risk factors of aseptic bone resorption: a study after

J Neurosurg / Volume 118 / May 2013

J Neurosurg 118:1141–1147, 2013

1141

©AANS, 2013

Decompressive craniotomy can be considered a therapeutic option in life-threatening space-occu-pying intracranial pathologies, such as intracere-

bral bleeding, and traumatic or ischemic brain swelling when aggressive medical management is not sufficient to alleviate the increasing ICP.14,15,17 In the surviving pa-tients, after resolution of the cerebral swelling and consol-idation of the clinical state, cranioplasty with reinsertion of the removed bone flap is performed for mechanical, cosmetic, and therapeutic reasons.4,5,9,11,12,16,18,23 While the

technique, its complications, and the possible benefit of decompressive cranial surgery have been studied exten-sively, there are still a lot of open questions concerning the subsequent skull remodeling. Apart from the correct storing of the bone flap, the optimal timing of reinsertion is still under debate. Furthermore, the incidence and the role of possible long-term complications, which require recurrent surgical interventions, need to be elucidated. In particular, progressive aseptic necrosis of the bone flap remains a matter of concern during follow-up, occurring in up to 50% of the patients after cranioplasty.1,3,6,7,13 In this context, it would be reasonable to define a “high-risk” patient group that might benefit from an initial al-lograft. Therefore, it was the aim of this study to report

Risk factors of aseptic bone resorption: a study after autologous bone flap reinsertion due to decompressive craniotomy

Clinical article*Pedro dünisch, M.d.,1 Jan Walter, M.d.,1 Yasser sakr, M.d., Ph.d.,2 rolf kalff, M.d.,1 albrecht Waschke, M.d.,1 and christian eWald, M.d.1

Departments of 1Neurosurgery and 2Anaesthesiology, Jena University Hospital, Friedrich Schiller University Jena, Germany

Object. In patients who have undergone decompressive craniectomy, autologous bone flap reinsertion becomes necessary whenever the cerebral situation has consolidated. However, aseptic necrosis of the bone flap remains a concern. The aim of this study was to report possible perioperative complications in patients undergoing autologous bone flap reinsertion and to identify the risk factors that may predispose the bone flap to necrosis.

Methods. All patients admitted to the authors’ neurosurgical department between September 1994 and June 2011 and who received their own cryoconserved bone flap after decompressive craniectomy were studied. The grade of the bone flap necrosis was classified into 2 types. Type II bone necrosis was characterized by aseptic resorption with circumscribed or complete lysis of tabula interna and externa requiring surgical revision. To define predisposing fac-tors, a multivariate analysis was performed using bone necrosis as the dependent variable.

Results. Among the 372 patients (mean age 48.6 years, 57.4% males) who received 414 bone flaps during the observation period, 134 (36.0%) had a diffuse traumatic brain injury, 69 (18.5%) had subarachnoid hemorrhage, 58 (15.6%) had cerebral infarction, 56 (15.1%) had extraaxial bleeding, 43 (11.6%) had intracerebral bleeding, and 12 (3.2%) had a neoplasm. Surgical relevant Type II bone flap necrosis occurred in 85 patients (22.8%) and 91 bone flaps, after a median time of 15 months (interquartile range [IQR], 10–33 months). In a multivariate analysis with Type II necrosis as the dependent variable, bone flap fragmentation with 2 (OR 3.35, 95% CI 1.59–7.01, p < 0.002) or more fragments (OR 24.00, 95% CI 10.13–56.84, p < 0.001), shunt-dependent hydrocephalus (OR 1.76, 95% CI 0.99–3.12, p = 0.04), and a younger age (OR 0.98, 95% CI 0.96–0.99, p = 0.004) was associated with a higher risk for the development of an aseptic bone flap necrosis.

Conclusions. In patients undergoing bone flap reinsertion after craniotomy, aseptic bone necrosis is an under-estimated problem during long-term follow-up. Especially in younger patients with an expected good neurological recovery and a fragmented bone flap, an initial allograft should be considered because of an increased risk for aseptic bone flap necrosis.(http://thejns.org/doi/abs/10.3171/2013.1.JNS12860)

keY Words • craniotomy • cranioplasty • bone flap • aseptic bone flap necrosis • bone resorption

1141

Abbreviations used in this paper: ICP = intracranial pressure; IQR = interquartile range; SAH = subarachnoid hemorrhage.

