| Abstract |
Nosocomial infections of the Central Nervous System (CNS) are a relatively small but important category of hospital-acquired infections. These infections span a spectrum from superficial wound infections, to ventricular shunt infections, meningitis and deep-seated abscesses of the brain parenchyma. These infections are usually serious, if not life threatening and can be associated with a poor outcome. Infections in neurosurgical patients have not been extensively described in Greece. In this extensive study in the University of Crete Medical Center, we attempted to define the risk factors associated with post-craniotomy meningitis (PCM) in Crete and compare them with the risk factors that were associated with post-craniotomy meningitis in a cohort in New York University Medical Center. Since traumatic brain injury was the main reason for admission in the UOC Medical Center-Department of Neurosurgery, we specifically analyzed the infections in this population. The author tried to confirm the results from the retrospective studies with a prospective study on risk factors associated with PCM. The studies performed are summarized as follows:
Study 1: Overview of the neurosurgical infections in the University Hospital of Crete based on a 3-year retrospective study
The medical records of the patients >18 years old that were admitted to the Department of Neurosurgery between 2004 and 2006 were reviewed. A total of 1,112 events were analyzed. Trauma was the most common cause for admission (56.3%). Craniotomy was the most common procedure performed (21.8%) but one-third of the patients admitted during that time period did not undergo any major procedures.
The prevalence of Surgical Site Infections (SSIs) during the aforementioned time period was 12.5% with superficial wound infections being the most common. VAP (Ventilator-Associated Pneumonia) was the most common non-SSI infection encountered in this cohort. The rate of SSIs was higher in patients who also developed VAP, urinary tract infections (UTI) and blood stream infections/catheter associated bacteremia (BSI/CAB).
In multivariate analysis, malignancy, surgery for a vascular reason, shunt replacement surgery, placement of any drain, and surgery through a sinus were all independent predictors for SSIs development. The development of any infection was independently associated with a history of malignancy, performance of surgical
13
procedures for trauma or cerebrovascular events, shunt replacement surgery and the placement of any drains or an ICP monitor device.
Infections, including SSIs, were associated with a prolonged hospitalization, both in the ICU and the wards, but not with an increased mortality. The most common pathogens isolated in SSIs were the gram-positive. Acinetobacter spp. were the most common isolates in VAP (58.3%), a fact with a particular significance, due to the increased resistance of the pathogens.
Therefore with this retrospective study, we demonstrated that the rate of SSIs was considerably high. Malignancy was for the first time demonstrated as a risk factor for infection development and SSIs in particular. Development of both SSIs and other infections were associated with a prolonged length of stay and consequently increased hospital costs but not an increased mortality.
Study 2: Retrospective study regarding the risk factors associated with post-craniotomy meningitis in New York University Medical Center- Department of Neurosurgery
This retrospective study was performed in New York University (NYU) Medical Center. The purpose was to determine the rate, bacteriology and risk factors for post-craniotomy meningitis (PCM). The study included patients >18 years old that underwent non-stereotactic craniotomies between January 1996 and March 2000. Operations for burr holes, ICP and shunt placements were excluded. Host factors, craniotomy type and pre- and postoperative variables were evaluated as risk factors for meningitis.
Four hundred and fifty three patients were included. Among them, there were 25 cases of meningitis. Most cases analyzed were operated for an oncological reason (45%), therefore they were elective cases. Ninety-two percent of the patients received antibiotic prophylaxis, mainly a first generation cephalosporin (78%). There was a doubling in the risk in the patients who did not receive prophylactic antibiotics but this was not statistically significant (p= 0.148)
The rate of meningitis was 5.5%, higher than noted in some studies, but not all. Eight out of 12 culture-positive cases revealed gram-positive cocci [mainly coagulase-negative staphylococci (CoNS)]. These results probably reflect the emergence of gram-positive as nosocomial pathogens during that time-period.
