Abstract
Background
Patients with obstructive sleep apnea (OSA) may be at increased risk for serious perioperative complications. The suitability of ambulatory surgery for patients with OSA remains controversial and several national guidelines call for more evidence that assess clinically significant outcomes. In this study, we investigate the association between OSA status (STOP-Bang risk, or previously diagnosed) and short-term outcomes and safety for patients undergoing cancer surgery at a free-standing ambulatory surgery facility.
Methods
We conducted a retrospective analysis of all patients having surgery at the Josie Robertson Surgery Center, a freestanding ambulatory surgery facility of the Memorial Sloan Kettering Cancer Center. Surgeries included more complex “Ambulatory extended recovery” (AXR) procedures for which patients typically stay overnight, such as mastectomy, thyroidectomy, and minimally-invasive hysterectomy, prostatectomy, and nephrectomy, as well as typical outpatient (OP) surgeries. Both univariate and multivariable analyses were used to assess the association between OSA risk and transfer to the main hospital, Urgent Care Center (UCC) visit, and hospital readmission within 30 days postoperatively (primary outcomes) and length of stay (LOS) and discharge time (secondary outcomes). Multivariable models were adjusted for age, ASA score, robotic surgery, and type of anesthesia (general or MAC), and also adjusted for surgery start time for LOS and discharge time outcomes. Chi-squared tests were used to assess the association between OSA risk and respiratory events and device use.
Results
5721 patients were included in the analysis. Of these, 526 (9.2%) patients were diagnosed or at moderate or high-risk for OSA. We found no evidence of a difference in LOS when comparing high-risk or diagnosed OSA patients to low or moderate risk patients whether they underwent OP (p=0.2) or AXR procedures (p=0.3). Though a greater frequency of postoperative respiratory events were reported in high-risk or diagnosed OSA patients compared to moderate risk (p=0.004), the rate of hospital transfer was not significantly different between the groups (risk difference 0.78%, 95% CI −0.43%, 2.0%, p=0.2). On multivariable analysis there was no evidence of increased rate of UCC visits (adjusted risk difference 1.4%, 95% CI −0.68%, 3.4%, p=0.15) or readmissions within 30 days (adjusted risk difference 1.2%, 95% CI −0.40%, 2.8%, p=0.077) when comparing high-risk or diagnosed OSA to low or moderate risk patients. Based on the upper bounds of the confidence intervals, a clinically relevant increase in transfers, readmissions and UCC visits is unlikely.
Conclusions
Our results contribute to the body of evidence supporting that patients with moderate, high-risk, or diagnosed OSA can safely undergo outpatient and advanced ambulatory oncology surgery without increased health care burden of extended stay or hospital admission and avoiding adverse post-operative outcomes. Our results support the adoption of several national OSA guidelines, focusing on pre-operative identification of OSA patients and clinical pathways for perioperative management and postoperative monitoring.
Introduction
Obstructive sleep apnea (OSA) is a significant public health concern affecting approximately 9–43% of the general adult population, with estimates varying by age, gender and other factors.1–5 Repeated upper airway obstruction may result in arterial desaturations, sympathetic activation, systemic inflammation, and cause a variety of cardiovascular and respiratory conditions with increased mortality6,7. OSA prevalence is reported to be even higher in the surgical population than the general population8, with up to 80–90% of surgical patients remaining undiagnosed.4,5 In the perioperative setting, depression of muscle activity of the upper airway following sedation or general anesthesia may exacerbate OSA symptoms, and patients may be at risk for increased perioperative complications, including hypoxemia and airway obstruction, respiratory arrest and death 9–16. Respiratory complications are further influenced by postoperative use of sedatives and opioids, residual muscle weakness, atelectasis, pain, and surgical trauma. OSA patients have been shown to be at increased risk for pulmonary and cardiac complications, more likely to use ventilatory support and intensive care, consume more economic resources, and have longer lengths of hospitalization17,18. Additionally, an increase in malpractice lawsuits related to OSA complications has been reported 19.
To address these clinical concerns, several professional societies published guidelines identifying risks and mitigating strategies for the perioperative setting 5,20–22. The American Society of Anesthesiologists 2014 updated practice guideline concluded that the literature is insufficient to offer guidance on which patients with OSA can be safely managed on an inpatient vs. outpatient basis 21. The Society of Anesthesia and Sleep Medicine provide a detailed review of the updated evidence and focus on preoperative evaluation, and though not specific for ambulatory surgery, establish the importance of properly screening at-risk patients and the benefit of positive airway pressure therapy in the management of these patients 5. The STOP-Bang Tool, developed in 2008 by Chung and colleagues to screen surgical patients for OSA 23, is a validated screening tool that identifies patients at high or moderate risk for OSA and may help reduce risk of perioperative complications 24. In surgical patients, a greater STOP-Bang score is associated with a greater probability of moderate-to-severe OSA.5
The suitability of ambulatory surgery for a patient with OSA remains controversial, and the evidence regarding safety is limited. Careful preoperative screening of patients to identify those at risk for OSA is an important first step to improving care. The Society for Ambulatory Anesthesia (SAMBA) issued a consensus statement supporting that patients with known or presumed diagnosis of OSA, with optimized comorbid conditions and whose postoperative pain can be managed predominantly with nonopioid analgesics, can be considered for ambulatory surgery 20. They call for more evidence that assesses clinically significant outcomes, (e.g. delayed discharge, unanticipated hospital admission, readmission, or serious morbidity and mortality) rather than surrogate outcomes (e.g. desaturation, hypoxemia, supplemental oxygen). Most importantly the impact of these recommendations on clinical perioperative outcomes in ambulatory surgery is unknown.
