Keywords
craniotomy - hair wash - intrawound vancomycin - surgical site infection
Introduction
Postoperative surgical site infections (SSIs) are a recognized complication following
craniotomy with an associated increase in morbidity and mortality, length of stay,
and treatment costs. Postoperative SSIs may also worsen overall prognosis, particularly
of oncological patients due to delays in further treatment such as chemotherapy and
radiotherapy.[1]
[2]
[3] Postoperative SSIs include superficial skin infection, wound dehiscence, bone flap
osteomyelitis, meningitis, subdural empyema, and/or brain abscess. Depending on the
severity and nature of the infection, treatments include oral or intravenous antibiotics
and surgical reintervention.[1] The incidence of neurosurgical postoperative infections varies between 0.8 and 7%
in the published literature with several studies attempting to identify the most common
risk factors.[1]
[4]
[5] The mean infection rate for cranial neurosurgery according to a National Survey
of NHS Trusts in England is 4.3%.[6] However, it should be noted that some published studies reporting postoperative
infections include both surgical and medical infections, such as pneumonia or urinary
tract infections.[1] The most common organism responsible according to previous series is Staphylococcus aureus, followed by Coagulase-negative staphylococci and Propionibacterium species.[4] Public Health England has recommended a bundle of care to reduce SSI rates; this
bundle includes among other measures preoperative antibiotics within 60 minutes from
skin incision and a repeat dose for prolonged procedures, intraoperative temperature
monitoring, blood glucose control, and preoperative showering using soap.[7]
Our study aims to clarify which perioperative measures play a role in reducing surgical
infection rates further.
Materials and Methods
The study was registered with our audit department and was compliant with local policy
and national regulations. All data have been anonymized and no item of information
that would enable the identification of any subject was recorded, hence informed consent
of the participants was not sought. We retrospectively included all patients who underwent
an elective craniotomy between August 2018 and August 2019. Emergency craniotomies,
burr hole operations, endoscopic approaches, and spinal procedures were excluded.
The primary endpoint of this retrospective observational study was the SSI rate at
30 days and 4 months after the procedure. SSIs were defined as per Section 3 of the
Protocol for Surveillance of SSI published by Public Health England.[8] This includes superficial incisional infection, deep incisional infection, and space/organ
infection as per the criteria described in the guidelines. For each SSI, we determined
the time from surgery to onset, the pathogen involved, if readmission was needed,
the length of stay, and the treatment required including any further surgical intervention.
Furthermore, we collected data regarding the surgical preparation used and use of
antibiotics both at induction and, for long procedures, the use of repeated doses
of antibiotics.
The craniotomy size was measured on the first postoperative magnetic resonance (MR)
scan or computed tomography scan where MR was not performed, by two of the authors.
Maximum dimensions were measured for height and anteroposterior diameter of the craniotomy.
The larger of the two measures was taken as maximum diameter of the craniotomy (mm)
for the purposes of statistical analysis.
Perioperative Measures
All patients received antibiotic prophylaxis, as per trust policy, consisting of a
single dose of intravenous cefuroxime 1.5 g 30 minutes before the skin incision. Patients
allergic to penicillin received teicoplanin 400 mg intravenously. According to our
local guidelines, a repeat dose of cefuroxime should be administered every 4 hours
during long procedures.
Perioperative Hair Wash with Chlorhexidine Gluconate 4%
Hair wash and skin antiseptic techniques varied across the cohort and depended on
the preference of the operating surgeon. Patients who received perioperative hair
wash, underwent hair wash with chlorhexidine gluconate 4% and lukewarm sterile water
in the anesthetic room postintubation or sedation in case of awake craniotomies. The
hair was washed twice over a 5 minutes period using the above technique, until visibly
clean with no scalp debris. Skin preparation involved the use of alcoholic chlorhexidine,
alcoholic povidone–iodine, or both depending on the operating consultant.
Vancomycin Powder
When vancomycin was used intraoperatively, at the end of the procedure, following
copious irrigation, the bone flap and subgaleal space were dusted with 1 g or 500 mg
vancomycin powder depending on the craniotomy size. This subgroup of patients had
their sutures removed 10 days postoperatively to minimize the risk of wound break
down in view of the known osmotic effect of vancomycin.
