Update on Shigella and Nontyphoidal Salmonella Antimicrobial Drug Resistance: Implications on Empirical Treatment of Acute Infectious Diarrhea in Cambodia

ABSTRACT

Information on causative diarrheal pathogens and their associated antimicrobial susceptibility remains limited for Cambodia. This study describes antimicrobial resistance patterns for Shigella and nontyphoidal Salmonella isolates collected in Cambodia over a 5-year period. Multidrug resistance was shown in 98% of Shigella isolates, with 70%, 11%, and 29% of isolates being resistant to fluoroquinolones, azithromycin, and cephalosporin, respectively. As many as 11% of Shigella isolates were resistant to nearly all oral and parenteral drugs typically used for shigellosis, demonstrating extreme drug resistance phenotypes. Although a vast majority of nontyphoidal Salmonella isolates remained susceptible to cephalosporins (99%) and macrolides (98%), decreased susceptibility to ciprofloxacin was found in 67% of isolates, which is notably higher than previous reports. In conclusion, increasing antimicrobial resistance of Shigella and nontyphoidal Salmonella is a major concern for selecting empirical treatment of acute infectious diarrhea in Cambodia. Treatment practices should be updated and follow local antimicrobial resistance data for the identified pathogens.

INTRODUCTION

The treatment of acute infectious diarrhea usually relies on empirical antibiotics because the results of stool culture and susceptibility data are not readily available to guide initial treatment decisions. However, the efficacy of this practice may be limited by variation in antimicrobial susceptibility patterns across regions related to selective pressure from prescribing. Information on the local resistance patterns must be taken into account to aid in the selection of best antimicrobial agents for patients presenting with acute infectious diarrhea. In 2018, the World Health Organization (WHO) published the first global list of antibiotic-resistant “priority pathogens,” with 12 families of bacteria identified as posing the greatest threat to human health. Salmonella spp. and Shigella spp. are included in the high and medium threat levels, respectively, and account for a large proportion of diarrheal cases worldwide (1). Similarly, in 2019, the Centers for Disease Control and Prevention (CDC) reported that antibiotic resistance is one of the greatest threats facing the United States (2). The report classified 18 antibiotic-resistant bacteria and fungi into three threat categories based on risk to human health, including urgent, serious, and concerning threats. Included among serious threats are drug-resistant nontyphoidal Salmonella and Shigella (2), with both showing increased resistance to the fluoroquinolone class of antibiotics that are commonly used as first-line treatment in acute infectious diarrhea. Although most enteric infections are self-limiting and antibiotic treatment is typically not required, antibiotics have shown benefit in shortening the duration and reducing the severity of diarrheal illness in patients infected with Shigella (3) and in a subset of patients with nontyphoidal Salmonella who have predisposing factors for invasive disease, to include pediatric and geriatric populations.

In Cambodia, diarrheal disease is a significant public health concern, with a documented rise in drug resistance to commonly prescribed antibiotics. Currently, Salmonella spp., Shigella spp., Vibrio cholerae, rotavirus, and enterotoxigenic Escherichia coli (ETEC) are the most common agents responsible for diarrhea-associated mortality among children under 5 years old (4). A study published in 2016 on drug resistance of Shigella revealed a decreased efficacy of multiple classes of antibiotics, including fluoroquinolones, cephalosporins, and azithromycin (AZM) (5). These are frequently prescribed for acute diarrhea as first-line agents. This has raised concerns about the optimal empirical treatment options in the areas where susceptibility testing is not available (5) and where the drug choices are increasingly more limited. There is a critical need for the latest information regarding the causative organisms responsible for acute diarrhea in Cambodia and their associated antimicrobial susceptibility patterns (6). This report study describes the antimicrobial resistance patterns of Shigella and nontyphoidal Salmonella isolated in Cambodian stool samples as part of a 5-year prospective surveillance study. This information will aid in selecting appropriate antimicrobial therapies for treatment of acute infectious diarrhea originating from this region.

