Bacterial Tonsillitis and Antimicrobial Resistance Proles Among Children Within Five Years of Age At Hargeisa Group of Hospital, Somaliland: A Cross-Sectional Study

Tonsillitis is the third most frequently diagnosed infection in the pediatrics age group around the world. It causes signicant morbidity and loss of school attendance in children. The emergence of drug resistance in bacterial tonsillitis is getting higher every year. However, data on the drug resistance proles of bacterial causes of tonsillitis among children within ve years of age is not available in Somaliland.Therefore; this study determined the bacterial causes of tonsillitis and their antimicrobial resistance proles among children within ve years of age at Hargeisa Group of Hospital, Somaliland. Antimicrobial susceptibility was variables were structured regression analysis was factors associated with bacterial tonsillitis. P-values were taken as statistically acid (20/10 µg), gentamicin (10 µg), clarithromycin (15 µg), erythromycin (15 µg), vancomycin (30 µg), ooxacin (5 µg) and ciprooxacin (5 µg). A loop full of culture was taken from a pure culture colony and transferred to a tube containing 5 ml of normal saline and mixed gently until it forms a homogenous suspension. The turbidity of the suspension was then adjusted to the turbidity of McFarland 0.5 (which carries 10 8 CFU/ml) and was swabbed on a dry surface of MHA plate (150 mm) using a sterile cotton swab. Antibiotic discs were dispensed using a single disc dispenser. Plates were then incubated for 24 h at 37 o C. Diameters of the zone of inhibition around the discs were measured using a digital caliper. The results of the zone of antibiotics were interpreted based on the 2019 CLSI guideline [7]. These antimicrobial drug discs were selected based on the frequent prescriptions of these drugs for the treatment of tonsillitis infection in the study area and using the 2019 CLSI [7].

The situation is critical in Somaliland, where antimicrobials are vastly and frequently used irrationally. This might increase the emergence of resistance to the commonly used antibiotics for the treatment of tonsillitis [5].
In Hargeisa Group of Hospitals, like other similar health settings in Somaliland, routine culture and antibiotic susceptibility testing are not usually performed as essential part of patient care and treatments are mostly based on empirical therapy. However, there are no studies conducted and published on the culture con rmed burden of bacterial tonsillitis and antimicrobial resistance pro les as well as the extent of MRSA in children within ve years of age in Somaliland. Moreover, asymptomatic children can be the sources of dissemination of bacteria causing tonsillitis to non-infected children at home or at school settings that can lead to wide range of tonsillar infections. Therefore, we present the rst report of the prevalence of bacterial causes of tonsillitis and antimicrobial resistance pro les of the isolates in children within ve years of age at Hargeisa Group of Hospital, Somaliland.

Methods
Study design, period and setting A hospital based cross-sectional study was conducted between March and July 2020 in Hargeisa Group of Hospital (HGH), Somaliland. Hargeisa Group of Hospital is located in Maroodi Jeex Region, the capital city of Somaliland known as Hargeisa. According to the 2019 census report from Central Statistics Department of Somaliland, Hargeisa has a total population of 1.2 million. Hargeisa Group of Hospital is the largest referral public hospital with more than 200 healthcare professionals. It is one of the health hubs in Somaliland. The pediatric department of HGH has outpatient and inpatient departments with 15 Pediatricians and two Nurses. Daily 50 outpatients and 1 to 4 hospitalized children attend the Pediatric Clinic for different medical conditions. All children within ve years of age with tonsillitis at Ear, Nose and Throat (ENT) of HGH were the study population.

Inclusion and Exclusion Criteria
Children within ve years of age presumptive for tonsillitis with sore throat, red swollen tonsils, pain when swallowing, fever, cough, tiredness, malaise, and white pus-lled spots on the tonsils, swollen lymph nodes, pain in the ears or neck, and weight loss were included in the study. On the other hand, children who were on antibiotic treatment within the previous two weeks of sample collection, and had tonsillectomy were excluded from the study Sample size and Sampling The sample size was calculated using single population formula n = (Zα/2) 2 P (1-P)/d 2 where, n = sample size, Z = level of con dence according to the standard normal distribution, P = sample proportion and d = tolerated margin of error. Therefore; by taking Z(α/2) = 1.96 for a level of con dence of 95 %, P = 0.5 which is the maximum proportion of bacterial tonsillitis and 5% margin of error, the sample size was calculated as n = (1.96) 2 x 0.5 (1-0.5)/(0.05) 2 = 384. However, due to the lack of su cient throat swabs and incomplete questionnaire based data, only 374 children within ve years of age with tonsillitis were included in the study. Study participants were included conveniently. All children within ve years of age with tonsillitis attending at ENT department of HGH and who ful lled the inclusion criteria were included consecutively until the required sample size was reached.

