Comparison of Helicobacter pylori colonization on the tonsillar surface versus tonsillar core tissue as determined by the CLO test.

Abstract

We conducted a prospective study to determine the correlation between the presence or absence of Helicobacter pylori on the tonsillar surface and in the tonsillar core as determined by the Campylobacter-like organism (CLO) rapid urease enzyme test. Our study population was made up of 55 patients who underwent adenoidectomy, tonsillectomy, or both from December 2002 through April 2003 at Khalili Hospital in Shiraz, Iran. Of these 55 patients, 45 (82%) were positive and 10 (18%) were negative for H pylori colonization as determined by CLO testing. Analysis of samples obtained from individual patients revealed differences in H pylori colonization between tonsillar surface samples and the core tissue samples. Of 106 tonsils obtained from 53 patients who underwent adenotonsillectomy or tonsillectomy, H pylori was found on 56 tonsillar surface samples (53%) and 24 tonsillar core samples (23%); only 13 tonsils (12%) contained H pylori both on the surface and in the core. We conclude that a surface swab is neither specific nor sensitive as an indicator of the presence or absence of H pylori colonization in tonsillar core tissue.

Introduction

Tonsillar tissue is a component of mucosa-associated lymphoid tissue (MALT) that has evolved to protect the vulnerable mucosal surface. However, even as the tonsils contribute to the body's immune system function, (1) they may also serve as a reservoir for pathogens.

The role of MALT in the pathogenesis of gastrointestinal disease has been a subject of some interest. (2) Helicobacter pylori is the organism that has been implicated as an etiologic factor in the development of duodenal ulcers and gastritis. Because the MALT of Waldeyer's ring is the first line of mucosal defense against invading pathogens, it is plausible that H pylori induces a proinflammatory reaction that is both local and systemic. Such downstream priming of mucosal immunity by H pylori may account for the diverse pathologic tissue responses observed in infected patients; these responses include gastritis (3) and MALT lymphomas (4) of the salivary glands, stomach, and rectum.

Some authors have suggested that tonsillectomy may protect the host against H pylori infestation of the stomach because the tonsils may act as reservoirs for recurrent systemic infections. (5) If infection of tonsillar tissue does indeed have an impact on H pylori colonization of the gastric mucosa, clinicians would benefit from having access to an easy, sensitive, and specific method of verifying the presence of H pylori in tonsillar tissue. Eradicating known H pylori from tonsillar tissue might make it easier to eradicate it in gastric mucosa.

With this theory in mind, we undertook a study of H pylori colonization in the tonsils by obtaining samples from the tonsillar surface and from tonsillar core tissue and subjecting them to the Campylobacter-like organism (CLO) rapid urease enzyme test. Our goal was to determine the correlation between the presence or absence of H pylori on the tonsillar surface and in the tonsillar core.

Patients and methods

For this prospective study, we enrolled 55 patients--27 males and 28 females, aged 2 to 75 years (mean: 12)--who were scheduled for tonsillectomy, adenoidectomy, or both from December 2002 through April 2003 in the ENT Department at Khalili Hospital in Shiraz, Iran. Of the 55 patients, 28 underwent adenotonsillectomy, 25 underwent tonsillectomy only, and 2 underwent adenoidectomy only, all under general anesthesia. Their indications for surgery were routine--primarily chronic/recurrent tonsillitis and obstructive sleep apnea.

Information was obtained on each patient's use of antibiotics, metronidazole, omeprazole, and bismuth compounds during the 2 months prior to surgery.

Specimens. Preoperatively, we obtained a surface swab sample from each tonsil in all of the 53 patients who underwent adenotonsillectomy or tonsillectomy. No tonsillar surface swab was performed on the 2 patients who underwent adenoidectomy alone.

Following excision, all surgical specimens were washed with sterile normal saline. Then, three core samples were obtained from each of the 28 adenotonsillectomy patients (one from each tonsil and one from the adenoid), two from each of the 25 tonsillectomy patients (one from each tonsil), and one from the adenoid of each of the 2 adenoidectomy patients. Each core sample was 2 mm wide, and each was cut from the excised tissue with a newly sterilized blade; to further lower the risk of cross-sample contamination, gloves were changed after every cut. All core samples were then washed with physiologic saline.

CLO test. The tonsillar swab samples were finely dissolved in 1 ml of normal saline, and the solutions were placed in CLO test tubes. The tonsillar core tissue specimens washed with physiologic saline were also placed in CLO test tubes. All test tubes were prepared according to published guidelines and kept at room temperature.

