Keywords
intracranial infection - neurosurgery - risk factors - prognosis
Introduction
Intracranial infection is a common and serious complication of neurosurgery. The incidence
of intracranial infection after neurosurgery can vary from 0.7 to 8.9%, as reported
in different studies.[1]
[2]
[3]
[4] This discrepancy may derive from the choice of different diagnostic criteria, the
presence of multiple clinical factors that may interfere with the prognosis, and regional
differences. China has the largest population in the world, and the highest number
of craniotomies is performed in China. Retrospective studies conducted by Chinese
groups reported that the incidence of intracranial infection is 7.4[5] to 8.6%.[6] Postoperative intracranial infection leads to longer hospitalization, multiple surgeries,
and long-term usage of antibiotics, thus increasing the risks for the patients as
well as the medical costs.[7] Severe infections may lead to neurologic dysfunctions or death.[8] The mortality rate of postoperative intracranial infections varies from 8 to 25%
in the literature[9]
[10]
[11]
[12]
[13] and some medical centers reported a rate of 40.8%.[14] Only a few studies have examined the risk factors that may contribute to the prognosis
of patients affected by a postoperative infection.[9] Therefore, this study aimed to further investigate the possible risk factors correlated
to this clinical event and improve the prognosis for patients who suffer from infection
after neurosurgery.
Materials and Methods
Study Population
This retrospective study was conducted at the People's Hospital of Anyang, Henan Province.
The neurosurgery department of the hospital is a key medical unit in the Province
of Henan and nearly 2,000 neurosurgical operations are performed every year. The study
included data on 113 patients who were diagnosed with a postoperative intracranial
infection after neurosurgical treatment between June 2018 and September 2021. The
research protocol was approved by the hospital ethical committee (approval no. KY2018020),
and the research was carried out in accordance with the ethical standards of the 1964
Helsinki Declaration and its subsequent amendments or similar ethical standards. All
the patients enrolled in this study received distinct ethical approval from the hospital
ethical committee to participate in the retrospective study.
Inclusion and Exclusion Criteria of Patients
The patients were considered eligible for the study according to the following criteria:
(1) the patient underwent neurosurgery and (2) the patient developed a central nervous
system infection after the operation. According to the Centers for Disease Control
and Prevention Standard,[15] postoperative central nervous system infections (PCNSIs) include meningitis, ventriculitis,
and intracranial abscess.[16] The patients were excluded from the study if they had severe open craniocerebral
contaminated incisions or in case of incomplete clinical records that may affect the
analyses.
Data Collection
Demographic information and clinical data of the patients were recorded. Demographic
information included gender and age. Clinical data included the patient's primary
disease, the surgical approach used, the duration of the infection-related surgery
(in hours), the number of infection-related operations, if it was an emergency surgery
or not, the duration of the drainage (in days), the method used for the drainage,
the number of drainage tubes, presence of any leakage of the cerebrospinal fluid (CSF),
the number of antibiotics used for treatment, the duration of usage of antibiotics
(in days), the number of leucocytes in the CSF (×106/l), the number of erythrocytes in the CSF (×106/l), the erythrocytes-to-leucocytes ratio in the CSF, the mononuclear cell percentage
in the CSF, the multinucleated cell percentage in the CSF, the amount of glucose in
the CSF (mmol/l), the total amount of proteins in the CSF (mg/l), the amount of chlorine
in the CSF (mmol/l), the bacteria responsible for the infection, the type of cultured
bacteria (categorized), the time before the incurrence of the infection (in days),
and the duration of infection (in days).
The number of infection-related operations refers to the number of operations that
the patient underwent before the development of an intracranial infection during hospitalization.
The duration of the infection-related surgery is the duration of the most relevant
surgery before the infection. The duration of drainage refers to the length of time
the drainage tube was kept in place before the incurrence of infection. CSF samples
were taken at the time of diagnosis of the infection. The results from bacterial culture
derive from samples taken after the development of infection. The duration of the
infection refers to the time from the diagnosis to the time when the patient was considered
cured of the infection or to the time of death or neurologic deterioration. The time
of discharge from the hospital was selected as the clinical observation endpoint.
