General anesthesia in patients with severe chronic obstructive pulmonary disease (COPD) has many potential risks. Patients with COPD require increased expiratory times to avoid air trapping, reduce barotrauma, and the risk of bullae rupture. Slower ventilatory rates and increased dead space increase their chance of becoming hypercarbic. Furthermore, inhaled anesthetics decrease hypoxic pulmonary vasoconstriction and increase shunting, potentially resulting in hypoxia. The combination of hypoxia and hypercarbia can lead to acidosis and increased pulmonary vascular resistance leading to right ventricular strain. Patients with forced expiratory volume in 1 second/forced vital capacity (FEV1/FVC) ratios <0.5 or baseline PaCO2 over 55 mm Hg are at increased risk of postoperative mechanical ventilation.1 In addition, there is concern that patients with severe lung disease whose tracheas are intubated for surgical procedures may require prolonged mechanical ventilation after surgery.
Informed consent was obtained from the patient and her family for publication of this case report and accompanying images.
Our patient was a frail 62-year-old woman (42 kg, 157 cm) scheduled for right frontal craniotomy for resection of tumor. The patient was extremely anxious and cachectic. Her medical history was significant for a remote 45 pack-year smoking history, severe COPD (receiving nocturnal home oxygen), and a history of non–small-cell carcinoma with right middle lobe resection 6 years prior. After her previous lung resection, her trachea was extubated immediately after surgery, and she was discharged home on postoperative day 3. Seven months before the scheduled surgery, she experienced hypoxic cardiac arrest after becoming overly sedated on a small dose of clonazepam. Complications after her arrest included 10 days of tracheal intubation, stunned myocardium, renal failure, and possibly anoxic brain injury. Five months before the scheduled neurosurgery, she broke her wrist and heel from a fall. Her recovery from the fall was complicated by pneumonia. She presented for surgical consideration because of a symptomatic frontal brain tumor. Her symptoms included bilateral tremor, psychomotor slowing, and decreased fine motor function. Her mass was noted to have considerable surrounding edema.
At the time of her surgery, she had mostly recovered from the complications of her hypoxic arrest and subsequent fall. Her electrocardiogram showed normal sinus rhythm. A preoperative echocardiogram showed fully recovered cardiac function. She had normal biventricular function and no valvular pathology. There was no evidence of right ventricular strain. Her pulmonary function tests were severely abnormal, showing an FVC 49%, FEV1 28%, and FEV1/FVC 56%. Her chest radiograph showed postlobectomy changes and hyperinflation. There was no evidence of infection. Her creatinine was 0.99 mg/dL, demonstrating that her acute kidney injury had resolved. Her medications were Spiriva, citalopram, clonazepam, budesonide, and formoterol.
The patient’s pulmonologist strongly recommended that the craniotomy be performed under moderate sedation without endotracheal intubation. The patient’s family strongly concurred. After detailed discussions with the pulmonologist and neurosurgeon, we decided to proceed with moderate to deep sedation, resorting to general anesthesia with endotracheal intubation only if sedation proved inadequate to perform the surgical procedure. Our sedation plans were influenced by the requirement that her head be away from the anesthesiologist, held fixed in pins (Mayfield head holder; Integra, Plainsboro, NJ) and covered with drapes.
Our primary concern was that the medications usually used for awake craniotomies (remifentanil, propofol, dexmedetomidine) may cause respiratory depression in even healthy neurosurgical candidates. We obtained a bilevel positive airway pressure (BiPAP) machine for the procedure from our respiratory therapy department. In the operating room (OR), we tried fitting various BiPAP masks to the patient (Respironics nasal and small and medium-sized oronasal masks; Philips, Murrysville, PA). However, all the masks tested potentially interfered with either the surgical site or the application of the Mayfield frame holding the patient’s head in place (Fig. 1).
Instead, we placed a standard medium-sized adult mask (Portex; Smiths Medical, St. Paul, MN) on the patient with elastic straps holding it in place. The anesthesia machine was set to pressure support mode (triggered), with a small amount of positive end-expiratory pressure (1–3 cm H2O) and enough support pressure to generate tidal volumes of 300 mL. Her baseline SpO2 on room air was 97% and improved to 100% with supplemental oxygen. We set fresh gas flows at 10 L/min. The patient said she was comfortable with the assisted ventilation. We then started our anesthetic infusions: dexmedetomidine (0.1–0.3 μg/kg/h), remifentanil (0.02–0.05 μg/kg/min), and propofol (15–20 μg/kg/min). The patient continued to trigger the pressure support as she drifted in and out of consciousness. PaO2 on FIO2 of 100% ranged from 584 to 596 mm Hg. She tolerated application of the Mayfield head frame (pins) with local anesthesia provided by the surgeon. With stable ventilation, we turned the operating table 90° away from us, fixing the patient’s head to the OR table, and draping as for a standard craniotomy. A Mayo stand was placed above the patient’s head to provide tenting of the drapes, allowing us to communicate with the patient and periodically confirm proper placement of the facemask. We then reduced her FIO2 to 60%, on which her PaO2 was 310 mm Hg.
