Background

Staphylococcus aureus is a well-recognized human pathogen that is implicated in a wide array of superficial, invasive and toxigenic infections [1]. Meta-analyses of published studies have provided evidence that S. aureus nasal carriage is an important risk factor for subsequent infection among patients with surgical site infections and atopic dermatitis [2, 3]. Other high-risk groups include patients colonized with methicillin-resistant Staphylococcus aureus (MRSA) undergoing dialysis, and patients admitted in the intensive care unit [4, 5]. Consequently, infection prevention strategies such as nasal decolonization are employed to minimize the occurrence of staphylococcal infection and reduce the risk of transmission in healthcare settings [6, 7]. Mupirocin (2%) nasal ointment alone or in combination with 4% chlorhexidine (CHG) based body wash is considered as the main decolonization strategy for S. aureus carriage [8, 9]. Mupirocin is a naturally occurring antibiotic produced by Pseudomonas fluorescens that interferes with protein synthesis by competitive inhibition of the bacterial isoleucyl-tRNA synthetase (IRS) [10, 11]. It gained prominence in the mid-1990s for the eradication of S. aureus nasal carriage due to its effectiveness, safety and cost [12].

Mupirocin-resistant (mupR) S. aureus was first reported in the United Kingdom in 1987 [13]. Since then, it has been reported in several countries worldwide [14–17]. The emergence of mupR S. aureus has been associated with unrestricted policies and use of mupirocin for long periods in health care settings [8, 18]. Decolonization failure in patients with S. aureus carriage is associated with high-level mupirocin resistance (HmupR - minimum inhibitory concentration [MIC]: ≥512 μg/ml), while that of low-level mupirocin resistance (LmupR – MIC: 8-64 μg/ml) is still unclear [7, 19]. LmupR is mediated through point mutation (largely V588F and V631F) in the native isoleucyl-tRNA synthetase (ileS) gene [20]. In contrast, HmupR is mainly attributed to the acquisition of plasmids with the mupA (or ileS2) gene encoding an additional IRS with no affinity for mupirocin [11, 21]. Another determinant for HmupR is the acquisition of a plasmid-mediated mupB gene [22].

There is no data summarizing reports on screening, prevalence, characterization, and geographic spread of mupR S. aureus in Africa. This systematic review evaluated published articles that assessed for mupirocin resistance in African S. aureus isolates. The findings from this systematic review highlight the need to develop an early warning system, including harmonized strategies for the prompt screening and identification of mupR S. aureus in Africa.

Methods

Literature search strategy

The relevant English articles from human and animal investigations were retrieved by three authors (YA, SA, and AS) from five electronic databases (EBSCOhost, Google Scholar, ISI Web of Science, MEDLINE, and Scopus). The search terms for each database are reported in Table 1. The literature search was concluded on 3 August 2016.

Table 1 Keywords used to identify eligible studies available in five biomedical databases
Full size table

Eligible article identification

The identification of the eligible articles was conducted according to the guidelines for preferred reporting items for systematic reviews and meta-analyses (PRISMA) [23]. We defined an eligible article as a peer-reviewed publication that (i) included mupirocin in the antibiotic susceptibility testing of S. aureus isolates, and (ii) employed phenotypic ((disc diffusion, E-test, minimum inhibitory concentration (MIC), VITEK and other automated methods)), and/or molecular ((conventional or real-time polymerase chain reaction (PCR)) techniques. International multicentre studies that included African countries were also eligible for inclusion.

Data extraction and analysis

The relevant data were extracted from each of the eligible articles included in this systematic review. A study that analysed S. aureus isolates from another investigation but answered a different research question were both considered as one study (Table 2). We performed three levels of analysis (Fig. 1). First, to understand the characteristics and geographic spread of mupR S. aureus in Africa, studies that included mupirocin in the antibiotic susceptibility testing and employed phenotypic and/or molecular techniques were identified. Secondly, the prevalence of S. aureus with the mupA gene, isolates that expressed LmupR and HmupR, and mupR-MRSA in Africa were derived from each eligible study as follows:

$$ MupA\hbox{-} \mathrm{positive}\ S.\kern0.5em aureus=\frac{\mathrm{Number}\ \mathrm{of}\ MupA\hbox{-} \mathrm{positive}\ S.\kern0.5em aureus\ \mathrm{isolates}}{\mathrm{Total}\ \mathrm{number}\ \mathrm{of}\ \mathrm{isolates}\ \mathrm{screened}\ \mathrm{with}\ \mathrm{mupirocin}} $$
$$ S.\kern0.5em aureus\ \mathrm{that}\ \mathrm{expressed}\ \mathrm{LmupR}=\frac{\mathrm{Number}\ \mathrm{of}\ S.\kern0.5em aureus\ \mathrm{isolates}\ \mathrm{with}\ \mathrm{LmupR}}{\mathrm{Total}\ \mathrm{number}\ \mathrm{of}\ \mathrm{isolates}\ \mathrm{screened}\ \mathrm{with}\ \mathrm{mupirocin}} $$
$$ S.\kern0.5em aureus\ \mathrm{that}\ \mathrm{expressed}\ \mathrm{HmupR}=\frac{\mathrm{Number}\ \mathrm{of}\ S.\kern0.5em aureus\ \mathrm{isolates}\ \mathrm{with}\ \mathrm{HmupR}}{\mathrm{Total}\ \mathrm{number}\ \mathrm{of}\ \mathrm{isolates}\ \mathrm{screened}\ \mathrm{with}\ \mathrm{mupirocin}} $$
$$ \mathrm{MupR}\hbox{-} \mathrm{MRSA}=\frac{\mathrm{Number}\ \mathrm{of}\ \mathrm{mupR}\hbox{-} \mathrm{MRSA}\ \mathrm{isolates}}{\mathrm{Total}\ \mathrm{number}\ \mathrm{of}\ \mathrm{isolates}\ \mathrm{screened}\ \mathrm{with}\ \mathrm{mupirocin}} $$
Table 2 Characteristics of the 43 eligible studies on screening for mupirocin resistance in Staphylococcus aureus from various sources in Africa
Full size table
Fig. 1
figure 1

The Preferred Reporting Items for Systematic Review and Meta-analysis flow diagram

Full size image

Thirdly, to estimate the prevalence of mupR S. aureus in humans, studies that employed at least one of the screening methods with defined breakpoint for mupirocin resistance were included in the meta-analysis. The StatsDirect statistical software version 3.0.165 (England: StatsDirectLtd.2016) was utilized to assess the heterogeneity of the eligible studies included in the meta-analysis (Cochran Q-test) [24], and to ascertain the inconsistency across the studies (I2 statistic) [25]. The random effects model was used to determine the pooled prevalence of mupR S. aureus in Africa. The criterion for statistical significance for heterogeneity was set at alpha = 0.05. The risk of publication bias was assessed and visualized by a Funnel plot [26, 27].

Results

Eligible studies from electronic database search

We identified 43 reports (Table 1) of which 34 studies investigated only human samples. The remaining nine studies assessed samples from only animals (n = 5), human and environmental sources (n = 2), human and animal sources (n = 1), and cockroaches (n = 1). Most of the eligible studies (32/43; 74%) were obtained from EBSCOhost, ISI Web of Science, MEDLINE, and Scopus. The remaining studies (11/43; 26%) were obtained only through Google Scholar and consisted of studies conducted in Egypt [28–31], South Africa [32–34], Nigeria [35, 36], Ethiopia [37] and Kenya [38].

Screening and identification of mupR S. aureus in Africa

Only 12 of the 54 (22%) African countries reported data on screening for mupR S. aureus (Fig. 2). The first published article indicated that mupirocin had been in use in Africa, at least from the late 1980s [39]. Most of these studies (28/43; 65%) were conducted in Nigeria (10/43; 23%), Egypt (7/43; 16%), South Africa (6/43; 14%) and Tunisia (5/43; 12%) (Fig. 2). MupR S. aureus was mainly identified through the disk diffusion method (29/43; 67%). The guidelines by the Clinical and Laboratory Standards Institute (CLSI), previously known as National Committee for Clinical Laboratory Standards (NCCLS), were broadly used in Africa (Table 2). However, a number of studies [28, 29, 31, 33, 36, 40–46] utilized the disk diffusion method with CLSI guidelines that had no zone diameter breakpoint for mupirocin. Moreover, some studies [47–49] did not provide information on the year of publication of the CLSI guidelines. MupR S. aureus was reported in six African countries including South Africa [32–34, 46, 50, 51], Egypt [29–31, 52], Nigeria [36, 44, 53], Ghana [54, 55], Libya [56, 57] and Ethiopia [37] (Fig. 2; Table 2). The mupA-positive S. aureus was detected in five studies from Egypt [30, 52], South Africa [33, 50] and Nigeria [53]. LmupR and HmupR were both reported in six human studies conducted in South Africa [32, 33, 50], Egypt [30, 52] and Libya [57]. The mupR-MRSA isolates were identified in South Africa [32, 34, 50, 51], Egypt [30, 31, 52], Libya [56, 57], Ghana [55] and Nigeria [36] (Table 3). MupR-MRSA was not reported from MRSA isolates recovered from studies conducted in Egypt [28, 58], Tunisia [59, 60] and Algeria [47].

Fig. 2
figure 2

Studies on screening for mupirocin-resistant Staphylococcus aureus in Africa

Full size image
Table 3 Prevalence of mupirocin-resistant S. aureus from various sources in Africa based on phenotypic and molecular methods
Full size table

An assessment of data on mupR S. aureus at the regional level is described as follows (Fig. 3).

Fig. 3
figure 3

Geographic distribution of mupirocin-resistant (mupR) Staphylococcus aureus in Africa. Countries (in green) in which mupR S. aureus have been investigated but not reported. Countries (in red) in which mupR S. aureus have been investigated and reported

Full size image

North Africa

Seventeen eligible studies were recorded from this region, including Egypt [28–31, 40, 52, 58], Tunisia [41–43, 59, 60], Libya [56, 57, 61], Algeria [47] and Morocco [62]. MupR S. aureus was reported in six studies conducted in two North African countries: Egypt [29–31, 52] and Libya [56, 57]. PCR detection of the mupA gene was performed in only two studies conducted in Egypt [30, 52]. In addition, one of the reports identified two mupA positive MRSA that exhibited LmupR [30]. MupR S. aureus was not detected in Tunisia [41–43, 59, 60], Algeria [47], and Morocco [62].

West Africa

S. aureus resistance to mupirocin was investigated in Nigeria [35, 36, 44, 48, 49, 53, 63–66] and Ghana [54, 55, 67, 68]. Only two studies from Ghana reported on mupR S. aureus [54, 55]. In Nigeria, three studies (including two from only human sources and one from both animal and human samples, respectively) reported on S. aureus isolates that demonstrated HmupR [36, 44, 53].

Central Africa

MupR S. aureus was not detected in studies conducted in Gabon [69], and São Tomé and Príncipe [70].

East Africa

In this review, we identified four eligible studies conducted in Kenya [38, 71, 72] and Ethiopia [37]. A report on the role of cockroaches as potential vectors of foodborne pathogens in Ethiopia identified 17 mupR S. aureus isolates [37]. All the S. aureus isolates (one animal and two human studies) from Kenya were susceptible to mupirocin [38, 71, 72].

Southern Africa

The six studies reported in this geographical area were from South Africa and consisted of two single centre studies [34, 46] and four multicenter studies [32, 33, 50, 51]. MupR S. aureus was identified in all the reports, while mupA-positive S. aureus isolates were noted in only two studies [33, 50].

Prevalence of mupR S. aureus in Africa

The random-effects pooled prevalence of mupR S. aureus in Africa is 14% (95% CI =6.8 to 23.2%). This was calculated based on 11 heterogeneous human studies (Figs. 4 and 5) conducted in South Africa [32, 33, 50, 51], Ghana [54, 55], Egypt [30, 52], Libya [56, 57] and Nigeria [53]. In Africa, the proportion of S. aureus isolates with the mupA gene, and those that expressed LmupR and HmupR ranged between 0.5 and 8%, 4 and 47%, 0.5 and 38%, respectively. The frequency of mupR-MRSA isolates ranged between 5 and 50% (Table 3).

Fig. 4
figure 4

Bias assessment (Funnel) plot for studies assessing rates of mupirocin-resistant Staphylococcus aureus in Africa. Random effects (DerSimonian-Laird). Pooled proportion = 0.139303 (95% CI = 0.067511 to 0.23165). Bias indicators, Begg-Mazumdar: Kendall’s tau = 0.2 P = 0.4454, Egger: bias = 4.771137 (95% CI = −2.517874 to 12.060148) P = 0.1728, Harbord: bias = 2.014783 (92.5% CI = −5.90181 to 9.931377) P = 0.6208

Full size image
Fig. 5
figure 5

Pooled estimate of proportions (human studies) for mupirocin-resistant Staphylococcus aureus in Africa

Full size image

Association of MupR S. aureus with mupirocin use in Africa

There is no data on the use of mupirocin as an agent for S. aureus decolonization and its association with mupR S. aureus in Africa.

MupR S. aureus and biofilm production

A report from Egypt noted that mupR-MRSA were moderate to strong biofilm producers [52].

MupR S. aureus and co-resistance to other antibiotics

In this systematic review, two studies (conducted in Egypt and South Africa) showed that mupR S. aureus was associated with multi-drug resistance [30, 33].

Molecular characterization of mupR S. aureus in Africa

Only three studies provided molecular data on mupR S. aureus in Africa [45, 54, 55]. A report provided evidence of a 35 kb (non-conjugative) and 41.1 kb (conjugative) plasmid encoding mupA in S. aureus isolates from Nigeria and South Africa [45]. It also described an MRSA clone that demonstrated LmupR in South Africa. LmupR was also identified among MRSA isolates assigned with ST36, ST88, and ST789 in Ghana [55]. A cross-sectional S. aureus study identified a methicillin susceptible S. aureus (MSSA) strain with HmupR from a 51-year-old hospital staff in Ghana [54]. Molecular characterization indicated that the strain (spa type t4805) was PVL-positive.

Discussion

This is the first systematic review on mupR S. aureus in Africa and clearly showed the paucity of data on the continent. Nevertheless, this study indicated a high prevalence ((14% (95% CI =6.8 to 23.2)) of mupR S. aureus in Africa. These observations support the need for mupR S. aureus surveillance data to provide information on its epidemiology and clinical significance in Africa. It is noteworthy that Google Scholar was valuable in the identification of several eligible studies [28–38]. We observed that 26% (11/43) of the eligible studies were identified from African journals which were not indexed in commonly used electronic databases. Google Scholar has been considered as a useful supplement with other electronic databases for systematic review search [73] including recent meta-analyses of published studies on S. aureus in Africa [74, 75].

The phenotypic methods for the screening and identification of mupR S. aureus include disc diffusion (two-disc strategy: 5 μg and 200 μg), agar dilution, broth micro-dilution and E-test [19]. In this study, the disk diffusion method and the CLSI (formerly NCCLS) guidelines were strategies mainly applied to detect mupR S. aureus in Africa. However, we observed certain inconsistencies [28, 29, 31, 33, 36, 40–49]. For instance, a number of studies [28, 29, 31, 33, 36, 40–42, 44–46] applied the disk diffusion method with the CLSI guidelines that had no breakpoint values for mupirocin. The 2017 CLSI guidelines recommend the use of the 200 μg disk to differentiate between HmupR and the absence of HmupR (i.e. no zone = HmupR; any zone = absence of HmupR) [76]. The 200 μg disk with a different breakpoint (Susceptible ≥30 mm, Resistance < 18 mm) is also endorsed for the differentiation between HmupR and the absence of HmupR in the latest versions (accessed 28th May, 2018) of the European Committee for Antimicrobial Susceptibility Testing (EUCAST) and Comité de l’antibiogramme de la Société Française de Microbiologie (CA-SFM) [77, 78]. The breakpoint values for the detection of LmupR and differentiation from HmupR are not provided in these documents (CA-SFM, CLSI, and EUCAST). Despite this limitation, the disk diffusion method in conjunction with any of these guidelines could at least be valuable for the preliminary screening and identification of HmupR S. aureus in Africa. MRSA decolonization failure is of clinical significance as it is often attributed to persistence or re-colonization associated with isolates exhibiting HmupR, while that of LmupR is not clear [7, 19, 79]. In this review, the prevalence of S. aureus that exhibited LmupR, HmupR and mupR-MRSA in Africa was predicated on a range of methods using different guidelines. We suggest that surveillance data from Africa is established on harmonized guidelines to enhance quality assurance and comparison at the continental and global level.

We noted a prevalence of mupR-MRSA ranging between 5 and 50% in Africa (Table 3). This is of serious concern. Specifically, the relationship between mupirocin resistance and MRSA has important consequences on infection control measures and effectiveness of decolonization strategies [8]. MupR-MRSA could limit the choices available for the control and prevention of healthcare-associated MRSA infections (7, 8). Therefore, surveillance studies are important to investigate the emergence and spread of mupirocin resistance in hospital settings in Africa. This is important among patients at high risk of MRSA infections, including patients in the dermatology, dialysis and the Intensive Care Units. In addition, there is the need for more data on the molecular characterization of mupR S. aureus in Africa [45, 54, 55]. For instance, whole genome sequencing (WGS) will assist in understanding the transmission dynamics of mupR S. aureus in Africa. Moreover, WGS data will allow comprehensive investigation of the genetic basis for LmupR mutation (which is largely due to V588F and V631F in the native gene (ileS)) and mupB-positive S. aureus in Africa.

Language bias was the main limitation of this systematic review as we did not include studies published in French, Portuguese, Arabic and Spanish.

Conclusions

This study showed the need for more epidemiological data to understand the transmission, burden and risk factors associated with mupR S. aureus in Africa. In addition, there is a need for data on administration and use of mupirocin in community and hospital setting in Africa. This is important in antibiotic stewardship to mitigate the emergence and spread of mupR S. aureus in Africa. Finally, this systematic review highlighted the need for harmonized guidelines to facilitate the comparison of data on mupR S. aureus from Africa.