- Email: [email protected]
Tonsillectomy—clinical consequences twenty years after surgery?
International Journal of Pediatric Otorhinolaryngology (2003) 67, 981 /988
www.elsevier.com/locate/ijporl
*
Tonsillectomy clinical consequences twenty years after surgery? /
Ewa Johansson, Elisabeth Hultcrantz* Department of Neuroscience and Locomotion, Division of Otorhinolaryngology, Linko ¨pings Universitet, SE 58185 Linko ¨ping, Sweden Received 18 January 2003; received in revised form 30 May 2003; accepted 1 June 2003
KEYWORDS Tonsillectomy; Children; Long-term effect of surgery; Chronic disease
Summary Tonsillectomy (T) is one of the most common surgical procedures performed on children. Long-term follow-up studies concerning its consequences are lacking. This study is the first study done on a group of patients that underwent T in their childhood, about 20 years ago. The investigation is a cohort study, which followed-up 18 patients who were tonsillectomized 20 years ago. It was to be determined whether these subjects suffer from more respiratory tract infections (or other infections) today, than people who are not tonsillectomized. A group of 54 agematched subjects were selected for comparison. A questionnaire was mailed to the study population. No significant differences were found between the groups in the frequency of upper respiratory tract infection (URI). The mean number of URI’s was approximately [MSOffice1]2.5 per year in both groups. The duration of the URI’s was identical in each group. A high temperature was present to the same extent in each group. Absence from work, number of visits to physicians and the use of antibiotics were the same in each group. However, the prevalence of chronic disease was greater in the T-group than in the comparison group. The difference was significant with a Relative Risk of 9.41 and a Confidence Interval differing from 1 (1.13 B/RRB/78.14) for the T-population to develop chronic disease. Because of the small number of the present study population, the results must be validated by further immunological and epidemiological studies on long-term effects of tonsillectomy. – 2003 Elsevier Ireland Ltd. All rights reserved.
1. Introduction For the last 100 years, tonsillectomy has been one of the most common surgical procedures, especially on children. The indications for surgery have varied but have, for the majority, been to remove the tonsils in order to prevent recurrent infections or to remove an obstacle to breathing.
*Corresponding author. E-mail address: [email protected] (E. Hultcrantz).
Different methods have been used, but the intention has been to remove as much as possible of the tonsillar tissue to avoid recurrences. If and how such radical surgery on part of the Waldeyer’s ring has influenced the immune system in the growing individual, has come under more and more discussion. In the short run, no negative consequences from the procedure have been noticed. But what happens in the long run? We really do not know very much about the genealogical consequences of tonsillectomies. Do we introduce any changes in the individual’s ability to resist an infectious
0165-5876/03/$ - see front matter – 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0165-5876(03)00196-4
982
disease? Is it possible that the person who is tonsillectomized may be prone to suffer URI’s later in life?
1.1. Background The palatine tonsils, which are lymphoepithelial structures with a center of lymphoid tissue (primarily lymphocytes) are strategically placed between the oral cavity and the throat, around the entrance to both the windpipe and the esophageal opening. They should have an important role in the mucous membranes ‘first line of defense’ for invading microorganisms transported in the air and by food. Antigens are there processed and presented to the immune system and both a local and a systematic response can be mobilized during infectious situations. In Sweden, tonsillectomy is one of the most common surgical procedures on children, with 4612 tonsillectomies performed in 1999 on children under 16 years old. The purpose of this investigation is to consider, whether there are any differences in tendencies for infection between adult individuals who have been tonsillectomized in their childhood compared with people of the same age who have not been operated upon. Another object is to study the prevalence of snoring.
2. Method Eighteen people who were tonsillectomized during childhood were compared with 54 people of the same age and sex who were not operated upon. The study was performed using questionnaires and was approved by the ethics committees at Uppsala University and University of Linko ¨ping, Sweden.
2.1. The T-group The T-group consisted of 18 persons, 7 women and 11 men (median: 25 years; age interval 23 / 30), living in the middle of Sweden. This group of patients was originally a randomized sample from a larger group of children referred to the Ear, Nose and Throat Clinic in Uppsala between the years 1980 /82, to be tonsillectomized due to obstructive problems. All of them were snorers and were mouth breathers during sleep with different degrees of breathing problems. Some had also had problems eating due to their hypertrophied tonsils. One of the inclusion criteria was that they had not had recurrent tonsillitis before the surgery and that they did not have an enlarged adenoid. At the time of surgery they had a median age of 5 years and 11
E. Johansson, E. Hultcrantz
months, age interval 3 years and 5 months to 9.5 years. Nine children were below 5 years, five were 5 /7 years and four 8 /9.5 years. This patient material has been examined in a previous study about bite development after tonsillectomy [9] where examination follow-ups were done after 6 months and again after 2 years. The group consisted originally of 22 children; four that were omitted. One of these had died and two that had moved from Sweden and could not be located and one had already reached puberty at the time for surgery (15 years) and could not be regarded as a child.
2.2. The comparison group The comparison group consisted of 54 subjects, 21 women and 33 men. They were taken from the Swedish population register. Matching three individuals of the same sex, born the same day, to each tonsillectomized person’s birth data did the selection. At this stage, no selection could be done of only those subjects that had not been tonsillectomized. The subjects chosen came from communities and social situation similar to the patients who were in the T group.
2.3. Questionnaire The questionnaire, which consisted of 21 questions, was mailed to each subject together with a letter of information. Two reminders were sent and the people who had been operated upon were also contacted by telephone if they had not answered the letter. The questions included the number of infections during the last 12 months especially for the upper airways and aspects of the person’s life, which might be important for the person’s risk to acquire an infection. ‘Serious respiratory infections’ without time limit were asked for. Questions about personal contacts at work and during free time, contact with children, smoking, allergies, and medication. Most questions could be answered choosing among several alternatives. The possibility was also given to answer open-ended questions about their own experience of the frequency and severity with which they had upper airway infections.
2.4. Statistics The risk quota was calculated using the program Epi Info [10]. Epi Info is widely used in epidemiological research. Relative risks, confidence intervals, and stratified analysis were calculated from a 2 /2 table.
Tonsillectomy */clinical consequences twenty years after surgery? The confidence interval is 95% and calculated according to the Taylor Series for Relative Risk (RR). P -values are Fisher exact two-sided for calculations that include small numbers ( B/5). Stratified analysis for adjustment of possible confounding is shown with Mentel/Haenszels RR and Greenland/Robins confidence interval.
3. Results The reply frequency in the T-group was 94 (17/ 18) and 83% (45/54) in the comparison group. Five people were already tonsillectomized in the comparison group (one of them at the age of 3.5 years) and were therefore omitted. The resulting reply frequency was 80% (40/50). There was a small overrepresentation of men in each group. Demographic data are presented in Table 1. The questionnaire had 2 /7 possible answers for each question. In cases where the answers were divided evenly, the number in each group could be too few and were therefore added into larger groups by compiling the answer alternatives. For example, the question about sick leave, with 7 answer alternatives from 0 to /4 weeks, was shown as two alternatives: B/2 and /2 weeks.
3.1. Contagion sources Regarding risk for contagion due to working with children, there was no difference between the groups in the extent to which they worked within the day-care system. One-fifth of the subjects in every group had
983
children at home, constantly or just on weekends. No significant increased risk for frequent colds ( / 4/year) was found among individuals that had children or had contact with children. The amount of contacts with others during the day: at work, to and from work, and during leisure time, was equal in both groups. The majority of respondents had most contact with others (/20 persons) at work, fewer ( B/7 persons) during transportation to and from work, and 7 /20 contacts with others during leisure time. No significant increased risk for upper respiratory tract infections existed among those subjects that had many contacts with others (see Table 2).
3.2. Potential confounding factors Smoking was found equally in both groups: 17% in the T-group were smokers, all women. In the comparison group there were four men and three women smoking or 17%. A certain, though not significant, correlation could be seen between smoking and an increased number of URI’s in this age group. During stratification for smoking there was an increased risk, though it was not significant. Allergies (with other symptoms than asthma) were less frequent in the T-group than in the comparison group. This difference was not significant. On the other hand, the asthma symptoms with allergies were more common in the T-group than in the comparison group. Once again there was no significant difference between the groups in this respect. There was no significant increased risk for URI’s among the individuals who had allergies or asthma.
3.3. Chronic disease
Table 1 Demographic data
T n/17
Controls n/40
Women Men Occupation (-99,-00) Employed Student Parental leave Employed/student/ parental leave Other (sick, unemployed) Children in the home
7 (41.2%) 10 (58.8%)
17 (42.5%) 23 (57.5%)
9 3 3 2
20 (50.0%) 4 (10.0%) 2 (5.0%) 11 (27.5%)
/ 3 (17.6%)
3 (7.5%) 8 (20.0%)
Number of daily person contacts At work /20 To and from work B/7 During free time 7 /20
8 (47.1%) 12 (70.6%) 10 (58.8%)
26 (65.0%) 30 (75.0%) 20 (50.0%)
(52.9%) (17.6%) (17.6%) (11.8%)
Chronic illness occurred with 4/17 (23.5%) of the subjects in the T-group and with only 1/40 (2.5%) in the comparison group. Increased disposition for acute infections could not be shown with these individuals.The chronic diseases in the T-group were asthma (1), ulcerative colitis (1), goiter (1), and arterial hypertension (1). In the comparison group, only one case of asthma occurred. All of these subjects used medication daily for their illnesses. The relationship between T and chronic disease had a clear significance with a relative risk at 9.41 and the confidence interval distinguished from 1 (see Table 3).
3.4. Infections (URI) The number of URI the last 12 months was on average 2.5 in both groups. The proportion that
984
E. Johansson, E. Hultcrantz
Table 2 Contacts with children at home or many contacts at work versus increased URI-risk /4 URI/12 months
Children at home Contacts at work /20
T
Controls
2 2
1 5
had more than 4 colds was 4/17 (24%) in the Tgroup and 6/40 (15%) in the comparison group. The difference suggests a certain risk increase for the tonsillectomized to be affected by URI’s, but it is not significant. The age at tonsillectomy for those more severely affected were B/4 years for one and /4 years for 3 (see Table 4). For approximately 77% of the subjects, the duration of the URI period 1 week or less in each group. For 17% the URI lasted 2 weeks, and for 5%, 3 weeks in each group. Four subjects in the T-group experienced more serious respiratory tract infections (RI) during childhood, split between pneumonia (3), and epiglottiditis (1). Five subjects in the comparison group had had serious respiratory tract infections, divided between pneumonia (3), bronchitis (1) and laryngitis (1). No sign increased risk for the tonsillectomized was shown in this case. There was no significant difference between the two groups in the extent to which they developed a high temperature during URI. There were no major differences in the descriptions of symptoms between the groups. The symptoms from the nose included stuffed nose, and sneezing. The throat symptoms were swelling, sore throat and mucus. Common symptoms included tiredness, chills, headaches, nausea, and body pains. Other symptoms were runny and itchy eyes, itchy ears and stuffed ears (‘ear wax’), plus aches in the sinuses. The question on symptoms provided no predefined answer alternatives, therefore no risk calculation was done.
3.5. Consequences of infections About half of the T-group had taken sick leave due to URI’s during the last 2 years compared to
RR
Confidence-interval
4.71 1.18
0.46B/RRB/48.49 0.36B/RRB/4.64
65% of the comparison group. Sick leave lasting longer than 2 weeks occurred more frequently in the T-group, but the increased risk was not significant. The periods of use of antibiotics during the last 2 years were also the same. Almost half of the members of each of the groups had used penicillin V or another antibiotic. During these periods, most admitted they had had an URI, mainly a throat infection that was treated with antibiotics. There was a slightly increased risk for the tonsillectomized group, but the difference was not significant. Other infections that were treated with antibiotics were UVI, gynecological infections, and tooth infections. Only one person (6%) in the T-group and 7 (17.5%) in the comparison group had been given antibiotics for other than URI’s. Approximately 1/5, in both groups, had sought a doctor’s help because of a URI during the last 2 years. The personal experience (feeling) of having more URI’s or having worse sore throats than other people around them was the same in both groups. Twelve percent in the T-group had that feeling and 15% in the comparison group.
3.6. Snoring With regard to snoring, which for the individuals in the T-group was part of the original reason for the surgery, it occurred ‘now and then’ for 53% and two reported that they were snoring every night. This can be compared with an occurrence of 42.5% in the comparison group, in which all said that they just snore ‘now and then’. An increased risk was found for those tonsillectomized, though it was not significant (see Table 5). These results shall be investigated further.
Table 3 Chronic disease, daily medication T Age at T
Occurrence of chronic disease
B/5
/5
0
4 (23.5%)
Controls
RR
Confidence-interval
P
1 (2.5%)
9.41
1.13 B/RR B/78.14
0.024
Tonsillectomy */clinical consequences twenty years after surgery?
985
Table 4 URI frequency the last 12 months
/4 colds URI in average
T
Controls
RR
Confidence-interval
4 (23.5%) 2.65
6 (15%) 2.42
1.28 /
0.40B/RRB/4.01 /
4. Discussion This investigation follows up a number of individuals 20 years after they as children went through a tonsillectomy. These children had originally problems exclusively of obstructive nature (snoring, sleep apnea) and not of recurrent tonsillitis. The fact that the recurring infections were not a problem for these individuals possibly makes it easier to study the immunological importance of the tonsils. The reply frequency was good in both of the groups. The omissions in the comparison group were nine people, which constitute 17% of the material, not large enough that an omission analysis was needed. Sex was not shown to have influence on infection disposition in this study. An examination at a health center close to the area where the tonsillectomized group lived in 1998, showed that among the applicants, the occurrence of URI’s was higher among women 15 /44 years old than among men of the same age, with 7.9% the women and 4.6% for the men [11]. However, in that case it concerned individuals that actively sought the health center, in contrast with the present study that examined epidemiological occurrences of URI’s. An analysis was done to see if ‘exposure to children’ resulted in more URI’s, but no difference in prevalence was found between the groups. The number of children at home every week was found at the same rate with the tonsillectomized and the reference group. This indicates that the subjects in these groups were exposed to ‘day-care exposure’ at the same rate and had given the same result concerning URI’s. That should argue in favor of tonsillectomies not effecting the immune system’s capability to fight off URI’s to any significant degree. Smoking must always be taken into consideration, as a confounding factor that could worsen an
infection or lessen the person’s defense against infections. There was no difference in smoking between the groups; 17.5% of each group smoked. The equivalent number of daily smokers for the adult population (16 /84 years) in Sweden, year 2001, was 19%. (Source: ULF, the Swedish Statistical Central Bureau.) Thus, the occurrence of smoking in the groups was lower than in the background population and these young smokers had not increased frequency of URI’s. That allergies would worsen infections was not supported by results in either of the groups. Regarding the allergy sufferers and the asthmatics, it has been shown that symptoms and medicine use lessen greatly after tonsillectomies [12]. That study was done a short time after the operation and lacks long-term follow-up. In the present study, the children had no allergy symptoms before the surgery; it is therefore difficult to make a direct comparison. The number and the length of URI periods were the same in both of the groups. The tonsillectomized showed no significantly higher occurrences of serious respiratory tract infections during childhood than those in the comparison group. These results indicate that a T should be of little consequence for these people’s tendencies to become ill with an URI 20 years after the operation. The palatine tonsils are regarded as playing an important role in the normal development and function of the immune system [1 /3]. When a child is born, he or she is essentially agammaglobulinemic. The production of Ig G and Ig M starts. However, the Ig A production is not completely developed until school age and for that development the tonsils are necessary. When the child grows, especially during 4 /8 years of age, there is a temporary physiological hypertrophy of the tonsils. During that phase, many authors mean that the tonsils have an important task in fighting
Table 5 Frequency of snoring
Now and then Every night Total
T
Controls
RR
Confidence-interval
9 (52.9%) 2 (11.8%) 11 (64.7%)
17 (42.5%) / 17 (42.5%)
1.52
0.92 B/RRB/2.52
986
throat infections (‘large tonsils /busy tonsils’) before certain resistance for infections develops in 8 /10 years of age [4,5]. For that reason some authors advise that tonsils should not be removed before 8 years of age unless you have absolute indications for surgery. A tonsillectomy in early childhood might have negative effects on the local immunological response in the same way removal of MALT (mucous associated lymphoid tissue) has on rodents and might have generally negative consequences for the body’s ability to rid the airways of infections [2,3]. In our study there was a slight increased proportion of individuals who experienced more than four URIs during the last year in the tonsillectomized group, compared to the controls (see Table 4). Those were operated at the age of 3.5, 6, 8.5 and 9.5 years, thus evenly spread within the examined age span. In our small study-group, we could not find that the risk for increased number of infection was higher for the youngest children. Another opinion about the tonsils’ function is that they can be a focus of infection that harbors the pathogens. According to that opinion, recurrent tonsillitis has a blocking effect on the immune system, e.g. an immunosuppressive effect, which makes it more difficult to get rid of infections [6]. This immunosuppressive effect could be reduced by tonsillectomy and that is why tonsillectomy reduces the patient’s visits to the doctor, the use of antibiotics, and of course, episodes with sore throats [7,8]. The well known study by Paradise et al. [13], on the long-term effect of tonsillectomy on throat infections, showed a smaller frequency of URI’s in the tonsillectomized up to 2 years after the T. However, in the third year there were no differences in the infection frequency between the two groups. At this point both groups had a lower infection frequency. In the present study, that deals with an initially healthier patient material, a similar result was seen, i.e. no difference in the mean number of infection 20 years after a T compared with individuals that had their tonsils intact. Snoring was reported more often in the T-group than in the reference group. The operation seems not to have had a lasting effect for all subjects. This is the first follow-up study of patients as much as 20 years after a T because of obstruction. Many studies have shown good effects of T’s for almost all children with OSAS problems [19 /21,24]. Yet, they have seldom been followed up for more than at most, a few years. In this study, the majority (65%) of the T-group had experience at least partial recurrence of the symptoms. More research will be
E. Johansson, E. Hultcrantz
done with these individuals concerning breathing patterns and bite development. A new finding in the present study was the significantly increased occurrence of chronic disease in the tonsillectomized group: 23.5% in the Tgroup compared with 2.5% in the comparison group, RR /9.41 (1.13 B/RR B/78.14) P /0.024. These diseases were not apparent at the time of the surgery, based on medical records from the time of the patients’ original treatment. To be mentioned is also that one of the four persons in the control group, omitted due to earlier tonsillectomy, had answered that she was operated at the age of 3.5 years and that she to day suffered from ‘chronic pain in joints and muscles’. The illnesses asthma, goiter and ulcerative colitis, and high blood pressure which were represented in the T-group, in many cases have an immunological background. Therefore the question arises whether there could be a connection between tonsillectomies and later development of immunological diseases? The children who later developed a chronic disease did not belong to the youngest age-group, B/5 years, at the time for surgery, they were all between 5 and 9 years. There are several studies [1,6,14 /16] that show that immunoglobulins in serum and saliva decrease directly after a tonsillectomy. Bo ¨ck et al. [17] carried out a study in 1994, of different infection variables 6.5 years after T in children. They found no increase of infection frequency, but they did find changes in lymphocytic structure and a decrease of Ig A. However, they would not assign any clinical relevance to this. They concluded that the changes were an effect of decreased antigen load after a T, and therefore decreased Ig-production. Other studies have mainly pointed out that the concentrations of Ig A has decreased and remained low after the surgery. Serum Ig A and saliva Ig A have been lower compared with the time before the operation [6,14,15,17]. However, the concentration has never been under normal reference values and, therefore has not been as low as with Ig A hypogamma-globulinemia. It has been stressed that the decrease does not have any clinical significance [1,3,15]. If one draws a parallel to the patients with Ig A deficiency, it has been found that these individuals do not necessarily have more infections than others. On the other hand, it has been seen that people with Ig A deficiency have an increased risk for developing autoimmune diseases [18]. That tends to support the findings in the present study, although these are based only on questionnaires and not on laboratory data. In the present study, the basis for the differences in the occurrence of chronic disease is not clear:
Tonsillectomy */clinical consequences twenty years after surgery? can T have such an effect on the immune system that the individual runs a risk to develop a disease with an immune origin? Or is it that these people have the predisposition to suffer both from tonsil hypertrophy and chronic diseases? It would be helpful in answering these questions if one could get information about these subject’s backgrounds */for example on heredity for immune diseases */but since they were promised that data would not be analyzed individually, this is not possible ethically. The question is sufficiently interesting that a larger study should be done analyzing the occurrence of tonsillectomy in people with different chronic and autoimmune diseases. A T should be done on strict indications with thorough examinations where the pros and cons for the surgery are weighed carefully [19,22,23]. Individual therapeutic solutions can be recommended. For example, an alternative to a T is tonsillotomy in which only the obstructive part of the tonsil is removed [25] for cases with mostly obstructive problems. Another alternative, which has been tried, is long time antibiotic prophylaxis for recurrent tonsillitis [5,22]. Both alternatives are intended to decrease the potential negative effects on the immune system of the surgery.
5. Conclusion Tonsillectomies in childhood of individuals that have not shown increased disposition for infections before the surgery do not seem to give more infections 20 years later in comparison with nonoperated individuals in the same age group. On the other hand, an increased occurrence of chronic, immune-mediated diseases was observed in the operated group. This finding argues for extended epidemiological and immunological research within the area.
References [1] R. Donovan, J.F. Soothill, Immunological studies in children undergoing tonsillectomy, Clin. Exp. Immunol. 14 (1973) 347 /357. [2] C. Frieke Kuper, P.J. Koornstra, D.M.H. Hameleers, J. Bbiewenga, B.J. Spit, A.M. Duijvestijn, P.J.C. van Breda Vriesman, T. Sminia, The role of nasopharyngeal lymphoid tissue, Immunol. Today 13 (6) (1992) 219 /224. [3] C. Paulussen, J. Claes, G. Claes, M. Jorissen, Adenoids and tonsils, indications for surgery and immunological consequences of surgery, Acta Oto-Rhino-Laryngol. Belg. 54 (2000) 403 /408. [4] L.P. Gray, The T’s and A’s problem */assessment and reassessment, J. Laryngol. Otolog. 91 (1977) 11.
987
[5] Ming-de Ying, Immunological basis of indications for tonsillectomy and adenoidectomy, Acta Otolaryngol. (Stockholm) Suppl. 454 (1998) 279 /285. [6] M. Sainz, F. Guterrez, P.M. Moreno, C. Mun ˜oz, M. Ciges, Changes in immunologic response in tonsillectomized children. I. Immunosuppression in recurrent tonsillitis, Clin. Otolaryngol. 17 (1992) 376 / 379. [7] S. Mui, B.M. Rasgon, R.L. Hilsinger, Efficacy of tonsillectomy for recurrent throat infection in adults, Laryngoscope 108 (1998) 1325 /1328. [8] A.K. Soreide, J. Olofsson, (Adeno)tonsillektomi hos barn, Tidsskr. Nor. Laegeforen. 115 (3) (1995) 352 / 354. [9] E. Hultcrantz, M. Larsson, R. Hellquist, J. AhlquistRastad, H. Svanholm, O.P. Jakobsson, The influence of tonsillar obstruction and tonsillectomy on facial growth and dental arch morphology, Int. J. Pediatr. Otorhinolaryngol. 22 (1991) 125 / 134. [10] CDC */Centers for Disease Control and Prevention, Epi Info 6 for DOS version 6.04d; http://www.cdc.gov/epiinfo. [11] J. Cederholm, M. Korpela, Diagnosredovisning fo la a 1998 vid va i ¨r ¨karbeso ¨k ˚r ˚rdcentralen Tierp; http://www.sos.se/fulltext/0046-001/S7_87.PDF. [12] H. Saito, K. Asakura, M. Hata, A. Kataura, K. Morimoto, Does adenotonsillectomy affect the course of bronchial asthma and nasal allergy, Acta Otolaryngol. (Stockholm) 523 (Suppl.) (1996) 212 / 215. [13] J.L. Paradise, C.D. Bluestone, et al., Efficacy of tonsillectomy for recurrent throat infection in severely affected children, N. Engl. J. Med. 310 (1984) 674 /683. [14] A. Cantani, P. Bellioni, F. Salvinelli, L. Businco, Ann. Allergy 57 (1986) 413 /416. [15] B.E. Del Rio-Navarro, S. Torres, L. Barraga ´n-Tame, C. De la Torre, A. Berber, G. Acosta, J.J.L. Sienra-Monge, Immunological effects of tonsillectomy/adenectomy in children, Adv. Mucosal Immunol. (1995) 737 /740. [16] A. Morag, P.L. Ogra, Immunologic aspect of tonsils, Ann. Otology Rhinology Laryngol. (1975) Suppl. 19, vol. 84 (2), 37 /43. [17] A. Bo ¨ck, W. Popp, K.R. Herkner, Tonsillectomy and the immune system: a long-term follow up comparison between tonsillectomized and nontonsillectomized children, Eur. Arch. Otorhinolaryngol. 251 (1994) 423 / 427. [18] J.T. Cassidy, R.E. Petty, Textbook of Pediatric Rheumatology, 3rd ed., pp. 472 /473. [19] J. Ahlquist-Rastad, E. Hultcrantz, H. Svanholm, Children with tonsillar obstruction: indication for and efficacy of tonsillectomy, Acta Paed. Scand. 77 (1988) 831 /835. [20] F. Kudoh, A. Sanai, Effect of tonsillectomy and adenoidectomy on obese children with sleep-associated breathing disorders, Acta Otolaryngol. (Stockholm 1996), Suppl. 523: 216 /218. [21] M. Sato, M. Suzuki, et al., Long-term follow-up of obstructive sleep apnea syndrome following surgery in children and adults, Tohoku J. Expt. Med. 192 (3) (2000) 165 / 172. [22] T. Marshall, A review of tonsillectomy for recurrent throat infection, Br. J. General. Practice 48 (1998) 1331 / 1335.
