Antibacterial resistance in sub-Saharan Africa: an underestimated emergency

Abstract

Antibacterial resistance–associated infections are known to increase morbidity and mortality and cost of treatment and to potentially put others in the community at higher risk of infections. In high-income countries, where the burden of infectious diseases is relatively modest, resistance to first-line antibacterial agents is usually overcome by use of second- and third-line agents. However, in developing countries where the burden of infectious diseases is high, patients with antibacterial-resistant infections may be unable to obtain or afford effective second-line treatments. In sub-Saharan Africa (SSA), the situation is aggravated by poor hygiene, unreliable water supplies, civil conflicts, and increasing numbers of immunocompromised people, such as those with HIV, which facilitate both the evolution of resistant pathogens and their rapid spread in the community. Due to limited capacity for disease detection and surveillance, the burden of illnesses due to treatable bacterial infections, their specific etiologies, and the awareness of antibacterial resistance is less well established in most of SSA, and therefore the ability to mitigate their consequences is significantly limited.

Antibacterial resistance among some pathogens of critical importance in sub-Saharan Africa

This review on antibacterial resistance in sub-Saharan Africa (SSA) focuses on available published literature on the subject dating from 1990 to the present and unpublished reports from ongoing projects looking at surveillance of antibacterial resistance in the region. The review does not include tuberculosis, HIV, or malaria, which are better addressed in other specialized reviews. To illustrate the importance of the problem of antibacterial resistance in SSA, this review will focus on the current status among some key Gram⁻ and Gram⁺ bacterial pathogens affecting communities in SSA, factors that may be responsible for driving antibacterial resistance and then attempt to chart the way forward in addressing these challenges. The review covers reports and published data from 1990 to the present.

Neisseria gonorrhoeae

Gonorrhea remains one of the most common sexually transmitted infections (STIs) in SSA countries, with the high prevalence of HIV, especially in urban centres, being a major risk factor. The emergence of resistance in Neisseria gonorrhoeae to commonly available antibacterial agents including, most recently, third-generation extended-spectrum cephalosporins, is a major obstacle in the control of infections in several countries in the region. For instance, in a recent cross-sectional study in Northern Uganda, gonococcal isolates (n = 151) showed decreased susceptibility to ampicillin, tetracycline and erythromycin, and ciprofloxacin, and intermediate resistance to chloramphenicol. Fortunately, gentamicin and cefotaxime have remained as a single-dose effective treatment for N. gonorrhoeae in this region.¹ Similarly, in a study in rural western Kenya investigating STIs in men from 2002 to 2009, penicillin resistance was found in 65% of 168 isolates obtained from 142 patients and plasmid-mediated tetracycline resistance in 97%, while 11% of the N. gonorrhoeae isolates were ciprofloxacin resistant. This study also noted that quinolone-resistant N. gonorrhoeae first appeared in 2007, increasing from 9.5% in 2007 to 50% in 2009. However, resistance was not detected to spectinomycin, cefixime, ceftriaxone, or azithromycin, but minimum inhibitory concentrations (MICs_for cefixime (P = 0.018), ceftriaxone (P < 0.001), and azithromycin (P = 0.097) also increased over the period 2002 to 2007.² In a survey in South Africa, among 209 gonococcal isolates, 54 (25.8%) penicillinase-producing N. gonorrhoeae (PPNG), and 154 (73.3%) plasmid-mediated tetracycline-resistant N. gonorrhoeae (TRNG) were detected, thus rendering the two previously commonly used antibiotics largely ineffective.³ In addition, the increasing frequency of isolation of ciprofloxacin-resistant N. gonorrhoeae across the urban centers in South Africa⁴ and the recent detection of two cases of extended-spectrum cephalosporin (ESC)–resistant N. gonorrhoeae, one of which was associated with verified cefixime treatment failure, will only exacerbate the situation in the region.^5,6

With the rapid spread of this highly resistant pathogen, it is likely that an era of untreatable gonorrhoea may be approaching, posing a major public health problem. It will be important to seek concerted efforts in the implementation of plans globally and nationally for control strategies, including enhancing surveillance of gonococcal resistance, treatment failures, and antibacterial use/misuse. Efforts aimed at improving prevention, especially of HIV transmission, early diagnosis, and treatment of gonorrhoea, will be critically important. Novel treatment strategies, new antibacterials (or other compounds), and novel vaccine strategies must be developed in order to combat this STI.

