Results from the Survey of Antibiotic Resistance (SOAR) 2015–18 in Tunisia, Kenya and Morocco: data based on CLSI, EUCAST (dose-specific) and pharmacokinetic/pharmacodynamic (PK/PD) breakpoints

Abstract

Objectives

To determine antibiotic susceptibility of community-acquired respiratory tract infection (CA-RTI) isolates of Streptococcus pneumoniae and Haemophilus influenzae collected in 2015–18 from Tunisia, Kenya and Morocco.

Methods

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

Results

S. pneumoniae isolates from Tunisia (n = 79), Kenya (n = 44) and Morocco (n = 19) and H. influenzae isolates (n = 74) from Tunisia only were collected and analysed. Low antibiotic susceptibility was observed in S. pneumoniae from Tunisia, with >90% susceptible only to the fluoroquinolones (all breakpoints), penicillin (CLSI IV and EUCAST high-dose) and ceftriaxone (CLSI, EUCAST high-dose and PK/PD breakpoints). In addition, isolate susceptibility in Kenya was >90% to amoxicillin and amoxicillin/clavulanic acid (CLSI and PK/PD breakpoints). Antibiotic activity was highest in Morocco, where ≥89.5% of pneumococci were susceptible to most antibiotics, excluding trimethoprim/sulfamethoxazole (68.4% by CLSI or PK/PD and 79%–84.2% by EUCAST), macrolides (79%–84.2% by all breakpoints) and cefaclor (0% by EUCAST and 52.6% by PK/PD). The majority (≥86.5%) of H. influenzae isolates from Tunisia were susceptible to most antibiotics by all available breakpoints, except ampicillin and amoxicillin (almost one-third were β-lactamase positive), trimethoprim/sulfamethoxazole (51.4%–56.8%), cefaclor (1.4% by PK/PD), cefuroxime (4.1% by EUCAST), macrolides (1.4%–2.7% by PK/PD) and cefdinir (66.2% by PK/PD). The application of different EUCAST breakpoints for low and higher doses for some of the antibiotics (amoxicillin, amoxicillin/clavulanic acid, ampicillin, penicillin, ceftriaxone, clarithromycin, erythromycin, levofloxacin and trimethoprim/sulfamethoxazole) allowed, for the first time in a SOAR study, the effect of raising the dosage on susceptibility to be quantified.

Conclusions

Low antibiotic susceptibility was observed in S. pneumoniae from Tunisia, but susceptibility was higher in isolates from Kenya and highest in those from Morocco. H. influenzae from Tunisia were highly susceptible to most antibiotics. These factors are important in decision making for empirical therapy of CA-RTIs.

Introduction

Community-acquired bacterial pneumonia is an important world health problem associated with mortality and significant morbidity, with much of the burden occurring in the hospital setting.¹ Lower respiratory tract infections (LRTIs) (including community- and hospital-acquired infections) have been ranked as the fifth overall cause of death in the world, despite being largely preventable by vaccination.² Antibiotic therapy plays an important role in the treatment of most LRTIs and management of these infections very much relies on appropriate empirical antibiotic therapy based on local and international guidelines,^2,3 especially in this era of increasing antibiotic resistance in many parts of the world.¹ However, the WHO has estimated that in low- and middle-income countries about 80% of antibiotics are used in the community and of these 20%–50% are used inappropriately.⁴ Much of this is down to self-medication due to poor accessibility, affordability and generally poor health facilities in these countries.⁵

The major bacteria associated with community-acquired respiratory tract infections (CA-RTIs) are Streptococcus pneumoniae and Haemophilus influenzae, and both pathogens have shown increasing resistance to first-line antibiotics such as penicillin and ampicillin.^1,6 As rates of resistance are variable from country to country, surveillance data can provide useful information to guide local antibiotic policies.

The Survey of Antibiotic Resistance (SOAR) is an international antibiotic resistance surveillance study that focuses on key respiratory pathogens from community-acquired infections and has been running since 2002 in the Middle East, Africa, Latin America, Asia-Pacific and Commonwealth of Independent States countries. For this report, recent SOAR data from hospitals in Tunisia, Kenya and Morocco were analysed to provide a picture of the current state of antibiotic susceptibility of S. pneumoniae and H. influenzae associated with CA-RTIs.

Materials and methods

Collaborating centres

The following four centres took part in the study: one site in Tunisia (Habib Bourguiba University Hospital, Sfax), two sites in Kenya (Kenyatta National Hospital and Aga Khan University Hospital) and one site in Morocco (Centre Hospitalo Universitaire Ibn Rochd, Casablanca).

Clinical isolates

Isolates of H. influenzae and S. pneumoniae from CA-RTIs (patients with respiratory tract infections in the community not hospitalized for more than 48 h) were sent to a central laboratory (LGC, Fordham, UK) in transport swabs, where they were subcultured 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 not accepted.

Susceptibility testing

Isolates were evaluated for antibiotic susceptibility using broth microdilution methodology recommended by CLSI.⁷ Both pathogens were assessed for susceptibility to amoxicillin, amoxicillin/clavulanic acid (2:1), ampicillin, azithromycin, cefaclor, cefdinir, cefditoren, cefixime, cefpodoxime, ceftriaxone, cefuroxime, clarithromycin, levofloxacin, moxifloxacin and trimethoprim/sulfamethoxazole (1:19). S. pneumoniae was also tested for susceptibility to penicillin and erythromycin.

Susceptibility to the study drugs was calculated based on CLSI breakpoints, EUCAST (dose-specific) breakpoints and pharmacokinetic/pharmacodynamic (PK/PD) breakpoints.^8–10 These breakpoints are shown in Tables 1–3. To fully assess antibiotics where high-dose therapies are available, susceptibility using EUCAST criteria was also calculated by combining percentage susceptible and percentage intermediate (susceptible, increased exposure) into the susceptible category.⁹ The antibiotics with high-dose availability assessed in this way were as follows: amoxicillin (0.75–1 g oral, 3 × daily), amoxicillin/clavulanic acid (0.875 g amoxicillin/0.125 g clavulanic acid oral, 3 × daily), ampicillin (2 g IV 4 × daily), penicillin (2.4 g IV, 2 MU 4–6 × daily), ceftriaxone (2 g IV, 2 × daily), clarithromycin (0.5 g oral, 2 × daily), erythromycin (1 g oral or IV, 4 × daily), levofloxacin (0.75 g oral 2 × daily, or 0.4 g IV 3 × daily) and trimethoprim/sulfamethoxazole (0.24 g trimethoprim/1.2 g sulfamethoxazole oral or IV, 2 × daily).⁹

Table 1.

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

Antibiotic	S. pneumoniae			H. influenzae
	S	I	R	S	I	R
Amoxicillin	≤2	4	≥8	–	–	–
Amoxicillin/ clavulanic acida	≤2	4	≥8	≤4	–	≥8
Ampicillin	–	–	–	≤1	2	≥4
Penicillin (2.4 g 2 MU × 4–6 IV)	≤2	4	≥8	–	–	–
Penicillin (oral)	≤0.06	0.12–1	≥2	–	–	–
Cefaclor	≤1	2	≥4	≤8	16	≥32
Cefdinir	≤0.5	1	≥2	≤1	–	–
Cefditoren	–	–	–	–	–	–
Cefixime	–	–	–	≤1	–	–
Cefpodoxime	≤0.5	1	≥2	≤2	–	–
Ceftriaxone	≤1	2	≥4	≤2	–	–
Cefuroximeb	≤1	2	≥4	≤4	8	≥16
Azithromycin	≤0.5	1	≥2	≤4	–	–
Clarithromycin	≤0.25	0.5	≥1	≤8	16	≥32
Erythromycin	≤0.25	0.5	≥1	–	–	–
Levofloxacin	≤2	4	≥8	≤2	–	–
Moxifloxacin	≤1	2	≥4	≤1	–	–
Trimethoprim/ sulfamethoxazolec	≤0.5	1–2	≥4	≤0.5	1–2	≥4

S, susceptible; I, intermediate; R, resistant; –, not applicable.

a

Amoxicillin/clavulanic acid was tested at a 2:1 amoxicillin to clavulanic acid ratio; breakpoints are expressed as the amoxicillin component.

b

Breakpoints used are for cefuroxime axetil (oral).

c

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

Table 2.

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

Antibiotic	S. pneumoniae		H. influenzae
	S	R	S	R
Amoxicillin (0.5 g × 3 oral)	≤0.5	>1	–	–
Amoxicillin (0.75–1 g × 3 oral)	≤1	>1	≤2	>2
Amoxicillin/clavulanic acida (0.5 g/0.125 g × 3 oral)	≤0.5	>1	–	–
Amoxicillin/clavulanic acida (0.875 g/0.125 g × 3 oral)	≤1	>1	≤2	>2
Ampicillin (2 g × 3 IV)	≤0.5	>2	≤1	>1
Ampicillin (2 g × 4 IV)	≤2	>2	≤1	>1
Penicillin (0.6 g 1 MU × 4 IV)	≤0.06	>2	–	–
Penicillin (2.4 g 2 MU × 4–6 IV)	≤2	>2	–	–
Cefaclor	≤0.03	>0.5	–	–
Cefdinir	–	–	–	–
Cefditoren	–	–	–	–
Cefixime	–	–	≤0.12	>0.12
Cefpodoxime	≤0.25	>0.5	≤0.25	>0.25
Ceftriaxone (1 g × 1 IV)	≤0.5	>2	≤0.12	>0.12
Ceftriaxone (2 g × 2 IV)	≤2	>2	≤0.12	>0.12
Cefuroximeb	≤0.25	>0.5	≤0.12	>1
Azithromycin	≤0.25	>0.5	–	–
Clarithromycin (0.25 g × 2 oral)	≤0.25	>0.5	–	–
Clarithromycin (0.5 g × 2 oral)	≤0.5	>0.5	–	–
Erythromycin (0.5 g × 2–4 oral or 0.5 g × 2–4 IV)	≤0.25	>0.5	–	–
Erythromycin (1 g × 4 oral or 1 g × 4 IV)	≤0.5	>0.5	–	–
Levofloxacin (0.5 g × 2 oral or 0.4 g × 2 IV)	≤2	>2	≤0.06	>0.06
Levofloxacin (0.75 g × 2 oral or 0.4 g × 3 IV)	≤2	>2	≤0.06	>0.06
Moxifloxacin	≤0.5	>0.5	≤0.12	>0.12
Trimethoprim/sulfamethoxazolec (0.16 g/0.8 g × 2 oral or IV)	≤1	>2	≤0.5	>1
Trimethoprim/sulfamethoxazolec (0.24 g/1.2 g × 2 oral or IV)	≤2	>2	≤1	>1

The I category is not listed but is interpreted as the values between the S and the R breakpoints. If the S and R breakpoints are the same value there is no I category.⁹

S, susceptible; R, resistant; –, not applicable.

a

Amoxicillin/clavulanic acid was tested at a 2:1 amoxicillin to clavulanic acid ratio; breakpoints are expressed as the amoxicillin component.

b

Breakpoints used are for cefuroxime axetil (oral).

c

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

Table 3.

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

Antibiotic	S. pneumoniae and H. influenzae
	S only
Amoxicillin (1.5 g/day)	≤2
Amoxicillin (4 g/day)	≤4
Amoxicillin/clavulanic acida (1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children)	≤2
Amoxicillin/clavulanic acidb (4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children)	≤4
Ampicillin	–
Penicillin	–
Cefaclor	≤0.5
Cefdinir	≤0.25
Cefditoren	–
Cefixime	≤1
Cefpodoxime	≤0.5
Ceftriaxone	≤1
Cefuroximec	≤1
Azithromycin	≤0.12
Clarithromycin	≤0.25
Erythromycin	≤0.25
Levofloxacin	≤2
Moxifloxacin	≤1
Trimethoprim/sulfamethoxazoled	≤0.5

S, susceptible; –, not applicable.

a

Amoxicillin/clavulanic acid for low dose in adults/children.

b

Amoxicillin/clavulanic acid for high dose in adults/children.

c

Breakpoints used are for cefuroxime axetil (oral).

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 35218 were included on each day of testing. Results of susceptibility testing were accepted if the results for the control strains were within published limits. Differences in susceptibility (using CLSI criteria only) across countries and penicillin susceptibility (S. pneumoniae only) were assessed for statistical significance with Fisher’s exact test using XLSTAT version 2019.1.3.57796. A P value <0.05 was considered statistically significant.

Ethics

SOAR studies are not human subject studies. During the study, only microorganisms were tested.

Results

S. pneumoniae isolates

A total of 142 S. pneumoniae isolates were collected from four centres in Tunisia, Kenya and Morocco from 2015 to 2018. Most pneumococci came from sputum (n = 58; 40.8%), transtracheal aspirate (n = 31; 21.8%) and blood (n = 30; 21.1%). Other isolates were from middle ear effusion (n = 12; 8.5%), pleural fluid (n = 5; 3.5%), bronchoalveolar lavage (n = 4; 2.8%) and sinuses (n = 2; 1.4%). Most isolates (n = 60; 42.3%) came from adult patients (aged 13–64 years), 48 (33.8%) were from paediatric patients (≤12 years) and 34 (23.9%) were from elderly patients (aged ≥65 years).

Summary MIC and susceptibility data for all 142 S. pneumoniae isolates by country of origin are shown in Tables 4–12 and Figures 1–3. MIC distribution data are shown in Tables S1–S3 (available as Supplementary data at JAC Online).

Figure 1.

Percentage susceptibility rates (with 95% CI) based on CLSI breakpoints for antibiotics against all S. pneumoniae isolates collected in the three African countries in the period 2015–18. ^aSusceptibility was significantly higher among isolates from Kenya and Morocco than Tunisia (P < 0.05). ^bSusceptibility was significantly higher among isolates from Morocco than Kenya and Tunisia (P < 0.05). ^cSusceptibility was significantly higher among isolates from Morocco than from Kenya, and higher from both countries compared with Tunisia (P < 0.05). ^dSusceptibility was significantly higher among isolates from Tunisia than Morocco (P < 0.05). ^eSusceptibility was significantly higher among isolates from Morocco and Tunisia than Kenya (P < 0.05).

Figure 2.

Percentage susceptibility rates based on EUCAST (dose-specific) breakpoints for antibiotics against all S. pneumoniae isolates collected in the three African countries in the period 2015–18.

Figure 3.

Percentage susceptibility rates based on PK/PD breakpoints for antibiotics against all S. pneumoniae isolates collected in the three African countries in the period 2015–18. Low-dose amoxicillin/clavulanic acid = 1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children. High-dose amoxicillin/clavulanic acid = 4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children.

Table 4.

