Antimicrobial Susceptibility and Mechanisms of Resistance in Shigella and Salmonella Isolates from Children under Five Years of Age with Diarrhea in Rural Mozambique

Abstract

The antimicrobial susceptibility and mechanisms of resistance of 109 Shigella and 40 Salmonella isolates from children with diarrhea in southern Mozambique were assessed. The susceptibility to seven antimicrobial agents was tested by disk diffusion, and mechanisms of resistance were searched by PCR or colorimetric method. A high proportion of Shigella isolates were resistant to chloramphenicol (Chl) (52%), ampicillin (Amp) (56%), tetracycline (Tet) (66%), and trimethoprim-sulfamethoxazole (Sxt) (84%). Sixty-five percent of the isolates were multidrug resistant. Shigella flexneri isolates were more resistant than those of Shigella sonnei to Amp (66% versus 0.0%, P < 0.001) and Chl (61% versus 0.0%, P < 0.001), whereas S. sonnei isolates presented higher resistance to Tet than S. flexneri isolates (93% versus 64%, P = 0.02). Resistance among Salmonella isolates was as follows: Tet and Chl, 15% each; Sxt, 18%; and Amp, 25%. Only 3% of Salmonella isolates were resistant to nalidixic acid (Nal), and none to ciprofloxacin or ceftriaxone (Cro). Among Salmonella isolates, multiresistance was found in 23%. Among Shigella isolates, antibiotic resistance was related mainly to the presence of oxa-1-like β-lactamases for Amp, dfrA1 genes for Sxt, tetB genes for Tet, and Chl acetyltransferase (CAT) activity for Chl. Among Salmonella isolates, resistance was conferred by tem-like β-lactamases for Amp, floR genes and CAT activity for Chl, tetA genes for Tet, and dfrA1 genes for Sxt. Our data show that Shigella isolates are resistant mostly to the most available, inexpensive antibiotics by various molecular mechanisms but remain susceptible to ciprofloxacin, Cro, and Nal, which is the first line for empirical treatment of shigellosis in the country.

Diarrhea disease is one of the main causes of childhood mortality worldwide. It is estimated that 17% of the 10.8 million deaths of children under 5 years of age worldwide is due to diarrhea, with developing countries being the most affected (6). Diarrhea can be caused by different agents, such as bacteria, parasites, and virus (1, 24, 35). The severity of the illness is mediated by different factors related to both the patient (nutritional status, presence of concomitant illness, and human immunodeficiency virus status) and the etiological agent (specific bacterial virulence and antimicrobial resistance). Salmonella spp. and Shigella spp. remain among the bacteria most frequently isolated from stool samples obtained from diarrhea patients, especially in rural areas from developing countries (24, 35). Salmonella spp. usually produce a self-limited illness, whereas infections due to Shigella are likely to be more severe.

The management of acute diarrhea is based on replacement of fluids (8). However, antibiotics might be required for management of the most severe cases or cases involving malnourished children. With shigellosis, appropriate antimicrobial therapy can reduce the duration of fever and the period of shedding of the pathogens (26), which is relevant to transmission of the pathogen to susceptible contacts. At the Manhiça District Hospital (MDH), the combination of ampicillin (Amp) plus gentamicin used with children younger than 2 months of age and chloramphenicol (Chl) used alone with older children are the most available therapeutic options for treating bacterial infections (32), while nalidixic acid (Nal) is recommended for cases of dysentery.

Until recently, Salmonella spp. were highly susceptible to the most commonly used antibiotics (34). However, in the last decade, the emergence of multidrug-resistant nontyphoidal Salmonella strains, including isolates resistant to quinolones, has been described worldwide (14, 33, 36). Also, the increase in the number of Shigella isolates resistant to most of the antibiotics available in countries where the choice of treatment is limited (14, 16, 21, 25, 36) represents an important health problem.

