Complication Avoidance in Deep Brain Stimulation Surgery

Death

The reported mortality rate from DBS surgery is low, ranging from 0.3% to 0.9%. Most investigators have reported no mortality at all. Most deaths are related to the lead implant procedure, and the most common cause of death is intracranial hemorrhage. ^, Additional causes of death include cardiac complications, ^, aspiration pneumonia, ^, “multifactorial” causes, ^, or unknown causes. Pulmonary embolus may also prove fatal.

Neurological Deficit

Neurological deficit related to DBS surgery may arise from one of three main causes: a poorly targeted DBS lead that traverses and injures an unintended structure; intracranial hemorrhage; or venous infarction. Of these, intracranial hemorrhage is by far the most common.

Intracerebral Hemorrhage

Although rare, death or disability caused by intracerebral hemorrhage is the most disconcerting potential complication of DBS surgery and is probably the most common reason why patients either are not referred for surgery or refuse to undergo the procedure. In case series with more than 100 DBS procedures, the reported incidence of intracranial hemorrhage (including subarachnoid, subdural, and epidural hemorrhages) ranges from 0% to 9.5% of patients ^{, , , ,} and 0% to 3.3% of implanted leads. ^a

a References 4, 32, 34, 37, 39, 42–44.

The incidence of just intraparenchymal and intraventricular hemorrhage ranges from 0% to 8.4% of patients (mean, 3.2%) ^b

b References 2, 8, 10, 11, 18, 24, 30, 32, 39, 42, 45–49.

and 0% to 3% of leads (mean, 1.7%). ^c

c References 2, 10, 11, 17, 24, 32, 39, 42, 45, 49.

Intraparenchymal hemorrhages accounted for 58% of the intracranial bleeds, and intraventricular hemorrhages accounted for 42%. ^, Among the intraparenchymal hemorrhages, 50% to 77% occurred at or near the target, and 23% to 50% occurred along the implantation trajectory. ^{, ,} Of intracranial hemorrhages, 37.5% to 71.5% were symptomatic, ^{, ,} and up to 50% of patients sustained permanent deficits. Mortality secondary to procedure-related intracerebral hemorrhage ranged from 8.3% to 50%. ^, Intracranial hemorrhage also increased length of hospital stay. ^, In one study the average postoperative stay for patients without hemorrhage was 2.0 days, in contrast to 4.2 days for those with hemorrhage.

Patient-specific risk factors for procedure-related intracranial hemorrhage are poorly defined. Some authors have reported a higher risk of intracerebral hemorrhage in older patients, ^{, ,} whereas others have found no such correlation. ^{, ,} In particular, a retrospective study of 1757 patients revealed that patients older than 75 years exhibited the same risk of intracerebral hemorrhage as did younger patients. Chronic hypertension has also been proposed as a risk factor, ^{, , ,} although some authors have found no statistical relation. Some have found that patients with Parkinson disease have an increased risk of hemorrhage, but others have not. ^{, ,}

Technical factors that may increase hemorrhage risk are also controversial. Some authors have reported that procedures targeting the globus pallidus interna (GPi) are associated with a greater incidence of hemorrhage (up to 7%), ^, whereas others have found no correlation between intracerebral hemorrhage and target. ^, The risks associated with intraoperative microelectrode recording (MER) remain unclear. Theoretically, each electrode pass carries a risk of hemorrhage, and so more passes should confer more risk. Some series have demonstrated this correlation. ^{, , ,} One multicenter study involving 143 patients demonstrated that procedures in which a hemorrhage occurred were performed with a mean of 4.1 ± 2.0 microelectrode penetrations versus 2.9 ± 1.8 penetrations in procedures without a hemorrhage. Park and colleagues reported six intraparenchymal and two intraventricular hemorrhages in 106 DBS procedures during which they employed five simultaneous MER trajectories, in comparison with no hemorrhages in 65 procedures during which they employed only two. In contrast, many other studies have revealed no correlation between MER and hemorrhage, ^d

d References 31, 32, 39, 43, 48, 52, 53.

which suggests that the key factor may not be the use of MER per se, but how MER is employed. Binder and associates suggested advancing microelectrodes no faster than 0.5 mm/s and advancing no more than 2 mm beyond the inferior border of the GPi to avoid blood vessels within the choroidal fissure. Other authors have suggested limiting the number of MER penetrations in patients in whom nonspecific white matter hyperintensities and lacunar infarcts are visible on MR images. To reduce the risk of hemorrhage, Ben-Haim and colleagues proposed a modified microelectrode with a smaller electrode tip and no guide tube covering the electrode’s distal 25 mm.

