Extensive Counseling

First of all, patients should be comprehensively informed about conservative alternatives to surgery. Devices such as positive airway pressure (PAP) machines, dental appliances, or positional devices (in positional obstructive sleep apnea [POSA]) can be recommended. If the patient feels fine with the device, he/she can decide to keep on using it. A conservative treatment can usually be stopped without long-lasting side effects. In contrast, surgery is not reversible in most cases. Once a tonsil is removed, it cannot be replaced. Furthermore, some complications of surgery are serious and may not be completely resolved. The individual effect of sleep surgery arises 2 to 3 months after surgery. In case it is not sufficient, further treatment might be necessary. This is why we advise our patients to try conservative treatment first (if there is any promising option) before undergoing surgery.

On the other hand, the best thing about successful surgery is that the disease gets completely resolved. In case of PAP users, obstructive sleep apnea (OSA) is only corrected as long as the patient uses his/her machine.

Therefore, in our opinion, it is crucial to inform the patient about advantages and disadvantages of any treatment beforehand.

The treatment of SDB is not necessarily a question of surgical versus nonsurgical measures. Often a combination of treatments finally helps the patients.

Comorbidities

As there are nonsurgical alternatives to surgery in OSA treatment, a surgeon should focus on the patient’s specific comorbidities first. There is an increased perioperative risk for OSA patients undergoing any kind of surgery. ^{1 , 2 , 3 , 4} Various complications have been described in (▶Table 5.1).

In sleep apnea surgery, an additional factor to be dealt with is any wound(s) within the upper airway that might cause bleeding, swelling, and airway compromise. This implicates that OSA surgery should be indicated with care in patients with severe comorbidities, and only in case of failure of conservative treatment.

If a patient with severe cardiorespiratory comorbidities is selected for upper airway surgery, surgery should only be performed as an inpatient procedure. If possible, the patient should be put on CPAP treatment prior to surgery with the treatment being continued during the perioperative phase. Overnight monitoring should be available in these cases. Please refer to Chapter 10 for further information.

Apart from these aspects, it is also mandatory to check for anticoagulative medication that might be stopped or bridged before surgery. Another important point is to look for signs of difficult intubation, such as micrognathia, male gender, age older than 50 years, body mass index (BMI) greater than 35 kg/m², and neck circumference greater than 40 cm. ⁶ In case a difficult intubation is expected, the anesthesiologist should be informed occasionally.

First- or Second-Line Treatment

As stated previously, all kinds of sleep apnea treatments have certain failure rates. These failure rates vary substantially between different treatment modalities. In the case of surgery, we know that cure rates decrease with baseline BMI and apnea–hypopnea index (AHI) (▶Fig. 5.1).

As a consequence authors recommend surgery as a first-line procedure if the AHI is below 30 and the BMI is below 34 kg/m². These cut-off values are not supported by evidence-based medicine data from the literature but are based on authors’ clinical experience of more than 20 years. However, exceptions to the rule are always possible. For example, the authors recently performed combined surgery of uvulopalatopharyngoplasty (UPPP) and tonsillectomy, and radiofrequency treatment (RFT) of the base of tongue in a patient with an AHI of 73 and cured the disease. In other words, if there is a pathoanatomical finding (e.g., massively enlarged tonsils) that is regarded as the origin of the airway obstruction and can be easily treated by surgery, first-line surgery may work well.

Multilevel or Staged Surgery

The authors use drug-induced sleep endoscopy (DISE) in addition to the ear, nose, and throat (ENT) examination (when the patient is awake) to determine the site(s) of obstruction. We prefer the VOTE-classification ⁸ that distinguishes four levels of obstruction, namely, velo-, oro-, hypopharynx, and epiglottis. If multilevel obstruction is detected, we need to decide whether surgeries can be combined at different levels within the same operation or if staged surgery should be recommended for treating every level separately.

