Neurological syndromes
Prolonged coma
Posterior fossa syndrome/cerebellar mutism
Dysphagia including lower cranial nerve palsy
Dysarthria
Apnea and other neuro-respiratory disturbances
Vomiting
Ataxia
Limb weakness
Diplopia and other visual disturbances
Hearing loss
Facial weakness
Headache and pain
Cognitive-affective dysfunction
Regional complications
CSF leak
Pseudomeningocele
Hydrocephalus progression
Intracranial hemorrhage and infarction
Cerebellar edema
Meningoencephalitis
Superficial and deep wound infection
Systemic complications
Hemodynamic instability
Systemic infections including urinary and pulmonary infections
Deep vein thrombosis
Salient events and “rescue treatments”
Unplanned return to theater
Unplanned admission to intensive care unit
Prolonged intubation
Gastrostomy requirement
Large volume blood transfusion
Prolonged length of stay in hospital
The definition may include complications due to closely related interventions to surgery, for example, general anesthesia – this may be implied by using the more general term “postsurgical complications.” Clinical events and “rescue treatments” such as unexpected return to theater can also be considered a sentinel feature of a complication having occurred and can be useful in monitoring purposes (see section 62.3).
The complication may be reported as an externally observable clinical syndrome (e.g., prolonged coma, CSF leak, posterior fossa syndrome, and dysphagia) or the pathophysiological etiological mechanism that led to the clinical syndrome. The etiological mechanism may be observable only using investigations such as radiology or detailed examination (e.g., edema, hydrocephalus decompensation, hemorrhage, cranial nerve lesion). These distinct categories of complications – mechanisms and clinical syndromes – are often conflated but should be distinguished. Table 62.2 lists the most common etiological mechanisms along with syndromes they can cause.
Table 62.2
Potential pathophysiological mechanisms underlying neurological complications. Multiple mechanisms are likely to be present in any individual child
Pathophysiological mechanisms | Examples |
|---|---|
Global cerebral ischemic injury | Raised intracranial pressure, systemic hemodynamic instability |
Focal ischemic-contusion injury | Retraction, edema, vasospasm |
Direct tumor-related focal injury | Invasion, hemorrhage |
Direct local injury by surgery | Neurotmesis |
Remote injury mechanisms | Diaschisis, transsynaptic degeneration |
As an example, a clinical syndrome such as dysarthria can arise from multiple causative mechanisms such as direct cranial nerve section (neurotmesis) or parenchymal edema causing transient dysfunction of the brainstem cranial nerve nuclei. These have different implications for avoidance, early management, and prognosis for recovery. Clinical syndromes are also more readily observed than etiological mechanism and are thus more likely to be accurately recorded in some situations. For example, dysphagia as a clinical syndrome is more likely to be recorded as a complication in a questionnaire-based survey or retrospective case note-based review of complication rates than an underlying bulbar cranial nerve lesion which would require formal neurological or speech therapist examination in a prospective study design. A related issue is that clinical symptoms which are not accompanied by observable signs such as pain may be under ascertained, for example, surgical incision site pain, allodynia, and taste disturbance.
Some authors include in their definition that a complication should be an unexpected event [1]. Using this criterion, events are excluded as complications if they are held to be unavoidable and thus are accepted as part of the postoperative course. The “unexpectedness” of an event is however a somewhat subjective assessment and this is likely to lead to further systematic bias. Without a reliable definition of “unexpectedness,” events or problems that may be expected postoperatively but are nevertheless more severe than expected may not be noted as a surgical complication even if they are potentially avoidable. An example is ataxia, which is a typical presenting sign of the tumor, but often unquantified in its severity. The ataxia may be more severe or qualitatively different after surgery, for example, the child may be able to walk with ataxia pre surgery, but unable to even sit independently post surgery. It should also be noted that a postoperative problem simply being expected does not mean it is unavoidable with appropriate refinements in surgical technique, preoperative planning including use of alternative nonsurgical strategies, etc. Accepting the problem as unavoidable a priori runs the danger of becoming a self-fulfilling prophecy and lack of innovative solutions.
