Introduction
Tonsillectomy (i.e. removal of the tonsils) is one of the most common operations performed worldwide.Reference Burton and Glasziou1 Although chronic or recurrent tonsillitis is not a life-threatening condition, it can significantly affect the patient's quality of life (QoL): chronic sore throat can cause many days of missed work, halitosis can cause social embarrassment, and treatment can be a significant economic burden.
Quality of life is a measure of wellness.Reference Calman2 Wellness depends not only on good health but also on social and physical factors. It should not be confused with ‘patient satisfaction’, which is closely related to how satisfied the patient is with the level of service received.
Reliable measurement, quantification and interpretation of QoL after tonsillectomy is of the utmost importance as it can indicate the effectiveness of the procedure.Reference Robinson, Gatehouse and Browning3 Moreover, patients' appraisal of the effectiveness of their tonsillectomy should not be ignored. Such appraisal can help determine whether tonsillectomy is worth having, especially since significant post-operative pain is involved. Thus, it is important to have robust evidence in order to confidently state to patients whether or not a tonsillectomy will improve their QoL.
However, although the effects of tonsillectomy have been studied extensively in children, in adults and adolescents this is not the case. Existing studies attempting to measure post-tonsillectomy QoL changes in adults are few, small and mostly retrospective. There are even fewer studies of adolescents. This means that adult tonsillectomy due to chronic or recurrent tonsillitis is still a controversial procedure, without fully established benefits or (long-term) cost-effectiveness.
The benefits of tonsillectomy for certain conditions (e.g. malignancies) are undisputed. Hence, this review is limited to tonsillectomies performed for chronic or recurrent tonsillitis, with or without coexisting conditions.
This systematic literature review analysed the evidence from studies which directly measured QoL changes. It aimed to determine whether tonsillectomy improves QoL in adults suffering from chronic or recurrent tonsillitis.
Methods
An initial database search was conducted to exclude the possibility of existing or planned reviews on our study topic.
We then conducted a systematic search of the Medline and PubMed databases up until the third week of November 2011. We utilised a search strategy using combinations of text words and Medical Subject Heading terms relating to ‘tonsillectomy’, ‘quality of life’ and ‘adult’ (defined for the purposes of this review as older than 15 years) (see Table I). We limited accepted studies to those published in the English language. Further relevant papers were then identified in an iterative fashion by searching through the bibliographies of the identified studies. Articles which cited the included studies were also identified and assessed. The search string was reviewed and refined several times in this way. No attempt was made to hand-search journals or theses or to find unpublished literature. Figure 1 shows the process of study selection, as well as the inclusion and exclusion criteria.
Study selection criteria.
Search string
Retrospective studies were deemed appropriate for inclusion. We believed it was necessary to obtain patients' subjective, retrospective views of their QoL post-tonsillectomy in order to determine the long-term effectiveness of the procedure. Furthermore, one of the most widely used, validated and ENT-specific surveys, the Glasgow Benefit Inventory, assesses QoL changes retrospectively.Reference Robinson, Gatehouse and Browning3
Studies that did not assess QoL directly (e.g. those assessing patient satisfaction only) were excluded.
Results
General
Eight studies were identified, the earliest published in 2002 and the latest in 2010 (Table II).Reference Bhattacharyya and Kepnes4, Reference Senska, Ellermann, Ernst, Lax and Dost5 We identified no randomised, controlled trials published on our review topic.Reference Witsell, Orvidas, Stewart, Hannley, Weaver and Yueh11
Review results
For Glasgow Benefit Inventory (GBI), scale of –100 to +100, positive values indicate quality of life (QoL) improvement; for Short Form questionnaires (SF; i.e. SF-12 and longer version SF-36), positive values indicate QoL improvement. *Compared with baseline; †compared with 6-month individual component summary (CS) score. Resp = response rate; CI = confidence interval; Δ = change from baseline; HT = ‘hot tonsillectomy’ subgroup (method of tonsillectomy not analysed); CT = ‘cold tonsillectomy’ subgroup (method of tonsillectomy not analysed); – = data not supplied; NS = not significant
Table III shows an evaluation of the surveys used.