MIC and susceptibility data for S. pneumoniae isolates (n = 79) from Tunisia using CLSI breakpoints

Antibiotic	MIC (mg/L)				CLSI susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin	2	8	≤0.015	8	69.6	15.2	15.2
Amoxicillin/clavulanic acid	2	8	≤0.015	8	72.2	12.7	15.2
Ampicillin	4	8	≤0.03	8	–	–	–
Penicillin (2.4 g 2 MU × 4–6 IV)	1	2	≤0.06	4	93.7	6.3	0.0
Penicillin (oral)	1	2	≤0.06	4	13.9	36.7	49.4
Cefaclor	>32	>32	0.5	>32	17.7	5.1	77.2
Cefdinir	4	8	0.03	16	26.6	1.3	72.2
Cefditoren	0.5	1	≤0.015	1	–	–	–
Cefixime	16	32	0.12	64	–	–	–
Cefpodoxime	2	4	≤0.015	8	26.6	8.9	64.6
Ceftriaxone	1	1	≤0.015	2	91.1	8.9	0.0
Cefuroxime	8	>16	≤0.015	>16	24.1	5.1	70.9
Azithromycin	>32	>32	≤0.03	>32	20.3	0.0	79.8
Clarithromycin	>16	>16	≤0.015	>16	20.3	0.0	79.8
Erythromycin	>16	>16	≤0.015	>16	20.3	0.0	79.8
Levofloxacin	1	1	0.5	2	100	0.0	0.0
Moxifloxacin	0.12	0.25	0.06	0.25	100	0.0	0.0
Trimethoprim/sulfamethoxazole	2	8	≤0.015	8	60.8	13.9	25.3

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 5.

MIC and susceptibility data for S. pneumoniae isolates (n = 79) from Tunisia using EUCAST (dose-specific) breakpoints

Antibiotic	MIC (mg/L)				EUCAST susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin (0.5 g × 3 oral)	2	8	≤0.015	8	26.6	5.1	68.4
Amoxicillin (0.75–1 g × 3 oral)	2	8	≤0.015	8	31.6	–	68.4
Amoxicillin/clavulanic acid (0. 5 g/0.125 g × 3 oral)	2	8	≤0.015	8	25.3	10.1	64.6
Amoxicillin/clavulanic acid (0.875 g/0.125 g × 3 oral)	2	8	≤0.015	8	35.4	–	64.6
Ampicillin (2 g × 3 IV)	4	8	≤0.03	8	27.9	21.5	50.6
Ampicillin (2 g × 4 IV)	4	8	≤0.03	8	49.4	–	50.6
Penicillin (0.6 g 1 MU × 4 IV)	1	2	≤0.06	4	13.9	79.8	6.3
Penicillin (2.4 g 2 MU × 4–6 IV)	1	2	≤0.06	4	93.7	–	6.3
Cefaclor	>32	>32	0.5	>32	0.0	12.7	87.3
Cefdinir	4	8	0.03	16	–	–	–
Cefditoren	0.5	1	≤0.015	1	–	–	–
Cefixime	16	32	0.12	64	–	–	–
Cefpodoxime	2	4	≤0.015	8	25.3	1.3	73.4
Ceftriaxone (1 g × 1 IV)	1	1	≤0.015	2	46.8	53.2	0.0
Ceftriaxone (2 g × 2 IV)	1	1	≤0.015	2	100	–	0.0
Cefuroxime	8	>16	≤0.015	>16	15.2	5.1	79.8
Azithromycin	>32	>32	≤0.03	>32	20.3	0.0	79.8
Clarithromycin (0.25 g × 2 oral)	>16	>16	≤0.015	>16	20.3	0.0	79.8
Clarithromycin (0.5 g × 2 oral)	>16	>16	≤0.015	>16	20.3	–	79.8
Erythromycin (0.5 g × 2–4 oral or × 2–4 IV)	>16	>16	≤0.015	>16	20.3	0.0	79.8
Erythromycin (1 g × 4 oral or × 4 IV)	>16	>16	≤0.015	>16	20.3	–	79.8
Levofloxacin (0.5 g × 2 oral or 0.4 g × 2 IV)	1	1	0.5	2	100	–	0.0
Levofloxacin (0.75 g × 2 oral or 0.4 g × 3 IV)	1	1	0.5	2	100	–	0.0
Moxifloxacin	0.12	0.25	0.06	0.25	100	–	0.0
Trimethoprim/sulfamethoxazole (0.16 g/0.8 g × 2 oral or IV)	2	8	≤0.015	8	69.6	5.1	25.3
Trimethoprim/sulfamethoxazole (0.24 g/1.2 g × 2 oral or IV)	2	8	≤0.015	8	74.7	–	25.3

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 6.

MIC and susceptibility data for S. pneumoniae isolates (n = 79) from Tunisia using PK/PD breakpoints

Antibiotic	MIC (mg/L)				PK/PD susceptibility
	50%	90%	min	max	%S
Amoxicillin (1.5 g/day)	2	8	≤0.015	8	69.6
Amoxicillin (4 g/day)	2	8	≤0.015	8	84.8
Amoxicillin/clavulanic acid (1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children)	2	8	≤0.015	8	72.2
Amoxicillin/clavulanic acid (4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children)	2	8	≤0.015	8	84.8
Ampicillin	4	8	≤0.03	8	–
Penicillin	1	2	≤0.06	4	–
Cefaclor	>32	>32	0.5	>32	12.7
Cefdinir	4	8	0.03	16	20.3
Cefditoren	0.5	1	≤0.015	1	–
Cefixime	16	32	0.12	64	20.3
Cefpodoxime	2	4	≤0.015	8	26.6
Ceftriaxone	1	1	≤0.015	2	91.1
Cefuroxime	8	>16	≤0.015	>16	24.1
Azithromycin	>32	>32	≤0.03	>32	20.3
Clarithromycin	>16	>16	≤0.015	>16	20.3
Erythromycin	>16	>16	≤0.015	>16	20.3
Levofloxacin	1	1	0.5	2	100
Moxifloxacin	0.12	0.25	0.06	0.25	100
Trimethoprim/sulfamethoxazole	2	8	≤0.015	8	60.8

min, minimum; max, maximum; S, susceptible; –, not applicable.

Table 7.

MIC and susceptibility data for S. pneumoniae isolates (n = 44) from Kenya using CLSI breakpoints

Antibiotic	MIC (mg/L)				CLSI susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin	0.5	2	≤0.015	8	90.9	2.3	6.8
Amoxicillin/clavulanic acid	0.5	2	≤0.015	8	90.9	2.3	6.8
Ampicillin	1	4	≤0.03	16	–	–	–
Penicillin (2.4 g 2 MU × 4–6 IV)	0.5	2	≤0.06	4	90.9	9.1	0.0
Penicillin (oral)	0.5	2	≤0.06	4	20.5	56.8	22.7
Cefaclor	8	>32	0.5	>32	13.6	20.5	65.9
Cefdinir	0.5	8	0.06	16	50.0	9.1	40.9
Cefditoren	0.25	1	≤0.015	2	–	–	–
Cefixime	4	32	0.25	64	–	–	–
Cefpodoxime	0.5	4	0.03	8	52.3	6.8	40.9
Ceftriaxone	0.25	1	≤0.015	2	95.5	4.6	0.0
Cefuroxime	1	8	≤0.015	16	54.6	6.8	38.6
Azithromycin	0.06	>32	≤0.03	>32	61.4	2.3	36.4
Clarithromycin	0.03	>16	≤0.015	>16	63.6	2.3	34.1
Erythromycin	0.06	>16	≤0.015	>16	63.6	0.0	36.4
Levofloxacin	1	1	0.5	1	100	0.0	0.0
Moxifloxacin	0.12	0.25	0.12	0.25	100	0.0	0.0
Trimethoprim/sulfamethoxazole	8	16	0.25	>32	6.8	2.3	90.9

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 8.