Variables
Bacterial causes of tonsillitis was the dependent variable while demographic variables (child's age, mother's age, father's age, gender, residence, maternal education, paternal education, and parental occupation), clinical related variables (history of tonsillitis, current type of tonsillitis, number of previous tonsillitis, body temperature, sore throat, swollen tonsils, headache, swollen lymph nodes, di culty in swallowing, white exudates on the throat, weight loss, tonsillar structural change and history of drug use) were the independent variables. Moreover, patient related variables such as history of contact with someone who had cough, type of breast feeding, attending day care and school, living in overcrowded environment and exposure to wood biofuel were taken as the independent variables.

Data Collection
A structured questionnaire was used to collect data on demographic characteristics, clinical pro les and other variables. Demographic and other related pro les of the children were collected with face -to -face interviews of their caregivers. Clinical data of children with tonsillitis were screened by the attending pediatricians.
Throat Swab Sample Collection and processing Throat swabs were taken by the attending pediatricians from each patient using a sterile cotton swab. Visible exudates or hyperemic areas on tonsillar walls were swabbed with a sterile cotton swab while the tongue depressed by a wooden spatula when necessary.
All the swab samples were immediately transported to the Microbiology Department of HGH using Amie's transport medium (Oxoid, England). Swabs were simultaneously plated onto Blood Agar (BA), Chocolate Agar (CA), and MacConkey (MAC) Agar and incubated for 48 h at 37 o C. Chocolate Agar was incubated in a candle jar to get 5% CO 2 while BA and MAC were incubated at a normal atmosphere.

Identi cation of bacterial isolates
Pure colonies of the bacterial isolates were identi ed in to species following standard enzymatic and biochemical tests [6]. Small colonies, Gram positive cocci arranged in chain, forming complete hemolysis on BA and both coagulase and catalase negative were taken as S. pyogenes isolates. S. pneumoniae isolates were identi ed by Gram positive alpha-hemolytic small colonies on BA and were susceptible to optochin. S. aureus isolates were identi ed by Gram positive cluster forming glistering golden yellow colonies on BA and Mannitol Salt Agar (MSA) which were coagulase, catalase and oxidase positive. Moraxella catarrhalis were identi ed by large kidney shaped diplococci Gram negative grey to white hemispheric colonies on BA with both oxidase and catalase positive. K. pneumoniae and P. aeruginosa isolates were identi ed by standard manual biochemical tests.

Antimicrobial susceptibility testing
Susceptibilities of all the identi ed bacterial isolates to different antimicrobials were performed according to the criteria of Clinical and Laboratory Standards Institute (CLSI, 2019) using the Kirby-Bauer disc diffusion method on Mueller-Hinton Agar (MHA) (Himedia, India). The following drug discs were tested: ampicillin (10 µg), amoxicillin-clavulanic acid (20/10 µg), gentamicin (10 µg), clarithromycin (15 µg), erythromycin (15 µg), vancomycin (30 µg), o oxacin (5 µg) and cipro oxacin (5 µg). A loop full of culture was taken from a pure culture colony and transferred to a tube containing 5 ml of normal saline and mixed gently until it forms a homogenous suspension. The turbidity of the suspension was then adjusted to the turbidity of McFarland 0.5 (which carries 10 8 CFU/ml) and was swabbed on a dry surface of MHA plate (150 mm) using a sterile cotton swab. Antibiotic discs were dispensed using a single disc dispenser. Plates were then incubated for 24 h at 37 o C. Diameters of the zone of inhibition around the discs were measured using a digital caliper. The results of the zone of antibiotics were interpreted based on the 2019 CLSI guideline [7]. These antimicrobial drug discs were selected based on the frequent prescriptions of these drugs for the treatment of tonsillitis infection in the study area and using the 2019 CLSI [7].