Statistical Analysis
To study the influence of the factors correlated to the prognosis of patients with
intracranial infection, we analyzed each factor by univariate analysis. Specifically,
two independent sample t-tests and two independent sample rank-sum tests (Mann–Whitney U test) were used for continuous variables, whereas the chi-squared test and Fisher's
exact test were used for categorical variables. A difference of p < 0.05 was considered significant. A student's t-test was used to analyze age and glucose content in the CSF, given their normal distribution.
Mann–Whitney U test was used to analyze several parameters, including the number of infection-related
operations, duration of infection-related surgeries, number of drainage tubes placed
before the infection, duration of drainage before infection, number of antibiotics
used, duration of antibiotic treatment (in days), number leucocytes in the CSF, number
of erythrocytes in the CSF, ratio between erythrocytes and leucocytes, percentage
of monocytes, percentage of multinucleated cells, total protein content, chlorine
content, number of bacteria, time before infection. A chi-squared test was used to
analyze sex, emergency operations versus nonemergency operations, and the occurrence
of CSF leakage. Fisher's exact test was used to analyze the primary disease, the surgical
approach, the method of drainage, and the results of bacterial culture. All data were
analyzed using the SPSS software (version 25.0). Multivariate logistic regression
analyses were performed on the influencing factors that presented significant differences.
Results
Patient Demographics
Data on 113 patients, including 58 males (51.3%) and 55 females (48.7%), were collected
and investigated. The minimum age of the patients was 15 years and the maximum age
was 85 years, with an average age of 56.07 ± 12.65 years (see [Table 1] for details). These patients were affected by 15 primary diseases (see [Table 2] for details).
Table 1
Incidence and recovery rate in 113 patients with postoperative infection
|
Age group (y)
|
Infected patients
|
Proportion (%)
|
Cumulative ratio (%)
|
Cured patients
|
Cure ratio (%)
|
|
15∼
|
1
|
0.9
|
0.9
|
1
|
100
|
|
25∼
|
6
|
5.3
|
6.2
|
5
|
83.3
|
|
35∼
|
10
|
8.8
|
15.0
|
9
|
90
|
|
45∼
|
31
|
27.4
|
42.5
|
26
|
83.9
|
|
55∼
|
37
|
32.7
|
75.2
|
31
|
83.8
|
|
65∼
|
21
|
18.6
|
93.8
|
14
|
66.7
|
|
75–85
|
7
|
6.2
|
100
|
6
|
85.7
|
Table 2
Incidence and cure rate in 113 patients grouped by primary disease
|
Primary
|
Infected patients
|
Proportion (%)
|
Cured patients
|
Cure ratio (%)
|
|
Cerebral hemorrhage
|
58
|
51.3
|
43
|
74.1
|
|
Aneurysm
|
22
|
19.5
|
18
|
81.8
|
|
Glioma
|
8
|
7.1
|
8
|
100
|
|
Trigeminal neuralgia
|
6
|
5.3
|
6
|
100
|
|
Brain injury
|
4
|
3.5
|
3
|
75
|
|
Pituitary tumor
|
3
|
2.7
|
3
|
100
|
|
Facial spasm
|
2
|
1.8
|
2
|
100
|
|
Meningiomas
|
2
|
1.8
|
1
|
50
|
|
Acoustic neuroma
|
2
|
1.8
|
2
|
100
|
|
Arteriovenous malformation
|
1
|
0.9
|
1
|
100
|
|
Cavernous hemangioma
|
1
|
0.9
|
1
|
100
|
|
Craniopharyngioma
|
1
|
0.9
|
1
|
100
|
|
Ependymal tumor
|
1
|
0.9
|
1
|
100
|
|
Pineal tumor
|
1
|
0.9
|
1
|
100
|
|
Arachnoid cyst
|
1
|
0.9
|
1
|
100
|
Due to the diversity in etiology, pathology, and site of diagnosis, diagnosis of the
primary diseases was divided into the following categories: 81 cases (71.7%) of spontaneous
intracranial hemorrhage, 19 cases of intracranial tumor (16.8%), 4 cases of traumatic
brain injury (3.5%), 6 cases of trigeminal neuralgia (5.3%), and others. Similarly,
the types of surgeries that preceded the infection were classified according to the
surgical method utilized: 38 cases of craniotomy (33.6%), 36 cases of ventricular
drainage (31.9%), 13 cases of craniotomy plus ventricular drainage (11.5%), 9 cases
of decompressive craniectomy (8.0%), 7 cases of lumbar drainage (6.2%), 5 cases of
ventricular drainage plus lumbar drainage (4.4%), 4 cases of craniotomy plus lumbar
drainage (3.5%), and 1 case (0.9%) of decompressive craniectomy plus ventricular drainage.