* Drs. Dünisch and Walter contributed equally to this work.

P. Dünisch et al.

1142 J Neurosurg / Volume 118 / May 2013

possible perioperative complications in patients undergo-ing bone flap reinsertion after craniotomy and to identify the risk factors predisposing to bone flap necrosis in these patients.

MethodsPatients

We included 372 patients undergoing bone flap re-insertion after craniotomy (414 bone flaps) admitted to our institution between September 1994 and June 2011. The study was approved by the local ethics commit-tee. Informed consent was not required because of the anonymous, observational, and retrospective nature of the analysis.

Operation and Follow-UpAfter craniectomy, the removed bone flap was cleaned

of adjacent soft tissue, wrapped in sterile cloths, and stored at -80°C. Eighty-five bone flaps were fragmented by a fracture or a former, smaller craniotomy. In these cases, the flap was reconstructed using tiny osteosynthetic titani-um plates before freezing it. All bone flaps were routinely cultured during the operation after surgical removal. We reinserted exclusively the flaps that showed no evidence of bacterial contamination. Infected flaps were discarded and not included in the study. All flaps were stored for a median of 78 days.

For reinsertion, the bone flap was passively defrosted in 37°C sodium chloride solution. After reopening the former skin incision, the bony rims of the skull defect and a dural layer were prepared. The bone flap was reinserted and fixed with rigid titanium plates (Biomet). If neces-sary, CSF was drained from the frontal horn periopera-tively. The dural layer was fixed to the bone flap with 3–5 central sutures in all cases to prevent postoperative epi-dural hematoma. At least 2 nonsuction drains were placed below and above the bone flap so that an outflow of epi-dural and subgaleal fluid was ensured. These drains were removed within the first 2 postoperative days. All patients received an antibiotic prophylaxis as long as these drains were in place.

Follow-up was scheduled in our outpatient depart-ment, according to an institutional protocol, with clinical and CT examinations at 6 weeks, 3 months, 6 months, 12 months, and yearly thereafter. In cases of a suspected or manifest complication, this regimen was varied depend-ing on the clinical and radiological findings. The end point of the data acquisition was the CT diagnosis of a Type II bone flap necrosis (see below). If there were no radiological signs of bony destruction or Type I bone flap necrosis was obvious, the last available CT scan defined the follow-up result.

Data AcquisitionPatients’ files and radiological images were screened

retrospectively. We created a data set containing informa-tion on sex, age at the first operation, diagnosis, period between removal and replacement of the bone flap, size of the bone flap, duration of follow-up, and the presence of a CSF shunt. The diagnosis leading to craniotomy was clas-

sified as diffuse brain injury including traumatic brain swelling and traumatic intracerebral hematoma, brain swelling caused by SAH including vasospastic infarc-tion, ischemic stroke (malignant middle cerebral artery infarction), tumor, other spontaneous intracerebral bleed-ing, and extraaxial hematomas like subacute or chronic subdural hematomas.

The relative 2D size of the trephination defect was calculated by the senior author (C.E.) on a preoperative sagittal radiograph or a sagittal CT scan using the follow-ing formula: A = p/4 × B × b, where p is a constant and B and b are the longest diameters of an elliptic area (Fig. 1).

To quantify the extent of a possible aseptic bone ne-crosis, we defined 2 different types of necrosis according to the CT features with W/L 2500/500 HU. A thinning of the bone flap and/or a beginning resorption along the rims of the flap was classified as Type I necrosis, whereas a Type II bone flap necrosis was characterized by a cir-cumscribed, complete lysis of the bone within the flap, including tabula interna and externa with loss of the bony protection of the brain (Fig. 2). According to our institu-tional protocol, Type II bone flap necrosis was considered to be an indication for surgical revision, with resection of the necrotic bone and implantation of an allograft. The median follow-up duration was 11 months (IQR 2–25 months), with 265 flaps being observed for more than 6 months.

Statistical AnalysisData were analyzed using SPSS version 17.0 software

for Windows (SPSS Inc.). Discrete variables are expressed as total number and percentage and continuous variables as mean ± SD or median and IQR unless stated otherwise. Categorical data were compared using the chi-square test or by Fisher exact test, as appropriate. Continuous vari-ables conforming to a normal distribution were compared using Student t-test. Otherwise the Mann-Whitney U-test was applied. Because of its clinical impact, we focused on the surgical relevant Type II bone flap necrosis. To iden-tify possible risk factors, a multivariate logistic regression analysis was performed using Type II bone flap necrosis as the dependent variable. A univariate analysis was per-formed prior to modeling, and variables were considered if the p value was < 0.2. Colinearity between variables was excluded before modeling (R2 > 0.7). The variables cho-sen for the multivariate analysis included age (per year), admission primary diagnosis, period to reinsertion of the bone flap (in months), number of fragments (categorized as 2 or > 2 fragments), bilateral versus unilateral bone flap reinsertion, shunt-dependent hydrocephalus, and the size of the bone flap. A Hosmer-Lemeshow test was used to assess the goodness of fit of the multivariate model. Co-variates were retained in the model if the p value was < 0.2, and the OR and 95% CI were computed. All statistics were 2-tailed, and a p value < 0.05 was considered to be statistically significant.

ResultsPatient Characteristics

Among 372 patients (mean age 48.6 ± 18.4 years;


Aseptic bone necrosis

1143

57.4% males) who underwent autologous bone flap rein-sertion in our institution during the observation period, 134 (36.0%) had a diffuse traumatic brain injury, 69 (18.5%) had SAH, 58 (15.6%) had cerebral infarction, 56 (15.1%) had extraaxial bleeding, 43 (11.6%) had intracere-bral bleeding, and 12 (3.2%) had a neoplasm. One hundred forty-one patients (37.9%), in whom 169 flaps were placed, had a shunt-dependent hydrocephalus, with 84 having a CSF shunt implanted before the bone flap reinsertion. In all others the drain was placed within 3 months. In 42 patients (11.3%), bilateral bone flap implantations were necessary. The characteristics of this study group are pre-sented in Table 1.

A total of 414 bone flaps were reinserted after a me-dian of 78 days (IQR 54–113 days), and the mean bone flap size was 85.7 ± 23.06 cm2. In 85 cases, fragmented bone flaps were removed during the initial operation due to the severity of trauma and resulting multiple skull frac-tures. In 43 patients (10.4%) the bone flap consisted of 2 fragments, and in 42 patients (10.1%) there were 3 or more fragments. At the time of reimplantation, the individual parts were rejoined together and placed again as a single bone flap.

Early Complications and ReoperationsConcerning the perioperative complication rate

(within the period of hospital stay), 5 patients died after the operation, the causes being pulmonary embolism (n = 3), myocardial infarction (n = 1), and brainstem ischemia (n = 1). The 30-day mortality rate was 1.2%. Furthermore, there were 67 complications during the hospital stay lead-ing to a second surgical intervention. In 24 cases (6.5%), operative revision was necessary for epidural, subdural, or intracerebral bleeding (n = 23). One patient had to undergo reoperation for a subgaleal hematoma. Subdu-

TABLE 1: Patient and bone flap characteristics

Characteristic No. of Patients/Value

no. of patients 372mean age in yrs ± SD 48.6 ± 18.4male/female ratio (%) 214 (57.4) to 158 (42.6)diagnosis (%) diffuse brain injury 134 (36.0) SAH 69 (18.5) cerebral infarction 58 (15.6) extraaxial bleeding 56 (15.1) intracerebral bleeding 43 (11.6) tumor 12 (3.2)unilat/bilat ratio bone flap (%) 330 (88.7) to 42 (11.3)shunt-dependent hydrocephalus (%) 141 (37.9)bone flap characteristics no. of bone flaps 414 median days to replacement (IQR) 78 (54–113) mean bone flap size (cm²) ± SD 85.7 ± 23.06 ≤70 (%) 87 (21.0) 71–90 (%) 160 (38.6) 91–110 (%) 117 (28.3) >110 (%) 50 (12.1) fragmentation (%) 0 fragments 329 (79.5) 2 fragments 43 (10.4) >2 fragments 42 (10.1)median mos to follow-up CT (IQR) 11 (2 ± 25) ≤6 (%) 149 (36.0) 7–12 (%) 73 (17.6) 13–24 (%) 83 (20.0) 25–48 (%) 51 (12.3) >48 (%) 58 (14.0)

Fig. 2. Different types of bone flap resorption according to our clas-sification: Type I (left) with left-sided thinning of the flap and Type II (right) with a complete lysis of the bone within the flaps.

Fig. 1. Measuring the 2D size of the bone flap as A = p/4 × B × b.

P. Dünisch et al.


roperitoneal shunting was necessary in 16 cases (4.3%), wound debridement for wound necrosis or deep infection in 26 cases (7.0%), and a re-removal of the bone flap for progressive brain swelling in 1 case (0.3%) (Table 2). A dislocation of the flap during follow-up, requiring a repeat fixation, occurred in 3 patients. Thus, the overall short-term complication rate, referring to operative revisions (n = 67) and the refixation of loosened flaps (n = 3), was 18.8%.

Aseptic Bone NecrosisOn the last CT scans that were available, 131 (31.6%)

of 414 bone flaps exhibited the radiological features of Type I bone flap necrosis after a median duration of 19 months (IQR 11–46 months). An aseptic necrosis with complete lysis of the tabula interna and externa (Type II necrosis) occurred in 85 patients (mean age 41.7 ± 19.3, 56.5% males) and 91 bone flaps (21.9%) after a median duration of 15 months (IQR 10–33 months), from the time of bone flap reinsertion onward (Table 3). In all cases, we observed a clinical indication to remove the necrotic bone flap. In patients diagnosed with necrotic flaps, cosmetic disfiguring was the main complaint—for example, for hypermobility of the flap, displacement of osteosynthet-ic material, or aesthetical reasons. Thirty-nine necrotic bone flaps were resected and replaced by an allograft (ti-tanium in 30, Bioverit in 7, and Palacos in 2 cases). In the remaining 52 cases, the reasons for nonreplacement of the Type II necrotic bone flap were as follows: patient or his/her representative refused the operation (n = 24), patient in a vegetative neurological state (n = 20), and patient lost to follow-up after the initial CT scan (n = 8). Patients in whom necrosis of the reinserted bone flap developed were significantly younger (41.7 ± 19.3 vs 50.6 ± 17.7 years, p = 0.009) (Table 3). Necrotic bone flaps (n = 91) were more likely to be fragmented than those that were not necrotic (Table 2). As a quality control, regular follow-up CT scanning of all nonnecrotic flaps was performed.

Risk Factors for Bone Flap NecrosisIn a multivariate logistic regression analysis with

bone flap necrosis Type II as the dependent variable (414 procedures were considered), increasing age was associ-ated with low risk (OR 0.98 [per year], 95% CI 0.96–0.99, p = 0.004) of bone flap necrosis, whereas shunt-dependent hydrocephalus (OR 1.76, 95% CI 0.99–3.12, p = 0.04) and fragmentation into 2 fragments (OR 3.35, 95% CI 1.59–7.01, p < 0.002) or more than 2 fragments (OR 24.00, 95% CI 10.13–56.84, p < 0.001) were associated with a signif-icantly higher risk for a Type II bone flap necrosis (Ta- ble 3).

