Elsevier

Child Abuse & Neglect

Volume 30, Issue 1, January 2006, Pages 7-16
Child Abuse & Neglect

Age-related incidence curve of hospitalized Shaken Baby Syndrome cases: Convergent evidence for crying as a trigger to shaking

https://doi.org/10.1016/j.chiabu.2005.06.009Get rights and content

Abstract

Objective

To determine whether there is an age-specific incidence of hospitalized cases of Shaken Baby Syndrome (SBS) that has similar properties to the previously reported “normal crying curve,” as a form of indirect evidence that crying is an important stimulus for SBS.

Design and setting

The study analyzed cases of Shaken Baby Syndrome by age at hospitalization from hospital discharge data for California hospitals from October 1996 through December 2000.

Patients

All cases of children less than 18 months (78 weeks) of age for whom the diagnostic code for Shaken Baby Syndrome (995.55) in the International Classification of Disease, Ninth Edition, Clinical Modification was assigned.

Results

There were 273 hospitalizations for SBS. Like the “normal crying curve,” the curve of age-specific incidence starts at 2–3 weeks, has a clear peak, and declines to baseline by about 36 weeks of age. In contrast to the normal crying curve that peaks at 5–6 weeks, the peak of SBS hospitalizations occurs at 10–13 weeks.

Conclusions

The age-specific incidence curve of hospitalized SBS cases has a similar starting point and shape to the previously reported normal crying curve but the peak occurs about 4–6 weeks later. Of the likely predisposing causes, this pattern is only consistent with the properties of early crying. There are numerous explanations for the lag in the peaks between crying and SBS hospitalizations, including the possibility of repeat shakings prior to hospitalization. The importance of crying as a stimulus to SBS may provide an opportunity for preventive intervention.

Résumé

French language abstract not available at the time of publication.

Resumen

Spanish language abstract not available at the time of publication.

Introduction

Shaken Baby Syndrome (SBS) is a form of intentional injury to infants and children inflicted by violent shaking with or without concomitant contact with a hard surface, resulting in head trauma including subdural hematomas, diffuse axonal injury, and retinal hemorrhages but also often fractures of the long bones or ribs, with little or no external evidence of trauma. Although usually attributed to John Caffey (Caffey, 1972, Caffey, 1974) the first explicit report of shaking resulting in such lesions was published by the British pediatric neurosurgeon Dr. Norman Guthkelch (Guthkelch, 1971). In one of the case histories in Guthkelch's series and occasionally in Caffey's original reports (Caffey, 1974, Caffey, 1972), crying as a proximal stimulus for the shaking is explicitly mentioned. In subsequent articles and reviews of Shaken Baby Syndrome, crying is often mentioned as a precipitant (Dykes, 1986; Levitt, Smith, & Alexander, 1996; Ludwig, 1984; Reijneveld, van der Wal, Brugman, Sing, & Verloove-Vanhorick, 2004) usually based on anecdotal reports. This is sometimes supported by the observation that the median age of the cases occurs in the first few months of life when crying is greatest. In the first report of incident factors limited to fatal cases garnered from investigative reports in the United States Air Force, Brewster, Nelson, and Hymel (1998) reported that perpetrators mentioned crying as a stimulus in 58% of the cases. The role of crying as a precipitating stimulus for shaking has also been incorporated in policy statements concerning Shaken Baby Syndrome. In the Policy Statement of the American Academy of Pediatrics on “Shaken Baby Syndrome: Rotational Cranial Injuries” (Committee on Child Abuse and Neglect of the American Academy of Pediatrics, 2001), “crying or irritability” is described as “often” the proximal cause of shaking, and pediatricians are encouraged to ask about “response to the crying infant” as part of anticipatory guidance to prevent Shaken Baby Syndrome. Similarly, in the Canadian “Joint Statement on Shaken Baby Syndrome” (2001) co-signed by eight organizations committed to its prevention, infant demands and “especially crying” are cited as triggers for shaking in exhausted or frustrated caregivers.

Despite the reasonableness and acceptance of the assumption that crying is a trigger for shaking, the objective data supporting its role is limited at best. However, in the last 40 years, increasingly careful investigations of infant crying behavior have demonstrated specific and robust properties of crying that contribute to the frustration that caregivers experience in the first few months of life (Barr, Paterson, MacMartin, Lehtonen, & Young, 2005; Barr, St. James-Roberts, & Keefe, 2001). Furthermore, many of these properties have characteristics that provide an opportunity to acquire indirect convergent evidence of the importance of crying as a stimulus for shaking. The most important of these is a robust age-related pattern of crying. This pattern is manifest as increases in the average daily duration of crying in the first few weeks, a peak sometime in the second month of life, and then decreases to more stable levels by about the fifth month.

This pattern was first described as typical of normally developing infants by Brazelton in 1962 who asked parents in his Cambridge, MA, practice to keep daily diaries of distress for the first 12 weeks of life (Brazelton, 1962). It has come to be known as the “normal crying curve.” Since then, a number of investigators have replicated this pattern (Alvarez & St. James-Roberts, 1996; Hunziker & Barr, 1986; St. James-Roberts & Halil, 1991). The replications have indicated how robust the pattern is, and that there has been very little secular change in either the pattern or the amounts of distress behavior since Brazelton's original report. This is illustrated in Figure 1. The median amounts of distress behavior from Brazelton's 1962 study (Brazelton, 1962) is represented along with the mean amounts of distress (crying and fussing) reported in Hunziker and Barr's Montreal study in 1986 (Hunziker and Barr, 1986), and in a more recent Montreal study in 2001 using the same diary (Kramer et al., 2001). The important features of this pattern are that there is not simply “more” crying in the first weeks of life, nor is it just a generalized, linear decrease in crying from birth to older ages. Importantly, it manifests the properties of an “n-shaped” developmental curve, with a phase of increasing crying, a discernible peak or maximum, and a phase of decreasing crying. Although not illustrated by the data sets in Figure 1, this phase is followed by a period of relatively stable, low level crying throughout the remainder of the first year of life at less than half the overall amounts achieved in the first 3 months (St. James-Roberts & Halil, 1991).

