Shunt infections in children: presentation and management.
Infection has been a significant source of morbidity and mortality in the surgical management of hydrocephalus since the earliest attempts to divert cerebrospinal fluid (CSF). Early surgical procedures for this problem were first reported in the 1890s using material such as glass wool, gold tubes, silver wires and linen thread. Subsequent attempts involved the use of autologous or homologous donor veins to drain CSF into the superior sagittal sinus or the superficial temporal vein.[2,9,10] Unfortunately, immediate postoperative complications such as infection resulted in significant mortality rates for most early procedures. The development of modern fabricated plastic substances and antireflux valves in the 1950s coupled with the advent of systemic antibiotics brought about the modern era of surgical management of hydrocephalus. The initial success of ventriculo-atrial shunting was complicated by blood-borne infections which spread to the kidneys and presented significant problems to renal function (shunt nephritis) as well as the need for frequent surgical procedures to lengthen the tubing.[1,9] These problems led to the current surgical preference for the peritoneum as a site for the distal tubing.[12,14]
Establishing the diagnosis of shunt infections was also a problem. The identification of bacteria which were usually thought of as merely colonizing organisms of low virulence such as Staphylococcus epidermidis further complicated the early management of shunt infection. The recognition of normal skin bacteria as the true pathogens in most infections suggests that shunt tubing may be easily contaminated through either the patient's skin flora or during the operative procedure. The propensity for causative organisms to adhere to shunt tubing has led many authors to recommend the removal of all infected hardware in patients who are able to tolerate such procedures. The bacteria is able to adhere to the tubing through the secretion of a mucopolysaccharide, or slime, known as the glycocalyx (Fig 1). Accumulation of this substance within the lumen of the shunt tubing can result in obstruction of the system. Phagocytosis of bacteria is also inhibited due to this substance.[1,3,7,8] Although bacteria may be difficult to isolate, the initial cell count of CSF from infected shunts may show increased numbers of neutrophils and eosinophils as an early indication of infection. The introduction of infected material into the peritoneal cavity can result in the formation of cysts around the shunt tip or pseudocysts (Fig 2).[1,4,5] Thus the infection of shunt hardware presents a number of complex problems which are not easily treated without surgical intervention.
The age of the patient is also a factor in outcome. Shunt infections in patients less than a year of age are thought to be more virulent and have a higher mortality rate in some series.[3,7,11] These elements underscore the importance of early recognition of shunt infection in children. Presenting symptoms such as fever, signs of increased intracranial pressure, wound changes and abdominal pain are essential elements in nursing assessment.[13,14] The neuroscience nurse is in a unique position to assist in detecting this problem through daily assessments of the patient and familiarity with presenting symptoms. Problems in management of shunt infection can be averted by understanding the usual care of such patients. The average length of stay of up to 14 days, the need for multiple procedures and concern related to the effect of infection on the nervous system can make the stress of this diagnosis an acute problem for the patient and family.
This was a retrospective review of 68 consecutive cases of shunt infection over an eight year period in 61 children at a university-affiliated pediatric hospital. All shunt infections were studied regardless of length of time the system had been implanted. Data related to shunt management at time of placement and previous infections were obtained from the patient's medical record. Presenting symptoms, laboratory data, clinical management and characteristics of the infecting organism were recorded. A sample data sheet utilized for each patient is contained in Figure 3.
Goals of the study were to:
* identify presenting symptoms
* identify most frequently encountered bacteria
* develop an algorithm for guidance in the usual management of shunt infections in children as practiced in this setting
The statistical analysis was descriptive in nature. Data on continuous variables were evaluated for arithmetic mean, standard deviation, median, mode and range. The discrete variables were evaluated for overall frequency within identified subsets.
All patients with suspected shunt infections were evaluated according to the following parameters:
1. History and physical examination with attention to complaints of fever, abdominal pain, irritability, wound changes, meningismu, or sign of shunt malfunction.
2. Diagnostic imaging studies including brain computed tomography, shunt series, and in the last 36 months of study, ultrasound of the abdomen to rule out pseudocyst.
3. Laboratory studies including complete blood count (CBC) and routine urinalysis and urine culture. Shunt taps were done to obtain CSF for culture, cell count with differential and glucose/protein.
Patient management was divided into three groups based on risk of shunt removal or externalization and presence of an open fontanelle (Fig 4). Group 1 consisted of suspected cases or patients judged to be poor risk for externalization due to self-destructive behaviors, presence of other serious medical/surgical problems, or extreme dependence on shunting devices as demonstrated by previous experience. This was a very small number of patients who nevertheless present a serious problem in treatment options. A course of conservative management with close observation and appropriate antikiotic therapy was pursued. Observation for possible improvement after 3-4 days of intravenous antibiotics was done; patients who continued to show clinical signs at that point were consigned to groups 2 or 3 below. Discussion of potential failure to sterilize the shunt system with antibiotics alone was held with families in this group. Two such cases (3%) occurred in this series of patients.
Group 2 consisted of patients with diagnosis of shunt infection confirmed having open fontanelle. These were usually infants less than eight months of age. Removal of infected tubing with management of hydrocephalus by serial lumbar punctures or ventricular taps was undertaken. Appropriate intravenous antibiotics were given based on sensitivity of organisms cultured from CSF and tubing. Replacement of shunt tubing on the opposite side was carried out after ten days of treatment when CSF cultures were negative and confirmation of need for reshunting was demonstrated by ultrasound or CT imaging.
Group 3 consisted of patients with confirmed diagnosis and closed fontanelle. Shunt tubing was externalized from the abdomen and appropriate intravenous antibiotics were begun based on sensitivity of organisms cultured from CSF, abdominal fluid (obtained at time of externalization) or tubing. Daily samples of CSF from the externalized system were analyzed for cell count, culture, glucose and protein. Decreasing CSF neutrophilia and negative cultures after 5-7 days were criteria for replacement of shunt tubing on opposite side. Patients who did not exhibit improvement in these studies were considered for intraventricular antibiotics, removal of all old tubing with placement of extraventricular drainage (EVD) and careful review for other sources of infection. Proper intervention on these problems was generally sufficient to clear such cases and the shunts were replaced after 46 days of negative cultures. All patients were carefully evaluated regarding the need for shunt replacement.
Patients who required EVD during the last 2 years of data collection had an inline shunt valve of slightly higher pressure than their usual requirements placed into the ventricular catheter during routine ventriculostomy. The collection system consisted of sterile empty macrodrip intravenous tubing and fluid bag (Abbott # 11682 and 7951-19). Care was taken that microdrip IV tubing not be used in this system since the microdrip chamber did not allow free drainage of CSF. The collection bag was kept at the level of abdomen or upper chest depending on the amount of drainage desired. By using the roller clamp on the macrodrip tubing set, the system could be clamped as needed. The ventriculostomy was tunnelled beneath the skin from the skull opening to a site approximately 2-3 centimeters posteriorly and sutured in the usual manner. The ventriculostomy tubing was cut and the appropriate pressure valve was inserted into the system, utilizing 2-0 silk to secure all connections (Fig 5). With careful taping of the ventriculostomy and the additional precaution of securing the drainage bag to the patient's clothing, children having EVD for shunt infection were allowed up with supervision (Fig 6). This technique avoided the need for manual opening of stopcocks to drain CSF by utilizing the shunt valve to drain at a preset pressure. Intraventricular antibiotics were injected through the valve reservoir as needed. Children managed with this technique did not require the intensive nursing management to keep parts of the external drainage system at the level of the foremen of Monro as is necessary in stopcock-regulated systems[13,15,16]
A cost analysis of this technique showed that the valve regulated system was also beneficial since less nursing time was required, the child did not require restraint or bedrest and the problems of overdrainage associated with nonvalved systems were largely avoided. This system has allowed most patients with EVD for shunt infection to be managed outside the intensive care unit, thus generating additional savings as well as avoiding the trauma of ICU stay and prolonged bedrest in young children.
A total of 68 shunt infections occurred with 659 shunt operations during eight years. This produced a rate of 10.3% on all patients without regard to the length of time the device was implanted. When infections in the first six months after surgery were found producing a rate of 8%. Children weighing less than ten kilograms made up 44% of cases (30 patients). The mean age of patients with infections was 5.9 years with nearly 50% less than 2 years of age. A second significant group was children ten years or older who comprised 30% of the study population (Fig 7).
Presenting symptoms in order of decreasing frequency were shunt malfunction (33%), fever (26%), localized wound or shunt tract inflammation (22%) and abdominal pain or pseudocyst (19%). Approximately 60% of patients had two or more symptoms at the time of diagnosis (Fig 8). This cluster of indicators was especially important if they occurred within the first eight months after surgery when 80% of cases were diagnosed. Significant laboratory findings are shown in Figure 9. Elevated count in the CSF on shunt tap was by far the best laboratory indicator of shunt infection (70% of cases). Although the total white blood cell (WBC) count on peripheral blood sample was elevated in approximately 30% of cases, a left shift was not consistently present.
Confirmation of diagnosis was made by growth of bacteria from CSF, shunt tubing or abdominal pseudocyst. These results are shown in Fig 10. Staphylococcus epidermidis was by far the most frequently identified organism (56% of cases), followed by Staphylococcus aureus. Propionobacter acnes was present in 9% of cases, all of whom were adolescents. Other bacteria covered a wide spectrum of organisms including Francisella tularensis cultured from a patient exposed to a rabbit which presumably was infected. This finding was confirmed with serologic testing. One case of Hemophilus influenza was found in a child with chronic draining otitis media. The time from surgery to presentation in cases not associated with Staphylococcus or Proprionobacter species averaged 29 months, probably reflecting a hematogenous spread of these organisms. Despite the fact that patients in this study were frequently premature infants or children with multiple problems requiring prolonged hospital stays, few highly resistant organisms were encountered.
Abdominal pseudocysts were found by ultrasonography in seventeen patients (25%). These patients were older (mean age = 9.3 years vs 4.6 years for patients without pseudocyst) and presented at longer intervals after surgery (14.1 vs 8.2 months). When pseudocysts were identified efforts were made to aspirate fluid from the cyst at the time of externalization or removal of shunt tubing. This fluid was frequently positive for bacteria and neutrophils even when CSF sampled from shunt tap was not.
A total of 23 children required EVD for an average of 14 days per patient. Secondary infections occurred in three cases (4%). One death (1.4%) occurred in the series of patients in this report; this occurred two weeks after shunt replacement secondary to abdominal obstruction and adhesions. Five patients did not require replacement of their shunts despite a lengthy period of follow-up after clearing of their infections.
The review was able to identify a group of symptoms associated with shunt infection in children, especially if the patient presentation occurred within six months of surgical manipulation of the system. Monitoring for fever, wound changes, abdominal discomfort and signs of increased intracranial pressure is central to neuroscience nursing assessment. This allows the nurse to play a crucial role in the early detection of shunt infection as well as in family teaching regarding the signs of infection.
Bacteria detected in this series were not unlike those reported in other series. The predominance of the Staphylococcus species coupled with the increased risk of infection in infancy are also similar to other reports. Probable contamination of shunt hardware with skin flora seemed to be a likely cause of colonization. While the total shunt infection rate is not as low as some series which have reported 2-5% rates, the time period for monitoring patients for infection was open ended in this series. Most studies use the CDC criteria of one year or less for post operative infections with implantable devices. In this study, no time limit was placed on identification of infection in shunt devices. It was felt that the total spectrum of shunt infection should be investigated since the neuroscience nurse may care for a patient with hydrocephalus several years after shunt placement. Infection in that instance, although less frequently encountered, is just as important to detect as in the recent surgical case. Detection of Propionobacter acnes as the causative organism in several adolescents has led to use of penicillin as preoperative prophylaxis in patients with recognized acne infection.
The use of a shunt valve in EVD for patients in this series was found to be beneficial in allowing greater activity for patients and simplifying requirements for nursing care activities. Most patients had valves of one level higher in pressure for their EVD than for their usual shunt system since the fluid was draining to an external collection system and not to the peritoneal cavity which placed some inherent limits on the amount of CSF drained. While extra care was required to ensure adequate stabilization of the system, only one case of displacement of EVD due to patient activity occurred. Many patients with EVD are now being treated in intensive care units and are managed with bedrest while attempting to keep the child's head at a predetermined level.[13,15,16] The system described in our study affords greater safety and less restraint for such patients. We also found that with greater mobility allowed many children were less inclined to try to pull out their drainage system. In general, patients were allowed up in chairs or ambulating short distances with supervision. Although no review of complications related to prolonged bedrest was included in this study, this would certainly be a topic for future investigation.
This study also demonstrates the importance of the working relationship with infection control nurse practitioners. The expertise in bacterial types and possible sources of infection as well as familiarity with various antibiotics makes the contribution of this nursing role invaluable to the neuroscience nurse. One example of such collaboration in our study was the placement of incubating vials in the operating room for culture of shunt tubing removed during surgery. By avoiding additional handling of tubing removed during shunt operations, the rate of contamination of specimens has decreased. Under previous arrangements, the tubing was transferred from the patient to a sterile container in the operating room, then to the incubating medium in the microbiology laboratory. Elimination of this additional transfer of shunt tubing for incubation was done as a result of collaboration with the infection control nurse practitioner. Any bacteria cultured from shunt tubing is carefully incubated and the sensitivities of the organism recorded, even if the patient shows no signs of shunt infection. This information is available for immediate use if the patient subsequently develops an infection. Systematic review of such results with the infection control nurse practitioner can assist the neuroscience nurse in identifying trends in patterns of infection.
Although the management of shunt infection remains a source of considerable controversy, it is essential that the neuroscience nurse become familiar with the clinical presentation and management of this problem. Awareness of the frequency of infection and the types of bacteria identified assist the nurse in suggesting possible avenues of investigation to decrease the rate of infection. Collaboration with infection control nursing colleagues offers the opportunity for both groups to increase their knowledge and improve patient outcome from the problem. Implications for future practice include emphasis on symptoms of infection in discharge teaching for patients with shunts. The possible development of methods for external ventricular drainage during shunt infection which allow greater patient mobility should also be explored.
A limitation in this study was the exclusion of adult patients since the research was based in a pediatric hospital. It would be interesting to compare these results to a similar review of adults with shunt infections. Additional topics of investigation could include a review of complications associated with the management of shunt infections and the bedrest usually required in traditional approaches to EVD.
We would like to thank Dr. Thomas Pittman for his support and encouragement in the development of this manuscript. We would also like to thank Ms. Pat Thon, RN, MSN, for her enthusiasm and assistance in formulation of the study as well as chart review.
[Figures 1 to 10 ILLUSTRATION OMITTED]
[1.] Bayston R: Hydrocephalus Shunt Infections. Chapman and Hall, 1989. [2.] Carey C: Hydrocephalus: Etiology, pathologic effects, diagnosis, and natural history. Pages 185-201 in: Pediatric Neurosurgery, 3rd ed, Cheek W (editor). WB Saunders, 1994. [3.] Fan-Havard P: Treatment and prevention of infections of cerebrospinal fluid shunts. Clin Pharmacy 1987; 6:866-880. [4.] Gaskil S: Pseudocysts of the abdomen associated with ventriculoperitoneal shunts: A report of twelve cases and a review of the literature. Pediatr Neurosci 1989; 15:23-27. [5.] Hahn Y: Abdominal CSF pseudocyst: Clinical features and surgical management. Pediatr Neurosci 1986; 12:75-79. [6.] Hirsch B: Instillation of vancomycin into a cerebrospinal Quid reservoir to clear infection: Pharmacokinetic considerations. J Inf Dis 1991;163:197-200. [7.] Klein D: Shunt infections. Page 87-97 in: Concepts in Neurosurgery, Vol 3, Scott R (editor). Williams and Wilkins, 1990. [8.] Marlin A: Cerebrospinal fluid shunts: Complications and results. Pages 221-233 in: Pediatric Neurosurgery, 3rd ed, Cheek WR (editor). WB Saunders, 1994. [9.] McCullough D: History of the treatment of hydrocephalus. Pages 1-10 in: Concepts in Neurosurgery, Vol 3, Scott R (editor). Williams and Wilkins, 1990. [10.] Pudenz R: The surgical treatment of hydrocephalus-an historical review. Surg Neurol 1981, 15(1):15-26. [11.] Quigley M: Cerebrospinal Quid shunt infections: Report of 41 cases and a critical review of the literature. Pediatr Neurosci 1989; 15:111-120. [12.] Schiff S: Delayed Cerebrospinal fluid shunt infection in children. Pediat Neurosci 15:131-135. [13.] Scheinblum S: Treatment of children with shunt infections: Extraventricular drainage system care. Pediatr Nurs 1990; 16(2): 139-143. [14.] Shiminski-Maher T: Current trends in the diagnosis and management of hydrocephalus in children. J Pediatr Nurs 1994; 9(2):74-82. [15.] Terry D: Nursing care of the child with external ventricular drainage. J Neurosci Nurs 1991; 23(6): 347-355. [16.] Tilem D: Nursing care of the child with a ventriculostomy. Pediatr Nurs 1988; 3(3):188-193. [17.] Tung H: Ventricular cerebrospinal Quid eosinophilia in children with ventriculoperitoneal shunts. J Neurosurg 1991;75: 541-544. [18.] Venes J: Infections of CSF shunt and intracranial pressure monitoring devices. Inf Dis Clinics N Amer 1989; 3(2): 289-299.
Questions or comments about this article may be directed to: Dianne G. Williams, RN, MSN, Cardinal Glennon Children's Hospital, 1465 South Grand Boulevard, St. Louis, Missouri 63104. She is a pediatric neurosurgical clinical specialist. Joyce Hayes, RN, MPH, CIC, is an infection control practitioner at Cardinal Glennon Children's Hospital. Susan McCool, RN, MSN, CPNP, is the director of nursing practice/development at Cardinal Glennon Hospital.
|Printer friendly Cite/link Email Feedback|
|Author:||Williams, Dianne G.; Hayes, Joyce; McCool, Susan|
|Publication:||Journal of Neuroscience Nursing|
|Date:||Jun 1, 1996|
|Previous Article:||Replacement therapy: arginine vasopressin (AVP), growth hormone (GH), cortisol, thyroxine, testosterone and estrogen.|
|Next Article:||Learned helplessness in the multiple sclerosis population.|