Among all the surgeries performed, 75 cases (66.4%) were categorized as emergency
surgeries and 38 cases (33.6%) as nonemergency surgeries (see [Table 3] for details).
Table 3
Incidence and cure rate in 113 patients grouped by surgical method
|
Surgical approach
|
Infected patients
|
Proportion (%)
|
Cumulative ratio (%)
|
Cured patients
|
Cure ratio (%)
|
|
Craniotomy
|
38
|
33.6
|
33.6
|
35
|
92.1
|
|
Ventricular drainage
|
36
|
31.9
|
65.5
|
26
|
72.2
|
|
Craniotomy and ventricular drainage
|
13
|
11.5
|
77.0
|
9
|
69.2
|
|
Decompressive craniectomy
|
9
|
8.0
|
85.0
|
7
|
77.8
|
|
Lumbar drainage
|
7
|
6.2
|
91.2
|
7
|
100.0
|
|
Ventricular drainage and lumbar drainage
|
5
|
4.4
|
95.6
|
4
|
80.0
|
|
Craniotomy and lumbar drainage
|
4
|
3.5
|
99.1
|
3
|
75.0
|
|
Decompressive craniectomy and ventricular drainage
|
1
|
0.9
|
100.0
|
1
|
100.0
|
Microbiology
Samples from 108 patients were collected for bacterial culture. Fifty-one cases (45.1%)
were negative and 57 cases (48.7%) were positive for bacterial contamination, and
7 cases (6.2%) were not analyzed. Among the bacteria species detected, the most common
bacteria were Acinetobacter baumannii (11.5% with 14 cases), Klebsiella pneumonia (7.1% with 8 cases), the gram-positive bacilli by smear (2.7% with 3 cases), and
Pseudomonas aeruginosa (2.7% with 3 cases).
Risk Factors
After univariate analyses, the parameters that emerged as possible risk factors affecting
the prognosis were whether the surgery was an emergency operation or not, the number
of infection-related operations, the duration of the drainage (in days) before the
infection, the number of antibiotics used for the treatment, the duration of antibiotic
treatment (in days), the number of leucocytes in the CSF, and the glucose content
in the CSF (mmol/l), the total protein content in the CSF (mg/l), the cultured bacteria,
the number of cultured bacteria, and duration of infection (p < 0.05; see [Table 4] for univariate analysis results).
Table 4
Univariate analysis results
|
Outcome
|
X
2/t/Z
|
p value
|
|
Deteriorated (n = 21)
|
Cured (n = 92)
|
|
Sex
|
|
Female
|
9 (16.4%)
|
46 (83.6%)
|
0.349
|
0.555
|
|
Male
|
12 (20.7%)
|
46 (79.3%)
|
|
Age (y)
|
58.33 ± 11.45
|
55.55 ± 12.91
|
0.908
|
0.366
|
|
Primary disease
|
|
Intracranial hemorrhage
|
19 (23.5%)
|
62 (76.5%)
|
–
|
0.107
|
|
Intracranial trauma
|
1 (25.0%)
|
3 (75.0%)
|
|
Intracranial tumor and cysts
|
1 (5.0%)
|
19 (95.0%)
|
|
Mimic convulsion and trigeminal neuralgia
|
0
|
8 (100.0%)
|
|
Surgical approach
|
|
Craniotomy
|
3 (7.9%)
|
35(92.1%)
|
–
|
0.197
|
|
Decompressive craniectomy
|
2 (22.2%)
|
7(77.8%)
|
|
|
Ventricular drainage
|
10 (27.8%)
|
26 (72.2%)
|
|
|
Craniotomy + ventricular drainage
|
4 (30.8%)
|
9 (69.2%)
|
|
|