Non-typhi Salmonella (NTS)

In developed countries, non-typhi Salmonella (NTS) are a cause of foodborne illness clinically manifesting mainly as self-limiting gastroenteritis, and less than 1% of cases would progress to cause invasive disease. However, in SSA countries, NTS have emerged as important and sometimes dominant contributors to invasive bacterial disease (iNTS) in HIV-seropositive adults and young children below 5 years of age with HIV, malnutrition, and/or co-infection with malaria as major risk factors.^7–10 Indeed, iNTS are the second-most common cause of neonatal meningitis and the third-most common cause of bacterial meningitis in children over 2 months of age in Malawi and Kenya.^11–13 For instance, in single-site studies in rural Kilifi, Kenya, in 2000–2005, the local minimum incidence of community- acquired NTS was estimated at 166 per 100,000 per year for children less than 5 years of age,^14,15 while in another rural site in western Kenya from 2006–2009 among 55,000 persons, incidence was estimated at 568/100,000 person-years of observation (pyo).¹³ However, true incidence in both sites was thought to be underestimated owing to incomplete blood culturing of febrile patients, community cases that never reach the hospital, and the general insensitivity of blood culturing. At the National Referral Hospital in Nairobi, among over 1300 patients over the period 1988–1997, an incidence of 13.7% was reported among HIV-positive adult patients, compared to 3.1% among HIV-negative patients of NTS-associated invasive bacterial disease (IBD), accounting for half of the bacteremic cases,¹⁶ with mortality ranging between 18.5 and 40%. In Malawi, the overall case-fatality rate associated with NTS bacteremia was 22.3%.⁸

Comparative whole-genome analysis revealed distinct genomic signatures that suggested that Salmonella Typhimurium ST313 isolates commonly found in iNTS cases in SSA may be undergoing adaptation to a particular restricted pathogenic lifestyle, that of invasive disease in the human host^17–19 (Fig. 1). Clearly, the use of chloramphenicol as a drug of choice for treatment for suspected severe bacterial infections and cases of iNTS infection confirmed by blood culture in endemic countries such as Malawi led to the emergence of resistance around 2001–2004. The acquisition of chloramphenicol resistance may have afforded resistant clones a greater opportunity to survive treatment and transmit.¹⁸ Multidrug resistance (MDR) to commonly available antibacterials including chloramphenicol, trimethoprim/sulphamethoxazole, several beta lactams used as first-line agents and fluoroquinolones is now widespread among iNTS causing IBD in Kenya and Malawi^8,15,17,19 as well as other parts of SSA,^20–22 posing a major challenge to the treatment and management of IBD.

Figure 1.

Radial phylogram based on chromosomal single nucleotide polylmorphisms (SNPs) showing the phylogenetic relationship of S. Typhimurium isolates. Branch lengths indicate the number of SNPs, scale as indicated, separating the sequenced isolates from outside Africa and those from SSA. ST19 isolates (red circles), DT2 reference strain (blue circle) and ST313 isolates (yellow circles) are shown. Adapted from Ref. 17.

With the introduction of conjugate Hib vaccine in 2004, IBD caused by Haemophilus influenzae has reduced by over 80% in Kenya, and the introduction in 2011 of conjugate pneumococcal vaccine into the national Expanded Program on Immunization is expected to lead to a major fall in IBD due to S. pneumoniae.²³ iNTS may consequently become the principal cause of community-acquired IBD in Kenya and in other African countries as they introduce similar vaccination schedules.

Typhoid Fever

Typhoid fever caused by Salmonella enterica serovar Typhi (S. Typhi) remains an important global public health problem, causing 22 million cases of typhoid and approximately 200,000 deaths annually worldwide.²⁴ It is an imported disease in industrialized countries, since improved sanitation and water supply has eradicated endemic typhoid, but the disease remains endemic in South East/Central Asia and parts of SSA.^25–27 In Africa, the true burden of typhoid fever is largely unknown, mainly because credible measures of disease incidence inherently require confirmed diagnosis based on blood or bone marrow culture and such facilities are limited or non-existent on the continent. Typhoid incidence rates of 39/100,000^24,28 and 59/100,000²⁹ have been reported for East Africa and Egypt, respectively but these figures may again be an underestimate due to underreporting, as only very severely ill patients will usually seek treatment in hospital.

It is possible that the rapid growth of informal settlements in SSA, as more people moved into the cities looking for job opportunities and water and sanitation services became overwhelmed, led to increased rates of typhoid. In recent studies on the epidemiology of the disease, typhoid fever patients showed a markedly contrasting age distribution from that seen in iNTS, with typhoid affecting mainly school-age children and younger adults, while iNTS was more prevalent among children less than 5 years of age (although these rates differed slightly between the two study sites in Kenya)^13,30 Similarly in Malawi, 15 of 75 typhoid cases reported at the main National Referral facility were in preschool-age children, compared with only 5 of 105 of cases in a South Africa hospital.¹⁰ Again, the reasons for these differences will require further investigation.

