5.1 OBJECTIVE DEFINITIONS OF OUTCOME

These can be defined as:

Success is most commonly objectively assessed by polysomnography. Over the past decades various cut-off points have been used in different studies estimating UPPP outcome of which a brief review follows. Fujita was the first to define the cut-off value for success as an AI reduction of 50%. Other variables taken into account included the level of oxygenation (SaO₂>85%) and the number of arousals.³ The large meta-analysis of Sher et al. used an AI reduction with an absolute AI of 10 and/or an RDI reduction of 50%, with an RDI of less than 20 to evaluate treatment outcome.¹¹ Larsson et al. also used an RDI reduction of 50% together with a reduced RDI of 20 or less for their long-term evaluation of UPPP outcome.¹⁶ A 50% reduction of the AHI was also used by Janson et al. and combined with an absolute AI value of 10 or less.¹⁷

Another frequently used outcome measure is the Oxygen Desaturation Index (ODI) where responders are defined as having a postoperative ODI of <20 with an ODI reduction of at least 50% compared to preoperative values.^16,18

We propose this subdivision into responders (success and response) and non-responders (comprising both no change and deterioration) to achieve a simplified subdivision in treatment outcome and ease study comparison.

Issues that require attention in patient classification by PSG measurements are night-to-night variability and the influence of different types of leads when interpreting PSG values close to the cutoff point between responders and non-responders.⁵

For UPPP, another clinical distinction can be proposed:

Fig. 60.1 A, Situation before UPPP. B, Clinically desired situation after UPPP.

Fig. 60.2 A, Mild to moderate stenosis after UPPP; clinically undesired situation. B, Severe stenosis after UPPP; clinically undesired situation.

Nasopharyngeal stenosis is a much feared iatrogenic complication of UPPP. Severe stenosis is estimated to occur in less than 1% of post-UPPP cases.¹⁹ Based on the degree of severity, nasopharyngeal stenosis can be divided into mild/moderate stenosis (Figure 60.2A) or severe stenosis (Figure 60.2B), each requiring a different approach. Mild to moderate stenosis can be managed with Z-palatoplasty according to Friedman.²⁰ In severe cases of NPS, Krespi and Kacker found that the use of CO₂ laser combined with a custom-made nasopharyngeal obturator gives satisfactory long-term results in severe NPS management.¹⁹

6.1 UPPP FOR SOCIALLY UNACCEPTABLE SNORING

Simple snoring is often the result of soft palate vibration. UPPP targets this area and it is therefore no surprise that initial success rates in this situation are high at approximately 87%. Unfortunately this percentage is almost reduced to half after long-term follow-up.²³A retrospective review of snoring surgery by Jones et al. cites a wide range in success rate of 45% (at 2 years) to even 90% (at 5 years) but the general trend is a regression of positive outcome with time. The estimated overall failure rate of snoring surgery (including UPPP and irrespective of type of surgery) is ∼66% at 2 years.²⁴

Of equal importance is the level of partner satisfaction after UPPP. Prasad et al. found a significant reduction in sleep disturbance and an improved quality of life as measured by a questionnaire-based survey with a minimum of a 1-year follow-up.²⁵

6.2 UPPP FOR OSAS

The range of success rates reported in various studies has a wide variation of 25–85%.²⁶

In 1981 Fujita described the first cohort to be treated with UPPP and arbitrarily defined criteria for success which have since been widely applied. However, this value was based on the study group as a whole with large variation among the individuals. Even in this first study, two out of the total of 12 patients had an increase in apneas, although none indicated subjective worsening after UPPP. (Objective) UPPP ‘failure’ has been described from the beginning.³ Sher et al. did the first meta-analysis which showed a success rate of 50% when using the criteria defined by Fujita. This rate dropped to 40.7% in unselected patients when adjusted to include an AI of <10. The success rate was 52.3% in the group of patients clinically suspected of unilevel oropharyngeal obstruction, indicating the importance of anatomical staging.¹¹

