AUTHOR(S): Bogdahn, U., M.D.; Lau, W., M.D.; Hassel, W., Grad. Eng.; Gunreben, G., M.D.; Mertens, H. G., M.D.; Brawanski, A., M.D. Department of Neurology (UB, WL, WH, GG, HGM) and Department of Neurosurgery (AB), University of Würzburg, Würzburg, and Department of Neurosurgery (AB), University of Regensburg, Regensburg, Germany Neurosurgery 31; 898-904, 1992 ABSTRACT: EXPERIENCE WITH A continuouspressure controlled, external ventricular drainage system (EVD) in 100 patients (n = 49 female, n = 51 male; mean age, 56.3 yr) with acute hydrocephalus is reported. Cerebrospinal fluid circulation disturbances resulted from hemorrhages caused by subarachnoid hemorrhage (n = 45), parenchymal hemorrhages from angioma (n = 4), anticoagulants (n = 7), or hypertension or other reasons (n = 30); in addition, hydrocephalus developed from infections (n = 3), tumors (n = 2), infratentorial infarction (n = 5), or unknown reasons (n = 4); 52 patients had ventricular hemorrhages. No patient died of system-associated morbidity. Mean time of EVD treatment was 9.5 days, with 40 patients being treated for 10 to 29 days; routine refobacin (5 mg) flushing of the system was performed three times a day. Patients without cerebrospinal fluid leakage had a 2% rate of secondary infection compared with 13% in patients with cerebrospinal fluid leakage due to ventricular catheter placement (P < 0.05; overall infection rate, 5%). A clinical mortality rate of 29% during EVD treatment was observed in subarachnoid hemorrhage patients (Hunt and Hess Grades II, III, IV, and V; n = 9, 9, 18, and 9, respectively); recurrent hemorrhages during EVD treatment occurred in 19 patients (26 hemorrhages), and of these, 10 patients died. System occlusion was seen in 19 cases (12 of 45 patients with subarachnoid hemorrhage), requiring catheter and system renewal in 1 case; system extraction was seen in 3 cases, misplacement was seen in 11 cases, and disconnection was seen in 5 cases. We conclude that EVD treatment with a modern aseptic system and continuous pressure monitoring has no additional mortality, as well as an acceptable rate of secondary infections and technical complications; the poor prognosis patient selection may be responsible for a relatively high frequency of recurrent hemorrhages. Continuous pressure monitoring allows regular monitoring of intracranial pressure, early recognition of system occlusion, and recurrent hemorrhage. KEY WORDS: Aneurysms; Antibiotic therapy; Central nervous system infection; Central nervous system tuberculosis; Hydrocephalus;
Refobacin; Shunt infections; Subarachnoid hemorrhage; Ventricular drainage External ventricular drainage (EVD), as described earlier by Pampus (18,19), Lundberg (15), and Merrem (17) is frequently required in the treatment of adults with acute cerebrospinal fluid (CSF) circulation disturbances: most patients have intracranial or subarachnoid hemorrhages (SAH); some patients have acute space-occupying lesions requiring preoperative CSF drainage. EVD is widely employed, and data on systems and their complication rates are frequently reported for pediatric patients (1,5-7,10,20,22, 25,28) . Pressure recording and concomitant CSF drainage, as described originally (17), in children (5) and in adults (2,16,23,24), have been reported more recently. Continuous drainage of CSF may result, however, in a number of additional risks, especially subsequent bleeding. Continuous-pressure monitoring, prolonged handling, and regular screening of CSF probes may be associated with a chance of system contamination. We report our experience with an EVD system allowing continuous pressure monitoring; results of safety and complications are presented. PATIENTS AND METHODS Patients Data of 100 intensive care unit patients (n = 49 female, mean age, 56.4 yr; n = 51 male, mean age, 56.1 yr) with ventricular catheter treatment were analyzed. Treatment was indicated in all cases by acute CSF circulation disturbances (acute hydrocephalus) caused by the following diagnoses: SAH from aneurysm (n = 17) or unknown source (n = 28) classified according to Hunt and Hess (12), parenchyma hemorrhages (n = 30), angioma (n = 4), anticoagulant hemorrhages (n = 7), and rare cases of infection (n = 3), tumor (n = 2), ischemia (posterior fossa infarctions with hydrocephalus; n = 5), or unknown reasons (n = 4). Among these, 52 patients had ventricular hemorrhages. All patients received heparin at approximately 15,000 IU/day subcutaneously or intravenously. The EVD system was removed whenever the intracranial pressure values read up to a maximum of 20 torrs for 2 consecutive days. Ventricular drainage system A ventricular drainage system (B. Braun Melsungen A.G., Melsungen, Germany; Figure 1) has been used in our department with the following characteristics: it has a rigid CSF collector with quantitative volume measurement and a CSF collecting bag with automatic overflow, as well as a stable holding device keeping the manifold plate including CSF pressure transducer away from the patient (for hygiene reasons, disconnection, handling). A Codman ventricular Silicon catheter (Codman & Shurtleff, Inc., Randolph, MA) is placed into the lateral ventricle by a standard frontal approach in the nondominant hemisphere and is connected to the drainage system (150-cm-long pressure-resistant tubing) at a three-way stopcock
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 November 1992, Volume 31, Number 5 898 Continuous-Pressure Controlled, External Ventricular Drainage for Treatment of Acute Hydrocephalus--Evaluation of Risk Factors Clinical Study
Data analysis Patients' data have been evaluated for systemic signs of infection (fever, leukocytosis >10,000/cbm, creactive protein, ESR, microbiology) and CSF infections (pleocytosis, microbiology, clinical signs). Catheter misplacement, dislocation, disconnection, and occlusion were analyzed, as was secondary neurosurgical treatment. In addition, clinical parameters, such as mortality and morbidity, and signs of recurrent hemmorhage were documented. Data were accumulated during the time patients were monitored at the neurological intensive care unit. Neuroradiological procedures were performed as clinically indicated and were also retrieved for evaluation. Standard statistical analysis was performed, mainly as a χ2 test. RESULTS General results The mean time of EVD was 9.5 days (range, 1 to 30 days); the complete EVD system was exchanged only twice, either when blood clots irreversibly occluded the tubing or in a case with catheter extraction requiring continuous drainage (Figure 2). Cutaneous wound reactions were registered in two cases, but without morbidity. The overall patient mortality was 35%; 14 patients (40%) died of recurrent hemorrhages; the remainder died of complications of their primary disease, mainly brain edema. Mortality during EVD therapy was 21%. No patient died of EVD-related morbidity, especially EVD-related infections or catheter dislocation. In no single case did technical complications lead to serious patient morbidity. After removal of the EVD system in six cases, a permanent ventriculoperitoneal shunt system was installed because of permanently raised CSF pressure and/or recurring hydrocephalus. Safety In 10 patients (10%), technical complications
occurred without patient morbidity. These involved system disconnections (patients, n = 5; treatment team, n = 2), system leakage of the tubing (n = 1), and patient-induced EVD extraction (n = 3). System occlusions occurred in 19 cases, and in nearly all cases, flushing with saline reopened the catheter. The entire system (including catheter) had to be exchanged in one case of tuberculous meningitis, after it became extracted. In a second case, a recurrent hemorrhage caused an occlusion, which could not be reopened; a new EVD had to be placed. In 11 cases, the EVD catheter was initially misplaced and had to be corrected. In four cases, more than one puncture was necessary; no puncture-associated hemorrhage occurred. Cerebrospinal fluid leakage CSF leakage, as defined by leakage beside the catheter, was observed during EVD treatment in 26 patients (26%) at least once, in 15 patients twice, and in 7 patients three times (Figure 3); leakage was observed from Day 0 to Day 16 postoperatively. Secondary suturing was required in 21 patients. After removal of EVD catheters in 13 cases, CSF leakage was documented once, and in 2 cases, it was documented twice; the longest interval of CSF leakage after removal was 4 days. Secondary suturing after EVD removal was performed in eight cases. In four of five patients with secondary infections, CSF