Clinical Chemistry 51: 202-207, 2005.
First published November 18, 2004; 10.1373/clinchem.2004.039719
(Clinical Chemistry. 2005;51:202-207.)
© 2005 American Association for Clinical Chemistry, Inc.
|
General Clinical Chemistry |
Serum S-100B as an Indicator of Early Postoperative Deterioration after Meningioma Surgery
George Stranjalis1,
Stefanos Korfias1,a,
Christina Psachoulia2,
Efstathios Boviatsis1,
Andreas Kouyialis1,
Despina Protopappa3 and
Damianos E. Sakas1
Departments of 1
Neurosurgery, University of Athens, 2
Biochemistry, and 3
Pathology, Evangelismos Hospital, Athens, Greece.
aAddress correspondence to this author at: 80 Agias Varvaras St., Halandri 15231, Greece. Fax 30-210-7249986; e-mail skorfias{at}mail.gr.
 |
Abstract
|
|---|
Background: S-100B protein is an established serum marker of primary and secondary brain damage in head injury and stroke. Despite major progress in neurophysiologic monitoring, there are still difficulties in the early identification and quantification of evolving edema or trauma after craniotomy for tumor. In this study we aimed to correlate serum S-100B values with early postoperative neurologic course as well as late outcome in meningioma surgery.
Methods: We enrolled 50 consecutive patients who underwent meningioma resection. Serum S-100B was measured preoperatively and postcraniotomy for 7 consecutive days. Twenty-five patients (50%) developed immediate postoperative neurologic deterioration, and 15 (30%) had unfavorable 6-month outcomes. We used the Mann–Whitney U-test to assess the association of S-100B with all variables of interest. We used multiple logistic regression to search for the most significant predictor of postoperative deterioration.
Results: Increased S-100B was highly correlated with larger tumors, intraoperative difficulties, postcraniotomy acute deterioration, and long-term poor outcome. In addition, multiple logistic regression showed that age, sex, site, preoperative edema, history of meningioma resection, extent of resection, and histologic type did not correlate with postoperative increases in S-100B. Furthermore, patients with postoperative S-100B values >0.4 µg/L had increased risk of deterioration (relative risk = 9.0; 95% confidence interval, 2.4–34; P <0.0001) and of poor ultimate outcome (relative risk = 11; 95% confidence interval, 1.6–77; P = 0.002).
Conclusions: After meningioma excision, postcraniotomy increases in serum S-100B appear to be an early indicator of short-term postoperative neurologic deterioration and of a poor longer-term outcome.
|
Introduction
|
|---|
Meningiomas, which constitute 15% of primary brain tumors in adults, are extraaxial, slow-growing, and histologically mostly benign. Although total resection and cure seems plausible, patients with meningiomas may still have an unsuccessful outcome because of postoperative exacerbation of a preexisting peritumoral vasogenic edema (PTVE)1
(1)(2)(3)(4)(5) or the development of a completely new edema or postoperative hematoma. The exact pathogenesis of PTVE remains unclear; it may be present in various degrees, and it may cause intraoperative difficulties as well as postoperative deterioration with increased mortality and morbidity. PTVE occurs in 45–92% of patients (2)(3).
On the other hand, a simple blood test indicative of ongoing organic brain damage could be more sensitive than a decrease in Glasgow Coma Score or an increase in intracranial pressure in ventilated patients for early detection of postcraniotomy deterioration of preexisting PTVE or the development of edema or hematoma (as a consequence of surgical trauma). Despite the fact that serum S-100B protein is an established neurochemical marker of organic brain damage in head injury and stroke (6)(7)(8)(9)(10)(11)(12), there have been no studies involving neurochemical monitoring in the early postoperative period after meningioma surgery.
We designed a study to correlate the serum concentrations of S-100B with postcraniotomy neurologic deterioration in patients undergoing meningioma resection.
|
Materials and Methods
|
|---|
We prospectively studied 50 consecutive patients who underwent surgery for meningioma. We recorded age, sex, neurologic status, tumor site and size, extent of PTVE according to Trittmatcher’s criteria (13) (none, perifocal, or hemispheric), previous history of meningioma resection, extent of present resection (total or subtotal), intraoperative technical difficulties (difficult approach because of tumor vascularity and invasiveness with subsequent increased surgical manipulation and trauma), pre- and postoperative 7-day course of serum S-100B values, postoperative neurologic deterioration and computerized tomography (CT) scan, cause of deterioration (hematoma with shift, increase of edema, or ischemic changes), histologic type based on WHO classification (2)(14), and 6-month Glasgow Outcome Scale (15). The study was conducted from January 2002 until June 2003 and after approval and authorization from the Ethics Committee of our hospital.