Increase in the prevalence of MDR in S. Typhi (often associated with resistance to ampicillin, chloramphenicol, and trimetoprim/sulphamethoxazole) has been linked to greater severity of illness and higher case-fatality rates. In Kenya, multidrug-resistant S. Typhi in adults and school-age children has previously been observed in sporadic outbreaks in resource-poor settings, especially in congested urban informal settlements.^27,30 For the treatment of multidrug-resistant S. Typhi, fluoroquinolones, macrolides, and cephalosporins are the second-line treatment choices, but these antibacterials are not readily available in most of rural SSA. In addition, there are reports of S. Typhi isolates that are now resistant to nalidixic acid and show increasing fluoroquinolone resistance in a number of SSA countries where data is available. For instance, between 2000 and 2010 reports from the Democratic Republic of Congo consistently showed that the proportion of S. Typhi isolates from blood cultures were multidrug resistant and also showed decreased susceptibility to fluoroquinolones, with an increasing trend, and in 2010 prevalence of MDR isolates stood at 30.3% while 15.4% showed decreased susceptibility to fluoroquinoles.^31,32

In 2001–2002 and in 2009 in Ghana, S. Typhi was the most prevalent pathogen isolated in blood cultures, accounting for 40.7% of 100 and 31.3% of 24 Salmonella enetrica isolates obtained during the two study periods. In addition, 88.3% of the isolates were resistant to chloramphenicol and to other commonly available second-line antibacterial agents including trimethoprim/sulphamethoxazole and ampicillin.³³ In Kenya, since the first report of multidrug resistant S. Typhi outbreaks in 1998,³⁴ when the prevalence of the MDR phenotype was 50–65%, the levels of MDR S. Typhi have been rising steadily, and at present over 75% of all S. Typhi from the main referral hospital and private clinics in Nairobi are multidrug resistant. In addition, the proportion of S. Typhi that are both multidrug resistant and resistant to nalidixic acid with decreased susceptibility to fluoroquinolones has risen from 1% in 2000 to nearly 25% in 2008 (Fig. 2). Similar patterns of multidrug resistant S. Typhi have been reported in epidemics in South Africa³⁵ and Egypt,²⁹ and treatment failures with ciprofloxacin were reported in 2005 in Cameroon.²⁶ In SSA, molecular epidemiological studies have shown that multidrug resistant S. Typhi outbreaks are dominated by haplotype 58,^27,36 a clade that is commonly associated with outbreaks in Southeast Asia.

Figure 2.

Distribution of multidrug resistant S. Typhi in Kenya during the period 1988–2008. Bar heights indicate the total number of drug-susceptible and multidrug resistant isolates included in the study from each surveillance period. Black and red bars indicate the number of these isolates that were included in SNP typing. Adapted from Ref. 27.

Antibiotic-resistant diarrheagenic E. coli in community-acquired infections

Although in general in SSA, mortality due to childhood diarrheal illness has been greatly reduced as a result of the widespread use of oral rehydration treatments (ORT) at health centers, morbidity rates still remain high.³⁷ In several studies worldwide that have sought a broad range of pathogens as etiologies of diarrhea, enterotoxigenic Escherichia coli (ETEC) have repeatedly been cited as predominant causes, especially in children less than 5 years of age and in travellers to developing countries. ETEC strains are consistently associated with acute and persistent diarrhea leading to nutritional faltering and, potentially, to death.^38–40 In the recent Global Enteric Multicenter Study (GEMS)⁴¹ to identify the etiology and population-based burden of pediatric moderate-to-severe diarrhoeal disease in over 1000 children below 5 years of age in SSA and South Asia, EPEC producing heat-stable toxin (STETEC) was among the four major pathogens significantly associated with moderate-to-severe diarrhoea at all seven study sites. In contrast, typical EPEC was significantly associated with moderate-to-severe diarrhea at one site in Kenya.⁴¹ Other pathogens associated with diarrhea included rotavirus, Cryptosporidium, and Shigella.⁴¹

