheim, Norway, the incidence of    ultrasonographic hip instability in breech presentation is 61 [245]. In Germany, the incidence of ultrasonographic (>Graf IIa) neonatal hip instability in 3739 newborns was 136 in 317 breech children and 64 in nonbreech children [242]. In another German study; however, there was no correlation between intrauterine presentation and sonographic hip instability [215]. Breech-type (frank breech or bilateral hip flexion/knee extension, nonfrank, or varying amounts of hip and knee flexion) is also important. The incidence of DDH in Hungarian breech children was 340 in nonfrank and 185 in frank breech [241]. In Norway [245], the incidence of DDH in frank breech was higher than other breech types. DDH in breech children may be decreased by elective Caesarean section [246]; of 941 breech presenting infants, the incidence of DDH was 3.69% (19 of 515) when delivered by elective pre-labor Caesarean section, 6.64% (26 of 241) when delivered by intrapartum Caesarean section, and 8.11% (15 of 185) when delivered vaginally. In New York, children born by Caesarean section had a 3.4 times higher chance of DDH when breech compared to vertex presentation, and those born vaginally had a 7.0 times higher chance of DDH when breech compared to vertex presentation [168].
The incidence of DDH is less in premature and lowbirth-weight infants [44]. Children <37 wk gestational age had an incidence of 3.4, 37-42 wks 6.3, and >42 wks 29.9; for those <2500 gms, the incidence was 4.1, 2500-3499 gms 6.7, and >3500 gms 6.4. In another study, all babies with DDH had a gestational age of 39 weeks or more, and 52% were firstborn [234]. In Northern Ireland, the risk of DDH was less when the birth weight was <3.0 kg [118]. Children born large for gestational age (LGA) have an increased risk of hip subluxation (OR 1.73) [247] and twice the incidence of ultrasonographic DDH (> Graf IIc) compared to normalbirth-weight newborns (6.1% versus 3.5%) [248]. In breech presentation, DDH is more common in those with higher birth weights (3.49 kg with DDH, 3.06 kg without DDH) [245]. Very-low-birth-weight infants are not at increased risk of DDH [249].
Primiparity increases the risk of DDH. In Hungary [142], the average birth order of 1767 children with DDH was 1.37 compared to 1.54 for 108966 control children showing that DDH children are skewed to the first born. In Finland, 63% of DDH children were first born compared to 55% in the normal population [105], in Southampton, England, 83% of children with subluxation were first born [112], and in Madrid, Spain, 50% of were firstborn [122]. In Western Australia [44], the incidence was 7.6 if first born and 5.9 if multiparous. At Christchurch Women's Hospital, first-borns accounted for 83% of DDH children but only 42% of all births [146]. In Utah, USA, the primiparity was 46% greater than expected in the 327 cases of DDH. In Singapore [56], 43.8% to 50% of DDH children were firstborn [54].
Consanguinity results in a very high incidence of DDH in Japan [81] and the Middle East where 25% [169], 40% [65], and 49% [225] of DDH cases were from consanguineous parents. In western Galilee, an area with a high incidence of DDH, marriage between 1st cousins is frequent [61,257]. Tight gene pools were implicated in the high incidence of DDH in Fort Apache Navajos [43].
To simplify these figures, epidemiologists use decision analysis/meta-analysis or multiple logistic regression to determine DDH risk factors. The clinical practice guidelines of the American Academy of Pediatrics [5] used a decision analysis model, concluding that the baseline incidence of DDH (not at risk children) was 11.5 (4.1 for boys and 19 for girls). The relative risk with a positive family history is 1.7 times higher (for an absolute incidence of 6.4 for boys and 32 for girls), and the relative risk for a breech presentation compared to vertex is 6.3 times higher (for an incidence of 29 for boys and 133 for girls).
In 51 Israeli DDH children (1994)(1995), ORs were 6.0 for breech presentation, 4.3 for female gender, and 2.7 for primiparity [258]. In 81 Nottingham children with DDH and 323 control subjects [259], the ORs for those born by Caesarean section or breech delivery were 3.29 and 4.57, respectively, and 0.55 and 0.53 for multigravida and multiparity; maternal age, gestational age, placental weight, race/ethnicity, and birth weight/height were not risk factors [259].
In 1,059,479 Norwegian children [179], predictors of neonatal hip instability (clinical exam) were gestational age, gender, and breech presentation; birth order had no effect. The overall incidence of neonatal hip instability overall was 10.3 with differences by birth weight (2.6-<2500 gms; 10.2-≥2500 gms) and presentation (

