Objective. In part 1 of this investigation, we demonstrated that children of multiple birth pregnancies are at higher risk for development of deformational plagiocephaly. In the current investigation, we explore whether certain prenatal and postnatal risk factors predispose one twin over the other by examining the occurrence and severity of plagiocephaly in both discordant (only 1 affected) and concordant (both affected) twin pairs.
Methods. Throughout 1999, we obtained detailed medical histories on 140 sets of twins who had presented for treatment at 1 of 9 treatment centers across the United States. The study cohort consisted of 46 concordant and 94 discordant twin pairs (a total of 280 study participants). Information about prenatal and postnatal history of each infant was obtained through detailed review of the children’s medical records as well as analysis of a patient database maintained on all infants who receive treatment. Follow-up interviews were performed to verify the information recorded and to obtain additional information about the child who had not received treatment (if appropriate) or use of reproductive assistance (fertility drugs, intracytoplasmic sperm injection, etc) that had not been previously recorded. Statistical analyses were performed to assess the effects of prenatal and postnatal risk factors (in utero position, in utero orientation, birth weight, neck involvement, sleeping position) with respect to which infant was affected in the discordant twin pairs and to which infant was more severely affected in the concordant twin pairs.
Results. Statistical analyses of both discordant and concordant twin pairs demonstrated that the lower in utero infant was significantly more likely to be affected (χ2 = 17.391). In addition, the more severely affected infant was significantly more likely to have some form of neck involvement (torticollis, neck tightness; χ2 = 46.380), as well as have been carried in a vertex position (χ2 = 7.408). Conversely, neither sleeping position nor gender was found to be associated with development of plagiocephaly.
Conclusions. The results of this investigation strongly support an in utero cause of plagiocephaly and demonstrate that intrauterine positioning may play a prominent role in determining both the occurrence and severity of deformational plagiocephaly in twins. These findings confirm that the lower in utero infant is at increased risk for the development of plagiocephaly, likely resulting from the more restrictive intrauterine environment encountered during the later part of the pregnancy.
Deformational plagiocephaly refers to a condition in which an infant’s head becomes deformed as the result of external molding forces being applied to the malleable cranium. This distortion often leads to an asymmetric cranium, ear misalignment, and facial asymmetry. In recent years, a growing body of evidence1–5 has documented an increase in the number of infants who present with this condition. Although the recent increase is often attributed to the American Academy of Pediatrics’ recommendation to sleep infants on their back to reduce the risk of sudden infant death syndrome, a restrictive intrauterine environment also plays a very significant role.6–12
One clear example of a restrictive condition is the intrauterine environment of multiple-birth infants, particularly when those infants are carried to full term. In part 1 of this investigation, we documented that infants of multiple-birth origin are several times more likely to develop deformational plagiocephaly than their singleton counterparts.13 In the current study, we explore whether certain prenatal and postnatal risk factors predispose one twin over the other by examining the occurrence and severity of plagiocephaly in both concordant and discordant twin pairs.
Between January and December 1999, we obtained detailed medical histories on 140 sets of twins. Informed consent was obtained from all study participants. The information was obtained as part of a larger survey of 159 sets of multiples (140 sets of twins, 16 sets of triplets, and 3 sets of quadruplets) identified from 9 treatment centers. Each set included at least 1 child who had been treated for deformational plagiocephaly and thus categorized as having a moderate-to-severe deformity. However, infants were classified as “affected” when any positional deformation (occipital flattening, ear misalignment, facial asymmetry) was noted and included mild, untreated cases as well as those who had shown improvement with repositioning and physical therapy. Diagnosis of plagiocephaly and ranking of the severity were made by the child’s physician, and in most cases computed tomographic scans were not required to differentiate deformational from synostotic plagiocephaly.5
Information about the prenatal and postnatal history of each infant was obtained through detailed review of the children’s medical records as well as analysis of a patient database maintained on all infants who receive treatment. Follow-up interviews were performed with all parents to verify the information recorded and to obtain additional information about the child who had not received treatment (if appropriate) or use of reproductive assistance (fertility drugs, intracytoplasmic sperm injection, etc) that had not been previously recorded. The information obtained may be broken down into the following categories:
Demographics: gender, age, race/ethnicity (reported by parents), date of birth, zygosity
Maternal factors: age of mother at delivery, primigravida, abnormal uterine or pelvic structure
Prenatal events: in utero position, in utero orientation (no information was recorded to document whether in utero position/orientation was determined by physical examination and/or routine ultrasound), gestational age, reproductive assistance, dropping early, engagement of fetus in pelvis or under ribs, twin-twin transfusion, excess/paucity of amniotic fluid, drugs to prolong pregnancy
Delivery: mode of delivery, presentation, complications, epidural or spinal administered, use of assistive devices (eg, forceps, suction)
Postnatal events: birth weight, birth length, time in the neonatal intensive care unit, congenital anomalies, sleeping position, neck involvement (torticollis, neck tightness), existence and severity of plagiocephaly, ear infections, developmental delays.
