✓The choice of outcome (or outcomes) and their measurement are critical for a sound clinical trial. Surgeons have traditionally measured simple outcomes such as death, duration of survival, or tumor recurrence but have recently developed more sophisticated measures of the effect of an intervention. Many outcome measures require a lengthy maturation process, which includes a determination of the instrument's validity, reliability, and sensitivity; thus, using established instruments rather than creating new ones is recommended. The authors illustrate several guidelines for the determination of appropriate outcome measures by using examples from their experience and describe several outcome measures that can be used in pediatric neurosurgery. These include general outcome measures such as the Pediatric Evaluation of Disability Inventory and the Functional Independence Measure for Children, which measure physical function and independence in chronically ill and disabled children as well as disease-specific measures for hydrocephalus (Hydrocephalus Outcome Questionnaire), cerebral palsy (gross motor function and performance measures), head injury (Pediatric Cerebral Performance Category and Children's Coma Scale), and oncology (Pediatric Cancer Quality-of-Life Inventory).
Paul Klimo Jr. and John R. W. Kestle
Paul Klimo Jr., Clinton J. Thompson, James Drake and John R. W. Kestle
Object. Most surgical procedures are associated with a learning curve in which the success rate is lower early in the experience before mistakes have been identified and modifications made to the procedure. Negative results obtained early in a trial's learning curve may be a matter of timing rather than a reflection of the procedure's effectiveness. The recently published results of the Endoscopic Shunt Insertion Trial (ESIT) represent the notion that endoscopically placed shunts were no more likely to survive than conventionally placed shunts. This negative result may be due to inexperience in performing endoscopic surgeries.
Methods. Surgical experience was assessed in two ways. Shunt survival rates were compared between cases treated endoscopically in the 1st and last years of the ESIT. The effect of center volume was evaluated using a Cox proportional hazard model in which the following variables were analyzed: age at registration, the diagnosis of myelomeningocele, head size, method of shunt placement (endoscopic compared with conventional), and center volume.
There was no difference in survival (endurance) of the shunt between patients enrolled in the 1st and last years (log rank = 0.08, p = 0.77). Likewise, no variable in the Cox multivariate model, including center volume, was a significant factor in predicting shunt survival.
Conclusions. The primary result of the ESIT was found to be internally valid. The fact that endoscopic shunt placement did not benefit patients evaluated in the study was not due to early timing of the trial. Any learning curve among the participating surgeons did not adversely affect the results.
Richard C. E. Anderson, Peter Kan, Paul Klimo, Douglas L. Brockmeyer, Marion L. Walker and John R. W. Kestle
Object. Intracranial pressure (ICP) monitoring has become routine in the management of patients with traumatic brain injury (TBI). Many surgeons prefer to use external ventricular drains (EVDs) over fiberoptic monitors to measure ICP because of the added benefit of cerebrospinal fluid drainage. The purpose of this study was to examine a consecutive series of children with TBI and compare the incidence of complications after placement of an EVD, a fiberoptic intraparencyhmal monitor, or both.
Methods. A retrospective chart review was conducted to identify children with TBI who met the criteria for insertion of an ICP monitor. All patients underwent head CT scanning on admission and after placement of an ICP monitor.
During a 5-year period 80 children met the criteria for inclusion in the study. Eighteen children (22.5%) underwent EVD placement only, 18 (22.5%) underwent placement of a fiberoptic device only, and 44 (55%) received both. A total of 62 fiberoptic devices (48%) were inserted, and 68 EVDs (52%) were placed. Overall, there was a fourfold increased risk of complications in children who received an EVD compared with those in whom a fiberoptic monitor was placed (p = 0.004). Hemorrhagic complications were detected in 12 (17.6%) of 62 patients who received an EVD compared with four (6.5%) of 62 patients who received a fiberoptic monitor (p = 0.025). Six (8.8%) of 68 EVDs were malpositioned and required replacement; in three (50%) of these cases a hemorrhagic complication occurred. Only one infection was noted in a patient with an EVD (1.5%).
Conclusions. In this retrospective cohort of pediatric patients with TBI, complication rates were significantly higher in those receiving EVDs than in those in whom fiberoptic monitors were placed. Although the majority of these complications did not entail clinical sequelae, surgeons should be aware of the different complication rates when choosing the most appropriate device for each patient.
Paul Klimo Jr., John R. W. Kestle, Joel D. MacDonald and Richard H. Schmidt
Cerebral vasospasm after subarachnoid hemorrhage (SAH) continues to be a major source of morbidity in patients despite significant clinical and basic science research. Efforts to prevent vasospasm by removing spasmogens from the subarachnoid space have produced mixed results. The authors hypothesize that lumbar cisternal drainage can remove blood from the basal subarachnoid spaces more effectively than an external ventricular drain (EVD). This nonrandomized, controlled-cohort study was undertaken to evaluate the effectiveness of a lumbar drain in patients with SAH compared with those in whom an EVD or no form of cerebrospinal fluid (CSF) drainage was used to prevent the development of clinical vasospasm and its sequelae.