Twin Studies.
In 589 children with DDH, there were five pairs of twins (1.7%): two monozygotic and three dizygotic [162]. The frequency of concordance (both twins having the same pathology) was ∼33% in monozygotic twins and ∼8% in dizygotic twins [260] supporting a polygenic mechanism. Even with this high of a concordance rate, the majority of monozygotic twins are not concordant, demonstrating that the same genetic background and intrauterine environment does not result in DDH most of the time [260]. The Norwegian Twin Registry found an incidence of 4.3 DDH in twins (38 of 8794). The OR of DDH in a relative of an effected twin was 10.0 : 35.8 for mothers, 12.7 for siblings, 8.1 for fathers, and 3.3 for subsequent offspring [261]. The increased OR for mothers over fathers suggests a maternal effect. In 4678 opposite sex twins, the relative risk of DDH was 0.13 for males compared to females [262]. Mirroring has been described in two sets of identical twins [263,264] with DDH; one twin had left DDH and the other right DDH. Both sets also had mirrored strabismus, and thumb sucking was mirrored in one set [263].

Genetic Patterns.
The prevalence of DDH in France in 1912 was 8 in Paris and 36-41 in Rennes, (Brittany region) [12,184] with an epicenter in Pont l'Abbé. DDH did not occur in this geographic area in the 1700s, and it appeared that a genetic mutation occurred in the 19th century [184]. The genetic mechanism was postulated to be autosomal dominant with incomplete penetrance and penetrance dependent upon gender [12]. In 16 Turkish family members with DDH, an autosomal dominant inheritance was also noted [265].
Most other investigators propose a two-gene system in DDH. In 589 children with DDH [162], one gene involved joint laxity (dominant) and the other acetabular dysplasia (polygenic). The risk of subsequent members of a family having DDH when one family member had DDH was 36% when there was one affected parent and child, 12% when there was one affected parent and no affected children, and 6% when both parents were normal and had one affected child with DDH [266]. A Hungarian study [267] confirmed the role of polygenic acetabular morphometry and monogenic joint laxity; neonatal and "late-diagnosis" cases did not seem to be different entities. In an Italian study of pedigrees from 171 patients with familial DDH [268] (1937 individuals in 507 nuclear families with 499 having DDH), segregation analysis demonstrated that a two-locus recessive-recessive model was the most appropriate fit. The major locus demonstrated a penetrance of 5.9% in males and 58.6% in females, and the modifier locus 0.3% in males and 15.6% in females.
Homozygous recessive conditions (attached earlobes, light eyes, color blindness, inability to roll the tongue, and 3 wrist creases) are increased in children with DDH [269]; 5.2 ± 0.1 in controls and 7.1 ± 0.2 in DDH.

HLA and ABO Blood Type Groups.
In Greek children with DDH [270] there was a significant increase in HLA A 1 in the DDH group (52 versus 26%). In Czechoslovakia, there was a significant decrease in the HLA B 7 type in the DDH group (10% versus 26%) [271]. In Japanese [272] the development of DDH and dysplastic osteoarthritis of the hip is associated with genes in the HLA-D regions, especially HLA DR4. No differences in ABO blood types [168,259,269,273] or Rh blood types [259] have been found in DDH.

Associated Genes and Chromosomal Locations. In Han
Chinese, DDH was associated with gene growth differentiate factor 5 (CDF5) (OR 1.40) [274]. The double von Willebrand factor A gene on chromosome 3p24.3 is strongly associated with osteoarthritis of the knee in Japanese and Chinese [275] but not DDH [275]. Estrogen receptor Xba I wild type (XX compared to Xx and xx) is more common in DDH than controls (55.8% versus 37.9%) [276]. Homozygosity for the mutant Taq1 vitamin D receptor t allele is associated with an increased acetabular index, and the Pvu II pp estrogen receptor genotype is associated with a low center-edge angle [276].
Several studies associate DDH with chromosome 17q21. This region contains a cluster of HOX genes that provide specific positional identities to mesenchymal cells in developing joints [277]. In French Caucasians, there was no increased association between DDH and the HOXB9 gene [270] while in the Chinese there is an association with the HOXB9 as well as the COL1A1 genes in DDH [278]. In Italian Caucasians, the COLL2A1 and vitamin D receptor genes are associated with nonsyndromic DDH [279] but not confirmed by others [280]. In an 18 member multigenerational family affected by DDH, a 4 Mb region on chromosome 17q21.32 was linked to DDH [277]. [235,266,267,[281][282][283]: 75% of boys and 33% of girls with DDH demonstrate joint laxity [235]. The prevalence of joint laxity (≥3 joints) in normal children ages 6-11 years was 10.5% and 47% in DDH children [282]. Increased joint laxity likely explains the increase in inguinal hernias in DDH children and their fathers and brothers [142,266,283]. Girls with DDH have a 5-fold increase and boys with DDH a 3fold increase in inguinal hernias; the hernia also develops earlier in life than normal [284]. Relaxin, which stimulates collagenase, alters the connective tissue and may lead to the development of both DDH and inguinal hernias.