Exposure to potential risk factors (eg, sleeping position, neck involvement, in utero position, in utero orientation, gender, birth weight) were discussed in the telephone survey and verified by careful review of the medical records. Statistical analyses were performed to 1) compare risk factors in discordant twin pairs (only 1 twin affected), 2) assess the effect of risk factors on relative severity in concordant twin pairs (both twins affected), and 3) compare the concordant and discordant twin groups as a whole on several prenatal parameters (maternal age, combined birth weight, weeks of gestation, primigravida). Differences were considered significant at P < .05. Because of our desire to detect risk factors that occur at greater frequencies than expected, our sample of 280 infants will provide power of 80% to detect proportional differences of 20% or greater at the 0.05 level of significance.
The study sample consisted of 140 sets of twins: 46 concordant pairs and 94 discordant pairs (Table 1). Gender was divided nearly equally, with boys being slightly overrepresented at 55.0%: 54 (38.5%) boy/girl pairs, 50 (35.7%) boy/boy pairs, and 36 (25.7%) girl/girl pairs. Mean age of the infants was 9.7 months (standard deviation: 4.4 months; range: 2–24 months). There was a relatively high rate of prematurity (73.6% were <38 weeks’ gestation) and associated low (<2500 g) or very low (<1500 g) birth weights (50.7% and 5.7%, respectively), but these values fall within normal ranges for the multiple-birth population. Mean maternal age at the time of delivery was 31.9 years (standard deviation: 4.9 months; range: 19–48 months).
Demographically, the population was 88.6% white, 2.1% Hispanic, 1.4% black; the remaining children were of mixed racial ancestry. A total of 116 sets of twins (82.8%) were reported to be dizygotic (fraternal), confirmed through placental evaluation or blood work or by being an unlike gender pair. Nineteen pairs (13.6%) were confirmed to be monozygotic (identical) through placental evaluation or blood work; in 5 cases (3.6%), the zygosity remained undetermined. The higher incidence of fraternal twins in this sample population (normal incidence is approximately 66.6%) may be explained by the fact that 48% of the parents in this study reported using some form of reproductive assistance: fertility drugs, in vitro insemination, or intracytoplasmic sperm injection. This increased use of fertility medicine may also explain why the multiples were the mother’s first-born children in 58% of the cases. Five couples reported a selective reduction that reduced the number of fetuses to 2.
Mothers reported few uterine or pelvic abnormalities, with the most common being a posteriorly tilted uterus (17.1%), bicornate uterus (1.4%), uterine tumor (0.7%), and android pelvis (0.7%). Considering the high incidence of tilted uterus in this population, it is important to clarify that this was a determination made by the mother’s physician. At this time, we cannot speculate on the overall significance or nonsignificance of this observation. The overall rate of Cesarean sections was high (63.0%) but, again, typical for the plural-birth population. An additional 16.4% of deliveries involved the use of assistive devices (forceps, suction). Mothers frequently reported that 1 of the infants was carried exceptionally low (30 reports) and that they experienced considerable pain in the hip or radiating into the leg, presumably related to the pressure of the fetus on the pelvic wall. In 17 cases, the mother cited that 1 infant (typically the lower of the 2) never changed position throughout the later part of the pregnancy.
Remarkably, the study participants reported few health problems among infants and mothers. We see the overall excellent health as a positive reflection of the standard of care currently provided for mothers of multiples. More than two thirds (187) of the twins were never required to spend any time in the neonatal intensive care unit, whereas only 35 (22.5%) required more than 2 weeks. Three cases of twin-to-twin transfusion and 9 cases of congenital anomalies were reported, including encephalocele (1), clubfoot (2), achondroplasia (2), Klippel-Feil sequence (1), hydrocephalus (1), cleft palate (1), and extra digit (1). Twelve cases of an excess or paucity of amniotic fluid were documented and ranged from noticeable but not concerning to significant (associated with the achondroplasia). One female fraternal twin infant was reported to have had an in utero stroke. Four cases (2.9%) of grade 2 growth discordance (>25% difference) were identified.
