Antibiotic Resistance in Diarrheagenic Escherichia coli and Shigella Strains Isolated from Children in Hanoi, Vietnam

Abstract

The MICs for 162 diarrheagenic Escherichia coli strains and 28 Shigella strains were determined on the basis of NCCLS guidelines. More than 75% of the strains were resistant to ampicillin, chloramphenicol (53.6% of Shigella strains), and trimethoprim-sulfamethoxazole. Multiresistance was detected in 89.5% of E. coli strains and 78.6% of Shigella strains.

Antibiotics have revolutionized the treatment of common bacterial infections and play a crucial role in reducing mortality. Antimicrobial therapy should be used in severe cases of diarrheal disease to reduce the duration of illness and may be used to prevent traveler's diarrhea (12). However, the progressive increase in antibiotic resistance among enteric pathogens in developing countries is becoming a critical area of concern. In addition, the overuse and misuse of antibiotics in the treatment of diarrhea could lead to an increase of antibiotic resistance.

In 1986, an economic reform, known as Doi Moi, was launched in Vietnam, introducing a market economy with privatization in all sectors, including drug provision (5). People could easily buy antibiotics without a doctor's prescription for treating themselves. One study performed in 1999 in Hanoi showed that 90% of drug dispensing was without a prescription and 94.9% of customers decided by themselves which drugs to buy (6). As a result, many children with symptoms of illnesses (respiratory infections, diarrhea, etc.) may have been empirically treated with antibiotics without advice from medical staffs (10, 16, 21). According to epidemiological studies in Vietnam, Escherichia coli is most commonly isolated in clinical samples from patients with diarrhea and shows a high prevalence of resistance to antibiotics (2, 9). Knowledge of local antimicrobial therapy is important in selecting the appropriate therapy. Moreover, there are few recently reported studies about antimicrobial resistance among diarrheagenic E. coli and Shigella strains in children in Vietnam (2, 9).

In this study, the susceptibilities of 28 Shigella strains and 162 diarrheagenic E. coli strains to different antibiotics were evaluated. Specifically, 28 Shigella strains, including 1 Shigella boydii strain, 7 Shigella flexneri strains, 20 Shigella sonnei strains, and 162 diarrheagenic E. coli strains belonging to 86 enteroaggregative E. coli (EAEC) strains, 12 enteroinvasive E. coli (EIEC) strains, 50 enteropathogenic E. coli (EPEC) strains, and 14 enterotoxigenic E. coli (ETEC) strains were examined. E. coli (ATCC 25922) and Staphylococcus aureus (ATCC 29213) were used for quality control of antibiotic susceptibility testing.

The following antibiotics were used for susceptibility testing: ampicillin (AMP; AstraZeneca), chloramphenicol (CHL; Sigma), trimethoprim-sulfamethoxazole (SXT; Sigma), imipenem (IPM; Merck Shark & Dohme B.V.), cefuroxime (CXM; Sigma), cefotaxime (CTX; Sigma), nalidixic acid (NAL; Sigma), and ciprofloxacin (CIP; Bayer AG).

MICs were determined by the agar dilution method following the recommendations of the National Committee for Clinical Laboratory Standards (NCCLS) (14, 15). Data from antibiotic susceptibility testing were analyzed using WHONET 5.1 software. The data were analyzed using Kruskal-Wallis H test for multiple comparison and Mann-Whitney U test for comparing two groups of E. coli in terms of antibiotic resistance to each antibiotic. A P value of <0.05 was considered significant.

The project was approved by the Ethical Committees of both the Karolinska Institutet, Stockholm, Sweden, and Hanoi Medical University, Hanoi, Vietnam.