Decompressive craniectomy + ventricular drainage
|
0
|
1 (100.0%)
|
|
|
Lumbar drainage
|
0
|
7 (100.0%)
|
|
|
Craniotomy + lumbar drainage
|
1 (25.0%)
|
3 (75.0%)
|
|
|
Ventricular drainage + lumbar drainage
|
1 (20.0%)
|
4 (80.0%)
|
|
|
Emergency
|
|
No
|
3 (7.9%)
|
35 (92.1%)
|
4.324
|
0.038
|
|
Yes
|
18 (24.0%)
|
57 (76.0%)
|
|
Number of infection-related operations
|
2 (1, 2)
|
1 (1, 2)
|
–2.285
|
0.022
|
|
Duration of the infection-related surgery
|
1.00 (0.59, 3.71)
|
3.25 (1.00, 4.25)
|
–1.618
|
0.106
|
|
Type of drainage
|
|
No
|
0
|
18 (100.0%)
|
–
|
0.180
|
|
Ventricular drainage
|
11 (28.9%)
|
27 (71.1%)
|
|
Subcutaneous drainage
|
4 (16.0%)
|
21 (84.0%)
|
|
Lumbar drainage
|
1 (14.3%)
|
6 (85.7%)
|
|
Ventricular drainage + subcutaneous drainage
|
2 (15.4%)
|
11 (84.6%)
|
|
Ventricular drainage + lumbar drainage
|
2 (33.3%)
|
4 (66.7%)
|
|
Subcutaneous drainage + lumbar drainage
|
1 (20.0%)
|
4 (80.0%)
|
|
Ventricular drainage + subcutaneous drainage + lumbar drainage
|
0
|
1 (100.0%)
|
|
Number of tubes placed before infection
|
1.0 (1.0, 1.5)
|
1.0 (1.0, 1.0)
|
–1.378
|
0.168
|
|
Duration of the drainage before infection (days)
|
9 (4, 19)
|
4 (2, 8)
|
–2.914
|
0.004
|
|
CSF leakage
|
|
Yes
|
17 (18.5%)
|
75 (81.5%)
|
0.000
|
0.994
|
|
No
|
2 (14.3%)
|
12 (85.7%)
|
|
Number of antibiotics
|
2 (2, 2)
|
2 (1, 2)
|
–2.050
|
0.040
|
|
Duration of antibiotic treatment (in d)
|
10 (3, 18)
|
15 (9, 24)
|
–2.339
|
0.019
|
|
Leucocytes in CSF (106/l)
|
38.40 (15.25, 122.40)
|
13.30 (6.00, 24.85)
|
–3.126
|
0.002
|
|
Erythrocytes in CSF (106/l)
|
2,680 (160, 7,900)
|
2,370 (83, 22,660)
|
–0.532
|
0.595
|
|
Erythrocytes-to-leucocytes ratio
|
0.61 (0.02, 2.79)
|
1.65 (0.11, 11.06)
|
–1.613
|
0.107
|
|
Monocyte percentage
|
22.5 (20.0, 30.0)
|
22.5 (20.0, 30.0)
|
–0.370
|
0.712
|
|
Multinucleated cell percentage
|
77.5 (70.0, 80.0)
|
77.5 (70.0, 80.0)
|
–0.370
|
0.712
|
|
Glucose content (mmol/l)
|
1.12 ± 1.44
|
2.07 ± 1.17
|
–3.176
|
0.002
|
|
Total protein content (mg/l)
|
37.08 (21.35, 67.31)
|
19.77 (11.37, 34.06)
|
–2.964
|
0.003
|
|
Chlorine (mmol/l)
|
116.65 (110.23, 122.00)
|
119.10 (115.03, 124.00)
|
–1.466
|
0.143
|
|
Bacterial culture
|
|
Negative
|
2 (3.9%)
|
49 (96.1%)
|
–
|
0.000
|
|
Gram negative
|
14 (48.3%)
|
15 (51.7%)
|
|
Gram positive
|
1 (5.6%)
|
17 (94.4%)
|
|
Mixed bacteria
|
1 (16.7%)
|
5 (83.3%)
|
|
Candida albicans
|
1 (50.0%)
|
1 (50.0%)
|
|
Uncultivated
|
2 (28.6%)
|
5 (71.4%)
|
|
Number of bacteria
|
1 (1, 1)
|
0 (0, 1)
|
–2.958
|
0.003
|
|
Timespan before infection (in d)
|
9.0 (5.5, 21.0)
|
7.0 (4.3, 10.0)
|
–1.537
|
0.124
|
|
Duration of infection
|
10.0 (2.5, 16.5)
|
12.0 (8.0, 18.0)
|
–1.967
|
0.049
|
CSF; cerebrospinal fluid. Note: t-test: age and glucose content in CSF (mmol/l).
