Original Article

Vacuum Drains versus Passive Drains versus No Drains in Decompressive

CraniectomieseA Randomized Controlled Trial on Subgaleal Drain Complication
Rates (VADER Trial)
Jo Ee Sam1,2, Regunath Kandasamy1, Albert Sii Hieng Wong2, Abdul Rahman Izaini Ghani1, Song Yee Ang1,
Zamzuri Idris1, Jafri Malin Abdullah3

- OBJECTIVE: Subgaleal drains are generally deemed with traumatic brain injury, the SGH amount of the PD and
necessary for cranial surgeries including decompressive ND group was significantly higher than that of the VD group.
craniectomies (DCs) to avoid excessive postoperative However, these higher amounts did not translate as a sig-
subgaleal hematoma (SGH) formation. Many surgeries have nificant risk factor for poor functional outcome or mortality.
moved away from routine prophylactic drainage but the VD may have better functional outcome and mortality.
role of subgaleal drainage in cranial surgeries has not - CONCLUSIONS: In terms of complication rates, VD, PD,
been addressed.
and ND may be used safely in DC. A higher amount of SGH
- METHODS: This was a randomized controlled trial at 2 was not associated with poorer outcomes. Further studies
centers. A total of 78 patients requiring DC were random- are needed to clarify the advantage of VD regarding
ized in a 1:1:1 ratio into 3 groups: vacuum drains (VD), functional outcome and mortality, and if ND reduces PCH
passive drains (PD), and no drains (ND). Complications rates.
studied were need for surgical revision, SGH amount, new
remote hematomas, postcraniectomy hydrocephalus (PCH),
functional outcomes, and mortality.
- RESULTS: Only 1 VD patient required surgical revision to INTRODUCTION
evacuate SGH. There was no difference in SGH thickness
and volume among the 3 drain types (P [ 0.171 and
P [ 0.320, respectively). Rate of new remote hematoma and
PCH was not significantly different (P [ 0.647 and
P [ 0.083, respectively), but the ND group did not have any
P rophylactic subcutaneous drains in surgery have been used
as a conduit for detection and drainage of hematomas or
excessive secretions. Subgaleal hematomas (SGHs) have
been considered to prolong healing and promote surgical site
infection.1 As a result, subgaleal vacuum drains (VD) have
patient with PCH. In the subgroup analysis of 49 patients generally been deemed necessary for cranial surgeries to reduce

Key words PD: Passive drains

- Complications SGH: Subgaleal hematoma
- Decompressive craniectomy TBI: Traumatic brain injury
- Functional outcome VD: Vacuum drains
- Mortality
- Postcraniectomy hydrocephalus From the 1Department of Neurosciences, School of Medical Sciences, Jalan Hospital USM,
- Subgaleal drains Universiti Sains Malaysia Health Campus, Kota Bharu, Kelantan; 2Department of
- Subgaleal hematoma Neurosurgery, Hospital Umum Sarawak, Jalan Hospital, Sarawak; and 3Department of
Neurosciences & Brain Behaviour Cluster, Hospital Universiti Sains, Malaysia, Universiti
Abbreviations and Acronyms Sains Malaysia, Health Campus, Kota Bharu, Kelantan, Malaysia
CI: Confidence interval To whom correspondence should be addressed: Jo Ee Sam, M.S.-Neurosurgery
CT: Computed tomography [E-mail: joeesam@gmail.com]
DC: Decompressive craniectomy
Citation: World Neurosurg. (2021).
GCS: Glasgow Coma Scale
https://doi.org/10.1016/j.wneu.2021.09.074
GOS: Glasgow Outcome Scale
ICB: Intracerebral bleed Journal homepage: www.journals.elsevier.com/world-neurosurgery
mRS: modified Rankin Scale Available online: www.sciencedirect.com
ND: No drains 1878-8750/$ - see front matter ª 2021 Elsevier Inc. All rights reserved.
OR: Odds ratio
PCH: Postcraniectomy hydrocephalus

WORLD NEUROSURGERY -: e1-e11, - 2021 www.journals.elsevier.com/world-neurosurgery e1

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

the amount of postoperative SGH formation. Multiple surgical DC is a common procedure that allows expansion of the
disciplines including spinal surgery have found that many damaged brain.6-10 An accumulation of SGH with mass effect after
operations do not need prophylactic drainage.2-5 This matter has DC may compromise the benefits of the procedure, and so it
not been addressed for cranial surgery yet, whether it is crani- seems logical that a subgaleal VD would reduce the amount of any
otomy or decompressive craniectomy (DC). postoperative SGH. However, routine drain use is not without

Figure 1. Study flowchart. CT, computed tomography.

e2 www.SCIENCEDIRECT.com WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2021.09.074