The mechanisms of antimicrobial resistance are associated with intrinsic resistance, point mutations, and acquired or extrachromosomal resistance (31). A wide range of molecular mechanisms, such as the presence of β-lactamases, dihydrofolate reductase, Chl acetyltransferase (CAT) enzymes and many others (7, 10, 11, 22, 27), have been described. However, few studies have investigated molecular mechanisms of antimicrobial resistance among isolates from sub-Saharan Africa, due mainly to the limited number of laboratories and research facilities with an adequate infrastructure available on the continent (10, 21).

Currently, data about antimicrobial resistance among diarrheagenic bacteria in Mozambique are scarce. This study describes the antimicrobial susceptibility and the molecular mechanisms of resistance in Salmonella and Shigella isolates from children with diarrhea in a rural hospital in Mozambique.

MATERIALS AND METHODS

Study site and population.

The study was conducted by the Centro de Investigação em Saúde da Manhiça (CISM), Maputo, Mozambique, at the Manhiça District Hospital (MDH), a 110-bed referral health facility for Manhiça District, a rural area located 80 km north of Maputo in the Maputo province in southern Mozambique. The area has a subtropical climate with two distinct seasons, a warm and rainy season between November and April and a cool and dry season during the rest of the year. The district has an estimated population of 140,000 inhabitants. A full description of the geographical and socio-demographic characteristics of the study community can be found elsewhere (2). The CISM has been running a continuous demographic surveillance system since 1996 in an area that currently extends 500 km² and includes 82,000 inhabitants. The CISM is adjacent to the MDH, and since 1997 they have jointly operated round-the-clock surveillance of all pediatric visits to the outpatient department and admissions to the wards.

Patients and bacterial isolation.

Patients were children <5 years of age who presented to the outpatient department with diarrhea. Antibiotic therapy with Nal was started among children presenting bloody diarrhea, according to national guidelines. When shigellosis was confirmed, a field worker visited the household to provide medication.

A total of 109 Shigella isolates and 40 Salmonella isolates, which were recovered between July 2001 and July 2003 from children under the age of five who were attended to at the MDH, were investigated. Briefly, after a consent form granting permission for the use of the sample for research purposes was signed by the parent/legal guardian, rectal swab specimens were collected and cultured onto solid media (MacConkey's agar, Salmonella-Shigella agar, xylose lysin deoxycholate agar) for bacteria isolation. Presumptive bacterial identification was done on the basis of colony color and morphology. Putative Shigella or Salmonella organisms were selected and inoculated into Kliger's iron agar slants and urea agar. Shigella spp. and Salmonella spp. were characterized biochemically and confirmed by an API-20E (bioMerieux, Marcy l'Etoile, France).

Shigella serogroups were determined by slide agglutination using specific antisera according to the manufacturer's instructions (Difco Laboratories, Detroit, MI). Salmonella enterica serotyping was performed at Instituto de Salud Carlos III, Madrid, Spain, using somatic and flagellum antiserum, as previously described (19). Isolates were stored in the Microbank system (Pro-Lab Diagnostics, Cheshire, United Kingdom) at −70°C and recovered at a later stage for molecular characterization.

Susceptibility testing.

Antimicrobial susceptibility to Amp, ceftriaxone (Cro), Chl, Nal, tetracycline (Tet), and trimethoprim-sulfamethoxazole (Sxt) was performed by disk diffusion in accordance with the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (9). Additionally, Nal-resistant strains were tested for ciprofloxacin resistance.

In this study, multidrug resistance (MDR) is defined as the presence of resistance to at least two nonrelated antimicrobial agents.

Molecular mechanisms of antimicrobial resistance.

