Global Burden of Haemophilus influenzae Infections.

Haemophilus influenzae is a small, capnophilic, pleomorphic, facultatively anaerobic, Gram-negative coccobacillus, which belongs to the family Pasteurellaceae.(1) It is an opportunist that commonly colonises the nasopharynx in the human host and causes many infections including acute sinusitis, otitis media, conjunctivitis, community-acquired pneumonia, septic meningitis, septicemia, peritonitis and arthritis.(2, 3) This bacterium is a human pathogen of clinical and public health importance worldwide, only found in humans.(4)Depending on presence or absence of a polysaccharide capsule, H. influenzae is divided in 6 serotypes (Hia, Hif) and non-encapsulated non-typeable H. influenzae (NTHi). Severe infection by Hib strains of the bacterium Haemophilus influenzae type B (Hib) were dramatically reduced in vaccinated populations following its introduction in the 1990s. But, NTHi serotypes have become the preferred clinical isolates globally.(5, 6) NTHi currently causes most cases of otitis media, sinusitis, and pneumonia in immunized individuals, especially in children, elderly and those who have chronic respiratory diseases.(7) NTHi causes 78% of invasive cases of H. influenzae in adults greater than 65 years of age in the United States and 89% of NTHi cases among adults greater than 85 years of age.(8)

The impact of H. influenzae infection worldwide is still significant. Respiratory tract infections (RTIs) are one of the most common and hazardous infections, responsible for more than 50 million deaths worldwide annually.(9) H. influenzae remains a major cause of pneumonia and meningitis in children in low- and middle-income countries and is a major cause of under-five mortality. (8) The resistance mechanisms of H. influenzae are varied and continually changing; the most prevalent is the production of β-Lactamase, which is mediated mainly by TEM-1 and TEM-163 enzymes, and is the most common resistance mechanism against penicillins and cephalosporin.(10) Amino acid substitutions in the penicillin binding protein 3 (PBP3) due to point mutations in the ftsI gene are an emerging problem, such as β-lactamase-negative ampicillin resistant (BLNAR) strains.(11) Haemophilus influenzae is particularly clinically relevant, because the bacteria is sensitive to the effect of glucocorticoids used in asthma treatment, which involves inhibiting biofilm formation and blocking the resistance to antibiotics.(12).

