Results from the Survey of Antibiotic Resistance (SOAR) 2014–16 in Greece

Abstract

Objectives

To determine antimicrobial susceptibility in isolates of Streptococcus pneumoniae and Haemophilus influenzae collected in 2014–16 from patients with community-acquired respiratory tract infections in Greece.

Methods

MICs were determined by CLSI broth microdilution and susceptibility assessed using CLSI, EUCAST and pharmacokinetic/pharmacodynamic (PK/PD) breakpoints.

Results

A total of 99 S. pneumoniae and 52 H. influenzae isolates were collected. Overall, 36.4% of S. pneumoniae were penicillin susceptible by CLSI oral/EUCAST and 88.9% by CLSI intravenous (iv) breakpoints. All were fluoroquinolone susceptible with ≥94% of isolates also susceptible to amoxicillin, amoxicillin/clavulanic acid and ceftriaxone by CLSI and PK/PD breakpoints. Trimethoprim/sulfamethoxazole, cefuroxime, cefaclor and macrolides were less active, with rates of susceptibility of 83.8%, 69.7%, 50.5% and 49.5%, respectively, by CLSI. Generally susceptibility was the same or slightly lower by EUCAST, but the cefaclor difference was much greater. Among H. influenzae, 15.4% of isolates were β-lactamase positive. Susceptibility to amoxicillin/clavulanic acid, ceftriaxone, cefuroxime and the fluoroquinolones was seen in >95% of isolates by CLSI criteria. Susceptibility to azithromycin was seen in 94.2% of isolates using CLSI breakpoints, but clarithromycin susceptibility was lower (61.5%). However, susceptibility to both macrolides was seen in <5% of isolates by PK/PD and EUCAST criteria. Susceptibility to trimethoprim/sulfamethoxazole was seen in 71.2% of isolates.

Conclusions

Owing to the high prevalence of macrolide resistance among S. pneumoniae and the reduced activity of clarithromycin against H. influenzae, it appears that these agents are not appropriate as monotherapy for community-acquired pneumonia in Greece. Amoxicillin/clavulanic acid, on the other hand, maintained excellent in vitro activity and, as opposed to the similarly effective fluoroquinolones, is safe to use in paediatric patients.

Introduction

Acute respiratory tract infections are among the most common reasons for physician office visits and for antibiotic prescriptions.^1,2 Treatment is problematic because respiratory tract infections are caused by a wide variety of pathogens (viral as well as bacterial), and when initiation of antimicrobial treatment is deemed necessary, the bacterial pathogen or its resistance profile is usually unknown. Empirical therapy therefore tends towards broad-spectrum agents, but inappropriate antibiotic use is linked to the development of antimicrobial resistance. Increasing resistance is considered an urgent threat to public health, and drug-resistant pneumococci are among the species listed as serious threats by the Centers for Disease Control and Prevention.^1,3 Surveillance of antimicrobial susceptibility of respiratory pathogens over time and across geographic regions is important: it improves individual patient outcomes by helping the clinician to select appropriate antibiotic agents, which in turn affects public health by reducing antimicrobial resistance. One such surveillance study is the ongoing Survey of Antibiotic Resistance (SOAR), which has been monitoring key respiratory pathogens in the Middle East, Africa, Latin America, Asia-Pacific and Commonwealth of Independent States countries since 2002.⁴ Here we report data from SOAR for major respiratory tract pathogens collected in Greece between 2014 and 2016.

Materials and methods

Clinical isolates

Isolates of H. influenzae and S. pneumoniae from community-acquired respiratory tract infections were collected at the National and Kapodistrian University of Athens from 2014 to 2016 and sent to a central laboratory (International Health Management Associates, Inc., Schaumburg, Switzerland), where they were sub-cultured and re-identified. H. influenzae were re-identified by MALDI-TOF MS methodology and S. pneumoniae identity was confirmed by optochin susceptibility and bile solubility. β-Lactamase production was determined for each H. influenzae isolate by a chromogenic cephalosporin (nitrocefin) disc method. Duplicate isolates from the same patient were excluded from the analysis.

