Gunther, Wendy M. M.D.; Symes, Steven A. Ph.D.; Berryman, Hugh E. Ph.D.
Public attention to child abuse has increased markedly during the second half of the 20th century and has resulted in educating the public, as well as abusers, regarding the signs of child abuse (1). Abusers, aware that certain specific injuries are more easily recognized as inflicted, occasionally attempt to punish children by using variants of battering such as anteroposterior manual compression that may be less easily traced to abuse. In the cases presented here, the abuser's defense each time included attributing lateral rib fractures and internal injuries to cardiopulmonary resuscitation (CPR).
Cases 1 and 2
In 1997, the dead body of a 7-month-old girl was received at the Regional Forensic Center with the story that the baby stopped breathing after 10 P.M., and the family suspected a gas leak. The father was alone with the children, and he stated to the police that he checked the baby at 7:30 P.M., and "she was okay." The mother corroborated that when she left the apartment at about 9:15 P.M., the baby was in good health. The father stated that about 10 minutes after the mother left for the video store, he heard their baby boy cry, removed him from the crib, and prepared a bottle but did not check on the baby girl. Just prior to the mother's return at 10:15 P.M., he checked the baby girl and found her not breathing. He stated that he took the baby girl out of the crib where she was sleeping with her twin brother and ran down the apartment hallway to the neighbors' apartment, where CPR was performed. CPR was also administered by a police officer who was in the neighborhood on an unrelated domestic call and later in the emergency department where the baby was pronounced dead about 10 minutes after arrival.
Although an initial report stated that 911 was not called, a 911 tape investigation showed that a man's voice made a call from the apartment more than 30 minutes before the child was said to have been found. The caller stated, "I think my baby's dead," but hung up before a full report could be made or advice given.
The twins had been born about 2 weeks prematurely by normal, spontaneous, vaginal delivery. The first 6 months of life showed normal growth and development and were uneventful. The baby girl and her twin brother had experienced the usual childhood minor illnesses and were up to date on vaccinations. Recently, the girl exhibited a slowing of the growth curve and was below the 50th percentile for her age. The mother stated that the baby had been vomiting lately, and the doctor's advice to place her on a different brand of formula was being followed. She also said that the child had a "cold" and was due to see her personal physician again in 3 weeks.
The father corroborated the children's state of health. Interviewing officers noted that when describing the twins, the father referred to the male twin as "my son" but to the female twin only as "her." Not once did the father refer to her by a first name or by using the words "my daughter."
The apartment was examined by Memphis Fire Department Truck 12, and no gas leak was found. The female twin's body underwent autopsy and forensic investigation. The living male twin underwent skeletal survey at the local emergency department.
Autopsy findings showed an array of 12 finger-tip-sized bruises over the chest and abdomen roughly along the child's costal margins. These bruises could easily have been inflicted by an abuser gripping the child's body from behind with the abuser's fingers across her torso and the thumbs at her back (Figs. 1 and 2). This hand position is consistent with a mechanism of death caused by anteroposterior manual compression.
Beneath these bruises, the superior vena cava was ruptured in the anterior mediastinum with a 140-ml hemothorax; total estimated blood volume of the child by weight was about 385 ml. The rupture of the superior vena cava was associated with dissection of the hemorrhage through the mediastinum, between the strap muscles of the neck, around the thyroid gland, and superiorly through the throat structures as far as the inferior border of the left parotid gland.
In the abdomen, there was an irregular laceration of the liver that extended from the anterior to the posterior surface and into the right and left lobes with 60 ml of blood in the peritoneum. This laceration was associated with dissection of the hemorrhage into the mesentery over a 3.75-cm area and multiple, small, fresh hemorrhages in the jejunal wall. There were multiple fresh and healing lateral rib fractures (Fig. 3) but no costochondral junction fractures or fractures of the rib heads lateral to the spine. Some of the fresh fractures had broken through healing callus from old fractures. Significant negative findings included no evidence of sexual abuse and no tears of the frenula of the lips or tongue. Histologic examination showed all the soft tissue injuries to be fresh except for those of the mesentery, which contained fresh clots, organizing clots, granulation tissue, and areas of scarring with numerous hemosiderin-laden macrophages within the mesenteric fat. No evidence of bronchitis was found to explain the upper respiratory symptoms, but there were hemosiderin-laden macrophages in many alveoli. There was also modest hilar nodal hypertrophy with marked secondary germinal center formation. Splenic white pulp, mesenteric nodes, and gut-associated lymphoid tissue also showed moderate hypertrophy.
