Editorial: Intracranial pressure monitoring for brain injury

Abstract

Zeng and colleagues examine the impact of treatment based on monitored intracranial pressure (ICP) versus that delivered without monitoring on both 6-month neurologic outcome (Glasgow Outcome Scale score) and renal function.5 They report significantly better recovery and less renal dysfunction in monitored patients managed in accordance with the Brain Trauma Foundation guidelines. Unfortunately, these authors do not present data sufficient to support their claim of improved outcome. Neither do they securely establish that ICP monitoring would be associated with less renal failure if this study were repeated elsewhere. Patients were parsed into monitored versus non-monitored groups according to the Brain Trauma Foundation’s recommendations as to whom published evidence suggests should be monitored. These criteria include CT imaging characteristics, age, admission Glasgow Coma Scale [GCS] motor score, and history or presence of hypotension. By definition, then, these are dissimilar groups, and it is the authors’ responsibility to quantify and control the differences in their analysis. Unfortunately, these data were not presented; rather, the groups were deemed demographically similar based on examination of age, total admission GCS score, and sex alone. It remains unproven, and highly unlikely, that these groups are similar enough in terms of outcome-relevant demographic characteristics to allow attributing differences in outcome to ICPmonitor–based management. In our recent randomized controlled trial (RCT),1 when such variables were studied and controlled, we found no difference in outcome between monitored and non-monitored groups. Regarding renal toxicity, the major risk in trauma patients, particularly those receiving hyperosmotic agents, is hypovolemia. Given that Zeng et al. found (as did we in our RCT) that ICP monitoring is associated with significantly decreased use of brain-specific therapies, including mannitol, patients managed without monitoring were at increased risk of negative fluid balance and renal hypoperfusion. In contrast to our findings, this risk was associated with what they called increased acute kidney injury (AKI). Unfortunately, they did not clearly define AKI and their reference to that definition4 does not provide one. The currently accepted definition of AKI2,3 would not support their reported AKI incidences. Furthermore, lack of information as to daily fluid balance or fractional excretion of sodium make it difficult to believe that the group differences in the renal data that they present are not simply due to lack of volume replacement to balance iatrogenic diuresis. In our RCT, the management protocol for all included patients required close attention to maintaining normovolemia, and our incidence of renal dysfunction was globally low and not related to treatment group (3% and 4%, for monitored versus non-monitored patients [p = 1.0], presented in Table 10b in the online supplements to the RCT available at NEJM.org).1 The use of ICP monitoring remains a critical aspect of the care of patients with severe traumatic brain injury despite the body of recent evidence suggesting that our interpretation and management algorithms need refinement based on clinical research. However, the risk of confounding is very high in such studies and requires meticulous prospective study design and data collection, shortcomings in which greatly inhibit the value of this study. (http://thejns.org/doi/abs/10.3171/2013.5.JNS13254)