Detection of Methicillin Resistant Staphylococcus aureus
Methicillin resistant Staphylococcus aureus was screened using oxacillin disk diffusion suceptibility testing. Pure colonies of Staphylococcus aureus were inoculated on MSA and 30 µg of oxacillin discs were impregnated on the plate and incubated for 18 h. The zone of inhibiton was measured by caliper. Measurements from the CLSI 2019 standard was followed as reference for interpretation. Accordingly, a zone of inhibition of ≤ 21 mm of oxacillin disk against Staphylococcus aureus isolates were considered as mecA positive and reported as methcillin resistant while if the zone of inhibition of oxacillin disc towards Staphylococcus aureus is ≥ 25mm, it was considered as mecA negative, and reported as methicillin sensitive [7].

Quality Control
Specimens were collected properly following standard bacteriological procedures. In order to prevent contamination, all the throat swab specimens were analyzed within two hours of collection. Culture media were checked for sterility. The media were tested every time after preparation for sterility checking by incubating a plate of each medium overnight in a different incubator than one used for culture. The performance of all the prepared culture media were checked by using American Type Culture Collection (ATCC) standard reference strains (S. aureus ATCC 29213, S. penumoniae ATCC 49618, and P. aeruginosa ATCC 27853).

Data analysis
Data were coded and analyzed using IBM SPSS statistics for windows version 25 (IBM Corp, Armonk, NY, USA). Univariate analysis was made to generate summary values for the most important variables. Logistic regression analysis was made to determine the association between dependent and independent variables. The generated data were compiled with frequency tables and other Page 5/22 statistical summary measures. Stepwise logistic regression model was used to nd factors associated with culture positive bacterial tonsillitis and statistical signi cance was set at p < 0.05.

Ethical Considerations
An ethical approval letter was obtained from the Institutional Review Board (IRB) of College of Medicine and Health Science (CMHS), Bahir Dar University. A permission letter was obtained from the Ministry of Health, Somaliland, and Hargeisa Group of Hospital (HGH).
Following well-versed about the purpose and importance of the study, written informed consent was obtained from children parents/guardians before collecting data. Information obtained during this study was kept con dential and used only for the study purpose. Bacteriological positive results were submitted to health pediatricians.

Characteristics of the study participants
A total of 374 children within ve years of age with tonsillitis took part with a response rate of 97.4%. Among them, 200 (53.5%) were males. Most (81.6%) of the children were urban dwellers. The age range of children was 2 to 5 years. Majority (3.7%) of the children were ve years old (mean = 4.1, median = 4). The children mother's age ranged from 20-45 years. Most (69%) of the parents were employee (Table 1). Overall, 120 (32.1%) of the children had culture con rmed bacterial tonsillitis. The proportion of bacterial tonsillitis was higher in males 76 (36.5%) than females 47 (27%). It was higher in urban 101 (33.1%) than rural 19 (27.5%) residents. The percentage of bacterial tonsillitis was higher (41.9%) in children from mother's unable to read and write than other groups (5.9-26.7%). Children from fathers who had higher educational attainment had lowest percentage of bacterial tonsillitis compared to others (Table 1).
Bacterial tonsillitis and clinical pro les  The percentage of bacterial tonsillitis was higher among children with a history of tonsillitis (55.2%) than others (12.4%). The percentage of tonsillitis was the highest (55.7%) in children with symptoms of chronic tonsillitis. Moreover, the percentage of bacterial tonsillitis was higher in children with tonsillar structural change (57.7%) than those without (28%). Children with swollen tonsils had higher percentage of culture con rmed tonsillitis (32.4%) than those without swollen tonsils (14.3%). The proportion of culture con rmed bacterial tonsillitis was higher among children who had weight loss (46.9%) than the counters (22.9%) ( Table 2).