Ligamentous Laxity. Joint laxity is increased in DDH
Collagen metabolism is altered in DDH. In newborns with DDH, the amount of umbilical cord collagen was decreased (∼26%) [285] but not total protein [286]; this was not found in another study [173]. Umbilical cord type III/I collagen ratio is increased in DDH [286]. Hip capsule and skin biopsies in older DDH children (1 to 4 years old) show a decreased type III/I collagen ratio in the hip capsule but not the skin [287]. Opposite results were found in Japanese children with DDH [288] where the type III/I collagen ratio was increased in the hip capsule and ligamentum teres. The ligamentum teres in DDH children demonstrates thicker collagen fiber bundles and increased hyalinization in those with complete dislocation compared to subluxation [289].
In patients with hypermobility syndrome (including several cases of DDH) [290], growth hormone, insulin, and IGF-1 levels were elevated, leading us to endocrine abnormalities in DDH.

Hormonal/Endocrine and Metabolic
Associations. DDH occurs predominantly in females, and a hormonal/endocrine relationship has long been suspected. In newborns with DDH, there is an increase in urinary excretion of conjugated estrogen and 17 β-estradiol [291][292][293], suggesting that DDH may be due to abnormal estrogen metabolism. Abnormal fetoplacental estrogen metabolism in a mother and child with DDH has been described [294].
Others found no estrogen metabolism abnormalities. Urinary estrogen excretion in the first 3 days of life in 7 newborns with bilateral Ortolani instability and 4 normal newborns was not different [295]. Urinary excretion of oestradiol, oestrone, and oestriol in 16 neonatal hip instability, and 19-matched-control newborns [296] showed no differences. Similar conclusions were reached by Borglin [297]. Umbilical cord blood measurements of 17 β-estradiol levels in 2185 consecutive newborns [298] with neonatal hip instability showed conflicting data; in boys low levels (<10 nmol/L) increased the risk of neonatal hip instability (OR 54) while in girls high levels (>15 nmol/L) increased the risk (OR 2.2).
Women who develop pelvic joint instability and pain in pregnancy have an increased risk of a child with DDH [299]: incidence of 7-9 compared to the normal 2-3. In another study of Norwegian women with pregnancy-initiated pelvic pain, the incidence of DDH in the children was 45 [300], 7 times normal [179]. Women with pelvic joint instability and pain have higher serum relaxin levels compared to those without pelvic pain [301] in the third trimester which may explain these associations; others refute this [302,303]. There is no correlation between serum relaxin in umbilical cord blood and neonatal hip instability [304,305].
In adults previously treated for DDH without surgery, bone mineral density at the hip but not the spine is decreased with a concomitant increase in osteocalcin [306]. The OR of decreased bone mineral density at the hip was 6.3 for those with DDH. Some children with DDH demonstrate lower serum calcium and alkaline phosphatase but not serum phosphorus [307].

Seasonal Variation.
Most studies note an increase in the number of DDH births in the winter. In Tokyo, 549 of 1276 (43%) DDH children were born on November-January [308]; in Kochi, Japan, 55 of 106 (52%) were born on December-February [81]. Other Japanese studies confirm this winter predominance [83,90,309], likely due to the increase of extension diapering/swaddling in the colder months. In Navajo boys, a slight increase was noted in the midwinter [37] but no seasonal variation was noted in other Navajos [40].
In Caucasians many studies also note a predominance of DDH births in the fall/winter months: September-December in Scandinavia [97][98][99][100][101][102]; September-November throughout Sweden [176,310]; October-January in Israel [70]; Novem-ber in Alicante, Spain [243]; September-January [311] and November-March in Leipzig [312]; December-February in Berlin [313], Northern Ireland [118], and Scotland [162]; December in all of the United States [231]; September-February in Utah [164]; September-March in England [110,314,315]; December-January in Debrecen, Hungary [141]; October-February in Budapest, Hungary [142]; November-February in Dêčín andČeská Lípa, Czechoslovakia [134]; December-March in Tel Aviv, Israel [201,316]; and January-April in East Anglia, England [317]. In the southern hemisphere, the same was also noted: July-August in Southland, New Zealand [147], and in Australia, April-June in Victoria [318], June in Adelaide, and July in Brisbane [319]. Infants born in colder winter months demonstrate poorer acetabular development compared to those born in the warmer months [320] as measured by acetabular depth and acetabular angles; this may explain the increase in DDH in children born in the winter or may represent an effect of increased swaddling or tight clothing to protect the baby from the colder weather. Pelvic insufficiency also shows a seasonal variation [321] with a peak in November-December.