The surgical trajectory to the therapeutic target may also be important. For example, Elias and associates found that trajectories near or within sulci are associated with a higher rate of vascular complications (10.1% versus 0.7%). The risk of hemorrhage may also be increased with trajectories that traverse the ventricles, and up to 50% of intraventricular hemorrhages may result in focal neurological deficits. ^{, ,} In two studies the risks of hemorrhage associated with transventricular trajectories were 5% and 10.8%. ^, With regard to target-related risk, the risk of ventricular penetration increases when more medial structures are targeted. Elias and associates found that only 6% of trajectories targeted to the GPi traversed the ventricle, in comparison with 45% of ventral intermediate nucleus (VIM) trajectories and 65% of subthalamic nucleus (STN) trajectories. This percentage can go up close to 100% when mesial structures such as the fornix or the anterior nucleus of thalamus are targeted.

In the final analysis, the risk of hemorrhage may be multifactorial; nevertheless, careful attention to surgical planning, judicious use of MER, and cautious avoidance of vascular structures and the lateral ventricles may significantly reduce this risk.

Subdural Hematoma

Subdural hematoma (SDH) is an unusual complication of DBS surgery, occurring in 1% to 1.4% of patients and 0.1% to 4.2% of implanted leads. ^{, ,} The clinical manifestations of SDH usually occur days or weeks after surgery. The hematoma may be discovered incidentally on routine postoperative imaging, or patients may become symptomatic, exhibiting lethargy, confusion, or motor deficit. ^, Patients receiving chronic anticoagulant therapy before surgery may be at greater risk for developing SDH.

Management should be conservative if possible, and, if surgical therapy is needed, a delay is preferred, to allow the hematoma to liquefy. The surgical treatment of choice is bur-hole evacuation. ^, SDH may result in displacement of the DBS lead, negatively affecting stimulation therapy. It may take weeks or months after the hematoma is evacuated for stimulation to be effective again, but lead revision is usually unnecessary. ^,

It is wise to suspend anticoagulant and antiplatelet therapy at least 10 days before the procedure. During surgery, excessive cerebrospinal fluid (CSF) leakage should be avoided so that the bridging veins will not be significantly stretched, especially in patients with significant brain atrophy. ^, We recommend the smallest possible dural opening and sharp corticectomy before guide tube insertion in order to minimize CSF losses and downward traction on the brain, respectively. Parsimonious use of single-track MER will reduce operative time, brain trauma, and CSF losses.

Cerebral Venous Infarction

During trajectory planning, care must be taken to avoid superficial cortical veins, injury to which may result in cerebral venous infarction. The incidence of this is low, occurring in approximately 1.3% of patients and involving 0.8% of electrodes. Venous infarction manifests over the first few hours after surgery, typically with altered mental status and new motor deficits. Once it is identified via imaging, supportive measures are instituted, including airway protection if necessary, head elevation, blood pressure control to prevent secondary hemorrhage, hydration, seizure prophylaxis, and early motor and cognitive rehabilitation. Fortunately, outcome in most of the cases is favorable as neurological deficits are recovered over time, although neurological sequels and even death can occur. ^,

Ischemic Stroke

Transient ischemic attack and ischemic stroke are rare complications of DBS. ^{, ,} The mechanism of these infarctions is unclear, although it has been proposed that high-amplitude stimulation might lead to small-vessel vasospasm. Poorly planned or executed trajectories, which can lead to cortical or deep vascular injuries or vasospasm, may increase this risk. ^, In the senior author’s experience, small infarctions may result from injury to perforating vessels within the lentiform nucleus during procedures targeting the GPi.

Perioperative Confusion

Perioperative confusion or agitation is one of the most commonly encountered complications of DBS surgery, with a reported incidence of up to 22%. ^e

e References 10, 15, 17, 18, 20, 30, 46, 58, 62.

The incidence of postoperative confusion increases with age (particularly in patients older than 70 years ^, ) and may be more common after contemporaneous bilateral electrode implantation. Consequently, this complication is most commonly observed in patients undergoing DBS for Parkinson disease.

Risk factors for postoperative confusion include a history of hallucinations and advanced age. Additional risk factors include complex medical comorbid conditions, more advanced Parkinson disease, the use of dementia-related medications before surgery, and the use of levodopa agonists. In patients with Parkinson disease and selected patients with essential tremor, detailed preoperative neurocognitive assessments may be used to detect patients who may be at risk for postoperative confusion. There is, however, no consensus regarding specific tests that are most useful for this task.

From a surgical standpoint, confusion may be more common after DBS surgery in the STN, especially with placement of the most ventral contact in or close to the substantia nigra pars reticulata. ^, Transventricular trajectories may also increase the risk of postoperative confusion, and it is far easier to avoid larger ventricles when targeting the GPi versus the more medial STN. Increased operative time, the number of microelectrode passes used to localize the intended target, and the introduction of bifrontal subdural air may all contribute to postoperative confusion, but no definitive relationship has been demonstrated.