leakage was found (see below). In most cases, CSF leakage was successfully treated by local compression for up to 24 hours. Subarachnoid hemorrhage According to clinical grading (Hunt and Hess 12), we observed 9 patients, each with Grades II, III, and V and 18 patients with Grade IV. Fourteen of 27 patients with unfavorable Hunt and Hess Grades IV and V died in the hospital. System occlusion occurred in 12 (27%) of 45 SAH patients, compared with 7 (15%) of 55 patients with other diagnoses (P < 0.1; χ2 test); however, no patient died from occlusion, because this was recognized early by continuous pressure monitoring (Figure 4a; see Figure 6). In two patients, the external part of the EVD systems had to be reinstalled (without ventricular catheter replacement). Recurrent hemorrhages (Figure 4b) were noted in 19 (42%) of 45 SAH patients while under EVD treatment, compared with 7 (13%) of 55 patients with other diagnoses (P < 0.01; χ2 test). Ten patients with recurrent hemorrhages died from their hemorrhage under EVD treatment (53%). A single recurrent hemorrhage causing an occlusion required an EVD replacement. The entire SAH mortality rate was 29% as compared to 17% for patients with other diagnoses, while patients were under EVD therapy (Figure 4). Ventricular hemorrhage In 52 patients, ventricular hemorrhage was initially diagnosed. Occlusions occurred in 12 cases (23%), and recurrent hemorrhages occurred in 11 cases (21%) (not significantly different compared to the remaining patients). Forty-nine patients were on
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
manifold plate. The manifold plate consists of four stopcocks: a first stopcock allows CSF drainage to the CSF collector, pressure adjustment to normal zero values, drainage with pressure control, pressure monitoring, and definition of holding pressure (No. 1). The second stopcock (No. 2) is used for injection of fluids and medications; a bacterial 0.22-µm-pore-size filter is designed to prevent unnecessary contamination. The third stopcock (No. 3) allows for the taking of CSF samples and fixation of the Combitrans transducer (B. Braun Melsungen AG). A fourth stopcock is connected to the Combitrans transducer for evacuation of air bubbles (No. 4). CSF volumes are monitored with a special four-chamber system, as depicted in Figure 1. A drip chamber serves as a barrier against infections; CSF volumes are monitored by four different chambers, and the CSF is collected in a collecting bag, which may be exchanged. The complete system is connected to the patient's bed by a holding device. The Combitrans transducer is connected to a pressure monitor with "alarm" and "trend" monitoring.
Recurrent hemorrhage As expected, the incidence of recurrent hemorrhage was much higher among patients with SAH than among those with other diagnoses (Figure 5). Among 23 patients with a first recurrent hemorrhage during EVD treatment, in 3 patients, a second hemorrhage occurred. In addition, 5 patients had recurrent hemorrhages after removal of the EVD; the overall mortality of patients with recurrent hemorrhages was 40%. Most hemorrhages occurred during the first week of treatment (Figure 4b). Recurrent hemorrhages were regularly accompanied by an increased intraventricular pressure, typically between 30 and 70 mm Hg (Figure 6). Infections Primary infections In six patients, EVD was necessary in the course of treatment of CSF infections, namely, pneumococcal meningitis (n = 1), tuberculous meningitis (n = 2), and shunt infections (n = 3), with staphylococci in two cases. In four cases, the primary infection could be treated and the CSF sterilized. Two patients died of their infections (pneumococcal and tuberculous infections). Leakage was observed in four of six cases (67%; P < 0.05; χ2 test) in comparison to 27% observed in the remaining patients (without primary infection). In four patients, secondary suturing was necessary; no additional morbidity was seen. In one patient, system occlusion was noted (initial catheter displacement), which could be reopened by system flushing, without any effect on the infection. In a second tuberculous patient, a system extraction was noted--the complete EVD had to be reestablished. Secondary infections In the remaining 94 patients, five secondary infections with clinical and laboratory signs of meningitis occurred; the infection was noted in two cases during EVD and in three cases 3, 5, and 8 days after removal. Time of EVD treatment had no influence on incidence of infection (data not shown). Patients without CSF leakage had a 1.6% rate of secondary infection, compared with 13% in patients with CSF leakage (P < 0.05; χ2 test) (Figure 7). In two cases, staphylococci were verified; in the remainder, no infectious agents could be identified. In all four cases, EVD was not removed until cessation of the initial indication for drainage. No patient died of EVD-related infection. Probable reasons for infection under EVD therapy were CSF leakage during (n = 1) and after (n = 3) EVD removal (Figure 7). Infections were treated with vancomycin (4 × 500 mg iv) and cefotaxim (3 × 2 g iv) or mezlocillin (4 × 4 g iv) for at least 14 days, and fever ceased within 2 to 5 days. CSF leakage required secondary suturing in three patients and local compression in the remaining patients. CSF pleocytosis decreased or normalized within 4, 5, 10, and 15 days; in one case, clinical signs of infection
were reduced and CSF was not investigated again. DISCUSSION An EVD system has been validated with the first 100 patients. This system has an integrated disposable pressure transducer allowing continuous intraventricular pressure monitoring, including continuous alarm alertness. CSF drainage is monitored in a closed system quantitatively, and probes can be drawn or injected under maximum hygienic conditions. Pressure adjustment to normal zero values and pressure relief do not require system disconnection. System handling may be performed distant from patient movements, excluding the danger of contamination or system dislocation. The drip chamber, the automatic overflow (multiple CSFcollecting chambers), and the collecting bag system prevent CSF reflux back into the ventricles. The entire system is fixed to the patient's bed by a stable holding device, preventing untoward manipulations with the system. With this system, the rate of infection is comparatively low (see below); patients with CSF leakage had a significantly higher risk of infection. However, the limitation of EVD system treatment may rest not so much in infections or technical complications, but in the elevated risk of recurrent hemorrhage, especially in SAH patients with acute hydrocephalus. Technical complications in EVD are mainly system disconnections and system occlusions, occurring during patient manipulations or drainage of grossly hemorrhagic CSF. Unfortunately, there are only few comparable data available in the literature covering these issues (20,22,26); in our population, we observed 10% technical complications (7 disconnections, 1 CSF leakage from the system, and 3 patient-induced EVD dislodgments)--none of these complications was associated with patient morbidity. In 19 patients (19%), system occlusions were observed, requiring a total EVD system renewal in 1 patient. The remaining occlusions could be reopened by flushing. No patient morbidity or mortality was observed related to system occlusions. Chan and Mann (4) did not observe occlusions requiring system renewal in 22 cases of ventricular hematoma; however, their external draining valve occluded in all cases, requiring new valves. In 37 children with posthemorrhagic hydrocephalus (mean age, 38 months), Rhodes et al. (22) found 41% occlusions and 13% dislodgments, with comparable rates for infections. It seems difficult to evaluate the question of secondary CSF infections because most authors apply prophylactic antibiotics (local or systemic). Rappaport and Shalit (21) reported a 10% rate of infection with 3% deaths in patients undergoing surgery for infratentorial tumors (EVD duration: mean, 2.3 d). A recent survey by Gerner-Smidt et al. (8) claimed a 17% infection rate with 7% infectionrelated mortality (depending on EVD duration). The highest infection rate of 50% (12 of 24) was observed by Hasan et al. (11), with necessary system removal in 5 patients and a 4% mortality rate; they also found a strong correlation of treatment duration with infection rate, which was not the case in our patients.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
respirator therapy, 12 patients (23%) died. No patient died of system occlusion.