The Josie Robertson Surgery Center (JRSC), of the Memorial Sloan-Kettering Cancer Center (MSKCC), is a free-standing surgical facility dedicated to outpatient procedures in cancer patients. It is unique in performing more advanced, non-traditional outpatient cancer procedures compared to most free-standing surgical centers, in part due to its overnight stay capability. Positive airway pressure support (PAP) can be provided by day and evening shift respiratory therapists. Since opening in 2016 JRSC has accumulated a large body of data and outcomes that offer an opportunity to provide additional evidence to evaluate consensus guidelines and examine the safety and suitability of ambulatory surgery for at-risk OSA patients.
In this study, we investigate the association between OSA status (STOP-Bang low, moderate, or high-risk, or previous OSA diagnosis) and short-term outcomes and safety for patients undergoing a variety of cancer surgery procedures in a free-standing ambulatory surgery facility.
Methods
After obtaining approval for the study from the MSKCC Institutional Review Board (IRB) and an IRB waiver on the requirement for written consent, we identified all patients that underwent surgical procedures at JRSC between January 1 and December 31, 2016. If patients had multiple procedures during the study period, the first procedure was included in the analysis and any subsequent procedures were excluded. Procedures were also excluded if the patient had an ASA score of 4 or the procedure used local anesthesia only. Data on patient age, gender, BMI, ASA score, anesthetic technique, type of surgical procedure, operative time, length of stay (LOS), transfer to main hospital or other acute care facility, and subsequent urgent care center (UCC) visits and hospital admissions or readmissions within 30 days were documented as part of routine care. OSA status was defined using the STOP BANG score, with scores <3 categorized as low risk of OSA, scores of 3–4 as moderate risk of OSA, and scores ≥5 as high-risk (Figure 1).
Figure 1.
STOP-Bang Scoring Tool. Patients are first asked the four STOP questions. If they receive a score of 2 or more, they are then asked the Bang Questions. The total score is used to determine obstructive sleep apnea risk. Adapted from Chung F, Abdullah HR, Liao P. STOP-Bang Questionnaire: a practical approach to screen for obstructive sleep apnea. Chest 2016;149(3):631–8.
We have defined ambulatory extended recovery (AXR) procedures as those that are more complex than typical outpatient (OP) surgeries and are scheduled to have a single overnight stay while still being considered as ambulatory surgery cases from a regulatory standpoint. These surgeries include mastectomy (unilateral, bilateral, and with or without immediate reconstruction); thyroidectomy and parotidectomy; and minimally-invasive hysterectomy, prostatectomy, and nephrectomy. To facilitate the performance of more complex cancer surgeries in the ambulatory setting, care, including patient and procedure selection, education and expectation setting, as well as intraoperative and postoperative management, is highly protocolized. We developed specific ERAS (enhanced recovery after surgery) protocols for anesthesia care for these AXR patients which include opioid-sparing approaches, including regional blocks and multimodal analgesia as a component of general anesthesia, but the specific choice of anesthetic is left to the anesthesia team’s discretion. There is no specific intraoperative anesthesia OSA protocol.
Data are presented for both AXR and traditional OP procedures. Post-operative length of stay (LOS) in hours was defined as time from entry into the post-anesthesia care unit (PACU) to discharge home for outpatient procedures, and did not include patients who had a reoperation or were transferred to the main hospital. Discharge time was defined as hours and minutes since midnight on the day of discharge, and we excluded patients who had a reoperation, were transferred to the main hospital, or did not stay overnight. All surgery patients wear a Real-Time Location System (RTLS) badge (Versus Technology Inc., Traverse City, MI). The badge is detected by sensors placed throughout the facility approximately 10 feet apart and provides continuous location updates to our facility visualization maps and database. The badge is collected from the patient at the time of discharge.
Patient identification and OSA Risk
All patients undergoing surgical procedures at JRSC are screened pre-operatively for OSA by a nurse practitioner within 1 month of the date of the surgery. Patients are asked whether they have been diagnosed with OSA, and if so, whether they own equipment, such as a CPAP machine or a mouth guard, and whether they use the equipment. If a patient has not been diagnosed with OSA, they are screened using the STOP-Bang tool. First, they are asked the four STOP questions (Figure 1). If they answer “No” to more than two questions, they are considered at low risk for OSA. If they answer “Yes” to two or more questions, the BANG questions are applied. As some patients with OSA go undiagnosed, patients identified as high-risk as well as those with previously diagnosed OSA are distinguished on the OR schedule with a custom icon to alert of this risk. This information is available to perioperative staff so preop nurses inquire about their PAP devices, anesthesia staff are aware of associated anesthesia risks, and respiratory therapists conduct postoperative assessment and monitoring. Because STOP-BANG also includes a moderate OSA risk category, we retrospectively stratified the high-risk into moderate or high based on individual chart review. For the low risk, on whom the BANG score was not initially applied, we did data checks based on the patient’s BMI, age and gender and used those along with STOP score to confirm that there were no false negatives and did not miss any patients who would have been high-risk. The scoring system outlined in Figure 1 is then used to determine if the patient is at low, moderate, or high-risk for OSA.
Primary and Secondary Outcomes
Our primary outcome was safety as defined by three outcomes: transfer from JRSC after surgery, any MSKCC urgent care center visit within 30 days after surgery, and any readmission within 30 days after surgery. As a secondary outcome, we assessed LOS for patients undergoing outpatient procedures and discharge time for patients undergoing AXR procedures.
