Outcomes of Cranioplasty from a Tertiary Hospital in a Developing Country
Keywords: Complications, cranioplasty, decompressive craniectomy, delayed, early CP infections, timingKey Messages: Cranioplasty procedures may result in signiifcant morbidities. Early procedure (<12 weeks) seems to be beneficial in terms of reduction of complication rates. Poor GOS score at admission is associated with higher complications.
Although cranioplasty (CP) is a straightforward procedure, a significant number of complications may occur, ranging from 10.5% to 50%. These include infections, seizures, intracranial hematomas, and rarely mortality.,,, Many reports have stated that early CP is associated with higher complications; however, more recent articles have contradicted this.,,,,, We intend to share our tertiary center experience on the complications and the determinants of these complications after CP.
The authors retrieved the hospital data of patients undergoing CP over the past 3 years (Jan 2014–December 2016) at our university tertiary hospital. We excluded cranioplasties done after tumor removal because they form an entirely different subgroup. Routinely, after DC, the bone flap would be preserved in the anterior abdominal wall in a majority of the cases. In cases of compound head injuries (HIs) such as depressed fractures, the involved bone would be discarded. If an abdominal flap infection occurs before CP, such bone would be discarded and the CP procedure would be performed using polymethyl methacrylate (PMMA)/other allogenic material. All patients would get a computed tomography done on admission to look for the presence of ventriculomegaly. As a protocol, before CP, patient's scalp would be shaved and a thorough antiseptic solution scalp wash would be given at least 2 days prior.
During the procedure, after retrieval from the abdominal wall, the bone flap would be thoroughly cleansed with saline and hydrogen peroxide. In cases with no autologous bone flap, cranioplasties were performed using PMMA, alone or combined with a titanium mesh. A subgaleal suction drain was placed after every procedure and removed after 12–24 h. All procedures were performed by registrars-in-training (with at least 1 year experience in neurosurgery) under the supervision of a consultant. All patients received a minimum of 48 h up to a maximum of 5 days of antibiotics after CP. The majority were discharged on third postoperative day (POD) and asked to return for suture removal on POD 10.
The demographic profile, etiology of decompressive craniectomy (DC), time interval between the two surgeries, operative details, complications, and follow-up data were retrospectively analyzed. For the timing of CP, we arbitrarily divided the DC-CP interval into three categories: very early (<6 weeks), early (6–12 weeks), and late (>12 weeks) CP, and correlation of complications and outcome between these groups was studied. All statistical analyses were done using SPSS software (version 23.0). For statistical analysis, Fisher's exact test was applied to look for statistical significance. A P value of <0.05 was termed significant.
We analyzed 93 consecutive cases of cranioplasties during the study period. The etiologies of the initial DC were as follows: traumatic brain injury (TBI) (acute subdural hematoma/contusion/depressed fractures), n = 55 (59.1%); ischemic stroke, n = 25 (26.8%); cortical venous thrombosis (CVT), n = 6 (6.4%); lobar hemorrhage, n = 5 (5.4%); and vascular etiology, n = 2 (2.2%).
The mean and median ages were 38.3 years (range 1–-68 years) and 38 years, respectively. The male:female ratio was 2.57:1. The mean age for trauma was 33.6 years (17–60 years,) while the mean age for infarct presentation was 47.9 years (range 29–63 years). There were eight open/compound HIs and all were associated with frontal depressed fractures involving the frontal sinus. Ten were bifrontal, 1 bilateral hemispheric, and the remaining were unilateral cranioplasties. Among unilateral ones, there were 40 (43%) right-sided cases and 42 (45%) left-sided cases.