The emergence of MDR in E. coli has profoundly aggravated the problem of management of severe diarrheal illness and other presentations of E. coli. For instance, in a survey in Kenya⁴² a total of 916 E. coli isolates from urine, blood cultures, and diarrheal cases over a 1-year period were found to be resistant to ampicillin alone or a combination of ampicillin and other different classes of β-lactam antibiotics. A total of 247 were extended spectrum beta-lactamase (ESBL) producers, with 142 isolates exhibiting resistance to combinations of aztreonam and multiple cephalosporins including ceftazidime, while the second set of 105 isolates were resistant to the same panel of antibiotics but not to ceftazidime. The antibiotic-resistance patterns of E. coli isolated from outpatient clinics in Kenya reflected the pattern of commonly used antibacterials in health facilities, including ampicillin, trimethoprim/sulphamethoxazole, streptomycin, and amoxycillin/clavulanic acid, with lower levels of resistance to third-generation cephalosporins and ciprofloxacin (Table 1)⁴². We also noted a high level of resistance among E. coli in inpatients compared to outpatients (P < 0.001), which was a major hindrance in terms of costs of treatment and time spent in hospital for most facilities where the studies were conducted. Genotyping of the β-lactamase–resistance genes from E. coli from hospitalized patients and from community-acquired infections showed that there may be frequent exchange of resistance genes or E. coli strains between the community and hospital environments. Complex mutant TEM-like (CMT) and pAmpC producers were identified as the most predominant E. coli, whose spread in hospital and community settings posed a major challenge in management of such infections in Kenya. Several other countries in SSA have also reported on ESBL-producing E. coli as a major pathogen in carriage and in diarrheal disease. In a study in Guinea-Bissau, ESBL-producing E. coli were implicated in fecal carriage, a potential source of transmission in the community. Of the 408 children studied for fecal carriage of bacterial pathogens, 83 had ESBL-producing E. coli and nearly all isolates were multidrug resistant, with co-resistance to ciprofloxacin, trimethoprim/sulphamethoxazole, and aminoglycosides being common. A total of 38.5% were co-resistant to these classes of antibiotics plus extended-spectrum cephalosporins, encompassing nearly all agents that may be used for treatment of Gram⁻ sepsis in Guinea-Bissau.⁴³ In Cameroon, where a total of 358 fecal samples from healthy volunteers in the community were analyzed, 58 of the samples (16%) showed an ESBL phenotype, and E. coli was the species most frequently isolated among the ESBL producers in outpatients (66.7%) and student volunteers (90%). E. coli isolates were also resistant to gentamicin, ciprofloxacin, and trimethoprim/sulphamethoxazole.⁴⁴ A similar investigation to detect carriage in hospitalized patients at Yaoundé Central Hospital and at two hospitals in Ngaoundere, Cameroon found that E. coli represented 48% of all isolates (n =71), which were also ESBL producers, with CTX-M-15 (96 %) and SHV-12 (4 %) being the main genes detected. These study findings emphasize the need for surveillance at community and hospital settings for potential pathogens and the need for effective hospital infection-control programs to reduce possibilities of nosocomial spread of such pathogens.⁴⁵

Table 1.

Antimicrobial-susceptibility patterns of selected E. coli from various specimens from inpatients and outpatients in a surveillance study across hospitals in Kenya over a 10-year period from 1999

			Distribution (number (%)) of resistant strains in different specimen types			Distribution (number (%)) of resistant strains according to patient category
	Number of resistant strains n = 1327	% of resistant strains	Stool n = 505	Urine n = 451	Blood n = 371	Inpatient n = 654	Outpatient n = 673
Ampicillin	809	61	253 (31)	373 (46)	184 (23)	518 (64)	292 (36)
Aoxyclavulanate	478	36	143 (30)	249 (52)	86 (18)	329 (69)	148 (31)
Tazobactam	279	21	85 (30)	141 (51)	53 (19)	226 (81)	53 (19)
Aztreonam	385	29	121 (31)	191 (50)	73 (19)	258 (67)	127 (33)
Cephradine	411	31	121 (29)	256 (62)	34 (8)	234 (57)	177 (43)
Cefuroxime	358	27	97 (27)	184 (51)	78 (22)	266 (74)	93 (26)
Ceftriaxone	372	28	102 (27)	197 (53)	73 (19)	290 (78)	82 (22)
Cefoxitin	106	8	19 (18)	79 (74)	8 (6)	87 (82)	19 (18)
Nalidixic acid	239	18	86 (36)	132 (55)	21 (9)	163 (68)	77 (32)
Ciprofloxacin	106	8	19 (18)	79 (75)	8 (8)	65 (61)	41 (39)
Streptomycin	491	37	145 (30)	271 (55)	75 (15)	290 (59)	201 (41)
Kanamycin	305	23	85 (28)	167 (55)	53 (17)	195 (64)	110 (36)
Gentamicin	239	18	71 (30)	131 (54)	37 (16)	170 (71)	69 (29)
Chloramphenicol	478	36	167 (35)	233 (49)	78 (16)	320 (67)	158 (33)
Sulfamethoxazole	637	48	189 (30)	356 (56)	92 (14)	440 (69)	197 (31)
Tetracycline	703	53	218 (31)	353 (50)	132 (19)	478 (68)	225 (32)
Trimethoprim	557	42	167 (30)	290 (52)	100 (18)	379 (68)	178 (32)

Antibacterial resistant E. coli in urinary tract infections and extraintestinal infections