The final related issue that impacts on reliability is deciding that the postoperative problem was caused by the surgery, since this is part of the definition of a complication. Causality is usually assigned by “global introspection,” that is, informal and unstructured personal judgment based on the evidence at hand. The evidence for surgery being the cause of a problem noted in a child who is assessed some time later is not always clear. The distinction between an unwanted event following but unrelated to surgery and thus not a complication (e.g., a symptom of the natural progression of the tumor) and one due to surgery itself is not always straightforward. Causation may be inferred by factors such as how specific the clinical syndrome or event is to the postoperative situation. A clinical syndrome or event that is not seen as a presenting sign of the untreated tumor and is not seen following other treatments (e.g., radiotherapy) independent of surgery is more likely to be considered a surgical complication. An example is posterior fossa syndrome which is a qualitatively specific syndrome seen after posterior fossa surgery. On the other hand, where a postoperative syndrome is similar to the presenting features of the untreated tumor or the complications of another treatment, that is, it has low specificity as a surgical complication, it may not be considered to be a complication. An example is persistent postsurgical vomiting; this, in our experience, can be a debilitating problem which is underreported as a surgical complication in the literature. This underrecognition as a complication is presumably because vomiting appears similar to the presenting features of a posterior fossa tumor and vomiting is also a complication of radiotherapy and chemotherapy that the children often receive following surgery.
62.2.2 Methods of Improving Reliability and Validity: Lessons from Medicines Regulation and Adverse Drug Reactions
There are useful lessons to be drawn on how to improve reliability and validity of identifying surgical complications from analogous work in medicines regulation on adverse drug reactions. By analogy, a surgical complication can be regarded as an “adverse surgical reaction.” In European medicines regulation for human drug trials, there is a two-step process to identifying a suspected adverse drug reaction: first identify if an adverse event not necessarily caused by the drug has occurred, and then identify if the adverse event is caused by the drug, that is, an adverse drug reaction. Additional formal definitions are provided from medicines regulatory agencies to decide whether the adverse reaction is serious and unexpected.
Applying these ideas from the medicines regulation to the study of surgical complications, we propose that all adverse events following surgery should be recorded prior to deciding whether or not they are expected or a complication of surgery. A serious adverse event can be defined as one that is life threatening, results in prolongation of hospitalization, or results in significant disability without additional corrective treatment. We also propose that surgical complications should include expected and unexpected adverse events due to surgery. From a practical point of view, expected complications would be those which are listed as common or serious (based on current knowledge) in a patient/parent information sheet provided as part of consent prior to surgery. This corresponds to the definition of an unexpected adverse drug reaction as a reaction that is not described in the product information for the drug.
There are also lessons to be applied from medicines research to improve on the subjective nature of determining causality, that is, whether the adverse postoperative event is caused by surgery. An ongoing research in our institution on adverse drug reactions in children has led to the development of a decision analysis tool in the form of a flowchart with the aim of improving the reliability and validity of the judgment on whether a drug caused an adverse event. We have developed an analogous decision analysis tool for determining causality in adverse events following surgery, the Liverpool Neurosurgical Complication Causality Assessment Tool (Fig. 62.1).
Fig. 62.1
A proposed causality assessment tool to identify whether an event or syndrome identified in a child following brain tumor surgery may be a surgical complication (Reproduced by permission Copyright © 2011 Alder Hey Children’s NHS Foundation Trust). Notes: *Examples: intervening adjunctive treatments (e.g., radiotherapy and chemotherapy) prior to the problem being noted; problem more characteristic of adjunctive treatment than surgery (e.g., hearing deficit in non-cerebellopontine angle tumors).**Examples of evidence: radiology (MRI T2 hyperintensities in areas consistent with neurological deficit), neurophysiology (EMG revealing acute denervation), re-exploration surgery findings
The tool is yet to be validated, and we emphasize that the purpose is not for point-of-care use in the individual child, but rather to improve concordance between individual reviewers of a case and research workers on whether an adverse event is a surgical complication.
The causality tool provides a structured decision process involving the following key questions:
1.
Was the problem present prior to the surgery?
2.
If so, did it worsen or change qualitatively after surgery?
3.
Is the problem specific to the postsurgical situation or can it be caused by other treatments or the tumor itself?
4.
Is there evidence of a basic causal mechanism by which surgery caused the problem in this child?
Depending on yes or no answers, and taking into account that it may not be possible to answer these questions with the patient information available, the case is categorized into one of the following causality groups: unlikely, possible, probable, and definite surgical complication.
