Upper respiratory tract infections (URTIs) are the most frequent cause for patient visits in general practice, and in children account for a substantial proportion of family doctor consultations1. Sore throat, a predominant symptom of pharyngotonsillitis is one of the most common complaints in paediatrics2, inviting considerable antibiotic prescriptions. In the United States 10% of all antibiotic prescriptions for paediatric consultations are given for pharyngotonsillitis3. Most episodes of pharygotonsillitis are caused by viruses, and Group A β-haemolytic Streptococcus (GABHS) or S pyogenes accounts for just about 15% of all cases2.

Pharyngotonsillitis, one of the most common childhood URTIs remains clinically indistinguishable whether it is of viral or bacterial etiology. Variations in tonsillar microbial flora may increase the risk of repeated attacks of tonsillitis and adenotonsillar hypertrophy, and why some children are more susceptible to infections of Waldeyer”s ring is unclear. Tonsillectomy, alone or with adenoidectomy in selected children for symptoms of recurrent infection or for hypertrophy and snoring is one of the most frequent paediatric surgical procedures. At least 34-80% of patients with sore throat are estimated to have a ‘possible” bacterial aetiology4,5 and approximately 24-65%, are actually tested for ß-haemolytic Streptococcus (BHS) in the throat1,6.

In a current hypothesis, acute pharyngotonsillitis is caused by adherent bacteria on the tonsillar epithelium which remaing present in surface secretions on the tonsillar crypts7.

Surface bacteria on the tonsils may not represent resident bacterial carriage within the tonsils. Treatment based on organisms identified from surface swabs of tonsils after recurrent tonsillitis may only address the surface bacteria allowing resident bacteria in tonsil core tissue to persist. Hypertrophied tonsils also harbor bacteria, predisposing to recurrent infection. In Trinidad and Tobago antibiotics are widely used to treat children”s URTIs, and though general practitioners preferentially prescribe them while recognize their overuse in these conditions8. Information on bacterial carriage in childrens” tonsils in the Caribbean is not available. This study investigated the prevalence of aerobic organisms on tonsil surface and core tissues in children following elective adenotonsillectomy for recurrent and/or persistent symptoms of infection or sleep apnoea.

Tonsils were removed in 51% (52) of children for chronic infection and in 45.1% (46) for obstructive tonsillar hyperplasia causing snoring and/or sleep apnoea. By chance, boys and girls were equally represented (50.0% each). About half the numbers of children were either below 6 years (52.9%) or between 6-16 years (48.1%) (Table1). The majority of patients were of either African (43.1%) or East Indian (32.4%) lineage. Fifty-five children (53.9%) had suffered at least one attack of sore throat in the past 12 months and of these 45 (81.8%) had suffered between 4 and 16 episodes. Forty-seven (52.2%) patients, (38 with recurrent tonsillitis and 9 with hypertrophy) had taken antibacterial agents for the immediate previous attack of sore throat. Amoxicillin alone or with clavulanate (40.4%) was most frequently prescribed followed by a macrolide (21.3%) {azithromycin (7, 70.0%), clarithromycin (2, 20.0%) and erythromycin (1, 10.0%)}. First generation cephalosporins were preferentially prescribed in 6 (12.8%) patients and one patient received cefuroxime. Only three respondents received prescriptions for penicillin V.

We have found a high association of surface and core growth of AHS in adenotonsillectomized children with significant BHS carriage on tonsil surfaces and cores in older children. Approximately half the children studied were between 1-5 (53.3%) or (6-16) years (46.7%). Age is a predictor of the infective etiology of pediatric tonsillitis; viral infections are more common in children <3 years, and GABHS infections are more frequent in children ≥ 6 years12 Children most frequently received amoxicillin/co-amoxicillin and macrolides. In reports from Trinidad8 and Spain13,14 amoxicillin-clavulanate and amoxicillin were most frequently prescribed for URTIs in primary care and penicillin V was never recommended even though it is the drug of choice to treat GABHS3. Steinman et al15 reported increased use of broad-spectrum antibiotics, especially azithromycin, clarithromycin and amoxicillin-clavulanate, in American children in primary care between 1991-99. The expanding empirical use of these agents for URTIs which have predominant viral aetiology indicates their unnecessary utilization.

Streptococcus spp. and Staphylococcus spp. were the most prevalent aerobes on children”s tonsils regardless of the site. Overall, children harbored similar proportions of aerobic bacteria on surfaces and cores with a total average of 8.9 isolates per child and respective surface and core averages of 3.0 and 2.5 isolates. Our findings compare with British findings16 of 9.2 and 8.8 average surface and core isolates respectively. However the sample was not limited to children. Italian children with recurrent tonsillitis who had not received antibiotics in the 20 days before sample collection, demonstrated a similar pattern of recovery for aerobic bacteria from tonsil surfaces (69.2%) and cores (59.2%)17

Overall streptococcal growth, and that of AHS in comparison to BHS was higher on the surfaces compared with cores. The high prevalence of surface and core AHS compared with BHS in oral and nasopharyngeal flora is expected. Brook et al18 reported Staphylococcus aureus, AHS and BHS were the predominant Gram-positive aerobic isolates on tonsil surfaces and cores in children with recurrent tonsillitis. Surow et al19 reported surface cultures grew normal respiratory flora but Staphylococcus was most commonly isolated from the core. High oral and nasopharyngeal AHS inhibits the growth of BHS in tonsillitis, and offers a mechanism to lower recurrence rates of streptococcal throat infections20, which are altered from antibiotic induced disturbances. The BHS carriage was consistent with reported peak age incidence of GABHS infections above 6 years3. Children are the major reservoir of GABHS and the target population for GABHS-induced pharyngotonsillitis and its suppurative or non-suppurative complications21 GABHS infections are common in school-aged children who mix with others in school and the neighbourhood22 forming a pool of high colonizers who can infect their peers. The lower frequency of BHS in younger children can be explained by their restricted peer contact as well as the residual presence of maternal antibodies.