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

(international normalized ratio >1.5) or thrombocytopenia

(platelet <100
103 per mL), or recent antiplatelet or anticoagu-
lant use in the past 7 days were excluded.
The primary end point for this study was the need of surgical
revision as a result of mass effect of SGH and the SGH amount
after DC quantified by maximum thickness and volume. Secondary
end points were rates of new remote hematomas, PCH, favorable
functional outcomes, and mortality.
All DC were performed by senior residents supervised by the
attending surgeons. The large frontotemporoparietal craniec-
tomies were performed in accordance with recommended tech-
niques.24 The two-dimensional area of the bone flap was
estimated by using the longest diameter (D)
diameter perpen-
dicular to D (d)
p/4. Expansile duraplasty was performed with
pericranium. Block randomization with assignment groups con-
cealed within envelopes was used. The 3 groups were patients with
subgaleal VD, patients with subgaleal PD, and patients with ND.
Figure 2. Volume of hematoma was estimated by assuming that the Randomization was performed in blocks of 6, each of which
subgaleal hematoma is half an ellipsoid using the formula. contained 2 cards of each group. This strategy was to ensure equal
representation in a 1:1:1 ratio for each group during recruitment.
Placement of subgaleal drains in the VD group, PD group, and
ND group were standardized: the tip of the full length drain
complications.11 There is a possibility that VD could cause overlies the frontal lobe and the rest of the drain overlies the
complications including SGHs, postcraniectomy hydrocephalus temporal lobe and is placed underneath the temporalis muscle
(PCH), and new remote hematomas.12-14 There are a few cranial flap. The drains were brought out through a separate stab incision
case reports and series dealing with hemodynamic and suction toward the occipital region and attached to a standardized closed
complications of VD,15-21 and a few works dealing with paradoxic suction drain system (3.2 mm in outer diameter, 10F round and
brain herniation that could have been caused by VD.22,23 transparent polyvinyl chloride tube with a 400-mL Redon bottle
In our practice, subgaleal VD, subgaleal passive drains (PD), spring evacuator chamber [Hilfsmittel, Comcorde Medical, Kuala
and no drains (ND) use have been based according to surgeons’ Lumpur, Malaysia]). Randomization envelopes were revealed only
preferences. Many studies have been carried out to optimize the after surgery was completed to avoid surgeon bias during hemo-
technique of DC but the possible complications associated with stasis. For those patients in the PD group, the drains were con-
subgaleal drains in DC have been overlooked. There have been no nected to the same drain system but the vacuum effect was
randomized studies to compare the complication rates of these 3 released using an 18G branula and placed at bed level. For patients
alternative choices of subgaleal drains in DC. The primary end in the ND group, drains were placed but were clamped until they
point that we studied was the need for surgical revision as a result were removed. This strategy allows for surgeon blinding and acts
of mass effect of SGH and the SGH amount after DC. The other as a safety measure by which the drain may be unclamped to allow
complications of interest were rates of new remote hematomas, for SGH drainage if there is significant mass effect compressing
PCH, functional outcomes, and mortality. the brain. The surgical sites for patients of all groups were then
closed in a similar manner. The decision about external ventricular
drains or intraparenchymal monitors was left to the discretion of
METHODS the neurosurgeon.
A total of 119 patients underwent DC from January 2019 until All patients were monitored in high-dependency units for sig-
January 2020 at Hospital Universiti Sains Malaysia and Hospital nificant SGH that may require surgical revision. The surgical site
Umum Sarawak. Of these patients, 28 with a history of recent and dressing were inspected daily in all patients until sutures were
antiplatelet or anticoagulant use and patients with evidence of removed. The total amount of drainage was recorded and drains
coagulopathy or thrombocytopenia from laboratory results were were removed on the second postoperative day. Thromboembolic
excluded. Another 13 patients either did not have a legal repre- deterrent stockings or intermittent pneumatic compression de-
sentative to consent to participation or the legal representative was vices were used for every patient for prevention of deep vein
not keen for the patient to participate. A total of 78 patients with thrombosis; however, prophylactic anticoagulants were not
26 patients in each group were recruited (Figure 1). The study was administered until at least 3 days after the surgery. Initial intra-
approved by the ethics committee (USM/JEPeM/18070350). The cranial pathology characterization was obtained from the last
trial protocol was registered under ClinicalTrials.gov computed tomography (CT) brain before DC. Brain CT with 0.6
(NCT03777774). cm slice thickness was performed routinely on the first post-
Patients were eligible for participation if they were between 12 operative day (24  12 hours). Information on SGH thickness,
and 80 years old, were indicated for unilateral DC as decided by SGH volume, and presence of new remote hematomas was ob-
the attending neurosurgeon, and consent of participation was tained from this brain CT by 2 different investigators who were not
granted by a legal representative. Patients with coagulopathy aware of the randomization. If there was clinical suspicion of SGH

WORLD NEUROSURGERY -: e1-e11, - 2021 www.journals.elsevier.com/world-neurosurgery e3