We adjusted the pressure support settings on the anesthesia machine throughout the case, to maintain ETCO2 slightly <40 mm Hg. Her baseline PaCO2 was 50 mm Hg. She maintained respirations at a rate of 12 to 18 with tidal volumes approximately 300 to 400 mL. By the end of the 4-hour case, PaCO2 had increased to 54 mm Hg, with pH 7.34.
The patient was instructed to raise her hand to alert the team of any discomfort. Several times, she confirmed that she was comfortable by squeezing a hand. At one point during the case, she reported that she was too warm, and the warming blanket was turned off. At the end of the case, the patient awoke promptly and demonstrated that she was stable in the neurologic examination. She was transferred to the neurologic intensive care unit (ICU), with supplemental oxygen delivered by nasal cannula, in stable condition. She was discharged home on postoperative day 6 after being weaned from daytime supplemental oxygen.
Noninvasive positive pressure ventilation is a well- established mode of airway management for patients experiencing severe COPD exacerbation. BiPAP has been shown to be effective in correcting gas exchange abnormalities, decreasing PaCO2, decreasing respiratory rate, and increasing the ratio of PaO2 to FIO2.2 Perioperative BiPAP has also been described. El-Khatib et al.3 discussed the use of a Respironics BiPAP machine in the OR for preoxygenation in a morbidly obese patient. Similarly, Baillard et al.4 describe the use of an ICU ventilator to deliver noninvasive pressure support ventilation through a mask for the purposes of oxygen administration before intubation of patients with respiratory failure. Yu et al.5 were able to demonstrate decreased shunt fraction and ventilation-perfusion mismatch in patients on a Respironics BiPAP machine under moderate sedation compared with controls receiving positive pressure ventilation under general anesthesia.
Despite a great deal of evidence supporting its use, BiPAP has not come into widespread use in the OR. A few possible explanations could be the inconvenience of bringing a separate BiPAP machine into the OR, a lack of familiarity with the BiPAP machine, the additional expense occurred in setting up the machine, and/or the bulk of the Respironics BiPAP mask. Clinicians in our department are familiar with using the pressure support mode on the Draeger Apollo anesthesia machine (Draeger Apollo, Luebeck, Germany) for laryngeal mask airway use and for emergence from general endotracheal anesthesia. However, none of our colleagues had used this model with a mask for supporting patients under moderate to deep sedation.
The term “BiPAP” is a brand of portable respirator by Respironics, meant to be usable in ICUs, recovery areas, and hospital floors. BiPAP can just as easily be delivered by any anesthesia machine with a pressure support mode. It simply consists of delivery of a preset positive inspiratory pressure that is triggered by the patient’s breath, with a background of continuous pressure support.
Although the ventilation mechanics are nearly the same, there are some important differences between delivering BiPAP through a Respironics machine and the Draeger Apollo. Respironics BiPAP is a single-limb system that blends entrained room air with oxygen. The gas mixture is sent through a humidifier and then to the patient. Positive pressure is delivered when the patient’s respiratory effort is detected at a proximal site. Exhaled gases are prevented from reentering the inspiratory limb by a 1-way valve and are vented to room air through an aperture close to the mask. In contrast, the anesthesia mask in conjunction with a circle system confines exhalation to the anesthesia circuit, with excess volume sent to the scavenging system.
Clinicians who elect to deliver BiPAP using an anesthesia machine need to be aware of the technical differences between types of anesthesia machines, notably Draeger’s piston-driven machines with fresh gas decoupling and the more standard bellows-based machines (GE [Fairfield, CT], Mindray [Shenzhen, China], Penlon [Abingdon, UK], Spacelabs [Snoqualmie, WA], and others).