In authors’ opinion, multilevel surgery within a single operation is one step further than “staged surgery.” The authors do not recommend performing surgery only one level at a time in case DISE shows obstruction at more levels. Multilevel surgery has proven to be safe, if the patients are selected carefully and monitored thoroughly in the perioperative period. For perioperative monitoring please refer to Chapter 10. In mild OSA, however, the authors do not advocate the use of multilevel surgery in combination with unexpected and unexplained severe DISE findings.

In this context, the authors would like to stress that combinations of pharyngeal and nasal surgeries often lead to increased postoperative morbidity as the patients are forced to breath from their mouth. The loss of nasal climatization and humidification of the breathed air causes discomfort for a couple of days even without the use of nasal dressings. This is why authors do not recommend performing combined nasal and pharyngeal surgeries.

Invasiveness

Minimal invasiveness can be defined as surgery that can be performed under local anesthesia, as an outpatient procedure that causes little intra- and postoperative morbidity, and is affected with low complication rates. According to authors the following procedures can be regarded as minimally invasive: interstitially applied RFTs (turbinates, palate, tonsils, base of tongue, or any combination), and palatal implants. Other soft palate surgery such as uvulopalatoplasty, in its different modifications, causes substantial postoperative pain, but fulfills all the other criteria. Tongue implants (i.e., ReVent) require general anesthesia and is affected with low morbidity.

Minimal invasive techniques have limited effect on AHI. This fact limits the indication of these surgeries to either simple snoring or to mild OSA. The success rates of minimally invasive surgeries are low when the sleep apnea is severe. So the surgeons need to choose more invasive surgeries to be successful in case of severe disease.

Authors would like to exemplify this aspect. In a case with mainly oropharyngeal and limited hypopharyngeal obstruction during DISE authors would recommend a modified UPPP plus tonsillectomy in combination with an RFT of the base of tongue in mild OSA, while preferring a midline glossectomy instead of RFT in of severe OSA. Another example could be palatal and oropharyngeal obstruction only. In snorers and mild apneics authors would recommend radiofrequency-assisted uvulopalatoplasty (RF-UPP) in combination with RFT of the tonsils, while recommending UPPP plus tonsillectomy in moderate to severe OSA.

The surgical principle that more severe OSA requires more invasive surgery has been postulated by Moore in 2000. ⁹ The authors are still convinced that this approach is partly correct, although technical developments and new surgical instruments have helped reduce postoperative morbidity substantially.

To sum it up, authors recommend to select surgeries with limited invasiveness (i.e., limited complication rates, limited morbidity, limited complexity for patient and surgeon) whenever possible while keeping the number of surgeries lesser. A single surgery is often better than repeated surgeries. As outcome in sleep apnea surgery is very difficult to predict, selection of the best surgical technique for the individual patient remains one of the biggest challenges in sleep surgery.

Introduction

Patients who fail to use CPAP compliantly often visit surgeons for an alternative OSA treatment. Such patients might still use their CPAP machine until their visit to the outpatient clinic. Reevaluation of the severity of the disease is often necessary. This will include not only medical history taking and performing a thorough clinical ENT examination including DISE, but also polysomnographic (PSG) reevaluation in case no recent sleep study is available. The primary PSG often dates back to more than 1 year ago and one cannot assume the disease severity to remain stable over time.

The question then arises if patients should be advised to stop their CPAP use before the repeat PSG. Both ethical and medicolegal issues of CPAP cessation might arise. Recent literature suggests that CPAP might have a temporary residual effect on OSA after withdrawal from therapy. This is called the “washout period.” ¹⁰ Therefore, it has been claimed that patients should stop their CPAP at least several days before the PSG is repeated. It has been suggested that without this washout period, the PSG result could be an underestimated level of the OSA severity. This washout phenomenon is an important issue from both theoretical and clinical perspective. This chapter aims to provide an overview of the available literature on this topic.