Patients were operated on by 4 neurosurgeons and were routinely prescribed dexamethasone (12–16 mg/day) and valproic acid (20 mg · kg–1 · day–1) during the pre- and postoperative course. Patients with large meningiomas (
4 cm) were routinely ventilated in the intensive care unit for 24 h postoperatively. All patients underwent serial postoperative CT scans. Patients with an uneventful postoperative course were scanned once. All patients with postoperative neurologic deterioration had at least two CT scans. All scans were interpreted by independent neurologists. Postcraniotomy neurologic deterioration was recorded if there was (a) a definitive worsening of preexisting neurologic signs or (b) signs such as deterioration of level of consciousness, mono- or hemiparesis, aphasia, seizures, or death.
Serum S-100B protein either is present in very low concentrations (<0.12 µg/L) or is undetectable in healthy individuals. There have been no reports of serum S-100B values in patients harboring brain tumors or in the early clinical course after tumor resection. We collected one serum sample several hours preoperatively and for 7 consecutive days; all samples were centrifuged and stored (–70 °C). S-100B measurements were performed with a two-site immunoluminometric assay (LIAISON-Sangtec 100; AB Sangtec Medical). The intraassay imprecision of the assay was 6.4% 0.11 µg/L, 2.8% at 1.6 µg/L, and to 3.6% at 18.4 µg/L (n = 5 measurements at each concentration). The interassay CV was 11% at a mean concentration of 0.11 µg/L, 3.7% at 1.6 µg/L, and 3.2% at 18.4 µg/L (n = 5 measurements at each concentration). The stated detection limit (limit of the blank) for the assay (LIAISON-Sangtec 100) is 0.02 µg/L (mean value of blank + 3 SD).
Summary statistics were based on frequency distribution tables for all categorical variables (16). Postoperative S-100B values are presented as the median and interquartile range because of deviation from a gaussian distribution. The primary hypothesis of a positive association between postoperative S-100B and neurologic deterioration was based on the evaluation of the area under the curve (AUC) for S-100B and the maximum serum concentration of S-100B any time postoperatively. AUC were calculated according to the trapezoid rule (17), covering the period from 0 to 144 h postoperatively. The Mann–Whitney U-test was used to assess the association of S-100B AUC, maximum concentration, and the first postoperative measurement of S-100B with all variables of interest. In addition, the statistical dependence of the first postoperative S-100B measurement (considering the cutoff value of 0.4 µg/L) on neurologic deterioration was estimated by the Fisher exact test. A forward logistic regression model was used with postoperative clinical deterioration as the dependent variable (18). The set of independent variables comprised all of the variables that had been found to be significantly related to postoperative clinical deterioration in the univariate analysis. The critical value for entering a variable in the model was set at 0.05. By clinical deterioration we mean definitive postoperative neurologic worsening of preexisting neurologic signs or the presentation of new signs [deterioration in degree of alertness or awareness, difficulties in speech (aphasia) or motion (hemiparesis), or development of seizures] as described previously in the Materials and Methods. Finally, to demonstrate variability between the false-positive and false-negative rates for different cutoff points, we constructed ROC curves with S-100B values, using postoperative deterioration and long-term outcome as dichotomous variables.
|
Results
|
|---|
The clinical characteristics of the 50 patients are shown in Table 1
. A relatively high number of patients (50%) deteriorated in the early postoperative period. Four patients sustained an intracavitary surgical hematoma with shift, and 21 had ischemia/edema evident in the CT scan. Histologically, 96% of the tumors were benign. The long-term outcome was favorable in 70% of patients; the rest were severely disabled or dead at 6 months. In six patients, the preoperative serum S-100B value was slightly above 0.2 µg/L.
The first postoperative S-100B concentration was higher (Table 2
) in patients with larger (>4 cm) tumors, sites other than convexity, intraoperative technical difficulties, postcraniotomy deterioration, and unfavorable long-term outcome. The AUC of serial S-100B values were higher (Table 2
) in cases with larger tumors, intraoperative technical difficulties, postcraniotomy deterioration, and unfavorable long-term outcome. Age, sex, recurrence, extent of resection, preoperative edema, and histologic type were not statistically significantly correlated with S-100B values.
