Cerebrospinal Fluid Cultures in Traumatic Brain Injury: Is It Worth It? A Two-Center Study

Abstract

Background:

Patients with traumatic brain injury (TBI) frequently develop leukocytosis, fever, and tachycardia that may lead to extensive medical investigations to rule out an infectious process. Cerebrospinal fluid (CSF) is often acquired during this workup, however, the utility of this practice has not been studied previously. We hypothesized that CSF cultures would unlikely yield positive results in patients with TBI.

Patients and Methods:

A retrospective review was conducted of all patients with TBI admitted to two level 1 trauma centers at urban, academic institutions from January 2009 to December 2016. Data collected included patient demographics, presenting Glasgow Coma Score (GCS), injury profile, injury severity scores (ISS), regional abbreviated injury scale (AIS), hospital and intensive care unit (ICU) length of stay (LOS), ventilator days, and culture results. For purposes of the analysis, CSF cultures with Staphylococcus epidermidis, Staphylococcus aureus, or Candida underwent a chart review and were considered contaminates if indicated.

Results:

There were 145 patients who had CSF cultures obtained with a median age of 39 years; 77.2% were male. The majority of patients presented after blunt trauma with median GCS of 6, head AIS of 4, and ISS of 25. These patients had prolonged median ICU and hospital stays at 13 and 22 days, respectively. Six (4.1%) CSF cultures demonstrated growth. Four (2.8%) were deemed contaminants, with two growing Staphylococcus epidermidis only, one with both Staphylococcus epidermidis and Staphylococcus aureus, and one with Candida. Two cultures (1.4%) were positive and grew Enterobacter cloacae. Of note, both patients had prior instrumentation with an external ventricular drain.

Conclusion:

Obtaining CSF cultures in patients with TBI is of low yield, especially in patients without prior external ventricular drain. Other sources of infectious etiologies should be considered in this patient population.

Traumatic brain injury (TBI) contributes to substantial morbidity and mortality in the United States [1,2]. In addition to the consequences of the immediate injury, patients with TBI often have prolonged hospitalizations partially because of depressed mental and functional status [3,4]. Hospitalizations are further complicated by infectious complications secondary to the immunosuppressive effects of TBI [5,6]. A higher rate of sepsis is seen within this patient population [7]. These patients often develop leukocytosis, fever, tachycardia, and changes in mental status that may lead to extensive medical investigations to identify an infectious source for pneumonia, urinary tract infection, surgical site infection, bacteremia, and meningitis [8].

National health care expenditures have increased substantially and were estimated to be $3.6 trillion in 2018 alone [9]. Given the concomitant increase in health care costs over recent years, there has been a significant focus on identifying sources of resource overutilization [10,11]. One such source is with laboratory testing, which may be of equivocal clinical benefit in certain contexts [12,13]. Cerebrospinal fluid (CSF) is often acquired in patients with a neurologic insult to rule out meningitis, at times using invasive procedures. This has been shown to serve a limited role at best in critically ill neurologic or neurosurgical patients [14]. The value of acquiring CSF in patients with TBI has not been investigated to date. We hypothesized that CSF cultures would unlikely yield positive results in patients with TBI.

Patients and Methods

A retrospective review at two urban, academic level 1 trauma centers was conducted of all patients with TBI admitted from January 2009 to December 2016 who had a CSF sample taken. Patients who had a specimen retrieved from a subgaleal tap, shunt, or a lumbar drain were excluded. Data were extracted from the electronic medical record when available. Initial data collection included patient demographics, Glasgow Coma Score (GCS) at the time of presentation, mechanism of injury, regional abbreviated injury scale (AIS), and injury severity scores (ISS). Radiographic imaging and reports were reviewed to identify the nature of brain injury for each patient, specifically identifying the presence of a contusion, epidural hematoma, intra-parenchymal hemorrhage, intra-ventricular hemorrhage, subarachnoid hemorrhage, and/or subdural hematoma. The presence of concurrent instrumentation devices, such as an external ventricular drain, were noted, as well as the method by which the specimen was collected. Outcome data including hospital and intensive care unit (ICU) length of stay, need for mechanical ventilation, ventilator days, operations and procedures performed, and mortality were collected. Patients with missing radiographic or outcome data were included in the study if the results of their CSF cultures were available.