Swaddling.
Swaddling theoretically maintains the hip in an extended and adducted position; radiographs of hips in cradled infants demonstrate that they are not in true adduction, and up to 15-20 • of abduction is possible [40]. Newborn swaddling used in many cultures [323] is considered by many to be a significant factor in the development of DDH. In ∼2300 Navajos [40], hip dislocation was more prevalent in adults than children. It was postulated that this decrease in dislocation was due to the transition from exclusive use of the Navajo cradleboard to a more frequent use of diapers.
Swaddling/cradle boards in a few studies are not associated with DDH. In the Navajo, 9 of 13 cases of DDH which spontaneously improved showed improvement while on the cradleboard, and 2 cases with complete dislocation were never kept on the cradleboard [40]. Nearly all the Cree-Ojibwa swaddle their children (tihkinākan cradleboard) [38], yet in 427 untreated Cree-Ojibwa infants with hip dysplasia, frank improvement was seen in 113 with ≥1 year of followup. Most interestingly, 21 of the 113 had an initial diagnosis of complete dislocation [37]. Swaddling is also used in Saudia Arabia (mehad) [326] and Iran (ghondagh) [34] with no differences in the use of mehad/ghondagh for those with or without DDH. These studies question the role of swaddling.
Most other studies demonstrate a strong correlation between swaddling and DDH. There was a 10-fold increase in DDH (123.0 versus 12.6) in Canadian Native Americans 24 ISRN Orthopedics from Ontario who used the cradleboard [39]. Similar concerns were noted in the Cree from northern Quebec [327]. In Turkey [59], 98% of DDH cases were swaddled as infants compared to 87.1% not swaddled; the OR of DDH in swaddled children was 6.1 [223]. In a sonographic Turkish study [60], swaddling was used in 21.2% of those children with DDH ≥ Graf IIb (P < 0.001), which was the highest risk factor associated with DDH, greater than breech delivery (9.0%), positive family history (6.6%), and female gender (6.2%). The incidence of ultrasonographic dysplasia in children at high risk for DDH in Qatar [328] dropped from 20% to 6% after a community awareness program demonstrating the harmful effects of swaddling and discouraging its use; swaddling alone did not cause the hip dysplasia but had an unfavorable effect on the future course of a dysplastic hip. Another study from Saudi Arabia concluded that swaddling was an important factor in the etiology of DDH [329]. Swaddling is also believed to be responsible for the high incidence of DDH in the Arabic peoples in Western Galilee [61] and Iraqi immigrants in Israel [330].
In Europe, Békés County, Hungary [139], swaddling was believed to account for the high incidence (28.7) of DDH and supported by others [134]. In the Swedish Sámi, the cradleboard (gietka or komse) was also believed to account for the high incidence (24.6) of DDH [46]. The gietka is a cradle hollowed from a log tightly swaddling the lower extremities and allowing for minimal movement. This cradle was very practical in the Sámi nomadic culture and lifestyle and allowed mothers to carry the cradle across their shoulders onto a reindeer's pack saddle; as their nomadic reindeer herding culture and lifestyle decreases and health nurses forbid the use of the geitka/komse [324,331], the prevalence of DDH in the Sámi has fallen (Figure 4(a)).
DDH is very rare in cultures where swaddling is not used (Southern Chinese, African Bantu, Thailand, North Korea, Sri Lanka [332]); the absence of swaddling is believed to be responsible for this [333]. Two different northern circumpolar peoples (Sámi and Inuit/Eskimos) have markedly different incidences of DDH [324,334]; the Sámi, who swaddled their young in the past, had a very high incidence, while the Inuit/Eskimo's, whose mothers carried their young inside their parkas in a hood (amauti) abducting the hips around their backs have an incidence of DDH similar to Caucasians.
In Japan, swaddling/diapering in extension was strongly associated with DDH [85]. This is especially so for children born in the winter months, being more tightly wrapped to protect against the cold [335]. This effect of swaddling/ diapering with the hips in extension was initially noted in a small series of 5 normal newborn children; no hip instability was noted within the 1st 24 hours on physical examination, but after diapering in extension, 4 of 5 children developed Ortolani-positive hip instability [84]. This resulted in a larger study, and the creation of an educational campaign regarding the problems associated with extension diapering/swaddling in Kyoto, Japan [85]. The incidence of DDH dropped from 52.9 in 1971-97 to 5.6 in 1974-1976 after the educational campaign [85]. These results in Kyoto led to a national Japanese education and prevention campaign [335] with similarly striking results (Table 2(c), Figure 4(b)). This demonstrates the impact of epidemiological/demographic studies in reducing the occurrence of a particular condition.
Swaddling thus influences the development of DDH [336,337]. This is particularly relevant since there is a strong resurgence to return to swaddling to reduce crying and promote uninterrupted sleep in the baby [336,338,339]. Swaddled infants arouse less and sleep longer; preterm infants show improved neuromuscular development, selfregulatory ability, and less physiologic distress when swaddled [338]. Infants that are at an increased risk of DDH should probably not be swaddled unless imaging studies are absolutely normal [336]; appropriate swaddling also must be used [340].
3.9. "Late Diagnosis" of DDH. When the hip actually dislocates [341][342][343] has been debated for some time [344]. The studies of "late" diagnosis reveal several findings. In one study, all the children were female, and 40% (8 of 20) had recognized risk factors [345]. Canadian children [346] diagnosed with DDH at an older age (20 months) have more right hip involvement (31% versus 15% right hips) compared to those diagnosed younger (<20 months) who have more left hip involvement (57% versus 15% left hips); bilateral dislocations are also more common when diagnosed at an older age (44 versus 28%) [346]. In Norway the incidence of "late" diagnosis was 2.4; of these 197 Norwegian children diagnosed "late" [347], all hips were stable at birth; 86% were female, the left hip was involved in 48%, right hip in 31%, and both in 21%. Breech delivery occurred in only 6.5%. The incidence of late DDH in all of Norway was 2.4 [347]: 1.7 in Oslo and 0.76 in southern Finland [348]. In Glasgow it was 0.84 before and 0.57 after institution of a selective ultrasound screening program [349]. A positive family history was associated with an increased risk of "late" DDH [350]. Avascular necrosis of the opposite normal hip in children with "late" diagnosis was described (9 of 103 cases) and associated with high, free-riding dislocations [351].