Study Design and Setting A cross sectional study was designed and it was a prospective one, conducted at Microbiology Department, DHQ hospital Abbottabad from 1st January 2024 to 1st December 2025 for a duration of 12 months. The hospital acts as a tertiary care health centre for Abbottabad District and the neighbouring localities of Khyber Pakhtunkhwa province of Pakistan. Sample Collection One hundred and three patients with clinically diagnosed community-acquired respiratory tract infections (cARTI) were sputum or nasopharyngeal swab samples collected and 130 H. influenzae isolates were obtained. Inclusion Criteria Patients with clinical features and symptoms of Respiratory tract infection (RTI) like cough, fever, sputum production, shortness of breath, chest pain) Age ≥ 1 year Culture of H. influenzae from sputum or nasopharyngeal swab is positive Consent (obtained from patient/legal guardian) Exclusion Criteria Patients on antibiotic therapy within 48 hours of sample collection (to avoid culture negativity) Patients with limited systems (HIV/AIDS, chemotherapy, transplant, primary immunodeficiencies) Patients with hospital acquired infections (onset>48 hrs after admission) Patients with a history of CF or Bx. Pregnant women Patients who are not informed or unwilling to consent. Patients with limited clinical and/or demographic information The course will cover the processes involved in preparing specimens for analysis, as well as the identification of bacteria. Sputum and nasopharyngeal swab specimens were collected by using standard sterile procedure and carried to microbiology laboratory within 2 hours of collection. Chocolate agar was used for inoculation and grown at 37°C under a 5–10% CO₂ environment for 24–48 hours. Identification of H. influenzae was done through: Gram stain morphology (small gram-negative coccobacilli) On chocolate agar, the colony of a bacterium appears translucent and is surrounded by a zone of area of transparency, which often has a "fried egg" appearance. Use of porphyrin test and satellite test with S. aureus for the requirement of X (hemin) and V (NAD) factors. The API NH or VITEK 2 identification system confirmation is required as necessary. Antimicrobial Susceptibility Testing The antimicrobial susceptibility testing was done on Haemophilus Test Medium (HTM) by the Kirby-bauer disc diffusion method following the CLSI guidelines (M100, 34th edition). The following antibiotics were tested: ampicillin (10µg), amoxicillin (25µg), amoxicillin clavulanic acid (20/10µg), cefotaxime (30µg), ceftriaxone (30µg), cefuroxime (30µg), imipenem (10µg), ciprofloxacin (5µg), levofloxacin (5µg), azithromycin (15µg), clarithromycin (15µg), tetracycline (30µg), trimethoprim/sulfamethoxazole (1.25/23.75µg), and chloramphenicol (30µg).The minimum inhibitory concentration (MIC) for selected antibiotics was determined by broth microdilution method in HTM broth and two-fold antibiotic dilutions. CLSI breakpoints were used for the interpretation of results for H. influenzae. The strains H. influenzae ATCC 49247 and E. coli ATCC 25922 were used for quality control. β-Lactamase Detection The production of β-lactamase was measured by the chromogenic cephalosporin (nitrocefin) test. A positive result was suggested by a colour change from yellow to red in 5 to 10 minutes after adding nitrocefin solution to colonies of bacteria. Biofilm Formation Assay The crystal violet staining technique was used to evaluate the ability of the bacteria to form biofilm on 96-well polystyrene microplates. H. influenzae isolates were grown for 24 hours at 37°C with 5% CO₂ in HTM broth. Plates were washed after incubation and fixed with methanol, stained with 0.1% crystal violet and the optical density was measured at OD 570 nm. Isolates were categorized as non-biofilm forming (OD ≤ ODc), weak biofilm-forming (ODc < OD ≤ 2 × ODc), moderate biofilm-forming (2 × ODc < OD ≤ 4 × ODc) or strong biofilm-forming (OD > 4 × ODc). Pharmacokinetic/Pharmacodynamic Analysis Pharmacokinetic parameters were taken from the published literature for seven antibiotics; results from a Monte Carlo simulation (10,000 iterations) were used to calculate PTA and CFR values. The PK/PD targets used for Haemophilus influenzae were: T > MIC ≥30–35% for penicillins, ≥40–50% for cephalosporins, and ≥20–25% for carbapenems; and AUC/MIC ≥30–40 for fluoroquinolones and ≥25 for macrolides. Data Collection Data were collected on a structured proforma which included the following: age, gender, residence, clinical diagnosis, duration of symptoms, use of antibiotics within the last three months, comorbidities (diabetes, asthma, COPD), and history of corticosteroid use. Statistical Analysis Data were analysed with the SPSS version 26.0. Categorical data was presented as frequencies and percentages. Data for the continuous variables were presented as mean ± SD. Chi-square test was used to compare proportions between groups. The results were considered statistically significant if the p value was < 0.05. Ethical Considerations The study was approved by the Institutional Review Board (IRB/DHQ/2024-01) of DHQ Hospital Abbottabad. All subjects or their parents/guardians gave informed consent before sample collection. There was complete confidentiality of patients throughout the study.

During the study period (January 2024 to December 2025), 130 H. influenzae isolates from 130 patients with community-acquired respiratory tract infections (cATIs) were recovered. Demographic and clinical data on the study population are shown in Table 1.

Table 1. Demographic and Clinical Characteristics of Study Participants (N = 130)

Most patients (29.2%) were children (1-5 years) and another large proportion of patients (24.6%) were adults (19–40 years). There were more men than women (55.4% vs. 44.6%). The most frequent clinical diagnosis was CAP (40.0%) followed by acute sinusitis (16.9%) and acute otitis media (15.4%). The antibiotic use history was reported in 40.0% and corticosteroid use history was reported in 18.5% of the patients in the last 3 months.

Antimicrobial Susceptibility Profile

The antimicrobial susceptibility profile of the 130 H. influenzae isolates is presented in Table 2.