988 [23] B. Lo ¨fstrand-Tidestro ¨m, B. Thilander, J. Ahlqvist-Rastad, O. Jakobsson, E. Hultcrantz, Breathing obstruction in relation to craniofacial and dental arch morphology n 4-year-old children, Eur. J. Orthodontics. 21 (1999) 323 /332. [24] T. Nishimura, N. Morishima, S. Hasegawa, N. Shibata, K. Iwanaa, M. Yagisawa, Effect of surgery on obstructive sleep
E. Johansson, E. Hultcrantz apnea, Acta Otolaryngol. (Stockholm) Suppl. 523 (1996) 231 /233. [25] E. Hultcrantz, A. Linder, A. Markstro ¨m, Tonsillectomy or tonsillotomy? */a randomized study comparing postoperative pain and long-term effects, Int. J. Pediatr. Otorhinolaryngol. 51 (1999) 171 / 176.
www.elsevier.com/locate/ijporl
*
Tonsillectomy clinical consequences twenty years after surgery? /
Ewa Johansson, Elisabeth Hultcrantz* Department of Neuroscience and Locomotion, Division of Otorhinolaryngology, Linko ¨pings Universitet, SE 58185 Linko ¨ping, Sweden Received 18 January 2003; received in revised form 30 May 2003; accepted 1 June 2003
KEYWORDS Tonsillectomy; Children; Long-term effect of surgery; Chronic disease
Summary Tonsillectomy (T) is one of the most common surgical procedures performed on children. Long-term follow-up studies concerning its consequences are lacking. This study is the first study done on a group of patients that underwent T in their childhood, about 20 years ago. The investigation is a cohort study, which followed-up 18 patients who were tonsillectomized 20 years ago. It was to be determined whether these subjects suffer from more respiratory tract infections (or other infections) today, than people who are not tonsillectomized. A group of 54 agematched subjects were selected for comparison. A questionnaire was mailed to the study population. No significant differences were found between the groups in the frequency of upper respiratory tract infection (URI). The mean number of URI’s was approximately [MSOffice1]2.5 per year in both groups. The duration of the URI’s was identical in each group. A high temperature was present to the same extent in each group. Absence from work, number of visits to physicians and the use of antibiotics were the same in each group. However, the prevalence of chronic disease was greater in the T-group than in the comparison group. The difference was significant with a Relative Risk of 9.41 and a Confidence Interval differing from 1 (1.13 B/RRB/78.14) for the T-population to develop chronic disease. Because of the small number of the present study population, the results must be validated by further immunological and epidemiological studies on long-term effects of tonsillectomy. – 2003 Elsevier Ireland Ltd. All rights reserved.
1. Introduction For the last 100 years, tonsillectomy has been one of the most common surgical procedures, especially on children. The indications for surgery have varied but have, for the majority, been to remove the tonsils in order to prevent recurrent infections or to remove an obstacle to breathing.
*Corresponding author. E-mail address: [email protected] (E. Hultcrantz).
Different methods have been used, but the intention has been to remove as much as possible of the tonsillar tissue to avoid recurrences. If and how such radical surgery on part of the Waldeyer’s ring has influenced the immune system in the growing individual, has come under more and more discussion. In the short run, no negative consequences from the procedure have been noticed. But what happens in the long run? We really do not know very much about the genealogical consequences of tonsillectomies. Do we introduce any changes in the individual’s ability to resist an infectious
0165-5876/03/$ - see front matter – 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/S0165-5876(03)00196-4
982
disease? Is it possible that the person who is tonsillectomized may be prone to suffer URI’s later in life?
1.1. Background The palatine tonsils, which are lymphoepithelial structures with a center of lymphoid tissue (primarily lymphocytes) are strategically placed between the oral cavity and the throat, around the entrance to both the windpipe and the esophageal opening. They should have an important role in the mucous membranes ‘first line of defense’ for invading microorganisms transported in the air and by food. Antigens are there processed and presented to the immune system and both a local and a systematic response can be mobilized during infectious situations. In Sweden, tonsillectomy is one of the most common surgical procedures on children, with 4612 tonsillectomies performed in 1999 on children under 16 years old. The purpose of this investigation is to consider, whether there are any differences in tendencies for infection between adult individuals who have been tonsillectomized in their childhood compared with people of the same age who have not been operated upon. Another object is to study the prevalence of snoring.
2. Method Eighteen people who were tonsillectomized during childhood were compared with 54 people of the same age and sex who were not operated upon. The study was performed using questionnaires and was approved by the ethics committees at Uppsala University and University of Linko ¨ping, Sweden.
2.1. The T-group The T-group consisted of 18 persons, 7 women and 11 men (median: 25 years; age interval 23 / 30), living in the middle of Sweden. This group of patients was originally a randomized sample from a larger group of children referred to the Ear, Nose and Throat Clinic in Uppsala between the years 1980 /82, to be tonsillectomized due to obstructive problems. All of them were snorers and were mouth breathers during sleep with different degrees of breathing problems. Some had also had problems eating due to their hypertrophied tonsils. One of the inclusion criteria was that they had not had recurrent tonsillitis before the surgery and that they did not have an enlarged adenoid. At the time of surgery they had a median age of 5 years and 11
E. Johansson, E. Hultcrantz
months, age interval 3 years and 5 months to 9.5 years. Nine children were below 5 years, five were 5 /7 years and four 8 /9.5 years. This patient material has been examined in a previous study about bite development after tonsillectomy [9] where examination follow-ups were done after 6 months and again after 2 years. The group consisted originally of 22 children; four that were omitted. One of these had died and two that had moved from Sweden and could not be located and one had already reached puberty at the time for surgery (15 years) and could not be regarded as a child.
2.2. The comparison group The comparison group consisted of 54 subjects, 21 women and 33 men. They were taken from the Swedish population register. Matching three individuals of the same sex, born the same day, to each tonsillectomized person’s birth data did the selection. At this stage, no selection could be done of only those subjects that had not been tonsillectomized. The subjects chosen came from communities and social situation similar to the patients who were in the T group.
2.3. Questionnaire The questionnaire, which consisted of 21 questions, was mailed to each subject together with a letter of information. Two reminders were sent and the people who had been operated upon were also contacted by telephone if they had not answered the letter. The questions included the number of infections during the last 12 months especially for the upper airways and aspects of the person’s life, which might be important for the person’s risk to acquire an infection. ‘Serious respiratory infections’ without time limit were asked for. Questions about personal contacts at work and during free time, contact with children, smoking, allergies, and medication. Most questions could be answered choosing among several alternatives. The possibility was also given to answer open-ended questions about their own experience of the frequency and severity with which they had upper airway infections.
2.4. Statistics The risk quota was calculated using the program Epi Info [10]. Epi Info is widely used in epidemiological research. Relative risks, confidence intervals, and stratified analysis were calculated from a 2 /2 table.
Tonsillectomy */clinical consequences twenty years after surgery? The confidence interval is 95% and calculated according to the Taylor Series for Relative Risk (RR). P -values are Fisher exact two-sided for calculations that include small numbers ( B/5). Stratified analysis for adjustment of possible confounding is shown with Mentel/Haenszels RR and Greenland/Robins confidence interval.
3. Results The reply frequency in the T-group was 94 (17/ 18) and 83% (45/54) in the comparison group. Five people were already tonsillectomized in the comparison group (one of them at the age of 3.5 years) and were therefore omitted. The resulting reply frequency was 80% (40/50). There was a small overrepresentation of men in each group. Demographic data are presented in Table 1. The questionnaire had 2 /7 possible answers for each question. In cases where the answers were divided evenly, the number in each group could be too few and were therefore added into larger groups by compiling the answer alternatives. For example, the question about sick leave, with 7 answer alternatives from 0 to /4 weeks, was shown as two alternatives: B/2 and /2 weeks.