In Utero Environment
Exposures to Agents. Administration of progesterone in the 1st trimester for threatened or habitual abortion [352] increases the risk of DDH as does maternal hypothyroidism [353] and maternal phenylketonuria [354]. Maternal smoking during pregnancy reduces the risk of sonographic DDH ≥ Graf IIa in female but not male neonates [355], most likely due to the effects of maternal smoke on estrogen physiology. The incidence of DDH in children whose mothers were exposed to radiation at Hiroshima or Nagasaki, Japan (0.31 and 0.23) is higher than those not exposed to radiation in Kure, Japan (0.13) [356]. In Seveso, Italy, no increase in DDH was seen after exposure of a densely populated area to TCDD (2,3,7,8-tetracholorobenzo-p-dioxin) [357]. Intrauterine toxoplasmosis [358] or viral infection [132] does not increase the risk of DDH. Iron deficiency anemia in fetal life may been associated with DDH [359].  Figure 4: (a) The prevalence of DDH in the late 1970's by age group in the Sámi living in northern Finland, Lake Inari region. The overall prevalence was 16.3; for those ≥60 years of age it was 60.2, and dropped to 3.5 for those 0-19 years of age. This has been attributed to a decrease in the practice of newborn swaddling using the gietka or komse. Data from Eriksson et al. [324]. (b) The marked decrease in DDH incidence in Japan after introduction of a nationwide educational program for both neonatal hip instability and hip dislocation after 3 months of age. For neonatal hip instability the data was taken from [84,325] and for hip dislocation from [83][84][85][86][87][88][89][90].
Mechanical and/or Physical Exposures. Multifetal pregnancies (e.g., twins, triplets) are not at increased risk of DDH [142,[360][361][362][363][364][365][366][367][368]. There were no cases of DDH in 968 sets of twins and 18 sets of triplets in Tokyo [369]. Two studies noted less DDH in children born to multifetal pregnancies (relative risk 0.57 [370], 0.46 [371]). Although extrauterine pregnancy is rare (1 in 1100 to 1 in 50000 pregnancies) [372], it is often associated with DDH [373,374], suggesting DDH is associated with molding forces in those circumstances rather than a teratological abnormality [372]. Mid-trimester amniocentesis does not increase the risk of DDH [375] whereas 1st trimester likely does [376] (OR 1.22). Premature labor and/or threatened abortion increases the risk of DDH [132,377]. Prior miscarriage increases the risk of a subsequent child having DDH (OR 1.4) [378]. Preterm infants demonstrate a decrease in neonatal hip instability on physical examination compared to term infants [379] and do not demonstrate an increase in ≥Graf IIb DDH even though many are immature (Graf IIa) [380]. Preterm infants demonstrate no increase in DDH even after reaching normal development and gestational age [381].