Postoperative confusion is managed conservatively with mild sedation and physiologic support. Psychosis necessitating intubation is rare, and restraints should be avoided if possible. Fortunately the confusion is usually transient, with a return to baseline function over the course of days to weeks. Discharge to prior level of care is routine. Nevertheless, long-term cognitive sequelae have been described after DBS surgery, ^{, ,} and patients and families should be forewarned of this possibility during the informed consent process.

Venous Air Embolism

Venous air embolism (VAE) is a rare complication, occurring in approximately 1% of DBS procedures. ^, However, if VAE is not recognized and addressed promptly, the results can be catastrophic. Both patient position (head elevated 30 degrees for comfort) and the fact that the patient is awake may contribute to the development of VAE. Normal ventilation produces negative intrathoracic pressure, resulting in a lower venous pressure than with mechanical ventilation with positive end-expiratory pressure (PEEP). The initial symptoms of VAE are persistent cough and a sense of heaviness or discomfort in the chest. A decrease in the end-tidal CO ₂ typically precedes compromised blood oxygen saturation and hypotension. Morbidity and mortality can be significant with VAE; intake of just 50 mL of air can lead to hypotension and dysrhythmias, and more than 300 mL of intravenous air can result in death. Once clinical suspicion is high, the surgeon should flood the surgical field with saline and lower the patient’s head to avoid more air inflow. ^, Bleeding vessels should be cauterized, and the bur-hole margins should be waxed liberally. These measures will usually stabilize the patient so that the procedure can be resumed safely; however, if the cough persists and the end-tidal CO ₂ does not improve rapidly, it is wise to abort the procedure and close the incision. In our experience, patients stabilize rapidly once the incision is closed, and the surgery may be completed at a later date without incident.

Poorly Positioned Electrodes

Despite surgeons’ best efforts, DBS leads occasionally are poorly placed. ^f

f References 2–4, 7, 11, 15, 18, 20, 24, 25, 45, 69–72.

Clinical improvement that is poorer than expected or the induction of unwanted adverse stimulation effects should prompt two questions: (1) Is the device functioning properly? and (2) Is or are the leads well positioned? The first question may be answered in the clinic with a diagnostic assessment of impedances, which, if higher than 2000 ohms in unipolar mode may indicate an open circuit (see “Lead/Extension Wire Fracture” section for more details). The second question is best assessed with MRI. It is crucial that a surgeon be honest with the patient and himself or herself. If the lead is poorly positioned, do not waste the patient’s time with excessive attempts to “program around the problem”; the lead should be repositioned.

Seizure

The risk of seizure with DBS is low. ^{, , ,} Seizures occur during surgery, with a reported incidence of 0.3% to 2.3%, or after surgery, with an occurrence of 0.9% to 9.1%. Almost 90% of seizures occur within the first 48 hours after surgery, and most are generalized tonic-clonic seizures. ^{, ,} Seizures are reported more commonly after DBS than after ablative surgery, and most reported seizures occur after STN implantations. ^,

A specific mechanism underlying DBS-related seizures has not been proposed. The only significant risk factors identified on multivariate analysis are an abnormality on postoperative imaging (i.e., hemorrhage, edema, or ischemia) ^{, ,} and multiple sclerosis —factors that increase the risk of seizures 50-fold and 8-fold, respectively. Seizures are associated with higher rates of morbidity and mortality, so they must be managed actively. Because of the elective nature of DBS surgery, the surgeon must consider aborting a DBS procedure if the patient has a seizure and does not quickly return to his or her neurological baseline. It is also important to perform early postoperative imaging to rule out potential precipitating factors such as hemorrhage or infarction. Patients who experience a single seizure in the first 48 hours postoperatively may be treated with an oral antiepileptic for 1 to 4 weeks. Those who experience delayed or multiple seizures should continue oral treatment for 6 months with electroencephalographic follow-up. ^, There is one case report with persistent status epilepticus lasting for 2 months, but no published reports of patients who developed chronic epilepsy after DBS surgery.

Aborted Procedures

Few aborted procedures have been reported (0.9%–4.9%), but the incidence might be higher. Stated reasons have been an acute cardiac event, ^, agitation or altered mental status, ^{, , ,} respiratory failure, vasovagal syncope, VAE, intraoperative hemiparesis, severe dyskinesia, mechanical failure of the navigation system, head frame misalignment, and brain shift during second lead implantation, among others. Most of these patients were able to undergo a successful operation at a later date after stabilization.

Cerebrospinal Fluid Leak

The incidence of CSF leak after DBS surgery is low (occurring in approximately 1.3% of patients and involving 0.8% of electrodes); however, as in all cases, CSF leak impairs wound healing and may lead to infection. Meticulous wound closure and head elevation are the best preventive measures.