Susceptibility testing

Isolates were evaluated for antibiotic susceptibility using broth microdilution methodology recommended by CLSI.⁵

Both pathogens were assessed for susceptibility to amoxicillin, amoxicillin/clavulanic acid (2:1), azithromycin, cefaclor, ceftriaxone, cefuroxime, clarithromycin, erythromycin, levofloxacin, moxifloxacin and trimethoprim/sulfamethoxazole (1:19). S. pneumoniae was also tested for susceptibility to penicillin whereas H. influenzae was additionally tested for susceptibility to ampicillin.

Susceptibility to the study drugs was calculated based on CLSI breakpoints, EUCAST breakpoints and pharmacokinetic/pharmacodynamic (PK/PD) breakpoints.^6–8 These breakpoints are shown in Table 1.

Table 1.

MIC breakpoints (mg/L) used for S. pneumoniae and H. influenzae isolates

	S. pneumoniae						H. influenzae
	CLSI			EUCAST			CLSI			EUCAST
Antimicrobial	S	I	R	S	I	R	S	I	R	S	I	R	PK/PD (S only)
Amoxicillin	≤2	4	≥8	NA	NA	NA	NA	NA	NA	≤2	–	≥4	≤2 (≤4)
AMCa	≤2	4	≥8	NA	NA	NA	≤4	–	≥8	≤2	–	≥4	≤2 (≤4)
Ampicillin	NA	NA	NA	NA	NA	NA	≤1	2	≥4	≤1	–	≥2	NA
Azithromycin	≤0.5	1	≥2	≤0.25	0.5	≥1	≤4	–	–	≤0.12	0.25–4	≥8	≤0.12
Cefaclor	≤1	2	≥4	≤0.03	0.06–0.5	≥1	≤8	16	≥32	NA	NA	NA	≤0.5
Ceftriaxone	≤1	2	≥4	≤0.5	1–2	≥4	≤2	–	–	≤0.12	–	≥0.25	≤1
Cefuroximeb	≤1	2	≥4	≤0.25	0.5	≥1	≤4	8	≥16	≤0.12	0.25–1	≥2	≤1
Clarithromycin	≤0.25	0.5	≥1	≤0.25	0.5	≥1	≤8	16	≥32	≤1	2–32	≥64	≤0.25
Erythromycin	≤0.25	0.5	≥1	≤0.25	0.5	≥1	NA	NA	NA	NA	NA	NA	≤0.25
Levofloxacin	≤2	4	≥8	≤2	–	≥4	≤2	–	–	≤0.06	–	≥0.12	≤2
Moxifloxacin	≤1	2	≥4	≤0.5	–	≥1	≤1	–	–	≤0.12	–	≥0.25	≤1
Penicillin (iv non-meningitis)	≤2	4	≥8	NAc	NAc	NAc	NA	NA	NA	NA	NA	NA	NA
Penicillin (oral)	≤0.06	0.12–1	≥2	≤0.06	0.12–2	≥4	NA	NA	NA	NA	NA	NA	NA
SXTd	≤0.5	1–2	≥4	≤1	2	≥4	≤0.5	1–2	≥4	≤0.5	1	≥2	≤0.5

AMC, amoxicillin/clavulanic acid; SXT, trimethoprim/sulfamethoxazole; S, susceptible; I, intermediate; R, resistant; NA, not applicable.

a

Amoxicillin/clavulanic acid was tested at a 2:1 amoxicillin to clavulanic acid ratio; breakpoints are expressed as the amoxicillin component. PK/PD breakpoints based on high dose (4 g of amoxicillin with 250 mg of clavulanate per day for adults) are shown in parentheses,⁶ which is the same as CLSI for H. influenzae and one dilution higher for S. pneumoniae.

b

Breakpoints used are for cefuroxime axetil (oral).

c

EUCAST do not give iv breakpoints but dose-specific susceptible breakpoints are noted for pneumonia: 1.2 g × 4 , ≤ 0.5 mg/L; 2.4 g × 4 (or 1.2 g × 6), ≤1 mg/L and 2.4 g × 6,  ≤ 2 mg/L.

d

Trimethoprim/sulfamethoxazole was tested at a 1: 19 trimethoprim to sulfamethoxazole ratio; breakpoints are expressed as the trimethoprim component.

