Intratonsillar detection of 27 distinct viruses: A cross‐sectional study

Abstract Palatine tonsils have been observed to harbor several distinct respiratory and herpesviruses in separate studies. In this study, the presence of these viruses in palatine tonsils was comprehensively studied in both children and adults. A cross‐sectional analysis of 181 patients (median age 22 years; range, 2.6‐66) operated for a benign tonsillar disease was conducted. Real‐time polymerase chain reaction was performed to detect 27 distinct viruses in all: eight human herpesviruses, 16 respiratory viruses, parvo B19, and polyoma BK/JC viruses. Clinical characteristics of the patients and underlying conditions were evaluated. In total, 92% of patients had virus detected in tonsils (Epstein‐Barr virus 72%, human herpesvirus 7, and 6B 54% and 16%, respectively, enterovirus 18%, parvovirus B19 7% and the rest <4%). No herpes simplex virus 2, varicella zoster virus, polyoma JC virus, parainfluenza‐, metapneumo‐, or coronaviruses were found. Enterovirus was more common in children and was frequently observed in the presence of HHV6B. None of the viruses showed a positive association to the tonsillar disease. Respiratory symptoms were not associated with the prevalence of viruses. This study comprehensively reports a cross‐sectional view of intratonsillar virus infections in elective tonsillectomy patients in a wide age range cohort. Tonsils are a major virus reservoir for distinct herpes and respiratory viruses without a positive association with tonsillar disease or respiratory symptoms.

The palatine tonsils' tendency to harbor viruses asymptomatically has also led to suggestions that there might be a causal relationship between viral presence in tonsils and chronic adenotonsillar diseases, namely, tonsillar hypertrophy and recurrent or chronic tonsillitis. 6 This is indeed interesting, as the etiology of these common diseases is still not completely understood.
Nevertheless, conflicting results have been reported on these kinds of evaluations [6][7][8][9] and the lack of a healthy control group is a major drawback in virtually all of these studies.
Previous literature on viral prevalence in the tonsils is somehow fragmented since asymptomatic tonsillar viral infections have mainly been evaluated in protocols detecting single or a few viruses. Furthermore, very few studies have focused on tonsillar viral prevalence in adults, especially respiratory RNA viruses. In this study, the presence of multiple viruses was investigated in the tonsils of nonacutely ill patients. A setup of a large viral analysis panel was constructed, including the detection of eight herpesviruses, 16 respiratory viruses, parvovirus B19, and polyoma BK/JC viruses simultaneously.

| Patients
Human tonsil samples were consecutively obtained at Turku University Hospital in Turku, Finland between October 2013 and December 2014. Only nonacutely ill patients who underwent elective tonsillectomy due to a nonmalignant disease were eligible for the study. Patients were considered nonacutely ill, if they were free from respiratory symptoms both at the time of preoperative phone call a few days before surgery and on the operation day fulfilling the criteria for elective surgery. Exclusion criteria for the study were systemic use of anti-inflammatory medication (other than nonsteroidal anti-inflammatory drugs) within 4 weeks before surgery, suspicion of malignancy, and a chronic systemic disease affecting the immune system (such as an autoimmune disease, immune complex disease, or immunodeficiency). Nonetheless, patients with allergy, asthma, and atopic dermatitis were eligible. Written informed consent was obtained from all individual participants and/or their guardians. The study protocol was approved by the Ethics Committee of the Hospital District of Southwest Finland. All study procedures involving human participants were in accordance with the 1964 Helsinki Declaration and its later amendments.

| Patient data
Participating patients were thoroughly interviewed for clinical information using a standard questionnaire, which included general health condition, medication, atopic diseases, current smoking habits, and respiratory symptoms (sore throat, fever, sneezing, coughing, and breathing difficulties) 2 weeks before surgery. Moreover, the indication of surgery was recorded at the time of operation.
The study patients were assessed in three age groups due to the assumed age-related differences in tonsillar function and prevalence of chronic adenotonsillar diseases. Age group I consisted of pediatric patients (under 16 years old) and age group II of adolescents and young adults (16-30 years) with immunologically active tonsils. 10,11 Age group III included patients over 30 years of age.
The patients were characterized to have tonsillar hypertrophy if their tonsils occupied more than 50% of the oropharyngeal airway, corresponding to a Brodsky score of 3 or 4. 12 Chronic tonsillitis was diagnosed in patients who had had a sore throat and tonsillar inflammation for at least 3 months. 13 Recurrent tonsillitis referred to at least four episodes of acute tonsillar infection in a year or three episodes in 6 months according to the Finnish guidelines of indications for elective tonsillectomy. 14 In addition, a history of one or several peritonsillar abscesses was detected.