U test: Number of infection-related operations, duration of the infection-related surgery,
number of tubes placed before the infection, duration of the drainage before the infection,
number of antibiotics, duration of antibiotic treatment (in d), leucocytes in the
CSF, erythrocytes in the CSF, erythrocytes-to-leucocytes ratio, monocyte percentage,
multinucleated cell percentage, total protein content, chlorine content, number of
bacteria, and timespan before infection.
Chi-squared test: sex, whether it was an emergency surgery or not, and CSF leakage.
Fisher's exact test: primary disease, surgical approach, drainage method, and results
of bacterial culture.
The prognosis (cured/deteriorated: 1/0) was used as the dependent variable and the
influencing factors in the univariate analysis were the independent variables. Multivariate
logistic regression analysis was performed using the stepwise method (forward: Likelihood
Ratio test [LR]). The results indicated that the independent risk factors affecting
prognosis include the duration of drainage before infection, number of antibiotics
used for treatment, duration of antibiotic treatment (in days), and the number of
leucocytes in the CSF (×108; p < 0.05). In particular, the duration of drainage (B: –0.113; odds ratio [OR]: 0.893;
95% confidence interval [CI]: 0.805–0.991; p = 0.033), the number of antibiotics used for treatment (B: –1.470; OR: 0.230; 95%
CI: 0.072–0.738; p = 0.013), and the number of leucocytes in the CSF (B: –0.016; OR: 0.984; 95% CI:
0.970–0.998; p = 0.027) are all risk factors that inversely correlate with a positive prognosis
for the patient. On the contrary, the duration of antibiotic treatment (B: 0.176;
OR: 1.193; 95% CI: 1.063–1.339; p = 0.003) was shown to positively correlate with a good prognosis for patients affected
by an intracranial infection. The results of the logistic meaningful analyses are
shown in [Table 5].
Table 5
Logistic regression analysis results
|
B
|
S.E.
|
Wald
|
df
|
Sig.
|
Exp(B)
|
95% CI for Exp(B)
|
|
|
|
|
|
|
|
Lower
|
Upper
|
|
Drainage duration before infection
|
–0.113
|
0.053
|
4.567
|
1
|
0.033
|
0.893
|
0.805
|
0.991
|
|
Number of antibiotics
|
–1.470
|
0.595
|
6.104
|
1
|
0.013
|
0.230
|
0.072
|
0.738
|
|
Duration of antibiotic treatment
|
0.176
|
0.059
|
8.971
|
1
|
0.003
|
1.193
|
1.063
|
1.339
|
|
Leucocytes in CSF (108/l)
|
–0.016
|
0.007
|
4.893
|
1
|
0.027
|
0.984
|
0.970
|
0.998
|
Abbreviations: CI, confidence interval; CSF, cerebrospinal fluid; S.E., standard error.
Discussion
Postoperative Central Nervous System Infections (PCNSIs) are a serious complication
of neurosurgery that needs further understanding. However, little is known about the
clinical risk factors that may affect the prognosis of patients affected by a postoperative
intracranial infection. In a retrospective study of 115 patients with gram-negative
meningitis, it was reported that factors that correlated to a 30-day mortality or
neurologic deterioration included the following: days from admission to meningitis
development, consciousness at diagnosis, CSF glucose levels <50 mg/dL at diagnosis,
higher creatinine levels, and blood glucose levels lower than 180 mg/dL.[9] In another retrospective study involving 3,580 patients after neurosurgery, the
risk factors affecting the prognosis were low CSF glucose content (<20% of blood glucose),
increased Acute Physiology and Chronic Health Evaluation III (APACHE III) score, and
the presence of gram-negative bacteria.[17] In comparison, the results of multivariate logistic regression analyses performed
in our study showed that the duration of the drainage before the development of infection
(in days), the duration of the antibiotic treatment (days), the number of antibiotics
used for treatment, and the number of leucocytes in the CSF are significantly different
in patients with a positive or negative prognosis. Furthermore, the glucose content
in the CSF (mmol/L) showed a significant difference between the two patient groups
in the univariate analysis. However, the results of multivariate analysis showed that
glucose content was not a risk factor for a poorer course of postoperative infection,
which was inconsistent with the results presented in the aforementioned articles,
where low levels of glucose were correlated to a deterioration in the prognosis. The
discrepancy in the results may derive from the methodological differences, as in our
study we regarded the CSF glucose content as a continuous variable, while in the previous
articles it was analyzed in different range levels.