The mean time to CP after DC was 8.5 weeks (range 0–28 weeks) (0 days meant intraoperative placement) and the median duration was 8 weeks. On subdividing, it was found that the mean interval was 8.2 weeks among trauma presentation, while it was 9.7 weeks for infarct etiology. With regard to timing of CP, there were 27 (29%) cases in very early, 51 (54.8%) in early, and 15 (16.1%) in late categories. Of these, a majority (83.8%) of patients had CP done within 12 weeks of DC. Based on the material used, autologous bone flap CP was done in 84.9% (n = 79) cases, PMMA in 9.6% (n = 9), and autologous bone augmented with acrylic and titanium mesh or titanium alone in 5.4% (n = 5) cases. Preoperative ventriculomegaly/hydrocephalus was noted in 9 (9.8%) cases. They were managed with intraoperative lumbar drain and ventricular tap in three and two cases, respectively, and one underwent ventriculoperitoneal (VP) shunt insertion. The remaining three cases underwent no extra interventions. [Table 1] summarizes the demographic and clinical data of the study cohort.
After the primary DC surgery, four patients had developed wound infection that was managed with antibiotics and resuturing, and four patients had re-surgeries for various reasons (evacuation of postoperative intracranial hematomas and subgaleal collection). There were 10 complications noted after CP: wound infection (n = 3) (of which two cases subsequently required bone flap removal), wound dehiscence (n = 2), seizures (n = 2), epidural hematoma (n = 1), and parenchymal hematoma (n = 1). One case developed hydrocephalus 2 months after CP that was managed by VP shunt insertion. Seizures were controlled with addition of an antiepileptic drug and wound dehiscence was managed with resuturing. The complication rate was 10.7% (10 of 93). Of those with preoperative complications (n = 8), only one case developed superficial wound infection managed conservatively and one developed seizures. The patient with extradural hematoma had grossly deranged coagulation parameters to which he succumbed. One patient died 4 months after CP due to an unrelated acute abdominal problem (feeding jejunostomy had been done 2 months prior). The mean postoperative hospital stay was 4.8 days and the median duration was 3 days. The mean follow-up was 9.1 months (range 3—19 months).
[Table 2]a and [Table 2]b summarize the complications noted in this study after CP. On correlating the different variables with complications, we noted the following:
Of the 10 complications, 6 occurred in patients with poor neurological condition as assessed by Glasgow Outcome Scale (GOS). There were six males and four females. Eight were unilateral (8 of 82, 9.7%) and two were bilateral (bifrontal) procedures (2 of 11, 18%). With respect to CP material, three patients with allogenic CP and seven patients with autologous graft developed complications.
In considering the timing of CP, there were three, four, and three complications noted in the very early, early, and delayed groups, respectively, and this difference was not statistically significant. With respect to etiology of DC, there were five traumatic, three ischemic, and one each in hemorrhagic, vascular, and CVT categories.
On statistical analysis, the only factor that was significant was the neurological condition at the time of CP. A poor GOS prior to CP had a significantly higher rate of complications. There was no difference noted in incidence of complications with respect to gender, CP material, etiology of DC, or laterality of DC. However, it was observed that early groups had slightly fewer complications than delayed group, while bilateral ones had slightly higher complication rates compared with unilateral procedures (this might be because most of the bifrontal ones were open HI). [Table 3] summarizes the results of these statistical tests with respect to the above factors.
DC is regarded as a life-saving procedure to combat refractory raised intracranial pressure. The etiology may be TBI, malignant cerebral infarction, intracranial hemorrhage, CVT, and others.,,,, Usually, the bone flap is preserved in the abdominal wall or deep freezer. CP is a procedure wherein bone defect is replaced with either autologous bone or allogenic materials that include PMMA cement, titanium mesh, ceramic, hydroxyapatite, and others. Usually regarded as a simple procedure, however, CP may result in significant complications, some of which may be major requiring revision surgery or rarely can have life-threatening consequences. The complication rates for CP vary, ranging from 10.5% to 50% with risks of infections alone constituting a significant proportion.,, Other notable complications include seizures, wound dehiscence, hematoma, bone flap resorption, and others.,,,,,, In addition, there is an ongoing debate regarding the optimal timing of CP with most of the earlier reports stating that early CP is associated with increased complications.,, However, there is a trend in recent studies favoring early CP.,, Furthermore, many studies have noted the positive impact of CP on neurological function.,, We intend to present the outcome and complications of cranioplasties performed at our institute over the past 3 years and compare with the available literature.