Survey evaluation
✓ = strength; 
= weakness; QoL = quality of life; GBI = Glasgow Benefit Inventory
All eight studies reported a general improvement in QoL after tonsillectomy in adults (Table II).Reference Bhattacharyya and Kepnes4–Reference Witsell, Orvidas, Stewart, Hannley, Weaver and Yueh11 Several studies had a relatively low response rate, below 50 per cent.Reference Bhattacharyya and Kepnes4, Reference Baumann, Kucheida, Blumenstock, Zalaman, Maassen and Plinkert6–Reference Richards, Bailey, Hooper and Thompson8 This could possibly have contributed to a selection bias, since patients with an improvement in their QoL may be more likely to respond.
Studies using the Glasgow Benefit Inventory
Bhattacharya et al. demonstrated a significant improvement in QoL (total score, +27.54; p < 0.001).Reference Bhattacharyya and Kepnes4 This improvement was mostly accounted for by the general health component of the Glasgow Benefit Inventory (+35.74; p < 0.001). However, the patients' QoL change was measured only once. A ‘snapshot’ of overall change in QoL does not reflect internal fluctuations.Reference Allison, Locker and Feine13 Thus, a trend could not be established.
In contrast to Bhattacharya et al., Baumann et al. reported no significant change in the social functioning subscale of the Glasgow Benefit Inventory (p = 1).Reference Baumann, Kucheida, Blumenstock, Zalaman, Maassen and Plinkert6 This study had a higher response rate (42 per cent) and a bigger sample size (n = 109) than Bhattacharya et al., so its results are likely to be more reliable. However, this study was conducted in Germany, with surveys in German, so cultural and translation issues may have influenced the results.
Schwentner and colleagues' Glasgow Benefit Inventory results all showed an improvement in QoL. However, in the ‘hot tonsillectomy’ subgroup the p values for these changes (i.e. 0.214, 0.085, 0.848, 0.941; see Table II) were all above the accepted 0.05 threshold; thus, these results cannot be regarded as statistically significant (p values for the ‘cold tonsillectomy’ subgroup were not reported). It is also worth noting that the range of patient ages in this study was reported to be 60 years. Even though this might enable wide-ranging conclusions, it is unlikely that young adults in their 20s would have the same perceptions as patients in their 70s. Therefore, the clinical significance of some of this study's findings can be questioned.
Richards et al. narrowed their patient age range to young adults (i.e. 15–25 years).Reference Richards, Bailey, Hooper and Thompson8 This potentially allowed these authors to make specific observations on that particular age group. Agreeing with Baumann et al., and in contrast with Bhattacharya et al., Richard and colleagues' social functioning subscale results indicated a non-significant change, and thus the possibility of no social benefit from tonsillectomy. Even though this study was smaller in size, with only 40 patients, the resultant loss of statistical power may have been compensated for by its narrower age range. All other subscale scores, and the total score, indicated a significant QoL improvement. In particular, Pearson's coefficient for both the total score and the general health subscale score was 0.96, indicating a nearly linear correlation.
In contrast with some aspects of the above studies, Koskenkorva and colleagues' study had certain strengths, including a response rate of 89 per cent and a relatively narrow age range (15–46 years).Reference Koskenkorva, Koivumen, Penna, Teppo and Alho9 In addition, the study design also included some prospective emphasis, with diary data collected pre-operatively as well as post-operatively. Although this study reported the largest increase in QoL post-tonsillectomy (+35.2), confidence intervals (CIs) were not reported.
Similarly, Senska et al. had an 85 per cent response rate, which should, in theory, improve the validity of their results.Reference Senska, Ellermann, Ernst, Lax and Dost5 In addition, this study's statistical power was more substantial as it was one of the largest studies, with 97 participants. Similarly to most other studies, there was no substantial improvement in social functioning subscale scores. Senska and colleagues' report was the third study in our review to show very little change in social QoL post-tonsillectomy. It should be noted that this study was aimed at a German-speaking population within Germany. All other subscales scores, and the total score, indicated a substantial improvement in QoL; in particular, scores were +39 for physical functioning, compared with +19 for total score. Moreover, the authors' graphical representation of results showed that the 95 per cent CIs for the total score and the general health subscale score included 0, indicating that the statistical significance was uncertain. In contrast, original graphical representation of the 95 per cent CI for the physical functioning subscale score indicated a robust positive improvement.
