workers (1990), and Corrigan et al. (1995) reviewed the literature on TBI and recurrent injury and showed a strong association with alcohol abuse. Closely related to repeat TBI is what has been called the “second-impact syndrome,” in which a repeat mild TBI was catastrophic or even fatal (Kelly et al., 1991).

Gronwall and Wrightson (1975) concluded that the effects of concussion might be cumulative especially in sports, in which populations may be easily monitored. Recurrent head injury in sports has been the subject of several case reports and case-series studies (e.g., Kelly et al., 1991; Cantu and Voy, 1995). Their findings of risks posed by recurrent TBI have prompted recommendations on when players can return to games in the event of even a minor concussion (CDC, 1997).


One outcome of TBI is death. Whereas mortality is an ideal measure of the magnitude of severity of TBI in the general population, the CFR after hospital admission is a measure of the immediate gross consequences of brain injury. The CFR has been used for decades as an indicator of hospital quality of care, but its use is subject to biases as described below.

Case-Fatality Rates

CFR data are available from 15 US population-based incidence studies (Table 3.13). They range from 4.4/100 hospitalized patients in Maryland (MacKenzie et al., 1989a) to about 25/100 in the Bronx, New York (Cooper et al., 1983), and 23/100 in Oklahoma (Oklahoma State Department of Health, 1991). The range in rates may reflect gross differences in hospital patient-admission practices. That is, hospitals that admit a high proportion of patients with severe brain injury would be expected to have higher CFRs than hospitals that admit a large proportion of patients with mild brain injury, who are less likely to die. CFRs in the most recent reports in the United States show the effect of changes in hospital admission practices of the last decade: fewer of the mildly head-injured persons were admitted.

Table 3.14 summarizes CFR data from outside the United States. The rates in the 15 studies range from 0.8/100 hospitalized patients in a report from South Australia (Badcock, 1988) to 30/100 in a county in Denmark (Engberg and Teasdale, 2001); the latter CFR represented only hospital-admitted patients with ICD-9-CM codes 850–854. The very low rate in South Australia may reflect the fact that over 90% of the patient cohort admitted to the hospitals in the study region had mild TBI. The CFR in severe-TBI patients in the study was 55%, which is comparable with rates in other studies that focused on severe-TBI patients. Discounting the single high CFR from Denmark, all remaining rates are less than 10/100 admitted patients.

Occasionally, a total or general CFR appears in the literature (e.g., Servadei et al., 2002a). Such a rate would reflect both in-hospital and prehospital deaths and express the risk of death from the moment of injury to hospital discharge. It is often 2 or 3 times the in-hospital CFR. Examples are found in Kraus et al. (1984), Vazquez-Barquero et al. (1992), and Tiret et al. (1990).

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement