Abstract
Introduction
The β-lactam susceptibility of Haemophilus influenzae varies globally due to resistance mechanisms such as β-lactamase production or alteration in Penicillin-Binding Protein 3 (PBP3). Monitoring and understanding these resistance trends are crucial for guiding effective treatment strategies. We described the global antimicrobial susceptibility patterns of H. influenzae using the SENTRY Antimicrobial Surveillance Program, a large database designed to monitor antimicrobial resistance patterns.
Methods
Antimicrobial susceptibility testing was performed using broth microdilution as the reference method. Demographics and phenotypic antimicrobial resistance of H. influenzae from over 150 medical centers representing 51 countries were analyzed between 2013 and 2022. The guidelines established by the Clinical and Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) were adopted to interpret antibiotic breakpoints. A nitrocefin test was employed to assess the production of β-lactamase. For statistical analysis, we used Pearson’s Chi-square or Fisher’s exact test, considering p ≤ 0.05 as significant.
Results
A total of 13,869 H. influenzae isolates were analyzed; the majority came from the USA (52.2%) and the UK (4.3%). The most affected groups were males under 18 years old and those over 65 years old. β-lactamase was produced in 24.1% of the isolates, with significant variations in antibiotic resistance across regions. Asia and the West Pacific exhibited the highest resistance rates to β-lactams and other antibiotics, including β-lactamase-negative ampicillin-resistant (9.4%) and β-lactamase-positive amoxicillin-clavulanic acid-resistant (10.9%) isolates, while also having the lowest rates of intensive care unit (ICU) admissions (14.9%) and invasive infections (0.4% bloodstream infections and no central nervous system infections). Ceftriaxone and piperacillin-tazobactam were the most in vitro active antibiotics (100% susceptibility based on the CLSI breakpoints, 99.1% and 99.8% susceptibility, respectively, based on the EUCAST breakpoints). β-lactamase producing isolates had reduced susceptibility to amoxicillin-clavulanate (91.4% vs. 95.4%), trimethoprim-sulfamethoxazole (61.2% vs. 66.5%), clarithromycin (85.7% vs. 87.8%), azithromycin (97.1% vs. 98.9%), and tetracycline (95% vs. 99.6%) compared to β-lactamase negative isolates. Contrarily, susceptibility to fluoroquinolones was higher among β-lactamase-producing isolates compared to β-lactamase-negative isolates (99.6% vs. 94.1%, respectively). The in vitro susceptibility of H. influenzae isolates was higher to azithromycin compared to clarithromycin (98.4% vs. 87.2%), respectively.
Conclusions
Reliance on the nitrocefin test alone to predict susceptibility to β-lactams, i.e., ampicillin, could lead to false susceptibility and a higher chance of treatment failure. The high prevalence of BLNAR and BLPACR in Asia and the West Pacific, along with rising resistance to other antibiotics, highlights the need for reliable diagnostic testing and stringent containment protocols. Empirical therapy with clarithromycin for H. influenzae infections should be used with caution, especially in high resistance settings.
Keywords: Amoxicillin-clavulanate, Ampicillin, Antimicrobial resistance, Β-lactamase, Ftsl gene, Global epidemiology, Haemophilus influenzae
Introduction
Haemophilus influenzae is a small Gram-negative coccobacillus that can cause various infections, ranging from mild respiratory tract infections to more severe diseases such as meningitis, bloodstream infections, septic arthritis, osteomyelitis, and endocarditis [1]. While the H. influenzae type b (Hib) vaccine has significantly reduced the incidence of invasive H. influenzae type b disease, non-typeable strains of H. influenzae have emerged as important pathogens in adults and children [2].
Resistance to antibiotics, particularly ampicillin, has become a significant concern in managing H. influenzae infections [3]. The World Health Organization lists ampicillin-resistant H. influenzae as a medium-priority pathogen due to its high disease burden and the growing issue of antibiotic resistance [4]. Resistance to ampicillin in H. influenzae is often due to the production of TEM-1 (mostly) or ROB-1 (less common) type-β-lactamases [3]. In most cases, the nitrocefin test, which detects the presence of β-lactamase, represents a rapid chromogenic method for detecting resistance to ampicillin. However, there have been reports of β-lactamase-negative ampicillin-resistant (BLNAR) H. influenzae isolates, which exhibit ampicillin resistance through structural changes in Penicillin-Binding Protein 3 (PBP3) due to mutations in the ftsl gene [5]. Reliance on the nitrocefin test to predict susceptibility to ampicillin would miss BLNAR isolates. Diagnosing these BLNAR strains involves susceptibility testing to ampicillin and confirming the absence of β-lactamase production [3]. Molecular methods to detect the ftsl gene may also help diagnose the BLNAR strains, but are not widely available [6, 7]. These molecular methods could predict antimicrobial resistance but not susceptibility. The prevalence of β-lactamase production and antimicrobial susceptibility patterns among H. influenzae strains helps in assessing the effectiveness of treatment regimens and guiding empirical therapy choices. The absence of a global, large-scale surveillance study of the antimicrobial susceptibility of H. influenzae isolates, coupled with variations in antimicrobial susceptibility interpretation arising from breakpoint-dependent classification over the past decade, has hindered comprehensive assessments of resistance trends. These gaps underscore the significance of our study, which aims to provide robust and longitudinally consistent data critical for informing about global antimicrobial resistance surveillance and improving antibiotic prescribing for H. influenzae infections. We present herein the global epidemiology of antimicrobial susceptibility patterns of H. influenzae isolates in Asia and the West Pacific, Europe, Latin America, and North America, with emphasis on ampicillin and amoxicillin-clavulanate susceptibility based on β-lactamase production using the SENTRY Antimicrobial Surveillance Program.
Methods
Using the SENTRY Antimicrobial Surveillance Program, a globally recognized initiative for monitoring antimicrobial resistance, we included a total of 13,869 H. influenzae isolates from patients who received medical care in over 150 medical centers in the United States, Europe, Latin America, Asia, and the West Pacific between 2013 and 2022. In the SENTRY program, bacterial isolates are collected consecutively—one for each episode of infection—based on the type of infection. The analysis included only one isolate per patient. These isolates are then sent to a central monitoring laboratory, JMI Laboratories in North Liberty, IA, USA, for antimicrobial susceptibility testing using reference broth microdilution methods and nitrocefin test to detect β-lactamase production. Strict protocols that adhere to international biosafety regulations are followed to ensure the safe transport of clinical isolates from medical centers worldwide to JMI Laboratories. Isolates are placed in leak-proof, temperature-controlled containers to maintain their viability during transit. Isolates are transported via priority courier services to minimize transit time and maintain sample integrity. Each sample is carefully labeled and tracked to ensure proper handling and prevent contamination. Fifty-one countries were represented, including Argentina, Australia, Austria, Belarus, Belgium, Brazil, Canada, Chile, China, Colombia, Costa Rica, Croatia, Czech Republic, Denmark, Ecuador, Finland, France, Germany, Greece, Hong Kong, Hungary, Ireland, Israel, Italy, Japan, Korea, Malaysia, Mexico, Netherlands, New Zealand, Norway, Panama, Philippines, Poland, Portugal, Romania, Russia, Singapore, Slovakia, Slovenia, Spain, Sweden, Switzerland, Taiwan, Thailand, Turkey, United Kingdom, Ukraine, USA, Venezuela, and Vietnam.