MIC and susceptibility data for S. pneumoniae isolates (n = 44) from Kenya using EUCAST (dose-specific) breakpoints

Antibiotic	MIC (mg/L)				EUCAST susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin (0.5 g × 3 oral)	0.5	2	≤0.015	8	54.6	18.2	27.3
Amoxicillin (0.75–1 g × 3 oral)	0.5	2	≤0.015	8	72.7	–	27.3
Amoxicillin/clavulanic acid (0.5 g/0.125 g × 3 oral)	0.5	2	≤0.015	8	54.6	31.8	13.6
Amoxicillin/clavulanic acid (0.875 g/0.125 g × 3 oral)	0.5	2	≤0.015	8	86.4	–	13.6
Ampicillin (2 g × 3 IV)	1	4	≤0.03	16	45.5	34.1	20.5
Ampicillin (2 g × 4 IV)	1	4	≤0.03	16	79.5	–	20.5
Penicillin (0.6 g 1 MU × 4 IV)	0.5	2	≤0.06	4	20.5	70.5	9.1
Penicillin (2.4 g 2 MU × 4–6 IV)	0.5	2	≤0.06	4	90.9	–	9.1
Cefaclor	8	>32	0.5	>32	0.0	4.6	95.5
Cefdinir	0.5	8	0.06	16	–	–	–
Cefditoren	0.25	1	≤0.015	2	–	–	–
Cefixime	4	32	0.25	64	–	–	–
Cefpodoxime	0.5	4	0.03	8	40.9	11.4	47.7
Ceftriaxone (1 g × 1 IV)	0.25	1	≤0.015	2	68.2	31.8	0.0
Ceftriaxone (2 g × 2 IV)	0.25	1	≤0.015	2	100	–	0.0
Cefuroxime	1	8	≤0.015	16	34.1	11.4	54.6
Azithromycin	0.06	>32	≤0.03	>32	61.4	0.0	38.6
Clarithromycin (0.25 g × 2 oral)	0.03	>16	≤0.015	>16	63.6	2.3	34.1
Clarithromycin (0.5 g × 2 oral)	0.03	>16	≤0.015	>16	65.9	–	34.1
Erythromycin (0.5 g × 2–4 oral or × 2–4 IV)	0.06	>16	≤0.015	>16	63.6	0.0	36.4
Erythromycin (1 g × 4 oral or × 4 IV)	0.06	>16	≤0.015	>16	63.6	–	36.4
Levofloxacin (0.5 g × 2 oral or 0.4 g × 2 IV)	1	1	0.5	1	100	–	0.0
Levofloxacin (0.75 g × 2 oral or 0.4 g × 3 IV)	1	1	0.5	1	100	–	0.0
Moxifloxacin	0.12	0.25	0.12	0.25	100	–	0.0
Trimethoprim/sulfamethoxazole (0.16 g/0.8 g × 2 oral or IV)	8	16	0.25	>32	6.8	2.3	90.9
Trimethoprim/sulfamethoxazole (0.24 g/1.2 g × 2 oral or IV)	8	16	0.25	>32	9.1	–	90.9

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 9.

MIC and susceptibility data for S. pneumoniae isolates (n = 44) from Kenya using PK/PD breakpoints

Antibiotic	MIC (mg/L)				PK/PD susceptibility
	50%	90%	min	max	%S
Amoxicillin (1.5 g/day)	0.5	2	≤0.015	8	90.9
Amoxicillin (4 g/day)	0.5	2	≤0.015	8	93.2
Amoxicillin/clavulanic acid (1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children)	0.5	2	≤0.015	8	90.9
Amoxicillin/clavulanic acid (4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children)	0.5	2	≤0.015	8	93.2
Ampicillin	1	4	≤0.03	16	–
Penicillin	0.5	2	≤0.06	4	–
Cefaclor	8	>32	0.5	>32	4.6
Cefdinir	0.5	8	0.06	16	36.4
Cefditoren	0.25	1	≤0.015	2	–
Cefixime	4	32	0.25	64	18.2
Cefpodoxime	0.5	4	0.03	8	52.3
Ceftriaxone	0.25	1	≤0.015	2	95.5
Cefuroxime	1	8	≤0.015	16	54.6
Azithromycin	0.06	>32	≤0.03	>32	61.4
Clarithromycin	0.03	>16	≤0.015	>16	63.6
Erythromycin	0.06	>16	≤0.015	>16	63.6
Levofloxacin	1	1	0.5	1	100
Moxifloxacin	0.12	0.25	0.12	0.25	100
Trimethoprim/sulfamethoxazole	8	16	0.25	>32	6.8

min, minimum; max, maximum; S, susceptible; –, not applicable.

Table 10.

MIC and susceptibility data for S. pneumoniae isolates (n = 19) from Morocco using CLSI breakpoints

Antibiotic	MIC (mg/L)				CLSI susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin	≤0.015	2	≤0.015	8	94.7	0.0	5.3
Amoxicillin/clavulanic acid	≤0.015	2	≤0.015	8	94.7	0.0	5.3
Ampicillin	≤0.03	4	≤0.03	8	–	–	–
Penicillin (2.4 g 2 MU × 4–6 IV)	≤0.06	2	≤0.06	4	94.7	5.3	0.0
Penicillin (oral)	≤0.06	2	≤0.06	4	89.5	0.0	10.5
Cefaclor	0.5	>32	0.12	>32	89.5	0.0	10.5
Cefdinir	0.06	8	0.03	8	89.5	0.0	10.5
Cefditoren	≤0.015	1	≤0.015	1	–	–	–
Cefixime	0.25	32	0.12	32	–	–	–
Cefpodoxime	0.03	2	≤0.015	4	89.5	0.0	10.5
Ceftriaxone	≤0.015	1	≤0.015	1	100	0.0	0.0
Cefuroxime	0.03	8	≤0.015	16	89.5	0.0	10.5
Azithromycin	0.06	>32	≤0.03	>32	84.2	0.0	15.8
Clarithromycin	0.03	>16	≤0.015	>16	84.2	0.0	15.8
Erythromycin	0.03	>16	≤0.015	>16	79.0	0.0	21.1
Levofloxacin	1	4	0.5	16	89.5	5.3	5.3
Moxifloxacin	0.12	0.25	0.06	4	94.7	0.0	5.3
Trimethoprim/sulfamethoxazole	0.25	8	0.12	16	68.4	15.8	15.8

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 11.