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

or hydrocephalus, brain CT was repeated. Patients were seen in compare means of 3 groups. A value of P < 0.05 was considered
the clinic after discharge for detection of complications and to statistically significant. Multivariate analyses were used to identify
determine 6-month functional status according to the modified independent risk factors for higher SGH amount, poor functional
Rankin Scale (mRS) for patients who had infarcts or spontaneous outcome, and mortality. Only variables with a P value <0.1 in the
intracerebral bleeds (ICBs), and Glasgow Outcome Scale (GOS) for univariate analyses were analyzed further, and no variables were
patients who had traumatic brain injuries (TBIs). removed in the multivariate analyses to avoid overfitting and
underfitting, respectively. Because of limited sample size, the
Definitions forward logistic regression method was used for multivariate an-
A postoperative surgical site SGH was defined as extra-axial blood alyses. For assessment of interrater reliability, we used the intra-
collection underneath the skin flap on first postoperative brain CT. class correlation coefficient.
The maximum thickness of SGH was measured on the axial cut of
the brain CT. SGH volume was estimated by assuming that the RESULTS
SGH is half an ellipsoid using the formula (Figure 2):
Complications and Outcome
width
length
height Participants were equally matched in the 3 groups (Table 1).
Volume of subgleal hematoma ¼
2 Table 2 summarizes the complications, functional outcome, and
mortality findings among the VD, PD, and ND groups. Although
The longest SGH slice was identified, and then the length and the amount of SGH was slightly lower for the VD group, this
maximum width of the hematoma in this slice were measured. difference was clearly not significant. Only 1 patient required
The height is taken as the number of counted slices with the SGH surgical revision to evacuate the SGH that was causing
seen multiplied by the slice thickness of 0.6 cm. This method of significant mass effect. This patient had spontaneous
using half an ellipsoid for calculating SGH volume has not been intracerebral hemorrhage and was on a VD with SGH thickness
verified with computer-assisted volumetric analysis but has been of 25 mm and SGH volume of 70.3 mL. The VD group had a
previously proved to be accurate for subdural hematomas.25-27 higher percentage (23.1%) but nonsignificant new remote
New remote hematomas were defined as hematomas not previ- hematoma rate compared with the PD (15.4%) and ND (11.5%)
ously seen on earlier brain CT but seen on the first postoperative groups. The overall PCH rate was 12.1%, with a trend toward
day brain CT that cannot be explained by a direct connection or significance in favor of the ND group (0%). Patients in the VD
complication from the original hematomas. PCH was diagnosed if group were shown to have better functional outcomes (P ¼
patients had 1) increasingly bulging craniectomy site, 2) deterio- 0.011; PD, odds ratio [OR], 6.571, confidence interval [CI],
ration of neurologic function or failure of neurologic improve- 1.574e27.432; ND, OR, 3.6, CI, 1.038e12.481) and lower
ment, and 3) Evans ratio >0.3 and periventricular lucency.28 About mortality (P ¼ 0.032; PD, OR, 5.622, CI, 1.342e23.559; ND, OR,
80% of patients develop hydrocephalus within the first 50 days of 4.792, CI, 1.136e20.211) compared with the PD and ND groups.
undergoing DC.29 Therefore, we excluded patients who died The intraclass correlation between the 2 investigators for
before 60 days after DC in the statistical analysis for PCH measurements of SGH thickness and volume was 0.993
because these patients may have died before hydrocephalus had (P < 0.001) and 0.986 (P < 0.001) respectively, indicating high
time to develop. Functional status using mRS or GOS was reproducibility between the 2 investigators.
assessed at 6 months after DC. Based on functional
independence, mRS scores of 1e2 were considered good Risk Factors of Higher SGH Thickness and Volume
outcome and scores of 4e6 were considered poor outcome. To determine if there were other factors influencing the SGH
Similarly, GOS scores of 4e5 were considered good outcome thickness or volume other than drain type, we analyzed for
and scores of 1e3 were considered poor outcome. All-cause possible risk factors that could have contributed to the higher SGH
mortality at 6 months was recorded as well. Herniation through amounts (Table 3). Extracranial herniation of the brain was
the craniectomy defect was defined as herniating brain tissue >1.5 significantly associated with less SGH thickness (P ¼ 0.001) and
cm higher than the outer table of the skull defect. less SGH volume (P ¼ 0.009). The surgical indication did not
influence the SGH thickness or the SGH volume. Analysis using
Statistical Analysis a Pearson correlation did not find any correlation of SGH
Fixed-effects, omnibus, 1-way analysis of variance was calculated thickness or volume with the platelet counts, international
using G*Power Statistical software version 3.1.9.2 (https://www. normalized ratio, activated partial thromboplastin time, or bone
psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und- flap size.
arbeitspsychologie/gpower). Taking the effect size f as 0.4, with a
2-sided type I error of 0.05 (alpha), and a power of 80%, the total Risk Factors for Poor Functional Outcome and Mortality
sample size needed was 66 patients. We recruited 78 patients and To determine if the type of drain independently influenced func-
there were no dropouts. tional outcome and mortality, univariate and multivariate analysis
The data were analyzed using SPSS version 22 (IBM Corp., of other potential factors was performed (Table 4). Higher SGH
Armonk, New York, USA). Categorical data were analyzed using a amount did not contribute to poor functional outcome or
c2 or Fisher exact test. An independent t test was used to compare mortality. Other identified risk factors for poor functional
means for 2 groups. One-way analysis of variance was used to outcome were infarction and spontaneous ICB as indication of

e4 www.SCIENCEDIRECT.com WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2021.09.074

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

Table 1. Characteristics of the Patients

Characteristics Vacuum Drain (n [ 26) Passive Drain (n [ 26) No Drain (n [ 26) P Value

Gender, n (%)
Male 22 (84.6) 23 (88.5) 20 (83.3) 0.647*
Female 4 (15.4) 3 (11.5) 6 (16.7)
Glasgow Coma Scale score, n (%)
3e8 18 (69.2) 16 (61.5) 17 (65.4) 0.844y
9e12 8 (30.8) 10 (38.5) 9 (34.6)
13e15 0 (0) 0 (0) 0 (0)
Midline shift >5 mm, n (%), present 24 (92.3) 21 (80.8) 22 (84.6) 0.604*
Dominant hemisphere decompressive 11 (42.3) 15 (57.7) 14 (53.8) 0.513y
craniectomy, n (%), yes
Basal cisterns, n (%)
Patent 10 (38.5) 6 (23.1) 10 (38.5) 0.212y
Partially effaced 9 (34.6) 11 (42.3) 4 (15.4)
Fully effaced 7 (26.9) 9 (34.6) 12 (46.2)
Surgery indication, n (%)
Traumatic brain injury 19 (73.1) 16 (61.5) 14 (53.8) 0.239*
Middle cerebral artery infarct 1 (3.8) 6 (23.1) 6 (23.1)
Intracerebral bleed 6 (23.1) 4 (15.4) 6 (23.1)
Extracranial herniation, n (%), 8 (30.8) 9 (34.6) 9 (34.6) 0.944y
present
Age (years), mean  SD (range) 34.2  19.34 (12e77) 45.7  15.492 (2e68) 41.9  17.271 (4e77) 0.060z
Bone flap size (cm ), mean  SD
3
148.0  10.43 (128.3e169.7) 148.5  10.99 (127.6e169.7) 145.0  15.58 (112.9e165) 0.552z
(range)
Distance from midline (cm), 1.5  0.70 (0.5e3.0) 1.9  0.6 (10.6e2.8) 1.9  1.40 (0.2e8.0) 0.223z
mean  SD (range)
Platelet (
103), mean  SD (range) 253.7  78.18 (161e498) 296.2  149.66 (134e920) 285.2  103.32 (129e649) 0.383z
International normalized ratio, mean 1.2  0.15 (0.9e1.5) 1.1  0.18 (0.8e1.6) 1.1  0.18 (0.83e1.5) 0.135z
 SD (range)
Activated partial thromboplastin time 36.5  6.132 (5.2e49.0) 33.5  7.78 (15.3e42.0) 34.2  5.03 (23.3e40.4) 0.105z
(seconds), mean  SD (range)

SD, standard deviation.