Streptococcus pyogenes owes its major success as a pathogen to its ability to colonize, multiply rapidly and spread in the host while evading phagocytosis. We found considerably higher recovery of GABHS on tonsil surfaces (42.1%) and cores (57.1 %) than reported rates of 15.5%23, 21%24, and 28%25 from throats of apparently healthy children or those with recurrent tonsillitis (16.9%)26,27 or tonsillar hypertrophy (20%)21,23. Surow et al19 reported a 19.6% colonization rate in the tonsil core and Kielmovitch et al26 found high prevalence of S. pyogenes in cultured tonsil surfaces of children with obstructive tonsillar hypertrophy. The observed high rate of S. pyogenes colonization may have occurred because we excluded children who had received antibacterial treatment in the preceding 4 weeks, contrasting with Surow et al19 who limited this time to just one week before the study. S. pyogenes produces multiple adhesins with varied specificities which enable its ability to colonize niches in the upper respiratory tract27. The nature of these adhesins may differ in surface and core tissues in children, which can explain the non-concordance between surface and core growth of Streptococcus spp. and AHS. Researching the protein nature of these adhesins and their anchorage to the cell surface may elucidate the high rate of observed GABHS colonization in children. Nonetheless, our data suggest that Trinidadian children are high colonizers of S. pyogenes, and constitute a reservoir of GABHS providing a source of latent re-infection.

The most frequent non-GABHS isolates detected were Groups C and G, which are closely related, share virulence factors with GABHS and sporadically occur in acute pharyngotonsillitis as unusual pathogens28. They are generating interest as emerging nosocomial and opportunistic pathogens primarily for eye, ear or throat infections, cellulitis and rarely meningitis, septicemia, endocarditis, or glomerulonephritis29. Indian25 and Italian30 school children harboring BHS had high prevalence rates of 43.2% and 38.8% of Lancefield”s Group G BHS respectively. These findings in Caribbean children invite further research to elucidate BHS carriage in healthy children and those with pharyngotonsillitis. Low recovery of S. pneumoniae is supported by reports which failed to recover S. pneumoniae in pharyngotonsillitis31, URTIs and sore throat23.

The few studies reporting staphylococcal recovery in tonsillar flora in children do not demonstrate differences between pathological and normal tonsils. S. aureus was the commonest bacterium (9.9%) cultured from tonsils of 1000 healthy Israeli children <2 years32. Staphylococcal prevalence was similar with or without tonsil pathology in Brazilian33 (27% versus 28%) and Dutch4 children (6% in both groups). Gaffney et al34 isolated S. aureus from the tonsil core in 29% of children between 8-14 years with recurrent tonsillitis. Our observed higher prevalence of S. aureus on tonsil surfaces (68.9%) and cores (71.1%) has likely occurred because we cultured swabs from surfaces and cores of dissected tonsils in lieu of pharyngeal swabs which are neither reliable nor valid7,35. Although the current observed dissimilarity in the frequency of staphylococcal detection compared with other studies may represent a true difference, it is pertinent to mention that differences in detection methodology and specimen handling prior to processing cannot be ignored.

H. influenzae, an infrequent Gram-negative isolate, has been demonstrated in sore throat21, tonsil hypertrophy and recurrent tonsillitis17,23. In Sweden where routine throat swab analyses do not include H. influenzae, Gunnarsson23 has discussed its aetiological implications in pharyngotonsillitis. Mandatory immunization of children with Haemophilus influenzae type B vaccine in Trinidad and Tobago over the last decade would explain its absence in the present sample. Studies on large defined populations can clarify the aetiological role of H. influenzae in childhood tonsillitis in Trinidadian children. The several Gram-negative genera recovered in the present study have not been reported as far as these authors are aware. The isolation of Klebsiella and Pseudomonas species though at a low frequency, is worrying. Antibiotic exposure may have disturbed throat flora and facilitated re-colonization with these species. The precise reason why children acquired these organisms is unclear and their clinical relevance remains to be determined. Though the study was designed to examine aerobic bacteria, anaerobes and atypical bacteria cannot be ignored and need to be investigated.

The recovery of mixed aerobic bacteria from tonsil surface and core cultures raises the possibility that core bacteria may represent the surface bacteria. Exploring the relationship between core and surface aerobic bacteria Brodsky et al10 reported similar findings, but cautioned that surface bacteria may not truly represent the core growth. Other reports7,19 do not reflect core bacteria on the surface tissue. Recurrent bouts of tonsillitis, significant inflammation, and bacterial adherence to mucosal epithelium or infection within the crypts may explain our differing results.

We found higher streptococcal growth on tonsil surfaces than cores, and higher AHS than BHS yield from surfaces compared with the cores, Recovery of GABHS was high from both, surface and core tissues, particularly in older children. Trinidadian children are high colonizers of S. pyogenes, and constitute a reservoir of GABHS for latent re-infection. Studies to probe the mechanisms and influences of multiple adhesions of S. pyogenes in Trinidadian children are suggested.