less successfully. In our patients, flushing of the system usually removed blood clots, although once a new system (catheter) had to be installed because of permanent obstruction. However, in children, occlusion rates are higher (22,28), presumably because thinner catheters are selected. The main advantage of continuous-pressure monitoring, however, rests in early recognition of recurrent hemorrhages (pressure alarm) and beginning system occlusion (pulse amplitude reduction). It alerts the therapeutic team, whenever the system is obstructed, because the pulse modulation of the pressure curve is lost immediately. In addition, EVD pressure monitoring substitutes for epidural pressure monitoring and allows optimal treatment of brain edema in the disease course. No reliable data are available on the question of whether EVD provokes subsequent bleeding in SAH patients, which is frequently taken as an argument against EVD. In our study group of SAH patients, we observed a rate of 42% recurrent hemorrhages, which seems higher than that in comparable patient populations (27%) (14); however, 60% of our SAH patients were Hunt and Hess Grades IV and V--49% of these survived. Hasan et al. (11) observed identical high rates of recurrent hemorrhages (41%) in SAH patients with EVD treatment (28) and 55% mortality for all SAH patients (44% in our population). Does this mean that EVD treatment provokes recurrent hemorrhages (by releasing pressure on the hematoma or unclipped aneurysm), or is it merely a reflection of the negative patient selection process (1,9)? At this moment, the question may not be resolved, because neurosurgical and neurological SAH patient populations are different in their prognostic composition. It may be concluded that: 1) the indication for EVD therapy should be very strict; 2) an analysis should be performed including SAH patients cared for by neurosurgeons and neurologists; and 3) perhaps the relief pressure adjustment (counterpressure) should be elevated to 30 torr. ACKNOWLEDGMENT The superb personal and technical engagement of the neurological intensive care unit's nursing team, namely, Gertrud Moldenhauer, is highly appreciated. Received, September 3, 1991. Accepted, May 29, 1992. Reprint requests: Ulrich Bogdahn, M.D., Department of Neurology, Intensive Care Unit, JosefSchneider Str. 11, D-8700 Würzburg, Germany. REFERENCES: (1-28) 1.
2.
3.
Amato M, Guggisberg C, Kaiser G: Treatment of progressive posthemorrhagic hydrocephalus with temporary external ventricular drainage. Helv Paediat Acta 41:317-324, 1986. Auer ML, Mokry M: Disturbed cerebrospinal fluid circulation after subarachnoid hemorrhage and acute aneurysm surgery. Neurosurgery 26:804-809, 1990. Chan K-H, Mann KS: Intraventricular
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
In our population, we observed 5% infections with no associated mortality. However, most patients with infections had CSF leakage; if CSF leakage is excluded, the system-related rate of infection is 1% (primary infections are withdrawn from this evaluation). These infection rates have to be seen in the light of EVD treatment duration, which was performed for a mean of 9.5 (1 to 29) days in our patients (see below). From the literature, it is unknown what percentage of patients will have CSF leakage, which in our hands seemed to be a major risk factor for infection. In conclusion, infections may be controlled by systematic daily CSF controls and prophylactic antibiotics and are therefore not the limitation for EVD treatment, as long as CSF leakage is strictly avoided. Because most patients treated in our population were severely disabled (86% had been on artificial respirator therapy during the initial phase; mean, 10.2 d; ς = 8.1 d), the total mortality rate was relatively high, at 21% during EVD therapy. Considering merely SAH patients, mortality ranged at 29%, which would have been expected (27). Sixty percent of our patients were initially Hunt and Hess Grades IV and V, of whom 49% survived. In a smaller population of 22 patients, mortality was 23%; however, EVD was carried out for a mean of 12 days (8). In both populations, the initial neurological findings determined the final prognosis. Nonhemorrhagic populations have much lower mortality rates, e.g., of 8% in infratentorial tumors (21). No serious morbidity was observed, aside from a 5% infection rate. No serious technical complication occurred; however, the rate of any complication occurring was 58% (some patients had more than one incident). In our population, EVD with a single system and catheter was carried out for a mean time of 9.5 days with a maximum of 29 days. In 40 patients, the catheter rested for more than 10 days. With a different system (4), a mean drainage time of 16 days with a maximum of 44 days was achieved in 34 patients (noncontinuous drainage without pressure monitoring); although the infection rate was low, the necessary valve had to be exchanged frequently because of blockage. A few more authors (13,22,28), describe long-term application of EVD treatment, although complications are difficult to evaluate because of differences in diagnoses, systems, etc. In all cases, prophylactic antibiotic therapy had been applied. No correlation was found in our patients between time of drainage and infection rate. These data supply evidence that long-term (1 to 2 wk normally, up to 44 d; [4]) EVD treatment may be performed securely without having to change the catheter system. Prophylactic antibiotic treatment by flushing the system with 5 mg of gentamycin twice per day and performing regular daily CSF controls have proven successful in our hands. System blockage has been a major argument against EVD treatment, especially in SAH and hemorrhage; as has been described in the literature (4, 22) so far, system blockage could be managed more or
5. 6. 7. 8.
9.
10.
11.
12.
13.
14.
15.
16.
17. 18. 19. 20.
21.
22.
23.
24. 25.
26. 27.