Forensic anthropologic examination revealed numerous examples of acute and healing trauma. The vault, right radius, all ribs, and the thoracic vertebrae were retained and processed for examination. The ribs on the left showed only one possible healed fracture at the sternal end of rib 4, with possible acute fractures occurring on the sternal ends of left ribs 5, 6, 7, 8, and 9 and definite acute perimortem fractures occurring in the rib body near the sternal ends of ribs 7, 8, and 9. The fracture of left rib 7 was a bending fracture with plates of bone overlapping on the internal surface (indicating compression) and a separating fracture (indicating tension) on the external surface. The ribs on the right showed suspect fractures of the sternal end of 3, 4, and 5, with two types of healing callus. A large healing callus of the rib body near the sternal end of ribs 6, 7, 8, and 9 showed active healing. Ribs 5 and 10 had a smooth but buttressed area of bone modification on the internal surfaces near the sternal end, suggesting old healing fractures that were stabilizing and in the final stages of remodeling. Definite acute fractures had occurred through the callus of old healing fractures in right ribs 6 and 8. Toxicologic examination was negative for drugs, alcohol, or carbon monoxide.
The cause of death was determined to be internal exsanguination due to vascular and visceral ruptures from anteroposterior manual compression. The defendant stated in the initial police interview that the injuries had to have been inflicted by CPR.
The living twin, the male child referred to as "my son" by the father, was examined at a regional children's hospital on the day following his sister's death; he was found to be asymptomatic. Full body x-ray films were interpreted as negative for abuse by the resident on call, and the child was allowed to go home with his parents. On review by the attending radiologist on Monday, lateral rib fractures were detected, and the Department of Child Services immediately placed the child in protective custody.
The body of a 13-month-old girl was brought to the regional forensic center. The mother had left her 3 children for the day in the care of her boyfriend of 2 months. On her return about 6 P.M., she noted that the child appeared to be having trouble breathing. The boyfriend initially attempted to discourage her from calling for help on the basis that the child did not appear seriously ill to him. The police report stated that the mother believed the child might have been given "someone else's medicine." At about 7 P.M., the mother went to check the baby, saw feces on the back of the shirt, changed the infant's shirt, and in the process saw bruises and marks on the body that she thought were cigarette burns. The boyfriend denied any knowledge of their origin, and when the child began vomiting, the mother suggested they call 911, but he demurred. In the face of her insistence, the boyfriend reportedly told her, "If you call the ambulance, we'll both sleep in jail tonight." With continued insistence, he relented and called 911. The child became unresponsive in the interval, and the mother began CPR. The moribund child was transported to the emergency department, where resuscitative efforts were unsuccessful.
Postmortem external examination revealed numerous bruises and abrasions, including a freshly bruised 7.5- × 12.5-cm area over the left side of the forehead, indistinct bruising over the buttocks, and two 0.625-cm bruises over the left forearm. The most significant bruising was over the chest and abdomen. There were multiple 0.625-cm, 1.875 cm, and 2.5-cm bruises and three abrasions on the central chest, two of which were curvilinear and of the appropriate size to have been inflicted by adult fingernails (Fig. 4). Five 0.625-cm bruises were present over the right upper and lower quadrants. There was a minute, fresh tear of the base of the superior labial frenula. The marks interpreted by the mother as cigarette burns were antemortem abrasions. There was no evidence of sexual abuse.
Autopsy revealed rupture of the left lobe of the liver just lateral to the falciparum ligament, lying over a traumatic rupture of the pancreas at the junction of the tail and the body where it overlay the vertebral column. There was associated hemoperitoneum, 130 ml (total estimated blood volume of the child was ∼450 ml), extensive right-sided retroperitoneal and Gerota's fascia hemorrhage, hemorrhage occupying and distending the medulla of the right adrenal gland, dissecting hemorrhage into the diaphragm, fresh jejunal bruising, a tear in the mesocolon, and hemorrhage into the abdominal wall on the right side that dissected laterally between muscle layers. The right rib 7 and the left ribs 5, 6, and 7 showed fresh lateral fractures; there were no healing fractures.
Histology showed all injuries to be fresh. Some peritoneal clots showed more loss of red cell outlines and more white cell karyorrhexis than other injuries, suggesting an age slightly greater than that of the other hemorrhages. There were no other significant findings except for moderate activation of the splenic white pulp. Forensic anthropologic examination revealed numerous examples of acute trauma only. Right ribs 7, 8, and 9 and left ribs 5, 6, and 7 were retained and processed for examination. Three of the fractures were to the left ribs and one to right rib 7, and angular changes in the other left ribs suggested lesser trauma. The fractures were located in the rib bodies of left ribs 5, 6, and 7 and right rib 7 near the sternal end but not in the sternocostal junction. Examination of the rib fractures showed that all four ribs exhibited fractures to the internal surface that were produced by compressive forces (Fig. 5). The force was directed anteroposteriorly, causing the sternal end of the ribs to be displaced internally. The elastic component of the bone allowed the external rib surface to remain intact, whereas the internal surface showed a perpendicularly aligned crushing or fracturing of the bone. Left ribs 5, 6, and 7 showed a slight angular change in the bone near the sternal end. Toxicologic examination was negative for drugs or alcohol and included no trace of any medication.