The first long-term follow-up (46 months) was by Larsson et al. in a group of 50 UPPP patients. Using a postoperative RDI <20 and a RDI reduction of at least 50% as the criteria for success, they found 60% of patients to be responders at 6 months. This amount decreased to 38.8% by 21 months but climbed back to 50% at a mean of 46 months.¹⁶ Janson et al. found a positive response of 64% 6 months postoperatively but a similar percentageof responders (48%) to other studies after a 4 to 8-yearfollow-up using a 50% or more decrease in AHI and a postoperative AI of less than 10 as criteria for success.¹⁷ In 2000 Boot et al. published results of another long-term follow-up in unselected patients. Using ODI as criterion for success,this study showed that only ∼20% remained as responderover time.¹⁸

Senior et al. found UPPP to be effective in 40% of their cohort of mild OSAS patients. Their study supported the increasing evidence of sub-optimal UPPP treatment outcome. However, the non-responders showed objective worsening on polysomnographic evaluation combined with lack of subjective symptom improvement.²¹ This finding drew attention to an underreported aspect of UPPP outcome in the literature. Deterioration after UPPP was also noted by Hessel and de Vries who even found some failures in the Type 1 and 2 groups despite preoperative selection by sleep endoscopy. They reported better than average results of UPPP with use of sedated endoscopy for patient selection but encountered failures (deterioration) in approximately 10–15%. Results were negatively influenced by previous tonsillectomy (with relative short anterior–posterior diameter at retropalatinal level) and simultaneous obstruction at tongue base level (Type 2 does worse than Type 1).²²

6.3 COMPLICATIONS

Another aspect of UPPP failure is the occurrence of complications. Grontved and Karup administered a questionnaire to 69 patients 2 years post-UPPP. They found that 42% had complaints.²⁷Although the majority of these were minor, it shows that patient satisfaction deserves ample attention and underlines the importance of both objective and subjective rating of surgical success.

This aspect was studied by Walker-Engstrom and colleagues who assessed quality of life after treatment of mild to moderate OSAS patients with either a dental appliance or UPPP over a 1-year follow-up period. They found that although both interventions led to significant improvement in quality of life (evaluated by the dimensions vitality, contentment and sleep), the subjective improvement was greater for UPPP. However, the AI and AHI rate (<5 resp. <10) was more normalized in the dental appliance group than the UPPP group (78% vs. 51%), underlining once again the frequent discrepancy between subjective and objective treatment outcomes.²⁸

It is concerning to discover that long-term outcome of UPPP on OSAS shows a high level of relapse.^17,18,21 Furthermore, there are increasing reports of disease deterioration after surgical intervention.^21,22 This raises the question whether it is justified to put patients at risk of surgery for an intervention that in the long run is likely to give minimal to no improvement and possibly even worsening of the situation. UPPP indeed appears not to be the single-step answer but it does prove sufficient therapy for a specific group of patients. The subsequent question therefore is: which patients are these and how do we identify them preoperatively?

6.4 WHICH FACTORS ARE ASSOCIATED WITH THE VARYING RESPONSE RATE TO UPPP?

When understanding the factors associated with OSAS, it is first of all important to appreciate the multifactorial pathophysiology of this disease. The complex interplay between numerous factors gives way to varying degrees of OSAS in different individuals.

Multilevel involvement is probably one of the most complicating factors in OSAS management. UPPP targets the oropharyngeal level but collapse at the hypopharyngeal level, if present, is not altered.^21,26

The influence of preoperative weight or Body Mass Index is a much debated issue. Riley et al. found that morbidly obese patients had less successful UPPP outcome.²⁹ However, there are conflicting results, as Sher et al. found no significant effect of BMI on UPPP outcome in their large meta-analysis.¹¹ On the other hand, Larsson et al.6 found a significant lower preoperative BMI in their responder group but this was not confirmed by Boot et al.³⁰ Larsson et al. also found a negative influence of postoperative weight gain on UPPP efficacy which was partly reversed in the same group after weight reduction.¹⁶

As for the severity of the disease, Sher et al. found that UPPP is less successful in more severe cases of OSAS and this relationship remained apparent despite stricter adjustment of the definition of success.¹¹ Using the same criteria for success, Senior et al. found no difference in UPPP outcome between mild and severe OSAS patients.²¹ This finding was also supported by Friedman et al. who showed that their clinical anatomical staging system is superior to disease severity in predicting UPPP outcome.²⁶

The influence of gender is interesting, foremost because of the significant difference in the pattern of SDB between the sexes, such as presenting symptoms, the severity of disease and the association with BMI. This is an area that needs further research, as does the role of ethnicity, which shows differences in the distribution of risk factors among various ethnic groups without a clear explanation. Age shows a tendency to increase the prevalence of snoring and OSAS up until middle age after which there is no further change.¹ Once again the exact mechanism is unknown.