DiscussionSkull remodeling after decompressive craniotomy

makes a second operation necessary. For aesthetic, psy-chological, and economic reasons the autologous bone flap should be the preferred graft. In addition to cosmetic aspects, clinical and neurological improvements have also been described.3,4,11,12 However, a second operation always carries the risk of new complications, especially

in patients who are often weakened by the impact of the initial event, such as brain injury or ischemic stroke. Fur-thermore, aseptic necrosis of the reinserted bone flap has been reported in 7.2%–50% of the cases, although the reasons for its development are still unknown.7,9 Thus, a better understanding of risk factors for the development of bone necrosis after autologous cranioplasty is impor-tant.4,5,12,16,18,23

To our knowledge, this study is the most extensive survey of a large patient group concerning the complica-tions after autologous cranioplasty, paying special atten-tion to necrosis of the bone flap after reinsertion. Because of its retrospective single-institution design, however, there are some limitations that need to be considered. At first, there were multiple physicians involved in the han-dling of aseptic bone necrosis and in deciding on the re-placement of the necrotic bone flap. Until now there has been no standardization of the operative technique and postoperative treatment, but due to a rigid internal pro-tocol a certain scheme for the outpatient follow-up has been fixed. Nevertheless, because 50 patients were lost to follow-up after 1 month, the incidence of long-term com-plications may be higher than we have reported.

The methods of measuring the size of the bone flap and quantifying the extent of bone necrosis were not vali-dated, but the fact that a single investigator (senior author) has collected these data can minimize potential errors. It also seems problematic to measure the size of a 3D flap with a formula describing a 2D area. However, it was our aim to investigate the influence of the autograft size on the incidence of aseptic necrosis rather than to define the exact bone flap size. For that reason, it appears to be suf-ficient to have the relative size of a flap compared with smaller or larger ones.

In this cohort, almost one-quarter (18.8%) of the pa-tients had a procedure-related early postoperative compli-cation, leading to further surgical interventions. However, this rate is within the reported range,3,6,9,20,21,23 depending on the operative technique and the definition of surgery-

TABLE 2: Early complications requiring repeat surgical interventions after reinsertion of 414 autologous bone flaps in 372 patients

Complication

No. of Patients

(%) Therapy

epidural, subdural, or intracerebral bleeding*

23 (6.2) operative revision

space-occupying chronic epi- or subdural fluid collection

16 (4.3) subduroperitoneal shunt

deep infection/epidural empyema 16 (4.3) removal of bone flap, debridement, allo- plastic implant after 3–12 mos

wound necrosis 10 (2.7) operative revisionbrain swelling 1 (0.3) removal of bone flapsubgaleal hematoma 1 (0.3) operative revision

* Within 5 days of the operation.



1145

related complications. Compared with other neurosurgi-cal “standard” procedures, this is a notably high value, and in the authors’ opinion the seriousness of the bone flap reinsertion as a second surgical intervention is still underestimated. Regardless of prolonged aseptic bone necrosis, rebleeding and space-occupying epi- and sub-dural fluid collections were the most important acute postoperative complications. Especially in patients with a preexisting shunt, a possible negative pressure may fa-cilitate blood collection in a transiently created new space below the bone flap, leading to an acute space-occupying hematoma or chronic epi- or subdural hygroma. For this reason, we always place an epidural drain, as also recom-mended by other authors. Wound healing problems oc-curred in 10 patients. Deep infections, especially epidural empyema, occurred in only 4.3% of the cases. Compared with other studies reporting a rate of deep infections be-tween 7% and 12%, our infection rate is low, confirming our prophylactic antibiotic regimen.3,6,21

The main focus of our study, however, was to analyze the incidence of and possible risk factors for aseptic ne-crosis. Aseptic necrosis is a long-term complication lead-ing to progressive destruction of the bone, with successive cosmetic disfiguring and loss of bony brain protection. As we have mentioned, our data have to be interpreted carefully because there is no overall accepted definition

of “bone necrosis” after cranioplasty. In our cohort 222 flaps (53.6%) showed CT signs of an impaired osseous integration: Type I necrosis in 131 and Type II necrosis in 91. Bone necrosis requiring surgical intervention was retrospectively detectable in 21.9% of the implanted bone flaps. The median time to diagnosis in our group was 15 months. Interestingly the median follow-up duration in patients who developed Type I necrosis was 19 months, showing that a beginning resorption of the flap is not nec-essarily the beginning of a progressive destruction of the bone. However, only 39 flaps were removed and replaced by an allograft. In all other cases, the patients were in a poor neurological condition, were lost to follow-up, or refused a third intervention. In the literature the incidence of delayed aseptic bone necrosis approaches 50%, espe-cially in younger patients with a longer follow-up peri-od.7,10 Thus, time dependency has to be considered, and in the context of this background a follow-up of 2 years seems reasonable.