For those Shigella and Salmonella isolates presenting full resistance to antibiotics, genes encoding β-lactamases (tem-like, carb-like, shv-like, oxa-1-like, oxa-2-like, and oxa-5-like) associated with Amp resistance were investigated by PCR, as described previously (22, 30). Determination of the presence of the cmlA and floR genes associated with Chl resistance was carried out as described by Cabrera et al. (7). The colorimetric method was used to detect the presence of CAT activity (4). Detection of the tetA, tetB, and tetG genes associated with Tet resistance was performed as previously described (7, 13), and detection of genes associated with Sxt resistance (dfrA1, dfrA5, dfrA6, dfrA7, dfrA8, dfrA12, dfrA13, dfrA14, dfrA15, dfrA15b, dfrA16, dfrA16b, and dfrA17) was carried out as described previously (23). Briefly, one colony of the strain was boiled in 25 μl sterile distilled water for 10 min, and 25 μl of a mixture containing 1.0 mM of each primer, 0.4 mM dNTPs, 2× PCR buffer with Mg, and 2.5 units of Taq polymerase was added. The mixture was overlaid with mineral oil, and the reaction was carried out in an Eppendorf thermocycler. The program consisted of 30 cycles as follows: 1-min denaturation (94°C), 1-min annealing, and 1-min elongation (72°C), plus a final extension of 10 min at 72°C. Annealing temperatures varied for the different primers. The reaction products were run in 1.5% agarose gels with 0.5 μg/ml of ethidium bromide.

Statistical analysis.

Data were double entered into a FoxPro database using visual FoxPro version 2.6 (Microsoft Corporation, Redmond, WA). The two entries were compared, and discrepancies were resolved by referring to the original forms. The statistical analysis was performed using STATA software (version 9.0). Proportions were compared using the chi-square test or Fisher's exact test, as appropriate.

RESULTS

Antimicrobial resistance.

A total of 109 Shigella isolates (94 were S. flexneri and 15 S. sonnei) and 40 Salmonella enterica isolates (10 were S. enterica serovar Infantis, 9 S. enterica serovar Typhimurium, 5 S. enterica serovar Virchow, 3 S. enterica serovar Isangi, 3 S. enterica serovar Heidelberg, 2 S. enterica serovar Enteritidis, and 8 other serotypes) were assessed for their antimicrobial susceptibility patterns.

Shigella isolates showed high levels of resistance to Sxt (84%), Tet (66%), Amp (56%), and Chl (52%). No Shigella isolates were resistant to Nal or Cro. When the resistance was analyzed by species, it was observed that S. flexneri isolates were more frequently resistant to Amp (66% versus 0.0%, P < 0.001) and Chl (61% versus 0%, P < 0.001), whereas S. sonnei isolates were more frequently resistant to Tet (93% versus 64%, P = 0.02) (Table 1).

TABLE 1.

Antimicrobial resistance (%) of Shigella and Salmonella isolates from children under 5 years of age with diarrhea at the MDH

Organism (no. of isolates)	Amp				Chl				Sxt				Tet
	MIC range	MIC50	MIC90	% Resistant	MIC range	MIC50	MIC90	% Resistant	MIC range	MIC50	MIC90	% Resistant	Range	MIC50	MIC90	% Resistant
Shigella spp. (109)	0.5->128	64	128	56	0.5-128	32	32	57	1-128	>128	>128	84	1-64	32	64	66
S. flexneri (94)	0.5->128	64	128	66a	0.5-128	32	32	61a	4-128	>128	>128	82	1-64	32	64	64
S. sonnei (15)	0.5->128	1	32	0	1-128	2	32	0	1->128	>128	>128	100	1-64	32	64	93b
Salmonella spp. (40)	<0.5->128	≤0.5	>128	25	<0.5->128	≤0.5	>128	15	<0.5->128	≤0.5	>128	18	<0.5->128	≤0.5	1	15
S. Infantis (10)	≤0.5	≤0.5	≤0.5	0	≤0.5	≤0.5	≤0.5	0	≤0.5	≤0.5	≤0.5	0	≤0.5	≤0.5	≤0.5	0
S. Typhimurium (9)	≤0.5->128	>128	>128	67	≤0.5->128	1	>128	22	<1	1	>128	44	<0.5-1	≤0.5	1	0
S. Virchow (5)	≤0.5-32	≤0.5	≤0.5	0	≤0.5	≤0.5	≤0.5	0	≤0.5	≤0.5	≤0.5	0	<0.5	≤0.5	≤0.5	0
Other Salmonella isolates (16)	<0.5->128	≤0.5	≤0.5	25	<0.5->128	≤0.5	>128	25	<0.5->128	≤0.5	>128	19	<0.5->128	≤0.5	>128	25

^aS. flexneri isolates were more frequently resistant to Amp and Chl than S. sonnei isolates (P < 0.001; Fisher's exact test).

