EARLYDRAIN- outcome after early lumbar CSF-drainage in aneurysmal subarachnoid hemorrhage: study protocol for a randomized controlled trial

Study design

The present study is in compliance with the Helsinki Declaration. Ethical approval was obtained by the ethical committee of the medical faculty of the Friedrich-Alexander-University Erlangen-Nürnberg (reference number: 4171).

The EARLYDRAIN study is a 2-arm randomized controlled trial to compare an intervention group receiving early continuous lumbar CSF-drainage and standard neurointensive care to a control group receiving standard neurointensive care only. It is conducted by a German national study group consisting of neurosurgical centers treating at least 30 patients with aneurysmal subarachnoid hemorrhage per year. Data management and monitoring will be performed by the Center for Stroke Research Berlin (CSB) at Charité University Medicine, Berlin, Germany.

Patients suffering from an aneurysmal SAH and completed elimination of the causative aneurysm are being recruited for this study. The choice of the method of aneurysm treatment is at the discretion of the neurovascular team taking care of a patient and not specified by the study protocol. All medical treatment is performed according to local guidelines and standard operating procedures.

Subject Inclusion criteria

Subject Exclusion criteria

Interventions

In order not to provoke premature rupture of the aneurysm due to accidental drainage, randomization to the study and eventual placement of a lumbar drain takes place after securing the aneurysm by the preferred method of choice (Figure 1). Every patient in the lumbar drainage group (LD-group) receives a lumbar drain during anesthesia required for the aneurysm treatment. Insertion of a lumbar drain into the subarachnoid space is conducted in standard fully sterile technique. This is to be performed before anticoagulation or anti-platelet therapy is initiated, which sometimes is warranted after endovascular coiling. A post-procedural CCT scan of the brain is performed within 24 hours after aneurysm treatment. In case of neurological worsening after the procedure it is strongly recommended to perform the follow-up CCT-scan as soon as possible.

In patients in the LD-group, CSF drainage via the lumbar drain is started slowly and steadily at a rate of approximately 5 ml per hour after the post-interventional CCT. This leads to a planned daily CSF-drainage of 120 ml per day through the lumbar route. Patients in both groups may receive additional CSF drainage via a ventricular device. The amount of CSF drained via the ventricular route is determined according to clinical requirement and not specified by the study protocol.

In order to enhance accuracy of the amount of CSF drained, regular drainage control every other hour and stopping in case of excess drainage is strongly recommended by the principal investigators. In case of neurological decline suspiciously related to the lumbar drainage, the drain must be closed immediately. Drainage may be gradually restarted after 12 to 24 hours, after performing a CCT scan.

If the post-procedural CCT or any other follow-up CCT scan shows compressed basal cisterns or any signs of threatening herniation, lumbar CSF diversion in the LD-group must not be performed. It may still be feasible to carefully drain CSF via the lumbar route [14], but this is at the discretion of the local investigator and not recommended. In patients requiring sedation and mechanical ventilation, either due to neurological impairment or for other medical reasons, intracranial pressure monitoring is mandatory. This may be performed according to local policy either with parenchymal or ventricular devices. If the intracranial pressure exceeds 20 mmHg, further CSF drainage via lumbar route shall be interrupted until the ICP is below 20 mmHg again. Careful CSF-drainage via the lumbar route may still be feasible in case of high intracranial pressure [14], but again this is at the discretion of the local investigator. A method of detecting the safety of lumbar CSF diversion may be determining the gradient of lumbar versus intracranial CSF pressure [15]. Albeit promising, the Earlydrain investigators explicitly rate this approach preliminary and experimental. In case of doubt, lumbar CSF diversion must not be performed.

Further neuromonitoring with transcranial duplex sonography (TCD), electroencephalography (EEG), brain tissue oxygenation recordings, jugular bulb oxymetry, regional cerebral blood flow measurement, microdialysis or other methods is at the discretion of the center and according to its local guidelines. As far as possible, this data should be saved electronically for post-hoc analysis.

A CCT scan as well as conventional digital subtraction angiography (DSA), CT angiography or MR angiography for assessment of vasospasm in the larger vessels is routinely performed on day 7 to 10 after the initial hemorrhage, according to local guidelines. In case of the occurrence of a DND when vasospasm is assumed to be the cause, angiography may be performed at any time. If it is performed earlier than day 7 to 10 and the patient shows no clinical deterioration thereafter, the angiography on day 7 to 10 is omitted. Treatment of radiographically confirmed vasospasm is according to local guidelines and not specified in the Earlydrain study protocol. It may include augmentation of cerebral blood flow via hypertensive hypervolemia as well as endovascular balloon dilation or intraarterial infusion of vasodilators.

After cerebrovascular imaging on day 7 to 10 the lumbar drainage of CSF is stopped in the LD-group. If cerebrovascular imaging is carried out before day 7, lumbar drainage is stopped on day 8. It may be pursued on a clinical base, as required.