CLO test readings were taken at 30 minutes and 24 hours. Each solution that had turned yellow by 24 hours represented a positive result. Any color changes that occurred beyond 24 hours were not considered to indicate a positive result.

For comparisons of tonsillar and adenoid tissue, positivity in either the right or left tonsil was sufficient to classify a patient as tonsil-positive. For comparisons of tonsillar surface and core samples, a paired specimen analysis of each individual tonsil was performed.

The CLO tests were provided by the Microbiology Research Department at Nemazee Hospital in Shiraz.

Statistical analysis. The chi-square ([chi square]) test and the Fisher's exact test were used for statistical analysis. A p value of less than 0.05 was considered to be significant.

Results

Medication use. All but 6 patients (11%) had used an antibiotic during the 2-month period before surgery; the most common were penicillin derivatives. Only 1 patient had used metronidazole during this period, and none had used omeprazole or bismuth.

CLO test results. Of the 55 patients in our study, 45 (82%) were positive and 10 (18%) were negative for H pylori colonization as determined by CLO testing of either the tonsillar surface, the tonsillar core tissue, or adenoid tissue.

Of the 53 patients who underwent adenotonsillectomy or tonsillectomy, 44 (83%) had a positive surface swab and/or core sample; 9 patients (17%) were negative at both sites (table 1). Surface swabs were positive in 41 patients (77%) and negative in 12 (23%), while core tissues were positive in 20 (38%) and negative in 33 (62%). A total of 17 patients had positive results in both the tonsillar surface swab and the core sample; this group represented 41% of the 41 patients with a positive surface swab and 85% of the 20 patients with a positive core sample.

In the group of 53 patients who underwent adenotonsillectomy or tonsillectomy, we performed a paired analysis of core and surface samples in each patient (total number of specimens: 106). Of these 106 specimens, 56 surface swabs (53%) and 24 core samples (23%) were positive (table 2).

In the 28 patients who underwent adenotonsillectomy, we analyzed the relationship between the presence of H pylori in adenoid core tissue and in the tonsillar core. Half the patients were positive in at least one site; 7 (25%) were positive in both cores, 2 (7%) were positive in the adenoid core only, and 5 (18%) were positive in the tonsillar core only, leaving 14 patients (50%) who were negative in both cores (table 3).

Discussion

It has been reported that H pylori infection is the most common chronic bacterial infection in man, affecting more than half of the world's population. (6,7) H pylori has been detected in dental plaque (8)--with the oral cavity probably acting as a reservoir for systemic infection and reinfection--and more recently in human feces. (6,9) These findings support the view that infection is transmitted by oral-oral or fecal-oral routes and that H pylori does not require an acidic environment (i.e., the stomach) to survive. In addition, reports have described H pylori transmission from the stomach to the mouth via a reflux mechanism. (10) Also, H pylori has been found in nasal and maxillary sinus tissue specimens obtained from patients with both chronic sinusitis and gastric H pylori infection. H Finally, it has been reported that transmission of H pylori may also occur via gastroesophageal reflux from the stomach to the nose and/or sinuses. (12)

We chose to use the CLO test in our study because it is highly specific and sensitive for H pylori and because it is the most common method of finding H pylori used by gastroenterologists. Dye et al reported that the specificity of the CLO test for H pylori was 97% and the sensitivity was 98%. (13) Schnell et al reported similar results. (14)

As a general rule, patients who undergo CLO testing should not use antibiotics, metronidazole, omeprazole, or bismuth compounds for a few weeks prior to testing because these agents reduce the sensitivity of the test. In view of the fact that most patients in our study (89%) had already taken an antibiotic many times during the preceding 2 months, we were not able to stipulate "nonuse of antibiotics" as an inclusion criterion. Also, the location of H pylori bacteria in tissues following antibiotic use can be scattered, and this can lead to false-negative results. To accommodate for this possibility, multiple samples should be taken from each specimen, as we did in our study. The use of metronidazole, omeprazole, and bismuth was not an issue in our study.