^bS. sonnei isolates were more frequently resistant to Tet than S. sonnei isolates (P < 0.02; chi-squared test).

The resistance rates among Salmonella isolates were lower than those reported for Shigella isolates. Resistance to Amp and Sxt was detected in 25% and 18% of isolates, respectively. For Tet and Chl, resistance rates were 15% each. One isolate (3%) was found to be resistant to Nal, while for this same isolate, no resistance to Cro or ciprofloxacin was detected. By serotypes, 67% of Salmonella serovar Typhimurium isolates were resistant to Amp. Additionally, a higher proportion of S. Typhimurium isolates were resistant to Sxt and Chl compared to other serotypes (Table 1). All Salmonella serovar Infantis isolates were susceptible to all antibiotics tested, and one isolate of Salmonella serovar Virchow was resistant to Tet. Three isolates of S. serovar Isangi were found to be MDR (Table 2).

TABLE 2.

Profiles of Shigella and Salmonella resistance to at least two unrelated antibiotics^a

Antibiotic resistance phenotype	No. (%) of resistant isolates of:
	Shigella flexneri	Shigella sonnei	All Shigella spp.	All Salmonella spp.
Ampr Sxtr	1 (1)		1 (1)	2 (5)
Ampr Chlr	1 (1)		1 (1)	1 (3)
Ampr Tetr				1 (3)
Chlr Sxtr	1 (1)		1 (1)
Chlr Tetr	1 (1)		1 (1)
Sxtr Tetr		12 (80)	12 (11)
Tetr Nalr				1 (3)
Ampr Chlr Sxtr				2 (5)
Ampr Chlr Tetr	9 (10)		9 (8)
Ampr Sxtr Chlr Tetr	46 (49)		46 (42)	3 (8)
Total	59 (63)	12 (80)	71 (65)	9 (23)

^aValues represent isolates exhibiting resistance to two or more unrelated antimicrobial agents.

MDR was detected with 65% of Shigella isolates and 23% of Salmonella isolates. Resistance to two antibiotics was observed in 1% of S. flexneri isolates, 80% of S. sonnei isolates, and 5% of Salmonella isolates. Ten percent of S. flexneri isolates and 5% of Salmonella isolates presented resistance to three agents, while 49% of S. flexneri and 8% of Salmonella isolates were resistant to four antimicrobial agents (Table 2).

Mechanisms of resistance.

A great variety of molecular mechanisms of resistance were detected, although some of them in few isolates (Table 3). Among S. flexneri isolates, the resistance to Amp was related mainly to the presence of oxa-1-like genes (55 out of 61 Amp-resistant isolates; 90%), while Tet resistance was related mainly to the presence of the tetB gene (57 out of 58 Tet-resistant isolates; 98%). The presence of genes dfrA1 (40 out of 71 Sxt-resistant isolates; 56%) and dfrA14-like (25 out of 71 Sxt-resistant isolates; 31%) was detected mainly in the resistance to Sxt. A total of 27% (19 out of 71) of S. flexneri isolates, from which dfrA14-like was found, presented a “new” restriction fragment length polymorphism (RFLP) pattern. Resistance to Chl was associated mainly with the presence of CAT activity. Meanwhile, for S. sonnei isolates the resistance to Sxt and Tet was related mainly to the presence of the tetA and dfrA1 genes, respectively (Table 3).

TABLE 3.