*Fisher exact test.
yc2 test.
zOne-way analysis of variance test.

surgery (P ¼ 0.008), extracranial herniation (P ¼ 0.044), and older (OR, 3.586; CI, 1.131e11.371) were significant predictors of mor-
age (P < 0.001). Older age was the only other significant risk tality. The type of drain did not influence the outcome in the
factor associated with mortality (P ¼ 0.002). multivariate analyses.
Using multiple logistic regression, only older age (OR, 1.108;
CI, 1.049e1.171) and extracranial herniation (OR, 10.92; CI,
2.084e57.214) were significant predictors of poor functional Subgroup Analysis of Patients with TBI
outcome. Multiple logistic regression for mortality again showed Subgroup analysis of complications and outcomes according
that age (OR, 1.055; CI, 1.021e1.090) and extracranial herniation to surgical indication was performed because it was found
earlier that surgical indication did influence functional

WORLD NEUROSURGERY -: e1-e11, - 2021 www.journals.elsevier.com/world-neurosurgery e5

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

Table 2. Complications and Outcome

Complication and Outcome Vacuum Drain (n [ 26) Passive Drain (n [ 26) No Drain (n [ 26) P Value

Subgaleal hematoma thickness (mm), mean  SD (range) 6.6  4.46 (2e25) 8.6  4.14 (2e22) 8.3  3.8 (12e17) 0.171*
Subgaleal hematoma volume (mL), mean  SD (range) 26.4  25.61 (5.5e140.6) 29.5  14.45 (5.4e53.8) 34.3  14.5 (48.7e74.7) 0.320*
Surgical revision, n (%)
Yes 1 (3.8) 0 (0) 0 (0) 1.000y
No 25 (96.2) 26 (100) 26 (100)
New remote hematomas, n (%)
Yes 6 (23.1) 4 (15.4) 3 (11.5) 0.647y
No 20 (76.9) 22 (84.6) 23 (88.5)
Post craniectomy hydrocephalusz, n (%)
Yes 3 (13.0) 4 (23.5) 0 (0.0) 0.083y
No 20 (87.0) 13 (76.5) 18 (100.0)
Functional outcome at 6 months, n (%)
Good 12 (46.2) 3 (11.5) 5 (19.2) 0.011x
Poor 14 (53.8) 23 (88.5) 21 (80.8)
Mortality at 6 months, n (%)
No 23 (88.5) 15 (57.7) 16 (61.5) 0.032x
Yes 3 (11.5) 11 (42.3) 10 (38.5)

SD, standard deviation.

*One-way analysis of variance.
yFisher exact test.
zOnly 58 patients who were alive at least 60 days after decompressive craniectomy were included in the analysis.
xc2 test.

outcome. Only the subgroup analysis of the 49 patients with Patients in the VD group had significantly less SGH and
TBI showed significant results. Baseline characteristics for all surprisingly again, lower mortality (Table 5). Further analysis
3 drain groups were comparable in this subgroup analysis. for other risk factors of poor outcome and mortality showed

Table 3. Risk Factors of Higher Subgaleal Hematoma Thickness and Volume

Subgaleal Hematoma

Thickness (mm) Volume (mL)

Variable n Mean (SD) P Value n Mean (SD) P Value

Extracranial herniation of brain

Yes 26 5.7 (3.26) 0.001* 26 22.2 (13.00) 0.009*
No 52 8.9 (4.21) 52 34.0 (20.36)
Surgery indication
Trauma 49 7.8 (4.04) 0.441y 49 29.4 (14.35) 0.797y
Middle cerebral artery infarct 13 6.70 (3.28) 13 29.1 (18.30)
Intracerebral bleed 16 8.7 (5.22) 16 32.9 (30.20)

SD, standard deviation.

*Independent t test.
yOne-way analysis of variance.

e6 www.SCIENCEDIRECT.com WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2021.09.074

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

Table 4. Possible Risk Factors of Poor Functional Outcome and Mortality

Functional Outcome Mortality

Characteristics Independent (n [ 20) Dependent (n [ 58) P Value No (n [ 54) Yes (n [ 24) P Value

Gender, n (%)
Male 15 (75) 50 (86.2) 0.246* 43 (79.6) 22 (91.7) 0.188*
Female 5 (25) 8 (13.8) 11 (20.4) 2 (8.3)
Glasgow Coma Scale score, n (%)
3e8 14 (70) 37 (63.8) 0.615* 33 (61.1) 18 (75.0) 0.234*
9e12 6 (30) 21 (36.2) 21 (38.9) 6 (25.0)
Midline shift >5 mm, n (%), present 16 (80) 51 (87.9) 0.380* 46 (85.2) 21 (87.5) 0.786*
Extracranial herniation, n (%), present 3 (15) 23 (39.7) 0.044* 15 (27.8) 11 (45.8) 0.118*
Dominant hemisphere decompressive 10 (50) 30 (51.7) 0.894* 25 (46.3) 15 (62.5) 0.186*
craniectomy, n (%), yes
Basal cisterns, n (%)
Patent 4 (20) 22 (37.9) 0.085* 19 (35.2) 7 (29.2) 0.059*
Partially effaced 10 (50) 14 (24.1) 20 (37.0) 4 (16.7)
Fully effaced 6 (30) 22 (35.9) 15 (27.8) 13 (54.2)
Surgery indication, n (%)
Traumatic brain injury 18 (90) 31 (53.4) 0.008y 34 (63.0) 15 (62.5) 1.000y
Infarct 0 (0) 13 (22.4) 9 (16.7) 4 (16.7)
Intracerebral bleed 2 (10) 14 (24.1) 11 (20.4) 5 (20.8)
Age (years), mean  SD (range) 27.1  12.9 (14e53) 45.2  17.0 (12e77) <0.001z 36.5  16.40 (12e68) 49.8  17.9 (28e77) 0.002z
Subgaleal hematoma thickness (mm), 8.0  3.7 (3e17) 7.8  4.4 (2e25) 0.810z 8.0  4.4 (2.8e17.0) 7.5  3.7 (2.0e14.0) 0.649z
mean  SD (range)
Subgaleal hematoma volume (mL), 27.9  13.1 (5.5e56.9) 30.8  20.7 (5.4e140.6) 0.561z 30.6  20.7 (5.5e140.6) 28.8  14.6 (5.4e51.8) 0.694z
mean  SD (range)

SD, standard deviation.

