Outcomes of cranial repair after craniectomy

Transcription

1 J Neurosurg 112: , 2010 Outcomes of cranial repair after craniectomy Clinical article Vi c t o r Ch a n g, M.D., 1 Pa u l Ha rt z f e l d, M.D., 1 Mar i a n n e La n g l o i s, P.A.-C, 2 Asi m Ma h m o o d, M.D., 1 a n d Do n a l d Se y f r i e d, M.D. 1 1 Department of Neurosurgery, Henry Ford Health System, Detroit, Michigan; and 2 Division of Neurosurgery, Baystate Health System, Springfield, Massachusetts Object. Hemicraniectomy is a commonly practiced neurosurgical intervention with a wide range of indications and clinical data supporting its use. The extensive use of this procedure directly results in more cranioplasties to repair skull defects. The complication rate for cranial repair after craniectomy seems to be higher than that of the typical elective craniotomy. This finding prompted the authors to review their experience with patients undergoing cranial repair. Methods. The authors performed a retrospective review of 212 patients who underwent cranial repair over a 13-year period at their institution. A database tracking age, presenting diagnosis, side of surgery, length of time before cranial repair, bone graft material used, presence of a ventricular shunt, presence of a postoperative drain, and complications was created and analyzed. Results. The overall complication rate was 16.4% (35 of 213 patients). Patients 0 39 years of age had the lowest complication rate of 8% (p = 0.028). For patients years of age and older than 60, complication rates were 20 and 26%, respectively. Patients who originally presented with traumatic injuries had a lower rate of complications than those who did not (10 vs 20%; p = 0.049). Conversely, patients who presented with tumors had a higher complication rate than those without (38 vs 15%; p = 0.027). Patients who received autologous bone graft placement had a statistically significant lower risk of postoperative infection (4.6 vs 18.4%; p = 0.002). Patients who underwent cranioplasty with a 0 3 month interval between operations had a complication rate of 9%, 3 6 months 18.8%, and > 6 months 26%. Pairwise comparisons showed that the difference between the 0 3 month interval and the > 6-month interval was significant (p = 0.007). The difference between the 0 3 month interval and the 4 6 month interval showed a trend (p = 0.07). No difference was detected between the 4 6 month interval and > 6-month interval (p = 0.35). Conclusions. The overall rate of complications related to cranioplasty after craniectomy is not negligible, and certain factors may be associated with increased risk. Therefore, when evaluating the need to perform a large decompressive craniectomy, the surgeon should also be aware that the patient is not only subject to the risks of the initial operation, but also the risks of subsequent cranioplasty. (DOI: / JNS09133) Ke y Wo r d s cranioplasty craniectomy cranial repair Cr a n i e c t o m y has been practiced since antiquity.2 In modern neurosurgical practice, large frontotemporoparietal craniectomies are performed for a number of indications. In the setting of traumatic brain injury, craniectomy has been shown to be effective in the management of high intracranial pressure. 1,3 5,10,14,16,19,24 In the stroke literature there is increasing evidence that early hemicraniectomy also plays a role in decreasing the mortality rate, as well as improving overall neurological outcomes in patients with malignant edema after middle cerebral artery infarction. 12,18,25,26 Craniectomies may also provide protective benefits after intracranial aneurysm rupture with intracerebral hemorrhage. 20 Because a growing set of indications and clinical data support its use, the hemicraniectomy will probably remain a relatively common neurosurgical intervention for a variety of pathological processes. With the use of this procedure comes a corresponding number of cranioplasties performed to replace the bone defects created. Although a simple procedure conceptually, the replacement of a bone flap after cranioplasty is not without significant risks. In addition to the concomitant risks inherent in any cranial operation, there are also the risks to which the procedure is predisposed, such as postoperative infections (with foreign body implantation), subdural or epidural fluid accumulations, seizures, and fixed neurological deficits. The observation that complications occur more frequently after cranial repair than after other elective cranial procedures prompted our review of cranioplasty after craniectomy over the past 13 years. We identified 213 patients who underwent bone flap replacement after craniectomy for a variety of indications. Methods We generated a patient database by querying procedures with the current procedural terminology code for 1120 J Neurosurg / Volume 112 / May 2010