Discordant Twin Pairs
Our initial analyses focused on the 94 pairs of discordant twins. Here we considered whether any of the commonly identified risk factors were associated with the presence of plagiocephaly in one twin as opposed to the other. The significance of potential associations was determined by evaluating the cross-tabulated data using χ2 analysis.
Statistical analysis demonstrated that the lower in utero infant was far more likely to develop plagiocephaly (P < .001). At the same time, the affected infant was significantly more likely 1) to have some form of neck involvement (P < .001), 2) to be in the vertex position (P < .002), and 3) to weigh less at birth (mean difference: 115 g) than his or her unaffected twin (P < .03; Fig 1). Not surprising, because nearly all of the infants in this study were predominately back sleepers, no statistical significance was found for sleep position (P = .835). Similarly, because both infants (affected and unaffected) were of the same gestational age, prematurity could not be isolated as a risk factor. Gender was found not to be associated with plagiocephaly (P = .059), with the fairly equal numbers of boys (n = 54) and girls (n = 40) being somewhat at odds with the traditional 2:1 to 3:1 male-to-female ratio that is commonly reported.2,14,15
Concordant Twin Pairs
Supporting the above findings among discordant twins, analyses of concordant twin pairs demonstrated that the lower in utero infant was significantly more likely to be the more severely affected (P < .001) and significantly more likely to be in the vertex position (P = .006; Figs 2 and 3). Similarly, neither sleeping position (back versus other: P = .819) nor gender (P = .666) was associated with the severity of deformity. Unlike discordant twin pairs, birth weight was not found to be associated with severity. Neck dysfunction was found to occur uniformly and at high incidence in both affected infants.
Concordant Versus Discordant
Statistical analyses of prenatal variables including the maternal age, weeks of gestation, and combined birth weight between the concordant and discordant twin groups revealed no statistically significant difference between these 2 groups. However, concordant twin pairs were found to be statistically more likely to be identical (P = .036). Presumably, a single placenta may lead to a more restrictive intrauterine environment than that experienced by fraternal twins; however, the true extent of how this may influence the development of plagiocephaly remains unclear.
The results of this investigation demonstrate that in utero conditions, mainly position and orientation, seem to increase 1 twin’s risk for the development of plagiocephaly. Specifically, we found that the lower in utero infant seems to be at greater risk in terms of both occurrence and severity. One plausible explanation is that the lower in utero infant is commonly reported to become engaged in the maternal pelvis. This infant frequently bears the weight of the second and in doing so often has less mobility and ability to change orientation, most notably during the last several weeks of pregnancy. This inability to change orientation may adversely predispose the lower in utero infant by allowing the head to become deformed prenatally in the maternal pelvis but also by promoting the development of a congenital muscular torticollis or other neck dysfunction. Before 1992, in utero constraint was considered one of the leading risk factors for the development of plagiocephaly.6–10,16,17
The images presented in Figs 1 to 3 are important in that they present examples of plagiocephaly believed to be of in utero origin. However, it is important to recognize that rarely is the cause of plagiocephaly so easy to differentiate. As appropriately pointed out in review of this article, intrauterine constraint is often a predominate factor that when left unattended leads to significant plagiocephaly postnatally. We could not agree more. The recognition that many cases of plagiocephaly may have an in utero origin emphasizes the need for early diagnosis and intervention. When caught early, during the first several months of life, conservative intervention including repositioning, physical therapy, and supervised tummy time can have a significant impact in preventing the development of plagiocephaly.
Torticollis as a risk factor for plagiocephaly has been well-established in the medical literature (Fig 4).18–21 Although the relationship between these 2 clinical entities is well-understood, considerable debate continues regarding the current incidence of torticollis. The current prevalence of torticollis is difficult to establish because of the large spectrum in severity encountered in the clinical practice, which ranges from a mild, almost imperceptible favoritism to one side, to a severe contracture of the sternocleidomastoid muscle requiring surgical intervention. What is generally agreed on, however, is that torticollis commonly has an in utero cause secondary to intrauterine constraint.8,22 In this investigation, we discovered that when only 1 twin developed plagiocephaly, he or she was also statistically more likely to have some form of neck dysfunction; likewise, when both twins had developed plagiocephaly, neck dysfunction was found to occur uniformly and at high incidence in both infants.
