Clavicle fractures are the most common bony injuries of the newborn, occurring particularly during the vaginal delivery of large neonates, breech-presenting neonates, or when shoulder dystocia occurs.1 The incidence of clavicle fractures in newborns has been reported to range between 2 and 3 per 1,000 live births.2,3 Birth-related brachial plexus palsy is also associated with the same conditions with an incidence of 0.4 to 4 per 1,000 live births.4
We hypothesize that in cases of shoulder dystocia, a fracture of the clavicle allows mediolateral compression of the shoulder girdle and, thus, easier delivery. A clavicle fracture in this setting should lessen the severity of stretch of the plexus and should be a positive prognostic sign for neurologic recovery of birth-related brachial plexus palsy. To examine this hypothesis, a true incidence of birth-related brachial plexus palsy, clavicle fracture, and the combination of the two must be established. In addition, longitudinal tracking of these neonates, including those who have early spontaneous recovery, is needed. Our clinical setting is uniquely positioned to shed light on these issues.
MATERIALS AND METHODS
Birth records in Dallas County were obtained from the Texas Department of Health Statistics for the period of January 1, 1988, through August 31, 2012. Discharge records from the same period from Parkland Memorial Hospital were queried for the number of live births, the number of birth-related brachial plexus palsies, clavicle fractures, and the combination of the two as documented in the newborn nursery. Diagnosis was made clinically by either an experienced pediatric nurse practitioner or attending faculty at Parkland Memorial Hospital. Transient plexus palsies that resolved before discharge from the newborn unit were not included. Clavicle fracture was defined by clinical manifestations (swelling, tenderness, crepitus over the clavicle) with or without a radiograph. Radiographic confirmation was not a required inclusion criterion. The data were collected by an experienced registered nurse through a quality improvement database, in which all charts were reviewed before chart closure. After instituting electronic medical records in 2009, all newborn nursery charts were reviewed by the same experienced registered nurses.
Medical records for all patients with birth-related brachial plexus palsy, and brachial plexus palsy with an ipsilateral clavicle fracture, seen during the same period in the brachial plexus palsy or hand clinic at Texas Scottish Rite Hospital for Children were reviewed. The study was approved through the University of Texas Southwestern Medical Center institutional review board.
Parkland Memorial Hospital is the taxpayer-supported safety-net county hospital that provides a disproportionately large share of indigent care and includes one of the largest obstetric services in the country. Approximately 40% of all births in Dallas County occur at Parkland Memorial Hospital. Texas Scottish Rite Hospital, founded in 1921, is a Masonic hospital that provides orthopedic and related neurologic care for the children of Texas. In the early 1980s, a specialty clinic for infants and children with brachial plexus palsy was established at Texas Scottish Rite Hospital in conjunction with involved faculty from the University of Texas Southwestern and Parkland Memorial Hospital. When a neonate is born with a brachial plexus palsy, with or without a clavicle fracture, at Parkland Memorial Hospital, written instructions are given to the mother to protect the involved limb until it is comfortable and to begin gentle range-of-motion exercises as instructed by the newborn nursery staff. An appointment to the brachial plexus clinic at Texas Scottish Rite Hospital is made for the infant at 2 months of age unless there were special circumstances such as severe or associated injuries (in which case the infant would be seen earlier). If the infant had recovered completely by the 2-month well-baby pediatric visit, the appointment was cancelled or not kept. Those infants with residual upper limb dysfunction at approximately 2 months of age are seen in our brachial plexus palsy clinic. Monthly visits are scheduled for the first 6 months until recovery has plateaued or is complete or a treatment intervention planned.
These referral patterns were developed within the Parkland Health and Hospital System, and most neonates born at Parkland Memorial Hospital are seen at one of the 12 Parkland community pediatric clinics or by physicians who are knowledgeable about this system. Because this model for maternal–child care is geographically determined by the boundaries of Dallas County, and because Texas Scottish Rite Hospital is the only referral center treating brachial plexus palsy within the County, we believe that we can extrapolate from our data the percentage of infants who did and did not recover even if the infant was not seen in the brachial plexus palsy clinic. In addition, infants not seen in the newborn period are consistently referred to Texas Scottish Rite Hospital later if there are concerns regarding limb function.