Urinary tract infections (UTIs) are treated empirically in SSA, where patients often cannot afford to consult a doctor or have laboratory tests. In Ethiopia,⁴⁶ E. coli isolated from patients (n = 228) with UTIs exhibited resistance to ampicillin but low resistance to ciprofloxacin (14.3%). Klebsiella spp, which is a prevalent pathogen of UTIs, displayed a similar resistance pattern to E. coli, although these isolates were susceptible to ciprofloxacin. In a study on susceptibility patterns of pathogens responsible for both community- and hospital-acquired UTIs to antibacterials agents currently used to treat UTIs in Rwanda, a high proportion of outpatient and inpatient E. coli strains (n = 119) were resistant to amoxicillin (86% vs. 100%), nitrofurantoin (26.4% vs. 29.8%), nalidixic acid (45.1% v. 77.4%), amoxycillin/clavulanic acid (56% vs. 70.2%), and gentamycin (36.1% vs. 46.8%%).⁴⁷ In addition, more than 70% of the outpatient isolates and more than 90% of the hospital isolates were resistant to trimethoprim/sulphamethoxazole. Resistance rates in isolates from outpatients were also higher compared to studies from other countries for the majority of antibacterials except third-generation cephalosporins, ceftriaxone, and ceftazidime. In a separate study in a Kenyan University outpatient clinic, E. coli isolated from UTIs were resistant to trimethoprim/sulphamethoxazole and ampicillin, 39% were resistant to co-amoxycillin/clavulanic acid, and 25% were resistant to nalidixic acid⁴⁸. Resistance to nitrofurantoin, gentamicin, cefuroxime, norfloxacin, and ciprofloxacin was demonstrated in less than 15% of the isolates. In a hospital-based survey in Dakar, Senegal, most E. coli (n = 1010) isolates from UTIs were resistant to amoxicillin (73.1%), amoxycillin/clavulanic acid (67.5%), cephalothin (55.8%), and trimethoprim/sulfamethoxazole (68.1%). Extended-spectrum β-lactamase was detected in 38 strains. The overall resistance rates to nalidixic acid, norfloxacin, and ciprofloxacin were 23.9%, 16.4%, and 15.5%, respectively. However, most of the E. coli were susceptible to gentamicin, nitrofurantoin, and fosfomycin (respective susceptibility rates at 93.8%, 89.9%, and 99.3%).⁴⁹ E. coli has also been detected as a major pathogen in surgical site infections (SSIs). For instance, in a referral hospital in Uganda, from the 314 enrolled patients with SSIs (mean age 29.7 ±13.14 years), 304 bacterial isolates were obtained of which 23.7% (72/304) were E. coli showing multidrug resistance to all commonly available antibacterials and three-fourths were ESBL producers.⁵⁰ In another study to determine the presence of ESBL-producing E. coli from outpatients in two university teaching hospitals in southeastern Nigeria, 77% of 44 E. coli strains were from urine, 13.6% from vaginal swabs, and 9.0% from wound swabs; 63.6% were from female patients, 68% were from outpatients, and 95.5% from patients younger than 30 years. All ESBL producers were positive in a PCR for the bla(CTX-M-1) cluster, in exemplary strains bla(CTX-M-15) was found by sequencing, while PCR for bla(TEM) and bla(OXA-1) was positive in 93.1% of strains and aac(6′)-Ib-cr was found in 97.7% of strains.⁵¹

Antibacterial-resistant E. coli in livestock production

Although it is generally assumed that farmers in resource-poor settings in SSA are unlikely to use huge quantities of antibiotics for livestock production, recent studies on microbial contamination and antibacterial resistance in the meat value chain including beef, pork, and poultry observed that antibacterial agents were regularly used on small-scale farms in Kenya.⁵² Oxytetracycline was the most commonly used antibacterial among small-scale poultry farmers, and others included fluoroquinolones (norfloxacin and enrofloxacin), erythromycin, sulphonamides, and trimethoprim/sulphamethoxazole. In poultry, soluble tetracyclines are used for specific therapies as well as for growth promotion. Additionally, sulphonamides and nitrofurans are also popular for the control of coccidiosis and poultry colibacillosis.⁵³ The high prevalence of resistance (tetracycline (60%), trimethoprim/sulphamethoxazole (54%), and ampicillin (30%)) in E. coli isolates from retail chicken meats in Nairobi outlets were consistent with those from chickens on small-scale farms, and the reported antibiotic usage either in water or as feed additives. E. coli isolates from beef carcasses at three abattoir locations in the Kenya study were most frequently resistant to tetracycline, ampicillin, and trimethoprim/sulphamethoxazole, with 33% resistant to two or three antibiotics. E. coli isolates from retail beef samples also showed resistance to ampicillin (30–31%) and tetracycline (18–20%), and 2–4% were resistant to nalidixic acid and ceftriaxone (Fig. 3).

Figure 3.