2 Cranial repair after craniectomy cranioplasty and replacement of bone flap from January 1995 through May This database was generated through our institution s operative logs, and medical records were reviewed for each included patient to determine the circumstances involving each cranial repair or cranioplasty. Of these patients, only those who had undergone cranioplasty for previous craniectomies were entered into our database. The variables under consideration included sex, age at initial operation, presenting diagnosis, side of operation, age at date of bone flap replacement, time period between removal and replacement of bone, material used to secure the bone plate (sutures vs titanium plates), material used for the bone graft (autologous bone, methylmethacrylate, or titanium mesh), whether a ventricular shunt was in place, whether a postoperative drain was placed (either subgaleal, epidural, or both), length of follow-up, and the occurrence of complications. The full course of follow-up was evaluated in each patient so that complications after discharge from the cranioplasty procedure would also be included. Two hundred and eighteen patients were identified and catalogued with these variables. Operative techniques tended to vary among surgeons. The size of the bone flap removed depended on the type of procedure being performed. In cases of ruptured intracranial aneurysms, a standard pterional craniotomy was typically performed. In traumatic brain injury or stroke a larger frontotemporoparietal craniectomy was performed measuring cm. The bone flap created for tumor surgeries was typically dependent on the size of the lesion to be excised. Bone flaps removed during the initial craniectomy were frozen and stored in our institution s tissue bank. These flaps were later identified and thawed prior to reimplantation for subsequent cranial repair. In addition, the type of postoperative drain used also differed among surgeons, and different types were used depending on the procedure. With regard to ventricular shunts, programmable valves were used more frequently as the technology evolved during the course of this study. Statistical Analysis Statistical analysis was performed using both Wilcoxon 2-sample tests as well as chi-square tests. Wilcoxon 2-sample tests were used to analyze the continuous variables of age and the time interval between the original operation and subsequent cranioplasty. The chi-square test was used to analyze the categorical variables as mentioned above. In addition, complications were further categorized. Any epidural or subdural fluid collections that necessitated a repeated operation were combined into 1 category: postoperative fluid collections. Infections involving either repeated operations or intravenous antibiotic use were combined. Postoperative seizures and other miscellaneous complications were the other 2 categories. A chi-square test was used to compare whether the presence of a drain affected the outcome of postoperative fluid collections. Infections were also analyzed to determine whether there was a correlation with the type of material used for the bone plate. J Neurosurg / Volume 112 / May 2010 Results We identified a total of 218 patients with cranioplasty for previous hemicraniectomy, 4 of whom were lost to follow-up with no information on complications. In addition, 1 patient was listed twice, having undergone 2 cranioplasties. Another patient was removed from analysis because the time between removal and replacement was > 5 years, which we thought would be an outlier. Information from this patient s first operation was recorded and he was counted once subsequently. Of the remaining 212 patients, 34 (16%) had complications. Complications consisted of subdural or subgaleal fluid collections or hematomas requiring repeated operations, infections, seizures, and individual cases of a sunken bone flap, and 1 patient who became brain dead during the procedure. At cranioplasty, the mean age (± SD) of the 34 patients with complications was 48.6 ± 16.1 years, while the mean age of the 178 patients without complications was 42.7 ± 15.7 years, a statistically significant difference (p = 0.046). The mean interval between initial operation and cranioplasty for patients with complications was 5.6 ± 3.2 months (median 4 months, range 0 36 months). For patients without complications, the mean interval was 5.3 ± 5.1 months (median 4 months, range 0 months to 3 years). The difference here tended toward significance (p = 0.093). This result prompted further analysis, and the time between initial