Mechanisms of resistance of Shigella and Salmonella isolates from children with diarrhea in rural hospital in Mozambique

Antimicrobial or mechanism of resistance	No. of isolates with indicated mechanism of resistance/total no. of resistant isolates (%)
	S. flexneri	S. sonnei	All Shigella isolates	All Salmonella isolates
Amp
oxa-1-like	55/61 (90)		55/61 (90)	0/10 (0)
oxa-2-like	3/61 (5)		3/61 (5)	0/10 (0)
oxa-5-like	1/61 (2)		1/61 (2)	2/10 (20)
tem-like	3/61 (5)		3/61 (5)	7/10 (70)
shv-like	1/61 (2)		1/61 (2)	0/10 (0)
carb-like	0/61 (0)		0/61 (0)	0/10 (0)
Chl
cmlA	1/57 (2)		1/57 (2)	2/6 (33)
floR	0/57 (0)		0/57 (0)	5/6 (83)
CAT	51/57 (89)		51/57 (89)	4/6 (67)
Tet
tetA	1/58 (2)	10/14 (71)	11/72 (15)	2/6 (33)
tetB	57/58 (98)	0/14 (0)	52/72 (79)	0/6 (0)
tetG	13/58 (22)	5/14 (36)	18/72 (25)	0/6 (0)
Sxt
drfA1	40/71 (56)	11/13 (85)	51/84 (61)	2/7 (29)
dfrA5	6/71 (8)	1/13 (8)	7/84 (8)	0/7 (0)
dfrA7	4/71 (6)	0/13 (0)	4/84 (5)	1/7 (14)
dfrA8	1/71 (1)	0/13 (0)	1/84 (1)	0/7 (0)
dfrA12	6/71 (8)	2/13 (15)	8/84 (10)	0/7 (0)
dfrA14	25/71 (35)	1/13 (8)	26/84 (31)	0/7 (0)
dfrA14-likea	19/71 (27)	3/13 (23)	22/84 (26)	0/7 (0)

^aIsolates presenting a new RFLP pattern.

Amp resistance among Salmonella isolates was exclusively associated with tem-like genes, which were detected with 70% (7/10) of the isolates, of which six were determined to be S. Typhimurium. CAT activity and the presence of the floR gene, both associated with resistance to Chl, were found with 67% and 83% of isolates, respectively. Resistance to Tet was associated with the tetA gene (33%; 2/6), and resistance to Sxt was associated with the presence of dfrA1 (29%; 2/7).

With a number of isolates, the presence of more than one mechanism of resistance to the same antimicrobial agent was identified. This phenomenon was observed with 32 S. flexneri isolates resistant to Sxt; 18 of them involved dfrA14-like genes with a new RFLP pattern. With 11 isolates resistant to Tet, 5 resistant to Amp, and 1 resistant to Chl also, more than one resistance mechanism to these agents was identified. The same occurrence was observed with S. sonnei isolates, of which 11 presented more than one mechanism of resistance to Tet and 4 were nonsusceptible to Sxt and involved three dfrA14-like genes with an unrecognized RFLP pattern. Among Salmonella isolates, 4 of the 5 isolates resistant to Chl also presented more than one mechanism of resistance, with three showing the presence of floR and CAT activity, and one showing the presence of floR and cmlA (data not shown).

A number of antimicrobial-resistant isolates were without an identified mechanism of resistance. Thus, among S. flexneri isolates, seven Amp-resistant, three Chl-resistant, two Sxt-resistant, and two Tet-resistant isolates appeared to have mechanisms of resistance other than the ones investigated here. The same occurrence was observed with two Tet-resistant S. sonnei isolates. Among Salmonella isolates, the resistance mechanism was not identified in 3/7 (43%) isolates resistant to Sxt, in 2/6 (33%) to Tet, and 1/10 (10%) to Amp.