The cause of death was determined to be internal exsanguination due to vascular and visceral ruptures from anteroposterior manual compression. The boyfriend initially claimed that the injuries were due to CPR. Further investigation culminated in the boyfriend's spontaneous confession that he had squeezed or compressed the infant when irritated by her crying.
In each of these cases, a defense was initially put forward that the injuries were not inflicted as a result of abuse but were incidental to CPR by untrained persons. It is true that each child, except the surviving boy, was subjected to CPR by untrained persons (the abuser, the mother, a police officer, unrelated persons in an adjoining apartment) followed by CPR from trained personnel (ambulance attendants, emergency department nurses and physicians). Although posterior rib fractures in infants are considered characteristic of abuse (2,3), lateral rib fractures have not been as widely recognized as a typical concomitant of abuse and are more readily incurred during CPR, at least in adults (4).
In a series of 211 children examined by Bush et al. (4), rib fractures in one child occurred during prolonged CPR by inexpert as well as expert practitioners (e.g., babysitter, two prehospital care providers, and emergency department personnel), and these occurred at the sternochondral junction and not lateral. Walker et al. (2) state that compression of the chest of a victim by an adult fractures ribs near their vertebral end, which is opposite to what we found. Rib fractures often associated with anteroposterior compression in 31 fatally abused children reported by Kleinman et al. (3) were most frequently located at the costovertebral junction; however, when they occurred laterally or anteriorly, they were most often seen along the inner cortex of the rib, which correlates with our findings. Interestingly, their report of assailants' descriptions of how anteroposterior compression was inflicted-with the fingers at the child's back and the palms at the child's sides-are nearly the reverse of the mechanism suggested by the 12 over-lapping fingertip-like bruises in our case 1. Thomsen et al. suggest that histology is required to detect fractures in young children (5), but we believe more information about stress and direction of fractures is obtained through preserving the bony specimen intact for analysis by a qualified forensic anthropologist.
A number of signs suggest that the children in our study were killed or injured by anteroposterior manual compression rather than by cardiopulmonary resuscitation. Injuries in our cases, although primarily lateral and often near the sternochondral end of the bone, showed healing in two of the children. The living child in particular had never been previously resuscitated and so could not have sustained these injuries during CPR. In addition to the rib fractures, the mesenteric scarring and the hemosiderin-laden macrophages in both the alveoli and the mesentery in case 1 provide evidence that a similar mechanism had probably been used to inflict nonfatal damage on multiple previous occasions. The child's recent slowing in growth rate, apparent "colds," and episodes of vomiting, for which medical care had been obtained, may feasibly be attributed to incidents of manual compression. This finding suggests that rib fractures and mesenteric and pulmonary hemorrhage were missed on examination by qualified medical personnel, as routinely occurs in child abuse victims (6).
In the absence of old or healing injuries, such as in the child in the third case, the following criteria make it likely that the injuries in these cases resulted from inflicted anteroposterior compression of increasing intensity:
* placement of the assailant's hands as suggested by fingertip abrasions and bruising
* periods of temporary survival without treatment after initial symptoms and before CPR was initiated
* compressive nature of the rib fractures, with vertical crushing of the internal surface of the rib and a lack of fracture of the external surface due to the elastic component of the bone
* rupturing of incompressible vessels and organs, with dissection of hemorrhage to distances far from the rupture site
* toxicologic explanations offered by abusers that were unsubstantiated by postmortem analysis
* severity of injury inconsistent with mechanism proposed, even forcefully and incorrectly applied CPR
In particular, the need for building force to cause ruptures of solid organs and vessels is at variance with the mechanism of compression that occurs even in inexpertly performed CPR. To cause ruptures, an episode of compression must occur that continues to increase in force until the child's internal structures cannot bear the load and fail. Even in inexpertly performed CPR, pressure is applied and released without continuing to build to a crescendo that causes the failure of vascular walls and the rupture of solid organs across bony prominences. Fractures across the inner cortical surfaces of the highly elastic ribs of infants, as seen in these cases, are more consistent with the buckling produced by a continuously applied, steadily increasing force than by multiple single applications of lesser force.
Both fatal and nonfatal injury to infants can result from a mechanism of abuse involving manual anteroposterior compression or squeezing. This abusive maneuver may be repeated multiple times before it is severe enough to cause death, and its severity may escalate if undetected. Radiologists should be aware of the possibility of lateral rib fractures in such cases, and the medical examiner should be alert to the significance at external examination of multiple fingertip-sized bruises along the anterior or anterolateral chest and torso. These lateral rib fractures are difficult to detect radiographically and grossly at autopsy but can be easily observed as a dry bone specimen. Examination of the injured bones by a forensic anthropologist will not only identify fractures, but may provide information on bending direction and application of lateral stress which will clarify the method of inducing fracture. The presence of lateral rib fractures in infants in the absence of posterior rib fractures does not signify that the death was natural and is not diagnostic for injuries sustained during CPR.
© 2000 Lippincott Williams & Wilkins, Inc.