Respiratory Outcomes
All patients who are diagnosed with OSA, or screened as moderate or high-risk for OSA, are identified preoperatively and assessed by a respiratory therapist post-operatively. The respiratory therapist makes note of any post-operative respiratory events, such as repeated desaturations <90% SpO2 in an unstimulated environment or obstruction (apnea or snoring) lasting 20 seconds, and records the need for CPAP, BiPAP, or continued mechanical ventilation.
All patients diagnosed with OSA who own a home device are encouraged to bring their device with them on the day of the surgery. The device is examined by the Biomedical Engineering Department for electrical integrity, and if approved, the patient is encouraged to use their own machine in the post-operative period. If the device is not approved, or if the patient does not bring his or her device, a PAP device is provided by the facility for these patients. For patients who do not use a machine at home or who did not bring their device, the respiratory therapist monitors them and may place them on a PAP device if they experience post-operative respiratory events. We recorded the use of PAP devices on at-risk or diagnosed patients, regardless of whether they used home equipment or not.
Statistical Analysis
For this analysis, OSA risk was categorized as low or moderate risk of OSA vs. high-risk or diagnosed OSA, as high risk and diagnosed patients are identified to clinicians before surgery to alert them of possible increased risk. As a sensitivity analysis, we repeated all analyses comparing low risk patients to moderate risk, high-risk and diagnosed OSA patients. Wilcoxon rank sum tests were used to assess the association between OSA risk and LOS for outpatient procedures and between OSA risk and discharge time for AXR procedures. We also planned a multivariable linear regression model, adjusted for age, ASA score, robotic surgery, type of anesthesia (general or MAC) and surgery start time, to assess the impact of OSA risk on length of stay after controlling for factors known to be associated with LOS or discharge time.
We assessed whether there was an association between OSA risk and the probability of transfer to the main hospital after surgery, UCC visit within 30 days after surgery, or hospital readmission within 30 days after surgery using chi-squared tests. We then tested whether the association remained significant after adjusting for age, ASA score, robotic surgery, type of anesthesia, and procedure class (OP or AXR) using multivariable logistic regression and reported the adjusted risk difference as the adjusted risk in the lower OSA risk group (low or moderate risk) subtracted from the adjusted risk in the higher OSA risk group (high risk or diagnosed OSA). We also reported the absolute difference in risk of post-operative death within 30 days between patients at low or moderate risk of OSA and those at high risk or diagnosed with OSA, with 95% confidence interval. We did not have sufficient event numbers to perform a multivariable analysis for the outcome of death within 30 days of surgery.
We assessed whether there was a difference in intraoperative or postoperative opioid use in oral morphine milligram equivalents (MMEs) based on OSA status. Univariate and multivariable logistic regression models were used with the outcome of any vs no intraoperative or postoperative opioid use. We then assessed the association while adjusting for age, gender, BMI, surgical service, and procedure class. Patients with a reoperation on the same day were excluded from the comparison of post-operative opioid use, as post-operative opioid use included opioid use after subsequent surgeries on the same day.
Among patients diagnosed with or at moderate or high-risk for OSA, we reported the incidence of post-operative respiratory events with 95% confidence interval. We also investigated whether the incidence of post-operative respiratory events or the use of a post-operative respiratory device was associated with risk of or diagnosis with OSA using chi-squared tests. P values < 0.05 were considered statistically significant. All analyses were performed using Stata 15 (StataCorp, College Station, TX). This manuscript adheres to the applicable STROBE reporting guidelines.
We estimated that this study would include close to 6000 patients, and that the prevalence of OSA or high risk for OSA would be 10%. For an adverse outcome with a prevalence close to 5%, this would give a 95% confidence interval narrower than ± 2%, which was deemed to be sufficient precision. We did not specify clinical relevance a priori in terms of a non-inferiority margin and this limits the strength of our conclusions.
Results
We identified a cohort of 5,731 patients that underwent 6,522 surgical procedures at JRSC during the study period. We excluded 791 procedures from patients who had a second procedure as part of a planned breast reconstruction, 9 patients classified as ASA 4, and one patient who received only local anesthesia. A total of 5721 surgical patients were included in the analysis. In this cohort, 526 (9.2%) were at risk for OSA and included 233 (4.1%) patients previously diagnosed, 91 (1.6%) at moderate risk, or 202 (3.5%) high-risk for OSA. Other patient characteristics are presented in Table 1.
Table 1.
Patient characteristics, N=5721. Data are presented as number (%) or median (quartiles).