The authors collected data on 266 patients with nontraumatic SAH who were admitted to the University of Utah Health Sciences Center between January 1994 and January 2003. Of these, 167 met the study entry criteria. The treatment group consisted of 81 patients in whom a lumbar drain had been placed for CSF shunting, whereas the control group was composed of 86 patients who received no form of CSF drainage or who were treated solely with an EVD. Primary outcome measures were as follows: 1) clinically evident vasospasm; 2) the need for endovascular intervention; 3) vasospasm-induced infarction; 4) disposition at time of discharge; and 5) Glasgow Outcome Scale (GOS) score at 1 to 3 months postdischarge. Secondary outcomes included length of stay and the need for CSF shunting.
The presence of a lumbar drain conferred a statistically significant protective and beneficial effect across all outcome measures, reducing the incidence of clinical vasospasm from 51 to 17%, the need for angioplasty from 45 to 17%, and the occurrence of vasospastic infarction from 27 to 7% (all p ≤ 0.001–0.008). Patients in the treatment group were more likely to be discharged home (54% compared with 25%, p = 0.002) and to have a GOS score of 5 at follow up (71% compared with 35%, p < 0.001). The mean number of days spent in the intensive care unit and in the hospital overall was also fewer in the treatment group. A similar degree of benefit was found in patients with different Fisher grades and regardless of whether an EVD was needed on presentation, both by subgroup analysis and multivariate logistic regression modeling. There was no statistical difference between the groups in terms of patients requiring a shunt. Complications with lumbar drains were rare and yielded no permanent sequelae.
Shunting of CSF through a lumbar drain after an SAH markedly reduces the risk of clinically evident vasospasm and its sequelae, shortens hospital stay, and improves outcome. Its beneficial effects are probably mediated through the removal of spasmogens that exist in the CSF. The results of this study warrant a randomized clinical trial, which is currently under way.
Warren R. Selman
Paul Klimo Jr., John R. W. Kestle and Meic H. Schmidt
Spinal cord compression is one of the most dreaded complications of metastatic cancer. It can lead to a number of sequelae, including pain, spinal instability, neurological deficits, and a reduction in the patient's quality of life. Except in selected circumstances, treatment is palliative. Treatment options include surgery, radiation, and chemotherapy. The goal of this study was to summarize the existing data on the outcomes of various treatment methods for metastatic spinal epidural disease and to make appropriate recommendations for their use.
The authors used a search strategy that included an electronic database search, a manual search of journals, analysis of bibliographies in relevant review papers, and consultation with the senior author. There is good evidence, including Class I data, that steroid drugs constitute a beneficial adjunctive therapy in patients with myelopathy from epidural compression. Historically, conventional radiation therapy has been viewed as the first-line treatment because it has been shown to be as effective as a decompressive laminectomy, with a lower incidence of complications (Class II data). Nevertheless, in the last 20 years there has been remarkable progress in surgical techniques and technology. Currently, the goals of surgery are to achieve a circumferential decompression of the spinal cord, and to reconstruct and immediately stabilize the spinal column. Results in a large body of literature support the belief that surgery is better at retaining or regaining neurological function than radiation and that surgery is highly effective in relieving pain. Most of the data on the treatment of metastatic spinal disease are Class II or III, but the preliminary results of a well-designed, randomized controlled trial in which surgery is compared with standard radiation therapy represents the first Class I data.
As the number of treatment options for metastatic spinal disease has grown, it has become clear that effective implementation of these treatments can only be achieved by a multidisciplinary approach.
Case report and classification scheme
Paul Klimo Jr., Ganesh Rao, Richard H. Schmidt and Meic H. Schmidt
Nerve sheath tumors that involve the sacrum are rare. Delayed presentation is common because of their slow-growing nature, the permissive surrounding anatomical environment, and nonspecific symptoms. Consequently, these tumors are usually of considerable size at the time of diagnosis.
The authors discuss a case of a sacral nerve sheath tumor. They also propose a classification scheme for these tumors based on their location with respect to the sacrum into three types (Types I–III). Type I tumors are confined to the sacrum; Type II originate within the sacrum but then locally metastasize through the anterior and posterior sacral walls into the presacral and subcutaneous spaces, respectively; and Type III are located primarily in the presacral/retroperitoneal area. The overwhelming majority of sacral nerve sheath tumors are schwannomas. Neurofibromas and malignant nerve sheath tumors are exceedingly rare. Regardless of their histological features, the goal of treatment is complete excision. Adjuvant radiotherapy may be used in patients in whom resection was subtotal. Approaches to the sacrum can generally be classified as anterior or posterior. Type I tumors may be resected via a posterior approach alone, Type III may require an anterior approach, and Type II tumors usually require combined anterior–posterior surgery.