Quality control and data analysis

Quality control strains S. pneumoniae ATCC 49619, Escherichia coli ATCC 25922, H. influenzae ATCC 49247, H. influenzae ATCC 49766 and E. coli ATCC 32518 were included on each day of testing. Results of susceptibility testing were accepted if the results of the control strains were within published limits. Differences in susceptibility between age groups, source of infection and penicillin susceptibility (S. pneumoniae only) were assessed for statistical significance with Fisher’s exact test using XLSTAT version 2011.1.05. A P value <0.05 was considered statistically significant.

Results

S. pneumoniae isolates

A total of 99 S. pneumoniae isolates were collected from one centre in Greece from 2014 to 2016. Most pneumococci came from sputum (n = 74; 74.7%). Less frequently, isolates were from blood (n = 12; 12.1%), middle ear effusion (n = 8; 8.1%) and bronchoalveolar lavage (n = 5; 5.1%). Most isolates (n = 61; 61.6%) came from children (aged ≤12 years), 31 (31.3%) were from adult patients (13–64 years old) and 7 (7.1%) were from elderly patients (aged ≥65 years). Summary MIC and susceptibility data for all 99 S. pneumoniae isolates are shown in Table 2. MIC distribution data are given in Table 3.

Table 2.

MIC and susceptibility data for S. pneumoniae isolates (n = 99)

					Susceptibility using indicated breakpoints
	MIC (mg/L)				CLSI			PK/PD	EUCAST
Antimicrobial	50%	90%	min	max	%S	%I	%R	%S	%S	%I	%R
Amoxicillina	0.12	2	≤0.015	8	94.9	1.0	4.0	94.9 (96.0)	NA	NA	NA
AMCa	0.06	1	≤0.015	8	94.9	3.0	2.0	94.9 (98.0)	NA	NA	NA
Azithromycin	>2	>2	0.06	>2	49.5	0.0	50.5	45.5	49.5	0.0	50.5
Cefaclor	1	>4	0.12	>4	50.5	14.1	35.4	10.1	0.0	10.1	89.9
Ceftriaxone	0.12	1	≤0.015	2	99.0	1.0	0.0	99.0	69.7	30.3	0.0
Cefuroxime (oral)	0.25	4	≤0.12	>4	69.7	2.0	28.3	69.7	51.5	14.1	34.3
Clarithromycin	0.5	>1	≤0.015	>1	49.5	1.0	49.5	49.5	49.5	1.0	49.5
Erythromycin	>0.5	>0.5	0.03	>0.5	49.5	0.0	50.5	49.5	49.5	0.0	50.5
Levofloxacin	1	1	0.5	2	100	0.0	0.0	100	100	–	0.0
Moxifloxacin	0.12	0.25	≤0.03	0.25	100	0.0	0.0	100	100	–	0.0
Penicillin (oral)	0.25	4	≤0.06	4	36.4	42.4	21.2	NA	36.4	52.5	11.1
Penicillin (iv non-meningitis)	0.25	4	≤0.06	4	88.9	11.1	0.0	NA	NA	NA	NA
SXT	0.12	>2	0.12	>2	83.8	4.0	12.1	83.8	83.8	4.0	12.1

AMC, amoxicillin/clavulanic acid; SXT, trimethoprim/sulfamethoxazole; S, susceptible; I, intermediate; R, resistant; NA, not applicable; min, minimum; max, maximum.

a

Amoxicillin and amoxicillin/clavulanic acid PK/PD susceptibility at high dose shown in parentheses.

Table 3.