Table 2. Antimicrobial Susceptibility Profile of H. influenzae Isolates (N = 130)

High susceptibility (>85%) was observed for imipenem (98.5%), ceftriaxone (95.4%), cefotaxime (93.8%), levofloxacin (93.8%), azithromycin (90.8%), ciprofloxacin (90.8%), chloramphenicol (90.8%), and amoxicillin/clavulanic acid (89.2%). Cefuroxime demonstrated 86.2% susceptibility. Susceptibilities of 14.6% and 52.3% were alarmingly low for trimethoprim/sulfamethoxazole and tetracycline respectively. The test results for susceptibility to ampicillin and amoxicillin were 64.6% and 63.1%, respectively. Of these, 124 were β-lactamase negative (95.4%) and 6 were β-lactamase positive (4.6%). 100% of the isolates that were β-lactamase-positive were resistant to ampicillin and amoxicillin. Amoxicillin/clavulanic acid resistance was however found in 4 out of 6 isolates that were β-lactamase positive and 2 isolates that were β-lactamase intermediate.

Biofilm Formation and Associated Factors

Biofilm formation was assessed in all 130 isolates. The results are presented in Table 3.

Table 3. Biofilm Formation Capacity and Associated Clinical Factors

In general, 68 (52.3%) isolates showed biofilm formation ability. NTHi isolates had significantly higher biofilm formation capacity than encapsulated serotypes (56.4% vs. 37.5%, p = 0.018). In patients who had a history of corticosteroid use (n = 24), 18 (75.0%) formed biofilm, while 50 (47.2%) of non-users formed biofilm (p = 0.012). The isolates with the ability to form biofilms had significantly higher MICs for azithromycin (MIC₉₀: 8 vs. 2 μg/mL, p < 0.001) and clarithromycin (MIC₉₀: 16 vs. 4 μg/mL, p < 0.001).

Monte Carlo simulation was used to determine the probability of target attainment (PTA) and cumulative fraction of response (CFR) for commonly used antibiotics. The CFR was ≥95% for amoxicillin/clavulanic acid (95–96%), cefotaxime (98%), ceftriaxone (97%), imipenem (99%), ciprofloxacin (94%) and levofloxacin (97%). Amoxicillin and ampicillin had low CFRs (6–8%) because they were resitant to β-lactamase. Azithromycin and clarithromycin demonstrated suboptimal CFR (76% and 68%, respectively), in part because of biofilm mediated-tolerance.