| Characteristic | Low/Moderate Risk (N=5286) | High-risk/Diagnosed OSA (N=435) |
|---|---|---|
| Male | 719 (14%) | 180 (41%) |
| Age (years) | 55 (45, 64) | 61 (54, 67) |
| Procedure Class | ||
| Outpatient | 3301 (62%) | 194 (45%) |
| AXR | 1985 (38%) | 241 (55%) |
| BMI | 26 (23, 30) | 34 (29, 38) |
| ASA score | ||
| 1 | 211 (4.0%) | 1 (0.23%) |
| 2 | 2985 (56%) | 103 (24%) |
| 3 | 2090 (40%) | 331 (76%) |
| Type of anesthesia | ||
| General | 3458 (65%) | 353 (81%) |
| MAC | 1828 (35%) | 82 (19%) |
| Robotic-assisted procedure | ||
| No | 4522 (86%) | 300 (69%) |
| Yes | 764 (14%) | 135 (31%) |
| Service | ||
| Breast | 2762 (52%) | 162 (37%) |
| Gastric mixed tumor | 1 (<0.1%) | 0 (0%) |
| Gynecology | 816 (15%) | 45 (10%) |
| Head and neck | 412 (7.8%) | 46 (11%) |
| Plastics | 733 (14%) | 36 (8.3%) |
| Urology | 562 (11%) | 146 (34%) |
| Length of surgery, in minutes | 78 (43, 143) | 118 (51, 219) |
| OSA Risk | ||
| Low Risk (STOP BANG 0–2) | 5195 (98%) | 0 (0%) |
| Moderate Risk (STOP BANG 3–4) | 91 (1.7%) | 0 (0%) |
| High-risk (STOP BANG 5–8) | 0 (0%) | 202 (46%) |
| Diagnosed | 0 (0%) | 233 (54%) |
AXR= Ambulatory Extended Recovery
Outcomes by OSA risk group are presented in Table 2. On univariate analysis, we found no evidence of a difference in LOS by OSA risk among patients undergoing outpatient procedures (p=0.2) or AXR procedures (p=0.3, Table 3). We found no evidence that patients at high-risk or diagnosed with OSA were more likely to be transferred to the main hospital after surgery (p=0.2, Table 3). Patients at high-risk for OSA or diagnosed with OSA were more likely to have a UCC visit within 30 days of surgery (p=0.027) and to be readmitted to the hospital within 30 days (p=0.047, Table 3). As a sensitivity analysis, we repeated these analyses comparing low risk to moderate risk, high-risk and diagnosed OSA patients. The results of this sensitivity analysis had similar effect sizes and were consistent with the results of the main analysis (data not shown). There were 51 (0.9%) patients transferred to the main hospital after surgery at JRSC. Top reasons for transfer were surgical bleeding (29%), other surgical issues (24%), and cardiac or neurological events (20%). Three patients (5.9%) were transferred for pulmonary events and only one of these was diagnosed with OSA.
Table 2.
Post-operative outcomes by OSA status. Data are presented as number (%) or median (quartiles).
| Low Risk (N=5195) | Moderate Risk (N=91) | High-risk (N=202) | Diagnosed OSA (N=233) | |
|---|---|---|---|---|
| Length of stay for outpatient procedures, in hours | 2.5 (2.0, 3.5) | 2.3 (2.0, 3.3) | 2.7 (2.1, 3.5) | 2.6 (2.1, 3.5) |
| Discharge time for AXR procedures, in hours | 10:42 am (9:45 am, 11:43 am) | 10:40 am (9:48 am, 11:33 am) | 10:55 am (9:35 am, 12:14 pm) | 10:49 am (9:50 am, 11:56 am) |
| Transfer to main hospital after surgery | 43 (0.8%) | 1 (1.1%) | 2 (1.0%) | 5 (2.1%) |
| UCC visit within 30 days | 248 (4.8%) | 1 (1.1%) | 16 (7.9%) | 17 (7.3%) |
| Readmission within 30 days | 106 (2.0%) | 1 (1.1%) | 7 (3.5%) | 10 (4.3%) |
| Any intraoperative opioid use | 5079 (98%) | 88 (97%) | 199 (99%) | 228 (98%) |
| Any postoperative opioid use | 3055 (59%) | 53 (58%rea) | 125 (62%) | 146 (63%) |
AXR= Ambulatory Extended Recovery
UCC= Urgent Care Center
Table 3.
Univariate post-operative outcomes and intraoperative and postoperative opioid use and difference in post-operative outcomes and opioid use by OSA risk group. A positive difference indicates a higher rate or mean in the high-risk and diagnosed OSA group.
| Low or moderate risk of OSA (N=5286) | High-risk or diagnosed OSA (N=435) | Difference | 95% CI | p value | |
|---|---|---|---|---|---|
| Length of stay for outpatient procedures, in hoursa | 3.2 (2.8) | 3.5 (3.3) | 0.31 | −0.11, 0.72 | 0.2 |
| Discharge time for AXR procedures, in hoursa | 11 (1.7) | 11 (1.9) | 0.20 | −0.05, 0.44 | 0.3 |
| Transfer to main hospital after surgeryb | 44 (0.8%) | 7 (1.6%) | 0.78% | −0.43%, 2.0% | 0.2 |
| UCC visit within 30 daysb | 249 (4.7%) | 33 (7.6%) | 2.9% | 0.32%, 5.4% | 0.027 |
| Readmission within 30 days | 107 (2.0%) | 17 (3.9%) | 1.9% | 0.02%, 3.7% | 0.047 |
| Any intraoperative opioid useb | 5167 (98%) | 427 (99%) | 0.46% | −0.70%, 1.6% | 0.5 |
| Any postoperative opioid useb | 3108 (59%) | 271 (63%) | 3.5% | −1.2%, 8.3% | 0.15 |
Means (SD) and mean differences presented and Wilcoxon rank sum tests used for length of stay and discharge time.
Number (%) of events and risk differences presented and chi-squared tests used for transfer, UCC visits, readmission and any use of intraoperative or postoperative opioids.