*c2 test.
yFisher exact test.
zIndependent T test.

that male gender, basal cisterns patency, and older age were intracranial pressure measurements from the contralateral
associated with poorer outcomes. The amount of SGH did not hemisphere decreased to approximately e20 mm Hg at the
influence the functional outcome both in univariate and in moment of connection of the subgaleal suction drain and
multivariate analysis (Table 6). warned on the use of VD. Multiple reports show unexpected
bradycardia and hypotension events when connecting the drains
DISCUSSION to negative pressure.15,17,18 Yadav et al.21 attributed this result to
sudden intracranial hypotension or the trigeminocardiac reflex.
Literature Review of Studies of Subgaleal Drains Woo et al.23 reported harmful upward herniation syndrome after
The technique for DC has been improved over time but it is uneventful cranioplasty, likely related to suction drainage. Such
striking that the choice of subgaleal drains is not phenomena were not detected in our study. PDs allow for
mentioned.24,30,31 There have since been more advocates of PD gravitational drainage of SGHs minus the possible detrimental
and ND because the VD high negative pressures of e71 to e175 effects of the strong vacuum suctions on the delicate brain.
mm Hg32 have been suspected to cause complications. Van The studies to date are mainly of craniotomies and cranio-
Roost et al.,19 in their extensive search for the cause of plasties and there are no data on drain use for DC (Table 7). Chang
pseudohypoxic brain swelling of 17 patients, stated that et al.33 showed that use of VD in cranioplasties was protected

WORLD NEUROSURGERY -: e1-e11, - 2021 www.journals.elsevier.com/world-neurosurgery e7

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

against postoperative fluid collection as opposed to ND but Sobani SGH amounts (Table 5). However, these higher amounts did not
et al.34 found no significant difference in complication rates. Choi translate as a significant risk factor for poor functional outcome
et al.35 showed that subgaleal drains caused a higher rate of SGH or mortality (Tables 4 and 6).
(7.9%) compared with the ND group (2.4%). These investigators As in most of the previous studies, it is proved that ND were not
postulated that the surgeons of the ND group were more inferior to VD or PD for DC. There are several advantages of not
meticulous in hemostasis compared with those of the VD group. having a drain, including 1) saving the space occupied by the drain
Guangming et al.36 reported similar rates of epidural hematoma to allow for brain swelling after craniectomy especially in patients
in both groups. Recently, Hamou et al.37 further added evidence with extracranial herniation; 2) reducing the injury to the tissues
that presence of subgaleal drains in supratentorial craniotomies that have just begun to heal during blind removal of the drain
did not affect accumulation of SGH or operative revision. Most from the surgical site; and 3) cost savings of omitting the drain,
of the studies that have been performed so far have more dressings, and sutures.
patients in the ND cohort. When there is extracranial herniation, SGH thickness and vol-
ume are minimal (Table 3). This situation is likely caused by
Discussion of Complications reduced potential space between the galea and dura when the
This study showed that all 3 options of VD, PD, and ND may be swollen brain pushes against the dura and galea. From this
used safely in DC because there was no difference in operative finding, we may deduce that the amount of SGH can easily be
revision, SGH thickness, SGH volume, rates of PCH, or new reduced just by minimizing this potential space and thus
remote hematomas (Table 2). All DC have some degree of reducing the usefulness of drains. This potential space created
subgaleal blood collection. The more important question is during craniectomies and craniotomies may be reduced by using
whether the amount of SGH causes mass effect requiring tenting stitches between the galea and the pericranium or dura.38
evacuation with another surgery and whether the amount of In our study, the type of drain did not affect the rate of for-
SGH can affect long-term outcome. Only 1 patient needed mation of new remote hematomas. Although the mechanism for
another surgery for SGH evacuation. This patient was on a VD, new remote hematomas is largely attributed to the release of
which undermines the notion that VD helps deter SGH collection. tamponade effect and the presence of a skull fracture, we hy-
Good surgical technique and hemostasis cannot be replaced by pothesized that the negative pressure environment caused by the
placement of a drain. In the overall analysis of 78 patients, PD and VD may also contribute to the development of new hematomas.39
ND had a nonsignificant trend of higher SGH amounts (Table 2). In support of this theory, a recent randomized controlled trial
In the TBI subgroup analysis, PD and ND had significantly higher found that new hematomas occurred significantly more after DC

Table 5. Subgroup Analysis of Complications and Outcome in Patients with Traumatic Brain Injury Only
Complication and Outcome Vacuum Drain (n [ 19) Passive Drain (n [ 16) No Drain (n [ 14) P Value

Subgaleal hematoma thickness (mm), mean  SD (range) 5.9  2.6 (12e11.4) 9.2  4.95 (2e22) 8.9  3.67 (2e17) 0.022*
Subgaleal hematoma volume (mL), mean  SD (range) 22.7  11.28 (7.9e42.8) 30.8  16.6 (15.4e53.8) 36.7  11.89 (8.7e56.9) 0.016*
New remote hematomas, n (%)
Yes 3 (15.8) 4 (25.0) 3 (21.4) 0.904y
No 16 (84.2) 12 (75.0) 11 (78.6)
Post craniectomy hydrocephalusz, n (%)
Yes 3 (17.6) 3 (27.3) 0 (0.0) 0.364y
No 14 (82.4) 8 (72.7) 8 (100.0)
Functional outcome at 6 months, n (%)
Good 10 (52.6) 3 (18.8) 5 (35.7) 0.117x
Poor 9 (47.4) 13 (81.3) 9 (64.3)
Mortality at 6 months, n (%)
No 17 (89.5) 10 (62.5) 7 (50.0) 0.038y
Yes 2 (10.5) 6 (37.5) 7 (50.0)

SD, standard deviation.