Antibiotic susceptibility patterns among E. coli isolates from beef carcasses at three abattoirs in Nairobi, Kenya (n = 188). AMC, amoxycillin/clavulanic acid; SXT, trimethoprim/sulphamethoxazole.

In general in the Kenyan poultry supply chain study, E. coli isolates were more frequently resistant to a broader range of antibiotics than beef or pork value chains, with the additional resistance to streptomycin, quinolones and third-generation cephalosporins at varying frequencies⁵². Prevalence of resistance in E. coli isolates was higher in samples from a commercial abattoir sourcing chicken from medium- and large-scale commercial farms compared with samples from small-scale poultry farms in the case of ampicillin (44% vs. 18%), chloramphenicol (28% vs. 5%) and ceftazidime (14% vs. 2%), but similar in the case of trimethoprim/sulphamethoxazole (60% vs. 45%), nalidixic acid (20% vs. 24%) and fluoroquinolones (2% vs. 5%), and lower in the cases of tetracycline (45% vs. 65%). It was also found that a larger proportion of isolates from commercial operations were multi-drug resistant (Fig. 4).

Figure 4.

Antibiotic susceptibility patterns for E. coli isolated from poultry in small-scale farms in Thika, Kenya (n = 350). AMC, amoxycillin/clavulanic acid; SXT, trimethoprim/sulphamethoxazole.

Methicillin-resistant Staphylococcus aureus infections

The emergence of methicillin-resistant Staphylococcus aureus (MRSA) was first reported in the 1960s, and by the end of that decade MRSA was responsible for major hospital outbreaks in Western Europe, Australia, and the United States.^54,55 Currently more outbreaks are being confirmed from community-acquired strains. However, relatively little data is available on the prevalence and characteristics of MRSA in SSA, but recent studies indicate that this pathogen is becoming a major concern in hospital settings. Also in SSA, close human contact with animals provides more opportunity for zoonotic transmission of MRSA⁵⁶ and other strain types, but data on the zoonotic burden of MRSA coming from animal reservoirs is also scarce.

In a study in a South African hospital, community-onset S. aureus bacteremia was identified in 161 children, representing an incidence of 26/100,000, with 63 (39%) isolates identified as MRSA (10/100,000). MRSA isolates were frequently multidrug resistant, with high resistance rates to trimethoprim/sulphamethoxazole (94%), erythromycin (92%), gentamicin (77%), rifampicin (63%), and clindamycin (65%). No significant differences in antibacterial resistance patterns was found between HIV-infected and uninfected children, although rifampicin resistance tended to be higher in the HIV-infected group.⁵⁷

Reports indicate that the prevalence of MRSA gradually rose in SSA during the first years of the new millennium.⁵⁸ For instance in Tunisia, the prevalence of MRSA varied between 41% and 46% after 2005, compared with a prevalence of 12–18% in the years before. However in South Africa, the prevalence of MRSA in referral hospitals decreased from 36% in 2006 to 24% during 2007–2011, and this was attributed in part to the implementation of effective infection-control policies. In Botswana, two studies on hospital-acquired MRSA infections showed the prevalence of MRSA varied between 23% and 44% for the period of 2000–2007. In a study evaluating the major causes of infections in adults and children at a referral hospital in Gabon, the overall proportion of MRSA from blood cultures was 5.8% (n= 19) and fluctuated during the studied period between 3.9% (n=2, 2009), 10.9% (n=7, 2010), 5.2% (n=5, 2011) and 3.3% (n=4, 2012) in children below 18 years of age.⁵⁹

At the University Teaching Hospital in Ibadan, Nigeria, during a 1-year surveillance in 2007 involving 346 non-duplicate S. aureus isolates from various clinical specimens, 20.23% were MRSA, with 47.15% of those being community-associated MRSA (CA-MRSA) from surgical or pediatric patients.⁶⁰ Although there are differences in the designs of the different studies reporting on MRSA prevalence and therefore many potential confounders such as country demography, the few data reported reflect a clear increasing trend in the prevalence of MRSA in the region. Reversing this trend will require regular surveillance and monitoring aimed at providing data and evidence for implementation of effective prevention strategies.