Local side effects known to be related to vancomycin use such as aseptic wound breakdown,
seromas, seizures, ototoxicity, nephrotoxicity together with systemic side effects
such as cardiovascular compromise on administration of the antibiotic were recorded.[9]
Patients who had the incision site shaved perioperatively, had a non-adherent dressing
applied as well as head bandage, whereas patients who had hair sparing procedures
had their hair washed with chlorhexidine gluconate and dried with clean towels and
no dressing or bandages were applied. Standard sutures (Vicryl, Ethilon, Prolene,
surgical clips) were used for closures in all procedures depending on surgeon preference.
Sutures were removed on average 10 days postoperatively. No wound drains were placed
as a previous audit in our department showed an increased risk of infection in craniotomy
patients when a wound drain was inserted. Postoperative hair wash with either chlorhexidine
gluconate or warm saline was performed depending on surgeon's preference.
Statistical Analysis
Statistical analysis was performed using SPSS version 25. Patient characteristics
were described using descriptive statistics. Factors predictive of postoperative infection
(any time point) were evaluated in univariate analysis with Fisher's exact and chi-squared
tests. Multivariate regression analysis could not be performed given the low number
of infective events.
Results
344 patients were included in this study. The median age was 56 years (range: 17–87
years). There were approximately equal numbers of males and females (males = 181/344,
53%). More than half of patients were active or previous smokers (181/344, 53%) or
overweight/obese defined by a body mass index of more than 25 or 30 (186/333, 56%).
Most procedures were supratentorial (285/344, 83%) and 18% were redo procedures (61/344).
Other patient characteristics, including demographic and systemic factors that may
predispose to an infection, are shown in [Table 1].
Table 1
Patient characteristics
|
Patient characteristics
|
|
Age (years)
|
|
|
Median
|
56
|
|
Range
|
17–87
|
|
Gender
|
|
|
Male
|
181 (53%)
|
|
Female
|
163 (47%)
|
|
Surgery type
|
|
|
Primary
|
283 (82%)
|
|
Redo
|
61 (18%)
|
|
Approach
|
|
|
Supratentorial
|
285 (83%)
|
|
Skull base
|
32 (9%)
|
|
Posterior fossa
|
27 (8%)
|
|
Risk factors
|
|
|
Diabetes mellitus
|
10 (3%)
|
|
Smoking
|
181 (53%)
|
|
Overweight or obese (BMI ≥25)
|
186/333 (56%)
|
|
Laterality of surgery
|
|
|
Left
|
183 (53%)
|
|
Right
|
135 (39%)
|
|
Midline
|
25 (7%)
|
|
Bilateral
|
1 (< 1%)
|
|
Duration of surgery
|
|
|
Median
|
3 hours 13 minutes
|
|
Range
|
39 minutes – 13 hours, 55 minutes
|
|
Maximum diameter of craniotomy (mm)
|
|
|
Median
|
46
|
|
Range
|
15–200
|
|
Pathology
|
|
|
Low-grade glioma
|
40 (12%)
|
|
High-grade glioma
|
116 (34%)
|
|
Meningioma
|
49 (14%)
|
|
Metastasis
|
73 (21%)
|
|
Schwannoma
|
22 (6%)
|
|
Other neoplasm[a]
|
22 (6%)
|
|
Other[b]
|
22 (6%)
|
Abbreviation: BMI, body mass index.
a Other neoplasms included lymphoma, medulloblastoma, choroid plexus papilloma, hemangioblastoma,
xanthogranuloma, gliosarcoma, neurocytoma, pituitary tumors, pineal tumors, spindle
cell tumor, chondrosarcoma, and rhabdoid tumor.
b The other group included colloid cyst, arachnoid cyst, radiation necrosis, epidermoid
cyst, cortical dysplasia, gliosis/inflammation, dermoid cyst, textiloma, aneurysmal
bone cyst, tegmen defect, displaced bone flap, Rosai Dorfman, inconclusive, cavernoma,
and arteriovenous malformation.
Most procedures were performed for tumors (322/344, 94%), including gliomas (156/344,
45%) or metastases (73/344, 21%) ([Table 1]). All patients received a first dose of antibiotics at induction. The median length
of surgery was 3 hours, 13 minutes; however, a significant proportion of procedures
performed was longer than 4 hours (103/344, 30%) and of these, just over half of patients
(53/103, 51%) received a further dose of antibiotics.