Parental
Associations. DDH was more common in higher socioeconomic groups in Scotland [162] but more common in lower groups in Martin, Slovakia [137]; no differences in New York City [168] or England [110] were noted by socioeconomic class. Paternal plywood mill workers had an increased risk (OR 2.71) of DDH in 14415 British Columbian children [399].
Older paternal age has been associated with DDH [400]. Younger maternal age is a function of birth order and primiparity [164]. In one study, DDH and maternal age demonstrated a bell-shaped curve with a peak at 30 years of age [401]. Short maternal stature is associated with DDH in Canada [402] and Ireland [403] but not in Britain [403]. Early abduction treatment in children with neonatal instability is associated with increased maternal anxiety, while the use of ultrasound to manage infants with DDH is not [404].

Child Development.
The median walking age in 86 DDH children diagnosed at an average age of 16 months was not statistically different from a control group of 100 normal children with fractures (13.8 months-DDH; 12.5 monthscontrol) [405]. Menarche in girls with DDH was decreased by 6 months compared to others [406]. DDH children have lower skin temperature on the affected lower limb [407].

Spinal and Neuromuscular
Associations. Abnormal sensory evoked potentials in DDH children 10-13 months of age were present in 31% without spinal dysraphism and 56% with spina bifida occulta [451]. The incidence of DDH is 10 times greater in children with infantile scoliosis [452]. One study found no free nerve endings in either the ligamentum teres or hip capsule of DDH children [453] while the ligamentum teres had a higher frequency (66.6%) of type IVa free nerve endings [289] in DDH children, suggesting a role of abnormal nocioception/proprioception [454]. Vibration arthrometry demonstrates low-frequency vibrations in unstable (Ortolani-or Barlow-positive) neonatal hips [455,456].
Electrophysiological and histological examinations of the hip muscles in DDH demonstrate muscle degeneration which is more severe in younger children [457]; the iliopsoas, gluteus medius, and vastus lateralis muscles demonstrated hyaline granulous degenerative changes [458]. Others note minimal histological changes [459].

The Opposite
Hip in Unilateral DDH. The opposite hip in "unilateral DDH" children often develops abnormally: 14% [460], 31% [461], 34% [462], and 43% [233]. The center-edge angle is the most predictive single measurement of delayed acetabular development while the overall best predictor is the hip ratio (CE angle + 100(acetabular depth/acetabular diameter))/2 [461]. A low ratio indicates a more abnormal hip; <22 is abnormal for those <2 years old and <23 for those ≥2 years old. This makes it possible to predict if the opposite hip is "at risk" [461]. Contralateral abduction hip contracture occurs in children with unilateral DDH [463,464] as well as diffuse pelvic asymmetry if diagnosed after 4 months of age [465].