Klinger et al. analyzed 258 cranioplasties over a 10-year period and noted a 10.8% complication rate in their series. On analysis, male patients, tumor patients, and patients undergoing surgery at the county hospital sustained a statistically higher rate of infection. Among patients with TBI, complex injuries and involvement of the frontal sinus carried a significantly higher infection rate of 17% and 38.5%, respectively. They, however, found no difference in infection between autologous bone and acrylic bone cement.
Piedra et al. analyzed 74 patients who underwent CP over a 10-year period, including only cases of vascular etiology and divided into early (<10 weeks) versus late (>10 weeks) groups. Overall, a 19% complication rate was noted, predominantly infectious ones. They found no difference between age, timing of CP, sex, subgaleal drain, synthetic, or autologous bone in relation to complications. However, it was noted that presence of VP shunt increased infections. They concluded that early CP does not carry an excessively higher complication risk when compared with delayed CP in cases of vascular etiology.
Schuss et al. conducted a retrospective analysis of 280 CP procedures from 1999 to 2011, where CP was performed early (≤2 months) in 19% and late (>2 months) in 81%. The overall complication rate was 16.4%, which included epidural or subdural hematoma (6%), wound healing disturbance (5.7%), abscess (1.4%), hygroma (1.1%), and cerebrospinal fluid fistula (1.1%). Patients undergoing early CP had significantly increased complications compared with late CP (25.9% versus 14.2%). In addition, patients with VP shunt had a significantly higher rate of complications after CP. On multivariate analysis, early CP, presence of a VP shunt, and intracerebral hemorrhage as underlying pathology for DC were significant predictors of postoperative complications after CP. Chang et al. analyzed 212 patients over a 13-year period undergoing cranial repair and noted an overall complication rate was 16.4%. The complication rate for patients 0–39, 40–59, and >60 years of age was 8%, 20%, and 26%, respectively. They also noted that young age (<39 years) and traumatic injury etiology had a lower rate of complications compared with tumor etiology, and patients with autologous bone graft placement had low infection rates.
Chaturvedi et al. analyzed 74 cases of CP performed after DC over a 10-year period including only patients with traumatic etiology. They reported a complication rate of 31%, of which infections constituted 14% and 10% of them required removal of implant. They concluded that females, age >20 years, timing of CP <3 months, and operating time >90 min are associated with higher complication rates. Liang et al. analyzed 88 cranioplasties over a 7-year period and noted a 6.8% complication rate and concluded that there was no difference between drain, timing, and comorbidity in relation to complications. Walcott et al. conducted an 8-year retrospective analysis of 239 CP procedures and noted an overall complication rate of 23.85%, which included surgical site infection, hydrocephalus, and new-onset seizures. Previous reoperation and therapeutic indication for stroke were identified as statistically significant for development of CP infection.
Basheer et al. conducted a retrospective analysis of 114 cranioplasties performed over a 3-year period. Based on the timing of CP, they divided the cohort into three groups: <12 weeks, 12–24 weeks, and >24 weeks. The overall complication rate was 23% and 14% required reoperation. They observed that patients who underwent CP after 24 weeks had more complications compared to the other two groups, although not statistically significant. They did not find any difference in complications or reoperation rate with respect to age, gender, type of DC, type of CP material, or method of autologous bone storage.[Table 4] summarizes the major studies reporting on CP outcomes in world literature.
Timing of CP
Early CP may carry some advantages compared with delayed CP. Proponents of early CP believe that better tissue planes are maintained resulting in easier tissue dissection plane with resultant shorter operative times. Piedra et al. suggested that there may be a lower overall cost if patients undergo CP within the same hospitalization as craniectomy if the risk of malignant intracranial hypertension has passed. Another possible advantage is the decreased time at risk for another injury for patients recovering without an intact bone flap with rehabilitation therapists being overly cautious in their efforts while the patient is without a bone flap. In addition, bone resorption is a potential problem if CP is delayed.,,,,,,,,
Chun and Yi analyzed 30 patients during a 30-month period and concluded that early CP provides satisfactory securing dissection plane during operative procedures compared with later CP, without causing additional complications in selected cases. They also noted that early CP (within 3 months) had the lowest complication rate of 9%. Archavlis and Nievas noted that GOS improvement was significantly higher when CP was done within 7 weeks (78%) and lowest in >13 weeks group (12%). Wachter et al. noted that patients with early CP had significantly better functional outcome than patients with late CP.