Postoperative Infection
Postoperative infection was observed in 18 patients (5.2%). The majority (15/18, 83%)
were within 30 days, while three further infections occurred within 4 months. None
of the patients who had perioperative hair wash and vancomycin powder before closure
suffered an SSI infection within 4 months. [Table 2] shows factors that were predictive of postoperative infection at any time point.
In univariate analysis, craniotomy size (Fisher's exact test, p = 0.05), lack of perioperative hair wash, and vancomycin powder use (Fisher's exact
test, p = 0.01) were predictive of postoperative infection. Patients with a longer length
of stay showed a trend toward an increased rate of infection, although this was not
statistically significant.
Table 2
Evaluation of factors predictive of postoperative infection
|
Factor
|
Incidence of infection
|
Univariate analysis
|
|
Age
|
≤ 56 years
|
10/174 (5.7%)
|
Fisher's exact test, p = 0.81
|
|
> 56 years
|
8/170 (4.7%)
|
|
Gender
|
Male
|
13/181 (7.2%)
|
Fisher's exact test, p = 0.069
|
|
Female
|
5/163 (3.1%)
|
|
Approach
|
Supratentorial
|
14/285 (4.9%)
|
Chi-squared test= 0.38, p = 0.83
|
|
Skull base
|
2/27 (7.4%)
|
|
Posterior fossa
|
2/32 (6.3%)
|
|
Diabetes
|
No
|
17/334 (5.1%)
|
Fisher's exact test, p = 0.42
|
|
Yes
|
1/10 (10.0%)
|
|
Smoking
|
No
|
7/163 (4.3%)
|
Fisher's exact test, p = 0.48
|
|
Yes
|
11/181 (6.1%)
|
|
Overweight or obese (BMI ≥25)
|
No
|
5/147 (3.4%)
|
Fisher's exact test, p = 0.22
|
|
Yes
|
13/186 (7.0%)
|
|
Pathology
|
Glioma
|
8/156 (5.1%)
|
Chi-squared test = 0.19,
p = 0.67
|
|
Meningioma
|
4/49 (8.2%)
|
|
Metastasis
|
2/73 (2.7%)
|
|
Other[a]
|
4/66 (6.1%)
|
|
Length of stay
|
≤ 4 days
|
6/189 (3.2%)
|
Fisher's exact test, p = 0.09
|
|
> 4 days
|
12/155 (7.7%)
|
|
Redo
|
No
|
14/283 (4.9%)
|
Fisher's exact test, p = 0.54
|
|
Yes
|
4/61 (6.6%)
|
|
Duration of surgery
|
< 4 hours
|
12/241 (5.0%)
|
Fisher's exact test, p = 0.79
|
|
≥ 4 hours
|
6/103 (5.8%)
|
|
Maximum diameter of craniotomy
|
≤ 46 mm
|
5/181 (2.8%)
|
Fisher's exact test, p = 0.05
|
|
> 46 mm
|
13/163 (8.0%)
|
|
Perioperative hair wash[b] and vancomycin powder use
|
No
|
18/263 (6.8%)
|
Fisher's exact test, p = 0.01
|
|
Yes
|
0/81 (0.0%)
|
|
Prep (unknown type in one patient)
|
Chlorhexidine
|
0/81 (0.0%)
|
Chi-squared test= 7.17, p = 0.07
|
|
Alcoholic betadine
|
13/212 (6.1%)
|
|
Combination
|
5/50 (10.0%)
|
|
Hair removal
|
No
|
6/140 (4.3%)
|
Fisher's exact test, p = 0.63
|
|
Yes
|
12/204 (5.9%)
|
|
Drain
|
No
|
18/335 (5.4%)
|
Fisher's exact test, p > 0.99
|
|
Yes
|
0/9 (0.0%)
|
Abbreviation: BMI, body mass index.
Craniotomy size and lack of perioperative hair wash and lack of vancomycin powder
use were predictive of postoperative infection. Variables such as age, length of stay,
and craniotomy size were dichotomized by median values.
a Refers to all other pathology (see [Table 1] for list).
b Preoperative hair wash was performed immediately prior to proceed with surgery in
the anesthetic room using chlorhexidine gluconate.
