Tonsillectomy for adults with recurrent streptococcal throat infection:randomised controlled trial

Tonsillectomy has some benefits over watchful waiting. Jocalyn Clark examines a recent paper that found the net benefit unclear and concluded that research into longer term outcomes is needed

This month's paper

"Tonsillectomy versus watchful waiting in recurrent streptococcal pharyngitis in adults: randomised controlled trial" by Olli-Pekka Alho and colleagues (BMJ 2007;334:939-41, 8 March doi: 10.1136/bmj.39140.632604.55).

Some of you probably had your tonsils taken out when you were children-especially if you were born in the '70s in the United States, when you might have been one of millions of children who had their tonsils removed. However, you are more likely to have been born in the '80s, when far fewer tonsils were taken out. Nowadays there are concerns that too many tonsillectomies are being carried out unnecessarily and without enough information about the benefits and harms. One commentator estimates that the rate of tonsillectomy in the United Kingdom has fallen by 34% in the 10 years between 1994-5 and 2004-5.¹ Still, it may be an appropriate procedure for some patients for whom other treatments have failed, especially patients with recurrent tonsillitis, estimated to be up to 12% of the population.²

This month's paper looked at whether tonsillectomy was a good and safe treatment option in adults with recurrent streptococcal pharyngitis.

Why do the study?

Surprisingly little evidence shows how effective tonsillectomy might be in adults. A recent review of the evidence for the best treatments for recurrent streptococcal throat infections published by the Cochrane Collaboration found none that show that it is effective in adults.³ This seems important because according to Best Treatments (www.besttreatments.co.uk), about 50000 people in the UK have their tonsils taken out a year, half of whom are at least 15 years old.

What did the authors do?

The authors conducted a randomised controlled trial in which patients with recurrent streptococcal pharyngitis referred to their ear, nose, and throat clinic were assigned to receive immediate tonsillectomy or to a waiting list for removal of their tonsils in the future. To qualify for the trial the patients had to have had three or more episodes of streptococcal pharyngitis in the previous six months or four episodes in the previous year. At least one of these episodes needed to be confirmed as a group A streptococcal infection.

Randomised controlled trials are considered the optimum standard for testing medical treatments or surgical procedures, so it is a good method to study the effectiveness and safety of tonsillectomy. Most health professionals would agree that we cannot know for sure if a treatment or procedure works and is safe until it has been studied in a randomised controlled trial. Two main points are important. Firstly that patients are randomly assigned to receiving the procedure of interest or not, and, secondly, that there is a control group with which the procedure of interest is being compared.

This randomised controlled trial is not a placebo trial, in which the intervention (tonsillectomy in this case) would be compared with no treatment. Instead, the authors used a "waiting list control." This meant they assessed the patients receiving tonsillectomy immediately without denying the other patients their needed surgery; the other patients simply had to wait a little longer to receive it-at least 90 days in this case.

A limitation of using a waiting list control group is that you cannot leave them waiting too long. Therefore, the study period is often relatively short. On the other hand, providing a waiting list helps researchers recruit and retain patients who would otherwise not be interested in a clinical trial in which they had a 50% chance of not receiving the procedure being tested. In this tonsillectomy trial, patients who were enrolled, no matter how severe their streptococcal pharyngitis pain was, knew they would be operated on at some point.

During the study period, 298 patients were referred to the authors' ear, nose, and throat clinic, but only 70 patients were included in the study. Most of the excluded patients did not have recurrent streptococcal pharyngitis, which was required for enrolment. Of the 70 patients, 36 were assigned to have their tonsils removed straight away-this was called the tonsillectomy group. Thirty four patients were put on the waiting list-this was the control group.

The authors recorded whether patients in both groups had any episodes of streptococcal pharyngitis in the 90 day study period. They determined the presence of streptococcal pharyngitis by asking patients and confirmed this using a throat swab. They also asked all patients to keep diaries about their symptoms and any visits to the doctor and about any adverse effects after the surgery for the tonsillectomy patients.

What did they find?

At the end of the 90 day study period, eight patients, or 24%, in the control group and one in the tonsillectomy group (3%) had had a streptococcal pharyngitis episode. This is a 21% difference between groups. Based on these findings the authors calculated a measure called "number needed to treat" (NNT), which can be a helpful way of understanding how effective a treatment is. They reported an NNT of five. This means that five people had to have a tonsillectomy for one of them to avoid a streptococcal pharyngitis episode within three months.

In addition to looking at the differences in the percentage of patients in each group who got streptococcal pharyngitis in those 90 days, the authors also compared the time it took to have a streptococcal pharyngitis episode among those who did. They found that the times to first episode of streptococcal pharyngitis in the control group were much shorter that of the tonsillectomy group.

When it comes to studying how good medical treatments or surgical procedures are, we must also look at whether they have any harms. The authors assessed the adverse effects that the tonsillectomy had. They report that for most patients the tonsil removal operation took one hour; needed to stay one day in hospital; and on average needed 13 days' recovery at home, which involved throat pain. The authors do not report how severe this throat pain was. Among the patients who got tonsillectomy in this study, there were no severe consequences of the surgery. However, two patients (6%) had mild secondary bleeding.

What does the study mean?

The study is important because it is one of few trials that has considered tonsillectomy in adults. The authors conclude that their study supports removal of the tonsils to prevent immediate additional episodes of streptococcal pharyngitis in adults with a history of recurrent streptococcal pharyngitis. They also conclude that tonsillectomy is not only effective but also safe.

The study has several limitations that are important, many of which pointed out by Paul Little in the editorial accompanying the paper in BMJ and in the rapid responses published on bmj.com.⁴⁵ As mentioned above, this study was only able to tell us about streptococcal pharyngitis episodes in a short 90 day follow-up period, and we don't know from this study if removal of the tonsils has a longer effect on preventing streptococcal pharyngitis episodes. For many patients, having to endure a 13 day recovery period with throat pain may not be worth the chance that they have fewer streptococcal pharyngitis episodes in 90 days.

This trial is a step forward on a path that is not yet well paved. We need more evidence about the effectiveness and safety of tonsillectomy, and this study offers a good start. Until more studies provide longer term and more safety data, however, it is difficult to give sound advice to doctors and patients.

Jocalyn Clark, associate editor, BMJ
Email: jclark@bmj.com

Competing interests: None declared.

studentBMJ 2007;15:213-256 June ISSN 0966-6494

1. Hospital Episode Statistics. Trends in ENT admission rates. London: Department of Health. www.hesonline.nhs.uk/ Ease/servlet/ContentServer?siteID=1937&categoryID=410
2. Kvestad E, Kvaerner K, Roysamb E, Tambs K, Harris J, Magnus P. Heritability of recurrent tonsillitis. Arch Otolaryngol Head Neck Surg 2005;131:383-7.
3. Burton MJ, Towler B, Glasziou P. Tonsillectomy versus non-surgical treatment for chronic/recurrent acute tonsillitis. Cochrane Database Syst Rev 2000;(2):CD001802.
4. Little P. Recurrent pharyngo-tonsillitis. BMJ 2007;334:909. doi: 10.1136/bmj.39184.617049.80.
5. Electronic responses. Tonsillectomy versus watchful waiting in recurrent streptococcal pharyngitis in adults: randomised controlled trial. BMJ 2007. www.bmj.com/cgi/eletters/334/7600/909