*One-way analysis of variance.
yFisher exact test.
zOnly 36 patients who were alive at least 60 days after decompressive craniectomy were included in analysis.
xc2 test.

e8 www.SCIENCEDIRECT.com WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2021.09.074

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

when conventional rapid decompression was performed compared Discussion on Functional Outcomes and Mortality
with a controlled decompression group.40 The sudden high PD and ND were risk factors for poor functional outcome and
negative pressure suction by VD compared with the gradual mortality in the univariate but not multivariate analysis. Mortality
drainage by PD would be similar in concept. There is a was still higher in the PD and ND group in the subgroup analysis
possibility that a sudden negative pressure may cause traction of only patients with TBI. Because the multivariate analyses did
on the meningeal vessels, stripping the dura from the inner not find drain type to be an independent risk factor for mortality,
table of the skull, causing the extradural vessels to bleed.41 there is likely a confounding factor. The reason for this situation is
The ND group did not have any patients with PCH. Further not clear and needs further study but there are some possible
studies with more patients may clarify if ND may be protective explanations. One explanation could be that the VD group has
against PCH. It would be beneficial if PCH rate could be reduced slightly higher proportions of patients with TBI and a younger age-
just by not using a drain during DC. The other question is whether group. Second, the VD group has only 1 patient with middle ce-
gravitational drainage by PD is enough to alter brain compliance rebral artery territory infarction and all 13 patients with middle
and pulsatile intracranial pressure dynamics just by exposing the cerebral artery infarction have a poor mRS score because they are
intracranial contents to the lower ambient atmospheric pressures usually hemiplegic and dependent.45 Furthermore, comparing
even without strong negative pressure suction.29,42 This possibility functional outcome scores for a heterogenous set of diagnosis is
is highly likely because studies by Herbowski43,44 showed that not ideal. Nonetheless, if VD are truly associated with better
there is a direct positive relationship of atmospheric pressure outcomes, further adequately powered and specially tailored
and intracranial pressure, and a decrease in atmospheric studies of functional outcome and subgaleal drains are crucial in
pressure causes a decrease in intracranial pressure amplitude. the future.
Passive drainage may have a mild suction effect because of
gravity and siphoning when placed below head level. If so, even Limitations
the use of PD could increase risk of PCH and does not confer Manual calculation of SGH volume may be affected by the angle of
any benefit compared with VD in this aspect. axial slice and the level of cuts. Even if the accuracy of this method

Table 6. Subgroup Analysis of Possible Risk Factors of Poor Functional Outcome and Mortality in Patients with Traumatic Brain Injury
Only
Functional Outcome Mortality

Independent Dependent
Characteristics (n [ 18) (n [ 31) P Value No (n [ 34) Yes (n [ 15) P Value

Gender, n (%)
Male 14 (77.8) 31 (100) 0.014* 30 (88.2) 15 (100) 0.219*
Female 4 (22.2) 0 (0) 4 (11.8) 0 (0)
Glasgow Coma Scale score, n (%)
3e8 12 (66.7) 26 (83.9) 0.150y 26 (76.5) 12 (80.0) 0.550y
9e12 6 (33.3) 5 (16.1) 8 (23.5) 3 (20.0)
Midline shift >5 mm, n (%), present 14 (77.8) 27 (87.1) 0.320y 28 (82.4) 13 (86.7) 0.532y
Extracranial herniation, n (%), present 3 (16.7) 11 (35.5) 0.140y 8 (23.5) 6 (40.0) 0.201y
Basal cisterns, n (%)
Patent 4 (22.2) 9 (29.0) 0.032y 11 (32.4) 2 (13.3) 0.012*
partially effaced 10 (55.6) 6 (19.4) 14 (41.2) 2 (13.3)
fully effaced 4 (22.2) 16 (51.6) 9 (26.5) 11 (73.3)
Age (years), mean  SD (range) 26  12.2 (14e53) 41.6  19.2 (12e77) 0.001z 32.1  16.6 (12e77) 44.4  20.3 (18e77) 0.030z
Subgaleal hematoma thickness (mm), 8.2  3.8 (3e17) 7.6  4.2 (2e22) 0.604z 8.2  4.2 (2.8e22.0) 7.1  3.8 (2.0e14.0) 0.395z
mean  SD (range)
Subgaleal hematoma volume (mL), 14.3  6.7 (3.9e28.4) 14.9  7.5 (2.7e26.9) 0.786z 14.3  6.9 (3.9e28.4) 15.5  8.0 (2.7e25.9) 0.637z
mean  SD (range)

SD, standard deviation.

*Fisher exact test.
yc2 test.
zIndependent t test.

WORLD NEUROSURGERY -: e1-e11, - 2021 www.journals.elsevier.com/world-neurosurgery e9

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

Table 7. Summary of All Studies of Subgaleal Drains to Date

Number of Patients,
Reference Study Design Total (No Drain/Drain) Results

Guangming Prospective observational study of patients undergoing 342 (166/176) Not significant P ¼ 0.952
et al., 200936 supratentorial craniotomy for epilepsy surgery 20 epidural hematoma in no drain group
Divided into 2 groups based on even and odd dates of 22 epidural hematoma in drain group
craniotomy
Chang et al., Retrospective analysis of complications in patients 212 (128/84) 8.6% with fluid collection in no drain group versus 2.4% in
201033 undergoing cranioplasty (subdural or subgaleal fluid drain group
collections requiring repeated operations) Univariate P ¼ 0.069
Multivariate P ¼ 0.016
Sobani et al., Retrospective analysis of complications in patients 96 (69/27) 1 patient developed epidural hematoma (which group not
201134 undergoing cranioplasty mentioned)
39% complications in no drain group versus 26% in drain
group. P ¼ not significant (exact value not mentioned)
Choi et al., Retrospective analysis of patients undergoing pterional 607 (333/274) 2.4% subgaleal hematoma in no drain
201535 craniotomy for aneurysms 7.9% subgaleal hematoma in drain group
Hamou et al., Prospective analysis of patients undergoing supratentorial 150 (87/63) 34.5% early subgaleal swelling in no drain group
202037 craniotomy 39.7% early subgaleal swelling in drain group