3.1. Contagion sources Regarding risk for contagion due to working with children, there was no difference between the groups in the extent to which they worked within the day-care system. One-fifth of the subjects in every group had
983
children at home, constantly or just on weekends. No significant increased risk for frequent colds ( / 4/year) was found among individuals that had children or had contact with children. The amount of contacts with others during the day: at work, to and from work, and during leisure time, was equal in both groups. The majority of respondents had most contact with others (/20 persons) at work, fewer ( B/7 persons) during transportation to and from work, and 7 /20 contacts with others during leisure time. No significant increased risk for upper respiratory tract infections existed among those subjects that had many contacts with others (see Table 2).
3.2. Potential confounding factors Smoking was found equally in both groups: 17% in the T-group were smokers, all women. In the comparison group there were four men and three women smoking or 17%. A certain, though not significant, correlation could be seen between smoking and an increased number of URI’s in this age group. During stratification for smoking there was an increased risk, though it was not significant. Allergies (with other symptoms than asthma) were less frequent in the T-group than in the comparison group. This difference was not significant. On the other hand, the asthma symptoms with allergies were more common in the T-group than in the comparison group. Once again there was no significant difference between the groups in this respect. There was no significant increased risk for URI’s among the individuals who had allergies or asthma.
3.3. Chronic disease
Table 1 Demographic data
T n/17
Controls n/40
Women Men Occupation (-99,-00) Employed Student Parental leave Employed/student/ parental leave Other (sick, unemployed) Children in the home
7 (41.2%) 10 (58.8%)
17 (42.5%) 23 (57.5%)
9 3 3 2
20 (50.0%) 4 (10.0%) 2 (5.0%) 11 (27.5%)
/ 3 (17.6%)
3 (7.5%) 8 (20.0%)
Number of daily person contacts At work /20 To and from work B/7 During free time 7 /20
8 (47.1%) 12 (70.6%) 10 (58.8%)
26 (65.0%) 30 (75.0%) 20 (50.0%)
(52.9%) (17.6%) (17.6%) (11.8%)
Chronic illness occurred with 4/17 (23.5%) of the subjects in the T-group and with only 1/40 (2.5%) in the comparison group. Increased disposition for acute infections could not be shown with these individuals.The chronic diseases in the T-group were asthma (1), ulcerative colitis (1), goiter (1), and arterial hypertension (1). In the comparison group, only one case of asthma occurred. All of these subjects used medication daily for their illnesses. The relationship between T and chronic disease had a clear significance with a relative risk at 9.41 and the confidence interval distinguished from 1 (see Table 3).
3.4. Infections (URI) The number of URI the last 12 months was on average 2.5 in both groups. The proportion that
984
E. Johansson, E. Hultcrantz
Table 2 Contacts with children at home or many contacts at work versus increased URI-risk /4 URI/12 months
Children at home Contacts at work /20
T
Controls
2 2
1 5
had more than 4 colds was 4/17 (24%) in the Tgroup and 6/40 (15%) in the comparison group. The difference suggests a certain risk increase for the tonsillectomized to be affected by URI’s, but it is not significant. The age at tonsillectomy for those more severely affected were B/4 years for one and /4 years for 3 (see Table 4). For approximately 77% of the subjects, the duration of the URI period 1 week or less in each group. For 17% the URI lasted 2 weeks, and for 5%, 3 weeks in each group. Four subjects in the T-group experienced more serious respiratory tract infections (RI) during childhood, split between pneumonia (3), and epiglottiditis (1). Five subjects in the comparison group had had serious respiratory tract infections, divided between pneumonia (3), bronchitis (1) and laryngitis (1). No sign increased risk for the tonsillectomized was shown in this case. There was no significant difference between the two groups in the extent to which they developed a high temperature during URI. There were no major differences in the descriptions of symptoms between the groups. The symptoms from the nose included stuffed nose, and sneezing. The throat symptoms were swelling, sore throat and mucus. Common symptoms included tiredness, chills, headaches, nausea, and body pains. Other symptoms were runny and itchy eyes, itchy ears and stuffed ears (‘ear wax’), plus aches in the sinuses. The question on symptoms provided no predefined answer alternatives, therefore no risk calculation was done.
3.5. Consequences of infections About half of the T-group had taken sick leave due to URI’s during the last 2 years compared to
RR
Confidence-interval
4.71 1.18
0.46B/RRB/48.49 0.36B/RRB/4.64
65% of the comparison group. Sick leave lasting longer than 2 weeks occurred more frequently in the T-group, but the increased risk was not significant. The periods of use of antibiotics during the last 2 years were also the same. Almost half of the members of each of the groups had used penicillin V or another antibiotic. During these periods, most admitted they had had an URI, mainly a throat infection that was treated with antibiotics. There was a slightly increased risk for the tonsillectomized group, but the difference was not significant. Other infections that were treated with antibiotics were UVI, gynecological infections, and tooth infections. Only one person (6%) in the T-group and 7 (17.5%) in the comparison group had been given antibiotics for other than URI’s. Approximately 1/5, in both groups, had sought a doctor’s help because of a URI during the last 2 years. The personal experience (feeling) of having more URI’s or having worse sore throats than other people around them was the same in both groups. Twelve percent in the T-group had that feeling and 15% in the comparison group.
3.6. Snoring With regard to snoring, which for the individuals in the T-group was part of the original reason for the surgery, it occurred ‘now and then’ for 53% and two reported that they were snoring every night. This can be compared with an occurrence of 42.5% in the comparison group, in which all said that they just snore ‘now and then’. An increased risk was found for those tonsillectomized, though it was not significant (see Table 5). These results shall be investigated further.
Table 3 Chronic disease, daily medication T Age at T
Occurrence of chronic disease
B/5
/5
0
4 (23.5%)
Controls
RR
Confidence-interval
P
1 (2.5%)
9.41
1.13 B/RR B/78.14
0.024
Tonsillectomy */clinical consequences twenty years after surgery?
