A case series of early and late cranioplasty—comparison of surgical outcomes

Cranioplasty is a common neurosurgical procedure performed to reconstruct a skull defect. It is most commonly performed following a decompressive craniectomy (DC), a procedure performed for raised intracranial pressure (ICP), which can occur following traumatic brain injury (TBI), cerebral infarction, subarachnoid haemorrhage, intracerebral haemorrhage, encephalitis and venous sinus thrombosis [14]. Other indications for cranioplasty include following the removal of bone-invading tumours or an infected bone-flap [14].

An increasing number of cranioplasty procedures have been performed in recent years due to the rising popularity of DC to manage raised intracranial pressure following traumatic brain injury and malignant ischaemic stroke [14]. One recent large randomised controlled trial has demonstrated reduced mortality in TBI patients with refractory raised ICP following DC [11]. Several large randomised controlled trials have demonstrated reduced mortality and improved outcomes following DC in malignant ischaemic stroke [1, 9, 13, 24, 29], and decompression is now part of the national guidelines for stroke management [1]. All of this has resulted in a growing cohort of patients who subsequently require a reconstructive cranioplasty.

The aim of a cranioplasty is to restore cosmetic appearances, protect the underlying brain from further injury and facilitate neurological recovery and rehabilitation [21]. A number of mechanisms have been proposed for reversible neurological disability following craniectomy—this includes cerebrospinal fluid flow disruption, venous sinus congestion, abnormal atmospheric pressures which can lead to ‘syndrome of the trephined’ and alterations in cellular metabolism [4, 15]. There is evidence for reduction in cerebral blood flow following DC, demonstrated by perfusion CT and transcranial Doppler ultrasound, which is attributed to the effect of atmospheric pressure on the decompressed area of brain [8]. Consequently, cranioplasty has been shown to resolve the cerebral blood flow and lead to improvement in neurological function [8, 10, 18, 23, 25].

There is a considerable variation in practice in terms of timing of cranioplasty following DC [7, 21, 25]. Traditionally, a cranioplasty would be delayed to allow for cerebral oedema to resolve and for the patient's neurological status to improve, and to reduce the chance of wound infection and delayed hydrocephalus [14]. There is evidence from a number of retrospective studies suggesting that an early cranioplasty procedure may lead to similar or reduced complication rates [2, 3, 6, 12, 20, 22, 28], with an improvement in overall neurological outcome [18, 20, 23, 28]. This has been supported by two systematic reviews [18, 27]. On the contrary, there is also evidence that infection and the development of hydrocephalus are more prevalent when cranioplasty is performed later [5, 17, 19, 26]. Notably, these studies varied in their definition of ‘early’ cranioplasty ranging from 6 weeks to 3 months, however, both of the systematic reviews used 3 months as the divider between early and late. Another study looked at preoperative CT findings and found that the brain sunken ratio was a stronger predictor of postoperative complications [16], suggesting that cranioplasty timing should take into account individual patient factors rather than having a set time for all patients. Overall, there does not seem to be a clear consensus on whether cranioplasty can safely be performed early and whether it may confer additional benefits in terms of neurological recovery.

In our neurosurgical unit, a cranioplasty is traditionally performed at 6–12 months to allow the wound to heal and for the patient to recover from the acute insult. However, the growing evidence in support of an earlier cranioplasty has led to some surgeons electing to perform the cranioplasties at an earlier date. This allows a direct comparison between the groups to assess outcomes and complications.

We present a retrospective study of patients undergoing cranioplasty following DC. We compare the effect of timing of cranioplasty on complications and also demonstrate how an early cranioplasty procedure can be safely performed during the initial hospital admission, with reduction in the overall length of hospital stay.

A retrospective review was performed on all patients who underwent a reconstructive cranioplasty at a single regional neurosurgical unit over a 3-year period between October 2014 and November 2017. Data was collected from the hospital’s electronic database. Individual patient consent and IRB/ethics approval was not required as the study was undertaken as an audit with local institutional approval.