The second part of our study looked at the recovery rates for the infants with brachial plexus palsy with and without clavicle fracture. Infants referred to Texas Scottish Rite Hospital for these diagnoses included the neonates born at Parkland Memorial Hospital and also neonates born elsewhere within Dallas County or in the state of Texas. The total number of neonates born at other hospitals in Dallas County is also known for this period. Referrals to Texas Scottish Rite Hospital for the diagnoses of brachial plexus palsy, clavicle fracture, or both were tracked by patient zip codes and referring pediatrician data. We do not know the incidence of brachial plexus palsy and clavicle fractures for each of the other 13 hospitals providing maternity services in the county. The total number of Dallas County and non-Parkland Memorial Hospital births and the number of infants seen from this group at Texas Scottish Rite Hospital serves as a check on our index Parkland Memorial Hospital population.
At Texas Scottish Rite Hospital, all patients were seen and evaluated by one of three fellowship-trained hand surgeons with extensive experience in treating pediatric upper limb conditions, specifically including brachial plexus palsy. At each clinic visit, a licensed therapist (occupational or physical) also assessed every patient. After an initial evaluation, patients were followed on a monthly basis if continued treatment and monitoring was felt to be necessary by the treating physician.
The Texas Scottish Rite Hospital medical records were reviewed for information about sex, birth weight, mode of delivery, presentation at birth, and length of follow-up. Any additional neurologic and musculoskeletal trauma was noted. Specific details from the first and final physical examination, assessing the neurologic status of the involved upper limb, including initial Narakas (severity classification) level (Table 1),5 were collected. This included age of resolution of symptoms or, if neurologic deficits did not resolve, an assessment of residual deficit and need for and type of treatment.6,7 Risk of permanent residual neurologic deficit was calculated for those neonates born at Parkland Memorial Hospital and correlated with that for the children born in Dallas County but at other facilities.
Between January 1, 1988, and August 31, 2012, 913,105 neonates were born in Dallas County, Texas, of which 366,408 neonates were born at Parkland Memorial Hospital. Of the neonates born at Parkland Memorial Hospital, there were a total of 1,291 brachial plexus palsies (3.5/1,000 births). Of these, 320 (25%) had a concomitant ipsilateral clavicle fracture and 971 had no clavicle fracture (75%). During this interval, 3,739 neonates sustained clavicle fractures (10.2/1,000 births), but 3,419 (91%) of them had no brachial plexus palsy (Fig. 1).
During the study period, 1,383 new patients with isolated brachial plexus palsy were seen in the Texas Scottish Rite Hospital brachial plexus palsy clinic, including those born at Parkland Memorial Hospital and others born at other locations in Dallas County or from the statewide referral area.
Of the 971 neonates born at Parkland Memorial Hospital who had a brachial plexus palsy but did not have a clavicle fracture, 214 (22%) came for the appointed visit at the Texas Scottish Rite Hospital brachial plexus palsy clinic. The remaining 757 neonates who were not seen in follow-up were assumed to have fully resolved. Of the 214 neonates who were seen, there were 58 infants (27%) with a residual deficit at 6 months of age. Assuming resolution of those who were not seen in clinic, residual deficit occurred in 58 of 971 (6%) overall. Of the remaining 156 infants, all but nine resolved by 3 months; the remaining nine resolved at 6 months. Initial Narakas severity score was available for 51 of the 58 infants; a lower score indicates a less severe lesion and therefore a better prognosis. Of these, 41 (80%) were N-I, two (0.4%) were N-II, four (0.8%) were N-III, and four (8%) were N-IV.
Twenty-one of 320 (6%) neonates born at Parkland Memorial Hospital with both brachial plexus palsy and ipsilateral clavicle fractures presented to Texas Scottish Rite Hospital for the 2-month appointment or at any time after that for persistent concerns about the arm. Based on the assumptions described, there was a 94% recovery before 2 months of age. During this time, 27 infants were seen with a diagnosis of brachial plexus palsy and ipsilateral clavicle fracture that were born in Dallas County but not at Parkland Memorial Hospital. An additional 26 patients were referred in from outside the county. Therefore, a total of 74 infants presented at an average age of 2.4 months (range 2–10 months) and were followed for this combined diagnosis. This larger group comprises the cohort we examined to better understand the outcome (Fig. 2).