Because the infants in this investigation were all predominantly back sleepers, establishing a supine sleeping orientation as a risk factor was not statistically possible. However, the growing body of anecdotal evidence documenting the relationship between a supine sleeping orientation and the increased incidence of deformational plagiocephaly is hard to ignore.1–5 The recognition that a supine sleeping position may have increased the number of infants who develop deformational plagiocephaly should not be interpreted as criticism of this important recommendation. Since its inception, the “Back-to-Sleep” campaign has saved thousands of lives, and the incidence of sudden infant death syndrome has decreased by more than 40% since 1992.12
A supine orientation is often further exacerbated as parents tend to orient toys or other entertainment to accommodate (thus removing any stimulus for the infant to change his or her position); by the extended use of car seats, infant swings, and carriers; and the current lack of education about the importance of supervised tummy time while the child is awake. A supine sleeping orientation may also allow an otherwise mild torticollis to lead to a significant plagiocephaly as the head is maintained against the mattress in the same orientation each night. Once the head begins to flatten, it will return to the position that provides the most stable orientation, continuing to do so even with subsequent resolution of the torticollis. Today, there is even speculation of an increase in “postural torticollis,” believed to be secondary to a supine sleeping position; however, whether this is a true clinical entity or just previously undiagnosed cases of mild torticollis remains to be shown.
The findings of this investigation confirm that the lower in utero infant of a twin pair is at greater risk for the development of plagiocephaly. Analysis of discordant twin pairs demonstrated that the lower in utero infant was statistically more likely to develop plagiocephaly. Neck dysfunction was found to be a common occurrence in the affected infant. Analysis of concordant twin pairs demonstrated that the lower in utero infant was statistically more likely to be the more severely affected of the two. Neck dysfunction was found to occur uniformly and at high incidence in both infants. Statistical analysis comparing the concordant and discordant twin groups as a whole found no significant difference between the groups, with the exception that concordant twin pairs were statistically more likely to be identical.
We thank all of the parents who took the time to participate in this study.
- ↵Kane A, Mitchell L, Craven K, Marsh J. Observations on a recent increase in plagiocephaly without synostosis. Pediatrics.1996;97 :877– 885
- Littlefield TR, Beals SP, Manwaring KH, et al. Treatment of craniofacial asymmetry with dynamic orthotic cranioplasty. J Craniofac Surg.1998;1 :11– 17
- ↵Huang M, Gruss JS, Clarren SK, et al. The differential diagnosis of posterior plagiocephaly: true lambdoid synostosis versus positional molding. Plast Reconstr Surg.1996;98 :765– 774
- ↵Dunn PM. The Influence of Intrauterine Environment in the Causation of Congenital Postural Deformities With Special Reference to Congenital Dislocation of the Hip [thesis]. Cambridge, England: Cambridge University; 1969
- American Academy of Pediatrics, Task Force on Infant Positioning and SIDS. Positioning and SIDS. Pediatrics.1992;89 :1120– 1126
- ↵American Academy of Pediatrics, Task Force on Infant Sleep Position and Sudden Infant Death Syndrome. Changing concepts of sudden infant death syndrome: implications for infant sleeping environment and sleep position. Pediatrics.2000;105 :650– 656
- ↵Littlefield TR, Kelly KM, Pomatto JK, Beals SP. Multiple birth infants at higher risk for development of deformational plagiocephaly. Pediatrics.1999;3 :565– 569
- ↵Jones PG. Torticollis in Infancy and Childhood. Springfield, IL: Charles C. Thomas; 1969
- Kawamoto HK. Torticollis versus plagiocephaly. In: Marchac D, ed. Craniofacial Surgery: International Society for CranioMaxilloFacial Surgery. New York, NY: Springer-Verlag; 1987:105–109
- ↵Cooperman DR. The differential diagnosis of torticollis in children. In: Karmel-Ross K, ed. Torticollis: Differential Diagnosis, Assessment and Treatment, Surgical Management and Bracing. New York, NY: Haworth Press; 1977:1–12
- ↵Dunn PM. Congenital sternomastoid torticollis: an intrauterine postural deformity. Arch Dis Child.1974;49 :824– 825
- Copyright © 2002 by the American Academy of Pediatrics