The duration of the antibiotic treatment (in days) was positively correlated with
a good prognosis. Previous literature indicated that early treatment with antibiotics
for intracranial infection is beneficial for a patient's prognosis,[18] and the duration of antibiotic treatment is mainly based on the patient's needs
and on the types of pathogenic bacteria. It was previously reported that the duration
of antibiotic application needs to meet relevant clinical indicators to ensure the
achievement of a cure from the infection.[19]
The duration of drainage before the incurrence of infection (in days) is also a risk
factor for a good prognosis. It was shown before that placement of the drainage tube
and its duration are risk factors for intracranial infection after neurosurgery. In
particular, the longer the duration of the drainage, the higher the rate of infection.[20] Interestingly, our data indicate that the duration of drainage before the infection
is also a predictor of a poor prognosis, as the longer the duration of drainage, the
worse the prognosis. Due to the limited sample size, an accurate prediction model
on the duration of drainage has not been established. However, previous literature
suggested that the risk of external ventricular drainage-related meningitis increases
by 3.832-fold, if the drainage was used for 5 days, and that the risk of postoperative
meningitis increases by 11.492-fold, if used for >7 days. .[20] Therefore, surgeons should pay special attention to patients with >5 days of external
ventricular drainage or >7 days of lumbar drainage.
The number of leucocytes in the CSF (×108/l) at the time of diagnosis of infection was also a risk factor against complete
healing of the patients. To make the results of the logistic analyses meaningful,
we convert units from 106 to 108/l. If the number of leucocytes is increased when the infection is diagnosed, the
probability of recovery from infection becomes higher.
The number of antibiotics used for treatment represents the total number of antibiotics
administrated to the patient from the time of diagnosis to the clinical observation
endpoint. We found that when the number of antibiotics used increases, the prognosis
is more likely to worsen. This was not previously reported in the literature. Further
analyses of this group of patients revealed that these individuals were often seriously
ill. As the infection was not easy to control, multiple bacterial cultures and several
changes in the type of antibiotics were performed.
Diabetes is analyzed as a risk factor in some studies on intracranial infection after
neurosurgery. A retrospective study of 115 patients with gram-negative meningitis
conducted by Neuberger et al shows that blood glucose >180 mg/dL was associated with
an unfavorable outcome on multivariate analysis.[9] In another retrospective study involving 3,580 patients after neurosurgery, the
risk factors affecting the prognosis were low CSF glucose content, increased APACHE
III score, and the presence of gram-negative bacteria. In this study, diabetes was
not included in the risk factors analyzed. Unfortunately, there is no patient with
diabetes and obesity enrolled in our study currently, so diabetes was not included
in the risk factors analyzed. Indeed, the experience of surgeons is also a risk factor,
which is not often considered in the literature, probably because the experience cannot
be accurately quantified.
The main limitation of this study is that all the cases collected belonged to the
same hospital center. Therefore, further analyses of patients coming from other centers
are necessary to confirm our results.
Conclusions
According to our analyses, an increase in the duration of drainage before infection,
the number of antibiotics used for treatment, and the number of leucocytes in the
CSF are all factors correlated with a deterioration of the prognosis for patients
with postoperative cranial infection. On the contrary, when the duration of antibiotic
treatment increases, chances of recovery also become higher. This study will help
clinicians to identify patients at risk of poor prognosis and, thus, improve the rate
of recovery thanks to early intervention.