Fifty-five percent were boys (Table 2). The average birth weight for the entire study group was 8.69 pounds (3,950 g; range 6 pounds [2,727 g] to 11 pounds [5,000 g]). The mean birth weight for the neonates in the complete recovery cohort was 8.5 pounds (3,864 g;) range 6 pounds [2,727 g] to 10.5 pounds [4,773 g]). The mean birth weight for the neonates with residual deficits was 9.25 pounds (4,196 g; range 7.75 pounds [3,523 g] to 11 pounds [5,000 g]). The difference seen in the birth weights of the neonates who recovered compared with those who had persistent deficits was statistically significant (P=.005; Table 3).
In the entire group, all clavicle fractures healed uneventfully; no follow-up imaging was completed. Initial Narakas severity score was available for 66 of the 74 infants. Of these, 58 (88%) were N-I, two (3%) were N-II, five (7%) were N-III, and one (2%) was N-IV. Fifty-five (74%) infants had complete resolution with no neurologic or functional deficit. This group included the three neonates born in breech presentation. Of the 19 infants (26%) who did not attain full recovery, six have undergone surgical procedures to improve shoulder function.
Fifteen of the 21 neonates from Parkland Memorial Hospital with combined brachial plexus palsy and clavicle fracture recovered early and completely. Six infants had permanent neurologic or functional deficits. Therefore, for those born at Parkland with a brachial plexus injury and clavicular fracture, the estimated full recovery rate is 314 of 320 (98%).
Our data indicate a rate of spontaneous resolution of an isolated birth-related brachial plexus palsy of 94% when no clavicle fracture occurs and 98% when there is a combined injury (P=.005). A χ2 test shows that there is a statistically significant difference between the two groups. This resolution rate is much higher than what has been previously reported in the literature.
Foad et al8 reviewed published recovery rates in brachial plexus palsy based on the severity of the injury and concluded that there is a 35% likelihood of persistent dysfunction. None of the studies reviewed in this report included background population statistics or numbers of infants who recovered completely and early.
Concurrent ipsilateral clavicle fractures in neonates born with brachial plexus palsy have been reported.9–12 al-Qattan and colleagues10 reported no prognostic value of a clavicle fracture on an ipsilateral brachial plexus palsy in 13 patients. The patients in this study were referred for surgical care and those who did not have surgery were considered to have “satisfactory spontaneous recovery.” There were no details on brachial plexus palsy severity, surgical indications, or overall incidence or type of injury.
The mechanism for the improved outcomes of brachial plexus palsy with an ipsilateral clavicle fracture must be considered. During birth, the clavicle fails in response to lateral compression. In the early 1900s, Sever12 reproduced brachial plexus traction injuries in infant cadavers. When he resected the clavicle, he found that less traction was required to disrupt the plexus but he did not assess how this changed the cross-sectional diameter of the shoulders in compression. We postulate that a clavicle fracture that occurs during vaginal delivery allows the shoulder girdle to compress and facilitates delivery. This concept is supported by our data showing that infants with concurrent clavicle fractures have lower severity brachial plexus palsy injuries and a higher rate of complete resolution than the general brachial plexus population.8 In addition, infants who did not recover completely were statistically more likely to be those with larger birth weights.
In this study we have defined our patient population and have followed a group of these infants from a very early age. We conclude, in contrast to al-Qattan et al,10 that a concurrent clavicle fracture is a predictor of more complete recovery. In the patients with brachial plexus palsy and concomitant clavicle fracture for whom we had initial Narakas levels, 9% were found with the more severe III and IV lesions in comparison to 15% in a reported isolated brachial plexus palsy cohort.8
The primary limitations of this study relate to the degree to which the patient population reported accurately represents the entire population of neonates in Dallas County who sustained a brachial plexus injury at birth with or without a clavicular fracture during the time studied. Discharge reports from the newborn unit at Parkland Memorial Hospital likely have a high degree of accuracy. There is no incentive to overreport birth complications, and the reported incidence of brachial plexus palsy of 3.5 per 1,000 births is within the range that has been consistently reported in the obstetric literature. Although complete resolution is most likely the case for most of the patients, there may be an unknown subset of patients who have residual symptoms. Late referral to our institution of children born in Dallas County who have residual problems related to brachial plexus palsy is extremely rare.
Another limitation concerns those neonates who had asymmetric movement of the upper limbs that resolved within the first 24 hours. If this represents a mild neurapraxic brachial plexus stretch, and is not related to a clavicle fracture, the hypothesis of a protective effect of a clavicle fracture is less robust. The number of times that this actually happened in this group of neonates is unknown, and it has been of no clinical consequence.