Perioperative hair wash and vancomycin powder use were always used in conjunction
such that 81 patients underwent perioperative hair wash and also received topical
vancomycin powder. No complications related to the use of vancomycin powder were recorded
within 4 months, including no aseptic wound breakdown and no postoperative seromas.
Subgroup Analysis
Most cases (12/18, 67%) of postoperative infection were observed within the 267 patients
with supratentorial tumors. This group was, therefore, selected for subgroup analysis.
The median age was 58 years (range: 17–87 years). There were approximately equally
numbers of males and females (males = 138/267, 52%). More than half of patients were
active or previous smokers (141/267, 53%) or overweight/obese (139/257, 54%). Nineteen
percent of procedures were redo procedures (50/267).
[Table 3] shows factors that were predictive of postoperative infection in this subgroup.
Lack of perioperative hair wash and vancomycin powder use (Fisher's exact test, p = 0.04) was predictive of postoperative infection.
Table 3
Evaluation of factors predictive of postoperative infection in patients with supratentorial
tumors
|
Factor
|
Incidence of infection
|
Univariate analysis
|
|
Age
|
≤ 58 years
|
7/135 (5.2%)
|
Fisher's exact test, p = 0.77
|
|
> 58 years
|
5/132 (3.8%)
|
|
Gender
|
Male
|
9/138 (6.5%)
|
Fisher's exact test, p = 0.14
|
|
Female
|
3/129 (2.3%)
|
|
Diabetes
|
No
|
12/259 (4.6%)
|
Fisher's exact test, p > 0.99
|
|
Yes
|
0/8 (0.0%)
|
|
Smoking
|
No
|
4/126 (3.2%)
|
Fisher's exact test, p = 0.39
|
|
Yes
|
8/141 (5.7%)
|
|
Overweight or obese (BMI ≥25)
|
No
|
3/118 (2.5%)
|
Fisher's exact test, p = 0.23
|
|
Yes
|
9/139 (6.5%)
|
|
Pathology
|
Glioma
|
7/149 (4.7%)
|
Chi-squared test= 1.28,
p = 0.73
|
|
Meningioma
|
3/45 (6.7%)
|
|
Metastasis
|
2/61 (3.3%)
|
|
Other[a]
|
0/12 (0.0%)
|
|
Length of stay
|
≤ 4 days
|
6/152 (3.9%)
|
Fisher's exact test, p = 0.78
|
|
> 4 days
|
6/115 (5.2%)
|
|
Redo
|
No
|
10/217 (4.6%)
|
Fisher's exact test, p > 0.99
|
|
Yes
|
2/50 (4.0%)
|
|
Duration of surgery
|
< 4 hours
|
8/202 (4.0%)
|
Fisher's exact test, p = 0.49
|
|
≥ 4 hours
|
4/65 (6.2%)
|
|
Maximum diameter of craniotomy
|
≤ 48 mm
|
3/134 (2.2%)
|
Fisher's exact test, p = 0.08
|
|
> 48 mm
|
9/133 (6.8%)
|
|
Perioperative hair wash and vancomycin powder use
|
No
|
12/196 (6.1%)
|
Fisher's exact test, p = 0.04
|
|
Yes
|
0/71 (0.0%)
|
|
Prep (unknown type in one patient)
|
Chlorhexidine
|
0/71 (0.0%)
|
Chi-squared test= 4.85, p = 0.18
|
|
Alcoholic betadine
|
9/155 (5.8%)
|
|
Combination
|
3/40 (7.5%)
|
|
Hair removal
|
No
|
3/113 (2.7%)
|
Fisher's exact test, p = 0.25
|
|
Yes
|
9/154 (5.8%)
|
|
Drain
|
No
|
12/260 (4.6%)
|
Fisher's exact test, p > 0.99
|
|
Yes
|
0/7 (0.0%)
|
Variables such as age, length of stay, and craniotomy size were dichotomized by median
values.
a Refers to all other pathology (see [Table 1] for list).