is not proved, any error would be equivalent in all groups. In associated with poorer outcomes. Larger randomized studies are
surgical-based trials, it is difficult to fully standardize the tech- needed to clarify the advantage of VD regarding functional
niques of all surgeons. Knowing that there is a potential that the outcome and mortality and if ND reduces PCH rates.
patient may be in the PD or ND groups would encourage more
meticulous hemostasis and thus at least standardizing of this CRediT AUTHORSHIP CONTRIBUTION STATEMENT
aspect of the surgery. This study is not adequately powered to Jo Ee Sam: Conceptualization, Formal analysis, Investigation,
draw conclusions on functional outcome and mortality because Visualization, Writing e original draft. Regunath Kandasamy:
many other variables such as comorbidities, complications, and Conceptualization, Validation, Resources, Data curation, Writing
postoperative care that can affect outcome were not studied. e review & editing. Albert Sii Hieng Wong: Methodology,
Investigation, Writing e review & editing, Validation, Supervision.
CONCLUSIONS Abdul Rahman Izaini Ghani: Project administration, Formal
All 3 drain groups proved to be safe because only 1 patient in the analysis, Writing e review & editing, Supervision. Song Yee Ang:
VD group required surgical revision. There was no significant Investigation, Visualization, Validation. Zamzuri Idris: Writing e
difference in terms of SGH amount, rate of new remote hema- review & editing, Data curation. Jafri Malin Abdullah: Project
tomas, and rate of PCH. A higher amount of SGH was not administration, Formal analysis.

necessary for single-level lumbar decompression?: healthcare professionals from the American Heart
REFERENCES review of 560 cases. Clin Orthop. 2010;468: Association/American Stroke Association. Stroke.
2690-2694. 2015;46:2032-2060.
1. Roth J, Galeano E, Milla S,
Hartmannsgruber MW, Weiner HL. Multiple
5. Payne DH, Fischgrund JS, Herkowitz HN, 9. Carney N, Totten AM, O’Reilly C, et al. Guidelines
epidural hematomas and hemodynamic collapse
Barry RL, Kurz LT, Montgomery DM. Efficacy of for the Management of Severe Traumatic Brain
caused by a subgaleal drain and suction-induced
closed wound suction drainage after single-level Injury, Fourth Edition. Neurosurgery. 2017;80:6-15.
intracranial hypotension: case report. Neurosur-
lumbar laminectomy. J Spinal Disord. 1996;9:
gery. 2011;68:E271-E275 [discussion: E276]. 10. Wijdicks EFM, Sheth KN, Carter BS, et al. Rec-
401-403.
ommendations for the management of cerebral
2. Kosins AM, Scholz T, Cetinkaya M, Evans GRD. 6. Hutchinson PJ, Kolias AG, Timofeev IS, et al. Trial and cerebellar infarction with swelling: a state-
Evidence-based value of subcutaneous surgical of decompressive craniectomy for traumatic ment for healthcare professionals from the
wound drainage: the largest systematic review and intracranial hypertension. N Engl J Med. 2016;375: American Heart Association/American Stroke As-
meta-analysis. Plast Reconstr Surg. 2013;132:443-450. 1119-1130. sociation. Stroke. 2014;45:1222-1238.

7. Alexander P, Heels-Ansdell D, Siemieniuk R, et al. 11. Cruse PJ, Foord R. A five-year prospective study of
3. Hung P-I, Chang M-C, Chou P-H, Lin H-H,
Hemicraniectomy versus medical treatment with 23,649 surgical wounds. Arch Surg. 1973;107:
Wang S-T, Liu C-L. Is a drain tube necessary for
large MCA infarct: a review and meta-analysis. 206-210.
minimally invasive lumbar spine fusion surgery?
BMJ Open. 2016;6:e014390.
Eur Spine J. 2017;26:733-737. 12. Kurland DB, Khaladj-Ghom A, Stokum JA, et al.
8. Hemphill JC, Greenberg SM, Anderson CS, et al. Complications associated with decompressive
4. Kanayama M, Oha F, Togawa D, Shigenobu K, Guidelines for the management of spontaneous craniectomy: a systematic review. Neurocrit Care.
Hashimoto T. Is closed-suction drainage intracerebral hemorrhage: a guideline for 2015;23:292-304.

e10 www.SCIENCEDIRECT.com WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2021.09.074