28.
hydrocephalus secondary to infratentorial brain tumors. Acta Neurochir (Wien) 96:118121, 1989. Rhodes TT, Edwards WH, Saunders RL, Harbaugh RE, Little CLC, Morgan LJ, Sargent SK: External ventricular drainage for initial treatment of neonatal posthemorrhagic hydrocephalus: Surgical and neurodevelopmental outcome. Pediatr Neurosci 13:255-262, 1987. Richard KE: Liquorventrikeldruckmessung mit Mikrokatheter und druckkontrollierte externe Liquordrainage. Acta Neurochir 38:7387, 1977. Rosner MJ, Becker DP: ICP monitoring: Complications and associated factors. Clin Neurosurg 23:494-519, 1976. Scheinblum ST, Hammond M: The treatment of children with shunt infections: Extraventricular drainage system care. Pediatr Nurs 16:139-143, 1990. Smith RW, Alksne F: Infections complicating the use of external ventriculostomy. J Neurosurg 44:567-570, 1976. Torner JC, Kassell NF, Wallace RB, Adams HP: Preoperative prognostic factors for rebleeding and survival in aneurysm patients receiving antifibrinolytic therapy: Report of the cooperative aneurysm study. Neurosurgery 9:506-513, 1981. Weniger M, Simbruner G, Salzer HR, Rosenkranz M, Lesigang C: Externe Ventrikeldrainage bei Neugeborenen mit rasch wachsenden Hydrocephalus. Wien Klin Wochenschr Jg 100:561-564, 1988.
COMMENTS The strength of the analysis of long-term ventricular drainage in neurosurgical patients by Bogdahn et al. is their breakdown of complication by category. This provides a forum for interesting discussion as to whether or not ventricular drainage potentiates subsequent hemorrhage from aneurysm and other lesions. Unfortunately, the weakness of the article is that none of the discussion evoked is answered. We have known for 30 years that long-term ventricular drainage can be accomplished with relative safety. The authors claim the same relative finding. We have known for nearly 20 years that cerebrospinal fluid (CSF) leak, basal fracture, craniotomy, and other openings of the intracranial contents are associated with an increased risk of infection in association with a monitoring device. The current evaluation suggests that CSF leakage is similarly associated with an increase in such risk. However, these data are purely correlative. Cause and effect cannot be established from this type of study. Those patients with CSF leakage from the ventriculostomy are more likely to have complex
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
4.
haematoma management of comatose patients with valve regulated external ventricular drainage. Br J Neurosurg 2:465-469, 1988. Chan K-H, Mann KS: Prolonged therapeutic external ventricular drainage: A prospective study. Neurosurgery 23:436-438, 1988. Duncan CC: Management of proximal shunt obstruction. J Neurosurg 68:817-819, 1988. Fan-Havard P, Nahata MC: Treatment and prevention of infections of cerebrospinal fluid shunts. Clin Pharm 6:866-880, 1987. Gardner P, Leipzig T, Phillips P: Infections of central nervous system shunts. Med Clin North Am 69:297-314, 1985. Gerner-Smidt P, Stenager E, Kock-Jensen C: Treatment of ventriculostomy-related infections. Acta Neurochir (Wien) 91:47-49, 1988. Gijn JV, Hijdra A, Wijdicks EFM, Vermeulen M, Crevel HV: Acute hydrocephalus after aneurysmal subarachnoid hemorrhage. J Neurosurg 26:804-809, 1990. Gruber R, Bubl R, Britschgi U, Jehle B: Der Stellenwert der externen Drainage für die Therapie der intrakraniellen Blutung im Neugeborenen- und Kindesalter. Z Kinderchir 41:144-150, 1986. Hasan D, Vermeulen M, Wijdicks EFM, Hijdra A, Gijn JV: Management problems in acute hydrocephalus after subarachnoid hemorrhage. Stroke 20:747-753, 1989. Hunt WE, Hess RM: Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J Neurosurg 28:14-19, 1968. Itakura T, Yokote H, Ozaki F, Itatani K, Hayashi S, Komai N: Stereotactic operation for brain abscess. Surg Neurol 28:196-200, 1987. Kassell NF, Torner JC: Aneurysmal rebleeding: A preliminary report from the cooperative aneurysm study. Neurosurgery 13:479-481, 1983. Lundberg N: Continuous recording and control of ventricular fluid pressure in neurosurgical practice. Acta Psychiatr Scand 36(Suppl 149):1-193, 1960. Lundberg N, Troupp H, Lorin H: Continuous recording of the ventricular-fluid pressure in patients with severe acute traumatic brain injury. J Neurosurg 22:581-590, 1965. Merrem B: Die Ventrikeldrainage. Zbl Neurochir 31:127-148, 1970. Pampus F: Zur Technik der Ventrikeldrainage. Zbl Neurochir 13:219-223, 1953. Pampus F: Die Indikation zur Anwendung der Ventrikeldrainage. Zbl Neurochir 21:216-221, 1961. Pezzotta S, Locatelli D, Bonfanti N, Sfogliarini R, Brushi L, Rondini G: Shunt in high-risk newborns. Child Nerv Syst 3:114116, 1987. Rappaport ZH, Shalit MN: Perioperative external ventricular drainage in obstructive
Michael J. Rosner Birmingham, Alabama The authors present statistical information on 100 well-studied patients undergoing external ventricular drainage (EVD) procedures for a variety of indications. The system they used is of their own design, because some of the systems that are used frequently in the United States are not available in Germany. The patients treated in this way were seriously ill as attested by the 29% mortality rate in their subarachnoid hemorrhage patients and an overall 35% mortality unrelated to the EVD itself. The complication rate experienced in this series is gratifyingly low, with an overall infection rate of 5% and one of only 2% in patients in whom no leak was found in the system. Whether this low infection rate can be attributed to the routine injection of intrathecal antibiotics cannot be gleaned from this study. Recurrent hemorrhage that appears to be unrelated
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
problems, they are more likely to have had more severe hemorrhages, they are more likely to have catheter replacement(s) with continued manipulation of the catheter system itself, and all of these are likely to increase infection rate. A high correlation will then be established between CSF leakage and infective complication, but not cause and effect. Similarly, to begin to address the question of whether subarachnoid hemorrhage and recurrent hemorrhage are potentiated by ventricular drainage, patient groups similar in diagnostic and other categories must be established where the only variable is the presence or absence of the ventricular drain or the pressure gradient against which the ventricular drainage system is opened. This type of study would then begin to produce useful information. The authors state that they find no relationship between the time of drainage and CSF infection. This is an indirect conflict with the majority of the literature dealing with foreign body implantation, line infection, and previously published information. The answer is probably hidden in the relatively few patients who had CSF drainage prolonged beyond 12 days. Those patients most likely to be drained for very prolonged periods of time are those most likely to have nicely functioning but easily maintained drainage systems and do not necessarily represent the same population of patients who have large amounts of bloody drainage, easily blocked catheters, or other complicating factors such as those that increase the likelihood of CSF leakage, etc. The design of this "study" makes it almost impossible to make conclusions regarding the length of CSF drainage versus infection rate. Another unanswered question relates to the use of antibiotic solutions in the drainage system. These are often recommended, but there is no evidence that they are actually useful. It is conceivable that they may not only be useless, but even harmful. Do such solutions potentiate the development of hydrocephalus? Do they increase the likelihood of superinfection with gram-negative bacteria or fungi?