Analysis of our data reveals that the results of a tonsillar surface swab do not reliably reflect the presence or absence of H pylori colonization in the tonsillar core, as only 13 of 106 tonsils (12%) harbored H pylori both on the surface and in the core. In our study, tonsillar swab results had a sensitivity of 23% and a specificity of 77%. This finding is consistent with results reported by Mitchelmore et al, who also performed a paired-specimen analysis of the bacterial flora of the tonsillar surface and the tonsillar core in patients with recurrent tonsillitis. (15) Their goal was to determine the usefulness of surface swabs in predicting the presence of bacterial flora in the tonsillar core, and they concluded that these swabs are not useful. In that study, only 44% of aerobic isolates and 31% of anaerobic isolates were recovered on both the surface and in the core. However, others have found that tonsillar swabs are useful predictors of core infection. (16)

The high rate of H pylori positivity on CLO testing in our adenotonsillectomy and tonsillectomy patients (83 %) is consistent with the findings of Minocha et al (5) and Unver et al, (7) but they vary greatly with those of other investigators. For example, Yilmaz et al (9) found no H pylori colonization in 50 adenotonsillectomy specimens with CLO testing, and di Bonaventura et al (17) found no H pylori colonization on tonsillar swabs with polymerase chain reaction testing. Skinner et al concluded that H pylori does not appear to colonize the tonsil but that it primes the tonsils by inducing macrophages to express a proinflammatory reaction through the inducible nitric oxide synthase. (18)

We hypothesize that because of similarities between the gastric and tonsillar environment, H pylori may be able to colonize the tonsillar surface and reach the core by penetrating the tonsillar crypts. If this is true, several possibilities may explain the difference between the rate of H pylori positivity on the tonsillar surface and the rate in the tonsillar core:

* A failure to detect H pylori may be attributable, at least in part, to the presence of coccoid forms of H pylori, which develop as a temporary expression of the bacterium's adaptation to an extragastric environment. (19) These coccoid forms have less urease activity than do normal H pylori organisms. This phenomenon is more common on the surface of the tonsils than in the core tissue. (20)

* The appearance of coccoid forms with their reduced level of urease activity is also a result of antibiotic use, as is the previously mentioned scattered redistribution of H pylori. (19) Again, these consequences of antibiotic use have a more direct impact on the surface of the tonsil than on the core. (17)

* An overgrowth of aerobic and anaerobic flora (Staphylococcus, Streptococcus, Neisseria, Fusobacterium, and Peptostreptococcus spp) on the tonsillar surface (7) probably makes it more difficult to detect H pylori by tonsillar swabbing.

In conclusion, our findings support the view that a CLO test of the tonsillar core is more reliable than a surface swab. It seems safe to conclude that with a sensitivity of 23% and a specificity of 77%, surface swabs do not reliably reflect the presence or absence of H pylori in the tonsillar core.

References

(1.) Sennaroglu L, Onerci M, Hascelik G. The effect of tonsillectomy and adenoidectomy on neutrophil chemotaxis. Laryngoscope 1993; 103(12):1349-51.

(2.) Minocha A, Raczkowski CA. Role of appendectomy and tonsillectomy in pathogenesis of ulcerative colitis. Dig Dis Sci 1997; 42(7):1567-9.

(3.) Marshall BJ, Warren JR. Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. Lancet 1984; 1(8390):1311-15.

(4.) Alkan S, Karcher DS, Newman MA, Cohen P. Regression of salivary gland MALT lymphoma after treatment for Helicobacter pylori. Lancet 1996;348(9022):268-9.

(5.) Minocha A, Raczkowski CA, Richards RJ. Is a history of tonsillectomy associated with a decreased risk of Helicobacter pylori infection? J Clin Gastroenterol 1997;25(4):580-2.

(6.) Thomas JE, Gibson GR, Darboe MK, et al. Isolation of Helicobacter pylori from human faeces. Lancet 1992;340(8829):1194-5.

(7.) Unver S, Kubilay U, Sezen OS, Coskuner T. Investigation of Helicobacter pylori colonization in adenotonsillectomy specimens by means of the CLO test. Laryngoscope 2001;111(12):2183-6.

(8.) Nguyen AM, el-Zaatari FA, Graham DY. Helicobacter pylori in the oral cavity. A critical review of the literature. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1995;79(6):705-9.

(9.) Yilmaz M, Kara CO, Kaleli I, et al. Are tonsils a reservoir for Helicobacter pylori infection in children? Int J Pediatr Otorhinolaryngol 2004;68(3):307-10.

(10.) Nabwera HM, Logan RP. Epidemiology of Helicobacter pylori: Transmission, translocation and extragastric reservoirs. J Physiol Pharmacol 1999;50(5):711-22.