The minimum inhibitory concentration (MIC) breakpoints were analyzed according to the 2024 Clinical and Laboratory Standards Institute (CLSI) standard [8]. Ceftriaxone, cefepime, and piperacillin-tazobactam MIC breakpoints were also evaluated according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines [9]. H. influenzae ATCC® 49,247 was included as a quality control strain during antimicrobial susceptibility testing to ensure the accuracy and reliability of the results [8, 9]. Susceptibility testing for ampicillin, amoxicillin-clavulanic acid, piperacillin-tazobactam, ceftazidime, ceftriaxone, cefepime, meropenem, imipenem-cilastatin, clarithromycin, azithromycin, ciprofloxacin, levofloxacin, moxifloxacin, and tetracycline was recorded and analyzed. The EUCAST recommends lower MIC breakpoints compared to the CLSI for piperacillin-tazobactam (0.25/4 vs. 1/4 µg/ml), ceftriaxone (0.125 vs. 2 µg/ml), cefepime (0.25 vs. 2 µg/ml), imipenem-cilastatin (2 vs. 4 µg/ml), meropenem [0.25 µg/ml (meningitis indications) and 2 µg/ml (non-meningitis indications) vs. 0.5 µg/ml], ciprofloxacin [0.03 µg/ml (meningitis indications) and 0.06 µg/ml (indications other than meningitis) vs. 1 µg/ml], levofloxacin (0.06 vs. 2 µg/ml), moxifloxacin (0.125 vs. 1 µg/ml) [8, 9]. The EUCAST has no established breakpoints for ceftazidime, clarithromycin, or azithromycin for H. influenzae [9].
H. influenzae isolates were categorized as β-lactamase-negative, ampicillin-resistant (BLNAR) if they exhibited negative results in the nitrocefin test for β-lactamase activity and demonstrated resistance to ampicillin based on the broth microdilution MIC testing method. Conversely, isolates were classified as β-lactamase-positive, amoxicillin-clavulanate-resistant (BLPACR) if they tested positive for β-lactamase production and exhibited resistance to amoxicillin-clavulanate using the broth microdilution method. These categorizations facilitate accurate phenotypic differentiation, which is essential for understanding antibiotic resistance mechanisms and informing treatment strategies. In contrast to the CLSI, the EUCAST breakpoints do not incorporate the ‘‘intermediate’’ or ‘‘susceptible, increased exposure’’ category for ampicillin and amoxicillin-clavulanate [9]. Consequently, the classifications of β-lactamase-negative ampicillin intermediate (BLNAI) and β-lactamase-positive amoxicillin-clavulanate intermediate (BLPACI) are defined exclusively according to the CLSI criteria [8, 9].
Statistical analysis was done using MedCalc (version 23.1.7, Ostend, Belgium) and R Foundation for Statistical Computing (version 4.4.2, Vienna, Austria). The Pearson Chi-square test was used to compare the association between the proportions of two categorical variables, e.g., the association between β-lactamase production and antibiotic resistance. Fisher’s Exact test was used for small sample sizes where the expected frequencies in the contingency table cells are less than 5. We reported the test statistic (χ²) and p-value for Pearson’s Chi-square tests. For Fisher’s Exact test, which doesn’t produce a test statistic, the main reported result was the p-value. The association between two categorical variables is considered statistically significant if the p-value is ≤ 0.05.
Results
A total of 13,869 H. influenzae isolates were analyzed. The majority were from the USA (52.2%), followed by the United Kingdom (4.3%), France (4.2%), Australia (3.7%), and Spain (3.6%). The most vulnerable groups were those younger than 18 and older than 65 years old (p < 0.0001). Male preponderance was significant (58.7% vs. 41.3% were females, χ²=795, p < 0.0001). In Asia, the West Pacific region, and Europe, patients older than 65 years old were the most affected at 39.7% and 34.5%, respectively. The pediatric age group (< 18 years old) was predominant in North and Latin America at 33.1% and 43.4%, respectively (Table 2).
Table 2.
Geographical distribution of the antimicrobial susceptibility of Haemophilus influenzae isolates from 2013 to 2022 in the SENTRY antimicrobial surveillance program
| Asia and the West Pacific (n = 1,384) | Europe (n = 4,949) | North America (n = 6,991) | Latin America (n = 545) | Chi2 test or Fisher exact test, p-value | |
|---|---|---|---|---|---|
| Age (years) | (n = 1,271) | (n = 4,417) | (n = 6,847) | (n = 528) | |
| ≤ 18 | 230 (18.1%) | 896 (20.3%) | 2,269 (33.1%) | 229 (43.4%) | χ²=460.6, p < 0.0001* |
| 19–45 | 203 (16%) | 724 (16.3%) | 1,156 (16.9%) | 84 (15.9%) | |
| 46–65 | 333 (26.2%) | 1,272 (28.8%) | 1,838 (26.9%) | 90 (17%) | |
| >65 | 505 (39.7%) | 1,525 (34.5%) | 1,584 (23.1%) | 125 (23.7%) | |
| Sex | |||||
| Male | 763 (60.5%) | 2,648 (59.7%) | 3,974 (58.1%) | 290 (54.9%) | χ²=7.8, p = 0.04* |
| Female | 498 (39.5%) | 1,789 (40.3%) | 2,869 (41.9%) | 239 (45.1%) | |
| Site of infection | |||||
| Bloodstream | 6 (0.43%) | 121 (2.4%) | 167 (2.39%) | 5 (0.92%) | p < 0.0001* |
| Community-acquired respiratory tract | 1,216 (87.8%) | 4,048 (81.8%) | 5,015 (71.73%) | 528 (96.9%) | |
| Pneumonia in hospitalized patients | 154 (11.12%) | 701 (14.2%) | 1,432 (20.5%) | 11 (2%) | |
| Intraabdominal | 0 (0) | 0 (0) | 3 (0.04%) | 0 (0) | |
| Skin and soft tissue | 3 (0.2%) | 7 (0.1%) | 31 (0.44%) | 0 (0) | |
| Cerebrospinal | 0 (0) | 1 (0.02%) | 3 (0.04%) | 1 (0.18%) | |
| Other sites | 5 (0.4%) | 71 (1.4%) | 345 (4.9%) | 0 (0) | |
| ICU admitted | 180 (14.9%) | 911 (21.1%) | 1,609 (25.9%) | 128 (25.1%) | χ²=83.6, p < 0.0001* |
| Cystic Fibrosis | 0 (0) (n = 4) | 1 (1.2%) (n = 86) | 5 (8.2%) (n = 61) | 0 (0) (n = 0) | p = 0.09 |
