RESEARCH—HUMAN—CLINICAL TRIALS
Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019
Jehuda Soleman, MD∗ ‡
Katharina Lutz, MD∗
Sabine Schaedelin, MSc§
Maria Kamenova, MD‡
Raphael Guzman, MD‡
Luigi Mariani, MD‡
Javier Fandino, MD∗
∗
Department of Neurosurgery, Kantons-
spital Aarau, Aarau, Switzerland; ‡ De-
partment of Neurosurgery, University
Hospital of Basel, Basel, Switzerland;
§
Clinical Trial Unit, University Hospital of
Basel, Basel, Switzerland
Parts of this paper’s results were
presented as a plenary presentation at
the Joint Annual Meeting of the Swiss
Society of Neurosurgery and Swiss
Society of Neuroradiology, September
28–30, 2016, Basel, Switzerland. The
results of this paper were presented as a
plenary presentation at the 18th
European Congress of Neurosurgery of
the European Association of
Neurosurgery, October 21–23, 2018,
Brussels, Belgium.
Correspondence:
Jehuda Soleman, MD,
Department of Neurosurgery,
University Hospital of Basel,
Spitalstrasse 21,
4031 Basel, Switzerland.
Email: jehuda.soleman@gmail.com
Received, September 5, 2018.
Accepted, February 26, 2019.
Copyright 
C 2019 by the
Congress of Neurological Surgeons
NEUROSURGERY
Subperiosteal vs Subdural Drain After Burr-Hole
Drainage of Chronic Subdural Hematoma:
A Randomized Clinical Trial (cSDH-Drain-Trial)
BACKGROUND: The use of a subdural drain (SDD) after burr-hole drainage of chronic
subdural hematoma (cSDH) reduces recurrence at 6 mo. Subperiosteal drains (SPDs) are
considered safer, since they are not positioned in direct contact to cortical structures,
bridging veins, or hematoma membranes.
OBJECTIVE: To investigate whether the recurrence rate after insertion of a SPD is nonin-
ferior to the insertion of a more commonly used SDD.
METHODS: Multicenter, prospective, randomized, controlled, noninferiority trial analyzing
patients undergoing burr-hole drainage for cSDH aged 18 yr and older. After hematoma
evacuation, patients were randomly assigned to receive either a SDD (SDD-group) or a
SPD (SPD-group). The primary endpoint was recurrence indicating a reoperation within 12
mo, with a noninferiority margin of 3.5%. Secondary outcomes included clinical and radio-
logical outcome, morbidity and mortality rates, and length of stay.
RESULTS: Of 220 randomized patients, all were included in the final analysis (120 SPD and
100 SDD). Recurrence rate was lower in the SPD group (8.33%, 95% confidence interval [CI]
4.28-14.72) than in the SDD group (12.00%, 95% CI 6.66-19.73), with the treatment difference
(3.67%, 95% CI -12.6-5.3) not meeting predefined noninferiority criteria. The SPD group
showed significantly lower rates of surgical infections (P = .0406) and iatrogenic morbidity
through drain placement (P = .0184). Length of stay and mortality rates were comparable
in both groups.
CONCLUSION: Although the noninferiority criteria were not met, SPD insertion led to
lower recurrence rates, fewer surgical infections, and lower drain misplacement rates.
These findings suggest that SPD may be warranted in routine clinical practice
KEY WORDS: Burr-hole drainage, Chronic subdural hematoma, Subdural drain, Subperiosteal drain, Traumatic
brain injury
Neurosurgery 0:1–10, 2019 DOI:10.1093/neuros/nyz095 www.neurosurgery-online.com
C
hronic subdural hematoma (cSDH) establishing and implementing effective patient
is one of the most common neuro- management is essential.
surgical entities which effects mostly Surgical evacuation of the hematoma via
elderly people and is associated with substantial burr-hole drainage is generally the treatment of
morbidity and mortality1-3 ; a trend towards choice for patients with symptomatic cSDH.2,5,6
higher incidence rates is expected.4 Therefore, Recurrence rate after burr-hole drainage of
cSDH is estimated at approximately 10%; cSDH
recurrence causes considerable morbidity, and
ABBREVIATIONS: CRF, case report form; cSDH, is therefore a major focus of research.3,4,7 A
chronic subdural hematoma; CT, chronic subdural
hematoma; GCS, Glasgow Coma Scale; GOS,
randomized controlled trial by Santarius and
Glasgow Outcome Scale; mRS, modified Rankin colleagues8 demonstrated that the insertion of a
Scale; MLS, midline shift; SDD, subdural drain; SPD, subdural drain (SDD) left in situ after burr-hole
Subperiosteal drain evacuation significantly reduces cSDH recur-
rence requiring redrainage (9.3% vs 24.0%
Supplemental digital content is available for this article at
www.neurosurgery-online.com. recurrence with vs without drain), and improves
functional outcomes. These SDDs are positioned
VOLUME 0 | NUMBER 0 | 2019 | 1
SOLEMAN ET AL
between the dura and the cortex (underneath the calvarium), in
proximity to the cortical surface, bridging veins, and hematoma
membranes, which may lead to iatrogenic injury of these struc-
tures and postoperative morbidity.9,10 Therefore, some authors
have advocated the insertion of subperiosteal drains (SPD),
which are positioned between the periosteum and the calvarium,
whereby intracranial bleeds, drains placed accidentally within
the cortex, or seizures can be averted.9-16 However, further
evidence from randomized controlled trials is needed to guide
treatment.3,5,17,18
Most surgeons still prefer the insertion of SDD over SPD
after burr-hole drainage of cSDH,5 presumably due to a lack of
experience or knowledge of the correct positioning of SPD, and
insufficient evidence of the procedure.
The aim of this prospective study was to investigate whether
the recurrence rate after insertion of an SPD is non-inferior to the
insertion of an SDD in patients undergoing surgical evacuation
of cSDH through burr-hole drainage.
METHODS
Trial Design