| Sample collection, processing, and virus detection
Tonsillectomy was performed according to routine clinical practice. Immediately after removing the tonsils, several pieces of tonsillar core tissue were put into dry tubes and stored at −70°C for later analysis.
The nucleic acid extractions and polymerase chain reaction (PCR) tests were performed in the accredited clinical virus diagnostic laboratory according to standard daily protocols. The performance of nucleic acid extraction and PCR tests for human tissue samples have been analyzed earlier when applying the assays in clinical use. The analytical sensitivity of the in-house-tests were 5 to 10 cp/reaction, 1 to 5 cp/reaction for commercial multiplex herpesvirus test, and 50 cp/reaction for commercial multiplex respiratory virus test.
The tissue samples were homogenized and incubated with Proteinase K in +56°C for 1 hour before the nucleic acid extraction with NucliSens easyMag (bioMerieux, Lyon, France). Nucleic acids were stored at −70°C until analyzed. Rotor-Gene 3000 instrument (Corbett Research/Qiagen) was used for PCR cycling, except for herpesviruses, whose detection was based on broad-range PCR and microarray identification (Mobidiag, Espoo, Finland). The testing for entero-(EV), rhino-(RV), and respiratory syncytial (RSV) viruses were performed by an in-house triplex PCR test. 15 The primers for EV and RV were derived from the highly conserved 5′ noncoding regions of picornavirus genome are highly sensitive and eventually detect all entero-and rhinovirus genotypes. 16,17 F protein gene-specific primers were used for RSV detection. 18 Positive amplicons were identified as EV, RV, and RSV by melting curve analysis. 15 Polyoma BK (BKPyV) and polyoma JC (JCPyV) viruses were analyzed using an in-house PCR test as described earlier with slight modifications. 19

| Statistical analysis
Continuous variables are presented as medians and ranges, and categorical variables as frequencies and percentages. The associations between categorical variables were evaluated using the χ 2 test.
In cases of low-frequency variables, Fisher's exact test was used.
When assessing the associations between dependent variables (prevalence of viruses) and explanatory variables (age group, gender, smoking, prevalence of atopic diseases, respiratory symptoms, and surgical indications), logistic regression was performed and an odds ratio (OR) with a 95% confidence interval (95% CI) is expressed.
Multivariable analyses for the prevalence of EV and HHV6B were performed with the same method including age group and surgical indication in the model. P < .05 (two-tailed) was considered statistically significant. The statistical analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC).

| Characteristics of the study population
In all, 183 patients participated in the study. One patient was ex-  Table 1.
Female patients reported more atopic dermatitis than male patients (P = .01). The prevalence rates of allergic rhinitis and asthma were higher in older patients than in younger ones (P < .001 and P = .044, respectively), and atopic diseases were more frequently observed in the presence of each other (data not shown). In addition, there was an association between smoking and age group (P < .001; Table 1). Otherwise, no associations were observed between age group, gender, atopic diseases, and smoking when assessing the relations between each other. Indications for surgery were different among the age groups (Table 1); tonsillar hypertrophy was more frequently present in the age group I, and the rest of the indications were more frequent in age groups II and III (P < .01 for each comparison between an individual indication and age groups). In addition, tonsillar hypertrophy and a history of peritonsillar abscess were more common in male patients (P = .004 and P = .006, respectively).

| Symptoms before surgery
The rate of reported respiratory symptoms before surgery was related to the age of the patient (Table SA). Both sore throat (P = .037) and other respiratory symptoms (P = .021) were more frequently present in patients in age groups II and III than in the age group I. In addition, current smokers reported respiratory symptoms other than sore throat more frequently than those who did not smoke (P < .05), and female patients reported sore throat more frequently than the males (P < .05; Table SB). Otherwise, no associations were observed between the clinical characteristics and respiratory symptoms before surgery. Sore throat and other respiratory symptoms were associated with the presence of chronic tonsillitis (P < .001 and P = .018, respectively), whereas patients with tonsillar hypertrophy had fewer symptoms before surgery (P = .005).

| Virus detection
The prevalence rates of all detected viruses are presented in Table 2.