RESULTS

A total of 1,985 stool samples were collected from 1,061 Cambodian children and 924 adult military personnel during July 2014 to June 2019. Of these, 1,098 samples were from acute diarrheal cases and 887 samples were from nondiarrheal controls. Isolation of Shigella spp. from stool specimens was more common for children (49/1,061 [5%]) than adults (6/924 [1%]) (P < 0.01). Subjects with Shigella infections were 2.97 times more likely to have acute diarrhea (odds ratio [OR] = 2.97; 95% confidence interval [CI] = 1.55 to 5.67). Serotype distribution of 55 Shigella isolates showed that 62% (34/55) were Shigella flexneri and 38% (21/55) were Shigella sonnei isolates. S. flexneri 3a was the most common serotype, detected in 27% (15/55) of all Shigella isolates (Fig. 1). No Shigella dysenteriae or Shigella boydii isolates were detected during the study period.

FIG 1.

Serotype distribution of Shigella species isolated from stool samples in Battambang, Cambodia, 2014 to 2019 (n = 55).

Information on antibiotics used for diarrhea episodes in symptomatic subjects was also collected. Among 297 acute diarrheal cases with Salmonella and/or Shigella, only 15 subjects reported receiving previous antibiotic treatment prior to seeking medical care. Beta-lactam antibiotics were the most commonly reported antimicrobials, including amoxicillin (n = 6), amoxicillin-clavulanic acid (n = 1), cefixime (n = 1), and ceftriaxone (CRO) plus gentamicin (n = 1). Ciprofloxacin (CIP) treatment was reported for 3 subjects, metronidazole for 2 subjects, and trimethoprim-sulfamethoxazole (SXT) for one subject.

The percentages of Shigella isolates with confirmed resistance to AMP, SXT, TET, nalidixic acid (NAL), CIP, CRO, and AZM were 85%, 96%, 98%, 96%, 70%, 29%, and 11%, respectively. Only 15% of Shigella isolates were fully susceptible to CIP. NAL susceptibility was found only in the 2 isolates (100%) of S. flexneri 6. Overall, the percentage of resistant isolates was significantly higher among isolates of S. sonnei than of S. flexneri, especially for ciprofloxacin (100% versus 53%; P < 0.01) and cephalosporins (57% versus 12%; P < 0.01). For AZM, a similar trend was observed, but the difference between S. sonnei versus S. flexneri was not statistically significant (19% versus 6%; P = 0.128).

Resistance patterns of Shigella isolates are described in Table 1. The most common resistance patterns in S. flexneri are AMP-SXT-TET-NAL (14/55 [25%]) and AMP-SXT-TET-NAL-CIP (14/55 [25%]). The AMP-SXT-TET-NAL-CIP-CRO-cefotaxime (CTX) resistance pattern was most common in S. sonnei (9/55 [16%]), followed by the SXT-TET-NAL-CIP resistance pattern (7/55 [13%]). As many as 11/13 S. flexneri 3a isolates (73%) shared the same AMP-SXT-TET-NAL-CIP pattern of resistance, and 9/21 S. sonnei isolates (43%) had the same AMP-SXT-TET-NAL-CIP-CRO-CTX pattern of resistance. These isolates are unlikely to be from a single outbreak since they originated from 3 different hospitals spanning a period of more than 2 years.

TABLE 1.

Antimicrobial resistance patterns of 55 Shigella isolates in Battambang, Cambodia, 2014 to 2019

Resistance pattern	Antimicrobial resistance patterna	No. of Shigella isolatesb
		Total (n = 55)	SHF1a (n = 3)	SHF2a (n = 5)	SHF2v (n = 9)	SHF3a (n = 15)	SHF6(n = 2)	SHS (n  = 21)
A	AMP-SXT-TET-NAL-CIP-CRO-CTX-AZM	4				2		2
B	AMP-SXT-NAL-CIP-CRO-CTX-AZM	1						1
C	AMP-SXT-TET-NAL-CIP-AZM	1						1
D	AMP-SXT-TET-NAL-CIP-CRO-CTX	9						9
E	AMP-TET-NAL-CIP_CRO-CTX	2	1		1
F	AMP-SXT-TET-NAL-CIP	14		1	1	11		1
G	SXT-TET-NAL-CIP	7						7
H	AMP-SXT-TET-CIP	1					1
I	AMP-SXT-TET-NAL	14	1	4	7	2
J	AMP-SXT-TET	1					1
K	TET-NAL	1	1

^aAbbreviations for antimicrobial agents: AZM, azithromycin; NAL, nalidixic acid; CIP, ciprofloxacin; AMP, ampicillin; SXT, trimethoprim-sulfamethoxazole; TET, tetracycline; CRO, ceftriaxone; CTX, cefotaxime; CAZ, ceftazidime.

^bAbbreviations for organisms: SHF1a, Shigella flexneri 1a; SHF2a, Shigella flexneri 2a; SHF2v, Shigella flexneri 2 variant; SHF3a, Shigella flexneri 3a; SHF6, Shigella flexneri 6; SHS, Shigella sonnei.

All but 1 of 55 Shigella isolates (identified as S. flexneri 1a) in this study met the criteria for multidrug resistance (MDR) to three or more antimicrobial classes. The MICs of 13 CIP-susceptible, 8 CIP-intermediate, and 39 CIP-resistant Shigella isolates as determined by disk diffusion testing were 0.064 to 0.25 μg/ml, 0.12 to 0.25 μg/ml, and 2 to 8 μg/ml, respectively. Decreased susceptibility to AZM and MIC epidemiological cutoff values (ECVs) of ≥16 μg/ml for S. flexneri and ≥32 μg/ml for S. sonnei were noted in 11% (6/55) of Shigella isolates (2 S. flexneri 3a and 4 S. sonnei isolates). Additional information for the other antimicrobials tested is shown in Fig. 2. The 6 AZM-non-wild type (NWT) isolates had a wide range of MICs, 16 to ≥256 μg/ml. All AZM-NWT isolates showed coresistance to CIP. Moreover, 4 of these 6 isolates showed resistance to 8 other antimicrobial agents tested in this study, which are also commonly used antibiotics in patients with diarrhea.

FIG 2.

Patterns of resistance of S. sonnei and S. flexneri isolates to antimicrobial agents. The percentages of isolates that were characterized as resistant to the available antibiotics are shown. A high proportion of S. sonnei isolates resistant to the first-line treatments (AZM, CIP, CRO, and CTX) was demonstrated. Among all antimicrobials tested against Shigella isolates, AZM showed the greatest success rate in drug sensitivity profiles. Abbreviations for antimicrobial agents: AZM, azithromycin; NAL, nalidixic acid; CIP; ciprofloxacin; AMP, ampicillin; SXT, trimethoprim-sulfamethoxazole; TET, tetracycline; CRO, ceftriaxone; CTX, cefotaxime; CAZ, ceftazidime.

Overall, 507 of 1,985 (26%) stool samples were culture positive for nontyphoidal Salmonella. Salmonella spp. were detected in both nondiarrheal controls (243/887 [27%]) and diarrhea cases (264/1,098 [24%]) in similar percentages (OR = 0.84; 95% CI = 0.69 to 1.03; P = 0.09). Salmonella spp. were detected more commonly in adults (299/924 [32%]) than children (208/1,061 [20%]) (P < 0.01). Some stool samples had more than one serogroup of nontyphoidal Salmonella identified, making a total of 642 nontyphoidal Salmonella isolates identified in 507 positive stool samples. The most prevalent serogroups were group C (281/642 [44%]), B (165/642 [26%]), and E (112/642 [18%]). Other serogroups found were group D (26/642 [4%]), group I (22/642 [3%]), and groups G, K, O (36/642 [6%]). The data showed no significant correlation between diarrheal and nondiarrheal samples for specific serogroups or presence of multiple serogroups in a single patient.

Among 642 Salmonella isolates, serogroups B and C showed the highest proportion of samples with resistance to common antimicrobials (Fig. 3). CIP-resistant Salmonella, a resistance pattern included in the high priority list by the WHO, was predominantly in serogroups C (20%) and E (8%). Out of 66 CIP-resistant Salmonella isolates, 27 were from patients who had acute diarrhea symptoms, while others originated from asymptomatic carriers. The resistance rates between the symptomatic and asymptomatic Salmonella isolates were similar for all antimicrobials tested except for NAL (P = 0.0074), with higher rates of resistance demonstrated for AMP, SXT, and TET in children (P = <0.001).

FIG 3.

Percent resistance to antimicrobial agents of Salmonella species by serogroup. Abbreviations for organisms: SALB, Salmonella group B; SALC, Salmonella group C; SALD, Salmonella group D; SALE, Salmonella group E; Other Sal, Salmonella spp. other than group A, B, C, D, or E.

Although CIP resistance was found in 10% (66/642) of all Salmonella isolates, 56% (361/642) of all isolates showed intermediate susceptibility. Therefore, as many as 67% of nontyphoidal Salmonella cases were classified as having decreased susceptibility to CIP, increasing concerns for rapid rise in fluoroquinolone resistance in Cambodia. Among the 66 CIP-resistant nontyphoidal Salmonella isolates, 94% were susceptible to cephalosporins, and 98% were susceptible to AZM. Resistance to cephalosporins was demonstrated in only serogroup B and C isolates (4/165 [2.4%] and 4/281 [1.4%], respectively). Resistance to AZM was infrequent and found in 13/642 (2%) isolates, all in serogroup C. The MICs of AZM were determined for 4 of these resistant isolates and ranged from 64 to 128 μg/ml. Resistance patterns of Salmonella isolates are described in Table 2.

TABLE 2.

Antimicrobial resistance patterns of 55 Salmonella isolates in Battambang, Cambodia, 2014 to 2019

Resistance pattern	Antimicrobial resistance pattern	No. of organismsa
		Total (n = 642)	SALB (n = 165)	SALC (n = 280)	SALD (n = 26)	SALE (n = 112)	Other SAL (n = 59)
A	AMP-SXT-TET-NAL-CIP-CRO-CTX-CAZ-AZM	2	0	2	0	0	0
B	AMP-SXT-TET-CIP-CRO-CTX-CAZ	1	0	1	0	0	0
C	AMP-SXT-TET-CRO-CTX-CAZ	1	0	1	0	0	0
D	AMP-TET-CRO-CTX-CAZ	4	4	0	0	0	0
E	AMP-SXT-TET-NAL-CIP	5	0	1	0	4	0
F	AMP-SXT-NAL-CIP	5	0	0	0	5	0
G	AMP-TET-NAL-CIP	47	0	47	0	0	0
H	AMP-SXT-TET-AZM	9	0	9	0	0	0
I	AMP-SXT-TET-NAL	8	3	3	0	2	0
J	SXT-TET-AZM	1	0	1	0	0	0
K	AMP-TET-CIP	1	0	1	0	0	0
L	AMP-SXT-NAL	2	0	0	0	2	0
M	AMP-SXT-TET	94	38	43	1	12	0
N	AMP-TET-NAL	1	0	1	0	0	0
O	TET-NAL-CIP	2	0	2	0	0	0
P	SXT-TET-NAL	1	1	0	0	0	0
Q	TET-AZM	1	0	1	0	0	0
R	AMP-CIP	1	0	0	0	0	1
S	NAL-CIP	1	1	0	0	0	0
T	AMP-SXT	1	0	0	1	0	0
U	AMP-TET	59	22	14	1	19	3
V	AMP-NAL	11	1	5	3	0	2
W	SXT-TET	6	2	4	0	0	0
X	TET-NAL	21	5	16	0	0	0
Y	CIP	1	0	1	0	0	0
Z	NAL	24	3	17	1	3	0
AA	TET	49	24	21	0	3	1
AB	AMP	17	5	5	1	50	6
AC	No resistance	266	56	84	18	62	46

^aAbbreviations for organisms: SALB, Salmonella group B; SALC, Salmonella group C; SALD, Salmonella group D; SALE, Salmonella group E; Other Sal, Salmonella spp. other than group A, B, C, D, or E.