The study was a prospective, randomized, controlled, multicenter,
non-inferiority, clinical trial in patients with symptomatic cSDH
indicating surgical evacuation. We recruited patients presenting to the
emergency department of two neurosurgical clinics. The trial was done
and analyzed according to the CONSORT guidelines, and ethical
approval was obtained from the local ethics committee.
Trial Registration
The trial is registered with ClinicalTrials.gov, number
NCT01869855. The study protocol has been published and is available
online.19
Participants
Eligible patients were female or male over the age of 18 yr presenting
to one of the centers with a symptomatic cSDH diagnosed by computed
tomography (CT) and/or magnetic resonance imaging, indicating
surgical evacuation of the hematoma. Written informed consent of the
patient or the next-of-kin (in comatose or non-judicious patients) was
obtained by a member of the neurosurgical staff prior to randomization.
Patients with other surgery indications (eg, craniotomy due to subacute
hematoma, membranes, etc), with cSDHs caused by a previous under-
lying condition (eg, over drainage of a ventriculoperitoneal shunt), or
without informed consent were excluded.
Interventions
All patients received two-burr-hole trephination; a drain (SPD or
SDD; according to the arm of the study to which the patient was
randomized) was then inserted without suction. The surgical procedure
was standardized for both institutions. Prophylactic anticonvulsant
medication was not applied, while all patients received a single shot
of antibiotics (Cefuroxim) 30 min prior to skin incision. We started
with supine positioning of the patients on a horseshoe headrest. The
frontal and parietal areas of the head were shaved, and patients were
draped. After skin incision, we drilled two 13 mm burr-holes, usually
2 | VOLUME 0 | NUMBER 0 | 2019
about 7 to 8 cm apart from each other over the maximum width of
the hematoma. The dura mater was opened with a cruciate incision and
coagulated. The subdural hematoma was then washed out with warm
Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019
saline, at the surgeon’s discretion with or without a soft catheter. Once the
surgeon had completed the irrigation and was ready for drain insertion,
the randomization envelope was opened and the assigned drain (SDD
or SPD) was inserted. The SDD was inserted from the parietal burr-
hole in frontal direction reaching the frontal burr hole under visual
control (Figure 1). The SPD was inserted subperiosteally and placed
over both burr-holes, without sealing them off with any kind of material
(eg, PDS-film, Spongostan) in order to allow communication between
the subdural space and the SPD. A Jackson-Pratt drainage connected
to a fluid drainage gravity system was routinely inserted for 48 h in
all cases, apart from two cases where the surgeon decided to insert
a slimmer ventriculostomy catheter (Figure 1). Bilateral hematomas
were treated as one case; both sides received the same treatment.
Patients with crossover treatment (eg, a patient was randomized to
the SDD group but the surgeon felt it was unsafe to insert an SDD
because of potential brain injury and inserted an SPD instead) were
noted in the case report form (CRF) and were not excluded from the
study. If the surgeon intraoperatively identified the need to perform
a craniotomy (eg, due to clotted hemorrhage which did not evacuate
sufficiently through burr-hole drainage), the patient was excluded
from the trial.
Outcomes
Primary outcome was recurrence of cSDH indicating revision surgery
within 12 mo of the original drainage procedure. Recurrence was defined
as occurrence of symptoms attributable to an ipsilateral hematoma seen
on CT scan. The indication for reoperation was established by the
admitting consultant neurosurgeon, based on the CT and the patients’
neurological condition. Secondary outcome measurements included
medical and surgical complications, reoperation rates (other than surgery
for recurrent cSDH), mortality, clinical outcome measured by Glasgow
Coma Scale (GCS), modified Rankin Scale (mRS), Glasgow Outcome
Scale (GOS), and Markwalder score, length of hospital stay, and radio-
logical characteristics of postoperative CT images including hematoma
width, midline shift (MLS), drain misplacement, and intracranial hemor-
rhage other than recurrent cSDH. All of the outcome measures were
centrally assessed.
Patient data were prospectively collected and registered on CRFs.
Demographic data such as age and sex, and clinical data such as
prescribed blood thinners, medical history, GCS, and neurological
condition at admission were collected on the CRF prior to surgery.
Hematoma size (measured at the thickest hematoma area on axial
CT slices in mm), affected side, MLS (measured at the level of the
foramina of Monro in mm), and presence/absence of hydrocephalus
were also recorded on the CRF. Documented intraoperative variables
included type of drain randomized, type of drain inserted, reason
for group crossover (if present), priority of the procedure (elective
or emergency), type of hematoma intraoperatively (chronic, subacute,
acute), presence/absence of membranes, number of membrane layers
(if present), and functioning communication between the 2 burr-holes.
Patients received clinical follow-up examination 24 h after surgery, at
discharge, as well as 6 wk and 12 mo after surgery. Clinical follow-up
variables included GCS scores, mRS scores, GOS scores, Markwalder
scores, neurological condition (improved, unchanged, or worse), length
of hospital stay, reoperations, complications, and mortality. Radiological
www.neurosurgery-online.com