Management of Infection
Management of postoperative infection was purely medical in 8/18 (44%) patients. With
10/18 (56%) patients requiring a combination of surgical and medical treatment, seven
patients (39%) required removal of their bone flap. The most common organisms were
Staphylococcus aureus 7/18 (39%) and gram-negative organisms 4/18 (22%) such as Enterobacter aerogenes and Klebsiella oxytoca. In the remaining cases, no organism was identified.
Discussion
In this retrospective study of patients undergoing elective craniotomies, we demonstrated
that the use of multiple perioperative surgical measures can significantly reduce
SSI rates. These measures include perioperative hair wash performed immediately prior
to surgery combined with the use of vancomycin powder prior to wound closure.
Our overall infection rate up to 4 months postcraniotomy is 5.2%, which is consistent
with published literature for neurosurgical SSI.[4]
Staphylococcus aureus was the responsible organism in most SSI cases. Within the group of supratentorial
tumors, 12 (6.1%) had an infection postoperatively. Perioperative hair wash with chlorhexidine
and use of vancomycin powder prior to closure of the craniotomy significantly reduced
the risk of infection. None of the patients who underwent hair wash and had topical
vancomycin administered developed an infection within 4 months. Vancomycin is a broad-spectrum
antibiotic that provides cover for gram-positive pathogens that are usually involved
in SSI. Applied as a topical agent vancomycin powder has a local bactericidal action;
it is deemed to be a safe and not excessively expensive way to prevent infection mainly
related to skin flora. Intrawound administration of vancomycin has been extensively
described and demonstrated to be effective in spine and orthopaedic surgery.[10]
[11]
[12] A systematic review and meta-analysis by Bokhari et al found an overall beneficial
effect on SSI incidence with the use of intrawound vancomycin across all neurosurgical
subspecialties except cranioplasty insertion. The studies reported no complications
from the use of vancomycin powder and no evidence that use of vancomycin powder resulted
in the emergence of vancomycin resistant microorganisms.[12] A prospective cohort study by Mallela et al compared the efficacy of 1 g of vancomycin
powder sprinkled over the bone flap and in the subgaleal space after the final irrigation
versus standard of care alone in craniotomy patients. They demonstrated a significant
reduction in SSI in the vancomycin cohort with no side effects.[11] In our center, similarly to Mallela et al, we administered vancomycin powder over
the bone flap and in the subgaleal space above the bone flap after final irrigation
of the surgical cavity. We recorded no side effects related to use of vancomycin powder.
Administration of vancomycin does not prolong the surgical procedure and is cost-effective
in reducing SSI.
In our cohort, patients who received intraoperative vancomycin also underwent double
hair wash just before surgery with chlorhexidine gluconate and lukewarm water. The
Cochrane review by Webster and Osborne[13] published in 2015 concluded that preoperative showering or bathing with chlorhexidine
did not significantly reduce SSI rates compared with other wash products or no wash.
However, this review included over 10,000 patients across multiple surgical specialties
in different geographical locations. Results may not, therefore, be applicable to
elective neurosurgical patients. Furthermore, studies in other specialties such as
gynecology and colorectal surgery have suggested a reduced rate of infection in patients
that have undergone preoperative wash with antiseptic agents.[14]
[15] In our study, perioperative hair wash with chlorhexidine has been proven to be effective
in reducing SSI.
Other factors that showed a nonstatistically significant trend toward reduced infection
rates included small craniotomy size and shorter length of stay. These factors have
been found to be associated with SSI in other studies.[16] Given the small number of infections multivariate analysis was not performed, the
impact of the craniotomy size on the infection rate may have been underestimated.
Limitations of this study include its retrospective nature, the relatively small sample
size, and the fact that this is a single-center study. Data collection relied on the
operation notes; perioperative hair wash may, therefore, have been performed but not
documented in some cases. The low incidence of infections overall meant that robust
multivariate analysis was not possible, though we do not feel this had a significant
impact on results given the small number of variables that were significant at the
univariate level.
Conclusion
This study evaluated the impact of surgical measures on SSI rates in patients undergoing
elective craniotomies. We found that simple measures such as peri-operative hair wash
and use of intrawound vancomycin powder in addition to standard practice can help
to reduce infection rates with negligible risk to patients and acceptable costs. Our
results should be validated further in future prospective studies.