| Β-lactamase production | 412 (29.8%) (n = 1,384) | 805 (16.3%) (n = 4,949) | 2,003 (28.7%) (n = 6,991) | 113 (20.7%) (n = 545) | χ²=273.4, p < 0.0001* |
| Ampicillin susceptible | 59.5% (n = 1,384) | 80% (n = 4,949) | 68.1% (n = 6,991) | 75.4% (n = 545) | χ²=318.6, p < 0.0001* |
| BLNAI | 6.4% (n = 971) | 4.3% (n = 4,141) | 0.7% (n = 4,981) | 4.4% (n = 432) | p < 0.0001* |
| BLNAR | 9.4% (n = 971) | 0.5% (n = 4,141) | 4.1% (n = 4,981) | 0.5% (n = 432) | p < 0.0001* |
| Amoxicillin-clavulanate susceptible | 82.1% (n = 1,383) | 96.6% (n = 4,949) | 95.2% (n = 6,990) | 96.5% (n = 545) | χ²=4635.4, p < 0.0001* |
| BLPACI | 14.5% (n = 412) | 3.4% (n = 805) | 6.3% (n = 2,002) | 3.5% (n = 113) | p < 0.0001* |
| BLPACR | 10.9% (n = 412) | 0.4% (n = 805) | 1% (n = 2,002) | 0 (n = 113) | p < 0.0001* |
| Piperacillin-tazobactam susceptible (CLSI breakpoints) | 100% (n = 1,322) | 100% (n = 4,503) | 100% (n = 6,612) | 100% (n = 520) | NA |
| Piperacillin-tazobactam susceptible (EUCAST breakpoints) | 99.4% (n = 1,322) | 99.9% (n = 4,503) | 99.8% (n = 6,612) | 100% (n = 520) | p = 0.01* |
| Meropenem susceptible | 99.8% (n = 1,384) | 100% (n = 4,949) | > 99.9% (n = 6,988) | 100% (n = 545) | p = 0.02* |
| Imipenem susceptible | 99.3% (n = 757) | 99.9% (n = 2,373) | 99.5% (n = 3,753) | 100% (n = 272) | p = 0.01* |
| Cefepime susceptible (CLSI breakpoints) | 99.4% (n = 1,383) | 100% (n = 4,947) | > 99.9% (n = 6,988) | 100% (n = 545) | p < 0.0001* |
| Cefepime susceptible (EUCAST breakpoints) | 79.7% (n = 1,383) | 98.6% (n = 4,947) | 97.7% (n = 6,988) | 99.3% (n = 545) | χ²=1193.3, p < 0.0001* |
| Ceftriaxone susceptible (CLSI breakpoints) | 100% (n = 1,383) | 100% (n = 4,946) | 100% (n = 6,988) | 100% (n = 543) | NA |
| Ceftriaxone susceptible (EUCAST breakpoints) | 93.2% (n = 1,383) | 99.9% (n = 4,946) | 99.6% (n = 6,988) | 100% (n = 543) | p < 0.0001* |
| Ceftazidime susceptible | 99.6% (n = 1,382) | 100% (n = 4,947) | > 99.9% (n = 6,981) | 100% (n = 544) | p = 0.0005* |
| Trimethoprim-sulfamethoxazole susceptible | 59.3% (n = 1,384) | 66.4% (n = 4,947) | 65.3% (n = 6,987) | 67.9% (n = 545) | χ²=26.3, p < 0.0001* |
| Clarithromycin susceptible | 85.3% (n = 1,380) | 90.2% (n = 4,948) | 85.5% (n = 6,982) | 87% (n = 545) | χ²=62.7, p < 0.0001* |
| Azithromycin susceptible | 96.7% (n = 1,382) | 99% (n = 4,947) | 98.6% (n = 6,983) | 95.6% (n = 544) | χ²=63.6, p < 0.0001* |
| Ciprofloxacin susceptible | 97.8% (n = 1,384) | 99% (n = 4,949) | 99.6% (n = 6,990) | 99.8% (n = 545) | p < 0.0001* |
| Levofloxacin susceptible | 98.2% (n = 1,384) | 99.1% (n = 4,948) | 99.7% (n = 6,989) | 99.8% (n = 545) | p < 0.0001* |
| Moxifloxacin susceptible | 97.8% (n = 1,208) | 98.9% (n = 4,424) | 99.7% (n = 6,111) | 99.8% (n = 503) | p < 0.0001* |
| Tetracycline susceptible | 94.9% (n = 1,383) | 99.1% (n = 4,948) | 98.8% (n = 6,983) | 99.8% (n = 544) | p < 0.0001* |
The minimum inhibitory concentrations (MIC) were determined according to CLSI standards, unless stated otherwise; NA, not applicable; * Statistically significant difference (p ≤ 0.05); ICU, intensive care unit; CLSI, Clinical and Laboratory Standards Institute; EUCAST, European Committee on Antimicrobial Susceptibility Testing; BLNAI, β-lactamase negative ampicillin intermediate; BLNAR, β-lactamase negative ampicillin resistant; BLPACI, β-lactamase positive ampicillin amoxicillin-clavulanic intermediate; BLPACR, β-lactamase positive amoxicillin-clavulanic acid resistant; χ², Chi-square test statistic
A total of 13,858 (99.9%) H. influenzae isolates were tested for β-lactamase production, with 10,525 (75.9%) being β-lactamase-negative and 3,333 isolates (24.1%) producing β-lactamase. None of the H. influenzae isolates in Austria, Ecuador, Norway, Romania, and Venezuela produced β-lactamase. Among β-lactamase-positive isolates, 3,046 (91.4%) were amoxicillin-clavulanate susceptible, 217 (6.5%) were BLPACI, and 70 (2.1%) were BLPACR. Compared to non-β-lactamase-producing H. influenzae, β-lactamase-producing isolates were associated with lower susceptibility to ampicillin (0.9% vs. 94.1%), (91.4% vs. 95.4%), trimethoprim-sulfamethoxazole (61.2% vs. 66.5%), clarithromycin (85.7% vs. 87.8%), azithromycin (97.1% vs. 98.9%), and tetracycline (95% vs. 99.6%) but higher susceptibility to fluoroquinolones (ciprofloxacin 99.6% vs. 94.1%), The differences in susceptibility rates across antibiotics were statistically significant (p < 0.05) (Table 1).
Table 1.
Characteristics of the worldwide Haemophilus influenzae isolates and differences between β-lactamase positive and β-lactamase negative isolates in the SENTRY antimicrobial surveillance program from 2013 to 2022
| All H. influenzae isolates, n = 13,869, (%) | β-lactamase positive isolates, n = 3,333, (%) | β-lactamase negative isolates, n = 10,525, (%) | Chi2 test or Fisher’s Exact test comparing β-lactamase positive and negative isolates, p-value | |
|---|---|---|---|---|
| Region | ||||
| Asia and West Pacific | 1,384 (10%) | 412 (12.4%) | 971 (9.2%) | χ²=273.83, p < 0.00001* |
| Europe | 4,949 (35.7%) | 805 (24.2%) | 4,141 (39.3%) | |
| North America | 6,991 (50.4%) | 2,003 (60.1%) | 4,981 (47.3%) | |
| Latin America | 545 (3.9%) | 113 (3.4%) | 432 (4.1%) | |
| Age (years) | (n = 13,063) | (n = 3,136) | (n = 9,916) | |
| ≤ 18 | 3,624 (27.8%) | 935 (29.8%) | 2,688 (27.1%) | χ²=23.7, p < 0.0001* |
| 19–45 | 2,167 (16.6%) | 464 (14.8%) | 1,702 (17.2%) | |
| 46–65 | 3,533 (27.1%) | 788 (25.1%) | 2,739 (27.6%) | |
| >65 | 3,739 (28.6%) | 949 (30.3%) | 2,787 (28.1%) | |
| Sex | (n = 13,070) | (n = 3,152) | (n = 9,907) | |