In a bellows system, compression of the bellows by the driving gas is used to generate pressure during patient inspiration. In pressure control and pressure support mode, the presence of a leak will result in larger bellows excursions, because of the extra inspiratory flow needed to reach and maintain the selected pressure settings. The bellows will also not return to the top of its enclosure during exhalation, because with significant leak, the inspiratory bellows excursion will likely exceed the actual tidal volume forced into the lungs and the volume exhaled back into the circuit. Because pressure support is not a timed breath, but ends based on inspiratory flow decreasing to some fraction of peak inspiratory flow (generally 25%), a large leak may result in the bellows bottoming out (using all of its 1.5 L volume), without end of inspiration ever being triggered. Setting high fresh gas flows on the flowmeters may be used to compensate for some of the leak around the mask in a bellows ventilator because the ventilator pop-off valve is closed during inspiration and all fresh gas flow is forced into the patient circuit. However, this may not satisfy inspiratory flow demands early in inspiration when peak inspiratory flow is maximal.
Leak around the mask in a Draeger piston system would be detected differently, by observing the reservoir bag progressively emptying. The Draeger Apollo uses an internal piston that actively refills using fresh gas flow and the gas returned to the reservoir bag. Thus, even if the piston were visible, the inspiratory excursions would be larger in the presence of a leak, but the piston would still refill completely, using more of the volume in the reservoir bag. Because of fresh gas decoupling, high fresh gas flow would not compensate for leak around the mask. If the piston used all its volume (1.5 L in the Apollo) trying to maintain pressure support during a large leak, it would need to completely refill itself before the next breath. If the refill volume of the piston exceeds the volume in the reservoir bag plus the volume of fresh gas flow during the refill period, the Draeger piston will entrain room air through its negative pressure valve during refill. Thus, the next breath may have a lower FIO2. One way to minimize this risk would be to use a larger reservoir bag (e.g., 3 L) during cases involving BiPAP by mask. We used our standard 2-L reservoir bag during this procedure and had no issue with large leaks.
As with any airway management technique involving supplemental oxygen administered via facemask, the risk of surgical fire should be considered. The risk is increased when the surgical field is in close proximity to an oxygen-enriched environment. For these reasons, surgical fires are most common in nonintubated patients receiving supplemental oxygen delivered through open systems, and tonsillectomies and tracheostomies using an endotracheal tube.6 Neurosurgical cases may be at lower risk for surgical fires because they use specially designed craniotomy drapes that separate the cranium from the lower part of the head and the rest of the body, and thus partially mitigate the enrichment of oxygen concentration at the surgical site. In addition, neurosurgeons generally use fine-tipped bipolar cautery instead of unipolar cautery, with bipolar cautery causing less tissue damage and less sparking.
These factors decrease but cannot eliminate the risk of surgical fire. The Joint Commission on Accreditation of Healthcare Organizations recommends an FIO2 ≤ 30% for open delivery to prevent surgical fire.7 It is our view that BiPAP through the anesthesia machine may decrease the risk of fire in 2 ways. First, the use of BiPAP through the anesthesia machine may serve to limit the FIO2 requirement in patients who may not otherwise tolerate an FIO2 of 30%. Second, provided an adequate seal can be established between the mask and the patient, the use of a circle system should decrease the amount of oxygen in the field. As noted earlier, the Respironics BiPAP device vents exhaled gases to the room air, increasing the oxygen concentration of ambient air trapped under the surgical drapes. Exhaled gases are kept in the anesthesia circuit and wasted to scavenging when BiPAP is administered using an anesthesia machine. It is our view that BiPAP administered through an anesthesia machine may pose less risk for surgical fires than traditional BiPAP through a Respironics device and less risk than supplemental oxygen delivered by nasal cannula or non-rebreathing mask.
We found that a standard anesthesia mask with typical strap system was smaller and less cumbersome than the mask used with the Respironics BiPAP machine. A standard anesthesia mask allowed the neurosurgeons adequate access to perform a frontal craniotomy. We did not encounter any difficulty fitting our mask to our patient. Obviously, if the mask did not fit well, our efforts to provide BiPAP would not have been successful. Our institution stocks masks in various sizes with a port to inflate or deflate the cushion. In our particular case, we were able to establish a very good seal, so we were comfortable using high concentrations of oxygen.
Although true hospital costs are difficult to establish, in our institution BiPAP machines are provided by our respiratory therapy department at a charge of $896 for initial setup and $548 for each subsequent day. When anesthesia provides BiPAP using an existing anesthesia machine, the additional cost of the mask and strap is negligible.
After this case, one of us now routinely uses BiPAP administered by an anesthesia machine using a conventional anesthesia mask for outpatient vascular procedures where the patient has requested no intubation but cannot tolerate the surgical stimulation after the procedure has started. On the basis of this experience, it would seem worthwhile to prospectively evaluate whether intraoperative BiPAP may improve patient outcomes for those patients with severe lung disease for whom moderate to deep sedation may be adequate to perform major surgery.