Bacterial tonsillitis in relation to other variables
Overall, 96 (25.7%) and 228 (61%) of children were exclusively breastfed and had history of contact with coughing patients, respectively. On the other hand, 86 (23%) and 282 (75.4%) of children were daycare center attendees and school attendees, respectively. Most of the children lived in a crowded house (71.7%) and 88.5% had exposure to biofuel (Table 3). The proportion of culture con rmed bacterial tonsillitis was higher among children who had history of contact with coughing patients (61%) than the counters (24%). Daycare center attendee children had a higher (39.5%) percentage of bacterial tonsillitis than others (29.9%). Moreover, school attending children had a higher (52.2%) percentage of bacterial tonsillitis than the counterparts (25.5%).
The proportion of bacterial tonsillitis was higher among children who had exposure to biofuel (35.3%) than others (7%) ( Table 3).

Multiple Drug Resistant (MDR) pro les of bacterial isolates
Overall, 72 (50.4%) of the bacterial species were MDR and 52.6% of S. pyogenes were MDR. The MDR pro le of S. aureus, S. pneumoniae and Klebsiella pneumoniae isolates were 18 (42.9%), 6 (60%) and 3 (50%), respectively (Table 6). Based on multivariable analysis, bacterial tonsillitis was signi cantly associated with di culty of swallowing (AOR = 6.99, CI = 3.56-13.13), weight loss (AOR = 0.33, CI = 0.186-0.597), attending school (AOR = 2.98, CI = 1.64-5.42), history of tonsillitis (AOR = 0.12, CI = 0.06-0.21) and exposure to biofuel (AOR = 0.19, CI = 0.04-0.84). Children who had di culty of swallowing were 7 times more likely to become culture positive for bacterial tonsillitis, compared to children who did not have di culty of swallowing. Likewise, school attending children were 3 times more likely to have con rmed bacterial tonsillitis compared to non-attendees. Children with history of tonsillitis were more likely to had bacterial tonsillitis than those without history of tonsillitis. Similarly, children who had weight loss and exposure to biofuel were more likely to become culture con rmed tonsillitis compared to those who did not have weight loss and exposure to biofuel (Table 7).