Illustrative examples of problems where these questions arise following posterior fossa tumor surgery are hydrocephalus, ataxia, and vomiting. These are very common presenting features of posterior fossa tumors, for example, 71–90 % of posterior fossa tumors have hydrocephalus at presentation [5]; 90 % of children with infratentorial tumors had ataxia [6]. Vomiting can also be caused by adjunctive treatments such as analgesia, radiotherapy, and chemotherapy. Thus, hydrocephalus, ataxia, and vomiting noted after surgery have low specificity for being considered surgical complications, in contrast to problems such as posterior fossa syndrome and CSF leak which are very specific to being surgical complications. However, hydrocephalus, ataxia, and vomiting can worsen following surgery (e.g., ataxia but able to walk pre surgery, ataxia with inability to sit or stand post surgery) indicating that these may be surgical complications. There can also be qualitative changes in syndromes present before surgery, for example, vomiting may be positioning provoked after surgery limiting rehabilitation, rather than the early morning non-positioning-related vomiting that was present prior to surgery. Often these quantitative and qualitative changes in syndromes are not recorded. At best these problems are marked only as being present or absent; thus, there is insufficient information to judge if problems worsened after surgery. In this situation, other supporting evidences must be used to decide if these problems were surgical complications. This could include postoperative CT or MRI scan demonstrating hemorrhage or signal abnormalities consistent with surgical injury to key anatomical structures, for example, deep cerebellar nuclei related to vestibular function and balance.
Some postoperative syndromes will still be difficult to allocate causality to surgery and thus may not be considered a surgical complication even using the causality tool. This is important if modifications to surgical technique could reduce the severity of problems that can cause persistent disability, that is, avoidable problems. This is often the case with syndromes that cause “invisible” or “silent” deficits, that is, problems that are not externally visible and require detailed assessment. An example is higher cognitive function deficit, for example, the cerebellar cognitive-affective syndrome. These silent deficits may not be evident in the early postoperative period and if they have not been assessed in detail prior to surgery. For example, preoperative detailed cognitive assessment is often not possible because of the child’s age or degree of obtundation [2]. There may be intervening treatments after surgery, such as radiotherapy, given before the problem comes to light which obscure judgment on what portion of the silent deficit is specifically due to modifiable surgical factors [3].
Using the systematic methods described above of defining and identifying complications will be essential to improve the current rather poor evidence base on risk factors, described in the following section. A benefit of reconciling the definitions of adverse reactions/complications in medicines research with that in surgery is that it will aid in future comparative effectiveness research studies comparing surgery against pharmaceutical treatments for posterior fossa tumors.
62.3 Complications After Posterior Fossa Tumor Surgery: Risk Factors and Presenting Features
In this section, we discuss the surgical complications (Table 62.1) with a focus on those most specific to posterior fossa tumors in children, their predisposing risks and potential means of mitigation. In Table 62.1, we have used a widely recognized categorization of neurosurgical complications: neurological, regional, and systemic complications. The neurological and regional complications are those most specific to posterior fossa tumor surgery; thus, we do not provide a detailed description of the general risks of surgery and anesthesia although these are of course of major importance, for example, blood loss; obtundation; chest, skin, and urine tract infections in the postoperative period. Many of these systemic complications will account for the additional complication category of “salient events” and “rescue treatments” we have introduced in Table 62.1. “Salient events” include events such as unplanned admission to intensive care for resuscitation and unplanned return to theater within the initial postoperative period to deal with a regional complication such as CSF leak. The concept of a “rescue treatment” is that it is an additional postoperative procedure that was required to prevent an established complication from causing greater injury or death, for example, large volume blood transfusion from intraoperative blood loss in an infant. The rescue treatment may itself lead to iatrogenic harm.
Before discussing the individual complications in detail, we first provide an overview of the relevant predisposing and causative factors. Much of this has been covered in the earlier individual chapters on the different tumor histologies and the chapters on surgical approaches and intraoperative monitoring which have been motivated by the need to reduce surgical complications. The aim is to draw comparisons and contrasts in the specific complications across these areas of work. For the majority of presumptive risk factors, we are not able to provide absolute and relative risk figures for the complications, given the lack of reliable study design. Only a minority of studies are prospective consecutive cohort with predefined operational definitions for the risks and complications and both presurgical and postsurgical assessment. Where the frequency of postoperative complications is stated, this presumably indicates systemic complications and other neurological and regional complications evident in the early postoperative period. Figures of 35–42 % appear to be common for postoperative complications after posterior fossa tumor surgery in children [3, 4].