| Male | 7,675 (58.7%) | 1,809 (57.4%) | 5,861 (59.2%) | χ²=3, p = 0.08 |
| Female | 5,395 (41.3%) | 1,343 (42.6%) | 4,046 (40.8%) | |
| Site of infection | ||||
| Bloodstream | 299 (2.2%) | 58 (1.7%) | 241 (2.3%) | p < 0.0001* |
| Community-acquired respiratory tract | 10,807 (77.8%) | 2,582 (77.5%) | 8,217 (78.1%) | |
| Pneumonia in hospitalized patients | 2,298 (16.6%) | 576 (17.3%) | 1,719 (16.3%) | |
| Intraabdominal | 1 (0.01%) | 0 (0) | 1 (0.01%) | |
| Skin and soft tissue | 41 (0.3%) | 11 (0.3%) | 30 (0.29%) | |
| Cerebrospinal | 5 (0.04%) | 2 (0.06%) | 3 (0.03%) | |
| Other sites | 418 (3.01%) | 104 (3.1%) | 314 (2.98%) | |
| ICU admitted | 2,831 (23.1%) (n = 12,244) | 691 (23.5%) (n = 2,947) | 2137 (23%) (n = 9,286) | χ²=0.3, p = 0.6 |
| Cystic fibrosis | 6 (4%) (n = 151) | 1 (2.9%) (n = 34) | 5 (4.4%) (n = 115) | p = 1 |
| Ampicillin susceptible | 71.7% (n = 13,869) | 0.9% (n = 3,333) | 94.1% (n = 10,525) | χ²=10,828, p < 0.0001* |
| Amoxicillin-clavulanate susceptible | 94.4% (n = 13,867) | 91.4% (n = 3,332) | 95.4% (n = 10,524) | χ²=76.8, p < 0.0001* |
| Piperacillin-tazobactam susceptible (CLSI breakpoints) | 100% (n = 12,957) | 100% (n = 3,128) | 100% (n = 9,818) | NA |
| Piperacillin-tazobactam susceptible (EUCAST breakpoints) | 99.8% (n = 12,957) | 99.7% (n = 3,128) | 99.8% (n = 9,818) | χ²=0.4, p = 0.5 |
| Meropenem susceptible | > 99.9% (n = 13,866) | > 99.9% (n = 3,331) | > 99.9% (n = 10,524) | p = 1 |
| Imipenem susceptible | 99.6% (n = 7,155) | 99.6% (n = 1,685) | 99.6% (n = 5,469) | p = 1 |
| Cefepime susceptible (CLSI breakpoints) | 99.9% (n = 13,864) | 99.8% (n = 3,332) | > 99.9% (n = 10,521) | χ²=1.4, p = 0.2 |
| Cefepime susceptible (EUCAST breakpoints) | 96.3% (n = 13,864) | 94.5% (n = 3,332) | 96.9% (n = 10,521) | χ²=40.3, p < 0.0001* |
| Ceftriaxone susceptible (CLSI breakpoints) | 100% (n = 13,860) | 100% (n = 3,331) | 100% (n = 10,518) | NA |
| Ceftriaxone susceptible (EUCAST breakpoints) | 99.1% (n = 13,860) | 98.6% (n = 3,331) | 99.2% (n = 10,518) | χ²=9.5, p = 0.002* |
| Ceftazidime susceptible | > 99.9% (n = 13,854) | 99.9% (n = 3,329) | > 99.9% (n = 10,514) | p = 1 |
| Trimethoprim-sulfamethoxazole susceptible | 65.2% (n = 13,863) | 61.2% (n = 3,332) | 66.5% (n = 10,520) | χ²=31.2, p < 0.0001* |
| Clarithromycin susceptible | 87.2% (13,855) | 85.7% (n = 3,328) | 87.8% (n = 10517) | χ²=9.9, p = 0.002* |
| Azithromycin susceptible | 98.4% (n = 13,856) | 97.1% (n = 3,331) | 98.9% (n = 10515) | χ²=53.5, p < 0.0001* |
| Ciprofloxacin susceptible | 99.2% (n = 13,868) | 99.6% (n = 3,333) | 94.1% (n = 10,524) | χ²=174.8, p < 0.0001* |
| Levofloxacin susceptible | 99.3% (n = 13,866) | 99.8% (n = 3,333) | 94.2% (n = 10,522) | χ²=184.4, p < 0.0001* |
| Moxifloxacin susceptible | 98.5% (n = 13,858) | 99.6% (n = 2,963) | 94.1% (n = 9,272) | χ²=154.2, p < 0.0001* |
| Tetracycline susceptible | 98.5% (n = 13,858) | 95% (n = 3,331) | 99.6% (n = 10,515) | χ²=359.2, p < 0.0001* |
The minimum inhibitory concentrations (MIC) were determined according to CLSI standard unless stated otherwise; CLSI, Clinical and Laboratory Standards Institute; EUCAST, European Committee on Antimicrobial Susceptibility Testing; ICU, intensive care unit; *statistically significant (p ≤ 0.05), χ², Chi-square test statistic
Ninety-one (0.7%) isolates were levofloxacin-resistant; 20 (0.14%) were in the U.S., 45 (0.32%) in Europe (17 in Spain, 5 in Italy, 4 in Germany and Turkey), 25 (0.18%) in Asia-West Pacific (11 in Taiwan, 7 in Korea), and one (0.007%) in Argentina indicating a broad geographic distribution. Of these 91 levofloxacin-resistant isolates, 7 (7.7%) produced a β-lactamase and were located in the USA (2 [2.2%] isolates), Japan (2 [2.2%] isolates), and one (1.1%) isolate each in Korea, Belgium, and Turkey. The susceptibility profiles of these isolates to additional antibiotics were as follows: ampicillin (83.5%), ceftriaxone (100%), piperacillin-tazobactam (100%), carbapenems (100%), trimethoprim-sulfamethoxazole (11%), clarithromycin (58.2%), azithromycin (9.7%), and tetracycline (97.8%).
The lowest susceptibility to ampicillin and amoxicillin-clavulanate was in Asia and the West Pacific region (59.5% and 82.1%, respectively), and the highest was in Europe (80% and 96.6%, respectively). In comparison with the other continents, Asia and the West Pacific also accounted for the lowest susceptibility to different antibiotics, including ceftriaxone (93.2% per the EUCAST breakpoints), cefepime (79.7% per the EUCAST guidelines and 99.4% per the CLSI guidelines), ceftazidime (99.6%), piperacillin-tazobactam (99.4% per the EUCAST breakpoints), meropenem (99.8%), trimethoprim-sulfamethoxazole (59.3%), clarithromycin (85.3%), ciprofloxacin (97.8%), levofloxacin (98.2%), moxifloxacin (97.8%) and tetracycline (94.9%) (Table 2).
BLNAR and BLPACR isolates were associated with fewer ICU admissions compared to β-lactamase-negative ampicillin-susceptible and β-lactamase-positive amoxicillin-clavulanate-susceptible isolates, respectively (15.1% vs. 24.3% and 20.3% vs. 23.2%, p < 0.05, respectively). Unlike the β-lactamase producing isolates, which were associated with higher susceptibility to fluoroquinolones than the β-lactamase negative isolates, the BLNAR and BLPACR were associated with lower susceptibility to most antibiotics, including ciprofloxacin (98.2% vs. 99.2% and 98.6% vs. 99.7%), levofloxacin (98.4% vs. 99.3% and 99% vs. 99.9%), in comparison with β-lactamase-negative ampicillin-susceptible and β-lactamase-positive amoxicillin-clavulanate-susceptible isolates, respectively (Tables 3 and 4).
Table 3.