Figure 2. Mean time of the pressure-controlled EVD.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 1. Schematic drawing of the ventricular drainage system, as described in the text.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 3. Occurrence of first, second, and third CSF leakage along the catheter during EVD.
Figure 5. Time course and incidence of recurrent hemorrhage during EVD in all patients. Most hemorrhages occurred during the first 10 days.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 4. (A) Frequency of ventricular drainage system occlusion in patients with either SAH or with other lesions. In 27% of patients with SAH, occlusion occurred, compared with 13% in the remainder (P < 0.01; χ2 test) (B) Frequency of recurrent hemorrhage in patients with SAH and other diseases. In 42.2% of SAH patients, a recurrent hemorrhage occurred, whereas this was observed in only 12.7% of the remaining patients with other diagnoses.
Figure 7. Incidence of secondary CSF infections in patients with and without CSF leakage. Four patients out of five with secondary central nervous system infections had CSF leakage.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 6. On-line registration of ventricular pressure during two events of recurrent hemorrhages (1, 2) in a patient with SAH hemorrhage. In both episodes (1,2), ventricular pressure increased to 40 mm Hg.
Refobacin; Shunt infections; Subarachnoid hemorrhage; Ventricular drainage External ventricular drainage (EVD), as described earlier by Pampus (18,19), Lundberg (15), and Merrem (17) is frequently required in the treatment of adults with acute cerebrospinal fluid (CSF) circulation disturbances: most patients have intracranial or subarachnoid hemorrhages (SAH); some patients have acute space-occupying lesions requiring preoperative CSF drainage. EVD is widely employed, and data on systems and their complication rates are frequently reported for pediatric patients (1,5-7,10,20,22, 25,28) . Pressure recording and concomitant CSF drainage, as described originally (17), in children (5) and in adults (2,16,23,24), have been reported more recently. Continuous drainage of CSF may result, however, in a number of additional risks, especially subsequent bleeding. Continuous-pressure monitoring, prolonged handling, and regular screening of CSF probes may be associated with a chance of system contamination. We report our experience with an EVD system allowing continuous pressure monitoring; results of safety and complications are presented. PATIENTS AND METHODS Patients Data of 100 intensive care unit patients (n = 49 female, mean age, 56.4 yr; n = 51 male, mean age, 56.1 yr) with ventricular catheter treatment were analyzed. Treatment was indicated in all cases by acute CSF circulation disturbances (acute hydrocephalus) caused by the following diagnoses: SAH from aneurysm (n = 17) or unknown source (n = 28) classified according to Hunt and Hess (12), parenchyma hemorrhages (n = 30), angioma (n = 4), anticoagulant hemorrhages (n = 7), and rare cases of infection (n = 3), tumor (n = 2), ischemia (posterior fossa infarctions with hydrocephalus; n = 5), or unknown reasons (n = 4). Among these, 52 patients had ventricular hemorrhages. All patients received heparin at approximately 15,000 IU/day subcutaneously or intravenously. The EVD system was removed whenever the intracranial pressure values read up to a maximum of 20 torrs for 2 consecutive days. Ventricular drainage system A ventricular drainage system (B. Braun Melsungen A.G., Melsungen, Germany; Figure 1) has been used in our department with the following characteristics: it has a rigid CSF collector with quantitative volume measurement and a CSF collecting bag with automatic overflow, as well as a stable holding device keeping the manifold plate including CSF pressure transducer away from the patient (for hygiene reasons, disconnection, handling). A Codman ventricular Silicon catheter (Codman & Shurtleff, Inc., Randolph, MA) is placed into the lateral ventricle by a standard frontal approach in the nondominant hemisphere and is connected to the drainage system (150-cm-long pressure-resistant tubing) at a three-way stopcock
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 November 1992, Volume 31, Number 5 898 Continuous-Pressure Controlled, External Ventricular Drainage for Treatment of Acute Hydrocephalus--Evaluation of Risk Factors Clinical Study
Data analysis Patients' data have been evaluated for systemic signs of infection (fever, leukocytosis >10,000/cbm, creactive protein, ESR, microbiology) and CSF infections (pleocytosis, microbiology, clinical signs). Catheter misplacement, dislocation, disconnection, and occlusion were analyzed, as was secondary neurosurgical treatment. In addition, clinical parameters, such as mortality and morbidity, and signs of recurrent hemmorhage were documented. Data were accumulated during the time patients were monitored at the neurological intensive care unit. Neuroradiological procedures were performed as clinically indicated and were also retrieved for evaluation. Standard statistical analysis was performed, mainly as a χ2 test. RESULTS General results The mean time of EVD was 9.5 days (range, 1 to 30 days); the complete EVD system was exchanged only twice, either when blood clots irreversibly occluded the tubing or in a case with catheter extraction requiring continuous drainage (Figure 2). Cutaneous wound reactions were registered in two cases, but without morbidity. The overall patient mortality was 35%; 14 patients (40%) died of recurrent hemorrhages; the remainder died of complications of their primary disease, mainly brain edema. Mortality during EVD therapy was 21%. No patient died of EVD-related morbidity, especially EVD-related infections or catheter dislocation. In no single case did technical complications lead to serious patient morbidity. After removal of the EVD system in six cases, a permanent ventriculoperitoneal shunt system was installed because of permanently raised CSF pressure and/or recurring hydrocephalus. Safety In 10 patients (10%), technical complications
occurred without patient morbidity. These involved system disconnections (patients, n = 5; treatment team, n = 2), system leakage of the tubing (n = 1), and patient-induced EVD extraction (n = 3). System occlusions occurred in 19 cases, and in nearly all cases, flushing with saline reopened the catheter. The entire system (including catheter) had to be exchanged in one case of tuberculous meningitis, after it became extracted. In a second case, a recurrent hemorrhage caused an occlusion, which could not be reopened; a new EVD had to be placed. In 11 cases, the EVD catheter was initially misplaced and had to be corrected. In four cases, more than one puncture was necessary; no puncture-associated hemorrhage occurred. Cerebrospinal fluid leakage CSF leakage, as defined by leakage beside the catheter, was observed during EVD treatment in 26 patients (26%) at least once, in 15 patients twice, and in 7 patients three times (Figure 3); leakage was observed from Day 0 to Day 16 postoperatively. Secondary suturing was required in 21 patients. After removal of EVD catheters in 13 cases, CSF leakage was documented once, and in 2 cases, it was documented twice; the longest interval of CSF leakage after removal was 4 days. Secondary suturing after EVD removal was performed in eight cases. In four of five patients with secondary infections, CSF leakage was found (see below). In most cases, CSF leakage was successfully treated by local compression for up to 24 hours. Subarachnoid hemorrhage According to clinical grading (Hunt and Hess 12), we observed 9 patients, each with Grades II, III, and V and 18 patients with Grade IV. Fourteen of 27 patients with unfavorable Hunt and Hess Grades IV and V died in the hospital. System occlusion occurred in 12 (27%) of 45 SAH patients, compared with 7 (15%) of 55 patients with other diagnoses (P < 0.1; χ2 test); however, no patient died from occlusion, because this was recognized early by continuous pressure monitoring (Figure 4a; see Figure 6). In two patients, the external part of the EVD systems had to be reinstalled (without ventricular catheter replacement). Recurrent hemorrhages (Figure 4b) were noted in 19 (42%) of 45 SAH patients while under EVD treatment, compared with 7 (13%) of 55 patients with other diagnoses (P < 0.01; χ2 test). Ten patients with recurrent hemorrhages died from their hemorrhage under EVD treatment (53%). A single recurrent hemorrhage causing an occlusion required an EVD replacement. The entire SAH mortality rate was 29% as compared to 17% for patients with other diagnoses, while patients were under EVD therapy (Figure 4). Ventricular hemorrhage In 52 patients, ventricular hemorrhage was initially diagnosed. Occlusions occurred in 12 cases (23%), and recurrent hemorrhages occurred in 11 cases (21%) (not significantly different compared to the remaining patients). Forty-nine patients were on