In multivariate analysis the risk for meningitis was increased by surgery that entered a sinus (OR 4.49), an increased ASA score (OR 1.72) and increase in the
14
number of days of external ventricular drainage (EVD) (OR 1.21) and intracranial pressure monitoring (ICP) (OR 1.24).
On conclusion, access of upper airway bacteria to the surgical wound, host factors as expressed by the ASA score and duration of device-related postoperative communication of the cerebrospinal fluid (CSF) and the environment are major risk factors for PCM development.
Study 3: Retrospective study regarding the risk factors associated with post-craniotomy meningitis in University of Crete Medical Center
This retrospective study was performed in the University of Crete Medical Center between January 1999 and December 2005. The purpose was to determine the rate, bacteriology and risk factors for post-craniotomy meningitis (PCM) for the first time in a cohort in Greece. The study included patients >18 years old that underwent non-stereotactic craniotomy. Host factors and pre- and postoperative variables were evaluated as risk factors for meningitis.
Six hundred nineteen craniotomies performed in 479 patients were analyzed. Traumatic brain injury (TBI) was the most common cause for craniotomy. 26% of the patients developed at least one infection. VAP was the most common infection recorded (13.1%). Meningitis/ventriculitis was encountered in 37 procedures (6.1%). Seventy-seven percent of the LP samples were positive. Gram-negative pathogens represented 48% of the culture-documented cases and gram-positive represented 43%.
In the multivariate analysis the risk for meningitis was independently associated with the development of another SSI (odds ratio [OR] 4.5), VAP/pneumonia (OR 4.4), UTI (OR 6.2), malignancy (OR 3.6), presence of a ventricular drainage (OR 12.7), presence of a lumbar drainage (OR 91.8) and an emergent procedure (OR 2.9).
Device-related postoperative communication of the CSF and the environment, SSI other than meningitis and infections outside the surgical field were defined as major risk factors for PCM.
On comparison, this retrospective cohort differed from the NYU cohort, in the percentage of emergent procedures which represented 51.4% in the UOC cohort whereas they represented only 34.2% in the NYU population. This mainly reflects the difference in the TBI population included in each cohort. The difference in the meningitis cases between the two cohorts (6.1% vs 5.5%) was not statistically significant. There was a slight preponderance of gram-negative pathogens in the UOC cohort, probably representing the fact that UOC cohort included patients until the year
15
2005. The trend in the preponderance of gram-negative pathogens (especially Acinetobacter spp.) as PCM pathogens seem to be universal. A major drawback of the UOC cohort was the lack of ASA evaluation as a risk factor. For the first time malignancy was an independent risk factor for PCM in the UOC cohort. The presence of a lumbar drain carried an independent association with PCM in the UOC cohort but in the NYU cohort there was no association. In the UOC cohort the presence of other infections in the association with PCM development was underscored.
Study 4: Prospective study regarding the risk factors associated with post-craniotomy meningitis in University of Crete Medical Center
In this prospective study we attempted to determine the rate, bacteriology and risk factors for PCM. The need for a prospective cohort was imperative for the confirmation of the risk factors for PCM defined in the retrospective studies of the Thesis. Patients >18 years old that underwent non-stereotactic craniotomies between January 2006 and December 2008 were included. Host factors and pre- and postoperative variables were evaluated as risk factors for meningitis.
Three hundred thirty four craniotomies were analyzed. Men represented 65.6% of them and TBI was the most common cause for craniotomy. 39.8% of the patients developed at least one infection. VAP was the most common infection recorded (22.5%). Meningitis/ventriculitis was encountered in 16 procedures (4.8%). One hundred percent of the LP samples cultured in suspected meningitis cases were positive. Gram-negative pathogens (Acinetobacter spp, Klebsiella spp, Pseudomonas aeruginosa, E.cloacae and Proteus mirabilis) represented 88% of them. In the multivariate analysis the risk for meningitis was independently associated with the perioperative steroid use (OR 11.55), CSF leak (OR 48.03), and postoperative ventricular drainage (OR 70.52).