AXR= Ambulatory Extended Recovery
UCC= Urgent Care Center
On multivariable analysis, we found no evidence of a difference in LOS based on OSA risk for patients undergoing outpatient procedures (p=0.8, Table 4). While there was a significant difference in discharge time in AXR procedures, the discharge time for high-risk and diagnosed OSA patients was only 15 minutes later (95% CI 0.42 minutes, 30 minutes, p=0.044). We also saw no evidence of a difference in transfer rates or UCC visits (Table 4). While there was some evidence that readmission rates within 30 days were higher among diagnosed and high-risk patients than among low or moderate risk patients (adjusted risk difference 1.2%, 95% CI −0.40%, 2.8%, p=0.077, Table 4), the confidence interval does not include clinically meaningful differences. Again, results of the sensitivity analysis comparing low risk to moderate risk, high-risk or diagnosed patients were consistent with the primary analysis. One patient who was at low risk of OSA died within 30 days of surgery, and there were no other deaths within 30 days of surgery in this cohort. The risk of post-operative death within 30 days was 0.02% higher in low or moderate OSA risk patients than in high risk or diagnosed OSA patients (95% CI −0.02%, 0.06%), with the 95% confidence interval excluding a greater than 0.02% higher risk of 30-day postoperative death in high risk or diagnosed OSA patients.
Table 4.
Multivariable post-operative outcomes and intraoperative and postoperative opioid use and difference in post-operative outcomes and opioid use by OSA risk group. The adjusted means/rates in each group and the adjusted difference in means/rates are presented. A positive difference indicates a higher adjusted rate or higher adjusted mean in the high-risk and diagnosed OSA group.
| Low or moderate risk of OSA (N=5286) | High-risk or diagnosed OSA (N=435) | Adjusted Difference | 95% CI | p value | |
|---|---|---|---|---|---|
| Length of stay for outpatient procedures, in hoursa | 3.24 | 3.30 | 0.063 | −0.33, 0.46 | 0.8 |
| Discharge time for AXR procedures, in hoursa | 10.9 | 11.1 | 0.25 | 0.01, 0.50 | 0.044 |
| Transfer to main hospital after surgery b | 0.47% | 0.60% | 0.12% | −0.37%, 0.61% | 0.6 |
| UCC visit within 30 days b | 4.2% | 5.6% | 1.4% | −0.68%, 3.4% | 0.15 |
| Readmission within 30 days b | 1.9% | 3.1% | 1.2% | −0.40%, 2.8% | 0.077 |
| Any intraoperative opioid usec | 99% | 98% | −0.60% | −2.1%, 0.87% | 0.3 |
| Any postoperative opioid usec | 61% | 60% | −1.3% | −6.9%, 4.3% | 0.6 |
Mean differences in length of stay and discharge time were adjusted for age, ASA, robotic surgery, type of anesthesia and surgery start time.
For transfer to the main hospital and UCC visit or readmission within 30 days, risk differences were adjusted for age, ASA, robotic surgery, type of anesthesia and procedure class.
Risk differences in use of any intraoperative or postoperative opioid use were adjusted for age, gender, BMI, surgical service and procedure class.
AXR= Ambulatory Extended Recovery
UCC= Urgent Care Center
We found no evidence of a difference in any use of intraoperative opioids or postoperative opioids on univariate (p=0.5 and p=0.2, respectively, Table 3) or multivariable analysis (p=0.3 and p=0.6, respectively, Table 4). A sensitivity analysis comparing low risk to moderate risk, high-risk and diagnosed OSA showed results consistent with the main analysis.
While there was some evidence on univariate analysis that rates of adverse safety events were higher in high-risk and diagnosed OSA patients, there were not clinically important differences seen on multivariable analysis based on the upper bounds of the confidence intervals. OSA risk does not add predictive value in addition to other patient characteristics that differed by OSA risk, such as type of procedure and ASA score.
Among the 526 patients diagnosed with OSA or at moderate or high-risk for OSA, 13% (95% CI 10%, 16%) experienced a post-operative respiratory event. When comparing moderate risk, high risk and diagnosed OSA patients separately, we found no evidence of a difference in the rate of post-operative respiratory events between high risk patients (N=30 events, 15%) and diagnosed OSA patients (N=36 events, 15%, p=0.9). The rate in the combined high risk and diagnosed OSA group was significantly higher than the rate of post-operative respiratory events in moderate risk patients (N=2 events, 2.2%, p=0.001). Nearly half of patients with diagnosed OSA used a post-operative respiratory device after surgery (N=113, 49%), as compared to only 13% (N=27) of high-risk and 2.2% (N=2) of moderate risk patients (p<0.0001). Notably, 60% (n= 94) of patients diagnosed with OSA who use a home device (n=156) chose to use their device or a hospital device post-operatively, regardless of the occurrence of post-operative events. Among the 77 (33%) patients diagnosed with OSA who did not use a home device, 19 (25%) were placed on a PAP after experiencing a post-operative OSA event.
Discussion
We investigated the association between OSA status and short-term outcomes and safety for patients undergoing a variety of OP and more complex AXR cancer surgery procedures in a free-standing ambulatory surgery facility. After multivariable adjustment, there was no statistically significant association between the risk of OSA and length of stay, urgent care visits, readmission or risk of transfer for either OP or AXR procedures. Based on the upper bounds of the confidence intervals, clinically significant increase in the risk of these adverse events associated with OSA is unlikely.
We did find an increase in postoperative respiratory events among high-risk and diagnosed OSA patients, but these events did not delay discharge or increase postoperative transfers. Although we found 49% of diagnosed patients used a post-opererative respiratory device, this measure is not reflective of the occurrence of postoperative events that occurred since all patients diagnosed with OSA who use a home device are encouraged to bring and use their home device post-operatively, regardless of the occurrence of post-operative events, whereas all other patients are placed on a PAP device only if they experience a post-operative event. Memtsoudis et al. commented in a recent editorial that the increasing prevalence of OSA may lead facilities to develop protocols based on insufficient scientific evidence that could increase resource utilization for implementation25. Our data demonstrate the feasibility of managing at-risk OSA patients without increasing the burden of extended hospitalization or readmission.