View this table:
[in this window]
[in a new window]
|
Table 2. Relationships of increased first postoperative serum S-100B and of increased serial S-100B values with clinical variables.
|
|
Postoperative clinical deterioration was more common in patients with larger tumors (>4 cm), technical difficulties during the operation, and S-100B (first postoperative S-100B value) concentrations >0.4 µg/L (Table 3
). Similarly poor long-term outcome was associated with technical difficulties during surgery and with S-100B (first postoperative S-100B value) >0.4 µg/L. Age, sex, site, preoperative edema, recurrence, extent of resection, and histologic type were not statistically significantly correlated with postoperative clinical deterioration or long-term outcome.
View this table:
[in this window]
[in a new window]
|
Table 3. Associations of clinical variables and S-100B with postoperative deterioration and poor long-term outcome.
|
|
Patients with postoperative serum S-100B values >0.4 µg/L had a 9-fold greater risk of neurologic deterioration [relative risk = 9.0; 95% confidence interval (CI), 2.4–34; P <0.0001] and an 11-fold higher risk of poor 6-month outcome (relative risk = 11; 95% CI, 1.6–77; P = 0.002). The best results for the sum of sensitivity and specificity were obtained at the cutoff point of 0.4 µg/L. The ROC curves are shown in Fig. 1
. The areas under the ROC curves (19) were 0.80 (95% CI, 0.68–0.92) for acute postoperative deterioration and 0.77 (95% CI, 0.64–0.90) for 6-month outcome (Fig. 1
). The sensitivity and specificity for postoperative deterioration were 92% (95% CI, 84–100%) and 80% (95% CI, 64–96%) with positive and negative predictive values being 82% (95% CI, 68–96%) and 91% (95% CI, 81–100%), respectively. The corresponding values for 6-month outcome were 93% (95% CI, 85–100%), 60% (95% CI, 44–76%), 50% (95% CI, 40–63%), and 95% (95% CI, 89–100%), respectively.
View larger version (13K):
[in this window]
[in a new window]
|
Figure 1. ROC curves for acute postoperative deterioration and for 6-month outcome.
AUC for acute postoperative deterioration (black line) = 0.80; AUC for 6-month outcome (gray line) = 0.77.
|
|
Finally, multiple logistic regression with postoperative neurologic deterioration as the dependent variable and as potential prognostic factors all variables, such as size, intraoperative technical difficulties (tumor vascularity or invasiveness), and S-100B values, that were significantly associated with postoperative deterioration in the univariate analysis (Table 3
), showed that increased AUC144 h S-100B was the most significant predictor of postoperative deterioration (P = 0.0013).
|
Discussion
|
|---|
summary of findings
The aim of this study was to examine the correlation of serial serum S-100B protein measurements to postcraniotomy clinical deterioration in patients undergoing meningioma resection. We enrolled 50 patients who underwent either total or partial meningioma excision. Twenty-five patients (50%) deteriorated, and 35 (70%) had a favorable 6-month outcome. Analysis of the data showed that the "first postoperative value" of S-100B was highly correlated with postcraniotomy neurologic deterioration attributable to edema/ischemia or hematoma, larger tumors (>4 cm), sites other than convexity, intraoperative difficulties, and unfavorable 6-month outcome. In addition, serial (daily) increased serum S-100B showed values were significantly correlated with larger size (>4 cm), technical difficulties, postoperative deterioration, and 6-month neurologic outcome. Age, sex, location, preoperative edema, history of previous resection, extent of resection, and histologic type have no statistically significant correlation with postoperative clinical deterioration or long-term outcome. Multiple logistic regression analysis with postoperative neurologic deterioration as the dependent variable, and size, intraoperative technical difficulties, and S-100B values as prognostic factors showed that increased S-100B (AUC144 h) was the most significant predictor for postoperative deterioration. Finally, patients with postoperative values >0.4 µg/L had a 9-fold greater risk of postoperative neurologic deterioration and an 11-fold risk of poor outcome.
s-100b as a serum marker of brain trauma
Previous studies have shown that high serum S-100B protein concentrations correlate well with primary damage, secondary deterioration, or bad outcome in severe head injury and subarachnoid hemorrhage (6)(7)(8)(9)(10)(11)(12). In addition, there have been several studies that introduced S-100B as a quantitative measure of brain damage in head injury and stroke (4)(8)(9)(12). More specifically, Raabe et al. (9), in a study conducted with 84 severe head injury patients, concluded that S-100B is a promising serum marker for assessing the extent of primary injury and the time course of secondary damage. Woertgen (12) found that serum S-100B >2 µg/L measured within 6 h posttrauma has positive and negative predictive values of 87% and 77%, respectively, for favorable and unfavorable outcomes. In addition, persisting or increasing values seem to indicate ongoing brain damage, an important criterion that could guide us toward immediate therapeutic measures.