Device-related postoperative communication of the CSF and the environment, CSF leak and perioperative steroid use were defined as major risk factors for PCM in this prospective study. The great preponderance of gram-negative pathogens (especially Acinetobacter spp), reconfirmed their predominance as nosocomial pathogens after the year 2000. Postoperative ventricular drainage was reconfirmed as a risk factor for PCM as in the two retrospective cohorts. CSF leak was a major risk factor as reported in previous studies. Perioperative steroid use was described for the first time as an independent risk factor for PCM. A major advantage of this study is the description of the microbiology and the sensitivities of the pathogens involved in
16
the PCM and the other infections encountered in patients undergoing craniotomy (especially VAP). This will provide great help in the empirical choice of antibiotics upon the presentation of the infections.
Study 5: Retrospective study regarding the infections in traumatic brain injury (TBI) in the University of Crete Medical Center- Risk factors associated with the development of surgical site infections (SSIs) and meningitis in TBI population
Admission and surgery for TBI was the most common in UOC Medical Center- Department of Neurosurgery. The purpose of this study was to delineate the frequency, types and risk factors for infection in TBI patients. This was a retrospective surveillance for all TBI patients, aged ≥18 years, cared at the Department of Neurosurgery of the University Hospital of Heraklion between 1999 and 2005.
Seven hundred and sixty patients (75% men- median age 41) were included. Two hundred fourteen infections were observed. The majority were infections of the lower respiratory tract (47%), mainly ventilator associated pneumonia (VAP) (33%), followed by surgical site infections (SSI) (17%). Multivariate analysis has shown that SSI development was independently associated with performance of ≥2 surgical procedures, presence of concomitant infections, namely VAP and urinary tract infections, insertion of lumbar and ventricular drains and cerebrospinal fluid (CSF) leak. Meningitis was associated with prolonged hospitalization, and insertion of lumbar and ventricular drains. There was a predominance of Acinetobacter spp as a VAP pathogen, gram positive organisms remained the most prevalent in SSI.
Respiratory tract infections were the most common among TBI patients. Device-related communication of the CSF with the environment and prolonged hospitalization were independently associated with the development of SSIs and meningitis in this particular cohort. The prevalence of the pathogens must be determined upon institutional basis for the establishment of proper treatment of these serious infections
Study 6: Association between operative site microbial counts and procedure classification in neurosurgery: A prospective study-Interim Report
No association between bacterial skin counts at the operative site and SSI has been reported. In this study we investigated the bacterial skin counts at the operative site, the association with procedure classification in neurosurgery and the impact on the development of SSI.
17
This was a prospective study performed in UOC Medical Center. Over a period of 18 months (February 2007- July 2008), three samples from neurosurgical patients were obtained while the patient was in the operating room and cultured. All samples were obtained at about 1 cm from the surgical incision, one pre-preparation sample, one post-preparation and one pre-closure sample. Patients having any type of procedure were included.
Ninety-three sample sets from 83 procedures were analyzed in this interim report. CoNS were the most frequently isolated organisms irrespectively of the time of the sampling and independently of the procedure classification. P. acnes was the second most frequently isolated organism. There was no statistical difference in the sampling positivity according to the sampling site or in the sampling positivity according to the procedure classification. Bacterial colony forming units (CFU) irrespectively of sampling time were not associated with procedure classification or revision surgery. CFU counts in the pre-preparation samples did not correlate with post-preparation or pre-closure samples. Pre-closure sample counts had a trend to increase with increased duration of surgery, especially if this exceeded the 3-hour duration. There was a trend to increased numbers of P. acnes and diphtheroids in the pre-closure samples.SSI development did not carry a significant association with the skin microbial CFU counts at any sampling and for any procedure classification. The positivity of the pre-preparation and the pre-preparation samples did not differ from a previous report.
In this pilot prospective study which is on-going we were unable to detect an association between procedure classification in neurosurgery, CFU counts in three different sampling times and SSI development
|
Collections