Studies using the Short Form questionnaire
Ericsson and colleagues' study used the Short Form-36 questionnaire, and reported a significant improvement (10.1 units, p < 0.001) in the physical component summary score one year post-operatively.Reference Ericsson, Ledin and Hultcrantz10 In contrast, the mental component summary score showed no statistically significant change. This study had an astonishing response rate of 97 per cent and a narrow age range (16–25 years). Thus, there was greater potential for more accurate conclusions. However, the sample size was small (n = 43), even if it was prospectively observed.
Witsell et al. used the Short Form-12 questionnaire both at six months and one year post-tonsillectomy.Reference Witsell, Orvidas, Stewart, Hannley, Weaver and Yueh11 This study design enabled possible trends to be charted. The physical component summary score showed a similar improvement to that reported by Ericsson et al., and remained stable between 6 and 12 months (being 7.6 and 7.7, respectively). The mental component score showed a decrease in positive QoL change, compared with baseline (being +3.2 at 6 months but +1.3 at 1 year). Unlike all the other studies, this study involved multiple centres and therefore could be expected to have eliminated any influencing factors specific to individual institutions.
Discussion
Possible limitations
Our literature search included English language papers only and was confined to published studies; thus, a publication bias could have arisen. We identified no randomised, controlled studies of our research topic.Reference Witsell, Orvidas, Stewart, Hannley, Weaver and Yueh11 However, the retrospective views of patients are crucial in determining treatment effectiveness and should always be taken into consideration; for this reason, retrospective studies were deemed appropriate for inclusion. No statistical meta-analysis was attempted in this review because the small number of studies and missing CIs did not permit it.
Lessons learned
Performing tonsillectomy as a treatment for chronic or recurrent adult tonsillitis has been described as controversial.Reference Burton and Glasziou1 This is primarily because, from a medical point of view, the evidence supporting the benefits of the procedure is thin. However, most of the available studies, despite some limitations, show that patients rate their overall QoL as better after tonsillectomy. The studies that did not fully agree with this conclusion tended to have small sample sizes and low response rates.
In addition, different types of surveys demonstrated consistent improvement in QoL. This helps to remove the uncertainty introduced by each method's limitations.
More importantly, as shown by the relevant studies, these QoL improvements are long-standing and relatively stable over time.Reference Ericsson, Ledin and Hultcrantz10, Reference Witsell, Orvidas, Stewart, Hannley, Weaver and Yueh11 Some of the included studies demonstrated that the identified improvement in QoL was associated with a decrease in antibiotics usage, physician visits and missed work days.Reference Bhattacharyya and Kepnes4, Reference Senska, Ellermann, Ernst, Lax and Dost5 These effects contribute to the cost-effectiveness of the procedure.Reference Bhattacharyya and Kepnes4
Another interesting finding was the significantly larger QoL change in the younger patients (i.e. less than 30 years old) compared with older patients.Reference Baumann, Kucheida, Blumenstock, Zalaman, Maassen and Plinkert6, Reference Richards, Bailey, Hooper and Thompson8 This age-related difference could possibly be due to the value placed by younger patients on their physical well-being.
Furthermore, it is important for clinicians to recognise that individual circumstances and morbidity before tonsillectomy can affect patients' QoL perceptions. Schwentner et al. reported that the Glasgow Benefit Inventory score improvement for patients with concurrent chronic disease (i.e. total score, +6.7; 95 per cent CI, 0.7–12.7; p = 0.004) was significantly lower than that of patients without chronic disease (i.e. total score, +16.6; 95 per cent CI, 14.3–18.9; p = 0.004). Koskenkorva et al. compared their ‘least pleased group’ (i.e. the lower 30 per cent of patients) to their ‘more pleased group’. They found that days with fever and number of tonsillitis episodes prior to surgery were the only predictive factors of the former cohort's differences in QoL. This suggests that the infective aspects of chronic or recurrent tonsillitis are important predictors of tonsillectomy's effectiveness, from the patient's point of view.