Comparison of β-lactamase negative ampicillin susceptible and ampicillin resistant Haemophilus influenzae isolates from 2013 to 2022 in the SENTRY antimicrobial surveillance program
| β-lactamase negative ampicillin susceptible isolates (n = 9,912) | β-lactamase negative ampicillin resistant isolates (n = 613) | Chi2 test or Fisher’s Exact test, p-value | |
|---|---|---|---|
| Region | |||
| Asia and the West Pacific | 818 (8.3%) | 153 (25%) | p < 0.00001* |
| Europe | 3,942 (39.8%) | 199 (32.5%) | |
| North America | 4,741 (47.8%) | 240 (39.2%) | |
| Latin America | 411 (4.2%) | 21 (3.4%) | |
| Age (years) | |||
| ≤ 18 | 2,553 (27.3%) | 135 (23.9%) | χ²=9.46, p = 0.002* |
| 19–45 | 1,615 (17.3%) | 87 (15.4%) | |
| 46–65 | 2,591 (27.7%) | 148 (26.2%) | |
| >65 | 2,592 (27.7%) | 195 (24.5%) | |
| Sex | |||
| Male | 5,547 (59.4%) | 314 (55.6%) | χ²=0.00001, p = 0.99 |
| Female | 3,795 (40.6%) | 251 (44.4%) | |
| Site of infection | |||
| Bloodstream | 230 (2.32%) | 11 (1.79%) | p = 0.0001* |
| Community-acquired respiratory tract | 7,710 (77.77%) | 507 (82.7%) | |
| Pneumonia in hospitalized patients | 1,628 (16.42%) | 91 (14.84%) | |
| Intraabdominal | 1 (0.01%) | 0 (0) | |
| Skin and soft tissue | 29 (0.29%) | 1 (0.16%) | |
| Cerebrospinal | 3 (0.03%) | 0 (0) | |
| Other sites | 311 (3.14%) | 3 (0.49%) | |
| ICU admitted | 2,030 (23.2%) (n = 8,758) | 107 (20.3%) (n = 528) | χ²=2.38, p = 0.12 |
| Cystic Fibrosis | 4 (3.6%) (n = 112) | 1 (33.3%) (n = 3) | p = 0.1 |
| Piperacillin-tazobactam susceptible (CLSI breakpoints) | 100% (n = 9,224) | 100% (n = 594) | NA |
| Piperacillin-tazobactam susceptible (EUCAST breakpoints) | 100% (n = 9,224) | 99.5% (n = 594) | p = 0.0002* |
| Meropenem susceptible | 100% (n = 9,911) | 99.3% (n = 613) | p < 0.0001* |
| Imipenem susceptible | 99.7% (n = 5,204) | 97.7% (n = 265) | χ²=10, p < 0.0001* |
| Cefepime susceptible | 100% (n = 9,908) | 99.5% (n = 613) | p = 0.0002* |
| Ceftriaxone susceptible (CLSI breakpoints) | 100% (n = 9,906) | 100% (n = 612) | NA |
| Ceftriaxone susceptible (EUCAST breakpoints) | 99.9% (n = 9,906) | 88.1% (n = 612) | χ²=127.4, p < 0.0001* |
| Ceftazidime susceptible | 100% (n = 9,901) | 99.5% (n = 613) | χ²=49.52, p < 0.0001* |
| Trimethoprim-sulfamethoxazole susceptible | 68.2% (n = 9,908) | 39.9% (n = 612) | χ²=206.7, p < 0.0001* |
| Clarithromycin susceptible | 88.1% (n = 9,904) | 82.7% (n = 613) | χ²=1.5, p = 0.0002* |
| Azithromycin susceptible | 98.9% (n = 9,902) | 98.7% (n = 613) | χ²=1.5, p = 0.3 |
| Ciprofloxacin susceptible | 99.2% (n = 9,912) | 98.2% (n = 612) | χ²=2, p = 0.04* |
| Levofloxacin susceptible | 99.3% (n = 9,910) | 98.4% (n = 612) | χ²=3, p = 0.004* |
| Moxifloxacin susceptible | 99.1% (n = 8,694) | 98.4% (n = 578) | χ²=2, p = 0.06 |
| Tetracycline susceptible | 99.6% (n = 9,903) | 99.2% (n = 613) | χ²=3.6, p = 0.01* |
The minimum inhibitory concentrations (MIC) were determined according to CLSI standards, unless stated otherwise; CLSI, Clinical and Laboratory Standards Institute; EUCAST, European Committee on Antimicrobial Susceptibility Testing; ICU, intensive care unit; * statistically significant (p ≤ 0.05)
Table 4.
Comparison of β-lactamase positive amoxicillin-clavulanate susceptible and amoxicillin-clavulanate resistant Haemophilus influenzae isolates from 2013 to 2022 in the SENTRY antimicrobial surveillance program
| β-lactamase positive amoxicillin-clavulanate susceptible isolates (n = 3,045) | β-lactamase positive amoxicillin-clavulanate resistant isolates (n = 287) | Chi2 test or Fisher’s exact test, p-value | |
|---|---|---|---|
| Region | |||
| Asia and the West Pacific | 307 (10.1%) | 105 (36.5%) | p < 0.0001* |
| Europe | 775 (25.5%) | 30 (10.5%) | |
| North America | 1,854 (60.9%) | 148 (51.6%) | |
| Latin America | 109 (3.5%) | 4 (1.4%) | |
| Age (years) | (n = 2,864) | (n = 271) | |
| ≤ 18 | 853 (29.8%) | 82 (30.3%) | p < 0.0001* |
| 19–45 | 426 (14.9%) | 38 (14%) | |
| 46–65 | 729 (25.4%) | 58 (21.4%) | |
| >65 | 856 (29.9%) | 93 (34.3%) | |
| Sex | (n = 2,882) | (n = 269) | |
| Male | 1,658 (57.5%) | 150 (55.8%) | χ²=1.1, p = 0.6 |
| Female | 1,224 (42.5%) | 119 (44.2%) | |
| Site of infection | |||
| Bloodstream | 54 (1.8%) | 4 (1.4%) | p < 0.0001* |
| Community-acquired respiratory tract | 2,340 (76.8%) | 241 (84%) | |
| Pneumonia in hospitalized patients | 541 (17.8%) | 35 (12.2%) | |
| Intraabdominal | 0 (0) | 0 (0) | |
| Skin and soft tissue | 10 (0.3%) | 1 (0.4%) | |
| Cerebrospinal | 98 (3.2%) | 0 (0) | |
| Other sites | 2 (0.1%) | 6 (2.1%) | |
| ICU admitted | 652 (24.3%) (n = 2,687) | 39 (15.1%) (n = 259) | χ²=1.8, p = 0.001* |
| Cystic Fibrosis | 1 (3.2%) (n = 31) | 0 (0%) (n = 3) | p = 1 |
| Piperacillin-tazobactam susceptible (using CLSI breakpoints) | 100% (n = 2,859) | 100% (n = 268) | NA |
| Piperacillin-tazobactam susceptible (using EUCAST breakpoints) | 99.9% (n = 2,859) | 98.4% (n = 268) | χ²=54.1, p = 0.00002* |
| Meropenem susceptible | 100% (n = 3,043) | 99.7% (n = 287) | p = 0.09 |
| Imipenem susceptible | 99.9% (n = 1,581) | 96.1% (n = 103) | χ²=63.2, p < 0.0001* |
| Cefepime susceptible | > 99.9% (n = 3,044) | 98.3% (n = 287) | χ²=17.6, p = 0.0002* |
| Ceftriaxone susceptible (using CLSI breakpoints) | 100% (n = 3,043) | 100% (n = 287) | NA |
| Ceftriaxone susceptible (using EUCAST breakpoints) | 99.9% (n = 9,224) | 85.4% (n = 594) | p < 0.0001* |
| Ceftazidime susceptible | 100% (n = 3,041) | 99% (n = 287) | p < 0.0001* |
| Trimethoprim-sulfamethoxazole susceptible | 63.2% (n = 3,044) | 39% (n = 287) | χ²=2.7, p < 0.0001* |
| Clarithromycin susceptible | 86.3% (n = 3,040) | 78.4% (n = 287) | χ²=1.7, p = 0.0003* |
| Azithromycin susceptible | 97.7% (n = 3,043) | 91.3% (n = 287) | χ²=3.9, p < 0.0001* |
| Ciprofloxacin susceptible | 99.7% (n = 3,045) | 98.6% (n = 287) | χ²=6.4, p = 0.03* |
| Levofloxacin susceptible | 99.9% (n = 3,045) | 99% (n = 287) | p = 0.01* |
| Moxifloxacin susceptible | 99.7% (n = 2,684) | 98.6% (n = 278) | χ²=4.9, p = 0.04* |
| Tetracycline susceptible | 95.1% (n = 3,043) | 94.1% (n = 287) | χ²=1.2, p = 0.5 |
The minimum inhibitory concentrations (MIC) were determined according to CLSI standard unless stated otherwise; NA, not applicable; CLSI, Clinical and Laboratory Standards Institute; EUCAST, European Committee on Antimicrobial Susceptibility Testing; ICU, intensive care unit; * statistically significant (p ≤ 0.05); χ²=Chi-square test statistic
Asia and the West Pacific harbored the highest resistance to β-lactams, including BLNAR and BLPACR (Table 2). Nonetheless, this region was the least associated with intensive care unit (ICU) admissions (14.9% vs. 21.2% in Europe, 25.9% in North America, and 25.1% in Latin America), bloodstream and central nervous system infections (0.4% bloodstream infections compared to 2.4% in Europe and North America, and no central nervous system infections vs. 0.01% in Europe and 0.04% in North America) (p < 0.05) (Table 2). Following the CLSI breakpoints, ceftriaxone and piperacillin-tazobactam demonstrated a 100% susceptibility rate, but, according to the EUCAST breakpoints, the susceptibility rates were 99.1% and 99.8%, respectively (Table 2). BLNAR and BLPACR H. influenzae isolates revealed higher resistance rates to ceftazidime, cefepime, carbapenem, clarithromycin, trimethoprim-sulfamethoxazole, tetracycline, and fluoroquinolones (Tables 3 and 4).
