244

F. Murillo-Cabezas M. A. MunÄoz-SaÂnchez J. M. Dominguez-RoldaÂn R. MartõÂn-Bermudez J. M. Flores-Cordero

Vasogenic oedema and brain trauma Received: 12 October 1998 Accepted: 13 November 1998 Sir: As stated by the authors of Naredi's interesting article [1], mortality and morbidity due to brain trauma is still very high. This explains the investigation of new physiopathological concepts and alternative treatment such as those developed by Asgeirsson et al. [2], aiming at a better outcome for the brain injured. The theoretical statement of this new approach, in disagreement with the more widely accepted concept of cerebral perfusion pressure (CPP) management [3], is mainly based on a reduction of vasogenic oedema through two mechanisms: diminution of the hydrostatic capillary pressure, achieved by a reduction of the mean arterial blood pressure (clonidine and metoprolol) and by precapillary vasoconstriction (dihydroergotamine). We will not discuss the relevance of vasogenic oedema in the initial phase of brain injury where other types of oedema, such as neurotoxic and cytotoxic oedema, may play a greater role, because of the great amount of evidence on the existence of early brain ischaemia [4]. Careful analysis of Naredi's study [1] raises many doubts. The first doubt is about what is considered new in the therapeutic approach proposed by the authors: is it the control of vasogenic oedema through the reduction of the mean arterial blood pressure and the maintenance of the coloidosmotic pressure (since the authors report that the patients were normotense; MAP 85  8 mm Hg), or simply the introduction of a new drug (dihydroergotamine) in the management of intracranial hypertension? Dihydroergotamine, besides producing precapillary vasoconstriction, provokes the reduction of cerebral blood volume, mainly the brain's venous volume, a goal also pursued (reduction of the cerebral blood volume) in the vasoconstrictor cascade of Rosner et al. [3]. Another doubt concerns the lapse of time between the moment of brain trauma and the initiation of infusion of hypotensive drugs, that is, the period of time the patients may have suffered intracraneal hy-

pertension without a reduction in the hydrostatic capillary pressure. On the other hand, the authors do not clarify either why or what type of surgery is carried out on the patients with diffuse lesions, mainly in those patients with type II lesions (23 % of those operated on), according to the classification of Marshall et al. [5]. Furthermore, the population could be biased given the fact that the patients who died in the first 4 days were not included. We question whether the good results reported are due to a reduction in vasogenic oedema or, on the other hand, the consequence of a more aggressive surgical approach and the effect of dihydroergotamine on the cerebral blood volume. Finally, we believe that there is too broad a spread (5±28 months) in the evaluation of patient outcome.

References 1. Naredi S, EdeÂn E, Zäll S, Stephensen H, Rydenhag B (1998) A standardized neurosurgical/neurointensive therapy directed toward vasogenic edema after severe traumatic brain injury: clinical results. Intensive Care Med 24: 446±451 2. Asgeirsson B, Grände PO, Norström CH (1994) A new therapy for post-trauma brain oedema based on haemodynamic principles for brain volume regulation. Intensive Care Med 20: 260±267 3. Rosner MJ, Rosner SD, Johnson AH (1995) Cerebral perfusion pressure: management protocol and clinical results. J Neurosurg 83: 949±962 4. Brown JI, Moulton RJ, Konasiewicz SJ, Baker AJ (1998) Cerebral oxidative metabolism and evoked potential deterioration after severe brain injury: new evidence of early post-traumatic ischemia. Neurosurgery 42: 1057±1063 5. Marshall LF, Marshall SB, Klauber MR, et al. (1991) A new classification of head injury based on computerized tomography. J Neurosurg 75: S 14±S 20

)

F. Murillo-Cabezas ( ) × M. A. MunÄoz-SaÂnchez × J. M. Dominguez-RoldaÂn × R. MartõÂn-Bermudez × J. M. Flores-Cordero Servicio de Cuidados Criticos y Urgencias, Hospital Universitario Virgen del RocõÂo, Avenida Manuel Siurot s/n, E-41013 Sevilla, Spain email: [email protected] Fax: + 3 49 54 24 81 11