In the reported studies, improvements in physical functioning subscale scores and general health subscale scores had an almost linear relationship with overall QoL improvements.Reference Richards, Bailey, Hooper and Thompson8 This is perhaps not surprising for a procedure aimed primarily at improving health. However, the social benefits of tonsillectomy appeared to be ambiguous.Reference Baumann, Kucheida, Blumenstock, Zalaman, Maassen and Plinkert6, Reference Richards, Bailey, Hooper and Thompson8, Reference Ericsson, Ledin and Hultcrantz10 This could reflect the fact that the morbidity of chronic or recurrent tonsillitis might not have a significant impact on patients' social lives.
Translation into clinical practice
A pragmatic dilemma for patients is whether the benefits of tonsillectomy outweigh the risks. Crucially, clinicians should now be able to advise patients that their overall QoL is likely to improve and the effects could be long-lasting. In particular, patients' general and physical health is likely to improve significantly; however, the effect on their social QoL is still uncertain. Patients with coexisting chronic conditions are likely to benefit less. More importantly, those patients with more severe infective symptoms due to tonsillitis will probably perceive a greater improvement in their QoL. Overall, tonsillectomy is likely to be cost-effective for adult patients, their employers and the health service treating them.
Currently, out of the aforementioned factors affecting QoL outcomes, one – severity of infective symptoms – is incorporated into national practice as an indication for tonsillectomy.14 The Scottish Intercollegiate Guidelines Network suggests the following as indications for tonsillectomy: sore throat due to tonsillitis; seven or more significant and adequately treated sore throat episodes in the preceding year, or five or more episodes in each of the preceding two years, or three or more episodes in each of the preceding three years; and adverse effect of sore throat episodes on normal daily functioning.14
Improvement in QoL is an important long-term factor, but it should not be the only factor directing the decision of whether to offer tonsillectomy (as a treatment for chronic or recurrent tonsillitis). Tonsillectomy, like any surgery, carries an inherent risk; therefore, consideration of the patient's safety should always be the first priority. Figure 2 presents the relative importance of factors affecting tonsillectomy decision-making over time.
Relative importance of factors affecting tonsillectomy decision-making over time. DM factors = factors influencing decision-making
Conclusion
Effective provision of health care is increasingly affected by the need for prudent allocation of scarce resources. Therefore, decisions on whether to perform tonsillectomy as treatment for chronic sore throat should be made based on the chances of surgical success. The long-term success of tonsillectomy largely depends on patients' QoL improvement.
Tonsillectomy has been described as a controversial procedure because its long-term benefits have not been established in adults suffering from chronic or recurrent tonsillitis; most of the current evidence has been extrapolated from children's data. However, several relatively small studies have tried to identify and measure the long-term benefits of tonsillectomy in terms of QoL changes in adult patients.
This review collates and critically appraises all the available studies that directly measure QoL, and attempts to present a complete assessment of evidence on QoL changes.
The identified literature provides consistent evidence that tonsillectomy is likely to produce a long-lasting, continuous improvement in adult patients' general QoL (even if the specific social benefits are ambiguous). The review also analyses factors contributing to this improvement: patients with coexisting chronic conditions are likely to benefit less, while younger patients and those with more severe infective symptoms due to tonsillitis are likely to benefit more.
• Adult tonsillectomy is controversial as its long-term benefits (re chronic tonsillitis) are not established
• This review appraises evidence on quality of life (QoL) after adult tonsillectomy
• There is evidence of prolonged, cost-effective QoL improvement
• Patients with coexisting chronic conditions benefit less
• Younger patients and those with more severe infective symptoms benefit more
These factors could be incorporated into routine hospital practice; they should also be communicated to patients to help them form their own opinions on the costs and benefits of tonsillectomy.
