Pituitary tumors are commonly resected using a minimally invasive, transsphenoidal approach.1 Previous research has shown that patients requiring pituitary tumor resection have an elevated risk of obstructive sleep apnea (OSA), as this disorder is present in up to 70% of patients with acromegaly and 33% of patients with Cushing disease.2 The high prevalence of OSA in this patient population requires a rational understanding of the risks as well as approach to management of these patients postoperatively.
Obstructive sleep apnea syndrome is a complex disorder caused by a narrowing of the upper airway during sleep. The disorder commonly occurs in the presence of tonsillar and adenoid hypertrophy, nasal obstruction, retrognathia and macroglossia.3 The use of continuous positive airway pressure (CPAP) has been proven to improve the apnea hypopnea index reducing the risk of progression to pulmonary hypertension.4 In a previous retrospective review, the incidence of hypoxemia was higher in patients with OSA (26%) compared to non OSA group (5%) after pituitary surgery.5 Although the American Society of Anesthesiologists recommends the use of CPAP postoperatively, the use of CPAP after transsphenoidal surgery creates unique considerations. The literature contains several reports of pneumoncephalus with the use of CPAP in this patient population.6,7 In addition, CPAP may carry a theoretical risk of introducing infection into the intracranial compartment,5 although this risk is poorly documented in the literature. Previous case series and cohort studies have attempted to determine the risks and benefits of the use of CPAP postoperatively following transsphenoidal surgery, although these studies are limited by small sample sizes and few patients in whom CPAP was applied immediately postoperatively.2,5,8
Overall the safety of the use of CPAP following transsphenoidal surgery remains unclear. At our institution, we have used CPAP permissively on patients presenting for transsphenoidal pituitary surgery and offer a unique opportunity to review the potential complications of the use of CPAP. Our study objective was to determine the incidence of complications related to the use of early continuous positive airway pressure following transsphenoidal pituitary resection.
After obtaining institutional ethics board approval from the University of British Columbia (H16-00711), we conducted a retrospective cohort study of patients who underwent transsphenoidal pituitary resection at our institution (Vancouver General Hospital, Vancouver, BC, Canada) between January 1, 2005 and March 24, 2016 (the date of ethics approval).
Patients were identified through our institutional neurosurgical database. Data was extracted from both paper and electronic medical records. First, we identified 2 groups of patients, those with and without OSA. We identified patients with OSA from the surgical or anesthesiologist’s consultation, including both those with suspected OSA documented in the medical record and those with formal sleep study documentation. If the anesthesiologist made note of a formal sleep study or overnight oximetry testing, we considered the patients to have a formal diagnosis of OSA based on an apnea-hypopnea index before the anesthetic consult. If the anesthesiologist documented a clinical suspicion of OSA, typically using the STOP-BANG screening tool9 at our institution, they were categorized as having OSA for the analysis. We also noted whether the patients with OSA used CPAP preoperatively and postoperatively. Second, we followed those who were on CPAP preoperatively and recorded reasons as to why CPAP was discontinued postoperatively, if applicable. Our typical practice is to allow early use of CPAP in all OSA patients, particularly those who were CPAP-compliant preoperatively. Although the type of closure varied depending on tumor size and presence of an intraoperative CSF leak, it did not influence the use of immediate CPAP postoperatively. We collected patient data including age, gender, body mass index (BMI), tumor size and type, date of admission, year of surgery, repeat surgery, and discharge date. Intraoperative data collected included Cormack-Lehane airway grade on laryngoscopy, duration of surgery and type of surgical closure. We also noted whether long-acting opioids were given during the procedure or in the recovery room (eg, morphine or hydromorphone).