FIGURE 1. Illustration showing the insertion of a subdural drain (on the left) and a subperiosteal drain
(on the right) after burr-hole drainage of a chronic subdural hematoma.
reassessment was conducted 24 h and 6 wk after surgery by cranial CT;
we documented hematoma size, MLS, drain misplacement (defined as a
drain penetrating the brain parenchyma or with clear contact to the brain,
with or without intracranial bleeding and/or neurological deficits), and
intracranial bleeding.
Sample Size
We set the threshold for non-inferiority of SPD for the primary
outcome at less than 3.5% difference from SDD in the percentage of
recurrence after burr-hole drainage of cSDH. This limit was based on
the clinical relevance assessed by our study group. When estimating the
sample size N by means of a sampling procedure, the recurrence rate
of cSDH was estimated at 7%, while a difference of 10% between the
2 groups was expected. Each sample size, Ni = 1; …; 101 = 50; …; 150,
was evaluated by sampling 9999 times Ni individual samples based on
the assumptions described above. We calculated that if 220 patients
were to be randomized ensuring 208 evaluable patients (ie, 110 patients
randomized per study arm), considering an overall drop-out rate of
5% after randomization (eg, death, lost to follow-up etc), we would
ensure with 80% power (1 –β = 0.8; ie, in 80 of 100 hypothetical
repetitions of the study) the estimation of a 95% confidence interval
which is entirely below the predefined non-inferiority margin of 3.5%.
Confidence intervals for the difference between proportions were calcu-
lated using a continuity-corrected modification of the Wilson’s score
method.20
NEUROSURGERY
CSDH-DRAIN TRIAL
Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019
Randomization
To ensure allocation concealment, we performed randomization with
blocks of 30 in a 1:1 allocation ratio by a web-based independent
randomization software (Random Allocation version 1.0). Patients were
randomly assigned to receive SDD or SPD after burr-hole drainage of
cSDH. Instruction on which drain (SDD or SPD) was to be used was
kept in sealed opaque envelopes labeled with sequential study numbers
and opened at surgery, right before the insertion of the drain.
Blinding
This randomization intervention did not allow for blinding of
treatment allocation. However, data was encoded and clinicians were
blinded to outcomes when possible.
Statistical Methods
Analysis was done on per protocol basis to include all randomized
patients with assessable data in all analyses. Due to the noninferiority
design of the study, sensitivity intention to treat analysis was undertaken
according to the ICHE9 Guidelines for a non-inferiority designed trial
(Tables, Supplemental Digital Content 1-4).21,22 In addition, major
protocol deviations were solely due to a change in type of drain insertion.
Therefore, the main analysis was performed according to the per-protocol
principle. The analysis was repeated according to the intention-to-treat
principle as sensitivity analysis. The difference in recurrence rates
was presented with 95% confidence interval, estimated using a
VOLUME 0 | NUMBER 0 | 2019 | 3
SOLEMAN ET AL
FIGURE 2. Trial profile.
continuity-corrected modification of Wilson’s score method.20 Outcome
scores were analyzed using a 2-sided t-test and are presented as
means and standard deviation. In case of violation of the normalized
assumption, median and interquartile range were calculated, while
proportions were compared using the Mann–Whitney U-test.
Categorical data are presented as frequencies and proportions and
compared using the chi-square or fisher’s exact test (in case of cell count
under 5).
All statistical analyses were done using R (Comprehensive R Archive
Network [CRAN], R Foundation for Statistical Computing, Vienna,
Austria, Version 3.2.2). The trial was overseen by an independent data
monitoring committee.
Funding Source
The funder of the study had no role in study design, data collection,
data analysis, data interpretation, or writing of the report. The corre-
4 | VOLUME 0 | NUMBER 0 | 2019
Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019
sponding author had full access to all data in the study and had final
responsibility for the decision to submit for publication.
RESULTS
A total of 220 patients were recruited between April 15,
2013, and December 9, 2015. The last follow-up examination
took place on December 6, 2016. A total fo 170 patients
were randomly assigned to receive SDD, 113 to receive SPD;
all patients contributed data for the primary outcome analysis
(Figure 2). Due to crossover of 14 patients from the SDD to
the SPD group and seven patients from the SPD to the SDD
group, the per protocol analysis consisted of 120 patients in the
SPD group and 100 patients in the SDD group. Demographic,
clinical, and radiological baseline characteristics are presented in
Table 1.
www.neurosurgery-online.com
 CSDH-DRAIN TRIAL