The resorption rate may appear at first seem higher than expected, but the incidence could be explained by the performance of repeated CT scanning over a short period of time. According to an institutional protocol, clinical and CT follow-up visits were scheduled at 6 weeks, 3 months, 6 months, and 12 months in our outpatient department. The majority of patients were profoundly weakened by

TABLE 3: Characteristics of the study group stratified by occurrence of necrosis

CharacteristicNo. of Patients

p ValueNo Necrosis Necrosis

no. of patients 287 85mean age in yrs ± SD 50.6 ± 17.7 41.7 ± 19.3 0.009male/female ratio (%) 166 (57.7) to 121 (42.2) 48 (56.5) to 37 (43.5) 0.496diagnosis (%) 0.233 diffuse brain injury 93 (32.4) 41 (48.2) SAH 55 (19.2) 14 (16.5) cerebral infarction 49 (17.1) 9 (10.6) extraaxial bleeding 46 (16.0) 10 (11.8) intracerebral bleeding 36 (12.5) 7 (8.2) tumor 8 (2.8) 4 (4.7)shunt-dependent hydrocephalus (%) 101 (35.2) 40 (47.1) 0.342unilat/bilat ratio bone flap (%) 251 (87.4) to 36 (12.6) 79 (92.9) to 6 (7.10) 0.478bone flap characteristics no. of bone flaps 323 91 median days to replacement (IQR) 80 (58–115) 74 (46–102) 0.068 mean size (cm²) ± SD 85.12 ± 23.06 87.68 ± 23.06 0.071 ≤70 (%) 67 (20.7) 20 (22.0) 71–90 (%) 140 (43.3) 20 (22.0) 91–110 (%) 81 (25.1) 36 (39.6) >110 (%) 35 (10.8) 15 (16.5) fragmentation (%) 11.8 51.7 <0.001 0 fragments 285 (88.2) 44 (48.4) 2 fragments 28 (8.7) 15 (16.5) >2 fragments 10 (3.1) 32 (35.2)median mos to follow-up CT (IQR) 9 (1–23) 15 (10–33) <0.001

P. Dünisch et al.


the initial trauma, making clinical status an insignificant diagnostic factor. Therefore, we performed regular CT follow-up. An added advantage was being able to monitor the ventricular size and possible increase in ICP. All inter-ventions based on radiographic findings correlated with a clinical complaint at the time of operation. All surgically treated patients exhibited pre-CT clinical symptoms in the presence of an aseptic necrosis.

There are many possible reasons being discussed for progressive necrosis of the reinserted bone flap. We found age, fragmentation, and shunt dependency to be significant and independent predictive parameters for the development of an aseptic bone necrosis. Hence, young patients with a fragmented flap and a CSF shunt in place have the highest risk. Fragmentation is the most critical factor triggering a nutritional deficit of the flap. This also explains the predisposition of patients after severe head injury. Fragmentation and a younger age are associated with a higher risk of bone flap necrosis because many of these patients initially had a serious accident or injury, like a car or bike accident, and in such patients we saw most of the fragmented flaps.

An exact congruity between the bone edge and the bone flap is supposed to be the most important factor for osseous integration of the reinserted necrotic bone.10 As recommended by Chang et al.,3 the fixation should be as rigid as possible, so we always use mini-osteosynthetic ti-tanium plates for bone flap refixation, as described above.

The impact of existent shunt-dependent hydroceph-alus is remarkable, but a definitive cause could not be isolated. We found a study that describes a higher com-plication rate in such operations after laying an external ventricular drain.20 Nonphysiological changes of the ICP, whether an increase or decrease, may be of importance. We suspect that the bur hole leads to a variation in pres-sure levels in the cranium, resulting in the incomplete ad-herence of the dura mater to the bone flap and, eventually, to insufficient blood supply. However, further studies are required to confirm this theory.