Our primary outcome was the incidence of postoperative anesthetic and surgical complications until discharge or 30 days postoperatively, whichever came first. Surgical complications included intraoperative hemorrhage, cerebrospinal fluid (CSF) leak, and pneumocephalus. Charts were screened for clinical evidence of hydrocephalus (clinical symptoms, diagnosis and treatment). Although we do not routinely screen for pneumocephalus, symptoms such headache, visual changes or decreased level of consciousness would typically prompt further assessment and imaging. Anesthetic complications included hypoxia and reintubation. Hypoxia was defined as an oxygen saturation of less than 92% documented in the medical record in the postoperative period until the time of discharge from hospital. A threshold of 92% was chosen both because it has been used previously in the literature10,11 and because it is the threshold for application of supplemental oxygen in our institution.
Descriptive statistics were used to describe the study population (eg, percentage, mean and standard deviation [SD], median and interquartile range [IQR], as appropriate). A P-value <0.05 was considered statistically significant. We first compared patients with and without OSA with respect to baseline characteristics and outcomes. We then described the characteristics and outcomes of patients with OSA who received CPAP to those who not. Data were compared using χ2 test, Fisher exact test, student t test, Wilcoxon rank sum test or analysis of variance, as appropriate depending the distribution, and type of data.
We identified a total of 427 patients who underwent transsphenoidal surgery between January 1, 2005 and March 24, 2016. The median year of surgery was 2012 (IQR, 2009 to 2014). The study population characteristics are summarized in Table 1. OSA was diagnosed or suspected in 64 of the 427 patients (15%). Of these 64 patients with OSA, 23 (36%) were suspected to have OSA with no formal diagnosis. Patients with OSA were more likely to be older, have a higher BMI and have a functional tumor than those without OSA. Specifically, acromegaly was more common amongst patients with OSA (42% vs. 10%; P<0.001) but not Cushing’s disease (19% vs. 16%; P=0.54). Patients with OSA were also more likely to be operated on in the later years of the cohort and have a longer duration of surgery. Patients with and without OSA had similar Cormack-Lehane airway grades (median, 2; IQR, 1 to 2 for both; P=0.14) although patients with acromegaly had overall higher airway grades (median, 2; IQR, 1 to 2 for both; P=0.012). In the acromegaly group 21% (n=10) were documented to have a grade 3 or 4 airway as compared with 6% (n=18) of the nonacromegalic patients. Patients with acromegaly were compliant with CPAP preoperatively more often than those without acromegaly (22% vs. 4%; P<0.0001).
All procedures utilized an endoscopic approach. The type of operative closure was the following: synthetic with or without fat (n=313, 75%); mucosal flap (n=9, 2%); nasoseptal flap (n=84, 20%); left open (n=10, 2%); and bovine pericardium graft (n=2, 0.5%). The type of closure used was similar both in those with and without OSA, and OSA patients with and without CPAP (Tables 1 and 2).
One patient with OSA was reintubated in the postanesthesia recovery room, and none were reintubated in the control group (2 vs. 0%; P=0.017). Hypoxia was more common in the OSA group (14 vs. 5%; P<0.001) although the incidence of intraoperative hemorrhage, CSF leak, and pneumocephalus were similar between the two groups.
Outcomes and Safety of Early CPAP in OSA Patients
Of the 64 patients with a diagnosis of OSA, 29 used CPAP preoperatively and 8 had CPAP applied postoperatively. All 8 of these patients received CPAP immediately in the postanesthesia recovery room and continued CPAP in neurosurgical intensive care unit. Patients who used CPAP preoperatively were more likely to receive CPAP postoperatively (Table 2). Generally, the patient characteristics and outcomes were similar between those with and without CPAP postoperatively. Documented pneumocephalus did not occur in any patients. One patient with OSA was reintubated in the recovery room for hypoxia and subsequently received CPAP postextubation. Although 8 patients in the non-CPAP group experienced hypoxia, none required reintubation.