biochemical and cellular properties of s-100b
According to some authors (7)(9), the ideal serum marker of brain injury should have high specificity for the brain, high sensitivity for brain injury, no age/sex variability, and rapid appearance in serum and should ensure a predictable relationship between the serum concentration and the tissue injury. The ß-form of S-100 is highly specific for nervous tissue. It is present in the cytosol of glial and Schwann cells and in adipocytes and chondrocytes, although in very low concentrations in the latter two. S-100B can be measured in arterial and venous serum, is not affected by hemolysis, and remains stable over hours without need for immediate analysis. Serum S-100B is not detectable in neurologically healthy patients undergoing spinal anesthesia (8). Hence, organs other than the brain are unlikely to be the source of increased serum S-100B. The role of protein S-100B is not yet fully understood. It has been suggested that S-100B has a neurotropic as well as a neurotoxic function. Another important feature is the short half-life (198 min) of S-100B, making its measurement crucial in the emergency and intensive care settings (20).
pre- and postoperative edema in meningiomas
Meningioma-associated edema ranged from 45% to 92% across different studies (1)(3)(5). It is the main cause of postoperative morbidity and mortality, not only because of increased intracranial pressure (a combination of mass effect and PTVE), but also because of increased intraoperative difficulties and subsequent surgical trauma to the adjacent ischemic brain. Xenon CT studies have confirmed these areas of ischemia, showing that cerebral blood flow is decreased by 28% in the peritumoral areas (21). Other factors, such as vascularity and invasiveness (disappearance of arachnoid layer), size, and location of the meningioma, might induce further vasogenic edema and contribute to further intraoperative technical difficulties (4). Larger size and increased vascularity obviously cause more extravasation and a larger pressure gradient from tumor to brain, with subsequent neuronal and glial damage, a mechanism that is exacerbated in the early postoperative period because of surgical manipulations. A recent study by de Vries (22) showed increased perioperative concentrations of S-100B in the tumor–brain contact areas, confirming the relationship between S-100B and surgical manipulation or trauma.
de Vries (2) reported a high correlation between edema and tumor location but not with tumor size, arachnoid invasion, histology, or vascularity. Tamiya et al. (5) suggested that edema formation depended on tumor infiltration and pial-cortical blood supply. The same authors found a weak correlation of PTVE with size, location (convexity-middle fossa), and histology (meningoendothelial, anaplastic). Salpietro et al. (4) correlated edema with size and cortical penetration. Unfortunately, the above authors did not examine the contribution of PTVE and its correlation to postoperative morbidity. In our study, we found a statistically significant correlation between size, intraoperative difficulties (combination of location and PTVE), and postoperative deterioration independent of S-100 concentrations. On the other hand, we found no correlation between age, sex, topography, PTVE, recurrence, totality of excision, histology, and postoperative morbidity.
tumor surgery, postoperative deterioration, and future applications of s-100b
Neurologic deficit after meningioma excision is detected by postcraniotomy clinical deterioration, with subsequent confirmation by neuroimaging. In patients who are ventilated and be monitored neurologically, a constant increase in intracranial pressure will detect the postoperative intracranial cause of deterioration (edema or hematoma). Meningiomas are well known for the postoperative deterioration of a preexisting edema, which might lead to severe disability or death. Routine early detection of this clinical event and subsequent therapeutic measures might reverse the injury to the nervous tissue.
S-100B is a well-established serum marker of primary or secondary brain damage in head injury and stroke (6)(7)(8)(9)(10)(12). The routine quantification of evolving structural damage, such as postoperative primary edema, deterioration of a preexisting edema, or development of a hematoma, in the early postcraniotomy period (usually 0–5 days) by serial serum S-100B measurements might be a valuable tool for earlier detection of an impending catastrophic event.