measures and the distribution of patients showing neurological

TABLE 1. Demographic, Clinical, and Radiological Baseline Charac- improvement at 24 h, discharge, 6 wk, and 12 mo after surgery
teristics of the Per Protocol Population were comparable in both groups (Table 2). While overall surgical

Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019

Variable SDD (n = 100) SPD (n = 120) morbidity and mortality rates were equally distributed amongst
the two groups, the surgical infection rate was significantly
Age 81.0 (74.0-85.0) 78.0 (70.7-83.2) lower in the SPD group than in the SDD group (n = 3,
Sex = female 32 (32%) 39 (32%)
2.5% vs n = 9 [9 infections in 8 patients], 9%; P = .0406;
Medical history
Diabetes mellitus 15 (15%) 15 (12%) Table 3). A wound infection was seen in 6 (6%) cases in the SDD
Hypertension 52 (52%) 64 (53%) group, as opposed to 2 (1.7%) in the SPD group. Postoperative
Coronary arterial disease 31 (31%) 32 (27%) seizure frequency, postoperative bleeds (such as acute subdural
Stroke/TIA 11 (11%) 16 (13%) or epidural hematoma or subarachnoid hemorrhage), medical
Atrial fibrillation 22 (22%) 24 (20%) complication rates, and new postoperative neurological deficits
Deep venous thrombosis 2 (2%) 6 (5%)
showed no significant differences between the 2 groups (Table 3).
Pulmonary embolism 3 (3%) 5 (4%)
Chronic renal failure 14 (14%) 15 (12%)
Misplacement of the drainage was seen in 17 patients (17%) in
Livercirrhosis 2 (2%) 1 (1%) the SDD group, as opposed to no misplacements in the SPD
Dementia 15 (15%) 11 (9%) group (P = .0001; Table 4). All other radiological measurements
Adipositas 2 (2%) 5 (4%) 24 h and 6 wk after surgery showed no significant difference
Other 48 (48%) 50 (42%) between the 2 groups (Table 4).
None 8 (8%) 14 (11·7%)
Blood thinners 65 (65%) 68 (57%)
Recurrent hematoma 7 (7%) 9 (7%)
History of trauma 29 (29%) 25 (21%)
GCS at admission DISCUSSION
Median [IQR] 14.0 (14.0-15.0) 15.0 (14.0-15.0)
GCS 14–15 81 (81%) 98 (82%) To date this is the largest RCT that compares recurrence rates
GCS 9–13 15 (15%) 20 (17%) of surgically drained cSDHs after the insertion of SPD or SDD.
GCS 3–8 4 (4%) 2 (2%) The mean recurrence rate in the SPD group was 8.3%, which is
Symptoms at presentation
3.7% lower than the recurrence rate of 12.0% in the SDD group.
Speech impediment 24 (24%) 23 (19%)
Limb weakness 52 (52%) 56 (47%) The 95% CI of difference in recurrence rates between the 2 groups
Sensory weakness 5 (5%) 9 (7%) was clearly in favor of the SPD group (12.6% in favor of SPD and
Incontinence 2 (2%) 4 (3%) 5.3% in favor of SDD), fell however just short from the rather
Headache 29 (29%) 38 (32%) strictly set noninferiority margin of 3.5%. However, our data
Gait disturbance 38 (38%) 37 (31%) shows good evidence that the insertion of an SPD causes at most
Mental deterioration 49 (49%) 54 (45%) 5.3% more recurrences compared to the insertion of an SDD. In,
Seizure 4 (4%) 8 (7%)
Coma 4 (4%) 2 (2%)
addition, patients treated with an SDD had significantly more
None 1 (1%) 3 (2%) surgical infections and brain injuries through misplaced SDDs.
Midline shift in mm 7.0 (4.0-10.5) 6.0 (4.0-10.0) Therefore, the results of our study support the use SPD after burr-
Hematoma width in mm 24.9 (9.1) 22.5 (9.9) hole drainage of cSDH.
Hematoma type Our findings are in accordance with results from two
Chronic 56 (56%) 65 (53%) prospective and some retrospective studies assessing the efficacy of
Acute on chronic 29 (29·0%) 29 (24%)
SPD vs SDD in the management of cSDHs.9,10,14,15,17,18 Most
Subacute 15 (15%) 25 (21%)
Hygroma 0 (0%) 1 (1%)
authors emphasize that SPD is less invasive, safe, and technically
Hematoma side easy.9,10,12,13,16
Right 37 (37%) 44 (37%) Overall surgical morbidity and overall mortality did not signif-
Left 39 (39%) 49 (41%) icantly differ between the SPD and SDD groups in our series.
Bilateral 23 (23%) 27 (22%) To date, only few retrospective studies have analyzed morbidity
Data are number (%) and median (IQR). SDD: subdural drain; SPD: subperiosteal drain; and mortality depending on the type of drain inserted after burr-
GCS: Glasgow coma scale hole drainage of cSDH.9,10,14,17 Similar to our observation, no
significant differences in surgical morbidity and overall mortality
Recurrence rate of cSDH at 12 mo was lower in the SPD were found. In contrast to these studies, our data showed signif-
group (n = 10, 8.33%, 95% CI 4.28-14.27) than in the SDD icantly more surgical infections and injury of brain parenchyma
group (n = 12, 12.0%, 95% CI 6.66-19.73). The 95% CI of the through misplaced drains in the SDD group. These results are
percentage difference in recurrence after burr-hole drainage of not surprising considering that SDD is more invasive than SPD,
cSDH between the groups (-12.6-5.3) exceeded the predefined which automatically leads to a higher risk of iatrogenic brain
noninferiority margin of 3.5% (Figure 3). Clinical outcome injury; furthermore, the SDD creates direct contact between the