In total, 21 patients who used CPAP preoperatively had it discontinued postoperatively. The reasons for discontinuing CPAP postoperatively were multifactorial. Two patients were intolerant of CPAP because of discomfort following surgery; in one case the surgeon withheld CPAP because of nasal packing in place; 1 patient refused CPAP because of perceived resolution of symptoms; 1 patient did not have access to their home CPAP in hospital. In 16 cases, the reasons for discontinuation were not documented in the medical record.
Our study results demonstrate that OSA is relatively common in the population presenting for transsphenoidal pituitary surgery (15%). Not surprisingly, our results demonstrate that OSA is particularly common in older patients with a higher BMI as well as those with acromegaly. Overall, patients with OSA undergoing transsphenoidal pituitary surgery experienced a higher incidence of respiratory complications, including reintubation and hypoxia, but similar rate of surgical complications, with or without CPAP. In our limited series of 8 patients, the use of CPAP immediately following surgery in the recovery room did not result in any surgical complications such as pneumocephalus, CSF leak and infection.
The overall incidence of OSA and characteristics of our study population was comparable with previous reports. The overall incidence of OSA (15%) in our study population was also consistent with 2 recent cohort studies in the transsphenoidal surgery population that found an incidence of 21%8 and 22%.5 Although we found the highest rate of OSA in patients with acromegaly, one of these prior studies found a higher rate of OSA with Cushing disease (33%) than acromegaly (18%) at their institution.5 Another study describing only acromegalic patients found a 37.5% incidence of OSA,1 similar to the incidence amongst acromegalics in our study (42%). It is interesting to note that, we did not document higher Cormack-Lehane airway grades in patients with OSA, consistent with a previous report.5 However, we documented overall higher airway grades in those with acromegaly (21% vs. 6%), a rate consistent with prior studies on acromegalic patients.12,13 Similarly, hospital length of stay was similar between those with and without OSA which likely reflects the relative importance of other factors on the hospital stay.
The higher incidence of hypoxemia and reintubation amongst OSA patients in our study was consistent with several prior publications, including in populations undergoing transsphenoidal pituitary surgery.1–3,5
Although several case reports exist in the literature documenting pneumocephalus in those who received CPAP,6,7 our chart review did not find any incidences of this complication in the smaller subset of patients who received CPAP in the immediate postoperative period. Our results are consistent with a recent, similar cohort study that failed to demonstrate an increased rate of pneumocephalus in those with CPAP although, importantly, very few patients in this study received CPAP in the immediate postoperative period (ie, the first 24 to 48 h).8 It is interesting to note that, acromegalic patients were overrepresented in the patients who received CPAP postoperatively; however, this was most likely because of their higher rate of compliance with CPAP preoperatively. As we only had 8 patients who received CPAP, we were underpowered to detect small differences, but our results add to the accumulating evidence supporting the application of CPAP in the immediate postoperative period.
Of the 29 patients using CPAP preoperatively, 21 (72%) discontinued CPAP postoperatively for a variety of reasons. Unfortunately, in a significant number, the reason for discontinuation was not documented in the medical record. We can speculate that patients may not have tolerated CPAP after transsphenoidal surgery because of edema and trauma to the nasal mucosa. It is interesting to note that, another large cohort study of OSA in pituitary surgery found that only 36% had a documented CPAP plan,8 suggesting that this component of postoperative management may be overlooked in the majority of patients.
Our study limitations include a retrospective study design at a single institution, which limits the generalizability of the results. As we relied on the medical record to determine the diagnosis of OSA, and formal testing was not available for many patients, we may have underestimated the overall incidence. The increased incidence of OSA in later years of the study may represent higher awareness of OSA in contemporary practice. We did not routinely screen for pneumocephalus; therefore, we cannot exclude an increase in asymptomatic pneumocephalus associated with CPAP although the clinical significance would be unknown. In addition, relatively few patients received CPAP (n=8), which limits our ability to determine the incidence of relatively rare complications such as pneumocephalus and CSF leak. Similarly, although the type of closure used was similar between those with and without CPAP, we cannot exclude an effect of closure type on the risk of postoperative pneumocephalus with CPAP because of low numbers. As mentioned, the reasons for discontinuation of CPAP in those using the device preoperatively were poorly documented in the medical record. We speculate based on our clinical experience that patients who were poorly tolerant of CPAP preoperatively likely did not continue it postoperatively, particularly those using nasal CPAP given the nature of the surgery. Finally, there was likely selection bias in those who received postoperative CPAP as they may have been higher risk for respiratory complications and lower risk for surgical complications.
