C-Reactive Protein in Prediction of Ventriculoperitoneal Shunt-Related Infection in High-Risk Patients

Abstract

Background:

C-reactive protein (CRP) is an inflammatory marker believed to be of value in the early detection of meningitis. We evaluated its potential as a marker for prediction of shunt-related infection in high-risk subjects.

Methods:

We conducted a prospective pilot study in 26 ventriculoperitoneal shunt procedures; 18 of the patients were considered to be at high risk of infection at the time of shunt insertion. All patients were screened for other disease that could cause, an increase in CRP.

Results:

The serum CRP medians were 3.90 mg/L in the whole sample and 5.36 mg/L in the high-risk participants. All four shunt infections occurred in the high-risk group (22.2% of the group), three (75%) of which were in patients with meningitis. The logistic regression model showed that CRP concentrations above the cut-off value of ≥7 mg/L were related to shunt infection (p=0.042). The receiving-operating characteristic curve revealed a cutoff point at ≥10.1 mg/L (sensitivity 0.75, 1 – specificity 0.18). The calculated area under the curve was 0.744. The sensitivity and specificity in the whole sample and high-risk group were not different (75% and 79%–80%, respectively). The positive post-test probability was 40% in the whole sample and 50% in the high-risk group. The negative post-test probability was 5% and 7%, respectively.

Conclusion:

Our data suggest that in a patient at high risk of shunt-related infection, the serum CRP concentration can be a valuable predictor of the risk of infection. Further studies in larger samples would be worthwhile.

Nowadays, placement of permanent internal cerebrospinal fluid (CSF) devices is a common neurosurgical procedure, the most common of which is ventriculoperitoneal shunt insertion. Because the procedure involves insertion of a foreign body, it is associated with a higher infection rate than other clean procedures. Previous studies showed a 5%–15% rate of shunt-related infection [1]. Patients who receive an external ventricular drainage (EVD) device, whether for drainage of CSF in acute hydrocephalus or for intracranial pressure (ICP) monitoring, face a yet higher rate of infection [2].

The protocol management of shunt infection may differ slightly from one institution to another. Although culture results often are negative, with normal CSF examination and chemistry, there nevertheless is a higher rate of repeated shunt infections RSI after re-insertion procedures [3]. We are still in need of a reliable diagnostic test to tell us a good time to do a reinsertion.

Patients who have received an EVD, regardless of the purpose of the insertion, develop infection in 10%–23% of cases [4,5]. Schade et al. [5] found that routine chemical and microbiologic analysis, which includes leukocyte count, protein and glucose concentration assays, and measurement of the CSF:blood glucose ratio and CSF interleukin-6 concentration, cannot detect early EVD-related meningitis accurately.

Shunt-infection patients who already have received adequate therapy also have a higher risk of RSI: As high as 20%. The risk factors do not include the initial organism, patient age, etiology of hydrocephalus, numbers of days of CSF drainage, or the therapy received at the time the first shunt infection was treated [6]. We hypothesized that the plasma C-reactive protein (CRP) concentration can show residual infection or inflammation that will carry a higher risk of shunt-related infection during insertion of the permanent device.

Patients and Methods

Participants

We conducted a cohort pilot study in 37 consecutive cases from December 2008 to September 2009. All participants had been admitted to King Chulalongkorn Memorial Hospital, Bangkok, Thailand, for insertion of a permanent ventricular device. We divided the participants into a “high-risk” and a “standard” group. The high-risk group included (1) patients who had an EVD or lumbar drain (LD) inserted for any purpose; (2) patients who had previous shunt infection; and (3) patients who had meningitis, ventriculitis, or other central nervous system (CNS) infection who had already started treatment with proper antibiotics given for at least 10 days, together with at least three negative culture results. The standard group consisted of the other patients who had to receive a permanent ventricular device.

All participants were screened to exclude other sites of infection or inflammation. If the patients had infection in another organ system, they must already have been treated with antibiotics for at least 72 h, with a good response, defined as: (1) No systemic inflammatory response syndrome (SIRS)[7] according to vital signs and laboratory values (Table 1); (2) disappearance or normalization of the number of white blood cells (WBC) in the urine if the patient had prior urinary tract infection; (3) improvement of respiratory status and secretion if the patient had pneumonia; (4) normal CSF cytostudy and chemistry if the patient had a CNS infection; (5) a culture from the infected organ showing either no growth or growth of a micro-organism with good sensitivity to the antibiotics given. Otherwise, patients were excluded. They had given their informed consent before recruitment.