These outcomes can be considered to reflect the impact of applied consensus and evidenced-based guidelines advanced by professional societies 5,20 in a real-world clinical setting. Our preoperative process includes STOP-Bang screening of all scheduled patients and assessing the optimization of comorbid conditions. In addition, all patients who were scheduled as AXR were on clinical pathways that include multimodal analgesic therapies including limiting, but not eliminating, opioids. Our results provide additional evidence for following current guidelines and maintaining safe use of opioids in the at-risk OSA population. Although we did not measure whether patients who received opioids had greater frequency of respiratory events, no patients with OSA were transferred because of prolonged hypoxemia or respiratory complications.
Our findings confirm those of Stierer et al. who found that, among 2139 ambulatory surgery patients, those with diagnosed or at high-risk for OSA did not have an increased risk of unplanned hospital admission, life-threatening events such as re-intubation or cardiac arrhythmia, or death26, although they did report an increase in perioperative events such as difficulty of intubation resulting in additional anesthetic management. Notably, they used a questionnaire and demographic data to identify patients, whereas we systematically screened all patients using the STOP-Bang tool. Our larger study population, including more complex ambulatory surgery cases, extends Stierer et al.’s conclusion that patients with OSA may undergo ambulatory surgery without increased risk of major adverse outcome, although they may require additional perioperative interventions.
In our study sample, 4.1% were diagnosed with OSA and 3.5% screened high-risk for OSA compared to higher estimates in the adult population1–5. The lower prevalence of OSA among our population is likely due in part to the lower prevalence of males (16%) in our sample, as male gender is a risk factor for OSA. Indeed, among our male population, 20% screened as high-risk or had an OSA diagnosis, closer to general population estimates for males.1 Our at-risk patients had higher BMI, more patients scored as ASA 3, greater use of general anesthesia, longer duration of surgery, and increased proportion undergoing advanced robotic ambulatory surgery. Nevertheless, with proper preoperative optimization and perioperative management, these at-risk patients did not experience an increase in LOS or adverse outcomes.
A 2016 study conducted among 404 ambulatory surgical oncology outpatients undergoing low risk cystoscopy procedures found that patients who screened high-risk for OSA had a longer LOS compared to low risk patients 27, with a difference in median LOS of approximately 30 minutes. Our study found no difference in LOS between diagnosed or high-risk patients vs. moderate or low risk patients, and this held true for both patients undergoing outpatient procedures as well as for patients undergoing more extensive AXR procedures. Notably, this study was conducted on predominantly male (78%) urologic oncology patients, compared to our predominantly female population (84%). It is unclear whether the patients who stayed longer required increased monitoring and use of PAP devices, however our data demonstrate that even high-risk and diagnosed patients may be managed with CPAP/BIPAP without prolonging LOS.
A 2003 retrospective analysis conducted by Sabers et al. among 234 patients with polysomnography-confirmed OSA and 234 matched controls found no significant difference in the rate of unplanned hospital admission, including readmission within 24 hours, or other adverse events.28 Our study, with a larger sample size and different surgical population, drew similar conclusions regarding transfer to the main hospital after surgery (equivalent to unplanned admission), but we report a 0.8% transfer rate compared to their 24%. Our expanded quality outcome of readmissions within 30 days also showed no significant difference based on OSA status (Table 4).
Our findings regarding post-operative respiratory events confirm those of Fernandez-Bustamante et al., who reported that patients at high-risk for OSA had similar rates of post-operative adverse respiratory events as those diagnosed with OSA 29. They also found, however, that patients at high-risk for OSA had higher rates of some adverse events than diagnosed patients, including reintubation, mechanical ventilation, and direct admission to the ICU after surgery, with an absolute increase in high-risk patients of approximately 3% for all three outcomes. Despite this increased risk of adverse events, they did not find a clinically significant difference in LOS (median 3 days for both at-risk and diagnosed OSA). Conversely, our study found that there was no evidence of an increase in LOS, transfers, UCC visits, readmissions within 30 days, and rates of postoperative respiratory events for diagnosed and high-risk patients based on the effect size estimates and the upper bounds of the corresponding confidence intervals. These differences may reflect that Fernandez-Bustamante et al. studied an inpatient surgery population while our study was conducted on ambulatory surgery patients.
Our study has several strengths. Firstly, our large sample size of 5721 ambulatory oncologic surgery patients provided a diverse patient population and enabled us to adjust for a number of patient and surgical characteristics. Secondly, the use of respiratory therapists and clinical staff to provide PAP during the postoperative stay may have contributed to the uneventful postoperative course, especially for the high-risk and diagnosed OSA patients. A previous study found that involvement of respiratory therapists in the postoperative care of patients diagnosed with or at-risk for OSA helped prevent acute respiratory compromise in these patients30. Furthermore, the involvement of respiratory therapists increased the reliability and validity of data collected on respiratory events and devices used through their direct observations. Finally, STOP-Bang screening is part of the routine pre-operative evaluation for all patients at MSK and is consistently and reliably implemented and electronically documented by our trained nurse practitioners.
This study also has several limitations. STOP-BANG screening has its limitations with sensitivity and specificity of 83.6% and 56.3% respectively.23–24 However, it is likely that patients that have severe OSA, restricting their functional ability with associated co-morbidities, would be scored as ASA 4 which triggers a PST staff consult with an anesthesiologist about patient suitability for JRSC. We feel our results demonstrate the feasibility of universal screening for OSA and practical applications in the clinical setting.