985
Table 4 URI frequency the last 12 months
/4 colds URI in average
T
Controls
RR
Confidence-interval
4 (23.5%) 2.65
6 (15%) 2.42
1.28 /
0.40B/RRB/4.01 /
4. Discussion This investigation follows up a number of individuals 20 years after they as children went through a tonsillectomy. These children had originally problems exclusively of obstructive nature (snoring, sleep apnea) and not of recurrent tonsillitis. The fact that the recurring infections were not a problem for these individuals possibly makes it easier to study the immunological importance of the tonsils. The reply frequency was good in both of the groups. The omissions in the comparison group were nine people, which constitute 17% of the material, not large enough that an omission analysis was needed. Sex was not shown to have influence on infection disposition in this study. An examination at a health center close to the area where the tonsillectomized group lived in 1998, showed that among the applicants, the occurrence of URI’s was higher among women 15 /44 years old than among men of the same age, with 7.9% the women and 4.6% for the men [11]. However, in that case it concerned individuals that actively sought the health center, in contrast with the present study that examined epidemiological occurrences of URI’s. An analysis was done to see if ‘exposure to children’ resulted in more URI’s, but no difference in prevalence was found between the groups. The number of children at home every week was found at the same rate with the tonsillectomized and the reference group. This indicates that the subjects in these groups were exposed to ‘day-care exposure’ at the same rate and had given the same result concerning URI’s. That should argue in favor of tonsillectomies not effecting the immune system’s capability to fight off URI’s to any significant degree. Smoking must always be taken into consideration, as a confounding factor that could worsen an
infection or lessen the person’s defense against infections. There was no difference in smoking between the groups; 17.5% of each group smoked. The equivalent number of daily smokers for the adult population (16 /84 years) in Sweden, year 2001, was 19%. (Source: ULF, the Swedish Statistical Central Bureau.) Thus, the occurrence of smoking in the groups was lower than in the background population and these young smokers had not increased frequency of URI’s. That allergies would worsen infections was not supported by results in either of the groups. Regarding the allergy sufferers and the asthmatics, it has been shown that symptoms and medicine use lessen greatly after tonsillectomies [12]. That study was done a short time after the operation and lacks long-term follow-up. In the present study, the children had no allergy symptoms before the surgery; it is therefore difficult to make a direct comparison. The number and the length of URI periods were the same in both of the groups. The tonsillectomized showed no significantly higher occurrences of serious respiratory tract infections during childhood than those in the comparison group. These results indicate that a T should be of little consequence for these people’s tendencies to become ill with an URI 20 years after the operation. The palatine tonsils are regarded as playing an important role in the normal development and function of the immune system [1 /3]. When a child is born, he or she is essentially agammaglobulinemic. The production of Ig G and Ig M starts. However, the Ig A production is not completely developed until school age and for that development the tonsils are necessary. When the child grows, especially during 4 /8 years of age, there is a temporary physiological hypertrophy of the tonsils. During that phase, many authors mean that the tonsils have an important task in fighting
Table 5 Frequency of snoring
Now and then Every night Total
T
Controls
RR
Confidence-interval
9 (52.9%) 2 (11.8%) 11 (64.7%)
17 (42.5%) / 17 (42.5%)
1.52
0.92 B/RRB/2.52
986
throat infections (‘large tonsils /busy tonsils’) before certain resistance for infections develops in 8 /10 years of age [4,5]. For that reason some authors advise that tonsils should not be removed before 8 years of age unless you have absolute indications for surgery. A tonsillectomy in early childhood might have negative effects on the local immunological response in the same way removal of MALT (mucous associated lymphoid tissue) has on rodents and might have generally negative consequences for the body’s ability to rid the airways of infections [2,3]. In our study there was a slight increased proportion of individuals who experienced more than four URIs during the last year in the tonsillectomized group, compared to the controls (see Table 4). Those were operated at the age of 3.5, 6, 8.5 and 9.5 years, thus evenly spread within the examined age span. In our small study-group, we could not find that the risk for increased number of infection was higher for the youngest children. Another opinion about the tonsils’ function is that they can be a focus of infection that harbors the pathogens. According to that opinion, recurrent tonsillitis has a blocking effect on the immune system, e.g. an immunosuppressive effect, which makes it more difficult to get rid of infections [6]. This immunosuppressive effect could be reduced by tonsillectomy and that is why tonsillectomy reduces the patient’s visits to the doctor, the use of antibiotics, and of course, episodes with sore throats [7,8]. The well known study by Paradise et al. [13], on the long-term effect of tonsillectomy on throat infections, showed a smaller frequency of URI’s in the tonsillectomized up to 2 years after the T. However, in the third year there were no differences in the infection frequency between the two groups. At this point both groups had a lower infection frequency. In the present study, that deals with an initially healthier patient material, a similar result was seen, i.e. no difference in the mean number of infection 20 years after a T compared with individuals that had their tonsils intact. Snoring was reported more often in the T-group than in the reference group. The operation seems not to have had a lasting effect for all subjects. This is the first follow-up study of patients as much as 20 years after a T because of obstruction. Many studies have shown good effects of T’s for almost all children with OSAS problems [19 /21,24]. Yet, they have seldom been followed up for more than at most, a few years. In this study, the majority (65%) of the T-group had experience at least partial recurrence of the symptoms. More research will be
E. Johansson, E. Hultcrantz
done with these individuals concerning breathing patterns and bite development. A new finding in the present study was the significantly increased occurrence of chronic disease in the tonsillectomized group: 23.5% in the Tgroup compared with 2.5% in the comparison group, RR /9.41 (1.13 B/RR B/78.14) P /0.024. These diseases were not apparent at the time of the surgery, based on medical records from the time of the patients’ original treatment. To be mentioned is also that one of the four persons in the control group, omitted due to earlier tonsillectomy, had answered that she was operated at the age of 3.5 years and that she to day suffered from ‘chronic pain in joints and muscles’. The illnesses asthma, goiter and ulcerative colitis, and high blood pressure which were represented in the T-group, in many cases have an immunological background. Therefore the question arises whether there could be a connection between tonsillectomies and later development of immunological diseases? The children who later developed a chronic disease did not belong to the youngest age-group, B/5 years, at the time for surgery, they were all between 5 and 9 years. There are several studies [1,6,14 /16] that show that immunoglobulins in serum and saliva decrease directly after a tonsillectomy. Bo ¨ck et al. [17] carried out a study in 1994, of different infection variables 6.5 years after T in children. They found no increase of infection frequency, but they did find changes in lymphocytic structure and a decrease of Ig A. However, they would not assign any clinical relevance to this. They concluded that the changes were an effect of decreased antigen load after a T, and therefore decreased Ig-production. Other studies have mainly pointed out that the concentrations of Ig A has decreased and remained low after the surgery. Serum Ig A and saliva Ig A have been lower compared with the time before the operation [6,14,15,17]. However, the concentration has never been under normal reference values and, therefore has not been as low as with Ig A hypogamma-globulinemia. It has been stressed that the decrease does not have any clinical significance [1,3,15]. If one draws a parallel to the patients with Ig A deficiency, it has been found that these individuals do not necessarily have more infections than others. On the other hand, it has been seen that people with Ig A deficiency have an increased risk for developing autoimmune diseases [18]. That tends to support the findings in the present study, although these are based only on questionnaires and not on laboratory data. In the present study, the basis for the differences in the occurrence of chronic disease is not clear:
Tonsillectomy */clinical consequences twenty years after surgery? can T have such an effect on the immune system that the individual runs a risk to develop a disease with an immune origin? Or is it that these people have the predisposition to suffer both from tonsil hypertrophy and chronic diseases? It would be helpful in answering these questions if one could get information about these subject’s backgrounds */for example on heredity for immune diseases */but since they were promised that data would not be analyzed individually, this is not possible ethically. The question is sufficiently interesting that a larger study should be done analyzing the occurrence of tonsillectomy in people with different chronic and autoimmune diseases. A T should be done on strict indications with thorough examinations where the pros and cons for the surgery are weighed carefully [19,22,23]. Individual therapeutic solutions can be recommended. For example, an alternative to a T is tonsillotomy in which only the obstructive part of the tonsil is removed [25] for cases with mostly obstructive problems. Another alternative, which has been tried, is long time antibiotic prophylaxis for recurrent tonsillitis [5,22]. Both alternatives are intended to decrease the potential negative effects on the immune system of the surgery.