MIC distribution data for S. pneumoniae isolates (n = 99)

Drug	Number of isolates at MIC (mg/L)
	≤0.015	≤0.03	0.03	≤0.06	0.06	≤0.12	0.12	0.25	0.5	>0.5	1	>1	2	>2	4	>4	8	>8
Amoxicillin	18		17		10		22	6	2		11		8		1		4
AMC	35		8		17		11	3	3		17				3		2
Azithromycin					6		39	4						50
Cefaclor							1	2	7		40		14		5	30
Ceftriaxone	1		29		11		9	16	3		29		1
Cefuroxime (oral)						40		11	14		4		2		22	6
Clarithromycin	40		8				1		1		15	34
Erythromycin			45		4					50
Levofloxacin									2		89		8
Moxifloxacin		1			23		56	19
Penicillin				36			10	20	4		8		10		11
SXT							52	8	23				4	12

AMC, amoxicillin/clavulanic acid; SXT, trimethoprim/sulfamethoxazole.

Overall, 88.9% (88/99) of S. pneumoniae were penicillin susceptible, 11.1% (11/99) were penicillin intermediate and no isolates were penicillin resistant by CLSI penicillin intravenous (iv) (non-meningitis) breakpoints. However, based on CLSI penicillin oral breakpoints or EUCAST breakpoints only 36.4% (36/99) were penicillin susceptible and 21.2% were penicillin resistant (21/99).

All S. pneumoniae isolates were susceptible to levofloxacin and moxifloxacin by all breakpoints. Amoxicillin or amoxicillin/clavulanic acid susceptibility was 94.9%–98.0% by CLSI or PK/PD breakpoints (no EUCAST breakpoints are available for amoxicillin and amoxicillin/clavulanic acid). Of the cephalosporins, ceftriaxone was very effective with 99.0% (98/99) of isolates susceptible by CLSI and PK/PD breakpoints, but fewer were susceptible by EUCAST criteria (69.7%; 69/99). Cefuroxime (oral) showed lower activity with 69.7% of isolates susceptible by CLSI and PK/PD breakpoints and 51.5% (51/99) by EUCAST breakpoints. Cefaclor had low activity by CLSI standards (50.5% susceptible; 50/99) and even lower activity (10.1% susceptible; 10/99) by PK/PD breakpoints; no isolates were susceptible by EUCAST criteria.

Susceptibility to clarithromycin and erythromycin was 49.5% (49/99) by CLSI, EUCAST and PK/PD breakpoints. Azithromycin susceptibility was also 49.5% by CLSI and EUCAST breakpoints but slightly lower (45.5%; 45/99) using PK/PD breakpoints. Trimethoprim/sulfamethoxazole susceptibility was 83.8% (83/99) by all breakpoints.

Antibiotic susceptibility in S. pneumoniae was compared according to specimen source, age group and penicillin susceptibility. There was no significant difference in antibiotic susceptibility based on specimen source or patient age (data not shown). Susceptibility to the tested antibiotics against isolates stratified by penicillin susceptibility (CLSI oral breakpoints only) is shown in Figure 1. All penicillin-susceptible isolates were >90% susceptible to all agents with available CLSI breakpoints. The fluoroquinolones, amoxicillin, amoxicillin/clavulanic acid and ceftriaxone retained full activity against penicillin-intermediate isolates, whereas the other antibacterials (except trimethoprim/sulfamethoxazole, 86% susceptibility) showed significantly less activity against penicillin-intermediate isolates compared with penicillin-susceptible isolates (P ≤ 0.001). Susceptibility to cefaclor and the macrolides did not exceed 40%, while 76% of isolates were susceptible to cefuroxime (oral). Furthermore, all agents except fluoroquinolones and ceftriaxone were significantly less active against penicillin-resistant isolates compared with penicillin-intermediate isolates and/or penicillin-susceptible isolates, with <10% susceptible to the oral cephalosporins and the macrolides.

Figure 1.

Percentage susceptibility rates (with 95% CI) for antimicrobials against S. pneumoniae according to penicillin susceptibility (CLSI oral breakpoints). PSSP, penicillin-susceptible S. pneumoniae; PISP, penicillin-intermediate S. pneumoniae; PRSP, penicillin-resistant S. pneumoniae. *Susceptibility significantly lower in PRSP than in PSSP or in PISP (P = 0.003 for amoxicillin and amoxicillin/clavulanic acid and P = 0.03 for trimethoprim/sulfamethoxazole). **Susceptibility lower in PISP compared with PSSP [P = 0.001 for cefuroxime (oral) and P < 0.0001 for cefaclor] and in PRSP compared with PISP (P < 0.0001 for cefuroxime and P = 0.001 for cefaclor). ***Susceptibility significantly lower in PISP and PRSP compared with PSSP (P < 0.0001).