S. Naredi E. EdØn S. Zäll H. Stephensen B. Rydenhag

Reply Received: 11 November 1998 Accepted: 11 November 1998 Sir: Our paper presented clinical outcome data after severe traumatic brain injury with a therapy directed toward vasogenic edema [1]. The doubts raised by MurilloCabezas et al. are due to misunderstandings and questioning of the hypothesis. We do not find that Murillo-Cabezas et al. come up with arguments that make the results of our study less important. The patients that died during the first 4 days were not excluded. Of the five patients that died, three died within 48 h. As stated in the paper, all patients with a cerebral perfusion pressure (CPP) > 0 were included. Patients that left the intensive care unit within 5 days were excluded, since their head trauma was not considered to be severe. The range in evaluation period was due to the procedure of having all patients objectively evaluated by independent staff on one occasion. The initial phase after brain injury concerns resuscitation. This does not conflict with the ªLund therapyº, which is introduced when the patient is normovolemic and hemodynamically stable. The ªLund therapyº is based upon a hypothesis aimed at prevention and treatment of the vasogenic edema, and it has support in several experimental studies [2±4]. This is a different approach compared to traditional therapy concerning CPP management. The ªLund therapyº maintains normal colloid absorbing forces and reduces intracapillary pressure in order to counteract transcapillary filtration. The intracapillary pressure is reduced by normalization of the systemic arterial pressure. The ªLund therapyº aims at normalizing mean arterial pressure (MAP) and keeping a CPP above 50±60 mmHg, whereas in ªtraditional therapyº MAP is increased in order to increase CPP above 70 mmHg. It was not our intention to describe in detail the type of surgery performed, since we wanted to stress the intensive care management. But, as stated in the paper, active neurosurgery was mandatory. Significant hematomas and contusions were evacuated to keep CPP and intracranial pressure (ICP) within acceptable ranges. Less than half of the patients received dihydroergotamine

245

(DHE). DHE was introduced as a last step in the therapy when CPP and ICP, despite pharmacological and surgical treatment, were unsatisfactory. DHE is primarily a venous vasoconstrictor, but also induces precapillary vasoconstriction. There is always a risk when using vasoconstrictor drugs in the treatment of severe brain injury, and the need for DHE has decreased markedly in our hands in recent years. We believe that a standardized neurosurgical and neurointensive therapy based on the hypothesis of prevention and treatment of the vasogenic edema improves the outcome after severe traumatic brain injury. The clinical results from this study, and from a recently published study in Critical Care Medicine [5], indicate that the ªLund therapyº is superior to traditional treatments.

References 1. Naredi S, EdØn E, Zäll S, Stephensen H, Rydenhag B (1998) A standardized neurosurgical/neurointensive therapy directed toward vasogenic edema after severe traumatic brain injury: clinical results. Intensive Care Med 24: 446±451 2. Asgeirsson B, Grände P-O (1994) Effects of arterial and venous pressure alterations on transcapillary fluid exchange during raised tissue pressure. Intensive Care Med 20: 567±572 3. Nilsson F, Messeter K, Grände P-O, RosØn I, Ryding E, Nordström C-H (1995) Effects of dihydroergotamine on cerebral circulation during experimental intracranial hypertension. Acta Anaesthesiol Scand 39: 916±921 4. Grände P-O, Asgeirsson B, Nordström C-H (1997) Physiological principles for volume regulation of a tissue enclosed in a rigid shelter with application to the injured brain. J Trauma 45:S23-S31 5. Eker C, Asgeirsson B, Grände P-O, SchalØn W, Nordström C-H (1998) Improved outcome after severe head injury with a new therapy based on principles for brain volume regulation and preserved microcirculation. Crit Care Med 26: 1881±1886

)

S. Naredi ( ) × E. EdØn × S. Zäll Department of Anesthesiology and Intensive Care, Institiute of Surgical Sciences, Sahlgrenska University Hospital, Göteborg, Sweden H. Stephensen × B. Rydenhag Department of Neurosurgery, Institute of Clinical Neuroscience, Sahlgrenska University Hospital, Göteborg, Sweden