5. Conclusion Tonsillectomies in childhood of individuals that have not shown increased disposition for infections before the surgery do not seem to give more infections 20 years later in comparison with nonoperated individuals in the same age group. On the other hand, an increased occurrence of chronic, immune-mediated diseases was observed in the operated group. This finding argues for extended epidemiological and immunological research within the area.
References [1] R. Donovan, J.F. Soothill, Immunological studies in children undergoing tonsillectomy, Clin. Exp. Immunol. 14 (1973) 347 /357. [2] C. Frieke Kuper, P.J. Koornstra, D.M.H. Hameleers, J. Bbiewenga, B.J. Spit, A.M. Duijvestijn, P.J.C. van Breda Vriesman, T. Sminia, The role of nasopharyngeal lymphoid tissue, Immunol. Today 13 (6) (1992) 219 /224. [3] C. Paulussen, J. Claes, G. Claes, M. Jorissen, Adenoids and tonsils, indications for surgery and immunological consequences of surgery, Acta Oto-Rhino-Laryngol. Belg. 54 (2000) 403 /408. [4] L.P. Gray, The T’s and A’s problem */assessment and reassessment, J. Laryngol. Otolog. 91 (1977) 11.
987
[5] Ming-de Ying, Immunological basis of indications for tonsillectomy and adenoidectomy, Acta Otolaryngol. (Stockholm) Suppl. 454 (1998) 279 /285. [6] M. Sainz, F. Guterrez, P.M. Moreno, C. Mun ˜oz, M. Ciges, Changes in immunologic response in tonsillectomized children. I. Immunosuppression in recurrent tonsillitis, Clin. Otolaryngol. 17 (1992) 376 / 379. [7] S. Mui, B.M. Rasgon, R.L. Hilsinger, Efficacy of tonsillectomy for recurrent throat infection in adults, Laryngoscope 108 (1998) 1325 /1328. [8] A.K. Soreide, J. Olofsson, (Adeno)tonsillektomi hos barn, Tidsskr. Nor. Laegeforen. 115 (3) (1995) 352 / 354. [9] E. Hultcrantz, M. Larsson, R. Hellquist, J. AhlquistRastad, H. Svanholm, O.P. Jakobsson, The influence of tonsillar obstruction and tonsillectomy on facial growth and dental arch morphology, Int. J. Pediatr. Otorhinolaryngol. 22 (1991) 125 / 134. [10] CDC */Centers for Disease Control and Prevention, Epi Info 6 for DOS version 6.04d; http://www.cdc.gov/epiinfo. [11] J. Cederholm, M. Korpela, Diagnosredovisning fo la a 1998 vid va i ¨r ¨karbeso ¨k ˚r ˚rdcentralen Tierp; http://www.sos.se/fulltext/0046-001/S7_87.PDF. [12] H. Saito, K. Asakura, M. Hata, A. Kataura, K. Morimoto, Does adenotonsillectomy affect the course of bronchial asthma and nasal allergy, Acta Otolaryngol. (Stockholm) 523 (Suppl.) (1996) 212 / 215. [13] J.L. Paradise, C.D. Bluestone, et al., Efficacy of tonsillectomy for recurrent throat infection in severely affected children, N. Engl. J. Med. 310 (1984) 674 /683. [14] A. Cantani, P. Bellioni, F. Salvinelli, L. Businco, Ann. Allergy 57 (1986) 413 /416. [15] B.E. Del Rio-Navarro, S. Torres, L. Barraga ´n-Tame, C. De la Torre, A. Berber, G. Acosta, J.J.L. Sienra-Monge, Immunological effects of tonsillectomy/adenectomy in children, Adv. Mucosal Immunol. (1995) 737 /740. [16] A. Morag, P.L. Ogra, Immunologic aspect of tonsils, Ann. Otology Rhinology Laryngol. (1975) Suppl. 19, vol. 84 (2), 37 /43. [17] A. Bo ¨ck, W. Popp, K.R. Herkner, Tonsillectomy and the immune system: a long-term follow up comparison between tonsillectomized and nontonsillectomized children, Eur. Arch. Otorhinolaryngol. 251 (1994) 423 / 427. [18] J.T. Cassidy, R.E. Petty, Textbook of Pediatric Rheumatology, 3rd ed., pp. 472 /473. [19] J. Ahlquist-Rastad, E. Hultcrantz, H. Svanholm, Children with tonsillar obstruction: indication for and efficacy of tonsillectomy, Acta Paed. Scand. 77 (1988) 831 /835. [20] F. Kudoh, A. Sanai, Effect of tonsillectomy and adenoidectomy on obese children with sleep-associated breathing disorders, Acta Otolaryngol. (Stockholm 1996), Suppl. 523: 216 /218. [21] M. Sato, M. Suzuki, et al., Long-term follow-up of obstructive sleep apnea syndrome following surgery in children and adults, Tohoku J. Expt. Med. 192 (3) (2000) 165 / 172. [22] T. Marshall, A review of tonsillectomy for recurrent throat infection, Br. J. General. Practice 48 (1998) 1331 / 1335.
988 [23] B. Lo ¨fstrand-Tidestro ¨m, B. Thilander, J. Ahlqvist-Rastad, O. Jakobsson, E. Hultcrantz, Breathing obstruction in relation to craniofacial and dental arch morphology n 4-year-old children, Eur. J. Orthodontics. 21 (1999) 323 /332. [24] T. Nishimura, N. Morishima, S. Hasegawa, N. Shibata, K. Iwanaa, M. Yagisawa, Effect of surgery on obstructive sleep
E. Johansson, E. Hultcrantz apnea, Acta Otolaryngol. (Stockholm) Suppl. 523 (1996) 231 /233. [25] E. Hultcrantz, A. Linder, A. Markstro ¨m, Tonsillectomy or tonsillotomy? */a randomized study comparing postoperative pain and long-term effects, Int. J. Pediatr. Otorhinolaryngol. 51 (1999) 171 / 176.