MIC and susceptibility data for S. pneumoniae isolates (n = 19) from Morocco using EUCAST (dose-specific) breakpoints

Antibiotic	MIC (mg/L)				EUCAST susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin (0.5 g × 3 oral)	≤0.015	2	≤0.015	8	89.5	0.0	10.5
Amoxicillin (0.75–1 g × 3 oral)	≤0.015	2	≤0.015	8	89.5	–	10.5
Amoxicillin/clavulanic acid (0.5 g/0.125 g × 3 oral)	≤0.015	2	≤0.015	8	89.5	0.0	10.5
Amoxicillin/clavulanic acid (0.875 g/0.125 g × 3 oral)	≤0.015	2	≤0.015	8	89.5	–	10.5
Ampicillin (2 g × 3 IV)	≤0.03	4	≤0.03	8	89.5	0.0	10.5
Ampicillin (2 g × 4 IV)	≤0.03	4	≤0.03	8	89.5	–	10.5
Penicillin (0.6 g 1 MU × 4 IV)	≤0.06	2	≤0.06	4	89.5	5.3	5.3
Penicillin (2.4 g 2 MU × 4–6 IV)	≤0.06	2	≤0.06	4	94.7	–	5.3
Cefaclor	0.5	>32	0.12	>32	0.0	52.6	47.4
Cefdinir	0.06	8	0.03	8	–	–	–
Cefditoren	≤0.015	1	≤0.015	1	–	–	–
Cefixime	0.25	32	0.12	32	–	–	–
Cefpodoxime	0.03	2	≤0.015	4	89.5	0.0	10.5
Ceftriaxone (1 g × 1 IV)	≤0.015	1	≤0.015	1	89.5	10.5	0.0
Ceftriaxone (2 g × 2 IV)	≤0.015	1	≤0.015	1	100	–	0.0
Cefuroxime	0.03	8	≤0.015	16	89.5	0.0	10.5
Azithromycin	0.06	>32	≤0.03	>32	79.0	5.3	15.8
Clarithromycin (0.25 g × 2 oral)	0.03	>16	≤0.015	>16	84.2	0.0	15.8
Clarithromycin (0.5 g × 2 oral)	0.03	>16	≤0.015	>16	84.2	–	15.8
Erythromycin (0.5 g × 2–4 oral or × 2–4 IV)	0.03	>16	≤0.015	>16	79.0	0.0	21.1
Erythromycin (1 g × 4 oral or × 4 IV)	0.03	>16	≤0.015	>16	79.0	–	21.1
Levofloxacin (0.5 g × 2 oral or 0.4 g × 2 IV)	1	4	0.5	16	89.5	–	10.5
Levofloxacin (0.75 g × 2 oral or 0.4 g × 3 IV)	1	4	0.5	16	89.5	–	10.5
Moxifloxacin	0.12	0.25	0.06	4	94.7	–	5.3
Trimethoprim/sulfamethoxazole (0.16 g/0.8 g × 2 oral or IV)	0.25	8	0.12	16	79.0	5.3	15.8
Trimethoprim/sulfamethoxazole (0.24 g/1.2 g × 2 oral or IV)	0.25	8	0.12	16	84.2	–	15.8

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 12.

MIC and susceptibility data for S. pneumoniae isolates (n = 19) from Morocco using PK/PD breakpoints

Antibiotic	MIC (mg/L)				PK/PD susceptibility
	50%	90%	min	max	%S
Amoxicillin (1.5 g/day)	≤0.015	2	≤0.015	8	94.7
Amoxicillin (4 g/day)	≤0.015	2	≤0.015	8	94.7
Amoxicillin/clavulanic acid (1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children)	≤0.015	2	≤0.015	8	94.7
Amoxicillin/clavulanic acid (4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children)	≤0.015	2	≤0.015	8	94.7
Ampicillin	≤0.03	4	≤0.03	8	–
Penicillin	≤0.06	2	≤0.06	4	–
Cefaclor	0.5	>32	0.12	>32	52.6
Cefdinir	0.06	8	0.03	8	89.5
Cefditoren	≤0.015	1	≤0.015	1	–
Cefixime	0.25	32	0.12	32	89.5
Cefpodoxime	0.03	2	≤0.015	4	89.5
Ceftriaxone	≤0.015	1	≤0.015	1	100
Cefuroxime	0.03	8	≤0.015	16	89.5
Azithromycin	0.06	>32	≤0.03	>32	79.0
Clarithromycin	0.03	>16	≤0.015	>16	84.2
Erythromycin	0.03	>16	≤0.015	>16	79.0
Levofloxacin	1	4	0.5	16	89.5
Moxifloxacin	0.12	0.25	0.06	4	94.7
Trimethoprim/sulfamethoxazole	0.25	8	0.12	16	68.4

min, minimum; max, maximum; S, susceptible; –, not applicable.

S. pneumoniae susceptibility in Tunisia

Of the 79 pneumococcal isolates from Tunisia, only 13.9% were susceptible to penicillin by CLSI oral or EUCAST low-dose IV breakpoints, but susceptibility increased to 93.7% using CLSI IV and EUCAST high-dose breakpoints. All isolates were susceptible to the fluoroquinolones by all breakpoints. The majority of isolates (91.1%–100%) were susceptible to ceftriaxone by CLSI, EUCAST high-dose and PK/PD breakpoints but only 46.8% were susceptible by EUCAST low-dose criteria. Based on CLSI and PK/PD low-dose criteria, susceptibility to amoxicillin and amoxicillin/clavulanic acid was 69.6%–72.2%, but susceptibility was 84.8% by PK/PD high-dose breakpoints; however, only 25.3%–35.4% of isolates were susceptible by EUCAST low- or high-dose breakpoints. Susceptibility of isolates to trimethoprim/sulfamethoxazole was low at 60.8%–69.6% by CLSI, PK/PD and EUCAST low-dose criteria and 74.7% by EUCAST high-dose breakpoints. The proportion of pneumococcal isolates susceptible to the remaining cephalosporins ranged from 17.7% (cefaclor) to 26.6% (cefpodoxime/cefdinir) by CLSI breakpoints, from 12.7% (cefaclor) to 26.6% (cefpodoxime) by PK/PD breakpoints and from 0% (cefaclor) to 25.3% (cefpodoxime) by EUCAST breakpoints. Susceptibility to macrolides was seen in 20.3% of isolates irrespective of breakpoints used (Tables 4–6 and Figures 1–3).

S. pneumoniae susceptibility in Kenya

All 44 pneumococcal isolates collected in Kenya were susceptible to the fluoroquinolones levofloxacin and moxifloxacin by all breakpoints. Susceptibility of isolates to ceftriaxone was high (95.5%–100%) by CLSI, EUCAST high-dose and PK/PD breakpoints, but this was reduced to 68.2% when using EUCAST low-dose criteria. The majority of isolates (90.9%) were susceptible to penicillin by CLSI IV and EUCAST high-dose breakpoints, but by CLSI oral or EUCAST low-dose IV breakpoints susceptibility reduced to 20.5%. Susceptibility to amoxicillin and amoxicillin/clavulanic acid was 90.9%–93.2% based on CLSI and PK/PD breakpoints, but 72.7% and 86.4%, respectively, by EUCAST high-dose and 54.6% for both antibiotics by EUCAST low-dose criteria. Susceptibility of isolates to cefpodoxime, cefuroxime (oral) and cefdinir ranged from 34.1% to 54.6% by all available breakpoints (EUCAST does not publish breakpoints for cefdinir). Susceptibility to cefaclor was more variable, at 0% by EUCAST, 4.6% by PK/PD and 13.6% by CLSI breakpoints. By PK/PD breakpoints, susceptibility to cefixime was 18.2% (EUCAST and CLSI do not publish breakpoints for cefixime). Susceptibility to macrolides was seen in 61.4%–65.9% of isolates by all criteria. The least active antibiotic was trimethoprim/sulfamethoxazole, with 6.8%–9.1% of isolates susceptible by all breakpoints (Tables 7–9 and Figures 1–3).

S. pneumoniae susceptibility in Morocco

All 19 pneumococcal isolates from Morocco were susceptible to ceftriaxone by CLSI, EUCAST high-dose and PK/PD breakpoints and 89.5% were susceptible by EUCAST low-dose criteria. High susceptibility (94.7%) was observed to penicillin using CLSI IV or EUCAST high-dose breakpoints and 89.5% of isolates were susceptible by CLSI oral and EUCAST low-dose IV criteria. Susceptibility to the other antibiotics tested using CLSI breakpoints was relatively high, ranging from 68.4% for trimethoprim/sulfamethoxazole to 94.7% for amoxicillin, amoxicillin/clavulanic acid and moxifloxacin. Similar susceptibility was observed with PK/PD and EUCAST breakpoints except for cefaclor, where susceptibility was 89.5% by CLSI and 52.6% by PK/PD criteria, but no isolates were susceptible by EUCAST breakpoints (Tables 10–12 and Figures 1–3).