62.3.1 Risk Factors for Complications
The main categories of risk factors and potential mitigating factors for surgical complications are shown in Fig. 62.2. These factors are not independent of each other, but rather are codependent, for example, surgical approach routes are dependent on the location of the tumor, which is related to the tumor histology.
Fig. 62.2
The broad categories of predisposing risk factors and mitigating factors for surgical complications in posterior fossa tumor surgery
As discussed in the preceding section, unless there has been a detailed preoperative assessment, it is not always clear from reviewing the literature whether a neurological finding after surgery was caused by the tumor itself or by the surgery. Given the interrelated nature of the risks, for example, large tumors distorting the regional anatomy and thus predisposing to further surgical injury, both tumor-related factors and surgical factors will need to be considered together.
The predisposing risk factors for complications are important to consider for a number of reasons:
1.
At a practical level, it allows the neurosurgeon to counsel the child and parent appropriately regarding the nature and likelihood of complications as part of informed consent.
2.
More generally, it allows consideration of whether complications are modifiable, using, for example, mitigating techniques such as preoperative planning and intraoperative monitoring. In Table 62.3, we highlight examples of how risk factors for complications may be modified.
Table 62.3
Examples of associations between specific risk factors and complications, with potential interventions to mitigate or prevent these complications from occurring
Risk factors | Examples of risk complications | Potential mitigating interventions |
|---|---|---|
Baseline presenting features | − | − |
Age | Infants have higher risk of hydrocephalus progression | Preoperative CSF diversion |
Previous treatment (e.g., surgery, radiotherapy) | Infections, wound breakdown, hemorrhage | Achieve definitive surgery first time, intraoperative MRI |
Hydrocephalus | CSF leak | Preoperative CSF diversion |
Length of history | Invasion and extension of tumor, hydrocephalus | Earlier diagnosis |
Tumor location | − | − |
Midline cerebellum and floor of the fourth ventricle | Posterior fossa syndrome/cerebellar mutism, hydrocephalus progression, truncal ataxia, vomiting | Surgical approaches including avoiding retraction |
Cerebellar hemisphere | Appendicular ataxia, focal cognitive deficits | Surgical approach |
Peduncle involvement | Increased risk of ataxia (unilateral) or posterior fossa syndrome (bilateral) | Preoperative and intraoperative monitoring |
Cerebellopontine angle involvement | Hearing loss (unilateral), vertigo, facial palsy, facial sensory disturbance, bulbar palsy (dysphagia) | Intraoperative electrophysiological monitoring |
Brainstem | Bulbar palsy (dysphagia), obstructive and central apneas, peripheral motor weakness | Intraoperative electrophysiological monitoring, limiting goal for extent of resection |
Tumor characteristics | − | − |
Vascularity (e.g., choroid plexus tumors, cystic astrocytoma) | Hemorrhage | Preoperative embolization, chemotherapy, hemostasis |
Size | Potential for retraction injury for large tumors | Retraction techniques, modified surgical approach |
Operative techniques | − | − |
Intracranial access | CSF leak rate may be higher with craniectomy than craniotomy | Craniotomy, address hydrocephalus preoperatively, closure technique |
Extent of resection/goal of resection | Gross total resection requirement (e.g., ependymoma) may predisposes to more neurological sequelae | Multidisciplinary presurgical treatment planning |
Surgical approach | Retraction injury, cerebellar incision with nucleus and tract injury, skull base surgery with lower cranial nerve injury | Preoperative planning and intraoperative monitoring |
3.
Risk factor adjustment (“adjusting for casemix”) is also necessary if surgical complication rates are to be validly monitored and compared across surgeons and provider institutions.