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
manifold plate. The manifold plate consists of four stopcocks: a first stopcock allows CSF drainage to the CSF collector, pressure adjustment to normal zero values, drainage with pressure control, pressure monitoring, and definition of holding pressure (No. 1). The second stopcock (No. 2) is used for injection of fluids and medications; a bacterial 0.22-µm-pore-size filter is designed to prevent unnecessary contamination. The third stopcock (No. 3) allows for the taking of CSF samples and fixation of the Combitrans transducer (B. Braun Melsungen AG). A fourth stopcock is connected to the Combitrans transducer for evacuation of air bubbles (No. 4). CSF volumes are monitored with a special four-chamber system, as depicted in Figure 1. A drip chamber serves as a barrier against infections; CSF volumes are monitored by four different chambers, and the CSF is collected in a collecting bag, which may be exchanged. The complete system is connected to the patient's bed by a holding device. The Combitrans transducer is connected to a pressure monitor with "alarm" and "trend" monitoring.
Recurrent hemorrhage As expected, the incidence of recurrent hemorrhage was much higher among patients with SAH than among those with other diagnoses (Figure 5). Among 23 patients with a first recurrent hemorrhage during EVD treatment, in 3 patients, a second hemorrhage occurred. In addition, 5 patients had recurrent hemorrhages after removal of the EVD; the overall mortality of patients with recurrent hemorrhages was 40%. Most hemorrhages occurred during the first week of treatment (Figure 4b). Recurrent hemorrhages were regularly accompanied by an increased intraventricular pressure, typically between 30 and 70 mm Hg (Figure 6). Infections Primary infections In six patients, EVD was necessary in the course of treatment of CSF infections, namely, pneumococcal meningitis (n = 1), tuberculous meningitis (n = 2), and shunt infections (n = 3), with staphylococci in two cases. In four cases, the primary infection could be treated and the CSF sterilized. Two patients died of their infections (pneumococcal and tuberculous infections). Leakage was observed in four of six cases (67%; P < 0.05; χ2 test) in comparison to 27% observed in the remaining patients (without primary infection). In four patients, secondary suturing was necessary; no additional morbidity was seen. In one patient, system occlusion was noted (initial catheter displacement), which could be reopened by system flushing, without any effect on the infection. In a second tuberculous patient, a system extraction was noted--the complete EVD had to be reestablished. Secondary infections In the remaining 94 patients, five secondary infections with clinical and laboratory signs of meningitis occurred; the infection was noted in two cases during EVD and in three cases 3, 5, and 8 days after removal. Time of EVD treatment had no influence on incidence of infection (data not shown). Patients without CSF leakage had a 1.6% rate of secondary infection, compared with 13% in patients with CSF leakage (P < 0.05; χ2 test) (Figure 7). In two cases, staphylococci were verified; in the remainder, no infectious agents could be identified. In all four cases, EVD was not removed until cessation of the initial indication for drainage. No patient died of EVD-related infection. Probable reasons for infection under EVD therapy were CSF leakage during (n = 1) and after (n = 3) EVD removal (Figure 7). Infections were treated with vancomycin (4 × 500 mg iv) and cefotaxim (3 × 2 g iv) or mezlocillin (4 × 4 g iv) for at least 14 days, and fever ceased within 2 to 5 days. CSF leakage required secondary suturing in three patients and local compression in the remaining patients. CSF pleocytosis decreased or normalized within 4, 5, 10, and 15 days; in one case, clinical signs of infection
were reduced and CSF was not investigated again. DISCUSSION An EVD system has been validated with the first 100 patients. This system has an integrated disposable pressure transducer allowing continuous intraventricular pressure monitoring, including continuous alarm alertness. CSF drainage is monitored in a closed system quantitatively, and probes can be drawn or injected under maximum hygienic conditions. Pressure adjustment to normal zero values and pressure relief do not require system disconnection. System handling may be performed distant from patient movements, excluding the danger of contamination or system dislocation. The drip chamber, the automatic overflow (multiple CSFcollecting chambers), and the collecting bag system prevent CSF reflux back into the ventricles. The entire system is fixed to the patient's bed by a stable holding device, preventing untoward manipulations with the system. With this system, the rate of infection is comparatively low (see below); patients with CSF leakage had a significantly higher risk of infection. However, the limitation of EVD system treatment may rest not so much in infections or technical complications, but in the elevated risk of recurrent hemorrhage, especially in SAH patients with acute hydrocephalus. Technical complications in EVD are mainly system disconnections and system occlusions, occurring during patient manipulations or drainage of grossly hemorrhagic CSF. Unfortunately, there are only few comparable data available in the literature covering these issues (20,22,26); in our population, we observed 10% technical complications (7 disconnections, 1 CSF leakage from the system, and 3 patient-induced EVD dislodgments)--none of these complications was associated with patient morbidity. In 19 patients (19%), system occlusions were observed, requiring a total EVD system renewal in 1 patient. The remaining occlusions could be reopened by flushing. No patient morbidity or mortality was observed related to system occlusions. Chan and Mann (4) did not observe occlusions requiring system renewal in 22 cases of ventricular hematoma; however, their external draining valve occluded in all cases, requiring new valves. In 37 children with posthemorrhagic hydrocephalus (mean age, 38 months), Rhodes et al. (22) found 41% occlusions and 13% dislodgments, with comparable rates for infections. It seems difficult to evaluate the question of secondary CSF infections because most authors apply prophylactic antibiotics (local or systemic). Rappaport and Shalit (21) reported a 10% rate of infection with 3% deaths in patients undergoing surgery for infratentorial tumors (EVD duration: mean, 2.3 d). A recent survey by Gerner-Smidt et al. (8) claimed a 17% infection rate with 7% infectionrelated mortality (depending on EVD duration). The highest infection rate of 50% (12 of 24) was observed by Hasan et al. (11), with necessary system removal in 5 patients and a 4% mortality rate; they also found a strong correlation of treatment duration with infection rate, which was not the case in our patients.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
respirator therapy, 12 patients (23%) died. No patient died of system occlusion.