H. influenzae isolates

A total of 52 H. influenzae isolates were collected for this study. The origins of the isolates included sputum (n = 49; 94.2%) and bronchoalveolar lavage (n = 3; 5.8%). Most isolates (n = 34; 65.4%) came from adults, 16 (30.8%) were from elderly patients, 1 isolate came from a child, and for 1 isolate the age of the patient was not specified. Eight isolates were β-lactamase-positive (8/52; 15.4%) and 44 (84.6%) were β-lactamase-negative. None was β-lactamase-negative-ampicillin-resistant (BLNAR) by CLSI or EUCAST breakpoints.

Summary MIC and susceptibility data for all 52 H. influenzae isolates are shown in Table 4. MIC distribution data are given in Table 5.

Table 4.

MIC and susceptibility data for H. influenzae isolates (n = 52)

					Susceptibility using indicated breakpoints
	MIC (mg/L)				CLSI			PK/PD	EUCAST
Antimicrobial	50%	90%	min	max	%S	%I	%R	%S	%S	%I	%R
Amoxicillina	≤2	64	≤2	>128	NA	NA	NA	84.6 (84.6)	84.6	–	15.4
AMCa	1	2	0.25	4	100	–	0.0	96.2 (100)	96.2	–	3.8
Ampicillin	1	64	0.12	>128	84.6	0.0	15.4	NA	84.6	–	15.4
Azithromycin	2	4	0.12	16	94.2	–	5.8	3.8	3.8	90.4	5.8
Cefaclor	4	16	1	32	88.5	7.7	3.8	0.0	NA	NA	NA
Ceftriaxone	≤0.03	≤0.03	≤0.03	0.5	100	–	–	100	96.2	–	3.8
Cefuroxime (oral)	1	2	0.25	8	98.1	1.9	0.0	69.2	0.0	69.2	30.8
Clarithromycin	8	>16	4	>16	61.5	26.9	11.5	0.0	0.0	100	0.0
Levofloxacin	0.015	0.06	0.008	>2	96.2	–	–	96.2	96.2	–	3.8
Moxifloxacin	0.015	0.12	0.008	>1	96.2	–	–	96.2	96.2	–	3.8
SXT	0.12	>2	≤0.015	>2	71.2	0.0	28.8	71.2	71.2	0.0	28.8

AMC, amoxicillin/clavulanic acid; SXT, trimethoprim/sulfamethoxazole; S, susceptible; I, intermediate; R, resistant; NA, not applicable; min, minimum; max, maximum.

a

Amoxicillin and amoxicillin/clavulanic acid PK/PD susceptibility at high dose shown in parentheses.

Table 5.

MIC distribution data for all H. influenzae isolates (n = 52)

Drug	Number of isolates at MIC (mg/L)
	0.008	≤0.015	0.015	≤0.03	0.03	0.06	0.12	0.25	≤0.5	0.5	1	>1	≤2	2	>2	4	8	16	>16	32	64	128	>128
Amoxicillin													44							1	4	1	2
AMC								9		14	9			18		2
Ampicillin							3	3		2	36									2	1	2	3
Azithromycin							2				16			28		3		3
Cefaclor											2			18		13	13	4		2
Ceftriaxone				49		1		1		1
Cefuroxime (oral)								1		24	11			13		2	1
Clarithromycin																3	29	14	6
Levofloxacin	4		28		9	9									2
Moxifloxacin	3		29		9	4	5					2
SXT		2			2	19	6	7		1					15

AMC, amoxicillin/clavulanic acid; SXT, trimethoprim/sulfamethoxazole.