Transition from Childhood to Adult Hip Dysplasia.
Persistent childhood dysplasia [466] and neonatal hip instability predisposes to adult hip disease. The Norwegian Medical Birth Registry was correlated with the Arthroplasty Registry [467]; when adjusting for gender and year of birth, there was a 2.6 times increased risk (95% CI 30-105) for children with neonatal hip instability to undergo total hip replacement. Of the 442 patients undergoing hip replacement, 95 had the surgery due to degenerative joint disease from residual hip dysplasia, yet only 8 had neonatal hip instability. This underscores the need for continued vigilance in the screening of neonatal hip instability and long-term followup. It also confirms that there is a significant amount of hip dysplasia, with no physical findings in childhood, that later becomes symptomatic in adult life (e.g., asymptomatic radiographic dysplasia). Underlying joint laxity may also predispose to the development of adult hip osteoarthritis [468,469].
3.13. Adult Hip Dysplasia. The incidence and demographics/epidemiology of adult hip dysplasia is easier to study yet there are fewer and less detailed studies. The diagnosis of dysplasia, aside from complete dislocation, involves relatively simple radiographic measurements [470]. The center-edge (CE) angle of Wiberg is most commonly used to determine hip dysplasia [398] but only quantifies lateral coverage, ignoring other methods such as anterior coverage assessment with the false-profile radiograph [470]. The literature defining acetabular dysplasia is confusing; some studies use a CE angle <20 • , while others use 22.5 • , 25 • , and 30 • as the cutoff (Table 6). Also, intra-and interobserver agreement in CE angle measurement is ∼ ±3% [471]. Considering these caveats, the prevalence of acetabular dysplasia ranges from 5.9% in Caucasians to ∼19% in the Sámi.
Adult hip osteoarthritis (OA) is either primary or secondary. Primary OA is a diagnosis of exclusion and defined as when neither an anatomic abnormality nor any specific disease is identified as the cause [475]. Secondary OA is due to preexisting conditions, such as DDH, Perthes' , and SCFE. Hoaglund [475] noted that normal hip anatomy was present in >80% of degenerative hip arthritis patients and that an underlying defect in articular cartilage or bone leads to the eventual development of OA in most Caucasians.
For the other 20%, what relations exist between mild dysplasia, ethnicity, and hip OA? The magnitude of adult acetabular dysplasia does not always correlate with degenerative hip disease in adults [385-387, 474, 475]. We must first know the prevalence of hip OA, and there are two ways to determine this number. One is the radiographic prevalence of hip OA. When using this method, the prevalence of radiographic hip OA is greater in men than in women and greater in Caucasians, intermediate in Blacks, and rare in Indo-Malays [472] (Figure 5(a)). The second is to determine total hip replacement rates, since truly symptomatic hip OA will likely come to THR. THR rates in San Francisco [473], in a population-based study, showed significant variation by ethnicity ( Figure 5(b)), with Caucasians having a rate of THR ∼10 times that of Indo-Malay peoples (Chinese, Japanese, Filipino), ∼1.5 to 2 times that of Blacks, and ∼3 to 4 times that of Hispanics.
Variations in the frequency of hip OA are not necessarily caused by differences in the occurrence of hip dysplasia. There are marked differences in the etiology of hip osteoarthritis between Japanese Indo-Malays and American Caucasians [474]. In patients with osteoarthritis of the hip, secondary osteoarthritis was present in ∼82% of the Japanese patients but only 10.1% of Caucasian Americans ( Figure 5(c)). There is more acetabular dysplasia in Japanese adults but more hip OA in British Caucasians [385]. The prevalence of acetabular dysplasia (defined as a CE < 25 • ) was 0.5% in British Caucasians and 17.3% in Japanese, yet the prevalence of hip OA was lower in Japan (0% men, 2% women) compared to Britain (11% men, 4.8% women). In a random effects' statistical model, a negative relationship was actually discovered between acetabular dysplasia and joint space [385]; as acetabular depth increased (CE angle), joint space decreased. This was confirmed in Hong Kong where there is a lower prevalence of radiographic OA in Hong Kong men compared to British men, yet acetabular dysplasia was equal in prevalence in both groups [391]. It was also confirmed in France with a negative correlation between acetabular dysplasia and the risk of hip OA [390]. The highest prevalence of OA was in French men (5.8%) who had a low prevalence of hip dysplasia (1.8%), while the highest prevalence of acetabular dysplasia was in Japanese women (11.6%) who had a low prevalence of OA (3.5%). In Turkish adults [395] there is no direct correlation between acetabular dysplasia and hip OA. The prevalence of acetabular dysplasia is similar between British Caucasian men and Nigerian Black men, yet the prevalence of hip OA was less in the Nigerian men [382]. In the Sámi, there is no evidence that acetabular dysplasia is associated with OA [476].
In other populations; however, there is a definite correlation between acetabular dysplasia/residual childhood DDH and secondary hip OA [474,477]. This is especially true in Asian populations [478], where the development of OA is nearly guaranteed if there is preexisting acetabular dysplasia or subluxation. In Japan, 97% of hip OA cases were secondary, with 88% due to DDH [479]. In elderly United States Caucasian women, mild acetabular dysplasia is associated with an increased risk of hip osteoarthritis [388]. The exceedingly low prevalence of acetabular dysplasia in Saudi people has been suggested as a contributing factor to their very low prevalence of OA [394].