Another issue is the role of operative timing in the context of bone resorption. The majority of the authors prefer performing cranioplasty about 3 months after cra-niectomy when the neurological situation is stable and when there is a tight tissue layer that protects the underly-ing brain, facilitating the preparation and the placement of the flap.3,22 An early operation—within a few weeks of the first operation—has also been described with comparable results.2,13 We could not confirm a significant advantage of early reimplantation. Regarding the period of time to replacement, there is no significant distinction detectable; there was only a trend (p = 0.07), bearing in mind that we had fewer cases of aseptic bone necrosis in patients in whom bone flap were replaced between Months 3 and 12 after craniectomy. The reason for this discrepancy re-mains unclear, but in conclusion the issue of the timing seems to be less important than previously thought.

The last issue to be covered is the storage of bone flaps. In our patients, the grafts were stored at -80°C. Some authors presume that a higher incidence of asep-tic bone resorption occurs in cryoconservation compared with flaps stored in the abdominal subcutaneous fat.8,10

However, with the latter method exists the disadvantage of a second operative site, and the attendant complication rate has not yet been identified adequately. Also, lysis of the subcutaneous implanted flap by macrophages is also possible.19 As a result, freezer storage has been the favor-able method of storage in our clinic until recently.

Unfortunately, we have no accurate data about the costs of cranioplasty when using frozen bone flaps. We assume that it would be cheaper to start with an allograft in the “high-risk” group of patients. When we add the costs of removing, storing, culturing, and reimplanting the flap, in addition to the cost of extracting the necrotic flap, the initial implantation of an allograft seems to more cost effective, especially in patients with 2 or more risk factors for developing an aseptic necrosis.

ConclusionsWhile craniectomy has become an inherent part in

the treatment of life-threatening increased ICP, many unanswered questions concerning the following cranio-plasty remain. In this retrospective study of a large pa-tient group, we were able to show that complications are not uncommon and that even long-term problems like an aseptic bone necrosis should be considered during follow-up outpatient reassessment. Surgeons and neurologists should not only be aware of the risks of the first operation but also of the cranioplasty. These data have to be reflect-ed in the therapeutic regimen and also in future studies dealing with the effect of decompressive surgery.

This survey shows that bone flap fragmentation, younger age, and shunt dependency are possible risk fac-tors for the development of bone flap necrosis. Further study including long-term data is mandatory, especially concerning the question of whether there is a high-risk patient group that would benefit from an initial allograft.

Disclosure

The authors report no conflict of interest concerning the mate-rials or methods used in this study or the findings specified in this paper.

Author contributions to the study and manuscript preparation include the following. Conception and design: Dünisch. Critically re vising the article: all authors. Reviewed submitted version of man-u script: all authors. Approved the final version of the manuscript on behalf of all authors: Dünisch. Statistical analysis: Sakr. Ad min i-stra tive/technical/material support: Kalff. Study supervision: Ewald.

References

1. Beauchamp KM, Kashuk J, Moore EE, Bolles G, Rabb C, Seinfeld J, et al: Cranioplasty after postinjury decompressive craniectomy: is timing of the essence? J Trauma 69:270–274, 2010

2. Carvi Y Nievas MN, Höllerhage HG: Early combined cranio-plasty and programmable shunt in patients with skull bone de-fects and CSF-circulation disorders. Neurol Res 28:139–144, 2006

3. Chang V, Hartzfeld P, Langlois M, Mahmood A, Seyfried D: Outcomes of cranial repair after craniectomy. Clinical article. J Neurosurg 112:1120–1124, 2010



1147

4. Chieregato A: The syndrome of the sunken skin flap: a ne-glected potentially reversible phenomenon affecting recovery after decompressive craniotomy. Intensive Care Med 32: 1668–1669, 2006

5. Dujovny M, Fernandez P, Alperin N, Betz W, Misra M, Mafee M: Post-cranioplasty cerebrospinal fluid hydrodynamic chang-es: magnetic resonance imaging quantitative analysis. Neurol Res 19:311–316, 1997

6. Gooch MR, Gin GE, Kenning TJ, German JW: Complications of cranioplasty following decompressive craniectomy: analy-sis of 62 cases. Neurosurg Focus 26(6):E9, 2009