All isolates were susceptible to ceftriaxone by CLSI and PK/PD breakpoint guidelines, with 96.2% susceptible by EUCAST criteria. Amoxicillin/clavulanic acid susceptibility was also 100% by CLSI and high-dose PK/PD breakpoints, and was 96.2% (50/52) by lower-dose PK/PD and by EUCAST breakpoints. The rate of susceptibility to levofloxacin and moxifloxacin was also 96.2% and was consistent across all breakpoints. Susceptibility to cefuroxime (oral) was 98.1% (51/52) by CLSI, but this reduced to 69.2% (36/52) by PK/PD and to 0% by EUCAST criteria. A pronounced difference between breakpoints was seen with azithromycin, with a high rate of susceptibility by CLSI (94.2%; 49/52), but greatly reduced susceptibility by PK/PD and EUCAST breakpoints (3.8%; 2/52). This breakpoint difference was also seen with clarithromycin susceptibility, although activity was much lower: 61.5% susceptible (32/52) by CLSI and 0% susceptible by PK/PD and EUCAST.

Similar variability was seen for cefaclor between CLSI and PK/PD breakpoints: 88.5% (46/52) and 0%, respectively (no EUCAST breakpoints are available for cefaclor). Trimethoprim/sulfamethoxazole susceptibility was seen in 71.2% (37/52) of isolates by all three breakpoint categories.

Antibiotic susceptibility in H. influenzae was compared between adult and elderly patients and no significant difference was found (P value >0.05, data not shown). Sample sizes were too small to compare susceptibility by specimen source.

Discussion

Although antimicrobial therapy of community-acquired respiratory infections is usually empirical, information regarding local resistance patterns is often lacking and even treatment guidelines published by professional societies are infrequently based on pertinent resistance levels.⁹ To add to the knowledge base of resistance in respiratory pathogens in Greece, this report examined susceptibility of recently collected isolates of S. pneumoniae and H. influenzae.

For S. pneumoniae, high non-susceptibility rates in the region of 50% or greater were found for the macrolides, oral penicillin and cefaclor by all breakpoint standards. Only amoxicillin, amoxicillin/clavulanic acid and the fluoroquinolones showed susceptibility levels in the region of 95% or higher by all available breakpoints. Antimicrobial activity was especially low among penicillin-resistant isolates, with susceptibility to oral cephalosporins and macrolides <10% and amoxicillin and amoxicillin/clavulanic acid susceptibility reduced to 76%; only the fluoroquinolones and ceftriaxone remained active against all penicillin-resistant isolates. High prevalence of resistant S. pneumoniae in Greece has been reported in the past; however, much of the data are from the last decade.^10,–12 One more recent study examined isolates collected in 2011 and 2012 from patients with community-acquired respiratory tract infections in one university hospital in Greece. Susceptibility rates (using CLSI criteria) to penicillin (63.2%), cefuroxime (80.8%), macrolides (71.2%) and trimethoprim/sulfamethoxazole (83.2%) were higher than those in the current study with the exception of the last agent, which was almost the same.¹¹ This could represent an increase in the intervening years or a different resistance profile at this hospital or region in Greece. Compared with global isolates from 2014 collected as part of a worldwide surveillance programme from patients with respiratory infections and a range of other infections, the susceptibility rates in the current study were lower than the worldwide average for azithromycin (49.5% versus 62.2% globally, using CLSI criteria) and oral penicillin (36.4% versus 61.5% globally), but not for amoxicillin/clavulanic acid (94.9% versus 89.4% globally) or trimethoprim/sulfamethoxazole (83.8% versus 66.0% globally).¹³ Interestingly, Maraki and Papadakis¹¹ also found higher susceptibility levels to trimethoprim/sulfamethoxazole from 2009 to 2012 compared with an earlier study of isolates from Greece from 2001–08. They attributed this improvement to decreased outpatient consumption of the antimicrobial in their area, as a connection between consumption of trimethoprim/sulfamethoxazole and decreased resistance in S. pneumoniae has been shown.^11,14 A recent study of respiratory tract infections also confirmed that macrolide resistance in Greece was high compared with worldwide averages, with reported global resistance rates of 40% as well rates in Europe of in the region of 30%, much lower than the 50% found in the current study.¹⁵