Table 1.

Criteria for Systemic Inflammatory Response Syndrome (SIRS)

Established in 1992 as part of the American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference. The conference concluded that the manifestations of SIRS include, but are not limited to:

• Body temperature <36°C or >38°C;

• Heart rate >90 beats/min;

• Tachypnea, with >20 breaths/min; or an arterial partial pressure of carbon dioxide <4.3 kPa (32 mm Hg);

• White blood cell count <4,000 cells/mm³ (4×10⁹ cells/L) or >12,000 cells/mm³ (12×10⁹ cells/L); or the presence of >10% immature neutrophils (band forms).

SIRS is diagnosed when two or more of these conditions are present.

An infection was considered to be associated with a CSF shunt if at least one of the following two criteria was fulfilled (modified criteria for nosocomial infections of the U.S. Centers for Disease Control and Prevention [CDC])[8]: (1) Growth of a pathogen in the CSF, on the shunt tip, or in incision sites overlying the implanted shunt material (if the pathogen was interpreted as relevant); or (2) fever (temperature >38°C), headache, neck stiffness, cranial nerve signs, or irritability without another recognized cause(s) together with physician initiation of an appropriate antimicrobial therapy for shunt-associated infection and a laboratory finding of CSF leukocyte count >5×10⁶ cells/L, CSF total protein >0.45 g/L, a CSF:blood glucose >0.5, organisms seen on CSF gram stain, or organisms found in blood culture. The onset of infection was defined by the first positive culture of CSF, incision site swab, or shunt tip specimen; the initiation of an appropriate antimicrobial treatment for shunt-associated infection; or surgery at the site of the shunt (whichever occurred first) [9].

All participants underwent ventriculoperitoneal shunt insertion using non-impregnated antibiotic catheters. Peri-operative antibiotics were given to every participant. All procedures followed the standard protocol for prevention of VP shunt infection, including double gloving with the outer pair being removed before the shunt was handled.

Test methods

Every patient in the study had blood drawn on the morning of operating day or before going into the operating room. Blood samples were collected and sent to the laboratory for separation of the plasma by centrifugation. Samples were stored in the freezer at –70°C until the assays were done. The CRP assays were done using the nephelometry method with a commercial kit. The normal range of serum CRP is <5 mg/L. We used a CRP cut-off point of ≥7.0 mg/L, referenced from a previous study showing it to be a cut-off point for shunt infection [3].

All participants were followed for at least 120 days or until they died after shunt insertion. The dead participants had died no sooner than 30 days after the operation unless they died from shunt-related complications because we studied only the validity of the CRP test. The participants who did not follow up at the clinic were contacted by phone.

Statistical analysis

Data were collected on a Microsoft Excel spreadsheet and analyzed using SPSS for Windows version 17. The CRP data were transformed to logarithms to become a normal distribution. We tested the relation between variables with the Fisher exact test and Student t-test. The binary logistic regression method was used to test for the factors related to shunt infection. The receiver-operating characteristic (ROC) curves were drawn and the area under the curve (AUC) calculated to test the validity. The validity was calculated. We considered p<0.05 to be significant.

Reporting

We used the Standards for Reporting of Diagnostic Accuracy (STARD) criteria checklist [10], as described by the STARD Steering Group, for completeness and accuracy of reports of studies of diagnostic accuracy.

Results

In total, 36 patients with 37 ventriculoperitoneal shunt procedures were enrolled in the study. Five patients were excluded because they had other sites of infection, and another three had SIRS without an identifiable source of infection. There was one patient who developed diabetic ketoacidosis perioperatively, and another excluded patient was an elderly woman who developed pneumonia within two weeks after the operation. She died one week after that; therefore, it had been already proved that she had no shunt infection. Another patient was lost to follow-up after discharge from the hospital. The final trial set was 26 procedures in 25 patients, of whom 18 were considered at high risk of infection.

The baseline characteristics of the entire sample and of the participants at high risk of shunt-related infection are shown in Table 2. The median age was 54.5 years (interquartile range [IQR] 35.5–65.5). The most common type of disease leading to shunt insertion was CNS tumor in both the whole and the high-risk sample (53.8% and 66.7%, respectively), followed by vascular disease (30.8% and 27.8%, respectively). Most of the high-risk group were patients who underwent EVD insertion (38.5%) and patients with meningitis (34.6%).