In this prospective study of 1 year duration at DHQ Hospital Abbottabad, the first comprehensive evaluation of antimicrobial susceptibility and pharmacokinetic/pharmacodynamic profiles of Haemophilus influenzae in the Khyber Pakhtunkhwa region has been done. The number of 130 isolates obtained from community-acquired respiratory tract infections (CARTIs) demonstrates that H. influenzae is still an important bacterial pathogen, especially in children aged 1–5 years (29.2%) and adults with comorbidities including diabetes (16.9%), asthma (15.4%) and COPD (11.5%). The majority were NTHi strains (95.4% β-lactamase negative) consistent with the world situation after Hib vaccine introduction and similar to SOAR 2018-2021 Pakistan data (95.5% β-lactamase negative) indicating the impact of vaccination and the growing problem of NTHi as the clinical dominant strain. Their susceptibility to amoxicillin/clavulanic acid (89.2%), third generation cephalosporins (>93%) and fluoroquinolones (>90%) makes them viable for continued use as empiric therapy for community-acquired respiratory tract infections.(13) These results corroborate national SOAR data with a susceptibility of >83.6% for these agents. But, the susceptibility rate to trimethoprim/sulfamethoxazole and tetracycline are alarmingly low (14.6 and 52.3%, respectively) and so these are not suitable for empiric therapy.(14) Trimethoprim/Sulfamethoxazole resistance rate is almost the same as the national SOAR of 13.4% indicating high resistance rates throughout Pakistan, which is likely due to indiscriminate and overuse of this inexpensive antibiotic.(15) The same trend has been observed in other developing countries such as India (15-20% susceptibility) and Bangladesh (12-18% susceptibility), suggesting that there is a regional problem that needs coordinated antimicrobial stewardship measures.(16) Clinicians should be strongly discouraged from treating suspected H. influenzae infections with trimethoprim/sulfamethoxazole.(17) This is similar to the historical data from Pakistan (2.6–3.2%) and SOAR 2018–21 data (4.5%) with the β-lactamase positivity rate being 4.6%. This represents a marked difference from the findings of other studies from areas with high levels of β-lactamase production (more than 30%) such as in parts of Southeast Asia and Eastern European countries where antibiotic selection pressure is higher. Lower prevalence in Pakistan could be because of less antibiotic selection pressure, difference in strains circulating or more frequent circulation of NTHi strains which are less likely to acquire the β-lactamase gene.(18) All β-lactamase-positive isolates were ampicillin- and amoxicillin-resistant, and 4 of 6 were still amoxicillin susceptible, demonstrating the effectiveness of β-lactamase inhibitor combinations.(19) The presence of ampicillin resistant and β-lactamase negative isolates (around 8-10%) indicates the emergence of BLNAR strains as a result of mutations in the PBP3, which is a concern worldwide.(20) The findings from SOAR highlighted that majority of ampicillin resistant isolates in Pakistan were BLNAR, thereby highlighting the importance of molecular surveillance for identification of PBP3 mutations that cannot be detected by routine phenotypic testing.(21) The high proportion of isolates (52.3%) with the ability to form biofilm has clinical significance. Antibiotic tolerance and persistence in NTHi infections is a well-documented phenomenon and is responsible for chronicity and for the recurrence of respiratory infections, including in COPD and CF patients and chronic OM patients.(22) This is similar to worldwide studies which reported 45-60% biofilm formation with NTHi, and a recent study in China with pediatric respiratory isolates reporting a 48.5% biofilm formation rate.(23) The correlation between corticosteroid use and increased biofilm formation (75.0% vs 47.2%, p = 0.012) is especially interesting, as a recent study published in EMBO Molecular Medicine in 2023 showed that H. influenzae biofilm formation is modulated by glucocorticoids, while also inducing azithromycin resistance.(24) This can present a therapeutic challenge because steroids are often used for treating exacerbations of asthma and COPD which may lead to bacterial persistence and lower anti-microbial effectiveness. In corticosteroid treated patients, other antibiotics and/or adjunctive therapy may be considered.(25) The higher MICs of macrolides for biofilm producers (azithromycin MIC₉₀: 8 μg/mL vs. 2 μg/mL; clarithromycin MIC₉₀: 16 μg/mL vs. 4 μg/mL) indicate that routine susceptibility testing for these agents could underestimate clinical effectiveness when used for biofilm producers.(26) This result is in line with the observed high rates of bacteriologic failure of azithromycin (53%) and cefaclor (52%) in H. influenzae infections, which underscores the difference between in-vitro and in-clinical efficacy.(27) Our results are similar to the national data on SOAR, with almost the same rate of β-lactamase negativity (95.4% vs. 95.5%) and trimethoprim/sulfamethoxazole resistance (14.6% vs. 13.4%), reflecting resistance pattern in Abbottabad is representative of that of Pakistan. The increased susceptibility reports for cephalosporins and macrolides at the DHQ Abbottabad might be due to regional variation in use of antibiotics, lower antibiotic pressure, or variation in circulating strains.(28) Overall resistance trends, however, are consistent and national empiric therapy guidelines should be followed without fear of resistance, though they should be continually monitored in this region for the emergence of resistance.(29) As shown in this study, the pharmacological principles (PK/PD optimization) and the microbiological information (resistance mechanisms and biofilm formation) offer a comprehensive approach to fight H. influenzae infections. Combining these disciplines could break out of the traditional strategies and use more complex ones with respect to drug exposure, host-pathogen interactions and resistance mechanisms. Further work should include molecular characterization of resistance mechanisms, development of anti-biofilm strategies and potential clinical trials which correlate PK/PD parameters with clinical outcomes. Incorporation of these integrated strategies along with strong antimicrobial stewardship programming will play an important role in maintaining the effectiveness of antibiotics and improve patient outcomes at DHQ Hospital Abbottabad and beyond.

This prospective study of community-acquired respiratory infections for a period of one year at DHQ Hospital, Abbottabad, show that Haemophilus influenzae is still an important cause of the disease in the area and high susceptibility to Amoxicillin/Clavulanic acid, third generation cephalosporins, fluoroquinolones and Azithromycin was noted with alarmingly low efficacy of Trimethoprim /Sulfamethoxazole and Tetracycline. Our results indicate that the percentage of biofilm formation is significantly increased in NTHi strains and in antibiotic exposed patients as compared to non-NTHi strains and non-antibiotic exposed patients, respectively (52.3%). This highlights the importance of biofilm-targeted therapeutic strategies and careful antibiotic selection. The results of the PK/PD analysis indicate that amoxicillin/clavulanic acid, cefotaxime, ceftriaxone, imipenem, and levofloxacin are the most ideal first-line drugs. A combination of pharmacology and microbiological surveillance has been developed, which is an overall approach to improving antimicrobial treatment and maintaining the effectiveness of antibiotics in this area.