Vietnam (87.5%), Japan (70.8%), Korea (65.7%), and Taiwan (65.9%) exhibited the highest percentages of BLNAR (Fig. 1). BLNAI H. influenzae isolates were predominant in Ecuador (42.9%), Panama (14.8%), and Thailand (14.3%) (Fig. 1). Likewise, most BLPACI H. influenzae isolates were located in Vietnam (50%), Korea (30.4%), and Japan (28.6%). Most BLPACR isolates were present in Korea (39.1%), Japan (35.7%), and Taiwan (10.1%) (Fig. 2).
Fig. 1.
Stacked Column chart representing the percentages of β-lactamase negative ampicillin intermediate and ampicillin-resistant H. influenzae isolates in each of the 51 analyzed countries
Fig. 2.
Stacked Column chart representing the percentages of β-lactamase positive (BLP) amoxicillin-clavulanate intermediate and amoxicillin-clavulanate resistant H. influenzae isolates by each analyzed country. Only 37 countries with BLP H. influenzae isolates were included in this figure
The longitudinal difference in the in vitro susceptibility of H. influenzae isolates to multiple antibiotics decreased over the years (p < 0.0001) except for susceptibility to trimethoprim-sulfamethoxazole and levofloxacin among β-lactamase positive isolates (Fig. 3).
Fig. 3.
Trends of β-lactamase negative (BLN) and β-lactamase positive (BLP) H. influenzae isolates susceptibility to antibiotics over the years, 2013–2022. Dashed lines represent trends of susceptibility of BLN H. influenzae isolates to antibiotics, and non-dashed lines correspond to BLP isolates’ susceptibility to antibiotics. The χ² values for the longitudinal difference of susceptibility rates of BLN H. influenzae isolates to amoxicillin-clavulanic acid, ampicillin, clarithromycin, trimethoprim-sulfamethoxazole, and levofloxacin between 2013 and 2022 were 188.2, 360.4, 44.7, 256.5, and 41, respectively, with p-values of < 0.0001. For BLP H. influenzae isolates, the χ² values for the longitudinal difference of susceptibility rates to amoxicillin-clavulanic acid, ampicillin, and clarithromycin between 2013 and 2022 were 86.2, 53.5, and 80.9, respectively, with p-values < 0.0001. For BLP isolates, the longitudinal susceptibility to trimethoprim-sulfamethoxazole and levofloxacin revealed a χ² of 15.1 and 12.3, and a p-value of 0.09 and 0.2, respectively
H. influenzae isolates were more susceptible to azithromycin than clarithromycin (98.4% vs. 87.2%, respectively, p < 0.0001) (Table 1). Among the 216 azithromycin-resistant H. influenzae isolates, only 0.5% were susceptible to clarithromycin. Conversely, among the 1,768 clarithromycin-resistant isolates, 87.8% remained susceptible to azithromycin.
Of the 13,105 respiratory H. influenzae isolates, 71.5% were susceptible to ampicillin, 94.3% to amoxicillin-clavulanate, 100% to ceftriaxone, 64.9% to trimethoprim-sulfamethoxazole, 86.9% to clarithromycin, 98.4% to azithromycin, 99.3% to levofloxacin, and 98.5% to tetracycline. Five hundred ninety-eight (4.6%) of the respiratory isolates were BLNAR, among which 285 (47.7%) tested susceptible to amoxicillin-clavulanate, 598 (100%) to ceftriaxone following CLSI breakpoints (527 [88.1%] following EUCAST breakpoints), 492 (82.3%) were susceptible to clarithromycin, 590 (98.7%) to azithromycin, 587 (98.2%) to ciprofloxacin, 588 (98.3%) to levofloxacin, 589 (98.4%) to moxifloxacin, and 593 (99.2%) to tetracycline. Of the 276 (2.1%) respiratory isolates that were BLPACR, 276 (100%) were susceptible to ceftriaxone according to CLSI breakpoints, but 234 (84.8%) according to EUCAST breakpoints, 214 (77.5%) were susceptible to clarithromycin, 251 (90.9%) to azithromycin, 273 (98.9%) to levofloxacin, and 259 (93.8%) to tetracycline.
Discussion
Continuous monitoring of H. influenzae resistance trends is critical for optimizing treatment protocols and supporting antimicrobial stewardship efforts. Our study offers several notable advantages. It is a 10-year longitudinal surveillance study for H. influenzae antimicrobial resistance and β-lactamase production in Asia, the West Pacific, Europe, Latin and North America. To our knowledge, this is the first global study to analyze the BLNAR, BLNAI, BLPACP, and BLPACI using the broth microdilution method, the reference antimicrobial susceptibility testing method recommended by the CLSI and EUCAST [8, 9]. Another significant advantage of our study is the reliance on the 2024 CLSI and EUCAST breakpoints. Analyzing the bacterial isolates using the same updated breakpoints across the study period enhances the reliability and comparability of the results. This consistency enhances accurate conclusions about trends in antimicrobial resistance over time. Note that the ampicillin and amoxicillin-clavulanic acid breakpoints for H. influenzae were updated twice during our study period. The CLSI and EUCAST revised the ampicillin breakpoint in 2014, setting the susceptible cutoff at ≤ 2 µg/ml [8, 9]. For amoxicillin/clavulanate, the 2014 susceptible breakpoint was set at ≤ 4/2 µg/ml [8]. In 2020, the ampicillin and amoxicillin-clavulanate susceptible breakpoints for H. influenzae were lowered to ≤ 1 and ≤ 2 µg/mL, respectively [8, 9]. Furthermore, all H. influenzae isolates were sent to a central laboratory for standardized testing using reference testing methods, ensuring consistency and reliability in susceptibility testing. Unlike retrospective reviews, which rely on previously tested isolates following discrepant breakpoints, our study ensures a reliable antimicrobial susceptibility classification. Outdated breakpoints may result in antimicrobial categorical disagreement, misclassifying an organism as susceptible when resistant. These biased misclassifications can result in ineffective therapy and adverse patient outcomes. Moreover, our study’s data from diverse geographical regions in the five continents ensures representative findings of the global epidemiological variations of H. influenzae infections.
While susceptibility to ceftriaxone and piperacillin-tazobactam was 100% based on CLSI breakpoints, it was lower at 99.1% and 99.8%, respectively, when evaluated using the EUCAST breakpoints (Table 1). For these two antibiotics, the EUCAST breakpoints are more conservative (≤ 0.125 and ≤ 0.25 µg/mL, respectively) than CLSI (≤ 2 µg/mL and ≤ 1 µg/mL, respectively), leading to more isolates being interpreted as susceptible. Similarly, cefepime breakpoints interpretation according to the EUCAST breakpoints exhibited lower susceptibility rates compared to the CLSI breakpoints [8, 9].