Overall, patient with OSA who undergo transsphenoidal pituitary surgery present a conundrum; our cohort suggests these patients experience more respiratory complications that those without OSA but this risk must be balanced with the potential risks of pneumocephalus and CSF leak associated with noninvasive positive pressure ventilation in this population. Our study presents novel data on the use of CPAP immediately following transsphenoidal pituitary surgery, which has not been previously described. Previous cohort studies of OSA patients undergoing transsphenoidal pituitary surgery did not allow CPAP for days to weeks after surgery given the potential risk of pneumocephalus and CSF leak. In contrast, 8 patients in our study received CPAP immediately after surgery in accordance with our institutional policy with similar rates of surgical complications. Our case series supports further investigation into the role of immediate CPAP in high-risk OSA patients.
1. Friedel ME, Johnston DR, Singhal S, et al. Airway management and perioperative concerns in acromegaly
patients undergoing endoscopic transsphenoidal
surgery for pituitary
tumors. Otolaryngol Head Neck Surg. 2013;149:840–844.
2. Venkatraghavan L, Perks A. Postoperative management of obstructive sleep apnea
after transsphenoidal pituitary
surgery. J Neurosurg Anesthesiol. 2009;21:179–180.
3. Mickelson SA, Rosenthal LD, Rock JP, et al. Obstructive sleep apnea
syndrome and acromegaly
. Otolaryngol Head Neck Surg. 1994;111:25–30.
4. Caples SM, Gami AS, Somers VK. Obstructive sleep apnea
. Ann Intern Med. 2005;142:187–197.
5. Rahimi E, Mariappan R, Tharmaradinam S, et al. Perioperative management and complications in patients with obstructive sleep apnea
surgery: Our institutional experience. J Anaesthesiol Clin Pharmacol. 2014;30:351–354.
6. Kopelovich JC, de la Garza GO, Greenlee JD, et al. Pneumocephalus with BiPAP use after transsphenoidal
surgery. J Clin Anesth. 2012;24:415–418.
7. Zlotnik D, Taylor G, Simmoneau A, et al. Two cases of pneumocephalus following noninvasive continuous positive airway ventilation after transsphenoidal
neurosurgery. Ann Fr Anesth Reanim. 2014;33:275–278.
8. White-Dzuro GA, Maynard K, Zuckerman SL, et al. Risk of post-operative pneumocephalus in patients with obstructive sleep apnea
surgery. J Clin Neurosci. 2016;29:25–28.
9. Chung F, Yegneswaran B, Liao P, et al. STOP questionnaire: a tool to screen patients for obstructive sleep apnea
. Anesthesiology. 2008;108:812–821.
10. Pantel H, Hwang J, Brams D, et al. Effect of incentive spirometry on postoperative hypoxemia and pulmonary complications after bariatric surgery: a randomized clinical trial. JAMA Surg. 2017;152:422–428.
11. Russell GB, Graybeal JM. Hypoxemic episodes of patients in a postanesthesia care unit. Chest. 1993;104:899–903.
12. Sharma D, Prabhakar H, Bithal PK, et al. Predicting difficult laryngoscopy in acromegaly
: a comparison of upper lip bite test with modified Mallampati classification. J Neurosurg Anesthesiol. 2010;22:138–143.
13. Lee HC, Kim MK, Kim YH, et al. Radiographic predictors of difficult laryngoscopy in acromegaly
patients. J Neurosurg Anesthesiol. 2019;31:50–56.