| Virus positivity in relation to clinical features
The results of virus detection in relation to age, gender, atopic diseases, and smoking are presented in Table 2 and Table SC. The prevalence of EV was related to the age of the patient (Tables 2 and 3), and the relationship remained significant when adjusted for smoking and surgical indications (P < .01). EBV infection was most frequently observed in age group III and was seen least in age group I, but the difference was not statistically significant (P = .49). Age group I had more respiratory DNA viruses (AdV and HBoV) than age groups II and III (Table 2), but the low numbers of cases hamper statistical assessment. Otherwise, no associations were observed between specific viruses and age, nor was the presence of any specific virus independently associated with gender, smoking, or presence of atopic diseases.
In addition, EV and HBoV were more frequent in patients with tonsillar hypertrophy (P = .02 for both), but the associations did not remain significant after adjusting for the age group. No other associations were observed between specific viral infections and indications of surgery (Table SD).
All patients with intratonsillar HBoV infection (n = 7) were asymptomatic before surgery. Conversely, FluA infection was observed in patients (n = 2), both of which had had respiratory symptoms. However, due to the low numbers of cases, statistical assessment was not possible in either observation. Comparing patients that reported any respiratory symptoms within 2 weeks before the surgery and those who did not, none of the individual investigated viruses were more prevalent in either group, excluding HBoV and FluA (Table 2). There was also no presence of any respiratory viruses, herpesviruses, or any number of all detected viruses in significant association to the presurgery symptoms.

| DISCUSSION
This study was conducted to clarify the general view of multiple intratonsillar virus infections in nonacutely ill tonsillectomy patients.
The study showed high intratonsillar prevalence rates for a few human herpesviruses, which was expected, but a reasonably low rate for respiratory viruses when compared with previous reports that were virtually limited only to the pediatric study populations. Respiratory symptoms 2 weeks before surgery were not associated with virus positivity. In addition, EV was frequently observed in the presence of HHV6B.
The study population was a pragmatic patient cohort undergoing routine tonsillectomy for a benign tonsillar disease at a single institution. Age group I, mainly consisting of patients undergoing operations for tonsillar hypertrophy, included more male patients than females, which is in line with the available corresponding data from the Swedish Tonsil Register. 20 The reason for the male preponderance in   was close to that (12.7%) reported in Italy a few years ago. 23 The prevalence rates of EBV, HHV7, and HHV6B were not related to age, gender, or smoking, although the percentages of infected patients increased by age in EBV and HHV7 and decreased in HHV6B. These age-related trends in prevalence rates are in line with the results of the SPLIT study. 22 Other HHVs were observed in very few cases.
This was probably the second study published to evaluate the prevalence of VZV in tonsils, 22   Accordingly, there might be a difference in the intratonsillar prevalence of all these viruses between pediatric and adult patients, and further studies are also needed to characterize these differences in respiratory RNA viruses.
Enteroviruses have been considered to be highly cytolytic, but it is well-known that they are able to persist after acute infection in some tissues, such as the myocardium and pancreatic tissue. 25 Only minor efforts have been made so far to study the intratonsillar prevalence of enteroviruses in asymptomatic adult patients. In this study, age was the strongest factor related to the prevalence of EV.
It is noteworthy that no significant sex differences were observed in intratonsillar EV prevalence considering its higher general frequency and severe complications in males. [26][27][28] Tonsillar hypertrophy was not independently associated with the prevalence of EV in contrast to a single previous report. 6 Instead, EV was more frequently detected in the presence of HHV6B, which is a novel and interesting observation indicating a need for further investigation.
Parvovirus was detected in 7.2% of cases, which is a remarkably lower frequency compared with previous study populations reporting frequencies up to 16% to 45%. [29][30][31][32] However, methodological aspects to explain these differences cannot be ruled out considering the recent findings of parvovirus persistence in tonsils. 30,31 The differences in parvovirus persistence rates are likely due to differences in test sensitivity since parvovirus persists typically in very low copies. 33 Only two patients (1.1%) were positive for the assay detecting BKPyV, which is comparable with previous corresponding studies using modern methods. 34,35 In this study, the only independent association between viral prevalence and indication for surgery was the inverse relation between HHV6B and chronic tonsillitis. This is an interesting preliminary observation since there seems to be a trend toward a decreasing prevalence of intratonsillar HHV6B near middle age, 22 along with an increasing incidence of clinical manifestations due to cryptic chronic tonsillitis, such as tonsilloliths. 36 The role of viruses in the pathogenesis of chronic adenotonsillar diseases has been speculated previously. 6 However, it is explicitly challenging to draw any conclusions on the causality between tonsillar viral prevalence and chronic tonsillar diseases using this kind of study protocol. For ethical reasons, it is not possible to create an appropriate control group since tonsils cannot be obtained from healthy patients. In conclusion, this study provides a comprehensive crosssectional report of intratonsillar virus infections in a pragmatic cohort of nonacutely ill tonsillectomy patients. Especially the data on respiratory RNA viruses in the adult population is mostly unique in its 3836 | present form. None of the investigated viruses was positively associated with tonsillar diseases or respiratory symptoms before surgery. The tendency toward differences between intratonsillar viral prevalence in pediatric and adult populations, as well as the specific features of the prevalence of HHV6B and EV, indicate a need for further evaluation.