Limitations

The present study has several limitations: The study was conducted in a single hospital, the DHQ Hospital Abbottabad, thus limiting the generalizability of the results to other regions of Pakistan; molecular characterisation of the mutations of PBP3, the prevalence of the efflux pumps and serotyping was not performed, so that the resistance mechanisms could not be analysed comprehensively; the cross-sectional design was used which does not allow the analysis of the temporal trend of resistance; the PK/PD modelling was based on literature-derived pharmacokinetic parameters that may not reflect the characteristics of the local population; and the in vitro biofilm assays may not reflect the behaviour of in vivo biofilms in the respiratory tract. Secondly, the limited sample size of 130 isolates, adequate for descriptive analysis, restricts subgroup analyses and the ability to detect rare resistance phenotypes.

Recommendations

Based on this study, we suggest: implementing an antimicrobial stewardship program at DHQ Hospital Abbottabad for dissemination of empiric therapy guidelines (excluding trimethoprim/sulfamethoxazole for H. influenzae infection); starting an antimicrobial surveillance program at the regional level to monitor the resistance trends across Khyber Pakhtunkhwa province; molecular surveillance of PBP3 mutations and efflux pump genes to detect emerging BLNAR strains; clinical trials to evaluate anti-biofilm strategies as an adjunctive treatment for chronic NTHi infections; and local development of PK/PD models using population pharmacokinetic data of Pakistani patients to optimize dosing regimens. More cooperation among microbiology, pharmacology, and clinical medicine is needed to ensure the effectiveness of existing antibiotics.