Streptococcus pneumoniae

Although the introduction of the widespread use of the conjugate pneumococcal vaccine is expected to reduce the childhood burden of illness in most of SSA, Streptococcus pneumoniae remains an important pathogen in otitis, sinusitis, bronchitis, and community-acquired pneumonia, as well as a predominant cause of meningitis and bacteremia. Before 1967, this pathogen was uniformly susceptible to penicillin and most other antibacterial agents; the first case of resistance to penicillin and other antibiotics was reported in South Africa in 1977⁶¹ and these strains spread to other countries as the S. pneumoniae became an important cause of invasive infections in children and in HIV-infected adults. In a review of cases in Central Africa,⁶² data on antibiotic-resistant S. pneumoniae were found in 11 studies from seven countries conducted between 1990 and 2008, mainly in urban children. Both invasive and carriage isolates of intermediate resistance to penicillin were reported up to 67.0%; high-level resistance remained less than 6% overall. Resistance rates for trimethoprim/sulphamethoxazole and chloramphenicol were above 60%. In contrast, data in a review conducted during the same period from community-acquired S. pneumoniae isolates from the East Africa region⁶³ showed that isolates were susceptible to most commonly used antibiotics, with the exception of trimethoprim/sulphamethoxazole, exhibiting no resistance to penicillin. This is similar to findings from West Africa, especially in Gambia,⁶⁴ where surveillance studies have been carried out extensively. For instance, a study of isolates from invasive pneumococcal disease obtained in 2000–2003 showed that only 6.6% of the isolates had intermediate resistance and none had full resistance to penicillin.⁶⁴

Consequences of antibacterial resistance

Few studies in SSA have evaluated the economic and social costs of antibacterial resistance. In one such study in Tanzania that assessed the incidence of bloodstream infection and risk factors for fatal outcome in a prospective cohort study of 1828 consecutive admissions of children aged 0–7 years with signs of systemic infection, antibacterial resistance was found to be a major risk factor for poor outcome of patients with septicemia.⁶⁵ Similarly, in a Ugandan hospital study, ESBL-producing Enterobacteriaceae and MRSA were clearly major risk factors for poor outcome of patients with SSI.⁵⁰ Studies in resource-rich countries have also demonstrated that patients with infections caused by antibiotic-resistant bacteria have higher mortality rates, longer durations of hospital stays, and higher healthcare costs compared to patients who have antibiotic-susceptible infections.⁶⁶ For instance, in a 3-year study using an electronic database covering four sites in a large healthcare system in New York City, significantly higher charges ($15,626; 95% CI $4339–$26,913 and $25,573; 95% CI, $9331–$41,816, respectively) and longer hospital stays were associated with community-acquired infections (3.3; 95% CI, 1.5–5.4). Patients with resistant healthcare-associated infections also had a significantly higher death rate (0.04; 95% CI, 0.01–0.08).⁶⁷ In another single healthcare study on hospitalized patients in the United States, the additional cost of hospitalization and length of stay attributed to infection with antibiotic-resistant Gram⁻ pathogens was recorded at 29.3% (P < 0.0001; 95% CI, 16.23–42.35) and 23.8% (P = 0.0003; 95% CI, 11.01–36.56), respectively, higher than those attributable to infections caused by antibiotic-susceptible pathogens.⁶⁸

What are our challenges in fighting AMR in SSA?

High burden of infectious diseases

In most SSA countries, the increasing trends of antibacterial resistance could be partly attributed to the high disease burden of infectious diseases. The top five killer diseases are treatable infectious diseases, with acute respiratory infections, diarrheal diseases, and HIV/AIDS-related complications among the leading causes of mortality. HIV/AIDS has predisposed populations to undue increase in the utilization of antibiotics to prevent and treat opportunistic infections, raising concerns over the looming emergence and spread of resistance to cheap and well-tolerated antibiotics.

Access to healthcare and laboratory capacity for diagnostic services

The healthcare facilities in most of SSA have inadequate diagnostic capacity, and the populace has limited access to formal healthcare services, as shown by the prevalence of self-medication and inadequate knowledge about the appropriate use of antibiotics by both healthcare providers and the general population. Inadequate laboratory diagnosis means that resistance frequently goes unrecognized and may only be detected as clinical treatment failure. Few peripheral health facilities have any laboratory facility that can provide diagnostic services, and most diagnosis is based on physician assessments on patients. Where few laboratory services may be available, reliable supply of reagents is a major challenge, and often diagnostic imprecision spurs overprescribing, particularly of broad-spectrum agents, promoting emergence and spread of potentially epidemic resistant clones. Sources include itinerant street vendors and kiosks in many informal settlements in cities, whose main intention is to make a quick profit and have every incentive to encourage antibacterial drug misuse.⁶⁹

Retail pharmacies, frequently operating without a licence, appear to be more accessible to patients: they are located within the community, do not charge consultant fees, have shorter waiting times, and are willing to negotiate treatment protocols to meet the financial needs of clients.^70,71 Huge populations, especially in rural settings in SSA, use retail pharmacies as the first site for outpatient care. A number of pilot studies in Kenya estimated that 65% of pharmacies dispense antibiotics without a doctor’s prescription.^72,73 In fact, drug retailers also appeared to be an important source of information about illness in general.^70,74 In an investigation of ORT use in treating diarrhea, caregivers of young children emphasised their reliance on both chemists and healthcare workers to guide them in diagnosing the problem and preparing treatments at home. In addition, pharmacy employees were relied upon for advice on childhood diarrhea and case management by a third of caregivers interviewed, leading to confusion about whether ORT was appropriate for all cases of diarrhea, and this resulted in a reduction of home-based rehydration techniques.⁷⁴