The antibiotic susceptibility testing data are shown in Table 1. Of the 162 diarrheagenic E. coli isolates, 86.4% were resistant to AMP, 77.2% were resistant to CHL, 29.6% were resistant to CXM, 24.1% were resistant to CTX, 19.1% were resistant to NAL, 3.7% were resistant to CIP, 88.3% were resistant to SXT, and all were sensitive to IPM. Of the few E. coli strains resistant to CIP, five strains were EPEC and one strain was an ETEC. The traditional antibiotics, including AMP, CHL, and SXT, showed low activity against the diarrheagenic E. coli strains (MIC at which 90% of the isolates tested are inhibited [MIC₉₀] of 1,024 mg/liter for AMP and CHL; MIC₉₀ of 4/76 mg/liter for SXT). CXM, CTX, and NAL showed moderate activity. However, MIC₉₀s of CTX and CXM were rather high at >1,024 mg/liter. They were still more active than the traditional antibiotics but less active than NAL with a MIC₉₀ of 128 mg/liter.

TABLE 1.

Antibiotic susceptibilities of diarrheagenic E. coli and Shigella strains

Organism and agent	No. of isolates	% not susceptible	MIC
			MIC50	MIC90	Range
Diarrheagenic E. colia	162
AMP		86.4	>1,024	>1,024	2->1,024
CHL		77.2	256	1,024	2-1,024
CXM		29.6	4	>1,024	0.032->1,024
CTX		24.1	0.064	>1,024	0.032->1,024
NAL		19.1	2	128	1->1,024
CIP		3.7	0.016	0.25	0.008-128
IPM		0	0.125	0.125	0.064-1
SXT		88.3	>4/76	>4/76	0.032->4/76
EAEC	86
AMP		90.7	>1,024	>1,024	2-1,024
CHL		87.2	512	1,024	2-1,024
CXM		36	8	>1,024	0.032-1,024
CTX		26.7	0.064	>1,024	0.032-1,024
NAL		18.6	2	64	1-1,024
CIP		0	0.016	0.125	0.008-1
IPM		0	0.125	0.125	0.064-1
SXT		91.9	>4/76	>4/76	0.032-16
EIEC	12
AMP		91.7	1,024	>1,024	8-1,024
CHL		83.3	256	512	2-1,024
CXM		16.7	4	>1,024	0.064-1,024
CTX		25	0.064	>1,024	0.032-1,024
NAL		8.3	2	4	1-64
CIP		0	0.016	0.032	0.016-0.064
IPM		0	0.125	0.125	0.064-0.125
SXT		100	>4/76	>4/76	8-16
EPEC	50
AMP		82	1,024	>1,024	2-1,024
CHL		64	256	1,024	4-1,024
CXM		24	8	1,024	0.064-1,024
CTX		20	0.064	64	0.032-1,024
NAL		24	2	128	2-1,024
CIP		10	0.016	1	0.016-128
IPM		0	0.125	0.25	0.064-1
SXT		84	>4/76	>4/76	0.032-16
ETEC	14
AMP		71.4	512	>1,024	2-1,024
CHL		57.1	128	1,024	4-1,024
CXM		21.4	4	1,024	0.064-1,024
CTX		21.4	0.032	128	0.032-128
NAL		14.3	2	>1,024	2-1,024
CIP		7.1	0.016	1	0.008-32
IPM		0	0.125	0.125	0.064-0.25
SXT		71.4	>4/76	>4/76	0.064-16
Shigellab	28
AMP		75	256	1,024	2-1,024
CHL		53.6	16	256	2-1,024
CXM		14.3	2	128	0.032-1,024
CTX		10.7	0.064	128	0.016-256
NAL		7.1	2	8	1-32
CIP		3.6	0.016	0.064	0.016-2
IPM		10.7	0.125	64	0.064-128
SXT		89.3	16	16	0.064-16
S. sonnei	20
AMP		70	128	>1,024	2->1,024
CHL		40	4	256	2-1,024
CXM		15	0.064	128	0.032-1,024
CTX		10	0.064	0.25	0.032-128
NAL		5	2	8	1-32
CIP		0	0.016	0.032	0.016-0.25
IPM		10	0.125	0.25	0.064-128
SXT		90	>4/76	>4/76	0.064->4/76
S. flexneri	7
AMP		85.7	512	>1,024	4->1,024
CHL		85.7	128	256	8-128
CXM		0	0.125	2	0.064-4
CTX		0	0.064	0.25	0.016-0.25
NAL		0	2	2	1-2
CIP		0	0.016	0.016	0.016-0.016
IPM		0	0.125	0.25	0.064-0.25
SXT		85.7	>4/76	>4/76	1->4/76
S. boydiic	1

^aIncludes EAEC, EIEC, EPEC, and ETEC strains.