Comparison of the Parkland Memorial Hospital and non-Parkland Memorial Hospital numbers gives strikingly similar risks for persistent long-term neurologic deficits. Twenty-one of the total number of neonates born at Parkland Memorial Hospital (366,408 [0.006%]) had both brachial plexus palsy and a clavicle fracture and presented with persistent concerns at 2 months of age or later. This percentage is almost identical to the 27 neonates born in Dallas County, but not at Parkland Memorial Hospital (546,697) at 0.005% (P=.609). Because these numbers are so similar, we doubt that we are missing infants who have residual limb dysfunction, and we believe that we are statistically identifying that subset of infants who completely recover early and are never seen by our specialty hospital.
The biggest limitation is that there is no prospective, complete follow-up of all infants. It is also not clear that all the infants were evaluated by the pediatrician at the 2-month well-baby check. Based on our referral patterns, we believe it is a reasonable assumption that all infants with a persistent brachial plexus injury at 2 months of age would be seen in the brachial plexus palsy clinic. The finding that the relative number of referrals for this diagnosis from the non-Parkland Memorial Hospital maternity and pediatric facilities was identical lends credence to the conclusions here.
The strengths of this study are the large numbers of infants, the background population statistics against which we can calculate incidence, and the consistent systems-based approach to brachial plexus palsy over the past three decades. Because Parkland Memorial Hospital is our publicly funded county hospital, and Texas Scottish Rite Hospital is a Masonic institution, no financial burden was incurred by families for an infant to be seen.
In conclusion, the likelihood of complete recovery from a brachial plexus palsy is higher than has been published. More importantly, an ipsilateral clavicle fracture appears to have a protective effect on the brachial plexus and is a positive prognostic sign for both severity of the plexus lesion and neurologic recovery. Our data are reassuring in the cases in which the accoucheur is considering fracturing the clavicle, because this maneuver does not worsen outcomes should a brachial plexus palsy occurs. Lastly, the parents of an infant with a clavicle fracture should be reassured that the clavicle will heal and the fracture itself likely prevented a more serious injury.
1. Curran JS. Birth associated injury. Clin Perinatol 1981;8:111–29.
2. Levine MG, Holroyde J, Woods JR, Siddiqi TA, Scott M, Miodovnik M. Birth trauma: incidence and predisposing factors. Obstet Gynecol 1984;63:792–5.
3. Oppenheim WL, Davis S, Growdon WA, Dorey FJ, Davlin LB. Clavicle fractures in the newborn. Clin Orthop Relat Res 1990;250:176–80.
4. Hale HB, Bae DS, Waters PM. Current concepts in the management of brachial plexus birth palsy. J Hand Surg Am 2010;35:322–31.
5. Narakas AO. Birth-relatedal brachial plexus injuries. In: Lamb DW, editor. The Paralyzed Hand. Edinburgh (UK): Churchill Livingstone; 1987. p. 116–35.
6. Mallet J. Obstetrical paralysis of the brachial plexus. II. Therapeutics. Treatment of sequelae. Priority for the treatment of the shoulder. Method for the expression of results [in French]. Rev Chir Orthop Reparatrice Appar Mot 1972;58(suppl 1):166–8.
7. Curtis C, Stephens D, Clarke HM, Andrews D. The active movement scale: an evaluation tool for infants with birth-relatedal brachial plexus palsy. J Hand Surg Am 2002;27:470–8.
8. Foad SL, Mehlman CT, Foad MB, Lippert WC. Prognosis following neonatal brachial plexus palsy: an evidence-based review. J Child Orthop 2009;3:459–63.
9. Lam MH, Wong GY, Lao TT. Reapprasial of neonatal clavicle fracture: relationship between infant size and neonatal morbidity. Obstet Gynecol 2002;100:115–9.
10. al-Qattan MM, Clarke HM, Curtis CG. The prognostic value of concurrent clavicle fractures in newborns with birth-related brachial plexus palsy. J Hand Surg Br 1994;19:729–30.
11. Greenwald AG, Schute PC, Shiveley JL. Brachial plexus birth palsy: a 10-year report on the incidence and prognosis. J Pediatr Orthop 1984;4:689–92.
12. Sever JW. Birth-related paralysis: its etiology, pathology, clinical aspects and treatment, with a report of four hundred and seventy cases. Am J Dis Child 1916;12:541–78.