On the contrary, a number of studies have shown no difference/worse outcomes with early CP., However, in those studies, the groups were heterogeneous consisting of both traumatic and non-traumatic (ischemic stroke, hemorrhage, tumor) etiology, open and closed HIs, and pediatric and adult cases, which are obviously difficult to compare. The authors favoring late CP believe that risk of infection is increased with early procedure and adequate time needs to be given for brain edema to subside. Matsuno et al. noted that the mean time to CP was significantly shorter in infected cases (74 days) than non-infected cases (108 days). Thavarajah et al. concluded that minimum time for CP should be 6 months to reduce infection. Moreover, one of the major predisposing factors associated with increased complications proposed by Chaturvedi et al., De Bonis et al., and Schuss et al. was an early CP (within 3 months).,, However, a few authors have noted no difference in timing and infections or overall complications.,,,, Honeybul and Ho concluded that there was no correlation between timing of CP and the risk of infection. In a multicenter analysis on CP outcomes in England by Coulter et al., 67 of 166 patients (40.4%) experienced at least one complication during a median follow-up time of 15 months. They noted that timing of CP does not affect outcome. A recent systematic review by Yadla et al. analyzed 18 studies to determine the effect of early surgery (within 3 months of craniectomy) on rates of infection and other complications. They found no difference in infection rates or overall complication rates between early and late surgery or between allograft and autograft cranioplasties. In a recent prospective multicentric study involving 70 patients in five nations, Quah et al. concluded that early CP in non-infection-related craniectomy is relatively safe and there was no significant difference in infection rates or risk of hydrocephalus between the early and late cohorts. In our study, early groups (<12 weeks) had relatively fewer complications than delayed group.
In the following sections, we intend to discuss other possible risk factors for complications as proposed by earlier studies and compare with our observations.
Location of DC
Multiple authors have noted that bifrontal defect is consistently associated with increased complications. De Bonis et al. published their 5-year experience of 218 CP procedures and concluded that patients with a bifrontal CP had a 2-fold increased risk of complications. Gooch et al. noted that the only factor statistically associated with need for reoperation was the presence of a bifrontal cranial defect. In a multicenter analysis on CP outcomes in England by Coulter et al., 67 of 166 patients (40.4%) experienced at least one complication during a median follow-up time of 15 months. They noted that risk was higher with bifrontal defect. In our study, although not significant, CP done for bifrontal defects with allogenic materials was slightly more prone for infectious complications.
Material for CP
A few studies have correlated the type of material used for CP and the postoperative complications. Matsuno et al. analyzed 206 cases in patients age 6 months to 79 years and concluded that autoclaved and autogenous bone grafts and PMMA have a significantly higher rate of graft infection while titanium mesh has the lowest rate of infection. In one review, it was shown that CP with titanium implants provides the lowest rate of complications, reasonable costs, and acceptable postoperative imaging. On the contrary, Liang et al. and Chang et al. noted that autologous bone CP had the lowest complications., We did not find any significant overall difference between autologous and allogenic materials, probably because of fewer patients with the latter type.
GOS at CP with outcome
The general condition and GOS at the time of CP might have a bearing on the complications and outcome. Wachter et al. noted that GOS prior to CP and nicotine abuse were prognostic factors for surgery-associated complications. Similarly, Lee et al. noted that the presence of neurological deficits before CP was significantly predictive for infection. We too are of the opinion that CP needs to be deferred until there is improvement in neurological condition and should not be performed in those having poor GOS with severe neurological deficits because of the risk of wound dehiscence, infection, and other complications in them. In our study, as noted earlier, poor GOS prior to CP was the only statistically significant factor associated with higher complications. Six of 10 patients with complications had a poor GOS at the time of procedure.