less successfully. In our patients, flushing of the system usually removed blood clots, although once a new system (catheter) had to be installed because of permanent obstruction. However, in children, occlusion rates are higher (22,28), presumably because thinner catheters are selected. The main advantage of continuous-pressure monitoring, however, rests in early recognition of recurrent hemorrhages (pressure alarm) and beginning system occlusion (pulse amplitude reduction). It alerts the therapeutic team, whenever the system is obstructed, because the pulse modulation of the pressure curve is lost immediately. In addition, EVD pressure monitoring substitutes for epidural pressure monitoring and allows optimal treatment of brain edema in the disease course. No reliable data are available on the question of whether EVD provokes subsequent bleeding in SAH patients, which is frequently taken as an argument against EVD. In our study group of SAH patients, we observed a rate of 42% recurrent hemorrhages, which seems higher than that in comparable patient populations (27%) (14); however, 60% of our SAH patients were Hunt and Hess Grades IV and V--49% of these survived. Hasan et al. (11) observed identical high rates of recurrent hemorrhages (41%) in SAH patients with EVD treatment (28) and 55% mortality for all SAH patients (44% in our population). Does this mean that EVD treatment provokes recurrent hemorrhages (by releasing pressure on the hematoma or unclipped aneurysm), or is it merely a reflection of the negative patient selection process (1,9)? At this moment, the question may not be resolved, because neurosurgical and neurological SAH patient populations are different in their prognostic composition. It may be concluded that: 1) the indication for EVD therapy should be very strict; 2) an analysis should be performed including SAH patients cared for by neurosurgeons and neurologists; and 3) perhaps the relief pressure adjustment (counterpressure) should be elevated to 30 torr. ACKNOWLEDGMENT The superb personal and technical engagement of the neurological intensive care unit's nursing team, namely, Gertrud Moldenhauer, is highly appreciated. Received, September 3, 1991. Accepted, May 29, 1992. Reprint requests: Ulrich Bogdahn, M.D., Department of Neurology, Intensive Care Unit, JosefSchneider Str. 11, D-8700 Würzburg, Germany. REFERENCES: (1-28) 1.
2.
3.
Amato M, Guggisberg C, Kaiser G: Treatment of progressive posthemorrhagic hydrocephalus with temporary external ventricular drainage. Helv Paediat Acta 41:317-324, 1986. Auer ML, Mokry M: Disturbed cerebrospinal fluid circulation after subarachnoid hemorrhage and acute aneurysm surgery. Neurosurgery 26:804-809, 1990. Chan K-H, Mann KS: Intraventricular
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
In our population, we observed 5% infections with no associated mortality. However, most patients with infections had CSF leakage; if CSF leakage is excluded, the system-related rate of infection is 1% (primary infections are withdrawn from this evaluation). These infection rates have to be seen in the light of EVD treatment duration, which was performed for a mean of 9.5 (1 to 29) days in our patients (see below). From the literature, it is unknown what percentage of patients will have CSF leakage, which in our hands seemed to be a major risk factor for infection. In conclusion, infections may be controlled by systematic daily CSF controls and prophylactic antibiotics and are therefore not the limitation for EVD treatment, as long as CSF leakage is strictly avoided. Because most patients treated in our population were severely disabled (86% had been on artificial respirator therapy during the initial phase; mean, 10.2 d; ς = 8.1 d), the total mortality rate was relatively high, at 21% during EVD therapy. Considering merely SAH patients, mortality ranged at 29%, which would have been expected (27). Sixty percent of our patients were initially Hunt and Hess Grades IV and V, of whom 49% survived. In a smaller population of 22 patients, mortality was 23%; however, EVD was carried out for a mean of 12 days (8). In both populations, the initial neurological findings determined the final prognosis. Nonhemorrhagic populations have much lower mortality rates, e.g., of 8% in infratentorial tumors (21). No serious morbidity was observed, aside from a 5% infection rate. No serious technical complication occurred; however, the rate of any complication occurring was 58% (some patients had more than one incident). In our population, EVD with a single system and catheter was carried out for a mean time of 9.5 days with a maximum of 29 days. In 40 patients, the catheter rested for more than 10 days. With a different system (4), a mean drainage time of 16 days with a maximum of 44 days was achieved in 34 patients (noncontinuous drainage without pressure monitoring); although the infection rate was low, the necessary valve had to be exchanged frequently because of blockage. A few more authors (13,22,28), describe long-term application of EVD treatment, although complications are difficult to evaluate because of differences in diagnoses, systems, etc. In all cases, prophylactic antibiotic therapy had been applied. No correlation was found in our patients between time of drainage and infection rate. These data supply evidence that long-term (1 to 2 wk normally, up to 44 d; [4]) EVD treatment may be performed securely without having to change the catheter system. Prophylactic antibiotic treatment by flushing the system with 5 mg of gentamycin twice per day and performing regular daily CSF controls have proven successful in our hands. System blockage has been a major argument against EVD treatment, especially in SAH and hemorrhage; as has been described in the literature (4, 22) so far, system blockage could be managed more or
5. 6. 7. 8.
9.
10.
11.
12.
13.
14.
15.
16.
17. 18. 19. 20.
21.
22.
23.
24. 25.
26. 27.
28.
hydrocephalus secondary to infratentorial brain tumors. Acta Neurochir (Wien) 96:118121, 1989. Rhodes TT, Edwards WH, Saunders RL, Harbaugh RE, Little CLC, Morgan LJ, Sargent SK: External ventricular drainage for initial treatment of neonatal posthemorrhagic hydrocephalus: Surgical and neurodevelopmental outcome. Pediatr Neurosci 13:255-262, 1987. Richard KE: Liquorventrikeldruckmessung mit Mikrokatheter und druckkontrollierte externe Liquordrainage. Acta Neurochir 38:7387, 1977. Rosner MJ, Becker DP: ICP monitoring: Complications and associated factors. Clin Neurosurg 23:494-519, 1976. Scheinblum ST, Hammond M: The treatment of children with shunt infections: Extraventricular drainage system care. Pediatr Nurs 16:139-143, 1990. Smith RW, Alksne F: Infections complicating the use of external ventriculostomy. J Neurosurg 44:567-570, 1976. Torner JC, Kassell NF, Wallace RB, Adams HP: Preoperative prognostic factors for rebleeding and survival in aneurysm patients receiving antifibrinolytic therapy: Report of the cooperative aneurysm study. Neurosurgery 9:506-513, 1981. Weniger M, Simbruner G, Salzer HR, Rosenkranz M, Lesigang C: Externe Ventrikeldrainage bei Neugeborenen mit rasch wachsenden Hydrocephalus. Wien Klin Wochenschr Jg 100:561-564, 1988.