Other comorbidities such as disproportionate anatomy, allergy and hypoventilation syndrome have been observed but their precise contribution as potential risk factors is less studied.¹

Polysomnographic recordings are the most frequently used way of obtaining objective data on pre- and postoperativeoutcome. An aspect that until recently was not fully recognized and included in these recordings is the so-called ‘positional sleep apnea’. These so-called position-dependent OSAS patients have an AHI in a supine position that is twice the AHI in other positions. Awareness of this positional sleep apnea phenomenon and correction for it, if necessary, when interpreting polysomnographic results can influence overall UPPP outcome.^31,32

6.5 THE VALUE OF PREOPERATIVE PATIENT SELECTION FOR UPPP

Over the years various patient selection methods have been developed to influence the likelihood of surgical success. The majority are based on the above-mentioned risk factors thought to influence UPPP outcome. It is imperative to take these factors into account when considering UPPP as a surgical treatment option.

Patient selection methods can be divided into subjective and objective measurement tools. It needs emphasizing that the majority of assessment tools are in some way subjective due to symptom scoring by the patient and/or partner or inter-investigator variability. This not only leads to difficulty in developing (objective) diagnostic measures but also complicates comparison of different studies.

The growing insight into the pathophysiology of OSAS places high demands on preoperative selection methods. Especially the multilevel involvement combined with alteration in dynamic circumstances necessitates the use of multiple selection modalities, as no single test has proved to cover the vast area of OSAS physiology so far.

To date, polysomnography remains the backbone of the OSAS diagnostic work-up for disease severity and the most frequently used way of selecting patients for surgery and evaluating postoperative outcome. Although poly-somnography is a valuable instrument in obtaining objective data it does not give information on the levels involved in upper airway collapse. Topical diagnostic work-up is presently either done by clinical examination only (e.g. using the Mallampati–Friedman staging) or (in combination) with sleep endoscopy. Sleep endoscopy is in our own experience the most feasible and promising technique to address this issue, but the subject is highly controversial in whether all patients in whom surgery is considered should have sleep endoscopy routinely performed. Hessel and de Vries showed that UPPP success rate (defined as an AHI drop <20) almost doubles by preoperatively combining PSG and sleep endoscopy in the diagnostic work-up.⁹ This study revealed two interesting findings. The first and only statistically significant finding is that UPPP is less successful in patients with prior tonsillectomy as many of these patients have a relatively shorter anterior–posterior diameter between posterior pharyngeal wall and soft palate. This has led to an important change in our policy. Currently we more frequently perform ZPP than simple UPPP in these cases.

The other important finding in this study confirmed that cases of combined retropalatal–retrolingual obstruction did worse than sole palatal obstruction. Therefore we now manage Type 2 cases with a relatively low AHI by combined UPPP and radiofrequency of the tongue base, while in more severe obstruction (as assessed by sleep endoscopy and higher AHI) we prefer to combine UPPP, radiofrequency of the tongue base and hyoid suspension as primary surgery, instead of UPPP only (see Chapter 50: Multilevel surgery). Continuing on our experience with multilevel surgery, we found that secondary hyoid suspension after UPPP failure is less successful than primary hyoid suspension for patients with obstruction at tongue base level and moderate to severe OSAS (see Chapter 49: Hyoid suspension as the only procedure). In addition, a respectively relative or functional post-UPPP stenosis at palatal level can often not be salvaged by hyoid suspension.³³

Once again we emphasize the importance of careful preoperative assessment. Our preference remains the combination of PSG and sleep endoscopy despite understandable criticism over sleep endoscopy (the primary question being: is artificially induced sleep a reliable representation of natural sleep?). We feel that at present there are no better methods to preoperatively (dynamically) diagnose the level of obstruction in OSAS especially as recent investigation of the correlation of the Mallampati/Friedman staging system with sleep endoscopy showed a low correlation.¹⁰

In general, when attempting to evaluate UPPP outcome (and failure) the ongoing difficulty with the available literature is the lack of uniform definitions of success and the use of modified UPPP techniques with adjunctive measures, which makes comparison difficult and inconclusive and at times yields conflicting results.³⁴