craniectomy and cranioplasty was divided as those performed within 3 months and those performed more than 3 months apart. Patients who underwent replacement within 3 months of removal were less likely to have complications than those with replacement surgeries after 3 months (9 vs 21%; p = 0.015). Results of complications stratified by categorical variable are summarized in Table 1. There was no statistical significance with rate of complication based on sex (p = 0.547) or operative side (p = 0.581). For statistical analysis, patient ages were subdivided into 3 groups: 0 39 years, years, and older than 60. Patients in the 0 39 group had the lowest complication (7%; p = 0.015). The and older than 60 age groups had complication rates of 20 and 26%, respectively. Overall, patients younger than 40 were less likely to experience complications than patients older than 40 (7 vs 22%; p = 0.005). The original indications for craniectomy were divided into 7 general categories: 1) subarachnoid hemorrhage for ruptured aneurysm; 2) traumatic injuries such as epidural, subdural, intraparenchymal hematomas, skull fractures, intractably elevated intracranial pressure, or penetrating trauma; 3) ischemic stroke; 4) tumors; 5) infectious processes such as cerebral abscesses, subdural empyemas, infected bone plates from previous craniectomies; 6) spontaneous intracerebral hemorrhages, including ruptured arteriovenous malformations; and 7) elective treatment for aneurysms or arteriovenous malformations. Of these categories, patients who originally presented with trauma had a lower rate of complications than those in other categories (9 vs 20%; p = 0.042). Conversely, patients who presented with tumors had a higher complication rate than those who did not (38 vs 15%; p = 0.023). Other variables such as the presence of a shunt, type 1121

3 V. Chang et al. of material used for the bone flap, and material used to secure the bone did not show statistically significant differences. The complication rate among patients with shunts was higher, but this finding did not reach statistical significance (20 vs 15%; p = 0.494). One variable that tended toward significance was postoperative drain placement (p = 0.087), which prompted further analysis as to whether there would be an increased incidence of postoperative fluid collection. Thirteen patients (6.1%) experienced fluid collection complications. The difference in the rate of fluid collection complications between patients with and without drain placement showed a trend toward significance (p = 0.065, chi-square test). Patients with drain placement had a fluid collection complication rate of 2.4% (2/84) compared with 8.6% (11/128) in patients without drain placement. In addition, we wanted to consider whether the material used for bone plate had any bearing on the rate of postoperative infections. Fifteen patients (7%) experienced an infectious complication. The difference in the rate of infections between patients who received autologous bone compared with other materials used was significant (p = 0.002). Patients with bone had an infection rate of 4.6% (8/175) compared with 18.9% (7/37) among patients with other materials (such as methylmethacrylate or titanium mesh). Logistic regression using backward elimination was used to assess the best combination of variables associated with complications. The variables to be considered in this analysis had probability values < 0.20 in the univariate analyses. These variables included diagnosis, age younger than 40 years, time to bone replacement < 3 months, material used to secure the bone (mini-plates vs sutures), and drain placement. For inclusion in the final model, variables had to have probability values < Table 2 displays the ORs, corresponding CIs, and probability values from the logistic regression. Age younger than 40 years, time to bone replacement < 3 months, and postoperative drain placement all reduced the risk of complications. Discussion As with any retrospective review, the results of the present study are influenced by certain biases that cannot be easily eliminated. Thus, the conclusions drawn here represent an observation of our collective experience with cranioplasty after craniectomy at our institution. The indications for performing hemicraniectomy were not standardized among patients. The decision to undertake the procedure depended on the clinical judgment of each individual surgeon. In addition, because multiple surgeons were involved in this study, there was no standardization of technique, which introduces a confounding factor. This difference in technique may also exist among individual surgeons because one would expect a surgeon s technique to evolve over the