Comparative susceptibility of S. pneumoniae by country

Using CLSI breakpoints, S. pneumoniae isolates from Morocco were more susceptible to the antibiotics tested than isolates from Tunisia, except for levofloxacin, where susceptibility was significantly lower (P < 0.05), and moxifloxacin, ceftriaxone and trimethoprim/sulfamethoxazole, where susceptibility was no different statistically (Figure 1). Higher susceptibility of isolates from Morocco compared with Kenya was statistically significant (P < 0.05) for all tested cephalosporins except ceftriaxone. No significant difference in susceptibility between isolates from Morocco and Kenya was found for amoxicillin, amoxicillin/clavulanic acid, macrolides and fluoroquinolones. However, when comparing the susceptibility of pneumococci in Morocco with those from Tunisia, susceptibility was significantly higher in Morocco for amoxicillin, amoxicillin/clavulanic acid and macrolides. In contrast, isolates from Morocco were significantly less susceptible to levofloxacin than isolates from Tunisia.

Isolates from Tunisia showed statistically lower susceptibility (P < 0.05) than isolates from Kenya to amoxicillin, amoxicillin/clavulanic acid, macrolides, cefdinir, cefpodoxime and cefuroxime (oral), whereas susceptibility to trimethoprim/sulfamethoxazole was significantly higher in isolates from Tunisia than Kenya. Susceptibility to cefaclor, ceftriaxone, fluoroquinolones and oral penicillin was statistically the same in Tunisia and Kenya. Histograms indicating susceptibility by EUCAST and PK/PD breakpoints are shown in Figures 2 and 3, respectively.

Comparative susceptibility of S. pneumoniae by penicillin susceptibility

Among the 142 S. pneumoniae isolates combined from the three African countries, 37 (26.1%) were penicillin susceptible (PSSP), 54 (38%) were penicillin intermediate (PISP) and 51 (35.9%) were penicillin resistant (PRSP) according to CLSI oral breakpoints. Among PSSP isolates, all were susceptible to amoxicillin, amoxicillin/clavulanic acid, cefdinir, cefpodoxime, ceftriaxone, cefuroxime (oral) and moxifloxacin, 97.3% to levofloxacin, 89.2% to azithromycin, clarithromycin and cefaclor, 86.5% to erythromycin, and 64.9% to trimethoprim/sulfamethoxazole (using CLSI breakpoints) (Figure 4). PSSP isolates were significantly more susceptible than PRSP to all antibiotics tested (P < 0.05) except fluoroquinolones. This was because virtually all isolates were quinolone susceptible, irrespective of penicillin susceptibility. PSSP isolates were also significantly more susceptible than PISP to macrolides and all tested cephalosporins except ceftriaxone. In addition, PISP isolates were significantly more susceptible than PRSP to all antibiotics except cefaclor, the fluoroquinolones and trimethoprim/sulfamethoxazole.

Figure 4.

Percentage susceptibility rates (with 95% CI) based on CLSI breakpoints for antibiotics against PSSP, PISP and PRSP, combining results from Tunisia, Kenya and Morocco. Penicillin susceptibility categories are based on oral penicillin CLSI breakpoints. ^aSusceptibility was significantly higher among PSSP and PISP isolates than PRSP isolates (P < 0.05). ^bSusceptibility was significantly higher among PSSP than PISP isolates and among PISP than PRSP isolates (P < 0.05). ^cSusceptibility was significantly higher among PSSP isolates than among PISP and PRSP isolates (P < 0.05). ^dSusceptibility was significantly higher among PSSP isolates than among PRSP isolates (P < 0.05).

H. influenzae isolates

A total of 78 H. influenzae isolates, including 74 from Tunisia and 4 from Morocco, were collected during 2015–18. No H. influenzae isolates were collected from Kenya.

Most isolates came from sputum (n = 64; 82.1%) and transtracheal aspirate (n = 11; 14.1%). Less frequently, isolates were from bronchoalveolar lavage (n = 2; 2.6%) and middle ear effusion (n = 1; 1.3%). Approximately half of the isolates (n = 36; 46.2%) came from paediatric patients (≤12 years) while isolates from adults (aged 13–64 years) represented 37.2% (n = 29), with the remaining 16.7% (n = 13) from elderly patients (aged ≥65 years). Only H. influenzae isolates from Tunisia were assessed further.

Summary MIC and susceptibility data for all 74 H. influenzae isolates from Tunisia are shown in Tables 13–15 and Figures 5–7. MIC distribution data are shown in Table S4.

Figure 5.

Percentage antibiotic susceptibility rates of H. influenzae isolates (n = 74) from Tunisia based on CLSI breakpoints.

Figure 6.

Percentage antibiotic susceptibility rates of H. influenzae isolates (n = 74) from Tunisia based on EUCAST (dose-specific) breakpoints.

Figure 7.

Percentage antibiotic susceptibility rates of H. influenzae isolates (n = 74) from Tunisia based on PK/PD breakpoints. Low-dose amoxicillin/clavulanic acid = 1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children. High-dose amoxicillin/clavulanic acid = 4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children.

Table 13.

MIC and susceptibility data for H. influenzae isolates (n = 74) from Tunisia using CLSI breakpoints

Antibiotic	MIC (mg/L)				CLSI susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin	1	32	≤0.12	>128	–	–	–
Amoxicillin/clavulanic acid	0.5	2	≤0.06	4	100	–	0.0
Ampicillin	0.25	32	≤0.06	64	64.9	2.7	32.4
Cefaclor	4	16	0.5	32	89.2	8.1	2.7
Cefdinir	0.25	0.5	0.12	2	98.7	–	–
Cefditoren	≤0.03	0.06	≤0.03	0.25	–	–	–
Cefixime	0.03	0.25	≤0.015	4	98.7	–	–
Cefpodoxime	0.06	0.12	≤0.015	1	100	–	–
Ceftriaxone	≤0.015	≤0.015	≤0.015	2	100	–	–
Cefuroxime	0.5	1	0.06	2	100	0.0	0.0
Azithromycin	0.5	1	0.12	2	100	–	–
Clarithromycin	4	4	0.12	8	100	0.0	0.0
Levofloxacin	0.03	0.03	≤0.004	>8	96.0	–	–
Moxifloxacin	0.03	0.06	0.008	>8	96.0	–	–
Trimethoprim/sulfamethoxazole	0.5	16	≤0.015	>16	51.4	6.8	41.9

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 14.