Individual CSF culture results were reviewed for the presence of any microbial growth. Given the limitations of the data available through the electronic record for a majority of patients, cell counts were not used to correlate for an infection. Provider notes were also reviewed to determine the reason for obtaining the CSF culture and the clinical context when available. For the purposes of the analysis, CSF cultures with Staphylococcus epidermidis, Staphylococcus aureus, or Candida were considered contaminated and deemed negative.

Data are summarized as percentages for categorical variables and means with standard deviations (SD) and medians with interquartile ranges (IQR) for continuous variables. This study was approved by the Institutional Review Boards at both institutions.

Results

There were 145 patients with TBI who had a CSF culture during their hospitalization from January 2009 to December 2016. Of these, 85 (58.6%) patients had incomplete data available through the electronic medical record but were included in the analysis because the results of their CSF cultures were available. The median age was 39 years and 77.2% were male (Table 1). The median GCS at the time of presentation was 6 and the most common mechanism of injury was motor vehicle collision (21.4%), followed by ground-level fall (17.9%). The majority of patients (97.9%) had a blunt mechanism of injury. Nearly half the patients had a subarachnoid hemorrhage (51.7%) or subdural hematoma (50%). An intra-parenchymal hemorrhage was present in 35% of patients and a contusion in 25%. Median head/neck AIS and ISS were 4 and 25, respectively.

Table 1.

Patient Characteristics (n = 145)

Age, y
Mean ± SD	43.8 ± 20.8
Median (IQR)	39 (25–60.5)
Male, n (%)	112 (77.2)
Initial GCS at presentation
Mean ± SD	7.8 ± 4.7
Median (IQR)	6 (3–13)
Mechanism of injury, n (%)
Assault	7 (4.8)
Automobile vs. bicycle collision	2 (1.4)
Automobile vs. pedestrian collision	11 (7.6)
Ground-level fall	26 (17.9)
Fall from height	13 (9.0)
Found down	9 (6.2)
Motorcycle collision	17 (11.7)
Motor vehicle collision	31 (21.4)
Other	29 (20.0)
Blunt, n (%)	142 (97.9)
Type of brain injury, n (%)^a,b
Contusion	15 (25.0)
Epidural hematoma	5 (8.3)
Intra-parenchymal hemorrhage	21 (35.0)
Intra-ventricular hemorrhage	6 (10.0)
Subarachnoid hemorrhage	31 (51.7)
Subdural hematoma	30 (50.0)
Head/neck AIS
Mean ± SD	4.0 ± 1.3
Median (IQR)	4 (4-5)
Face AIS
Mean ± SD	0.8 ± 1.0
Median (IQR)	0 (0–2)
Chest AIS
Mean ± SD	1.0 ± 1.5
Median (IQR)	0 (0–3)
Abdomen/Pelvis AIS
Mean ± SD	0.3 ± 0.9
Median (IQR)	0 (0–0)
Extremity AIS
Mean ± SD	0.9 ± 1.2
Median (IQR)	0 (0–2)
External AIS
Mean ± SD	0.4 ± 0.5
Median (IQR)	0 (0–1)
ISS
Mean ± SD	25.1 ± 12.5
Median (IQR)	25 (17–34)

AIS = abbreviated injury scale; GCS = Glasgow Coma Scale; IQR = interquartile range; ISS = injury severity scale; SD = standard deviation.

More than one brain injury possible for each patient.

Data available for 60 patients.