1. Roberts L. Neisseria species and Moraxella catarrhalis. Textbook of Diagnostic Microbiology-E-Book: Textbook of Diagnostic Microbiology-E-Book. 2022:371.
2. Xiao J, Su L, Huang S, Liu L, Ali K, Chen Z. Epidemic trends and biofilm formation mechanisms of Haemophilus influenzae: Insights into clinical implications and prevention strategies. Infection and drug resistance. 2023:5359-73.
3. Murray PR. Murray's Basic Medical Microbiology E-Book: Foundations and Cases: Elsevier Health Sciences; 2023.
4. Mancuso G, Midiri A, Gerace E, Biondo C. Bacterial antibiotic resistance: the most critical pathogens. Pathogens. 2021;10(10):1310.
5. Lin J, Wang Y, Lin C, Li R, Wang G. High prevalence of group III-like mutations among BLPACR and first report of Haemophilus influenzae ST95 isolated from blood in China. Infection and drug resistance. 2023:999-1008.
6. Tønnessen R, Garcia I, Debech N, Lindstrøm JC, Wester AL, Skaare D. Molecular epidemiology and antibiotic resistance profiles of invasive Haemophilus influenzae from Norway 2017–2021. Frontiers in microbiology. 2022;13:973257.
7. Khattak Z, Anjum F. Haemophilus influenzae. 2020.
8. Slack M, Cripps A, Grimwood K, Mackenzie G, Ulanova M. Invasive Haemophilus influenzae infections after 3 decades of Hib protein conjugate vaccine use. Clinical microbiology reviews. 2021;34(3):10.1128/cmr. 00028-21.
9. Oliva J, Terrier O. Viral and bacterial co-infections in the lungs: dangerous liaisons. Viruses. 2021;13(9):1725.
10. Singh A, Shahid M, Sami H, Shadab M, Khan HM. Class A Type Β-Lactamases. Beta-Lactam resistance in Gram-Negative bacteria: threats and challenges: Springer; 2022. p. 35-80.
11. Yalçın M, Tristram S, Bozdoğan B. Use of trans-complementation method to determine the effects of various ftsI mutations on β-lactamase-negative ampicillin-resistant (BLNAR) Haemophilus influenzae strains. Archives of Microbiology. 2023;205(1):27.
12. Brown MA, Jabeen M, Bharj G, Hinks TS. Non-typeable Haemophilus influenzae airways infection: the next treatable trait in asthma? European Respiratory Review. 2022;31(165):220008.
13. Zhao C, Yang S, Zhang F, Wang Z, Zhang Y, Wang X, et al. Antimicrobial resistance trends of the most common causative pathogens associated with community-acquired respiratory infections in China: 2009–2018. Infection and Drug Resistance. 2022:5069-83.
14. Motaung MA. Antimicrobial sensitivity testing of stored bacterial isolates from 2000 to 2021 in the agricultural sector: University of South Africa (South Africa); 2023.
15. Fahlin C. Trends in antibiotic resistance amongst pathogens causing enteric fever in India, Nepal and Pakistan 2012-2021. 2023.
16. Sihombing B, Bhatia R, Srivastava R, Aditama TY, Laxminarayan R, Rijal S. Response to antimicrobial resistance in South-East Asia region. The Lancet Regional Health-Southeast Asia. 2023;18.
17. Su P-Y, Cheng W-H, Ho C-H. Molecular characterization of multidrug-resistant non-typeable Haemophilus influenzae with high-level resistance to cefuroxime, levofloxacin, and trimethoprim-sulfamethoxazole. BMC microbiology. 2023;23(1):178.
18. Farzana R. A genomic approach to understanding the molecular epidemiology and clinical burden of multi-drug resistant Enterobacterale Infections in Bangladesh: Cardiff University; 2020.
19. Vrancianu CO, Gheorghe I, Dobre E-G, Barbu IC, Cristian RE, Popa M, et al. Emerging strategies to combat β-lactamase producing ESKAPE pathogens. International journal of molecular sciences. 2020;21(22):8527.
20. Heliodoro CIM, Bettencourt CR, Bajanca-Lavado MP, Infection PGftSoHiI. Molecular epidemiology of invasive Haemophilus influenzae disease in Portugal: an update of the post-vaccine period, 2011–2018. European Journal of Clinical Microbiology & Infectious Diseases. 2020;39(8):1471-80.
21. Manandhar S. The molecular epidemiology of bloodstream infections caused by non-Salmonella Gram-negative bacteria in a tertiary hospital in Nepal with focus on neonatal sepsis and Enterobacter species: University of Oxford; 2022.
22. Weeks JR, Staples KJ, Spalluto CM, Watson A, Wilkinson TM. The role of non-typeable Haemophilus influenzae biofilms in chronic obstructive pulmonary disease. Frontiers in cellular and infection microbiology. 2021;11:720742.
23. Huang S, He J, Zhang Y, Su L, Tong L, Sun Y, et al. The correlation between Biofilm-Forming ability of community-acquired methicillin-resistant Staphylococcus aureus isolated from the respiratory tract and clinical characteristics in children. Infection and Drug Resistance. 2022:3657-68.
24. Alsayed AR, Abed A, Khader HA, Al-Shdifat LM, Hasoun L, Al-Rshaidat MM, et al. Molecular accounting and profiling of human respiratory microbial communities: toward precision medicine by targeting the respiratory microbiome for disease diagnosis and treatment. International Journal of Molecular Sciences. 2023;24(4):4086.
25. Lopez-Campos JL, Miravitlles M, de la Rosa Carrillo D, Cantón R, Soler-Cataluña JJ, Martinez-Garcia MA. Current challenges in chronic bronchial infection in patients with chronic obstructive pulmonary disease. Journal of clinical medicine. 2020;9(6):1639.
26. Yue C, Shen W, Hu L, Liu Y, Zheng Y, Ye Y, et al. Effects of tigecycline combined with azithromycin against biofilms of multidrug-resistant Stenotrophomonas maltophilia isolates from a patient in China. Infection and Drug Resistance. 2021:775-86.
27. Makuta G. The feasibility of measuring bacterial load in exhaled breath of children with pneumonia and empyema using a novel point-of-care test: The BALLOON Study: Cardiff University; 2022.
28. Waseem M, Rafiq M, Munir A, Kamal Z, Aziz N, Iqbal MJ. Antibiotics prescription pattern in COVID-19 patients presenting in DHQ Teaching Hospital Sahiwal; is Pakistan heading towards post-COVID antibiotic resistance era? Journal of Rawalpindi Medical College. 2021;25(1).
29. Auzin A, Spits M, Tacconelli E, Rodríguez-Baño J, Hulscher M, Adang E, et al. What is the evidence base of used aggregated antibiotic resistance percentages to change empirical antibiotic treatment? A scoping review. Clinical Microbiology and Infection. 2022;28(7):928-35.