AZM resistance was infrequent among Salmonella (2%) and Shigella (11%) samples. Conversely, CIP resistance was exceedingly high among Shigella (70%) isolates. Resistance to CIP with decreased susceptibility to AZM was demonstrated among 11% of Shigella isolates and 0.3% of nontyphoidal Salmonella isolates. Resistance to three or more classes of antimicrobials was demonstrated in 54/55 (98%) and 181/642 (28%) of Shigella and Salmonella isolates, respectively. For the most common Salmonella serogroups, MDR was found in 39% (109/281) of group C, 28% (46/165) of group B, and 22% (25/112) of serogroup E isolates. Extended-spectrum-beta-lactamase (ESBL) production was present in approximately 29% of Shigella samples and 1% of Salmonella isolates. Overall, the resistance to CIP and AZM as well as the percentage of isolates with MDR and ESBL was more common among Shigella isolates than Salmonella.

DISCUSSION

Shigella infections are a common cause of diarrheal illness in Cambodia (7–9), with recent observation of rising drug resistance in the region. In this study, 7% of children with acute diarrhea had symptoms attributed to Shigella gastroenteritis. Based on prior reports, S. flexneri is more common in the developing world, while S. sonnei is more commonly found in industrialized countries or economically developed regions (10–13). This is consistent with the results of this study in Cambodia, where S. flexneri accounted for the majority of Shigella infections.

Access to up-to-date information on antimicrobial resistance for diarrheal pathogens circulating in Cambodia (6) remains limited, especially for Shigella spp. In addition, limited information is available on drug resistance among the military populations that participated in this study. The results will be useful for guiding empirical therapy, particularly with regard to Shigella spp. This report shows that in Cambodia, several types of Shigella continue to be in circulation, with S. flexneri being the predominant type and less commonly associated with resistance to the recommended first-line treatments (fluoroquinolones, azithromycin, and third-generation cephalosporin antimicrobials). However, a high proportion of resistant isolates was identified among S. sonnei samples in the study population.

Based on previously published reports for samples collected from 2001 to 2014 from Asia, Africa, and South America, Shigella serogroups were commonly resistant to AMP but still susceptible to CIP (10). These findings were similar to data reported in Cambodia by Meng et al. in 2011 that showed no resistance to CIP and AZM, while 78% of Shigella isolates were resistant to AMP and almost 100% resistant to SXT and TET (7). In the current study, conducted 10 years later, there was a remarkable rise in antimicrobial resistance from previously published reports. We observed a significant proportion of Shigella samples resistant to CIP (70%), cephalosporins (29%), and AZM (11%), with all S. sonnei samples (100%) being resistant to CIP and 24% (8/34) of all S. flexneri isolates showing decreased susceptibility to CIP. Whereas S. flexneri isolates were largely susceptible to third-generation cephalosporins, more than 50% of the S. sonnei isolates showed resistance to CRO and CTX. In addition, ESBL production was detected by phenotypic screening in 29% of Shigella samples, whereas previous studies failed to identify ESBL-producing bacteria among the 23 isolates of Shigella and Salmonella that were collected in Cambodia from 2012 to 2015 (14). In another report from Vietnam and Laos, where Shigella-associated dysentery is endemic, multidrug resistance was uncommon, with 3.3% of S. flexneri and 5.4% of S. sonnei isolates resistant to AZM and 2% of S. flexneri and 30% of S. sonnei isolates demonstrating resistance to CRO (15). The majority of Shigella isolates (89%) were still susceptible to AZM in this study. However, the rates of AZM-resistant NWT isolates (13% of S. flexneri 3a and 19% of S. sonnei isolates) were exceptionally high. A rising rate of resistance was also observed for CRO, with 12% of S. flexneri and 57% of S. sonnei isolates meeting criteria for drug resistance. An even higher proportion of samples showed resistance to CIP (70%). These observations require reevaluation of the empirical treatments that are being recommended in Cambodia and to those traveling in the region. AZM or CRO may be more effective for empirical therapy of acute diarrhea in Cambodia in selected patients, especially if pathogen class is known. For gastroenteritis due to Shigella, AZM is more likely to be an effective first-line treatment than CIP. The use of CIP is anecdotally common in Cambodia, as the drug is easily accessible, while AZM usage is very low due to high costs and likely explains the preserved susceptibility of AZM in our study population.