Overall, the patients had prolonged ICU and hospital stays at 13 and 22 days, respectively (Table 2). The majority (74.5%) required mechanical ventilation for a median of 10.5 days and 41 patients (68.3%) underwent at least one procedure. The most commonly performed procedures included placement of an external ventricular drain (60.0%), tracheostomy (51.7%), and a gastrostomy (48.3%). Fewer patients required placement of an intra-cranial pressure monitor (11.7%). Craniotomy (21.7%) and craniectomy (22.0%) were performed in approximately one of five patients. Overall mortality was 6.2%.

Table 2.

Patient Outcomes (n = 145)

Hospital LOS, d
Mean ± SD	28.3 ± 26.5
Median (IQR)	22 (16–33.5)
ICU LOS, d
Mean ± SD	16.5 ± 13.0
Median (IQR)	13 (9–21)
Required mechanical ventilation, n (%)	108 (74.5)
Ventilator days
Mean ± SD	12.6 ± 9.1
Median (IQR)	10.5 (7–17)
Operations/Procedures, n (%)^a,b
Craniectomy	12 (20.0)
Craniotomy	13 (21.7)
External ventricular drain	36 (60.0)
Gastrostomy	29 (48.3)
Intra-cranial pressure monitor	7 (11.7)
Tracheostomy	31 (51.7)
Other	10 (16.7)
Mortality, n (%)	9 (6.2)

IQR = interquartile range; LOS = length of stay; SD = standard deviation.

More than one operation or procedure was possible for each patient.

Data available for 60 patients.

Cerebrospinal fluid cultures were most commonly acquired in the setting of fevers or leukocytosis (Table 3). Other reasons included concern for CSF leak (6.7%), changes in mental status (5.0%), and for a seizure workup (1.7%). These samples were most often acquired from an external ventricular drain (63.3%) followed by lumbar puncture (36.7%). Microbiology results demonstrated no growth in most the samples (95.9%). Four (2.8%) cultures grew Candida, Staphylococcus aureus, and Staphylococcus epidermidis, all of which were considered contaminants per chart review. There were two (1.4%) positive cultures that grew Enterobacter cloacae. Both patients with positive cultures had previous instrumentation with an external ventricular drain. In addition, they had undergone a craniectomy or craniotomy.

Table 3.

Culture Results (n = 145)

CSF results, n (%)
No growth	139 (95.9)
Contaminant	4 (2.8)
Candida	1 (0.7)
Staphylococcus aureus	1 (0.7)
Staphylococcus epidermidis	3 (2.1)
Positive	2 (1.4)
Enterobacter cloacae	2 (1.4)
Reason for obtaining, n (%)^a
Fevers and/or leukocytosis	54 (90.0)
Concern for CSF leak	4 (6.7)
Mental status	3 (5.0)
Seizures	1 (1.7)
Method used to obtain culture, n (%)^a
External ventricular drain	38 (63.3)
Lumbar puncture	22 (36.7)

CSF = cerebrospinal fluid.

Data available for 60 patients

Other cultures acquired within three days of the CSF culture were also reviewed. Blood cultures were sent in 90.3% of patients, whereas sputum and urine cultures were sent in 75.2% of patients. Of all 145 patients, 39 (26.9%) had at least one positive culture: 22.1% had a positive sputum, 9.0% had a positive blood, and 4.1% had a positive urine culture during that time frame (Table 4). Patients were more likely to have a positive sputum or blood culture than a positive CSF culture. One of the two patients with a CSF culture had two sputum cultures with a different Enterobacter organism, whereas the other patient had no additional concomitant positive cultures.

Table 4.

Concomitant Cultures Obtained Near the Time CSF Culture Obtained (±3 days; n = 145)

Culture	n (%)	p^a
Blood cultures
Blood culture sent, n (%)	131 (90.3)	NA
Blood culture positive, n (%)	13 (9.0)	< 0.01
CSF cultures
CSF culture sent, n (%)	145 (100.0)	NA
CSF culture positive, n (%)	2 (1.4)	NA
Sputum cultures
Sputum culture sent, n (%)	109 (75.2)	NA
Sputum culture positive, n (%)	32 (22.1)	<0.01
Urine cultures
Urine culture sent, n (%)	109 (75.2)	NA
Urine culture positive, n (%)	6 (4.1)	0.28

CSF = cerebrospinal fluid; NA = not applicable.