A recently published meta-analysis study evaluated the global prevalence of multidrug-resistant (MDR) H. influenzae, including 16 clinical studies in Bangladesh, China, Japan, Iran, Taiwan, Portugal, Spain, and Ethiopia from 2003 to 2023 [10]. None of the studies were from North or Latin America. A total of 19,787 H. influenzae clinical isolates were analyzed, indicating a global prevalence of H. influenzae of 34.9%, while MDR H. influenzae, defined as resistance to at least three antibiotic classes, represented 23.1%. Antibiotic resistance rates to ampicillin, azithromycin, and ceftriaxone were 36%, 15.3%, and 1.4%, respectively. The study, prone to antimicrobial susceptibility disagreement, did not provide specific rates for BLNAR, BLPACR, BLNAI, and BLPACI H. influenzae isolates [10].
β-lactamase-producing H. influenzae isolates were associated with higher susceptibility to fluoroquinolones but lower susceptibility to trimethoprim-sulfamethoxazole, tetracycline, and macrolides than the non-β-lactamase-producing isolates (Table 1). The potential for β-lactamase to alter bacterial cell wall and membrane permeability could impact the effectiveness of fluoroquinolones by facilitating antibiotic penetration into the bacterial cell. Thus, fluoroquinolones could provide an alternative treatment option for infections caused by β-lactamase-producing H. influenzae. On the other hand, BLPACR and BLNAR isolates exhibited reduced susceptibility to fluoroquinolones compared to β–lactamase positive, amoxicillin-clavulanate–susceptible and β–lactamase negative, ampicillin-susceptible isolates, respectively (Tables 3 and 4). These findings suggest that alterations in PBP3 may contribute more significantly to fluoroquinolone resistance in H. influenzae than β-lactamase production.
Our study revealed that the global rates of BLNAR varied over the years, with 0.9% resistance to ampicillin in 2013, 14.3% in 2016, and 87% in 2021 (χ²=360.4, p < 0.0001, Fig. 3). That is consistent with the findings by Tsang et al. who presented findings on the antibiotic resistance patterns of H. influenzae in Canada from 2007 to 2014 and showed the rates of BLNAR varied, ranging from 8.8% in 2007 to 19.2% in 2014 with an average of 14.6% (129/882), without showing a consistent annual increase [7]. In Japan, the prevalence of BLNAR strains has notably increased, from 5.8% in 2000 to 34.5% in 2004, and from 23.2% in 2013 to 60% in 2016 [3]. In the United Kingdom, H. influenzae isolates from blood and cerebrospinal fluid cultures collected between January 1985 and December 2004 revealed that BLNAR isolates first emerged in 1991, reached a peak of 8.5% in 1998, and subsequently stayed below 5% [11]. A longitudinal study analyzed 3,052 H. influenzae isolates in multiple European countries in 1997/1998, 2002/2003, and 2004/2005. The percentage of BLNAR strains remained relatively stable, with rates of 8.8% in 1997/98, 9.6% in 2002/2003, and 8.8% in 2004/2005. However, the rates of BLNAR strains varied significantly across different countries, ranging from 0% (France and the Netherlands in 2004/2005) to 33.9% (Spain in 2004/2005) [12].
Antimicrobial susceptibility testing for the BLNAR H. influenzae strains can be challenging due to the variability in resistance expression and the limitations of standard diagnostic methods [13]. Moreover, H. influenzae in vitro susceptibility testing requires specific media or panels not widely available in clinical microbiology laboratories. While broth microdilution is the reference method, availability and logistical challenges in clinical settings can hinder its practical implementation. The dilutions and incubation times can prolong the testing process compared to the automated systems, delaying results. Alternative methods, i.e., disk diffusion, E-test, and automated systems, may have limitations in accuracy and availability [14]. Key clinical microbiology and infectious disease textbooks recommend, especially for non-invasive H. influenzae isolates, only nitrocefin β-lactamase testing [15]. Relying on the nitrocefin β-lactamase testing to predict β-lactam susceptibility would not only miss BLNAR but also BLPACR isolates. That could result in an overestimation of susceptibility to amoxicillin and amoxicillin-clavulanate globally, particularly in Asia and the West Pacific region (Figs. 1 and 2).
Our study displayed high percentages of BLNAR, BLNAI, BLPACP, and BLPACI in Asia and the West Pacific, consistent with the published literature [3, 5, 6]. Following the EUCAST breakpoints, which do not include an intermediate category for ampicillin and amoxicillin-clavulanate, BLNAI and BLPACI H. influenzae isolates would be resistant. It would be wise to refrain from treating BLNAI and BLPACI with ampicillin or amoxicillin-clavulanate, respectively. Although the in vitro susceptibility to piperacillin-tazobactam was preserved, the CLSI recommends considering piperacillin-tazobactam, among other β-lactams, e.g., amoxicillin-clavulanate, ampicillin-sulbactam, cefaclor, cefamandole, cefprozil, and cefuroxime, as inactive against BLNAR [8].
Compared with β-lactamase negative ampicillin-susceptible H. influenzae isolates, BLNAR isolates revealed a reduced carbapenem susceptibility (Table 3). That may likely be due to altered PBP3 containing the amino acid insertion V525_N526insM [16]. Similarly, BLNAR were significantly less susceptible to ciprofloxacin, levofloxacin, trimethoprim-sulfamethoxazole, ceftazidime, cefepime, tetracycline, and clarithromycin (p ≤ 0.05) (Table 3). Likewise, BLPACR H. influenzae isolates revealed a statistically higher resistance to ceftazidime, cefepime, imipenem, ciprofloxacin, levofloxacin, moxifloxacin, trimethoprim-sulfamethoxazole, clarithromycin, and azithromycin when compared to β-lactamase positive amoxicillin-clavulanate susceptible isolates (p ≤ 0.05) (Table 4). Adopting the EUCAST breakpoints also demonstrated higher resistance to ceftriaxone and piperacillin-tazobactam among the BLNAR and BLNACR isolates (Tables 3 and 4). This complies with emerging data on H. influenzae isolates with increased MIC and/or resistance to ceftriaxone due to decreased affinity for β-lactams, primarily associated with S385T and L389F substitutions and altered PBP3 [17]. BLNAR could pose serious challenges to treatment options. This situation necessitates heightened vigilance in tracking these resistant strains and highlights the importance of researching alternative diagnostic and therapeutic approaches. Careful antibiotic stewardship is also essential to mitigate the rise of resistant strains.
According to CLSI, antimicrobial susceptibility testing of respiratory H. influenzae against clarithromycin, azithromycin, amoxicillin-clavulanate, and second- or third-generation cephalosporins is not necessary [8]. These antimicrobial agents can be used as empiric treatment for H. influenzae infections. Applying this CLSI recommendation to our study’s findings would result in very major errors (VMEs) of 5.7%, 13.1%, and 1.6% for amoxicillin-clavulanate, clarithromycin, and azithromycin, respectively, among the 13,105 H. influenzae respiratory isolates. The VME threshold should be less than 1.5% to prevent incorrect categorization of a susceptible isolate as resistant [18]. Thus, one should be aware of this bias while treating empirically with these three antibiotics.