In addition to the “peak” pattern, crying has other properties that are also specific to this age period that contribute to caregiver frustration. Most of these are included in clinical descriptions of “colic syndrome,” a syndrome that is now recognized as the upper end of a spectrum of crying phenomenology in otherwise normal infants (Barr, 1989, Barr, 1999, Barr, 2000; Ghosh & Barr, 2000; St. James-Roberts, 2001; St. James-Roberts et al., 1995, St. James-Roberts et al., 1996). These properties include prolonged, unsoothable crying bouts that occur unexpectedly, seemingly unrelated to anything in the environment, during which the infant manifests a facial grimace, increased motor tone, and curling of its legs up over its abdomen that raise concerns about gastrointestinal pain in many caregivers. These bouts do not occur randomly throughout the day, but tend to cluster in the late afternoon and evening hours. Although the prolonged, unsoothable crying bouts comprise less than 10% of the overall crying of infants with colic, they are specific to the first few months of life, and occur rarely thereafter (Barr et al., 2005).

Consequently, if crying is a stimulus behavior for shaking injuries as the anecdotal evidence and common sense suggests, the prediction would be that both the timing and the shape of the age-specific incidence of Shaken Baby Syndrome should be similar. However, a test of this hypothesis requires a different reporting of childhood injuries than is typical in order to test this hypothesis. In most reports, injury estimates are based on conventional age groupings of 1–4 years of age or 1 year age groupings (Agran et al., 2003). These are clearly too broad and too late to capture a developmentally based stimulant condition such as crying behavior. In a recent study designed to capture developmentally related risks of injury specific to young children, Agran et al. utilized E-codes (external cause of injury codes) that would be reflective of age-related developmental characteristics and 3-month age categories. Their study demonstrated that there was marked variability in both rates and specific causes of injury by 3-month age groupings. Although the leading cause of injury is falls for the 0–3 years age group, battering (E-codes 967.0–967.9) had the highest rates specifically between 0 and 5 months, three to four times higher than any rate following 9 months of age.

While the early “battering” cluster in the Agran study included Shaken Baby Syndrome cases, both the codes and the 3-month age grouping are too broad for a specific indirect test of the hypothesis that the properties of early infant crying could be a stimulus for Shaken Baby Syndrome. In our study, we take advantage of a specific code for Shaken Baby Syndrome introduced into the California version of the ICD-9CM and analyze age-specific incidences of hospitalizations of SBS to determine whether SBS incidence reflects the temporal properties of early infant crying behavior.

Section snippets

Methods

We used data from the California hospital discharges from October 1996 through December 2000 to identify the day of age and cause of injury to California children less than 18 months (78 weeks) of age. A new code for Shaken Baby Syndrome was implemented in California on October 1, 1996. We examined data for 3 and 1/4 years because this produces larger and more stable numbers of hospitalizations per age. Hospital discharge data for California hospitals in 1996–2000 were obtained from the Office

Results

There were 273 hospitalizations for Shaken Baby Syndrome between October 1, 1996 and December 31, 2000. The youngest cases (n = 3) are reported for infants of 2–3 weeks of age. After that, the number of cases rises rapidly, reaching a peak at 10–13 weeks of age, and then shows a more gradual decline (Figure 2). At about 36 weeks of age, the number of cases remains around 10 per 4-week period for the remainder of the first year, and then consistently less than 10 after the first year.

To determine

Discussion

Although it is widely assumed that crying is a trigger for Shaken Baby Syndrome, the objective data supporting its role is limited at best. Indeed, in Brewster et al.'s 1998 report on 32 maltreatment deaths in the armed forces, they note that there were no previous studies at that time that examined factors surrounding the final, fatal incident. It is understandable that systematic evidence implicating crying (or any other behavior) as the stimulus is rare for a number of reasons. First, the

Acknowledgements

The authors would like to thank Merry Holliday-Hanson for early contributions to data analysis, and Amy Wicks of the National Center for Shaken Baby Syndrome, Ogden, Utah for literature searches.

References (42)

  • R.G. Barr

    Recasting a clinical enigma: The problem of early infant crying

  • R.G. Barr

    Colic and gas

  • R.G. Barr

    The enigma of infant crying: The emergence of defining dimensions

    Early Development and Parenting

    (1995)
  • R.G. Barr

    Infant cry behaviour and colic: An interpretation in evolutionary perspective

  • R.G. Barr

    Excessive crying

  • R.G. Barr

    Changing our understanding of infant colic

    Archives of Pediatrics and Adolescent Medicine

    (2002)
  • R.G. Barr et al.

    Prolonged and unsoothable crying bouts in infants with and without colic

    Journal of Developmental and Behavioral Pediatrics

    (2005)
  • T.B. Brazelton

    Crying in infancy

    Pediatrics

    (1962)
  • J. Caffey

    On the theory and practice of shaking infants

    American Journal of Diseases of Children

    (1972)
  • J. Caffey

    The whiplash-shaken infant syndrome: Manual shaking by the extremities with whiplash-induced intracranial and intraocular bleedings, linked with residual permanent brain damage and mental retardation

    Pediatrics

    (1974)
  • Cited by (290)

    View all citing articles on Scopus

    This study was supported by the Canada Research Chair in Community Health Care Research to R.G.B. and by cooperative agreement R49/CCR919796 from the U.S. Centers for Disease Control and Prevention, National Center of Injury Prevention and Control.

    View full text