Material and Methods

A literature search was run electronically in the MED-LINE and EMBASE databases, based on the following main search terms: (CPAP OR nCPAP) and (OSA OR apnea) and (withdrawal OR residual), leaving 561 articles. In addition, articles were identified from the reference lists of these papers. For the comparison of pre-CPAP and post-CPAP outcome, relevant articles were included and analyzed when they enclosed information on OSA parameters before and after withdrawal from CPAP therapy. Studies in which long-term CPAP use (>1 month) was missing were excluded, resulting in a set of 13 studies that met the search criteria and were further evaluated. The studies within the present search criteria investigated different objective and subjective outcome parameters. In order to compare these studies, an analysis was performed focusing on the AHI as reported in the studies.

Results

Overview of the Evidence

The majority of the studies analyzed the severity of OSA during CPAP treatment and after CPAP withdrawal. Since not all studies provided the same data, the studies were split up into different sets. Studies in which there was no significant BMI change before and after CPAP therapy were separated from studies that found a statistically significant reduction of the BMI, or failed to report BMI changes entirely. Studies that calculated the significance for the mean AHI or respiratory disturbance index (RDI) or oxygen desaturation index (ODI) differences between pre- and post-CPAP were separated from those that did not report these calculations. An overview of these studies can be found in ▶Table 5.2, with a bubble chart (▶Fig. 5.2) visualizing the number of nights off CPAP (x-axis), delta AHI/RDI (change in AHI/RDI; y-axis), and number of patients included in each study (circle area).

▶ Table 5.2 Literature on AHI pre-CPAP versus post-CPAP

Study	Year	No. of patients studied (total)	BMI change (P < 0.05)	AHI/RDI pre-CPAP	Mean AHI/RDI on CPAP	Mean AHI/RDI post-CPAP	P value (mean AHI/RDI pre-CPAP vs. post-CPAP)	Duration of CPAP treatment (months)	Nights of withdrawal on which PSGs were performed	AHI definition according to AASM1 11
Young et al 10	2013	20 (42)	No	77.6	2.9	61.9	P < 0.005	4	2	No a
Kohler et al 12	2011	21 (41)	Unknown	45.3	2.2	36.0	Unknown	>12	14	No a
Phillips et al 13	2007	20	Unknown	46	0.7	26.7; 39.0	Unknown	>12	1, 7	Yes
Bonsignore et al 14	2002	10 (29)	No	82	–	63	P < 0.05	5.5	1	Yes
Yang et al 15	2006	20	No	47	2.7	50; 50	Unknown	>12	1, 7	Yes
Pankow et al 16	2004	12	Unknown	43.0	3.7	32.3	Unknown	35	7–9	Yes
Marrone et al 17	2003	13	Yes	80.1	–	64.6	P < 0.05	5	1	Yes
Fiz et al 18	1998	10	No	47.0	5.4	40.5; 44.1; 42.2; 35.8	Unknown	24	1–4 b	Yes
Boudewyns et al 19	1996	25	No	93.6	–	94.1	NS	12	1	No e
Sforza and Lugaresi 20	1995	30	Yes	74.4	1.2	61.1	P < 0.005	13	1	Yes
Kribbs et al 21	1993	15	No	56.6	2.5	36.8	P < 0.0001	2.5	1	Yes
Leech et al 22	1992	17	No	91	–	55	P < 0.0001	6 c	1 d	No f
Rauscher et al 23	1991	21	No	53.9	–	28.7	NS	8	4 h	No g
Abbreviations: AASM, American Academy of Sleep Medicine; AHI, apnea–hypopnea index; BMI, body mass index; CPAP, continuous positive airway pressure; NS, nonsignificant; PSGs, polysomnographies; RDI, respiratory disturbance index. aHypopnea defined as airflow reduction > 30% for more than 10 s with a fall in oxygen saturation (SaO2). bConsecutive nights. cMedian. dPresumably the PSG was conducted at the first night of withdrawal. eHypopneas were defined as a 50% drop in tidal volume from its value in quiet wakefulness prior to sleep onset. fDefinition of hypopnea not noted in study. gHypopneas were defined as a reduction in rib cage and abdominal movements to 50% or less compared to the preceding five breaths for longer than 10 s accompanied by a fall in SaO2 to 92% or lower if baseline was equal or above 94% or a fall in SaO2 of 3% or more if baseline was 93% or lower upon withdrawal (pre-CPAP: 77.6/h sleep, post-CPAP: 61.9/h sleep).