Among H. influenzae, the β-lactamase rate in Greece was 15.4%, similar to the findings in the 2009–12 study in Greece (13.8%),¹⁰ but significantly higher than the rate seen in neighbouring Turkey (6.8%) from 2011 to 2013 (P = 0.049),¹⁶ and higher than the rate of 8.4% in the nearby countries of Bulgaria, Croatia, Romania and Serbia in 2014–16.¹⁷ Nevertheless, susceptibility of H. influenzae was higher in Greece than for S. pneumoniae, with amoxicillin/clavulanic acid, ceftriaxone, cefuroxime and the fluoroquinolones exceeding 95% susceptibility by CLSI criteria. Azithromycin susceptibility was seen in 94.2% of isolates by CLSI breakpoints, but clarithromycin susceptibility was lower (61.5%), and both macrolides showed very poor susceptibility (<5%) by PK/PD and EUCAST criteria. Comparing these results with the 2011–12 study in Greece, amoxicillin/clavulanic acid and the fluoroquinolones were similarly effective (<2% resistance), and clarithromycin showed correspondingly high resistance levels of 29.2% using CLSI breakpoints. The 2014 global study also showed similar susceptibility levels for amoxicillin/clavulanic acid and fluoroquinolones (>99% effective), but additionally clarithromycin susceptibility was high at 94.5% using CLSI criteria. Again, however, <5% of isolates were susceptible to macrolides by EUCAST breakpoints.¹³ As has been shown in other SOAR studies that have systematically compared breakpoints,^18,19 this variability is significant, impacting not only clinicians trying to interpret antimicrobial resistance data but also public health officials, epidemiologists and microbiologists struggling to compare data across studies from different geographic regions and over time using different MIC interpretation criteria. All would be much better served if breakpoints were harmonized.

Two of the 52 H. influenzae isolates (3.8%) were found to be non-susceptible to fluoroquinolones. Such resistance is still rare in most regions of the world;^13,15,19,20 however, the overall sample size was small in this study and the significance of the finding is unclear. Interestingly, the rate of 3.8% non-susceptibility is almost exactly the same as the rate found in neighbouring Turkey in 2011–13 (3.5%).¹⁶ Further monitoring of susceptibility to fluoroquinolones is important in order to determine whether there is a significant problem.

Even though this study included only one centre in Greece, which is a limitation of the study, it was able to paint a picture of local resistance levels, confirming results seen in earlier studies (also from just one centre) and contrasting the findings with the surrounding region and with global susceptibility rates. However, it is possible that susceptibility rates elsewhere in the country may differ. Owing to the high resistance level of S. pneumoniae to macrolides and the reduced activity of clarithromycin against H. influenzae, it appears that these agents are not appropriate as monotherapy for community-acquired pneumonia in Greece.¹¹ Amoxicillin/clavulanic acid, on the other hand, maintained excellent in vitro activity and, as opposed to the similarly effective fluoroquinolones, is safe to use in paediatric patients. Amoxicillin/clavulanic acid also maintained activity against β-lactamase-positive H. influenzae, and is therefore able to cover the two most common respiratory pathogens in Greece. Continued surveillance of resistance trends, ideally expanded to more Greek sites, is necessary to track whether resistance levels increase further and whether the resistance patterns found in the current study apply to other parts of the country.

Acknowledgements

We thank Gudrun Maechler and Olena Sopko for reviewing the manuscript. We also thank Julia Vassiliadou and Despina Karidi-Andrioti for their help in recruiting laboratories. We thank Dr Sibylle Lob for her assistance in the writing of this manuscript.

Funding

This study was funded by GlaxoSmithKline.

Transparency declarations

This article forms part of a Supplement sponsored by GlaxoSmithKline. D. Torumkuney is an employee of GlaxoSmithKline and also holds shares in GlaxoSmithKline. I. Morrissey is an employee of IHMA, a medical communication and consultancy company, who participated in the exploration, interpretation of the results and preparation of this manuscript on behalf of GlaxoSmithKline. IHMA also provided medical writing support in the form of writing assistance, collating authors' comments, grammatical editing and referencing that was paid for by GlaxoSmithKline. J. Papaparaskevas has no conflicts of interest. Editorial assistance was provided by Tracey Morris, Livewire Editorial Communications.