MIC and susceptibility data for H. influenzae isolates (n = 74) from Tunisia using EUCAST (dose-specific) breakpoints

Antibiotic	MIC (mg/L)				EUCAST susceptibility
	50%	90%	min	max	%S	%I	%R
Amoxicillin (0.75–1 g × 3 oral)	1	32	≤0.12	>128	64.9	–	35.1
Amoxicillin/clavulanic acid (0.875 g/0.125 g × 3 oral)	0.5	2	≤0.06	4	93.2	–	6.8
Ampicillin (2 g × 3 IV)	0.25	32	≤0.06	64	64.9	–	35.1
Ampicillin (2 g × 4 IV)	0.25	32	≤0.06	64	64.9	–	35.1
Cefaclor	4	16	0.5	32	–	–	–
Cefdinir	0.25	0.5	0.12	2	–	–	–
Cefditoren	≤0.03	0.06	≤0.03	0.25	–	–	–
Cefixime	0.03	0.25	≤0.015	4	86.5	–	13.5
Cefpodoxime	0.06	0.12	≤0.015	1	91.9	–	8.1
Ceftriaxone (1 g × 1 IV)	≤0.015	≤0.015	≤0.015	2	96.0	–	4.1
Ceftriaxone (2 g × 2 IV)	≤0.015	≤0.015	≤0.015	2	96.0	–	4.1
Cefuroxime	0.5	1	0.06	2	4.1	89.2	6.8
Azithromycin	0.5	1	0.12	2	–	–	–
Clarithromycin	4	4	0.12	8	–	–	–
Levofloxacin (0.5 g × 2 oral or 0.4 g × 2 IV)	0.03	0.03	≤0.004	>8	91.9	–	8.1
Levofloxacin (0.75 g × 2 oral or 0.4 g × 3 IV)	0.03	0.03	≤0.004	>8	91.9	–	8.1
Moxifloxacin	0.03	0.06	0.008	>8	93.2	–	6.8
Trimethoprim/sulfamethoxazole (0.16 g/0.8 g × 2 oral or IV)	0.5	16	≤0.015	>16	51.4	5.4	43.2
Trimethoprim/sulfamethoxazole (0.24 g/1.2 g × 2 oral or IV)	0.5	16	≤0.015	>16	56.8	–	43.2

min, minimum; max, maximum; S, susceptible; I, intermediate; R, resistant; –, not applicable.

Table 15.

MIC and susceptibility data for H. influenzae isolates (n = 74) from Tunisia using PK/PD breakpoints

Antibiotic	MIC (mg/L)				PK/PD susceptibility
	50%	90%	min	max	%S
Amoxicillin (1.5 g/day)	1	32	≤0.12	>128	64.9
Amoxicillin (4 g/day)	1	32	≤0.12	>128	70.3
Amoxicillin/clavulanic acid (1.75 g/0.25 g/day adults; 45 mg/6.4 mg/kg/day children)	0.5	2	≤0.06	4	93.2
Amoxicillin/clavulanic acid (4 g/0.25 g/day adults; 90 mg/6.4 mg/kg/day children)	0.5	2	≤0.06	4	100
Ampicillin	0.25	32	≤0.06	64	–
Cefaclor	4	16	0.5	32	1.4
Cefdinir	0.25	0.5	0.12	2	66.2
Cefditoren	≤0.03	0.06	≤0.03	0.25	–
Cefixime	0.03	0.25	≤0.015	4	98.7
Cefpodoxime	0.06	0.12	≤0.015	1	97.3
Ceftriaxone	≤0.015	≤0.015	≤0.015	2	98.7
Cefuroxime	0.5	1	0.06	2	93.2
Azithromycin	0.5	1	0.12	2	2.7
Clarithromycin	4	4	0.12	8	1.4
Levofloxacin	0.03	0.03	≤0.004	>8	96.0
Moxifloxacin	0.03	0.06	0.008	>8	96.0
Trimethoprim/sulfamethoxazole	0.5	16	≤0.015	>16	51.4

min, minimum; max, maximum; S, susceptible; –, not applicable.

H. influenzae susceptibility in Tunisia

Of the 74 H. influenzae isolates from Tunisia, 33.8% (n = 25) were β-lactamase positive. One isolate was β-lactamase negative and ampicillin resistant (BLNAR) by EUCAST breakpoints (ampicillin MIC ≥2 mg/L) and CLSI breakpoints (ampicillin MIC ≥4 mg/L). The H. influenzae isolates were ≥91.9% susceptible to amoxicillin/clavulanic acid, cefpodoxime, ceftriaxone and fluoroquinolones by all breakpoints. Susceptibility to cefixime was also high by CLSI and PK/PD breakpoints (98.7%) but was reduced when using EUCAST breakpoints (86.5%). A greater contrast between breakpoint criteria was observed for cefuroxime (oral), where only 4.1% of isolates were susceptible based on EUCAST breakpoints compared with 93.2%–100% susceptible by PK/PD or CLSI breakpoints. Similarly, high susceptibility to cefaclor and cefdinir was seen using CLSI breakpoints (89.2% and 98.7%, respectively), but only 1.4% and 66.2% of isolates were susceptible, respectively, using PK/PD breakpoints. More markedly, all isolates were susceptible to the macrolides azithromycin and clarithromycin by CLSI criteria compared with 1.4%–2.7% by PK/PD breakpoints. In keeping with the PK/PD data, EUCAST does not provide breakpoints for cefaclor or macrolides. Susceptibility to amoxicillin and ampicillin was 64.9%–70.3% by all available breakpoints (no breakpoint is given for amoxicillin by CLSI or ampicillin by PK/PD). The lowest susceptibility observed was to trimethoprim/sulfamethoxazole: from 51.4% to 56.8% by all breakpoints (Tables 13–15 and Figures 5–7).

Discussion

SOAR is an ongoing global surveillance study that originated in 2002, focusing on the two main CA-RTI pathogens, S. pneumoniae and H. influenzae. The data presented here are an analysis of the antibiotic susceptibility of S. pneumoniae isolates collected from Tunisia, Kenya and Morocco and H. influenzae from Tunisia between 2015 and 2018.

Using EUCAST low-dose IV or CLSI oral breakpoints, susceptibility of S. pneumoniae isolates to penicillin was low in Tunisia (13.9%) and Kenya (20.5%) but higher in Morocco (89.5%). However, by CLSI IV or EUCAST high-dose breakpoints, 90.9%–94.7% of isolates from Tunisia, Kenya and Morocco were penicillin susceptible.

Overall, isolates of S. pneumoniae from Kenya were fully susceptible by all breakpoints to the fluoroquinolones, ≥95.5% susceptible to ceftriaxone, ≥72.7% susceptible to amoxicillin and ≥86.4% susceptible to amoxicillin/clavulanic acid (except by EUCAST low-dose breakpoints: 68.2%, 54.6% and 54.6%, respectively). The other antibiotics tested were less active, with isolate susceptibility ranging from 6.8% (trimethoprim/sulfamethoxazole by EUCAST low-dose breakpoints, CLSI and PK/PD) to 79.5% (ampicillin by EUCAST high-dose breakpoints). By contrast, susceptibility of isolates from Morocco was much higher, ranging from 89.5% to 100%, to penicillins, amoxicillin/clavulanic acid, ceftriaxone, cefpodoxime, cefuroxime (oral) and the fluoroquinolones by all available breakpoints. However, it should be noted that the sample size of isolates from Morocco was small (n = 19), a limitation of the study that may affect the results and conclusions presented here. The S. pneumoniae isolates from Tunisia were the least susceptible, with ≥91.1% susceptible only to ceftriaxone and the fluoroquinolones by CLSI, EUCAST high-dose or PK/PD breakpoints, and penicillin by CLSI IV and EUCAST high-dose breakpoints.