As with all retrospective observational studies, there are limitations on drawing firm conclusions. Naturally, the ideal study would be to randomize patients with OSA to inpatient vs ambulatory care. However, this is less likely primarily because of insurance designations. Our estimate of the increase in risk of UCC visits associated with OSA is 2.9%. Even if treating OSA patients as inpatients halved the risk of a OSA-related complication, this would mean a small absolute risk difference of about 1.5%. To detect this difference in a randomized trial would require about 8,000 patients, and would therefore be infeasible. Moreover, it is questionable whether this decrease in risk would be clinically relevant, that is, whether it would be worth treating 67 OSA patients on an inpatient basis to prevent one OSA-related UCC visit. Nonetheless, there is room for reasonable disagreement as to the clinical relevance or otherwise of the observed difference in the study adverse events.
Patients who screened as low risk were not initially monitored directly by respiratory therapists, unlike patients at higher risk or diagnosed. If a post-operative respiratory event occurred in a low-risk patient, respiratory therapists were notified by nurses and would then begin close monitoring of the patient. Thus, post-operative respiratory events may be underreported among low-risk patients. Conversely, outcomes such as LOS, UCC visits, and hospital readmissions and transfers were systematically collected. Nevertheless, a limitation of our 30 -day readmission and UCC visit outcomes is that those occurring outside our MSK system would not be automatically captured, resulting in potential underreporting. Additionally, respiratory therapists are not available after midnight and thus respiratory events and device use are potentially underreported for overnight patients. Nevertheless, all patients arrive in the PACU before the respiratory therapist leaves for the night and none were transferred after midnight for pulmonary or respiratory issues. Although data collection was not designed for study purposes but was part of routine care, our data are accurate and complete and this allows examination of a large cohort of patients with high external validity for other practices. The lack of OSA-related events requiring intervention while these patients continued under our care should lend further confidence to the safety of managing these patients in the ambulatory setting.
More recently, after the completion of this study, a detailed review on intraoperative management of OSA patients has been published. This report suggested that patients with OSA are at increased risk for difficult intubation and mask ventilation, and recommend precautions when using propofol, neuromuscular blocking agents, opioid medications, or benzodiazepine sedation for OSA patients as they may be at increased risk for adverse effects from the use of these agents.22
Our results contribute to the growing body of evidence supporting that patients with moderate, high-risk, or diagnosed OSA can safely undergo outpatient and advanced ambulatory oncology surgery without increased health care burden and adverse post-operative outcomes in a protocolized perioperative management environment. Our data support the thoughtful adoption of practices promoted by several national OSA guidelines, focusing on pre-operative identification of OSA patients and clinical pathways for perioperative management and postoperative monitoring.
Key points.
Question
What is the association between OSA status and short-term outcomes as well as safety for patients undergoing ambulatory surgery in light of current national OSA guidelines?
Findings
No significant association was found between the risk of OSA, and length of stay, urgent care visits, readmissions or risk of transfer for a variety of ambulatory surgery cancer procedures in a free- standing ambulatory surgery facility.
Meaning
OSA patients can safely undergo outpatient and advanced ambulatory oncology surgery without increased health care burden or increasing adverse post-operative outcomes in a protocolized perioperative management environment.
Acknowledgements
Research reported in this publication was supported in part by the National Cancer Institute of the National Institutes of Health (NIH) under Award Number R25CA020449, “Summer Research Experiences For Medical Students Supervised By Faculty Mentors” (Betsy Szeto). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Footnotes
No funding to disclose.
No clinical trial number applicable.
No conflicts of interest to disclose.
References
- 1.Peppard PE, Young T, Barnet JH, Palta M, Hagen EW, Hla KM. Increased prevalence of sleep-disordered breathing in adults. Am J Epidemiol. 2013;177(9):1006–1014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Heinzer R, Vat S, Marques-Vidal P, et al. Prevalence of sleep-disordered breathing in the general population: the HypnoLaus study. Lancet Respir Med. 2015;3(4):310–318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Franklin KA, Lindberg E. Obstructive sleep apnea is a common disorder in the population-a review on the epidemiology of sleep apnea. J Thorac Dis. 2015;7(8):1311–1322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Singh M, Liao P, Kobah S, Wijeysundera DN, Shapiro C, Chung F. Proportion of surgical patients with undiagnosed obstructive sleep apnoea. British journal of anaesthesia. 2013;110(4):629–636. [DOI] [PubMed] [Google Scholar]
- 5.Chung F, Memtsoudis SG, Ramachandran SK, et al. Society of Anesthesia and Sleep Medicine Guidelines on Preoperative Screening and Assessment of Adult Patients With Obstructive Sleep Apnea. Anesthesia and analgesia. 2016;123(2):452–473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Patil SP, Schneider H, Schwartz AR, Smith PL. Adult obstructive sleep apnea: pathophysiology and diagnosis. Chest. 2007;132(1):325–337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ramar K, Caples SM. Vascular changes, cardiovascular disease and obstructive sleep apnea. Future cardiology. 2011;7(2):241–249. [DOI] [PubMed] [Google Scholar]