Mailing address: Department of Anesthesiology and Intensive Care, University Hospital of Northern Sweden, S-901 85 Umeå, Sweden email: [email protected] Tel. + 46 (90) 7 85 00 00 Fax + 46(90)7 85 38 82

E. Pautas M. Wysocki

The Omega system can underestimate medical activity in the ICU Received: 17 July 1998 Accepted: 27 July 1998 Sir: We read with great interest the paper of Sznajder and coworkers [1] reporting the Omega system as a simple and relevant indicator to estimate the direct cost of intensive care unit (ICU) stays. The authors noted that in some units the Omega system systematically underevaluated the direct costs, probably because of the failure of the Omega system to account for drugs and blood products or very specialized procedures. In a recent prospective study including 200 ICU patient stays [2, 3] we found that the Omega system by itself and in routine practice may also underestimate medical activity. In this study, the Omega score (W) for each patient's stay was calculated by the physician in charge (W ph) as described by Sznajder and coworkers [1] and simultaneously controlled by an independent physician reviewing the patient's whole chart (W contr). In the first part of the study (phase I: 100 patients), physicians in charge were unaware of the study and we found that in 54 % of patient stays at least one item composing the score, in most cases more than one, was forgotten by the physician in charge. Subsequently, the amount of medical activity lost (MAL =

Wph × 100) Wcontr ÿ Wph

was 14 ± 13 %. In the second part of the study (phase II: 100 patients), we tested simple corrective measures (weekly stressing to the medical staff not to forget items, monitoring and weekly information on the MAL). We failed to reduce significantly the number of stays with at least one item lost (45 %, p = 0.41 vs phase I) and again the MAL was 16 ± 16 % (p = 0.19 vs phase I).

Multivariate analysis on the 200 stays found that a longer length of stay (12 ± 17 vs 6 ± 10 days, p = 0.002) and a high daily Omega II (items recorded every time they are performed) (0.8 ± 1.3 vs 0.4 ± 0.7, p = 0.006) favored the forgetting of items. Correlation analysis found that the amount of medical activity lost was negatively correlated with the length of stay (r = Ÿ 0.4, p < 0.001). These two results demonstrate, first, that the longer the length of stay, the higher the probability of forgetting one item; second, that the shorter the length of stay, the more ªcostlyº (by underestimating medical activity) will the forgotten item be. Conversely, we found only marginal overestimation (4 in phase I, none in phase II) even if the present study was not designed to assess specifically this point. Considering the close Omega score-direct costs relationship found by Sznajder and coworkers [1], forgetting items in the Omega score will probably underestimate the direct costs of the ICU stay. If direct costs amounted to 60 % of the overall costs [1], underestimation by almost 15 % of direct costs (as suggested by our results) may have great implications for resource allocation. Because routine collection of items for a score may be far from the experimental setting, quality control for data collection may be a reasonable recommendation before using such a score for resource allocation.

References 1. Sznajder M, Leleu G, Buonamico G, Auvert B, Aegerter P, Merli›re Y, Dutheil M, Guidet B, LeGall JR (1998) Estimation of direct cost and resource allocation in intensive care: correlation with Omega system. Intensive Care Med 24: 582±589 2. Pautas E, Wysocki M, Suzan F, Wolff MA, Millet H, Herman B (1995) Les erreurs de receuil dans le calcul du score Omega: frØquence et consØquences (abstract). Rean Urg 4: 702 3. Pautas E, Wysocki M, Suzan F, Wolff MA, Millet H, Herman B (1996) Frequence et importance des oublis dans le calcul du score Omega: des manoeuvres correctives simples permettent-elles de les rØduire (abstract). Rean Urg 5: 776

)

E. Pautas × M. Wysocki ( ) RØanimation Polyvalente, Institut Mutualiste Montsouris 42 Bd Jourdan, F-75674 Paris Cedex 14, France e-mail: [email protected] Tel. + 33 (1) 40 77 62 62 Fax + 33 (1) 40 78 24 76