With respect to operative duration, it has been generally noted that longer operative duration is an independent risk factor for complications.,, Although a few factors such as age, gender, presence of a subgaleal drain, and type of CP material have been noted to be associated with complications in sporadic reports, a majority of the authors believe that such correlations do not exist.,,,, We too noted that there was no difference in complications in relation to the above factors.
Seizures after CP
Lee et al. noted a 14.5% seizure rate after CP in their analysis of 243 cases and suggested that the use of antiepileptics is advised for seizure prophylaxis if the risk of seizure is more than 10%., Similarly, a high rate of postoperative seizures was documented by Honeybul and Ho (23.6%). We routinely continue antiepileptic medications for at least 4–6 weeks and noted only two cases (2.1%) of seizures in our study.
In all our 93 cases, CP was done within 28 weeks (around 85% done within 12 weeks) and we observed only 10 complications (10.7%) with one mortality (1.1%), which is lesser than most of the studies that quote a complication rate ranging from 15% to 35%.,,,, Bone resorption is a potential problem noted by many authors, especially in pediatric patients. We did not observe any case of bone resorption, probably because of absence of pediatric cases in our cohort. A few studies noting the impact of CP on neurological function have shown that CP promotes improvement in consciousness levels and early CP may hasten recovery in such patients.,,,,,., Honeybul et al. performed functional and neurocognitive assessments including activities of daily living assessment, functional independence measure (FIM) and the Cognitive assessment report (COGNISTAT) on 25 patients, 72 h before and 7 days after CP. They noted a mild overall improvement in the mean FIM score (P = 0.049), mainly due to an improvement in motor function and concluded that a small but significant number of patients appear to improve clinically following CP. It is believed that few patients manifest neurological symptoms as a result of large cranial defects and sunken scalps, the so-called “syndrome of the trephined” and there are reports of reversal of these symptoms after CP. In addition, there is also a possibility of derangement in CSF dynamics and cerebral perfusion in the absence of the skull plate., Although a formal neurocognitive assessment was not done, we generally noted a trend toward better neurological recovery in terms of motor weakness, cognitive function, and speech. A study is ongoing at our center to study these aspects in detail.
Ventriculomegaly is a common occurrence noted after DC that is often responsible for plateauing/progressive decline in neurological function.,, The effect of CP on hydrocephalus is still unclear as is the timing of CSF diversion in such patients. In their study on CP after stroke in 17 patients, Waziri et al. showed that delayed time to CP was linked with development of persistent hydrocephalus. They concluded that early CP might restore normal intracranial pressure dynamics and prevent the need for permanent CSF diversion. On the contrary, some authors have reported that early CP may increase risk of hydrocephalus. In addition, few authors have observed that presence of VP shunt increased infections after CP.,,, Recently, Schuss et al. analyzed the optimal timing for CP and VP shunt after DC and concluded that patients with simultaneous CP and VP shunt had a significant risk for infectious complications when compared with patients with staged procedures. In our series, nine cases had ventriculomegaly that was managed with either a lumbar drain or ventricular tap done intraoperatively for CSF release. This method appears to be safe with no added complications, as suggested by a few authors. In one case, VP shunt insertion was done 2 months after CP with no further complication.
According to us, preserving the autologous bone in the abdominal wall, early CP (within 6–8 weeks) as soon as brain edema subsides, thorough antiseptic scalp preparation starting 2–3 days prior to CP, a minimum of 48 h of postoperative antibiotics, continuation of regular antiepileptics for at least 4–6 weeks, and an abbreviated hospital stay could help reduce the complications of these procedures.
Patients with poor neurological condition at the time of CP had a significantly higher complication rate. There was no difference noted in the rate of complications with respect to gender, CP material, or etiology of DC. However, early CP was associated with relatively fewer complications than delayed group, while bifrontal ones had slightly increased complication rates. Although CP seems to be a straightforward procedure; it may result in significant morbidity and rarely mortality. Strict adherence to a few simple measures may help reduce the complication rates.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3], [Table 4]