In conclusion, our study showed that, after meningioma resection, pathologically increased serum S-100B concentrations in the early postcraniotomy period correlate significantly with neurologic deterioration, i.e., with evolving brain damage. Consequently, this simple and inexpensive blood test might be a powerful tool for detecting postoperative deterioration earlier than a decrease in Glasgow Coma Score, development of focal signs, or increases in intracranial pressure.
|
Footnotes
|
|---|
1 Nonstandard abbreviations: PTVE, peritumoral vascular edema; CT, computerized tomography; AUC, area under the curve; and CI, confidence interval. 
|
References
|
|---|
- Bitzer M, Nagele T, Geist-Brath B, Klose U, Gronewaller E, Morgalla M, et al. Role of hydrodynamic processes in the pathogenesis of peritumoral edema in meningiomas. J Neurosurg 2000;93:594-604.[ISI][Medline]
[Order article via Infotrieve]
- De Vries J, Wakhloo AK. Cerebral oedema associated with WHO I-II-III meningiomas. Acta Neurochir 1993;125:34-40.
- Mantle R, Lach B, Delgado MR, Baeesa S, Belanger G. Predicting the probability of meningioma recurrence based on the quantity of peritumoral brain edema on CT scan. J Neurosurg 1994;91:375-383.[CrossRef]
- Salpietro F, Alafaci C, Lucerna S, Iacopino DG, Todaro C, Tomasello F. Peritumoral edema in meningiomas. Neurosurgery 1994;35:638-642.[ISI][Medline]
[Order article via Infotrieve]
- Tamiya T, Ono Y, Matsumoto K, Ohmoto T. Peritumoral brain edema in intracranial meningiomas: effects of radiological and histological factors. Neurosurgery 2001;49:1046-1052.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- De Kruijk JR, Leffers P, Menheere PP, Meerhoff S, Rutten J, Twijnstra A. Prediction of post-traumatic complaints after mild traumatic brain injury: early symptoms and biochemical markers. J Neurol Neurosurg Psychiatry 2002;73:727-732.[Abstract/Free Full Text]
- Ingebrigtsen T, Romner B, Kongstad P, Langbakk B. Increased serum concentrations of protein S-100 after minor head injury: a biochemical serum marker with prognostic value?. J Neurol Neurosurg Psychiatry 1995;59:103-104.
- Ingebrigtsen T, Waterloo K, Jacobsen EA, Langbakk B, Romner B. Traumatic brain damage in minor head injury: relation of serum S-100 protein measurements to MRI and neurobehavioral outcome. Neurosurgery 1999;45:468-476.[ISI][Medline]
[Order article via Infotrieve]
- Raabe A, Grolms S, Sorge O, Zimmermann M, Seifert V. Serum S-100B in severe head injury. Neurosurgery 1999;45:477-483.[ISI][Medline]
[Order article via Infotrieve]
- Rothoerl RD, Woertgen C, Holzschuh M, Metz C, Brawanski A. S-100 serum levels after minor and major head injury. J Trauma 1998;45:765-767.[ISI][Medline]
[Order article via Infotrieve]
- Savola O, Hillbom M. Early predictors of post-concussion symptoms in patients with mild head injury. Eur J Neurol 2003;10:175-181.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Woertgen C. Comparison of serial S-100B and NSE serum measurement after severe head injury. J Trauma 1999;6:1126-1130.
- Trittmacher S. Observations on preoperative and postoperative changes in brain tissue surrounding meningiomas. Schmiedek HH eds. Meningiomas and their surgical management 1991:63-69 WB Saunders Philadelphia. .
- Kleihues P, Burger PC, Scheithauer BW. Histological typing of tumors of CNS, 2nd ed 1993:28-31 Springer Verlag Berlin. .
- Jennett B, Bond M. Assessment of outcome after severe brain damage. A practical scale. Lancet 1975;I:480-484.
- Conover WJ. Practical nonparametric statistics, 2nd ed 1980:584pp John Wiley & Sons Inc. New York. .
- Atkinson K. Elementary numerical analysis, 2nd ed 1993:448pp John Wiley & Sons Inc. New York. .
- Albert A, Anderson JA. On the existence of maximum likelihood estimates in logistic regression models. Biometrika 1984;71:1-10.[Abstract/Free Full Text]
- Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36.[Abstract/Free Full Text]
- Jackson RG, Samra GS, Radcliffe J, Clark GH, Price CP. The early fall in levels of S-100 ß in traumatic brain injury. Clin Chem Lab Med 2000;38:1165-1167.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- Tatagiba M, Mirzai S, Samii M. Peritumoral blood flow in intracranial meningiomas. Neurosurgery 1991;28:400-404.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
- De Vries J. Perioperative levels of S-100 protein in serum: marker for surgical manipulation and postoperative complications. Minim Invasive Neurosurg 2003;46:33-36.[CrossRef][ISI][Medline]
[Order article via Infotrieve]
Top
Abstract
Introduction