NEUROSURGERY VOLUME 0 | NUMBER 0 | 2019 | 5

SOLEMAN ET AL

Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019

FIGURE 3. Primary endpoint at 12 mo follow-up.

drain insertion point and the subdural space, thus elevating the
TABLE 2. Clinical Outcomes of the Per Protocol Population risk of deep infections. The reason for the higher rate of super-
ficial infections in the SDD group might be due to the fact,
SDD SPD
Variable (n = 110) (n = 120) P-value
that the insertion of an SDD is sometimes difficult leading to
more manipulation of the drain itself, also at the level of the
Length of stay in days 6.0 (5·0-8·0) 6.0 (5.0-9.2) .918 skin, which might influence superficial infection rates. Iatrogenic
GCS 14–15 brain injury can be devastating, leading to large bleedings in
24 h 91 (91%) 104 (82%) .576 the brain parenchyma, causing neurological deficits, seizures, and
Discharge 92 (93%) 108 (91%) .646 even death (Figure 4). Drain insertion techniques and types of
6 wk 87 (98%) 110 (97%) .40
12 mo 78 (99%) 93 (98%) 1
drains used vary from institution to institution. In our insti-
GOS 4–5 tutes, we place the SDD to the full length between the parietal
24 h 84 (84%) 96 (80%) .555 and the frontal burr hole within the subdural space, in order
Discharge 84 (84%) 102 (86%) .889 to achieve better drainage from within the subdural hematoma
6 wk 78 (87%) 97 (86%) 1 cavity (Figure 1). Others might only place the tip of the drain into
12 mo 65 (83%) 82 (86%) .74 the parietal burr hole to minimize the risk of brain injury. The
mRS 0–3
type of drain inserted into the subdural space might also affect
24 h 86 (86%) 97 (81%) .401
Discharge 88 (89%) 107 (91%) .834
brain injury, since a ventriculostomy tube, which is considered
6 wk 83 (93%) 100 (89%) .465 less traumatic and is slimmer in size than the Jackson-Pratt drain,
12 mo 73 (92%) 87 (92%) 1 might cause less iatrogenic brain damage. This might be the case;
Markwalder score 0–1 however, in our cohort one of the misplaced drains was a ventricu-
24 h 70 (70%) 82 (68%) .905 lostomy drain, showing that drain misplacement is probably
Discharge 72 (73%) 92 (78%) .462 independent of the drain type used. Interestingly, although SDD
6 wk 79 (88%) 100 (88%) 1
patients suffered more surgical infections and iatrogenic brain
12 mo 66 (83%) 87 (92%) .166
Improvement of injuries, it did not seem to affect outcome measurements (mRS,
neurological exam GOS, Markwalder score) or mortality. This might be due to the
24 h 87 (87%) 94 (85%) .696 fact that eight of the 12 surgical infections were superficial wound
Discharge 88 (88%) 109 (93%) .282 infections not influencing outcome or mortality, and only 3 of the
6 wk 69 (77%) 95 (86%) .205 5 iatrogenic bleeds caused by SDD led to neurological symptoms
12 mo 61 (78%) 68 (72%) .479 that influence the outcomes. Nevertheless, our data show a signif-
Data are number (%) and median (IQR). SDD: subdural drain, SPD: subperiosteal drain; icant benefit for the insertion of SPD for the prevention of
GCS: Glasgow coma scale; GOS: Glasgow outcome scale; mRS: modified ranking scale. surgical infections and iatrogenic brain injuries, while the fact
that this did not influence the outcomes might be random.

6 | VOLUME 0 | NUMBER 0 | 2019 www.neurosurgery-online.com

 CSDH-DRAIN TRIAL

TABLE 3. Morbidity and Mortality of the Per Protocol Population TABLE 4. Radiological Outcome of the Per Protocol Population

Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019

Variable SDD (n = 100) SPD (n = 120) P-value SDD SPD
Variable (n = 100) (n = 120) P-value
Surgical infections 9 (9%)μ 3 (2%)α .0406
24 h 0 (0%) 0 (0%) Midline shift in mm
Discharge 4 (4%) 0 (0%) 24 h 3.0 (1.0-5.0) 3.0 (0.0-5.0) .629
6 wk 4 (4%) 3 (3%) 6 wk 0.0 (0.0-1.0) 0.0 (0.0-0.0) .388
12 mo 1 (1%) 0 (0%) Remnant hemorrhage in mm
Leading to surgery 7 (7%) 3 (2%) 24 h 13.0 (9.0-17.0) 11.5 (7.7-17.0) .403
Seizures 7 (7%) 11 (9%) .7363 6 wk 6.0 (0.0-11.0) 5.0 (0.0-11.0) .549
24 h 0 (0%) 0 (0%) Misplaced draina
Discharge 5 (5%) 4 (3%) Total 17 (17%) 0 (0%) .0001
6 wk 1 (1%) 4 (3%) Causing hemorrhage 5 (5%) 0 (0%) .0184
12 mo 1 (1%) 3 (3%) Causing neurological deficit 3 (3%) 0 (0%) .0924
Postoperative bleeds 10 (10%) 10 (8%) .8472
Data are number (%) and median (IQR). SDD: subdural drain; SPD: subperiosteal drain,
aSDH 5 (5%) 7 (6%)
h: hours, mm: millimeters
IPH 2 (2%) 2 (2%) a
Defined as drain within the brain parenchyma or with clear contact to the brain
aEDH 1 (1%) 0 (0%) Italics: significant values
SAH 2 (2%) 1 (1%)
Leading to surgery 1 (1%) 2 (2%)
Medical complications 18 (18%)∞ 20 (17%)× .859
24 h 2 (2%) 4 (3%)
Discharge 10 (10%) 11 (9%)
within the groups. The management of discontinuation and
6 wk 2 (2%) 4 (3%) resumption of these medications were standardized for both
12 mo 4 (4%) 1 (1%) groups. Analysis of patients treated with blood thinners showed
New neurological deficit 9 (9%)β 10 (8%)δ 1 no significant difference in recurrent rates of cSDH between the
24 h 2 (2%) 2 (2%) two drain groups.
Discharge 2 (2%) 5 (4%) Crossover from the SDD to the SPD group (13.08%) was
6 wk 4 (4) 2 (2%)
approximately twice as likely as from the SPD to the SDD
12 mo 1 (1%) 1 (1%)
Overall surgical morbidity∗ 26 (26%) 24 (20%) .335
groups (6.19%). All cases of crossover from the SDD to the SPD
Mortality 10 (10%) 12 (10%) 1 group were due to intraoperative brain expansion whereby the
treating neurosurgeon felt it was too risky to place a drain into
Data are number (%). SDD: subdural drain; SPD: subperiosteal drain; h: hours; NA: not
applicable; aSDH: acute subdural hematoma; IPH: intraparenchymal hematoma; aEDH:
the subdural space. Cases with crossover from SPD to SDD were
acute epidural hematoma; SAH: subarachnoid hemorrhage· due to minimal brain expansion during surgery, leading to a large
μ
includes 6 wound infections and 3 subdural empyema in 8 patients subdural space which led the treating physician to decide the
α
includes 2 wound infections and 1 subdural empyema in 3 patients insertion of an SDD was beneficial for the patient. Based on a
∞
18 complications in 15 patients
×
20 complications in 18 patients
recently published international survey, the majority of neurosur-
β
9 new neurological deficits in 8 patients geons who do not insert a drain, do so if they feel the insertion
δ
10 new neurological deficits in 8 patients of SDD is too risky.5 The cross over rates from the SDD to the
∗
includes surgical infections, seizures and postoperative bleeds (excluding recurrent
SPD group in our study reflect the reluctance of neurosurgeons
cSDH)
Italics: significant values to insert a drain when it seems unsafe, while on the other hand we
provide firm evidence that SPD is a valid and effective alternative
to SDD. Therefore, and in light of the evidence based benefit of
drain insertion after burr-hole drainage of cSDH,8 the routine
Contrariwise to our data, Kaliaperumal et al18 showed signifi- insertion of SPD should be considered.
cantly better outcomes at 6 mo (measured by mRS) in the SPD
group when compared to the SDD group. As expected, overall Limitations
duration of hospital stay was similar in both groups. Surpris- The number of crossovers within the groups is a limitation
ingly, although 17.0% of the patients in the SDD group showed of our study; however, it reflects clinical reality and one of the
misplaced drains (compared to 0% in the SPD group), seizure many dilemmas neurosurgeons are confronted with when treating
rates did not differ significantly between the two groups. This patients with cSDH. Therefore, we decided not to exclude
finding might relate to the fact that some seizures were missed, crossover patients, and undertook a sensitivity intention to treat
since patients were monitored clinically and not through postop- analysis which revealed similar results. An additional limitation
erative electroencephalogram. of our trial is that surgeons were not masked for treatment.
Overall 60% of the study cohort patients were treated with Clearly in a surgical study masking the surgeon for treatment is
blood thinners, while these patients were equally distributed almost impossible. Whenever possible, the patients were masked

NEUROSURGERY VOLUME 0 | NUMBER 0 | 2019 | 7

SOLEMAN ET AL

Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019

FIGURE 4. Postoperative axial computed tomography demonstrating the insertion of a subdural drain after
burr-hole drainage of a chronic subdural hematoma into the brain parenchyma leading to intracranial
bleeding and neurological deficits in a study patient.

to the treatment arm, and the treating surgeons were masked to
outcomes. Lastly, our estimated recurrence rate of 7.0% was lower
than the actual overall recurrence rate of 10%. This might have
influenced the statistical power of our sample size and results.
Features such as being the largest trial to date analyzing recur-
rence rates after insertion of SPD as compared to SDD, being
a prospective, randomized, multicenter study involving a rather
big number of patients and neurosurgeons, a low rate of patients
lost to follow up (1.36%), and a long follow up of 12 mo are
important strengths of our trial.
comparable to SDD, while surgical infection rates and the occur-
rence of iatrogenic brain injuries are significantly reduced. These
findings suggest that SPD can be used in routine clinical practice.
Disclosures
Funded by the Research Foundation Kantonsspital Aarau. The funder of the
study had no role in study design, data collection, data analysis, data interpre-
tation, or writing of the report. The corresponding author had full access to all
data in the study and had final responsibility for the decision to submit for publi-
cation. The authors have no personal, financial, or institutional interest in any of
the drugs, materials, or devices described in this article.