7. Grant GA, Jolley M, Ellenbogen RG, Roberts TS: Failure of autologous bone-assisted cranioplasty following decompres-sive craniectomy in children and adolescents. J Neurosurg 100 (2 Suppl Pediatrics):163–168, 2004

8. Häuptli J, Segantini P: [New tissue preservation method for bone flaps following decompressive craniotomy.] Helv Chir Acta 47:121–124, 1980 (Ger)

9. Honeybul S, Ho KM: Long-term complications of decompres-sive craniectomy for head injury. J Neurotrauma 28:929–935, 2011

10. Iwama T, Yamada J, Imai S, Shinoda J, Funakoshi T, Sakai N: The use of frozen autogenous bone flaps in delayed cranio-plasty revisited. Neurosurgery 52:591–596, 2003

11. Joseph V, Reilly P: Syndrome of the trephined. Case report. J Neurosurg 111:650–652, 2009

12. Kemmling A, Duning T, Lemcke L, Niederstadt T, Minnerup J, Wersching H, et al: Case report of MR perfusion imaging in sinking skin flap syndrome: growing evidence for hemody-namic impairment. BMC Neurol 10:80, 2010

13. Liang W, Xiaofeng Y, Weiguo L, Gang S, Xuesheng Z, Fei C, et al: Cranioplasty of large cranial defect at an early stage after decompressive craniectomy performed for severe head trauma. J Craniofac Surg 18:526–532, 2007

14. Meyer MJ, Megyesi J, Meythaler J, Murie-Fernandez M, Au-but JA, Foley N, et al: Acute management of acquired brain in-jury part I: an evidence-based review of non-pharmacological interventions. Brain Inj 24:694–705, 2010

15. Sahuquillo J, Arikan F: Decompressive craniectomy for the treatment of refractory high intracranial pressure in traumat-

ic brain injury. Cochrane Database Syst Rev 1:CD003983, 2006

16. Schorl M: Sinking skin flap syndrome (SSFS)—clinical spec-trum and impact on rehabilitation. Cent Eur Neurosurg 70: 68–72, 2009

17. Schwab S, Steiner T, Aschoff A, Schwarz S, Steiner HH, Jan-sen O, et al: Early hemicraniectomy in patients with complete middle cerebral artery infarction. Stroke 29:1888–1893, 1998

18. Segal DH, Oppenheim JS, Murovic JA: Neurological recovery after cranioplasty. Neurosurgery 34:729–731, 1994

19. Shoakazemi A, Flannery T, McConnell RS: Long-term out-come of subcutaneously preserved autologous cranioplasty. Neurosurgery 65:505–510, 2009

20. Stephens FL, Mossop CM, Bell RS, Tigno T Jr, Rosner MK, Kumar A, et al: Cranioplasty complications following war-time decompressive craniectomy. Neurosurg Focus 28(5): E3, 2010

21. Stiver SI: Complications of decompressive craniectomy for traumatic brain injury. Neurosurg Focus 26(6):E7, 2009

22. Waziri A, Fusco D, Mayer SA, McKhann GM II, Connolly ES Jr: Postoperative hydrocephalus in patients undergoing de-compressive hemicraniectomy for ischemic or hemorrhagic stroke. Neurosurgery 61:489–494, 2007

23. Yamaura A, Sato M, Meguro K, Nakamura T, Uemura K: [Cranioplasty following decompressive craniectomy—analy-sis of 300 cases (author’s transl).] No Shinkei Geka 5:345–353, 1977 (Jpn)

Manuscript submitted May 1, 2012.Accepted January 17, 2013.Portions of this work were given in oral presentation form at the

14th European Congress of Neurosurgery, Rome, Italy, October 9–14, 2011.

Please include this information when citing this paper: pub-lished online March 1, 2013; DOI: 10.3171/2013.1.JNS12860.

Address correspondence to: Pedro Dünisch, M.D., Department of Neurosurgery, Hospital of the Friedrich Schiller University, Er langer Allee 101, 07747 Jena, Germany. email: [email protected].

risk factors of aseptic bone resorption: a study after

Documents