The CLSI recommends azithromycin or clarithromycin as empiric therapy for H. influenzae respiratory infections [8]. Our findings revealed a higher in vitro susceptibility of H. influenzae isolates to azithromycin compared to clarithromycin (98.4% vs. 87.2%), respectively. Moreover, 87.8% of clarithromycin-resistant isolates remained susceptible to azithromycin vs. only 0.5% of azithromycin-resistant isolates were susceptible to clarithromycin. Therefore, azithromycin is a better option for empiric therapy. Azithromycin’s unique structure allows it to bind differently to the bacterial ribosome, potentially avoiding the impact of certain mutations that confer resistance to clarithromycin [19]. Moreover, azithromycin is less susceptible to efflux pump mechanisms than clarithromycin [19]. Assessing the in vitro susceptibility of isolates to macrolides and other antibiotics, such as fluoroquinolones, which have documented resistance, remains essential to guide effective antibiotic therapy.
This study presents the advantages of centralized and standardized testing, high-quality, consistent data, and global longitudinal insights for 10-year data trends. However, molecular testing of the resistance genes, i.e., blaTEM, blaROB, and ftsl, and serotype data were unavailable. Moreover, we lacked data on the treatment used and clinical outcomes. Selection biases may not be neglected. The program relies on participating medical centers to submit samples, which may not fully represent all healthcare settings within a country. In countries with fewer participating centers, the data might not fully capture the national antimicrobial resistance landscape. Data might be more representative of urban areas where participating hospitals are often located, potentially missing trends in rural settings. However, the alignment of our findings with existing data suggests that these biases are not significant. Overall, the SENTRY Antimicrobial Surveillance Program is valuable for tracking trends in global antimicrobial resistance.
In conclusion, the emergence of BLNAR and BLPACR strains presents a global challenge. The elevated prevalence of BLNAR and BLPACR, particularly in Asia and the West Pacific, coupled with increased resistance to other antibiotics, i.e., trimethoprim-sulfamethoxazole, clarithromycin, ciprofloxacin, and levofloxacin, underscores the urgent need for effective containment strategies to mitigate the spread of resistant strains. Empirical therapy with clarithromycin should be used with caution, especially in Asia and the West Pacific region. The discrepancy between the CLSI and EUCAST breakpoints has significant implications for the interpretation of antimicrobial susceptibility. The stricter EUCAST criteria may lead to higher resistance rates, which may impact clinical decision-making, surveillance data, and laboratory testing methods. Understanding these differences is crucial for harmonizing resistance reporting and optimizing antimicrobial therapy.
Acknowledgements
I would like to express our sincere gratitude to the SENTRY Antimicrobial Surveillance Program and Dr. Helio S. Sader at the JMI Laboratories for the invaluable scientific data that made this study possible. I also want to acknowledge the Balamand Medical Library, especially Mrs. Amal Issa and Mariam Alam, for the interlibrary loans of scientific data.
Author contributions
M.M.S. has prepared, analyzed, drafted, and reviewed the manuscript.
Funding
No specific funding was received for this study.
Data availability
The data is available by the SENTRY Antimicrobial Resistance Program available online at: https://www.jmilabs.com.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.CDC. Clinical overview of haemophilus influenzae disease. Centers for Disease Control and Prevention; 2024.
- 2.Gilsdorf JR. Hib vaccines: their impact on haemophilus influenzae type b disease. J Infect Dis. 2021;224:S321–30. 10.1093/infdis/jiaa537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Honda H, Sato T, Shinagawa M, Fukushima Y, Nakajima C, Suzuki Y, Shiraishi T, Kuronuma K, Takahashi S, Takahashi H, Yokota S. Multiclonal expansion and high prevalence of beta-Lactamase-Negative haemophilus influenzae with high-Level ampicillin resistance in Japan and susceptibility to quinolones. Antimicrob Agents Chemother. 2018;62:e00851–18. 10.1128/AAC.00851-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.World Health Organization. WHO updates list of drug-resistant bacteria most threatening to human health. 2024. Accessed online September 2024.
- 5.Zhou Y, Wang Y, Cheng J, Zhao X, Liang Y, Wu J. Molecular epidemiology and antimicrobial resistance of haemophilus influenzae in guiyang, guizhou, China. Front Public Health. 2022;10:947051. 10.3389/fpubh.2022.947051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Li X, Xiao S, Gu F, He W, Ni Y, Han L. Molecular epidemiology and antimicrobial resistance of haemophilus influenzae in adult patients in shanghai, China. Front Public Health. 2020;8:95. 10.3389/fpubh.2020.00095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tsang RSW, Shuel M, Whyte K, Hoang L, Tyrrell G, Horsman G, et al. Antibiotic susceptibility and molecular analysis of invasive haemophilus influenzae in canada, 2007 to 2014. J Antimicrob Chemother. 2017;72:1314–9. 10.1093/jac/dkw565. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.CLSI. M100: Performance Standards for Antimicrobial Susceptibility Testing. Ed34. 2024.
- 9.EUCAST. European Committee on Antimicrobial Susceptibility Testing Breakpoint tables for interpretation of MICs and zone diameters Version 14.0. 2024.
- 10.Abavisani M, Keikha M, Karbalaei M. First global report about the prevalence of multi-drug resistant haemophilus influenzae: a systematic review and meta-analysis. BMC Infect Dis. 2024;24:90–5. 10.1186/s12879-023-08930-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ladhani S, Heath PT, Ramsay ME, Slack MPE. Changes in antibiotic resistance rates of invasive haemophilus influenzae isolates in England and Wales over the last 20 years. J Antimicrob Chemother. 2008;62:776–9. 10.1093/jac/dkn274. [DOI] [PubMed] [Google Scholar]
- 12.Jansen WTM, Verel A, Beitsma M, Verhoef J, Milatovic D. Longitudinal European surveillance study of antibiotic resistance of haemophilus influenzae. J Antimicrob Chemother. 2006;58:873–7. 10.1093/jac/dkl310. [DOI] [PubMed] [Google Scholar]
- 13.Skaare D, Lia A, Hannisdal A, Tveten Y, Matuschek E, Kahlmeter G, Kristiansen B. Haemophilus influenzae with Non-Beta-Lactamase-Mediated Beta-Lactam resistance: easy to find but hard to categorize. J Clin Microbiol. 2015;53:3589–95. 10.1128/JCM.01630-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Giger O, Mortensen JE, Clark RB, Evangelista A. Comparison of five different susceptibility test methods for detecting antimicrobial agent resistance among haemophilus influenzae isolates. Diagn Microbiol Infect Dis. 1996;24:145–53. 10.1016/0732-8893(96)00026-0. [DOI] [PubMed] [Google Scholar]
- 15.Morton AB, Vareechon C, Pettengill MA. Is beta-lactamase testing acceptably accurate for predicting haemophilus influenzae susceptibility to beta-lactams? Epidemiological data from philadelphia, USA, 2017–2023. Microbiol Spectr. 2023;11:129223. 10.1128/spectrum.01292-23japa. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kitaoka K, Kimura K, Kitanaka H, Banno H, Jin W, Wachino J, Arakawa Y. Carbapenem-Nonsusceptible haemophilus influenzae with Penicillin-Binding protein 3 containing an amino acid insertion. Antimicrob Agents Chemother. 2018;62:e00671–18. 10.1128/AAC.00671-18. Print 2018 Aug. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Lee S, Kim G, Kim JH, Kim M, Lee J. Characterization of Ceftriaxone-Resistant haemophilus influenzae among Korean children. J Korean Med Sci. 2024;39:e136. 10.3346/jkms.2024.39.e136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.CLSI M52. Clinical and Laboratory Standards Institute. Verification of commercial microbial identification and antimicrobial susceptibility testing systems. 2015.
- 19.Whitman MS, Tunkel AR. Azithromycin and clarithromycin: overview and comparison with erythromycin. Infect Control Hosp Epidemiol. 1992;13:357–68. 10.1086/646545. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data is available by the SENTRY Antimicrobial Resistance Program available online at: https://www.jmilabs.com.