Amount and duration of CSF drainage

Patients randomized to the lumbar drainage group shall receive a daily drainage of 120 ml CSF, or 5 ml per hour for seven days. If higher amounts of CSF need to be drained on clinical grounds as in patients with hydrocephalus, this is preferably performed via an external ventricular drain.

The drain is planned to remain in place until the control angiography on day 7 to 10 after the initial hemorrhage. The local investigator may decide to remove the drain earlier in patients fully mobilized without clinical necessity of CSF drainage. However, consecutive drainage should not be less than four days to achieve a valid study result. Lumbar CSF drainage may be prolonged beyond the control angiography on clinical requirement. The amount of CSF drainage may then be adjusted to clinical requirements and bears no further restriction.

Patients randomized to the control group should not receive a lumbar drain before the planned control angiography to be performed on day 7 to 10 after SAH. If the patient develops hydrocephalus, and no EVD was placed initially for CSF drainage, a lumbar drain may be installed at the discretion of the local investigator. These patients are analyzed in the intention-to-treat analysis, but are not suitable for the per-protocol analysis.

Consent to study participation

Consent to study inclusion is sought after explanation and agreement to a specific aneurysm treatment. Thus, patients capable of consenting to the aneurysm treatment will be informed about the study details themselves and may or may not agree to participate. If a patient is incapable of consenting to the proposed treatment, the legal representative should be informed on the conditions of treatment choices and afterwards, on the details of the EARLYDRAIN study. A patient may be randomized if the legal representative gives informed consent to the study, based on the presumed will of the patient. If neither the patient is capable of giving informed consent nor a legal representative is available in due time, an independent physician not involved in the patient's treatment nor in the trial may be asked for study approval. The latter option reflects a distinct characteristic of German law and the local ethic committee may or may not permit this.

This option was introduced into the consent procedure because the aforementioned retrospective data on lumbar drainage for treatment of aneurysmal SAH suggests a potentially beneficial effect of the measure for the patient. Therefore it shall not be categorically withheld from patients who are not capable of deciding whether to participate in the study or not and who do not have a legal representative.

However, in these cases of deferred consent, a legal representative needs to be established as soon as possible, according to German law. From our experience this legal procedure requires a time period from proposal to establishment of a legal representative of up to 72 hours, thus requiring the consent of an independent physician upfront for study inclusion. As soon as a legal representative is available and/or the patient is capable again to consent to the study, he or she must be asked to give informed consent. If the patient or his/her legal representative refuses consent after inclusion by advice of an independent physician, the patient's further study participation is no longer possible. In this case, however, the patient or his/her legal representative is asked to give consent for evaluation of already acquired data.

The detailed explanation of the study to the patient, legal representative or independent physician has to be carried out using appropriate explanations and words depending on the previous medical knowledge of the respective person and her/his level of education. During the explanations the respective person will be asked on a regular basis if she/he understands the conveyed information and if any questions have arisen. In addition to these verbal explanations the patient/legal representative/independent physician will be given a leaflet containing the study details. After reading the leaflet the respective person will be given as much time as she/he demands for the decision on study participation.

Randomization

Any patient meeting the inclusion criteria and not violating the exclusion criteria may participate in the EARLYDRAIN study and be randomized to either receive a lumbar drain or not, thus defining the two distinct groups LD and NoLD.

Randomization is performed via a dedicated internet site accessible for all local investigators of the participating trial centers http://www.randomizer.at. No stratification or minimisation is to be used. The security measures of the online randomization system "randomizer.at" include that 1. All transactions are logged, 2. The audit trail of the trial can be accessed and analysed any time by the trial monitoring committee. 3. Network traffic between the web-browser and the randomizer is encrypted using SSL (Secure Sockets Layer) with strong encryption.

Sample size calculation

In the ISAT trial, the largest trial on the treatment of aneurysmal subarachnoid hemorrhage so far, the mortality at one year follow-up was 8.1% to 10.1% [16]. Given the data from both retrospective studies on lumbar drains after SAH, a reduction from 15% to 2.1% after coiling and from 5% to 3% after clipping was shown. Thus, both studies were way lower in their mortality rate and, therefore, their external validity may be questioned.

In the two retrospective trials, 167 [12] and 107 [13] patients were studied, respectively. The effect of lumbar drainage was a decrease of the incidence of "clinical vasospasm" by 34% [12] and 40% [13], respectively.

In the above-mentioned studies the term "clinical vasospasm" includes neurological deterioration not explainable by hemorrhage, cerebral edema, hydrocephalus, hyponatremia, drug toxicity, infection or seizures. No distinction is made between delayed cerebral ischemia (DCI) and vasospasm as potential causes of the clinical worsening.

The following statistical calculations are based on the assumption that the clearly defined subtype of "delayed neurological deficit" measured in the above-mentioned two retrospective trials is highly correlated with clinical outcome 6 months after SAH, which constitutes the primary endpoint of the EARLYDRAIN study.