period of 13 years, especially in light of previous complications. In addition, the craniectomy size was variable and this could also have confounded the observed results, as intuitively larger bone flaps may lead to a higher complication rate. Given the above limitations, a significant finding of TABLE 1: Cranioplasty complication rates stratified by baseline variables* Variable No. of Patients No. W/ Complications (%) p Value sex F (14) M (17) age group (yrs) (7) (20) > (26) operative side lt (19) rt (14) both 11 2 (18) time from removal to replacement mos 89 8 (9) 4 6 mos (19) >6 mos (24) diagnosis SAH 65 9 (14) trauma 79 7 (9) ischemic stroke 14 4 (29) tumor 13 5 (38) infection 15 2 (13) ICH (includes ruptured AVM) 21 5 (24) elective aneurysm/avm 3 2 (67) other 2 0 (0) shunt placed no (15) yes 35 7 (20) drain placed no (20) yes 84 9 (11) material used bone (15) nonautologous bone 37 8 (22) material used to secure bone mini-plates (14) suture 35 9 (26) * AVM = arteriovenous malformation; ICH = intracerebral hemorrhage. the present study is the overall complication rate of 16% for cranial repair in a general neurosurgical practice. Con sidering the predisposing factors that might influence out comes, younger patients tended to do better, although an 8% rate for those younger than 40 years of age is still significant. The fact that patients with head trauma who underwent cranioplasty also had better outcomes is probably a reflection of age as well, because these patients were younger and probably had less brain atrophy or encephalomalacia resulting in less potential subdural space. In ad J Neurosurg / Volume 112 / May 2010

4 Cranial repair after craniectomy TABLE 2: Summary of multivariate logistic regression results Variable OR 95% CI p Value age <40 yrs time to replacement <3 mos postop drain placement dition, we also observed that patients who had a presenting diagnosis of tumor were at significantly higher risk of complications. In reviewing the complications among patients with tumors, we found no instances of concurrent chemotherapy or radiation therapy. One patient in particular had undergone 2 craniotomies previously for a recurrent meningioma. This patient had received radiosurgery 6 months after a previous resection, and was at least 2 years out from radiosurgery at the time she suffered complications of cranioplasty. A more likely factor could be that older age among patients with tumors increases the risk of complications after cranioplasty. The results for patients with ischemic stroke were also higher overall at 29%, although this difference was not statistically significant. Again, age combined with what would probably be an increase in comorbidities could also be a factor. With such a low number of stroke patients in the study (14 total), it is difficult to draw accurate conclusions. What does seem fairly certain is that when looking at predisposing factors that may influence the rate of complications after cranioplasty, age seems to be the central issue. Although certain presenting diagnoses may entail a lower (as with trauma) or higher risk (tumor) of complications, these differences could also be accounted for by age. This finding was supported by both in the univariate and multivariate analysis. To further our analysis we also wanted to evaluate which (if any) intraoperative factors would influence outcomes. Patients in whom sutures were used to secure the bone plate tended to have a higher rate of complications. After the introduction of titanium plating, sutures were no longer used for this purpose at our institution. Intuitively, one might expect an increased risk of subdural fluid collection after cranioplasty in patients who receive shunts because of the relative intracerebral hypotension that a shunt introduces. In addition, the presence of hardware could also introduce an increased risk of infection. The presence of a ventricular shunt did not seem to influence the complication rate after cranioplasty. The more recent use of programmable shunt valves may explain this result, as the ability to increase valve pressure in anticipation of bone flap replacement or after replacement can be a protective measure. The complication rates among patients with postoperative drains and those without tended toward statistical significance (10 vs 20%; p = 0.069). Because drains are meant to prevent the primary complications of postoperative fluid collection or hematoma we compared rates of postoperative fluid collection among patients with drain placement. Despite the fact that the risk of fluid collection among patients with drains was 2.4% compared with 8.6% in those without drains, the difference still tended J Neurosurg / Volume 112 / May 2010 toward statistical significance on univariate analysis. How ev er, multivariate analysis did illustrate drain placement as a statistically significant protective factor with an OR of 0.34 (95% CI ; p = 0.016). The use of autologous bone versus synthetic materials did not show any statistical difference. Using autologous bone after it has been frozen and stored has been thought to increase the risk of infection or transmission of infection; 27 however, there is evidence that this may not be the case. 11 In addition, the presence of foreign material may also predispose to infection. In our study, patients in whom autologous bone was used showed a significantly lower rate of complications (4.6 vs 18.9%; p = 0.002) compared with patients who received all nonautologous materials. At our institution, the preferred material for cranioplasty was autologous bone whenever possible, which is supported by the results of our analysis. Regarding the timing of surgery after craniectomy, the suggested practice is to wait 3 months after the initial operation before replacing the bone plate. 23 However, some patients have manifested symptoms as a result of large cranial defects and sunken scalps, the so-called syndrome of the trephined. 6,8,9 There are reports of reversal of these symptoms after cranioplasty. 21 In addition, the authors of other studies have demonstrated the possibility of derangement in CSF dynamics 7,8,17 and cerebral perfusion 15,22,28 in the absence of the skull plate. All of this has prompted some groups to evaluate whether a shorter time frame for cranioplasty may result in improved outcomes. 13 In our study the patients who underwent cranioplasty within 3 months had a lower overall complication rate of 9%. This was significant compared with patients who underwent cranioplasty > 6 months after craniectomy (p = 0.007), and tended toward significance in patients who underwent cranioplasty 4 6 months afterward (p = 0.07). This was also seen as a significant factor in a multivariate analysis with an OR of 0.28 (CI ; p = 0.004). In our practice, the interval between craniectomy and cranioplasty was based largely on individual patient recovery, and was not standardized. One observation is that patients with a long interval between the initial operation and cranioplasty tended to have more serious injuries and worse neurological outcomes after the initial operation. Therefore, such patients would be predisposed to have a worse outcome after cranioplasty. In summary, the overall rate of complications related to cranioplasty after craniectomy is greater than with other standard elective cranial procedures. Advanced age appears to correlate with increasing risk of complications. In addition, an interval of > 3 months between operations also appears to be a risk. The presence of a shunt did not have significant association with outcome. The use of a postoperative drain did show a benefit on multivariate analysis. The use of autologous bone graft material showed a decreased risk of infection compared with other materials. Although our study is retrospective in nature and as such does not provide extremely rigorous evidence to support any particular course of action, it does illustrate some important observations that we think are useful to consider. 1123

5 V. Chang et al. Conclusions Craniectomy is a procedure with wide indications and will continue to be a part of general neurosurgical practice in the foreseeable future. Complications related to this procedure are not uncommon as illustrated in our study, and certain factors may predetermine increased risk. Therefore, when evaluating the need to perform a large decompressive craniectomy, the surgeon should be aware that the patient is not only subject to the risks of the initial operation, but also the risks of subsequent cranioplasty. Most importantly, a heightened awareness of the patient s risk during elective cranial repair, typically after an extended recovery process, should be paramount in discussions with families as well as in the approach to perioperative care. Disclaimer The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Acknowledgments The authors wish to thank Lonni Schultz for her assistance with biostatistics, Farah Muller for help in compiling the patient database, and Sue MacPhee-Gray for editorial support. References 1. Aarabi B, Hesdorffer DC, Ahn ES, Aresco C, Scalea TM, Eisenberg HM: Outcome following decompressive craniectomy for malignant swelling due to severe head injury. J Neurosurg 104: , Bakay L: An Early History of Craniotomy: From Antiquity to the Napoleonic Era. Springfield, IL: Charles C Thomas, Bullock MR, Chesnut R, Ghajar J, Gordon D, Hartl R, Newell DW, et