COMMENTS The strength of the analysis of long-term ventricular drainage in neurosurgical patients by Bogdahn et al. is their breakdown of complication by category. This provides a forum for interesting discussion as to whether or not ventricular drainage potentiates subsequent hemorrhage from aneurysm and other lesions. Unfortunately, the weakness of the article is that none of the discussion evoked is answered. We have known for 30 years that long-term ventricular drainage can be accomplished with relative safety. The authors claim the same relative finding. We have known for nearly 20 years that cerebrospinal fluid (CSF) leak, basal fracture, craniotomy, and other openings of the intracranial contents are associated with an increased risk of infection in association with a monitoring device. The current evaluation suggests that CSF leakage is similarly associated with an increase in such risk. However, these data are purely correlative. Cause and effect cannot be established from this type of study. Those patients with CSF leakage from the ventriculostomy are more likely to have complex
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
4.
haematoma management of comatose patients with valve regulated external ventricular drainage. Br J Neurosurg 2:465-469, 1988. Chan K-H, Mann KS: Prolonged therapeutic external ventricular drainage: A prospective study. Neurosurgery 23:436-438, 1988. Duncan CC: Management of proximal shunt obstruction. J Neurosurg 68:817-819, 1988. Fan-Havard P, Nahata MC: Treatment and prevention of infections of cerebrospinal fluid shunts. Clin Pharm 6:866-880, 1987. Gardner P, Leipzig T, Phillips P: Infections of central nervous system shunts. Med Clin North Am 69:297-314, 1985. Gerner-Smidt P, Stenager E, Kock-Jensen C: Treatment of ventriculostomy-related infections. Acta Neurochir (Wien) 91:47-49, 1988. Gijn JV, Hijdra A, Wijdicks EFM, Vermeulen M, Crevel HV: Acute hydrocephalus after aneurysmal subarachnoid hemorrhage. J Neurosurg 26:804-809, 1990. Gruber R, Bubl R, Britschgi U, Jehle B: Der Stellenwert der externen Drainage für die Therapie der intrakraniellen Blutung im Neugeborenen- und Kindesalter. Z Kinderchir 41:144-150, 1986. Hasan D, Vermeulen M, Wijdicks EFM, Hijdra A, Gijn JV: Management problems in acute hydrocephalus after subarachnoid hemorrhage. Stroke 20:747-753, 1989. Hunt WE, Hess RM: Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J Neurosurg 28:14-19, 1968. Itakura T, Yokote H, Ozaki F, Itatani K, Hayashi S, Komai N: Stereotactic operation for brain abscess. Surg Neurol 28:196-200, 1987. Kassell NF, Torner JC: Aneurysmal rebleeding: A preliminary report from the cooperative aneurysm study. Neurosurgery 13:479-481, 1983. Lundberg N: Continuous recording and control of ventricular fluid pressure in neurosurgical practice. Acta Psychiatr Scand 36(Suppl 149):1-193, 1960. Lundberg N, Troupp H, Lorin H: Continuous recording of the ventricular-fluid pressure in patients with severe acute traumatic brain injury. J Neurosurg 22:581-590, 1965. Merrem B: Die Ventrikeldrainage. Zbl Neurochir 31:127-148, 1970. Pampus F: Zur Technik der Ventrikeldrainage. Zbl Neurochir 13:219-223, 1953. Pampus F: Die Indikation zur Anwendung der Ventrikeldrainage. Zbl Neurochir 21:216-221, 1961. Pezzotta S, Locatelli D, Bonfanti N, Sfogliarini R, Brushi L, Rondini G: Shunt in high-risk newborns. Child Nerv Syst 3:114116, 1987. Rappaport ZH, Shalit MN: Perioperative external ventricular drainage in obstructive
Michael J. Rosner Birmingham, Alabama The authors present statistical information on 100 well-studied patients undergoing external ventricular drainage (EVD) procedures for a variety of indications. The system they used is of their own design, because some of the systems that are used frequently in the United States are not available in Germany. The patients treated in this way were seriously ill as attested by the 29% mortality rate in their subarachnoid hemorrhage patients and an overall 35% mortality unrelated to the EVD itself. The complication rate experienced in this series is gratifyingly low, with an overall infection rate of 5% and one of only 2% in patients in whom no leak was found in the system. Whether this low infection rate can be attributed to the routine injection of intrathecal antibiotics cannot be gleaned from this study. Recurrent hemorrhage that appears to be unrelated
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
problems, they are more likely to have had more severe hemorrhages, they are more likely to have catheter replacement(s) with continued manipulation of the catheter system itself, and all of these are likely to increase infection rate. A high correlation will then be established between CSF leakage and infective complication, but not cause and effect. Similarly, to begin to address the question of whether subarachnoid hemorrhage and recurrent hemorrhage are potentiated by ventricular drainage, patient groups similar in diagnostic and other categories must be established where the only variable is the presence or absence of the ventricular drain or the pressure gradient against which the ventricular drainage system is opened. This type of study would then begin to produce useful information. The authors state that they find no relationship between the time of drainage and CSF infection. This is an indirect conflict with the majority of the literature dealing with foreign body implantation, line infection, and previously published information. The answer is probably hidden in the relatively few patients who had CSF drainage prolonged beyond 12 days. Those patients most likely to be drained for very prolonged periods of time are those most likely to have nicely functioning but easily maintained drainage systems and do not necessarily represent the same population of patients who have large amounts of bloody drainage, easily blocked catheters, or other complicating factors such as those that increase the likelihood of CSF leakage, etc. The design of this "study" makes it almost impossible to make conclusions regarding the length of CSF drainage versus infection rate. Another unanswered question relates to the use of antibiotic solutions in the drainage system. These are often recommended, but there is no evidence that they are actually useful. It is conceivable that they may not only be useless, but even harmful. Do such solutions potentiate the development of hydrocephalus? Do they increase the likelihood of superinfection with gram-negative bacteria or fungi?
Figure 2. Mean time of the pressure-controlled EVD.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 1. Schematic drawing of the ventricular drainage system, as described in the text.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 3. Occurrence of first, second, and third CSF leakage along the catheter during EVD.
Figure 5. Time course and incidence of recurrent hemorrhage during EVD in all patients. Most hemorrhages occurred during the first 10 days.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 4. (A) Frequency of ventricular drainage system occlusion in patients with either SAH or with other lesions. In 27% of patients with SAH, occlusion occurred, compared with 13% in the remainder (P < 0.01; χ2 test) (B) Frequency of recurrent hemorrhage in patients with SAH and other diseases. In 42.2% of SAH patients, a recurrent hemorrhage occurred, whereas this was observed in only 12.7% of the remaining patients with other diagnoses.
Figure 7. Incidence of secondary CSF infections in patients with and without CSF leakage. Four patients out of five with secondary central nervous system infections had CSF leakage.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 6. On-line registration of ventricular pressure during two events of recurrent hemorrhages (1, 2) in a patient with SAH hemorrhage. In both episodes (1,2), ventricular pressure increased to 40 mm Hg.