Discussion
Tonsillitis has considerably a negative impact on the patients' quality of life and has a signi cant burden on public health. Untreated childhood tonsillitis can leads to peritonsillar abscess, tonsillar stones, and rheumatic fever. Therefore, identi cation and antimicrobial susceptibility of bacterial causes of tonsillitis is essential to curtail for the treatment of tonsillitis. However, patients with tonsillitis managed empirically in health care's settings of Somaliland. Therefore, this study presents the rst report of the prevalence of culture con rmed bacterial tonsillitis and the antimicrobial resistance pro les of isolates in HGH.
In this study, 32.1% of children within ve years of age had culture con rmed bacterial causes of tonsillitis. Due to the lack of previous data in Somaliland, comparison of countrywide results was not possible. However, the prevailing magnitude of tonsillitis is higher than similar studies with a prevalence of 11.3% in Ethiopia [8], 20.6% in Tanzania [1], 21.6% in Norway [9], and 19% in Bangladesh [10]. On the other hand, the existing prevalence of bacterial causes of tonsillitis from this study was lower than studies done in the United Kingdom (79%) [11], Trinidad (62.5%) [12], India (72%) [4], Saudi Arabia (65%) [13], Benin (73.97%) [5] and Ethiopia (51%) [14]. The lower rate of bacterial tonsillitis in the present study compared to other developing countries might be attributed to differences in geography, community living status and hygienic practices, host factor and educational level of the parents.
The prevalence of bacterial causes of tonsillitis in children within ve years of age was higher in males than in females which is similar to studies from India [2] and Nigeria [15]. The variations on the percentage of tonsillitis between genders of the children could be due to the fact that males spend more time in outdoor than females. The percentage of tonsillitis was higher among children living in urban than rural areas. This was similar with studies done in India [2], and Ethiopia [14]. This might be due to variation in: encountering infected people, exposure with air pollution from biofuel use, schooling and house crowding.
In the present study, Staphylococcus aureus was the second most frequent isolate of bacterial causes of tonsillitis with a rate of 29%. This could be due to the persistence of S. aureus in the tonsillar tissues, treatment with antimicrobials and antibiotic resistance. Moreover, S. aureus has the potential to form bio lm which results recurrent and chronic infection as well as treatment failure. The isolation of S. aureus as the main agent of tonsillitis has been reported by several authors in Ethiopia [16], Brazil (40%) [24], Trinidad (68.9%) [12], and Nigeria (32.1%) [15].
In the present study, there is high proportion of mixed infections particularly with S. pyogenes and S. aureus, S. pneumoniae and S. aureus and S. pneumoniae and M. catarrhalis. These co-infections of the tonsils may contribute to the severe in ammatory process and the failure of penicillin and ampicillin therapy which nally results recurrent infection, tonsillectomy, rheumatic fever and other complications [34,38].
The resistance of the isolates to ampicillin was 91.6% and 14.7% to the association of amoxicillin and clavulanate. The higher resistance to ampicillin by all of the bacterial isolates might be due to production of beta lactamase enzyme as well as abuse and excessive use of cheap drugs, which can be afforded and administered without a physician's guidance. This is a major concern that limits the use of this common therapeutic option in clinical practice in developing countries. The rate of penicillin resistance is comparable with reports from Nigeria (100%) [5].
The resistance rate of 94.9% of S. pyogenes to ampicillin is worrisome. As B -lactam antibiotics are the drug of choice for strep throat. The percentage of S. pyogenes resistant to gentamicin (42.3%) and o oxacin (43.6%) in the present study was comparable to studies done in Iran (32.2%) [44]. The resistance of Streptococcus pyogenes to the above drugs might be due to the enzymatic inactivation mediated by aminoglycoside-modifying enzymes (AMEs), and point mutations in the quinolones resistance-determining region (QRDR).
One of the major worries when determining resistance pro les of isolates is the availability of MDR strains. In this study, half of the bacterial isolates were MDR. This is a series problem for children within ve years of age in Somaliland. Children involved in the study area were outpatients and they might have constant contact with other children and their family. Moreover, in the study area there is no routine culture and antimicrobial susceptibility testing and management of children with tonsillitis is empirical. These may results repeated infections of the tonsils, pyogenic meningitis, rheumatic fever, lower respiratory tract infections and di culty to select the effective antibiotics. Furthermore, existence of MDR isolates demonstrates persistence of the bacteria and possibility of antimicrobial resistance dissemination and recurrence of infection [34].
The percentage of MDR S. pneumoniae (60%) in this study was higher than studies from Poland (52.9%) [50], Lithuania (12.5%) [48] and Vietnam (35%) [51]. In this study, all isolates of Pseudomonas aeruginosa were MDR (100%) which is concurrent to a study in Brazil (100%) [31]. These high proportions of MDR among the isolates might be due to productions of beta-lactamase enzyme by Pseudomonas aeruginosa and production of Penicillin binding proteins in Streptococcus pneumoniae.
In the present study, di culty of swallowing is one of the predictors of bacterial tonsillitis in children within ve years of age. Similar ndings were reported in India [2] and Lithuania (48). History of tonsillitis was also a predictor variable in this study which was similar to studies done in Ethiopia [9], and Yemen [52]. These might be due to cohabitations of the tonsils by multiple bacterial isolates as depicted in Table 4 and failure of penicillin and ampicillin therapy.
Weight loss was also another predictor for bacterial tonsillitis in this study in which similar studies were reported in Iran [53], and Germany [54]. Furthermore, attending school was a risk factor for tonsillitis in this study similar to studies done in Uganda [55], and Australia [56]. This might be due to overcrowding during schooling among children where carrier children can easily interact with healthy children.

Limitations Of The Study
This study provided the rst report of data on bacterial causes of tonsillitis and antibiotic resistance pro les of the isolates from children within ve years of age with tonsillitis at Hargeisa Group of Hospital. However, the study limited to identifying nonbacterial causes of tonsillitis.

Conclusions
High prevalence of tonsillitis with MDR pathogens, MRSA and mixed isolates were found. S. pyogenes followed by S. aureus and S. pneumoniae were the most frequent isolates. Most of the bacterial isolates were resistant to ampicillin. However, amoxicillinclavulanic acid and cipro oxacin are the least resisted drugs. Therefore, the result points the signi cance of culture and antimicrobial susceptibility testing in the diagnosis and treatment of any form of tonsillitis for the selection of the effective antibiotics against several pathogenic agents of tonsillitis and rational use of antimicrobials.
Further investigation to identify nonbacterial causes of tonsillitis, conducting studies covering larger geographical areas to draw the magnitude and topographic variations are needed to control the spread of tonsillitis among children within ve years age.