^bIncludes S. sonnei, S. flexneri, and S. boydii strains.

^cOne S. boydii strain was resistant to all antibiotics tested.

The differences in the distributions of resistance and MICs were seen for individual antibiotics and each category of E. coli. The multiple comparison showed a significant difference in resistance to CHL and CIP (P = 0.004 and 0.021, respectively). For SXT, the P value was borderline significant (P = 0.062). When the comparisons of antibiotic resistance of two groups of E. coli were performed, EAEC were significantly more susceptible to CIP than EPEC and ETEC (P = 0.003 and 0.013, respectively). However, EAEC were more resistant to CHL (P = 0.002 and 0.006, respectively). This group of E. coli also showed higher resistance to AMP and SXT (P = 0.041 and 0.024, respectively) compared to ETEC. The other comparisons among E. coli types to antibiotics did not appear to be statistically different.

Different resistance patterns were defined in the four categories of diarrheagenic E. coli. The most prevalent multiresistance pattern (resistance to at least two antibiotics) was Amp^r Chl^r Cxm^s Ctx^s Nal^s Cip^s Ipm^s Sxt^r in all types of E. coli, namely, 34.8, 16.6, 28, and 21.4% for EAEC, EIEC, EPEC, and ETEC, respectively. The Amp^r Chl^r Cxm^r Ctx^r Nal^s Cip^s Ipm^s Sxt^r pattern in EAEC strains and the Amp^r Chl^s Cxm^s Ctx^s Nal^s Cip^s Ipm^s Sxt^r pattern in EPEC strains were the second most prevalent multiresistance patterns with a prevalence of 17.4 and 18% in each category, respectively. Multiantibiotic resistance was detected in 89.5% of all diarrheagenic E. coli strains, 91.8% of EAEC strains, 100% of EIEC strains, 86% of EPEC strains, and 78.6% of ETEC strains. There was no significant difference in antibiotic resistance in diarrheagenic E. coli strains isolated from children with diarrhea compared to the strains from healthy control children (data not shown).

Twenty-eight Shigella strains were isolated from children with diarrhea only. AMP and SXT showed very low activity against Shigella strains with MIC₉₀s of >1,024 and >4/76 mg/liter, respectively. One strain was susceptible to all tested antibiotics. There were 22 of 28 (78.6%) multiantibiotic-resistant Shigella strains. One S. boydii strain was resistant to all tested antibiotics. The most common multiresistance pattern was Amp^r Chl^s Cxm^s Ctx^s Nal^s Cip^s Ipm^s Sxt^r with a prevalence of 35%.

Acute or chronic diarrhea caused by the different categories of E. coli is an emerging problem in many parts of the world (13). In developing countries, AMP, CHL, and SXT are widely used to treat diarrhea because of their low cost and ready availability. However, some previous studies have shown high prevalences of resistance to these antibiotics in enteric pathogens, especially diarrheagenic E. coli (8, 17). In our study, more than 77% of diarrheagenic E. coli strains were resistant to the commonly used antibiotics with high MIC₉₀s. The percentages of ETEC strains susceptible to AMP, CHL, and SXT were higher than those of other E. coli types. NAL, CIP, and IPM were the most active agents overall. Compared to other E. coli types, EAEC seemed highly resistant not only to AMP, CHL, and SXT but also to CXM, CTX, and NAL. According to our study, AMP, CHL, and SXT should not be used for the treatment of diarrhea in this population. Therefore, local information about antibiotic resistance should be used in clinical management, and treatment guidelines should be updated (23).