Information analysed included patient age and sex, indication for craniectomy, cranioplasty material and construction method, interval between craniectomy and cranioplasty, length of hospital stay, perioperative complications and time to last follow-up. Patients who underwent a cranioplasty following removal of an infected bone flap after a craniotomy were excluded from the study.

Two groups were identified—those who had received an ‘early’ cranioplasty, defined as less than 3 months from craniectomy, and those who received a ‘late’ cranioplasty, defined as more than 3 months from craniectomy. Perioperative complications were compared across the two groups and statistical analysis was performed using a Fisher’s exact test. Length of hospital stay was calculated for each admission and compared between the groups.

This study provides further evidence that early cranioplasty performed at less than 3 months post-craniectomy does not lead to an increased risk of complications. We chose 3 months as the cut-off between early and late, as this is consistent with the recent major systematic reviews on this topic [17, 18, 26, 27].

Several retrospective studies which also compared early and late cranioplasty have reported similar findings, however one retrospective study found an increased risk of infection in early cranioplasty [5], and two systematic reviews found an increased risk of hydrocephalus in early cranioplasty compared to late [17, 26]. These results are contradicted by a further review which did not find any increase in complications with early cranioplasty [27]. All of the reviews were limited by the retrospective nature of all papers included and differing definitions of the timing of the early and late cranioplasties. This suggests that further high-quality evidence is needed to determine the risk of complications in early and late cranioplasty. Our study did not show a significant increase in the risk of hydrocephalus with early cranioplasty. Hydrocephalus requiring ventricular shunting occurred in 8% on the early group and 8% of the late group in our study.

An additional benefit to an early cranioplasty procedure is the neurological improvement that is associated with cranioplasty. Several studies and a systematic review have demonstrated an improvement in neurological function following cranioplasty [8, 10, 15, 18, 23] and these improvements may be enhanced by an early cranioplasty. This is demonstrated in a systematic review which looked at eight separate studies and found early cranioplasty is associated with greater neurological recovery across all outcome measures [15]. Proposed mechanisms for improvement in neurological function following cranioplasty include the restoration of normal CSF dynamics and normal cerebral blood flow following reconstruction of the skull defect [4, 8, 10, 15, 18, 23]. This raises the tantalising prospect that early cranioplasty can augment the rehabilitation potential of patients, reducing the time required for recovery and improving outcomes. Our study did not assess neurological outcomes comparing the early and late cranioplasty patients, however, this would be a useful area for further research.

Our study has demonstrated that when cranioplasty is performed early, it can occur during the initial admission while the patient is undergoing inpatient rehabilitation. This negates the need for a further admission, which would include a pre-admission assessment, further blood tests and a median hospital stay of 2 days. We have demonstrated that if the cranioplasty is performed early and during the same admission, it will not prolong the inpatient length of stay. We acknowledge that the ability to perform a cranioplasty during a patient’s initial admission while undergoing inpatient rehabilitation depends on how the individual healthcare system is set up; however, it is not unusual for traumatic brain injury patients to have a prolonged inpatient stay due to their rehabilitation needs or while they are awaiting a rehabilitation placement. This study has shown that in cases where the inpatient stay is prolonged due to rehabilitation needs, an early cranioplasty can prevent the need for future readmission without prolonging the length of stay for the initial admission.

The limitations of this study include its retrospective nature and relatively small sample size. All series of this sort are likely to suffer from selection bias given the complexity of rehabilitation following severe TBI. We would envisage that the only mechanism for definitively addressing this question is a randomised controlled trial, notwithstanding the additional resource required to complete this.

We propose that early cranioplasty performed less than 3 months from decompressive cranioplasty can be a safe alternative to delayed cranioplasty and can be performed in the initial admission while the patient is undergoing inpatient rehabilitation. The patients may benefit from enhancement of their neurological rehabilitation following the cranioplasty. Other benefits would include completing their treatment prior to discharge, reducing the risk of further injury to the cranial defect after discharge and improving cosmesis, and preventing the need for a further elective admission with preoperative assessments and a median length of stay of 2 days.