Table 2.

Baseline Characteristics of the Study Cohort

Characteristic	Whole Sample	High-risk participants
No. of subjects	26	18
Age (year)(IQR)^a	54.5 (35.5–65.5)	49 (35–62)
Sex (%)
Men	10 (38.5)	7 (38.9)
Women	15 (61.5)	11 (61.1)
Diagnosis (%)
Congenital	2 (7.7)	0
NPH	1 (3.8)	0
Tumor	14 (53.8)	12 (66.7)
Vascular	8 (30.8)	5 (27.8)
Infection	1 (3.8)	1 (5.6)

Data presented as medians.

IQR=interquartile range; NPH=normal pressure hydrocephalus.

The data collected are listed in Table 3. The median serum CRP concentration was 3.90 mg/L (IQR 1.60–13.47 mg/L) for the whole sample and 5.36 mg/L (IQR 1.60–15.67 mg/L) in high-risk participants. All four shunt infections occurred in the high-risk group (22.2%), of which three (75%) occurred in meningitis patients. Seven patients died during the study period, six of whom were in the high-risk group.

Table 3.

Data Collected in the Study

Factor	Whole sample	High-risk participants
Median serum CRP (mg/L) (IQR)	3.90 (1.60–13.47)	5.36 (1.60–15.67)
EVD (%)	–	10 (38.5%)
EVD (days) (IQR)		6.5 (4.75–8.75)
Meningitis (%)	–	9 (34.6%)
CRP ≥7 mg/L (%)	7 (26.9)	6 (33.3)
Shunt infection (%)	4 (15.4)	4^a (22.2)
Died (%)	6 (23.1)	5 (27.8)
Followup (days (IQR)	151 (109–238)	151 (75–503)

Three in meningitis patients, one in patient with EVD.

CRP=C-reactive protein; EVD=external ventricular device; IQR=interquartile range.

The Fisher exact test showed that a CRP concentration above the designated cut-off value (≥7 mg/L) was related to shunt infection (p=0.047). The binary logistic regression model also showed that a CRP concentration above this cut-off value was related to shunt infection (p=0.042). The ROC curve analysis showed a cut-off point at≥10.1 mg/L (sensitivity 0.75, 1 – specificity 0.18). We also repeated tests for the cut-off point at ≥10.1 mg/L, and the results were the same. The calculated AUC was 0.744. None of the other factors was significantly related to shunt infection; also, no factors in this study showed significant association with death.

There was only one in four shunt infections that gave positive culture results. This culture yielded coagulase-negative Staphylococcus, which was the same pathogen found in the patient's previous meningitis.

The tests of validity are shown in Table 4. The sensitivity and specificity were equal in the whole sample and the high-risk group (75% sensitivity and 79–80% specificity, respectively). There was a slightly higher false-positive rate (21% and 18%), positive predictive value (50% and 43%), and positive post-test probability (50% and 40%) in the high-risk group than in the whole sample, respectively. The negative predictive value was lower in the high-risk group (91.7% and 94.7%, respectively). The negative post-test probability was slightly lower in the high-risk group than in the whole sample (7% and 5%, respectively).

Table 4.

Test of Validity

Characteristic	Whole sample	High-risk participants
Sensitivity (%)	75	75
Specificity (%)	80	79
False-negative rate (%)	25	25
False-positive rate (%)	18	21
Positive predictive value (%)	43	50
Negative predictive value (%)	95	92
Positive likelihood ratio	3.75	3.57
Negative likelihood ratio	0.31	0.32
Positive post-test probability (%)	40	50
Negative post-test probability (%)	5	7
Prevalence (%)	15	22

Discussion

Produced by the liver and adipocytes, CRP is an acute-phase plasma protein first described by Tillett and Francis in 1930. The CRP concentration will be elevated while the body develops the inflammation process. The protein can be detected 6 h after infection begins, and the concentration decreases when the infection is controlled and inflammation subsides. Interleukin-6, which is produced by macrophages in response to inflammation, plays a large role in activating CRP production. The CRP will attach to phosphorylcholine on microorganisms to help the complement system destroy the bacteria. Altogether, it enhances macrophage phagocytosis. C-reactive protein also benefits innate immunity by acting on the early inflectional defense mechanism.