al: Surgical management of traumatic parenchymal lesions. Neurosurgery 58 (3 Suppl):S25 S46, Colohan AR, Ghostine S, Esposito D: Exploring the limits of survivability: rational indications for decompressive craniectomy and resection of cerebral contusions in adults. Clin Neurosurg 52:19 23, Compagnone C, Murray GD, Teasdale GM, Maas AI, Esposito D, Princi P, et al: The management of patients with intradural post-traumatic mass lesions: a multicenter survey of current approaches to surgical management in 729 patients coordinated by the European Brain Injury Consortium. Neurosurgery 57: , Dujovny M, Agner C, Aviles A: Syndrome of the trephined: theory and facts. Crit Rev Neurosurg 9: , Dujovny M, Fernandez P, Alperin N, Betz W, Misra M, Mafee M: Post-cranioplasty cerebrospinal fluid hydrodynamic changes: magnetic resonance imaging quantitative analysis. Neurol Res 19: , Fodstad H, Love JA, Ekstedt J, Friden H, Liliequist B: Effect of cranioplasty on cerebrospinal fluid hydrodynamics in patients with the syndrome of the trephined. Acta Neurochir (Wien) 70:21 30, Grant FC, Norcross NC: Repair of cranial defects by cranioplasty. Ann Surg 110: , Hutchinson PJ, Corteen E, Czosnyka M, Mendelow AD, Menon DK, Mitchell P, et al: Decompressive craniectomy in traumatic brain injury: the randomized multicenter RESCUEicp study ( Acta Neurochir Suppl 96:17 20, Iwama T, Yamada J, Imai S, Shinoda J, Funakoshi T, Sakai N: The use of frozen autogenous bone flaps in delayed cranioplasty revisited. Neurosurgery 52: , Jüttler E, Schwab S, Schmiedek P, Unterberg A, Hennerici M, Woitzik J, et al: Decompressive surgery for the treatment of malignant infarction of the middle cerebral artery (DESTI- NY). Stroke 38: , Liang W, Yang X, Liu W, Shen G, Zheng X, Cao F, et al: Cranioplasty of large cranial defect at an early stage after decompressive craniectomy performed for severe head trauma. J Craniofac Surg 18: , Meier U, Gräwe A, König A: The importance of major extracranial injuries by the decompressive craniectomy in severe head injuries. Acta Neurochir Suppl 95:55 57, Richaud J, Boetto S, Guell A, Lazorthes Y: Effects of cranioplasty on neurological function and cerebral blood flow. Neurochirurgie 31: , Sahuquillo J, Arikan F: Decompressive craniectomy for the treatment of refractory high intracranial pressure in traumatic brain injury. Cochrane Database Syst Rev 1:CD003983, Schiffer J, Gur R, Nisim U, Pollak L: Symptomatic patients after craniectomy. Surg Neurol 47: , Schwab S, Steiner T, Achoff A, Schwarz S, Steiner H, Jansen H, et al: Early hemicraniectomy in patients with complete middle cerebral artery infarction. Stroke 29: , Skoglund TS, Eriksson-Ritzen C, Jensen C, Rydenhag B: Aspects on decompressive craniectomy in patients with traumatic head injuries. J Neurotrauma 23: , Smith ER, Carter BS, Ogilvy CS: Proposed use of prophylactic decompressive craniectomy in poor-grade aneurismal subarachnoid hemorrhage patients presenting with associated large sylvian hematomas. Neurosurgery 51: , Stiver SI, Wintermark M, Manley GT: Reversible monoparesis following decompressive hemicraniectomy for traumatic brain injury. J Neurosurg 109: , Suzuki N, Suzuki S, Iwabuchi T: Neurological improvement after cranioplasty. Analysis by dynamic CT scan. Acta Neurochir (Wien) 122:49 53, Timmons RL: Cranial defects and their repair, in Youmans JR (ed): Neurological Surgery, ed 2. Philadelphia: WB Saunders, 1982, pp Timofeev I, Kirkpatrick PJ, Corteen E, Hiler M, Czosnyka M, Menon DK, et al: Decompressive craniectomy in traumatic brain injury: outcome following protocol-driven therapy. Acta Neurochir Suppl 96:11 16, Unterberg A, Jüttler E: The role of surgery in ischemic stroke: decompressive surgery. Curr Opin Crit Care 13: , Vahedi K, Vicaut E, Mateo J, Kurtz A, Orabi M, Guichard JP, et al: Sequential-design, multicenter, randomized, controlled trial of early decompressive craniectomy in malignant middle cerebral artery infarction (DECIMAL Trial). Stroke 38: , Vanaclocha V, Bazan A, Saiz-Sapena N, Paloma V, Idoate M: Use of frozen cranial vault bone allografts in the repair of extensive cranial bone defects. Acta Neurochir (Wien) 139: , Winkler PA, Stummer W, Linke R, Krishnan KG, Tatsch K: Influence of cranioplasty on postural blood flow regulation, cerebrovascular reserve capacity, and cerebral glucose metabolism. J Neurosurg 93:53 61, 2000 Manuscript submitted January 26, Accepted June 3, Please include this information when citing this paper: published online July 17, 2009; DOI: / JNS Address correspondence to: Donald Seyfried, M.D., Department of Neurosurgery, Henry Ford Health System, 2799 West Grand Boulevard, Detroit, Michigan nsdos@neuro.hfh.edu J Neurosurg / Volume 112 / May 2010