 ORIGINAL ARTICLE
JO EE SAM ET AL. SUBGALEAL DRAIN COMPLICATIONS IN DECOMPRESSIVE CRANIECTOMY

13. Honeybul S, Ho KM. Long-term complications of 25. Gebel JM, Sila CA, Sloan MA, et al. Comparison of 38. Goel A. “Tenting” stitches for the scalp. Br J
decompressive craniectomy for head injury. the ABC/2 estimation technique to computer- Neurosurg. 1992;6:357-358.
J Neurotrauma. 2011;28:929-935. assisted volumetric analysis of intraparenchymal
and subdural hematomas complicating the 39. Huang Y-H, Lee T-C, Lee T-H, Yang K-Y, Liao C-
14. Ding J, Guo Y, Tian H. The influence of decom- GUSTO-1 trial. Stroke. 1998;29:1799-1801. C. Remote epidural hemorrhage after unilateral
pressive craniectomy on the development of hy- decompressive hemicraniectomy in brain-injured
drocephalus: a review. Arq Neuropsiquiatr. 2014;72: 26. Kasner Scott E. Geometry and subdural hema- patients. J Neurotrauma. 2013;30:96-101.
715-720. toma volume. Stroke. 1999;30:188.
40. Chen J, Li M, Chen L, et al. The effect of
15. Hernández-Palazón J, Tortosa JA, Sánchez- 27. Won S-Y, Zagorcic A, Dubinski D, et al. Excellent controlled decompression for severe traumatic
Bautista S, Martínez-Lage JF, Pérez-Flores D. accuracy of ABC/2 volume formula compared to brain injury: a randomized, controlled trial. Front
Cardiovascular disturbances caused by extradural computer-assisted volumetric analysis of subdural Neurol. 2020;11:107.
negative pressure drainage systems after intracra- hematomas. PLoS One. 2018;13:e0199809.
nial surgery. Br J Anaesth. 1998;80:599-601. 41. Chung H-J, Park J-S, Park J-H, Jeun S-S. Remote
28. Evans WA. An encephalographic ratio for esti- postoperative epidural hematoma after brain tu-
16. Sudmann E. Subgaleal closed suction drainage in mating ventricular enlargement and cerebral at- mor surgery. Brain Tumor Res Treat. 2015;3:132-137.
craniotomies. J Oslo City Hosp. 1973;23:77-82. rophy. Arch Neurol Psychiatry. 1942;47:931-937.
42. Waziri A, Fusco D, Mayer SA, McKhann GM,
17. Gilsanz FJ, Vaquero J, Lora-Tamayo JI, Alvarez J, 29. Kaen A, Jimenez-Roldan L, Alday R, et al. Inter- Connolly ES. Postoperative hydrocephalus in pa-
Martínez R. [Epicranial drainage: hemodynamic hemispheric hygroma after decompressive cra- tients undergoing decompressive hemi-
changes]. Rev Esp Anestesiol Reanim. 1984;31:20-21 niectomy: does it predict posttraumatic craniectomy for ischemic or hemorrhagic stroke.
[in Spanish]. hydrocephalus? J Neurosurg. 2010;113:1287-1293. Neurosurgery. 2007;61:489-493 [discussion: 493-
494].
18. Karamchandani K, Chouhan RS, Bithal PK, 30. Huang X, Wen L. Technical considerations in
Dash HH. Severe bradycardia and hypotension decompressive craniectomy in the treatment of 43. Herbowski L. The major influence of the atmo-
after connecting negative pressure to the subgaleal traumatic brain injury. Int J Med Sci. 2010;7: sphere on intracranial pressure: an observational
drain during craniotomy closure. BJA Br J Anaesth. 385-390. study. Int J Biometeorol. 2017;61:181-188.
2006;96:608-610.
31. Brown DA, Wijdicks EFM. Decompressive cra- 44. Herbowski L. From paradigm to paradox: di-
niectomy in acute brain injury. Handb Clin Neurol. vergency between intracranial pressure and
19. Van Roost D, Thees C, Brenke C, Oppel F,
2017;140:299-318. intracranial pulse pressure during atmospheric
Winkler PA, Schramm J. Pseudohypoxic brain
pressure fall: a case study [e-pub ahead of print]. J
swelling: a newly defined complication after un- 32. Grobmyer SR, Graham D, Brennan MF, Coit D. Neurosurg Sci. https://doi.org/10.23736/S0390-5616.
eventful brain surgery, probably related to suction High-pressure gradients generated by closed- 19.04737-4, accessed October 6, 2021.
drainage. Neurosurgery. 2003;53:1315-1326 [discus- suction surgical drainage systems. Surg Infect.
sion: 1326-1327]. 2002;3:245-249. 45. Kamal Alam B, Bukhari AS, Assad S, et al.
Functional outcome after decompressive craniec-
20. Laurenson VG, MacFarlane M, Davis FM. Negative 33. Chang V, Hartzfeld P, Langlois M, Mahmood A, tomy in patients with dominant or non-dominant
pressure drainage after craniotomy. Neurosurgery. Seyfried D. Outcomes of cranial repair after cra- malignant middle cerebral infarcts. Cureus. 2017;9:
1985;17:868. niectomy. J Neurosurg. 2010;112:1120-1124. e997.
21. Yadav M, Nikhar SA, Kulkarni DK, Gopinath R. 34. Sobani ZA, Shamim MS, Zafar SN, et al. Cranio-
Cardiac arrest after connecting negative pressure plasty after decompressive craniectomy: an insti-
to the subgaleal drain during craniotomy closure. tutional audit and analysis of factors related to
Case Rep Anesthesiol. 2014;2014:146870. complications. Surg Neurol Int. 2011;2:123.
Conflict of interest statement: The authors declare that the
22. Nasi D, Dobran M, Iacoangeli M, Di Somma L, 35. Choi SY, Yoon SM, Yoo CJ, Park CW, Kim YB, article content was composed in the absence of any
Gladi M, Scerrati M. Paradoxical brain herniation Kim WK. Necessity of surgical site closed suction commercial or financial relationships that could be construed
after decompressive craniectomy provoked by drain for pterional craniotomy. J Cerebrovasc Endo- as a potential conflict of interest.
drainage of subdural hygroma. World Neurosurg. vasc Neurosurg. 2015;17:194-202.
2016;91:673.e1-673.e4. Received 11 July 2021; accepted 15 September 2021
36. Guangming Z, Huancong Z, Wenjing Z, Citation: World Neurosurg. (2021).
23. Woo PYM, Lo WHY, Wong H-T, Chan K-Y. The Guoqiang C, Xiaosong W. Should epidural drain https://doi.org/10.1016/j.wneu.2021.09.074
“negative” impact of a subgaleal drain: post- be recommended after supratentorial craniotomy
cranioplasty negative pressure subgaleal drain- for epileptic patients? Surg Neurol. 2009;72:138-141 Journal homepage: www.journals.elsevier.com/world-
induced ascending transtentorial herniation. [discussion: 141]. neurosurgery
Asian J Neurosurg. 2019;14:256-261. Available online: www.sciencedirect.com
37. Hamou HA, Kotliar K, Tan SK, et al. Surgical
24. Quinn TM, Taylor JJ, Magarik JA, Vought E, nuances and placement of subgaleal drains for 1878-8750/$ - see front matter ª 2021 Elsevier Inc. All
Kindy MS, Ellegala DB. Decompressive craniec- supratentorial proceduresea prospective analysis rights reserved.
tomy: technical note. Acta Neurol Scand. 2011;123: of efficacy and outcome in 150 craniotomies. Acta
239-244. Neurochir (Wien). 2020;162:729-736.

WORLD NEUROSURGERY -: e1-e11, - 2021 www.journals.elsevier.com/world-neurosurgery e11