- 8.Finkel KJ, Searleman AC, Tymkew H, et al. Prevalence of undiagnosed obstructive sleep apnea among adult surgical patients in an academic medical center. Sleep medicine. 2009;10(7):753–758. [DOI] [PubMed] [Google Scholar]
- 9.Chung SA, Yuan H, Chung F. A systemic review of obstructive sleep apnea and its implications for anesthesiologists. Anesth Analg. 2008;107(5):1543–1563. [DOI] [PubMed] [Google Scholar]
- 10.Kaw R, Chung F, Pasupuleti V, Mehta J, Gay PC, Hernandez AV. Meta-analysis of the association between obstructive sleep apnoea and postoperative outcome. Br J Anaesth. 2012;109(6):897–906. [DOI] [PubMed] [Google Scholar]
- 11.Hai F, Porhomayon J, Vermont L, Frydrych L, Jacoude P, El-Solh AA. Postoperative complications in patiehts with obstructive sleep apnea: a meta-analysis. J Clin Anesth. 2014;26(8):591–600. [DOI] [PubMed] [Google Scholar]
- 12.Opperer M, Cozowicz C, Bugada D, et al. Does obstructive sleep apnea influence perioperative outcome? A qualitative systematic review for the Society of Anesthesia and Sleep Medicine task force on preoperative preparation of patients with sleep disordered breathing. Anesth Analg. 2016;122(5):1321–1334. [DOI] [PubMed] [Google Scholar]
- 13.Proczko MA, Stepaniak PS, de Quelerij M, et al. STOP-Bang and the effect on patient outcome and length of hospital stay when patients are not using continuous positive airway pressure. J Anesth. 2014;28(6):891–897. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Memtsoudis S, Liu SS, Ma Y, et al. Perioperative pulmonary outcomes in patients with sleep apnea after noncardiac surgery. Anesth Analg. 2011;112(1):113–121. [DOI] [PubMed] [Google Scholar]
- 15.Vasu TS, Doghramji K, Cavallazzi R, et al. Obstructive sleep apnea syndrome and postoperative complications: clinical use of the STOP-BANG questionnaire. Arch Otolaryngol Head Neck Surg. 2010;136(10):1020–1024. [DOI] [PubMed] [Google Scholar]
- 16.Chung F, Subramanyam R, Liao P, Sasaki E, Shapiro C, Sun Y. High STOP-Bang score indicates a high probability of obstructive sleep apnoea. Br J Anaesth. 2012;108(5):768–775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Memtsoudis SG, Stundner O, Rasul R, et al. The impact of sleep apnea on postoperative utilization of resources and adverse outcomes. Anesth Analg. 2014;118(2):407–418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mokhlesi B, Hovda MD, Vekhter B, Arora VM, Chung F, Meltzer DO. Sleep-disordered breathing and postoperative outcomes after elective surgery: analysis of the nationwide inpatient sample. Chest. 2013;144(3):903–914. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Fouladpour N, Jesudoss R, Bolden N, Shaman Z, Auckley D. Perioperative complications in obstructive sleep apnea patients undergoing surgery: a review of the legal literature. Anesth Analg. 2016;122(1):145–151. [DOI] [PubMed] [Google Scholar]
- 20.Joshi GP, Ankichetty SP, Gan TJ, Chung F. Society for Ambulatory Anesthesia consensus statement on preoperative selection of adult patients with obstructive sleep apnea scheduled for ambulatory surgery. Anesth Analg. 2012;115(5):1060–1068. [DOI] [PubMed] [Google Scholar]
- 21.American Society of Anesthesiologists Task Force on Perioperative Management of patients with obstructive sleep apnea. Practice guidelines for the perioperative management of patients with obstructive sleep apnea: an updated report by the American Society of Anesthesiologists Task Force on Perioperative Management of patients with obstructive sleep apnea. Anesthesiology. 2014;120(2):268–286. [DOI] [PubMed] [Google Scholar]
- 22.Memtsoudis SG, Cozowicz C, Nagappa M, et al. Society of Anesthesia and Sleep Medicine guideline on intraoperative management of adult patients with obstructive sleep apnea. Anesth Analg. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Chung F, Yegneswaran B, Liao P, et al. STOP questionnaire: a tool to screen patients for obstructive sleep apnea. Anesthesiology. 2008;108(5):812–821. [DOI] [PubMed] [Google Scholar]
- 24.Chung F, Elsaid H. Screening for obstructive sleep apnea before surgery: why is it important? Curr Opin Anaesthesiol. 2009;22(3):405–411. [DOI] [PubMed] [Google Scholar]
- 25.Memtsoudis SG, Besculides MC, Mazumdar M. A rude awakening--the perioperative sleep apnea epidemic. N Engl J Med. 2013;368(25):2352–2353. [DOI] [PubMed] [Google Scholar]
- 26.Stierer TL, Wright C, George A, Thompson RE, Wu CL, Collop N. Risk assessment of obstructive sleep apnea in a population of patients undergoing ambulatory surgery. J Clin Sleep Med. 2010;6(5):467–472. [PMC free article] [PubMed] [Google Scholar]
- 27.Balachandran DD, Faiz SA, Hernandez M, et al. Length of stay in ambulatory surgical oncology patients at high risk for sleep apnea as predicted by STOP-BANG questionnaire. Anesthesiol Res Pract. 2016;2016:9425936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Sabers C, Plevak DJ, Schroeder DR, Warner DO. The diagnosis of obstructive sleep apnea as a risk factor for unanticipated admissions in outpatient surgery. Anesth Analg. 2003;96(5):1328–1335. [DOI] [PubMed] [Google Scholar]
- 29.Fernandez-Bustamante A, Bartels K, Clavijo C, et al. Preoperatively screened obstructive sleep apnea is associated with worse postoperative outcomes than previously diagnosed obstructive sleep apnea. Anesth Analg. 2017;125(2):593–602. [DOI] [PubMed] [Google Scholar]
- 30.Ramachandran SK, Kheterpal S, Haas CF, Saran KA, Tremper KK. Automated notification of suspected obstructive sleep apnea patients to the perioperative respiratory therapist: a pilot study. Respir Care. 2010;55(4):414–418. [PubMed] [Google Scholar]