For lack of previous studies assessing clinical outcome after lumbar drainage in SAH as a primary endpoint this assumption seemed justified.

To assess a decrease of the incidence of DND from 40% to 20% in a prospective clinical trial, 93 patients in each of the two study arms are required to gain a power of 85%, using an alpha error of 5%. To account for possible imbalances in the randomization procedure concerning severity of clinical and radiological grading of the SAH or the choice of treatment and to facilitate a preplanned analysis on the severity of the initial hemorrhage, the planned study size is to include and randomize altogether 300 patients. This results in a power of 85.2%, again, using an alpha error of 5%, to detect a decrease in the rate of severe disability on a dichotomized modified Rankin scale from 50% to 33%, which would be consistent with the effect size from the retrospective trials on a dichotomized GOS [12, 13]. The EARLYDRAIN investigators are aware of other, more conservative calculations for the sample size of patient-centered outcome studies targeting vasospasm, indicating that there may be the necessity to include more than 5000 patients in a single trial [17]. The power calculations, as described above and based on the available retrospective data, do not substantiate numbers this large. Besides feasibility issues, clinical experience from the principal investigators considering the expected effort-benefit ratio does not warrant enlargement of the trial to detect a rather small difference between groups.

Safety of lumbar drains after aneurysmal SAH

In the above-mentioned two retrospective studies, mortality was lower in the lumbar drainage group. Neither of the retrospective studies mentions procedural related complications for the lumbar drains [12, 13]. In patients with increased intracranial pressure, careful lumbar drainage of CSF may be a possible treatment even in case of compressed basal cisterns [14]. A feasible strategy to enhance safety is determination of the lumbar-cranial pressure gradient and cessation of lumbar CSF diversion in patients with increasing pressure difference [15].

However, in patients presenting increased intracranial pressure or compressed basal cisterns on CCT-scan the risk associated with lumbar drainage is unclear according to the current state of medical knowledge. In unclear cases the investigator must refrain from the insertion of a lumbar drain.

Outcome assessment

The primary endpoint is disability after 6 months, assessed by the modified Rankin Scale [18] dichotomized at a score of 0 to 2 versus 3 to 6 (6 = death). Assessment is performed by a blinded investigator of the local study center by personal visit. Alternatively, a telephone questionnaire is suitable for outcome assessment using the modified Rankin Scale [19]. Outcome assessment is planned to be done on the whole dataset as well as in preplanned stratified subsets (i.e. for example clinical SAH grade according to the Hunt&Hess scale 1-2 vs. 3-5 [20], CT grading according to Fisher I-III vs. IV [21]).

Secondary outcome criteria are:

The following parameters will be recorded and used in predictor-/association models concerning primary and/or secondary outcome parameters:

Adverse events (AE) and severe adverse events (SAE)

Apart from AE and SAE which may occur after the beginning of the trial (synchronous with the insertion of the LD) there are complications related to securing the aneurysm:

Surgery-related complications: Surgical treatment includes the known risks of surgical interventions.

Complications related to endovascular therapy: Endovascular therapy includes the risks known to be associated with it.

Definition of adverse events and severe adverse events

The term "adverse event" (AE) describes any sign, symptom, syndrome or any disease 1. occurring newly in a trial participant after consent to the trial and 2. being of particular interest for the assessment of the disease or the security of the therapeutic concept. In this trial AEs include:

The term AE does not implicate a causal correlation with the participation in the trial. Surgical or endovascular interventions are not necessarily considered as AE but can be necessary for the therapy of an AE. AEs are divided in severe (SAE) and non-severe (AE) adverse events.

An SAE is any AE occurring during the trial that is related to:

Documentation

Investigation of AEs is part of every assessment of the study participants. Any AE has to be documented in the CRF.

Every SAE has to be documented on a special documentation form and has to be reported within 24 hours after recording, but at least at the next working day, to the data monitoring center in Berlin.

Statistical Analysis

All data are described according to their mean, median or frequency, as applicable. The dichotomized modified Rankin score as primary outcome variable is investigated in using univariate analysis. Multivariate logistic regression modeling is performed accordingly, adjusted for clinical grade, fisher grade, ventricular hemorrhage, parenchymal hemorrhage, gender, nimodipine or other concomitant medical treatment. Analysis is planned as intention-to-treat as well as per protocol, excluding the patients who were treated with amounts of CSF drainage via lumbar drain deviating from the specified 5 ml/h or which needed a lumbar drain when randomized to the No-LD group.

Interim Analysis

An interim analysis after inclusion of 10 patients will address safety issues. This analysis focuses on the secondary endpoints and SAEs only, especially the rate of death during hospital stay. During the interim analysis, the recruitment for the EARLYDRAIN study is not stopped. The frequency of further safety analyses will be adjusted to the recommendations of the data and safety monitoring board (DSMB).