DISCUSSION

In this paper we have correlated the levels of antimicrobial resistance and their mechanisms among Shigella and Salmonella isolates from children with diarrhea. We report a high prevalence of antimicrobial resistance among Shigella isolates, with S. flexneri isolates being much more resistant than S. sonnei isolates; in contrast, the Salmonella isolates had a much lower incidence of resistance. Selection of resistant strains might be favored in Shigella, because this microorganism usually causes severe infections that are likely to require antibiotic treatment. Despite the high proportion of antimicrobial resistance observed among Shigella isolates, these organisms remain highly susceptible to quinolones (Nal) and Cro, supporting the current use of Nal as first-line treatment of dysentery in Mozambique. The use of quinolones in children has been controversial but has recently been demonstrated to be safely used with this population (20). The use of third-generation cephalosporines, which are also effective against other diarrhea agents, is limited by cost in resource-poor countries.

Salmonella isolates presented a lower level of resistance. Although, interestingly, a small number-all the Salmonella serovar Isangi isolates in this study-were more resistant to the antibiotics tested, which is consistent with other reports from the region (12, 15).

This study has also shown that resistance to at least two antibiotics was common, with the majority of MDR in S. flexneri focused on Amp/Sxt/Chl/Tet, while in S. sonnei it was almost exclusively focused on Sxt/Tet; in contrast, the Salmonella isolates showed little MDR. The reason why antibiotic resistance is differently distributed among Shigella spp. is unknown. An explanation could be provided by differences in the genetic locations of the antimicrobial resistance genes that might possess a species-specific association (22). This hypothesis might be supported by the fact that, consistent with other studies (3, 5, 17, 21, 22), S. sonnei has been less resistant to Chl compared to S. flexneri. In addition, differences in the prevalences of genes encoding mechanisms of resistance among S. sonnei and S. flexneri have been reported previously (18, 22).

The high prevalence of the oxa-1-like-type β-lactamase among the Amp-resistant S. flexneri isolates is consistent with other findings (10, 21, 22), but the detection of other β-lactamases shows that other mechanisms of resistance could be present in S. flexneri and future alterations in their prevalence may occur. The genetic location of the antimicrobial resistance genes that might possess a species-specific association may help to explain the differences found with the mechanisms of resistance to Tet among S. flexneri (tetB) and S. sonnei (tetA) isolates. We found that the dfrA1 genes were the main mechanism of Sxt resistance among Shigella isolates (22). Interestingly, a high proportion of dfr14-like genes (22/84) presented a new RFLP pattern. The DNA sequences of the new RFLP pattern are currently being determined.

Among Salmonella isolates, antimicrobial resistance was related to tem-like β-lactamase (Amp), CAT activity and the floR gene (Chl), and the tetA gene (Tet), findings which are in agreement with those published by others (7, 10, 29). The prevalence of dfrA1 found among Salmonella isolates is also similar to the findings reported by Cabrera et al. (7). Differences in the prevalences of different resistance genes among Salmonella serotypes cannot be inferred from our data due to the small number of isolates belonging to each of the serotypes recovered.

A high number of isolates presented more than one mechanism of resistance to the same antimicrobial agent. This could facilitate adaptation to the changes of antibiotic selection pressure, as in some cases, such as with β-lactamases, the patterns of resistance conferred by each mechanism are not equal. Antibiotic resistance in a number of isolates appeared to be conferred by mechanisms other than those analyzed in this study. It should be taken into account that other mechanisms, such as overexpressed efflux pumps, may be present (28).

In summary, the present study shows that the prevalence in Shigella isolates of antimicrobial resistance to the most commonly used antibiotic in Mozambique is high, with S. flexneri isolates being much more resistant than S. sonnei isolates, but quinolones and the third generation of cephalosporins remain effective; resistance among Salmonella isolates was low. Among Shigella isolates, there was a difference in MDR profiles between S. flexneri and S. sonnei. S. flexneri Amp resistance linked with the oxa-1-like gene, Tet resistance with tetB, and Sxt resistance with dfrA1, dfrA14, and a related sequence of dfrA14-like; in contrast, S. sonnei Tet resistance was associated with tetA, while Sxt resistance was linked primarily to drfA1. Data generated in this study are also important in guiding clinicians in selection of antibiotics for empirical treatment. Continuous monitoring of antibiotic resistance in developing countries is imperative to ensure that severe diarrhea infections remain treatable.