CONCLUSION
REFERENCES
We conclude that the insertion of SPD after burr-hole drainage 1. Markwalder TM, Steinsiepe KF, Rohner M, Reichenbach W, Markwalder H. The
of cSDH is an efficient and safe alternative to the insertion of course of chronic subdural hematomas after burr-hole craniostomy and closed-
an SDD. Recurrence rates of cSDH after inserting an SPD are system drainage. J Neurosurg. 1981;55(3):390-396.

8 | VOLUME 0 | NUMBER 0 | 2019 www.neurosurgery-online.com


2. Weigel R, Schmiedek P, Krauss JK. Outcome of contemporary surgery for
chronic subdural haematoma: evidence based review. J Neurol Neurosurg Psychiatry.
2003;74(7):937-943.
3. Soleman J, Taussky P, Fandino J, Muroi C. Evidence-Based Treatment of Chronic
Subdural Hematoma. In: Sadaka D, ed. Traumatic Brain Injury. IntechOpen,
2014; DOI: 10.5772/57336. https://www.intechopen.com/books/traumatic-brain-
injury/evidence-based-treatment-of-chronic-subdural-hematoma. Accessed March
18, 2019.
4. Baechli H, Nordmann A, Bucher HC, Gratzl O. Demographics and prevalent risk
factors of chronic subdural haematoma: results of a large single-center cohort study.
Neurosurg Rev. 2004;27(4):263-266.
5. Soleman J, Kamenova M, Lutz K, Guzman R, Fandino J, Mariani L. Drain
insertion in chronic subdural hematoma: an international survey of practice. World
Neurosurg. 2017;104:528-536.
6. Soleman J, Nocera F, Mariani L. The conservative and pharmacological
management of chronic subdural haematoma. Swiss Med Wkly. 2017;147:w14398.
7. Almenawer SA, Farrokhyar F, Hong C, et al. Chronic subdural hematoma
management. Ann Surg. 2014;259(3):449-457.
8. Santarius T, Kirkpatrick PJ, Ganesan D, et al. Use of drains versus no drains after
burr-hole evacuation of chronic subdural haematoma: a randomised controlled
trial. Lancet North Am Ed. 2009;374(9695):1067-1073.
9. Chih AN, Hieng AW, Rahman NA, Abdullah JM. Subperiosteal drainage versus
subdural drainage in the management of chronic subdural hematoma (a compar-
ative study). Malays J Med Sci. 2017;24(1):21-30.
10. Oral S, Borklu RE, Kucuk A, Ulutabanca H, Selcuklu A. Comparison of subgaleal
and subdural closed drainage system in the surgical treatment of chronic subdural
hematoma. North Clin Istanb. 2015;2(2):115-121.
11. Abouzari M, Rashidi A, Rezaii J, et al. The role of postoperative patient posture
in the recurrence of traumatic chronic subdural hematoma after burr-hole surgery.
Neurosurgery. 2007;61(4):794-797; discussion 797.
12. Gazzeri R, Galarza M, Neroni M, Canova A, Refice GM, Esposito S. Continuous
subgaleal suction drainage for the treatment of chronic subdural haematoma. Acta
Neurochir (Wien). 2007;149(5):487-493; discussion 493.
13. Zumofen D, Regli L, Levivier M, Krayenbuhl N. Chronic subdural hematomas
treated by burr hole trepanation and a subperiostal drainage system. Neurosurgery.
2009;64(6):1116-1122; discussion 1121-1112.
14. Sjavik K, Bartek J, Jr., Sagberg LM, et al. Assessment of drainage techniques
for evacuation of chronic subdural hematoma: a consecutive population-
based comparative cohort study. J Neurosurg. 2017:1-7. https://doi.org/
10.3171/2016.12.jns161713.
15. Ishfaq A. Outcome in chronic subdural hematoma after subdural vs. subgaleal
drain. J Coll Physicians Surg Pak. 2017;27(7):419-422.
16. Yadav YR, Parihar V, Chourasia ID, Bajaj J, Namdev H. The role of subgaleal
suction drain placement in chronic subdural hematoma evacuation. Asian J
Neurosurg. 2016;11(3):214-218.
17. Bellut D, Woernle CM, Burkhardt JK, Kockro RA, Bertalanffy H, Krayenbuhl
N. Subdural drainage versus subperiosteal drainage in burr-hole trepanation for
symptomatic chronic subdural hematomas. World Neurosurg. 2012;77(1):111-118.
18. Kaliaperumal C, Khalil A, Fenton E, et al. A prospective randomised study to
compare the utility and outcomes of subdural and subperiosteal drains for the
treatment of chronic subdural haematoma. Acta Neurochir. 2012;154(11):2083-
2089; discussion 2088–2089.
19. Soleman J, Lutz K, Schaedelin S, Mariani L, Fandino J. Use of subperiosteal
drain versus subdural drain in chronic subdural hematomas treated with burr-
hole trepanation: study protocol for a randomized controlled trial. JMIR Res Protoc.
2016;5(2):e38.
20. Newcombe RG. Interval estimation for the difference between independent
proportions: comparison of eleven methods. Statist. Med. 1998;17(8):873-890.
21. Piaggio G, Elbourne DR, Pocock SJ, Evans SJ, Altman DG, Group C. Reporting
of noninferiority and equivalence randomized trials: extension of the CONSORT
2010 statement. JAMA. 2012;308(24):2594-2604.
22. Agency EM. ICH Topic E 9 Statistical Principles for Clinical Trials. 1998;
www.ema.europa.eu/docs/en_GB/…/09/WC500002928.pdf. Accessed August
28, 2018.
Acknowledgments
We thank all our patients for participating in the trial, all of our neurosur-
gical colleagues who assisted with patient recruitment, operating the patients,
NEUROSURGERY
CSDH-DRAIN TRIAL
and following up on them, thereby contributing to the completeness of the data;