Both Child and Adult Hip Dysplasia-Archeological
Evidence. There has been significant interest in hip dysplasia and archeological remains (Table 7). A systematic review in this domain is quite difficult due to the many detailed, but hard to identify and access, theses and monographs. Thus, the papers discussed here are not included in the 422 of 2277 manuscripts reviewed. We attempted to review the literature as best as possible, well recognizing that some studies have been missed. Many archeological studies are not applicable being solely a case description of a skeleton with DDH with no data from which a prevalence can be calculated. In archeological studies, prevalence rather than incidence is calculated. As DDH begins some time in infancy, and since there were no diagnostic/treatment methods for children with DDH in the eras of these archeological studies, prevalence and incidence become identical. With anthropological materials, the exact diagnosis may be questioned due to many different types of hip dislocation. However, a critical review was performed, and all attempts were made to exclude posttraumatic, neuromuscular, or infectious dislocations.
In the Mediterranean areas, a study of 139 Herculaneum people buried by the Mt. Vesuvius volcano in 79 AD demonstrated one DDH [418]. In Padua, Italy, a study of 213 skeletons (Tedeschi Osteological Collection) from the late 19th and early 20th centuries demonstrated one DDH 28 ISRN Orthopedics      [425], (precolonization era, before 1500), there was one unilateral case of DDH. In 75 Caddo-Indian peoples who lived in what is now the Red River county, Eastern Texas (1100-1800 AD) [429], there was one DDH. In the Ontario Uxbridge Iroquoian ossuary (∼1490 AD), one DDH was seen in 312 skeletons [423]. In 286 skeletons from the Fort Erie, Niagara River, Ontario (precolonization era), two cases of DDH were noted [424]. In 486 Native American Crow Creek villagers massacred in the 14th century on the east bank of the Missouri River in South Dakota [426], one DDH was found (a 6-10 years old child). In 54 skeletons from the Kechipawan Pueblo, New Mexico (1300-1600 AD), no DDH was seen [422]. In two small series of 26 skeletons each from El Morro Valley, New Mexico (13th century) [430] and Sundown, Prescott, Arizona (1100-1200 AD) [431], there were no cases of DDH. The overall gender and laterality mix in the archeological samples is very similar to the composite results in this paper (Table 7). In those skeletons with DDH where the data was known, 74% were female and 35% demonstrated bilateral involvement, similar to the figures of 76% female and 27% bilateral in this paper. The prevalence of DDH in these archeological samples, even though most are small in size except for that from the Spitalfields cemetery, demonstrates differences compared to the incidence of DDH in this paper. Those for Western Europe are similar (4.9 archeological, 3.6 present United Kingdom). All others were different: Eastern Europe (6.0 archeological, 35.8 present), Mediterranean (3.5 archeological, 25.5 present), Middle East (10.6 archeological, 6.2 present), and indigenous Native Americans (6.1 archeological, 76.1 present). If the sampling of the archeological data set was adequate, then statistical analysis should determine if such differences were real. We subjected the data to the Fisher two-tailed exact test, analyzing for differences between the archeological data and the present data. These data (DDH n/total n) and (archeological versus present) were 25/5089 and 3232/906428 for Western Europe; 4/668 and 12452/348031 for Eastern Europe; 2/577 and 2240/87581 for Mediterranean; 2/189 and 732/118225 for Arabic Middle East; 44/8321 and 1108/14553 for indigenous Native Americans. There were no differences for Western Europe (P = 0.12) and the Middle East Arabic (P = 0.33), while there were differences for Eastern Europe and indigenous Native Americans (P < 0.001). These differences  [472]. (b1) Estimated rates of total hip replacements for primary coxarthrosis in San Francisco, 1984-1988, for men. Data from Hoaglund et al. [473]. (b2) Estimated rates of total hip replacements for primary coxarthrosis in San Francisco, 1984-1988, for women. Data from Hoaglund et al. [473]. (c) Differences in the etiology of adult hip arthritis between Japanese and Caucasian patients. Note the rarity of SCFE and Perthes' as a cause of adult hip arthritis. Data from Hoaglund et al. [474]. may represent changing epidemiology, changes in natural selection processes in the archeological era, or inadequate sampling of the archeological materials. Further inquiry will be needed in this subject area.

Conclusion
DDH demonstrates a predominance of left-sided (64.0%) and unilateral involvement (63.4%). The incidence per 1000 live births ranges from 0.06 in Africans in Africa to 76.1 in Native Americans with significant variability between and within racial groups and geographic location. The incidence of clinical neonatal hip instability at birth ranges from 0.4 in Africans to 61.7 in Polish Caucasians. Predictors of DDH are breech presentation, positive family history, and gender (female). Ligamentous laxity and abnormalities in collagen metabolism, estrogen metabolism, and pregnancyassociated pelvic instability are well-described associations with DDH. Many studies demonstrate an increase of DDH in the winter, both in the Northern and Southern hemispheres. Swaddling is strongly associated with DDH. Associated conditions are congenital muscular torticollis and congenital foot deformities. The opposite hip is frequently abnormal when using rigorous radiographic assessments. Archeological studies demonstrate that the epidemiology of DDH may be changing.