The SOAR study has previously investigated isolates from Kenya in 2011–14,¹¹ where susceptibility was very similar to the data presented here. The 2011–14 SOAR study included the Democratic Republic of Congo, Ivory Coast and the Republic of Senegal, which also demonstrated high variability in antibiotic susceptibility in Africa.¹¹ Surveillance data from Kenya are rare, but a study looking at pneumococcal carriage and susceptibility in children from Kenya in 2009–10 was published in 2017.¹² As in the current study, susceptibility to fluoroquinolones and ceftriaxone was high and susceptibility to penicillin was low, but the isolates differed in susceptibility to erythromycin (98.7%¹² versus 63.6%). This difference demonstrates a reduction in macrolide susceptibility over time, which is supported by SOAR Kenya 2007–09 data showing 85.4% susceptibility to erythromycin.¹³

The most recent SOAR study conducted in Morocco was in 2007–09.¹³ As would be expected from data in the current study, susceptibility to most antibiotics was high, but interestingly susceptibility to azithromycin was lower than in the current study (66.2% versus 84.2% by CLSI) and was also lower for fluoroquinolones (ofloxacin 44.9% versus levofloxacin 89.5% of isolates susceptible by CLSI).¹³ However, any comparisons should be qualified by the low number of isolates from Morocco included in the current study. SOAR surveillance was also conducted in Morocco in 2004–06, where susceptibility to azithromycin was 86.3% and that to ofloxacin was 77.5%.¹⁴ Also of note is another study of pneumococci from Morocco collected in 2007–09, which showed lower susceptibility to erythromycin (38%), penicillin (61.3%) and trimethoprim/sulfamethoxazole (42.7%).¹⁵

Of the three African countries studied here, Tunisia had the lowest antibiotic susceptibility among S. pneumoniae isolates, although susceptibility to the fluoroquinolones, high-dose penicillin and high-dose ceftriaxone was >90% by all available breakpoints. When analysed statistically, isolates from Tunisia had significantly lower susceptibility than isolates from Kenya or Morocco to amoxicillin, amoxicillin/clavulanic acid, macrolides, oral penicillin, cefdinir, cefpodoxime and cefuroxime (oral). Two publications focusing on pneumococci from immunocompromised patients in Tunisia from 2005–11 also showed low antibiotic susceptibility in this country.^16,17 Similarly, another analysis of S. pneumoniae in Tunisia between 2012 and 2016 indicated low antibiotic susceptibility.¹⁸ Interestingly, there was some evidence of fluoroquinolone resistance [two isolates with levofloxacin MIC ≥32 mg/L (0.7%)] in the pneumococci,¹⁸ unlike the data presented here. There is clearly a resistance issue in Tunisia, which, according to one study, most probably relates to the 38% increase in antibiotic use in Tunisia between 2005 and 2013.¹⁹

For H. influenzae, only isolates from Tunisia were analysed. Approximately a third of isolates were β-lactamase positive, accounting for most ampicillin resistance, and one isolate was BLNAR. In the SOAR 2004–06 study, 21.1% of H. influenzae from Tunisia were β-lactamase positive,²⁰ indicating an ∼13% increase in the last 10 years or so. In the current study, susceptibility to the antibiotics tested was generally high, with reduced activity only consistently observed by all available breakpoints to trimethoprim/sulfamethoxazole (51.4%–56.8%) and ampicillin (64.9%). There has been a reduction in susceptibility to some antibiotics since the SOAR 2004–06 study versus the current study: ampicillin (78.9% versus 64.9%), cefaclor (98.6% versus 89.2%) and trimethoprim/sulfamethoxazole (84.5% versus 51.4%).²⁰

A part of the SOAR study is to evaluate the effect of different national guidelines and breakpoints. Of particular interest are the lower breakpoints for S. pneumoniae published by EUCAST compared with CLSI breakpoints. For Morocco, where the isolates had lower MICs, the effect was minimal for all antibiotics, except cefaclor (89.5% of isolates were susceptible by CLSI criteria compared with 0% by EUCAST breakpoints). In contrast, for isolates from Kenya and Tunisia, where MICs were higher, there was no marked difference in susceptibility to cefaclor, despite the breakpoint difference, because susceptibility ranged from 0% to 17.7% by all breakpoints. In Kenya susceptibility of isolates using EUCAST low-dose breakpoints was 68.2% for ceftriaxone and 54.6% for amoxicillin and amoxicillin/clavulanic acid compared with 90.9%–95.5% using CLSI breakpoints. Using EUCAST high-dose breakpoints, susceptibility increased to 100%, 72.7% and 86.4% for ceftriaxone, amoxicillin and amoxicillin/clavulanic acid, respectively. In addition, isolate susceptibility to amoxicillin and amoxicillin/clavulanic acid increased to 93.2% using PK/PD high-dose breakpoints. Similarly, in Tunisia, susceptibility of isolates to ceftriaxone by CLSI breakpoints was 91.1%, but 46.8% by EUCAST low-dose breakpoints, increasing to 100% by EUCAST high-dose breakpoints. Susceptibility to amoxicillin and amoxicillin/clavulanic acid in Tunisia was ∼70% by CLSI breakpoints, but only 31.6%–35.4% by EUCAST high-dose breakpoints. By PK/PD high-dose breakpoints, 84.8% of isolates were susceptible to amoxicillin and amoxicillin/clavulanic acid.

For H. influenzae, only isolates from Tunisia were assessed in this study and although susceptibility was generally high there was a distinct difference in susceptibility to cefuroxime (oral) by EUCAST criteria (4.1%) versus 93.2%–100% by CLSI or PK/PD breakpoints. Similarly, EUCAST does not publish H. influenzae breakpoints for macrolides or cefaclor, which is consistent with 1.4%–2.7% susceptibility observed by PK/PD breakpoints, but not comparable with 89.2%–100% susceptibility using CLSI breakpoints.

Following discussion and collaboration with EUCAST, this study also employed, for the first time in the SOAR investigations, different EUCAST breakpoints for low and high doses of several antibiotics tested so that the effect of different dosages on the susceptibility of the CA-RTI pathogens could be observed. Correct antibiotic dosing remains a challenge for the clinician, particularly since PK/PD parameters may alter during serious illness.²¹ Personalized antibiotic treatment may be an option, currently only considered for patients in intensive care.²² The ability to now assess pathogen susceptibility at different antibiotic doses along with advances in diagnostics and monitoring could allow possible progress in this area and the subsequent benefit for a wider range of patient groups.

In summary, we have assessed antibiotic susceptibility in CA-RTIs in three countries in Africa. Low antibiotic susceptibility was observed in S. pneumoniae from Tunisia, but susceptibility was higher in isolates from Kenya and highest in those from Morocco. H. influenzae isolates from Tunisia had high susceptibility to the antibiotics tested except for ampicillin and trimethoprim/sulfamethoxazole by all available breakpoints, macrolides and cefaclor by PK/PD breakpoints and cefuroxime (oral) by EUCAST breakpoints. The SOAR studies in Africa provide susceptibility data in countries where there is a paucity of such information; continued surveillance of antibiotic susceptibility in Africa is required.

Acknowledgements

We thank the participating laboratories for supplying the isolates; Dr Federica Monti (IHMA) and Dr Cedric Charrier (IHMA) for editorial assistance; Dr Sibylle Lob (IHMA) for the statistical analysis; Livewire Editorial Communications for editorial support; Christopher Were (GSK-Kenya), Hiram Mwangi (GSK-Kenya), Nawal Alila (GSK-Morocco and Tunisia), Soukaina Ragragui (GSK-Morocco and Tunisia) for their help in facilitating SOAR studies in each country; Karen Langfeld (GSK) for reviewing the manuscript; and the SOAR Team (Subhashri Kundu, Cristiana Beltrame, Rendani Manenzhe, Nergis Keles) for their continuous support of SOAR.

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 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. All other authors: none to declare. Editorial assistance was provided by Livewire Editorial Communications.