Marianne Schulz from the research office of the University Hospital in Basel,
Switzerland and Anna Scrowther, Carline Perren, Sonja Schwenne, and Sylvie
Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019
Nuc from the Neurological Research Office (NRO) of the Kantonsspital in Aarau,
Switzerland for data collection, data management, and their administrative help;
Dr Selina Ackermann for the medical editing of this manuscript.
Supplemental digital content is available for this article at
www.neurosurgeryonline.com.
Supplemental Digital Content 1. Table. Demographic, clinical, and radiological
baseline characteristics of the intention to treat population.
Supplemental Digital Content 2. Table. Clinical outcome of the intention to
treat population.
Supplemental Digital Content 3. Table. Morbidity and mortality of the
intention to treat population.
Supplemental Digital Content 4. Table. Radiological outcome of the intention
to treat population.
COMMENTS
T he authors performed a randomized controlled trial (RCT)
comparing treatments for chronic subdural hematomas. Patients
were randomized to receive either placement of a subdural drain (SDD)
or subperiosteal drain (SPD). Each patient (and side, in bilateral cases)
received 2 burr holes. Following data analysis, the authors conclude
that placement of SPD results in fewer recurrences and fewer compli-
cations. The latter is not surprising, although the former is somewhat
surprising. Statistical analysis was performed. The SDD group appears to
have had these drains inserted into the subdural space to a rather signif-
icant distance, resulting in extensive contact between the drains and the
cortex (or subdural membranes). I am not certain that this is a widespread
practice, but the paper illustrates issues with that method, and why the
SPD system appears to be superior. Some people insert ventriculostomy
style drains, which are less traumatic, and my own practice has been to
just place the tip of a Jackson-Pratt drain just barely into the posterior
burr hole using a gravity system with a bile bag connected to the Jackson-
Pratt tubing. One vexing problem with burr hold drainage is the cosmetic
aspect. Using the SPD as described, one can likely place porous titanium
burr holes over the bone defects, to minimize cosmetic issues resulting
from the bone defects. I believe this paper provides strong evidence for
use of the subperiosteal drain method.
Craig H. Rabb
Salt Lake City, Utah
I would like to congratulate the authors for this study. I am glad to
see significant progress being made in this topic that we commonly
encounter in clinical neurosurgery. I would like to clarify few aspects that
is relevant from the above study and also about the future prospects that
would benefit the wider community across the globe. Duration of drain
is an important aspect and this is elucidated clearly and one could argue
about the length of drain duration post evacuation of chronic subdural
hematoma. A 48-hour drainage seems a plausible duration. We would
not risk the possibility of getting an infection if the drain is left too long.
This was similar to a study that was conducted by our team, albeit in a
smaller group of patients. Then comes the aspect of utility of suction to
the drain particularly in the subperiosteal group as it would be safer in
VOLUME 0 | NUMBER 0 | 2019 | 9
SOLEMAN ET AL
comparison to the sub dural group. These factors did not seem to affect
the outcome in this study and this topic can be relevant for the future
studies. The other major point this is not clarified in detail is regarding
the role of anticoagulants and the outcome in the patients who were
on them. With the advent of newer anticoagulants available for various
medical conditions, it is still unclear if these have a significant impact
on the patient outcome. There is clearly a role for subgroup analysis and
subsequent studies on how we can manage these groups of patients. We
may have answers regarding how safely we can reintroduce the anticoag-
ulants postoperatively, depending on the patient’s clinical needs. In the
subdural group there is a potential chance of brain parenchymal irritation
10 | VOLUME 0 | NUMBER 0 | 2019
and possible seizures due to placement of drain. It is reassuring that
anticonvulsants were not prophylactically used and the seizure rates were
comparable in the subgroups. It is worthy of note that adhering to strict
Downloaded from https://academic.oup.com/neurosurgery/advance-article-abstract/doi/10.1093/neuros/nyz095/5440572 by Lancaster University user on 12 April 2019
antibiotic prophylaxis administered within 30 minutes of induction to
skin incision in this study is relevant and may prove beneficial particu-
larly with drain being left in situ for 48 hrs. The role of prolonged antibi-
otics for that period is still debatable. Overall I feel this study would add
to the available literature without any iota of doubt.
Chandrasekaran Kaliaperumal
Edinburgh, United Kingdom
www.neurosurgery-online.com