We found that serum CRP values are sufficient for the prediction of shunt-related infection, especially in high-risk patients. The post-test probability increased to 50% in high-risk patients, compared with the pre-test prevalence, which was only 15%–22% in this study, a figure equal to that found in a previous study. This means the risk of infection will be increased to 2.2–3 times normal if the CRP concentration is rising. In contrast, the negative post-test probability is 5%–7%, which, although not a zero risk, gives us enough confidence to do a reinsertion procedure. Although the mortality rate did not relate to shunt infection (six patients die), it surely caused morbidity to the patients, together with a prolonged hospital stay and a rising cost of treatment. There was no cost–benefit analysis in this study, and further analysis is required.

FIG. 1.

Relation between presence of risk, plasma C-reactive protein (mg/L), and shunt infection. CRP=C-reactive protein.

In subgroup analysis, most shunt infections (75%) occurred in patients with meningitis, although they were already receiving proper treatment. There was one patient who had normal CSF chemistry and cytologic examination but a high-serum CRP concentration in whom we did a ventriculoperitoneal shunt procedure and recovered a coagulase-negative Staphylococcus in the intra-operative CSF culture. We had to externalize the shunt and treat the patient with antibiotics. We replaced it when the CSF was clear. This may be another potential use of CRP in the case of shunt malfunction: To decide whether to insert a new shunt.

The limitations of our study merit careful consideration. First, this was a pilot study. Hence, the statistical power is low. The findings may not reflect the true conditions; however, because of the benefit we saw, it is clear that further study of this issue is worthwhile. Also, we could define the best cut-off value in a larger population, something that was not possible in this small series. Second, our cohort study enrolled only adult patients, and the results may not apply to the younger age group. As we know, younger patients are more susceptible to shunt infection, so further study is warranted to answer the question whether the serum CRP value can be used as a marker. Third, we do not know whether the conclusion can be applied to procedures not involving ventriculoperitoneal shunts. Although the rates of infection among various shunts procedures were not different (e.g., ventriculopleural, ventriculoatrial, or lumboperitoneal), it would be wiser to adopt the use of CRP to predict shunt infection only after its value is confirmed by further studies involving more examples of various types of shunts.

Finally, in a trial such as this, every patient must be proved not to have any recent infection, SIRS, or other causes that would contribute to an increase in the CRP, because that would lead to a false conclusion. However, elevation of CRP means there is an inflammation somewhere, which would lead us to explore, or postpone the operation, and it definitely makes the operation safer.

Conclusion

This was a prospective evaluation of plasma CRP concentration; it has been proved to be a valuable marker in predicting shunt-related infection in high-risk patients. Further studies in a larger population would be worthwhile.

Footnotes

Author Disclosure Statement

The authors report no conflicts of interest concerning the materials or methods used in this study or the findings specified in this paper.

References

Brown

, Edwards

, Pople

. Conservative management of patients with cerebrospinal fluid shunt infections. Neurosurgery, 2006; 58:657–665.

Hoefnagel

, Dammers

, Ter Laak-Poort

, Avezaat

. Risk factors for infections related to external ventricular drainage. Acta Neurochir, 2008; 150:209–214.

Schuhmann

, Ostrowski

, Draper

et al. The value of C-reactive protein in the management of shunt infections. J Neurosurg, 2005; 103:223–230.

Korinek

. Risk factors for neurosurgical site infections after craniotomy: A prospective multicenter study of 2944 patients. The French Study Group of Neurosurgical Infections, the SEHP, and the C-CLIN Paris-Nord. Service Epidemiologie Hygiene et Prevention. Neurosurgery, 1997; 41:1073–1079.

Schade

, Schinkel

, Roelandse

et al. Lack of value of routine analysis of cerebrospinal fluid for prediction and diagnosis of external drainage-related bacterial meningitis. J Neurosurg, 2006; 104:101–108.

Kulkarni

, Rabin

, Lamberti-Pasculli

, Drake

. Repeat cerebrospinal fluid shunt infection in children. Pediatr Neurosurg, 2001; 35:66–71.

American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med, 1992; 20:864–874.

Garner

, Jarvis

, Emori

et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control, 1988; 16:128–140.

Conen

, Walti

, Merlo

et al. Characteristics and treatment outcome of cerebrospinal fluid shunt-associated infections in adults: A retrospective analysis over an 11-year period. Clin Infect Dis, 2008; 47:73–82.

10.

Bossuyt

, Reitsma

, Bruns

et al. Towards complete and accurate reporting of studies of diagnostic accuracy: The STARD initiative. BMJ, 2003; 326:41–44.