Discussion

A total of 13 studies on CPAP withdrawal were assessed, focusing on AHI and/or RDI, before and after withdrawal from long-term CPAP treatment (>1 month). Of these, three studies failed to report BMI changes during this period, while two studies found significant reductions in mean BMI after CPAP use. All these five studies showed an improvement in the AHI after withdrawal from CPAP compared with before treatment, of which two reported a significant drop in AHI. Importantly, these two studies were also the studies in which a significant decrease of the BMI was reported.

Of the remaining eight studies reporting no significant change in BMI, six showed improved AHI levels post-CPAP compared with AHI levels before treatment, of which four reported significant differences. ^{10 , 14 , 21 , 22} Young et al found statistical difference only in the severe, but not the mild/moderate OSA subgroup. ¹⁰ Two studies showed an increase in the AHI after CPAP withdrawal compared with before treatment. ^{15 , 19} The reported increase in AHI in these two studies was limited (0.5–3/h sleep).

A few studies have been performed with the goal of evaluating the difference in OSA severity (using AHI or RDI) before and after treatment with CPAP. A total of six studies ^{10 , 12 , 18 , 21 , 22 , 23} used AHI or RDI as primary outcome measure. Other studies reported the pre- and post-CPAP AHI level, but used MSLT, ²⁰ respiratory effort, ¹⁹ neurobehavioral performance, ¹⁵ or cardiovascular parameters ^{13 , 14 , 16 , 17} as primary outcome parameters.

Studies that did not mention the AHI or RDI before and after CPAP treatment were excluded. Some of these studies described pre- and post-CPAP differences but focused on parameters such as hypoxemia, ²⁴ stress hormones, ²⁵ blood pressure, ²⁶ snoring characteristics, ²⁷ and driving performance. ^{28 , 29} Another study ³⁰ was not included because patients did not receive long-term CPAP treatment before the investigation.

From the 13 studies matching the search criteria comparing AHI before and after CPAP use, long-term use of CPAP showed to have an acute residual effect after withdrawal from therapy. Of these 13 studies, decreased AHI after CPAP withdrawal was reported in 11 studies with significant decrease of AHI in 6 studies.

There are several important considerations to be made in order to interpret these results. Most studies were small (n = 10 to 30). Studies in which the BMI could have played a role in the AHI improvement were not excluded. Although we are aware that OSA severity is related to BMI, ^{31 , 32} for the sake of comprehensiveness these papers were also taken into account. All six studies that showed a significant difference in OSA severity between pre- and post-CPAP, conducted the PSG on the first or second night of withdrawal. This suggests that it is likely that the AHI is lower immediately after withdrawal and a washout effect is indeed to be expected.

There are three studies that report results of the AHI on multiple days after withdrawal from CPAP. Yang et al ¹⁵ conducted PSGs on the first and seventh night of withdrawal. Fiz et al ¹⁸ repeated PSGs on first 4 nights after withdrawal. Both studies did not show a deterioration of OSA after multiple days of withholding CPAP. In contrast to these results, Phillips et al ¹³ found a significant deterioration of OSA comparing the first and seventh night post-CPAP.

How long it takes before OSA reaches a reliable and stable level after quitting CPAP therapy remains unclear and might vary per individual. As one study showed a significant difference between first and seventh nights of CPAP withdrawal, it could be advised to withdraw treatment for at least 8 days before conducting a PSG.