Inappropriate use of antibacterials

Although the total amount of antibiotic consumption is an important issue, perhaps of greater concern in SSA, where the need for antibiotics exceeds available stocks, is the way these drugs are used. Inappropriate antibiotic use in settings in SSA is complex and requires an understanding of the incentives behind prescribing, dispensing, and consuming drugs in order to design effective policies to curb resistance.⁷⁵ Some of the factors that drive irrational antibiotic use relate to motivations to save money by purchasing insufficient dosages and avoiding the consultant fees required at formal healthcare facilities. Other factors involve the geographic inaccessibility of formal healthcare facilities, noncompliance with prescribed regimens by prematurely ending treatment when symptoms subside, sharing of medication with family members and friends who may be unwell, and generally hoarding drugs for future use. Lack of laboratory confirmation for diagnosis, improper training, and legitimate fears of bad treatment outcomes in cases of critical illness are also important in driving inappropriate antibacterial use. Targeted initiatives by governments will be needed in order to provide partnerships between research and public health that can enhance awareness for communities and healthcare workers about the consequences of inappropriate use and emergence of resistance to antibacterials. Examples of such initiatives have been implemented in Kenya, Mozambique, and South Africa through the Global Antibiotic Resistance Partnerships (GARP) forums, and these have initiated various activities towards the rational use of antibacterials and efforts in minimizing resistance at community and healthcare levels.⁷⁶

Economic incentives

How much profit goes to a healthcare facility from drug sales, and what proportion of these sales are antibiotics? Are expensive, second-line antibiotics prescribed more for insured patients versus uninsured patients? Do patients themselves demand newer antibiotics when their prescriptions are covered by insurance? Are dispensing clinicians who directly benefit from drug sales more likely to prescribe antibiotics than non-dispensing clinicians? Many questions remain about the economic motivations behind antibiotic use and possible responses to changed incentives.^74,77 Although economic factors for prescribing, dispensing, and purchasing particular drugs are often discussed in theory, few papers carefully investigate these drivers of antibiotic use. Information is not available about the supply and demand response to prices of antibiotics versus alternative treatments, payment mechanisms available to consumers by dispensing agents (insurance, credit, and exemption schemes), healthcare workers’ incomes and facility financing mechanisms, and norms governing intra-household distribution of resources and control of income. In a poor informal settlement in Kenya, for instance, three-quarters of pharmacies report that their customers either cannot afford full regimens or have difficulty raising money for medicines recommended to them for treatment.⁷⁴

Providers’ incentives for antibiotic use

Not much is known about the influence of drug prescribers and dispensers, including clinicians, nurses, pharmacists, pharmacy assistants, and small-shop owners and counter staff, on antibiotic use in Kenya. A few studies have examined the extent of providers’ knowledge in different contexts and the prescription or recommendation of antibiotics in practice. Other influences appear to drive supply, however, since even practitioners who understand the correct treatment may inappropriately prescribe antibiotics.^{73, 77} Inadequate or underutilized diagnostics can cause uncertainty about the best treatment and fear about the outcome if the healthcare provider withholds antibiotics, leading not only to overuse of antibacterials in general, but also to a heavy reliance on broad-spectrum antibiotics.

Conclusion

Antibacterial resistance among key bacterial pathogens of major public health significance in SSA remains a big hindrance to providing basic healthcare needs to communities. Although the paucity of data and lack of effective systems for routine surveillance and reporting make estimations of the magnitude of the problem of antibacterial resistance difficult in SSA, there is evidence of the effects of resistance, especially in rising cases of treatment failure in severe infections. Strategies aimed at the prevention of transmission of infectious diseases could lead to reduced disease burden and hence lower prevalence of resistance. Similar effects may be derived from widespread adoption of vaccination strategies to reduce disease burden in the community, such as those available for S. Typhi, H. influenzae type B, and S. pneumonia, which have already yielded positive results in countries that have rolled out the vaccinations. We expect that public health–directed measures aimed at improving the supply of clean water and environmental sanitation will also contribute significantly to reducing the burden of infectious diseases in most of SSA. Data from surveillance studies will be vital to showing resistance trends among key bacterial pathogens, monitoring rational use of antibiotics, and developing clinical management guidelines to standardize therapy in healthcare facilities aimed at curtailing development of resistance against the remaining effective antibacterials. For these efforts to succeed, governments in the SSA region will be required to implement integrated multi-sectorial routine surveillance and monitoring programs and to enforce policies that ensure the prudent use of available antibacterials.