(11.) Namavar F, Roosendaal R, Kuipers EJ, et al. Presence of Helicobacter pylori in the oral cavity, oesophagus, stomach and faeces of patients with gastritis. Eur J Clin Microbiol Infect Dis 1995;14(3):234-7.

(12.) Morinaka S, Ichimiya M, Nakamura H. Detection of Helicobacter pylori in nasal and maxillary sinus specimens from patients with chronic sinusitis. Laryngoscope 2003;113(9):1557-63.

(13.) Dye KD, Marshall BJ, Frierson HF, et al. Is CLO test alone adequate to diagnose Campylobacter pylori? [abstract] Am J Gastroenterol 1988;83:1032.

(14.) Schnell GA, Schubert TT, Barnes WG, et al. Comparison of urease, H&E and culture tests for Campylobacter pylori [abstract]. Gastroenterology 1988;94:A410.

(15.) Mitchelmore IJ, Reilly PG, Hay AJ, Tabaqchali S. Tonsil surface and core cultures in recurrent tonsillitis: Prevalence of anaerobes and beta-lactamase producing organisms. Eur J Clin Microbiol Infect Dis 1994;13(7):542-8.

(16.) Timon CI, Cafferkey MT, Walsh M. Fine-needle aspiration in recurrent tonsillitis. Arch Otolaryngol Head Neck Surg 1991;117(6): 653-6.

(17.) di Bonaventura G, Neff M, Neri G, et al. Do tonsils represent an extragastric reservoir for Helicobacter pylori infection? J Infect 2001;42(3):221-2.

(18.) Skinner LJ, Winter DC, Curran AJ, et al. Helicobacter pylori and tonsillectomy. Clin Otolaryngol Allied Sci 2001;26(6):505-9.

(19.) Steer HW. Ultrastructure of Campylobacter pylori in vivo. In: Rathbone BJ, Heatley RV, eds. Campylobacter Pylori and Gastroduodenal Disease. 2nd ed. Oxford, U.K.: Blackwell Scientific Publishers; 1992:233-45.

(20.) Almadori G, Bastianini L, Bistoni F, et al. Microbial flora of surface versus core tonsillar cultures in recurrent tonsillitis in children. Int J Pediatr Otorhinolaryngol 1988;15(2):157-62.

Bijan Khademi, MD; Nika Niknejad, MD; Behrooz Gandomi, MD; Firoozeh Yeganeh, MD

From the Department of Otolaryngology, Shiraz University of Medical Sciences, Shiraz, Iran.

Reprint requests: Bijan Khademi, MD, Department of Otolaryngology, Khalili Hospital, Khalili Ave., Shiraz, Iran. Fax: [phone]; e-mail: [e-mail]

Table 1. CLO test results in patients who underwent
adenotonsillectomy or tonsillectomy alone (n = 53);
tonsillar surface vis-a-vis tonsillar core tissue

                                    Tonsillar surface
                                       n (%)

                            Positive   Negative   Total

Tonsillar core   Positive   17 (32)      3 (6)    20 (38)
n (%)            Negative   24 (45)     9 (17)    33 (62)
                 Total      41 (77)    12 (23)    53 (100)

p = 0.28, Fisher's exact test, for the difference in positivity
between the surface and core samples.

Table 2. CLO test results in patients who underwent
adenotonsillectomy or tonsillectomy alone (n = 53):
paired analysis of each tonsillar specimen (n =106);
tonsillar surface vis-a-vis tonsillar core tissue

                                 Tonsillar surface

                                        n (%)

                             Positive   Negative    Total

Tonsillar core,   Positive   13 (12)    11 (10)    24 (23 *)
n (%)             Negative   43 (41)    39 (37)    82 (77 *)
                  Total      56 (53)    50 (47)    106 (100)

p = 0.93 and [chi square] = 0.01 for the difference in positivity
between the surface and core samples.

* Rounding errors.

Table 3. CLO test results in patients who underwent
adenotonsillectomy (n = 28); adenoid core tissue
vis-a-vis tonsillar core tissue

                                  Tonsillar core

                                      n (%)
                          Positive   Negative    Total

Adenoid core   Positive    7 (25)     2 (7)      9 (32)
n (%)          Negative    5 (18)    14 (50)    19 (68)
               Total      12 (43)    16 (57)    28 (100)

p = 0.016, Fisher's exact test, for the difference in positivity
between the tonsillar core and the adenoid core.