Comparison to prevalence of a positive CSF culture.

Discussion

Infectious complications are often present during the hospitalization of patients with TBI. Even though meningitis is considered in the differential diagnosis, only 1.4% of patients had a positive culture, implying the low utility of obtaining CSF cultures. Importantly, only patients with a prior external ventricular drain had a positive culture.

Although CSF cultures were most commonly obtained from external ventricular drains, a substantial proportion of patients required lumbar punctures to acquire the specimen. Several complications related to lumbar punctures including post-lumbar puncture headaches in up to one-quarter of patients, bleeding, and herniation [15]. Nearly one-third of the patients in this study required a lumbar puncture to obtain CSF, increasing the risk for these complications with low apparent clinical benefit.

Other sources of infectious etiologies in patients with TBI such as a positive sputum culture are much more likely than CSF infection. Patients with TBI are highly susceptible to nosocomial infections because they often need mechanical ventilation, prolonged ICU care, and hospitalizations [16]. In addition to this risk, neurogenic fevers are frequent with an incidence estimated to be between 4% and 37% among survivors of TBI [17]. Neurogenic fevers occur because of alterations in the hypothalamic set point temperature, resulting in a dysregulated, hyperthermic state. It has been suggested that neurogenic fevers may reflect some degree of injury to the hypothalamus [18 –20].

Although obtaining CSF cultures routinely may not be indicated in all patients with TBI, one may argue that obtaining these studies may be warranted in patients with prior instrumentation. Bota et al. [21] describe a six-year study wherein they identified 9% of patients who had external ventricular drains with bacterial growth in their CSF cultures, 11% with colonization, and 14% with contamination. A similar rate of positive cultures was found in a review by Lozier et al. [22]. Additional risk factors associated with a positive culture included intra-ventricular hemorrhage, subarachnoid hemorrage, facial fractures with concomitant CSF leak, catheter irrigation, craniotomy, and catheterization duration [22]. With the limited data available in our dataset, we were able to ascertain that one patient had both an intra-ventricular hemorrhage and subarachnoid hemorrhage, and three patients had either a craniectomy or craniotomy. It may be reasonable to consider acquiring CSF cultures in the patients with instrumentation with the previously mentioned risk factors if the suspicion for CSF infection remains high.

There are several limitations to this study, including the retrospective nature with a small sample size. Furthermore, data availability was limited for the majority of patients, limiting our ability to frame these patients within a clinical context or to draw further conclusions. The decision to send certain cultures was highly dependent upon the clinician. Although we did demonstrate low clinical utility with CSF cultures in TBI patients, we were unable to demonstrate any adverse events associated with obtaining them. The prevalence of concomitant positive cultures may have been underestimated because not all patients had related blood, sputum, or urine cultures drawn. However, any additional positive cultures would support our conclusion that CSF cultures are of low clinical value in patients with TBI. Further prospective, multicenter studies with a larger sample size are required to validate our findings.

Conclusions

Obtaining CSF cultures in patients with TBI is of low clinical yield. Routinely sending CSF for culture leads to additional laboratory cost and potentially a lumbar puncture. During the almost seven-year study period, only 1.4% of CSF cultures led to a related infectious diagnosis and these were in patients who had a prior external ventricular drain. Other sources of infectious etiologies, especially respiratory related, should be considered in this patient population.

Footnotes

Acknowledgments

This manuscript was a quickshot presentation at the 14th Annual Academic Surgical Congress in Houston, Texas, February 6, 2019.

Funding Information

No financial support was received in relation to this article.