62.3.1.1 Tumor Type, Location, and Surgical Approaches as Risk Factors
The anatomical location of a tumor is closely related to the specific neurological and regional complications likely after surgery, as indicated in Table 62.3. This relates to surgical injury to the neighboring structures during resection such as the eloquent cerebellar cortex, nuclei, white matter tracts, and blood vessels. In terms of the neurological deficits caused by the tumor as opposed to that due to surgery, a slow-growing lesion due to a tumor is likely to have distinctive compensatory plastic alterations in the normal surrounding brain compared to the rapid lesion induced by surgery. The compensatory plastic alterations in the surrounding brain may be a mechanism for the late presentation of large cerebellar tumors with minimal “cerebellar signs,” often only when the accompanying hydrocephalus decompensates.
Both the primary site of the tumor and the anatomical areas of extension also need to be considered for predicting the neurological and regional complications. The primary sites of the tumor can be categorized as the midline cerebellar vermis including floor of the fourth ventricle, cerebellar hemisphere, brainstem including diffuse and focal exophytic, and extrinsic including cerebellopontine angle tumors as well as other specific cranial nerve-related tumors. Extension may occur from the primary site: for example, from the cerebellar hemisphere to the midline or along one of the middle cerebellar peduncles, a midline tumor may cause splaying of bilateral cerebellar peduncles or invade the brainstem. These areas of involvement by extension are additional predictable risks for postoperative deficits. There is a predilection of specific tumor histologies for certain anatomical sites as detailed in earlier chapters, for example, ependymoma predilection for the lateral recess of the fourth ventricle and extension to the cerebellopontine angle and medulloblastoma predilection for the midline cerebellar vermis. Thus, there are associations between certain tumor histologies and specific postoperative complications.
For tumors of the midline cerebellar vermis and floor of the fourth ventricle or extension to these areas from a primary hemispheric site, a common regional complication is progression of hydrocephalus, which is already present prior to surgery in the majority of children. The most characteristic neurological complication for a tumor in this location is posterior fossa syndrome/transient cerebellar mutism [7–9]. The extension of the tumor to involve the middle cerebral peduncles, brainstem invasion, and brainstem compression appear to be additional risk factors. Posterior fossa syndrome (PFS) in its florid form remains an uncommon encountered sequela with approximately 25 % of patients detected in prospective series of varying cerebellar tumors [8, 10]. More frequent neurological complications for midline cerebellar tumors include worsening of gait ataxia and truncal instability, dysarthria including slow speech, dysphagia, prolonged vomiting independent of dysphagia, and gaze disturbances worsening diplopia [3, 4]. These neurological syndromes are likely to be related to injury to the deep cerebellar nuclei and proximal efferent cerebellar pathways in the peduncles to structures in the brainstem and thalami [11, 12]. The injury to the inferior vermis and dentate nuclei is related to worse neurocognitive outcomes [3]. Postoperative facial weakness and dysphagia may be more likely if there is brainstem invasion or adherence. These sequelae may be due to injury to the cranial nerve nuclei beneath the floor of the fourth ventricle (facial colliculus, sixth nerve, and hypoglossal nerves) during resection including the use of ultrasonic aspiration. However, this type of bulbar weakness can be expected to resolve, unlike cranial nerve lesions seen after resection of extrinsic tumors involving the cranial nerves themselves.
Modifications of surgical approaches have been suggested to avoid these neurological complications presumably caused by injury to the deep cerebellar nuclei and inferior vermis. Incision of the vermis, whether superior and inferior, and vermis resection for subvermian and transvermian approaches were previously thought to cause the posterior fossa syndrome (Pollack et al.). However, paravermian incisions and the telovelar approach (incision of the cerebellomedullary fissure) have not prevented these neurological complications including the posterior fossa syndrome occurring completely even if there has been an apparent reduction in frequency [13]. This may relate to the size of the tumor, with larger tumors of the fourth ventricle still requiring retraction causing concussive injury of sensitive structures such as the dentate nuclei and the cerebellar peduncles, which are already distorted by the tumor [14]. Additional attention to reducing retraction-related injury may reduce the neurological complications above. Intraoperative neurophysiological monitoring and navigation techniques are likely to be useful to reduce the risk of neurological complications for tumors involving the floor of the fourth ventricle (see discussion below).