Although the cephalosporins (CXM and CTX) and IPM are not indicated to treat diarrhea, we have tested the susceptibilities of diarrheagenic E. coli strains to these antibiotics, since they could be empirically or incidentally used. According to the national household survey in Hanoi, Vietnam, in 1997 to 1998, on average, about two-thirds of those who are ill treated themselves (4). In our study, the results showed that CXM and CTX had moderate activity against E. coli strains. All E. coli strains were susceptible to IPM. The fluoroquinolone (CIP) and quinolone (NAL) antibiotics are now commonly used to treat infections, including diarrhea. They have also been recommended for prophylaxis and treatment of traveler's diarrhea. Compared to other studies (7, 22), our study showed a higher prevalence of resistance to NAL and CIP in diarrheagenic E. coli. We have isolated five E. coli strains (four EPEC strains and one ETEC strain) resistant to both NAL and CIP. If these antibiotics are used widely as the first choice of treatment of diarrhea, especially in developing countries, where the usage of antibiotics is not effectively controlled, a rapid emergence of antibiotic resistance most likely will occur.

Four categories of diarrheagenic E. coli had a high prevalence of multiresistance. The multiresistance to AMP, CHL, and SXT was the most prevalent in diarrheagenic E. coli strains. Multiresistance has been reported in previous studies (8, 11, 22). Sang et al. (20) described four cases of diarrhea caused by multiantibiotic-resistant EAEC. Moreover, these E. coli also had other multiresistance patterns with different prevalences. In particular, there were three EPEC strains and one ETEC strain resistant to all tested antibiotics except IPM. As mentioned, NAL and CIP can be used as alternative antibiotics in case the enteropathogens are resistant to the traditional antibiotics. Our study showed that 30 diarrheagenic E. coli strains (18.5%) were resistant to AMP, CHL, SXT, and either NAL or CIP. That means that the patients infected with these E. coli strains could risk a treatment failure. It is also indicated that the multiresistance of different categories of diarrheagenic E. coli strains is emerging in different developing countries where these antibiotics (both classical and new) have been widely used (7, 22).

Antibiotic treatment is indicated for dysentery caused by Shigella species because it can limit both the clinical course of illness and the duration of fecal excretion of the causative organism (18). However, Shigella bacteria are becoming increasingly resistant to most antibiotics commonly used in the treatment of diarrhea (8, 19). In the present study, the prevalences of resistance to SXT and AMP were similar to those of other studies in developing countries (1, 3, 8). In contrast, the study of Replogle et al. in United States showed lower prevalence of resistance of Shigella, with 59% resistant to SXT and 63% resistant to AMP (18). The lower resistance rates to these two antibiotics have also been reported in the study by Prats et al. in Spain (17). The resistance rates may be lower because of the more appropriate usage of antibiotics in developed countries compared to the developing countries.

Many studies have reported multiresistance in Shigella especially to AMP and SXT, which are commonly used to treat shigellosis (1, 11). In our study, 21 of 28 (75%) of Shigella strains were resistant to both AMP and SXT. Except for CHL that showed moderate activity with shigella, other antibiotics were still active against this pathogen. However, these antibiotics may appear active in vitro but may not be effective clinically. The wide usage of these antibiotics and the increasing prevalence of resistance are matters of concern to general practitioners and pediatricians. Nearly 60% of Shigella isolates in the study of Chu et al. in Hong Kong were Nal^r in the period of 1994 to 1995 (3). In the present study, we found one S. sonnei strain that was resistant to AMP, CHL, SXT, and NAL and one S. boydii strain that was resistant to all antibiotics tested. The progressive increase in antibiotic resistance among enteric pathogens, particularly in developing countries, is becoming a special concern. Of greatest immediate concern is the need for an effective, inexpensive antimicrobial agent that can be used safely for treatment of children with diarrhea, especially in developing countries, such as Vietnam.