There may be several explanations for this residual effect of CPAP after withdrawal. Various structures in the upper airway can contribute to upper airway obstruction in OSA. The upper airway anatomy and its response to differences in the pharyngeal pressure is one of the key factors in the pathogenesis of OSA. ^{33 , 34} During CPAP treatment, pharyngeal size increases, ³⁵ thus facilitating airflow. Several articles suggest that a structural change in airway anatomy due to long-term CPAP use is the main reason for the presence of a residual effect after therapy. ^{10 , 23} Because of extended friction of the upper airway in OSA, pharyngeal edema might develop, thereby further decreasing the airway lumen. This edema would disappear with long-term use of CPAP. Ryan et al ³⁵ performed an MRI study pre- and post-CPAP in five OSA patients and found a significant increase in oropharyngeal volume following CPAP therapy. Another study ³⁶ screened 24 OSA patients using cephalometry before and after CPAP treatment. The mean posterior airway space was significantly increased in supine, but not in erect position. Unfortunately, this study did not use MRI. In contrast, Collop et al ³⁷ who investigated pharyngeal volumes using MRI, found no pre- and post-CPAP difference in 12 patients with OSA. As these results show, there is limited data on the change of upper airway anatomy due to CPAP use. As a result, no definitive conclusion can be drawn about the possible existence of a difference between pharyngeal volume before and after CPAP treatment.

Another possible mechanism for the extended effect of CPAP on the AHI is an increased ventilation control mechanism in response to CPAP therapy. According to this theory, long-term exposure induces low oxygen levels in OSA. Patients would blunt their arousal responses to hypoxemic episodes. This lowered threshold in the central nervous system would result in altered responses to obstructive events.

Kimoff et al ³⁸ showed that long-term OSA resulted in a reduction of breathing frequencies in response to hypoxia. Ventilation control mechanisms also seem to be impaired in individuals at high altitude with chronic hypoxemia. ³⁹ A study by White et al ⁴⁰ showed how sleep deprivation significantly decreased ventilation responses to hypoxia. These findings suggest that long-term use of CPAP could counter the altered neuronal threshold for arousal responses by reducing oxygen desaturation and facilitating sleep. This seems plausible regarding the dynamics of neuroplasticity and could play an essential role in the cause of a washout effect after acute withdrawal from long-term CPAP therapy.

It is interesting to evaluate this matter from the sleep surgeon’s perspective in particular, as residual effects of non-CPAP therapies, for example, upper airway stimulation ⁴¹ or oral appliances (OAs), are not observed during DISE. The effect of these therapies seems to immediately disappear after the upper airway stimulation therapy is switched off or the OA is taken out. The same holds true for jaw thrust, chin lift, or similar maneuvers—these effects do not last. One hypothesis could be that, in the studies evaluated in the present paper, CPAP was used for more than 1 month and therefore a more prolonged effect on upper airway structures and behavior could be anticipated.

Atkins et al presented in a study that assessment of the quality of evidence and strength of recommendations by means of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system accentuates the lack of randomized controlled trials, limited study populations, imprecision because of different outcome measures, and the inconsistency of study results. ⁴² Moreover, the main research focus of the present study was mostly a secondary outcome in the majority of the studies included. Furthermore, the balance between health benefits and harms of CPAP cessation and duration thereof was not explicitly considered in any of the included studies. This implies a relatively low quality of currently available evidence to substantiate a solid recommendation on this topic, resulting in a 2C grade of recommendation. Further research, preferable by means of randomized clinical trials, could contribute to more robust recommendations.

Generalization of study results is cumbersome because the studies included only small numbers of subjects, presented variable measurements of respiratory events, discussed different periods of time to assess CPAP washout, and only a few of the included studies took BMI into account. However, addressing the initial research question, even though within the limitations as previously mentioned, remains of clinical importance for any clinician treating OSA patients who failed or refused to use CPAP.

In conclusion, there is some evidence that CPAP washout exists in patients with a stable BMI throughout the follow-up period. However, the intensity and duration of this effect remain unclear.

The studies assessed in this section described rather small patient populations and to what extent other reasons for night-to-night variability were controlled for (e.g., sleep position in POSA, changes in the percentage of obstructive, mixed, and central events), is uncertain. Within these limitations, based on currently available literature, it might be reasonable to maintain a washout period, with approximately 1 week being a possible advisable period, in case alternative OSA treatment options are considered and especially when a baseline PSG (and a follow-up PSG after treatment) is needed in case of clinical trials.