Cerebellar hemisphere tumors appear less likely to produce overt neurological and regional complications. Hydrocephalus progression is less common than for midline tumors unless the tumor is large enough to involve the brainstem. The most characteristic tumor histology at this site is astrocytoma. Pilocytic astrocytoma is usually cystic and readily resectable, predisposing to low risk of additional neurological complications although cystic tumors have higher risk of hemorrhage. Fibrillary astrocytomas may show extension into the ipsilateral cerebellar peduncle. These are higher risk for complications, due to injury to the cerebellar peduncle such as an increase in limb ataxia and tremor. As predictable given the lateralization of function in the cerebellar hemispheres, right cerebellar tumors are associated with language deficits and left cerebellar tumors with visuospatial deficits although these may well be present prior to surgery [2, 15].
Brainstem tumors of the midbrain, pons, and medulla are often associated with hydrocephalus and multiple cranial nerve deficits at presentation. The main consideration is whether it is feasible to perform resective surgery given the high risk of life-threatening complications due to the life-sustaining (autonomic) functions in neighboring structures. These include cardiorespiratory centers as well as bulbar nerve nuclei and descending motor pathways. Intraoperative cardiovascular instability often occurs during surgery to tumors that invade the brainstem and may limit the extent of resection. Focal enhancing and exophytic tumors are typically resectable, whereas diffuse intrinsic gliomas are not due to such concerns. Intraoperative neurophysiological monitoring is likely to have an increasingly important part to play in reducing neurological complications by identifying relevant nuclei and tracts [16]. Neuronavigation, intraoperative MRI, and ultrasound are additional techniques that are likely to further reduce these complications by improving targeting of the tumor and reducing injury to surrounding normal brain structures.
Tumors of the cerebellopontine angle (CPA) in children are typically ependymomas with extension from the lateral recesses of the fourth ventricle or exophytic brainstem tumors. The main issue for tumors in this site in children is adherence of the tumor to the brainstem and exiting lower cranial nerves and thus the potential for cranial nerve section (neurotmesis) of all the bulbar cranial nerves on one side. The need for complete resection is particularly important for ependymoma to prevent recurrence and mortality in this tumor which is so insensitive to other treatment modalities. This leads to significant problems with postoperative dysphagia and risk of aspiration necessitating tracheostomy and gastrostomy [17]. This predilection of ependymomas to involve the cerebellopontine angle and lateral recess of the fourth ventricle may explain the relatively high incidence of tracheostomy and gastrostomy performed (16 and 28 %, respectively) in Morris and colleagues’ series of 96 children with infratentorial ependymoma. Other complications include injury to arterial vessels supplying the brainstem, facial weakness, and sensorineural hearing loss.
Extrinsic tumors of the posterior fossa are uncommon in children. These include vestibular schwannomas involving the facial (seventh) and the vestibulocochlear (eighth) nerves, as well as schwannomas of the ninth nerve around the jugular foramen and fifth nerve schwannomas near the trigeminal ganglion. The potential for additional cranial nerve-specific deficits after surgery is clear, and it is important to identify the degree of deficit prior to surgery. This is most reliably done for hearing and facial weakness, but other important symptoms and signs of facial sensation, lacrimation, taste, dizziness, and tinnitus should also be recorded and the possibility of worsening after surgery discussed [18]. Additional issues may arise from progression of regional problems present prior to surgery including brainstem compression and obstructive hydrocephalus. Specific neurological complications aside from sensorineural hearing loss and facial weakness include nervus intermedius lesion affecting taste, lacrimation dysfunction due to parasympathetic fiber injury, and facial numbness from fifth nerve lesions.
The surgical approach and access route to the brain will depend on factors such as tumor location and nearby critical structures, size of the tumor, and extent of resection required for effective treatment, but will also relate to individual surgeon preference. There may be important differences in complication rates for some approaches compared to others. Suboccipital craniotomy appears to have a lower incidence of postoperative CSF leak than suboccipital craniectomy [19]. Of 110 children with posterior fossa tumors operated on in Gnanalingham and colleagues’ retrospective case series, the CSF leak and pseudomeningocele incidence was 24 % and 23 %, respectively, in those who had craniectomy, compared to 4 % and 9 % in those who had craniotomy. For deep-seated tumors in and around the ventral surface of the brainstem, for example, chordomas of the clivus, in the region of crucial vascular and neural structures, specialized skull base approaches tailored to the individual are required. This includes bone removal to create adequate exposure and less risk of direct and retraction injury to the critical blood vessels and cranial nerves, but with consequently increased risk of CSF leak compared to minimally invasive surgery.
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