Author Disclosure Statement

The authors have no conflicts of interest to report.

References

Iorio-Morin

, Fortin

, Blanchard

. TBI prognosis calculator: A mobile application to estimate mortality and morbidity following traumatic brain injury. Clin Neurol Neurosurg, 2016; 142:48–53.

Taylor

, Bell

, Breiding

, Xu

. Traumatic brain injury-related emergency department visits, hospitalizations, and deaths—United States, 2007 and 2013. MMWR Surveill Summ, 2017; 66:1–16.

Colantonio

, Croxford

, Farooq

, et al. Trends in hospitalization associated with traumatic brain injury in a publicly insured population, 1992–2002. J Trauma, 2009; 66:179–183.

Corral

, Javierre

, Ventura

, et al. Impact of non-neurological complications in severe traumatic brain injury outcome. Crit Care, 2012; 16:R44.

Meisel

, Schwab

, Prass

, et al. Central nervous system injury-induced immune deficiency syndrome. Nat Rev Neurosci, 2005; 6:775–786.

Roquilly

, David

, Cinotti

, et al. Role of IL-12 in overcoming the low responsiveness of NK cells to missing self after traumatic brain injury. Clin Immunol, 2017; 177:87–94.

Mascia

, Sakr

, Pasero

, et al. Extracranial complications in patients with acute brain injury: A post-hoc analysis of the SOAP study. Intensive Care Med, 2008; 34:720–727.

Kourbeti

, Vakis

, Papadakis

, et al. Infections in traumatic brain injury patients. Clin Microbiol Infect, 2012; 18:359–364.

NHE Fact Sheet

CMS

. www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/NationalHealthExpendData/NHE-Fact-Sheet (Last Accessed September 30, 2020).

10.

Levine

, Mulligan

. Overutilization, overutilized. J Health Polit Policy Law, 2015; 40:421–437.

11.

Motti

. Health care overutilization in the United States. JAMA, 2008; 300:2251.

12.

Merlani

, Garnerin

, Diby

, et al. Quality improvement report: Linking guideline to regular feedback to increase appropriate requests for clinical tests: Blood gas analysis in intensive care. BMJ, 2001; 323:620–624.

13.

Chami

, Simons

, Sweetman

, et al. Rates of inappropriate laboratory test utilization in Ontario. Clin Biochem, 2017; 50:822–827.

14.

Venkatesh

, Scott

, Ziegenfuss

. Cerebrospinal fluid in critical illness. Crit Care Resusc, 2000; 2:42–54.

15.

Arevalo-Rodriguez

, Muñoz

, Godoy-Casasbuenas

, et al. Needle gauge and tip designs for preventing post-dural puncture headache (PDPH). Cochrane Database Syst Rev, 2017; 4:CD010807.

16.

Dziedzic

, Slowik

, Szczudlik

. Nosocomial infections and immunity: lesson from brain-injured patients. Crit Care, 2004; 8:266–270.

17.

Thompson

, Pinto-Martin

, Bullock

. Neurogenic fever after traumatic brain injury: An epidemiological study. J Neurol Neurosurg Psychiatry, 2003; 74:614–619.

18.

Childers

, Rupright

, Smith

. Post-traumatic hyperthermia in acute brain injury rehabilitation. Brain Inj, 1994; 8:335–343.

19.

Segatore

. Fever after traumatic brain injury. J Neurosci Nurs, 1992; 24:104–109.

20.

Agrawal

, Timothy

, Thapa

. Neurogenic fever. Singapore Med J, 2007; 48:492–494.

21.

Bota

, Lefranc

, Vilallobos

, Brimioulle

, Vincent

J-L

. Ventriculostomy-related infections in critically ill patients: A 6-year experience. J Neurosurg. 2005; 103:468–472.

22.

Lozier

, Sciacca

, Romagnoli

, Connolly

. Ventriculostomy-related infections: A critical review of the literature. Neurosurgery, 2002; 51:170–181.