In the United States, where AZM and CIP are the two most commonly prescribed antibiotics for shigellosis, CIP resistance (6%) is much lower than what was observed in this study, and yet antibiotic-resistant Shigella has been reported as a serious threat (2, 16). Due to this concern, in April 2017, the CDC released a health advisory recommending against the use of fluoroquinolones for shigellosis if the CIP MIC is ≥0.12 μg/ml. All but one of the Shigella isolates (98%) in our study had CIP MICs of ≥0.12 μg/ml, making fluoroquinolone unsuitable for treatment of shigellosis in Cambodia based on the CDC advisory. In addition, due to high rate of CIP resistance in Shigella isolates in Cambodia, all travelers to Cambodia should be aware of the high risk of treatment failure with CIP for acute diarrhea. Alternatives such as AZM should be considered for patients with dysentery or suspected Shigella gastroenteritis. Based on published data, AZM should also be very effective for fluoroquinolone-resistant Campylobacter and other bacterial causes such as enteroinvasive E. coli, Aeromonas spp., Plesiomonas spp., and Yersinia enterocolitica (17).

The high rate of MDR Shigella isolates (98%) limits the choice of treatments. Even more worrisome, 6 isolates (11%) in the present study (4 S. sonnei and 2 S. flexneri 3a isolates) met the criteria for extremely drug-resistant (XDR) strains of Shigella (18, 19), defined by resistance to AMP, SXT, and CIP and decreased susceptibility to AZM. Among these samples, 5 isolates (83%) were ESBL producers, indicating their decreased likelihood of response to cephalosporin treatment as well. A similar study from neighboring Laos and Vietnam (15) reported much lower rates of MDR Shigella than in samples collected in Cambodia, underscoring the importance of characterizing local antimicrobial resistance patterns for guiding treatment decisions. Treatment recommendations for acute diarrhea must be guided by local drug resistance patterns whenever data are available. Travel clinics must consider new and emerging data on antimicrobial resistance when prescribing empirical treatments. Although this was not investigated in this study, in addition to the risk of treatment failure, the rise of antimicrobial resistance has the potential to result in longer pathogen shedding times and contribute to transmission of resistant strains.

With regard to Salmonella spp., the Foodborne Disease Burden Epidemiology Reference Group established by the WHO reported nontyphoidal Salmonella as one of the leading causes of mortality from diarrheal illness worldwide (20). Its broad-host-range characteristics and potential to cause invasive extraintestinal infection in high-risk, immunocompromised patients, infants, and the elderly also lead to increased public health concern (21, 22). The emergence of invasive nontyphoidal Salmonella strains that are resistant to commonly used antimicrobial agents, including third-generation cephalosporins and fluoroquinolones, has been reported in Asia, the United States, and Africa (23–25). In this study, resistance to NAL (20%), CIP (10%), and AMP (42%) was higher than in other areas (2, 25). In addition, an increase in the rate of resistance among nontyphoidal Salmonella isolates against fluoroquinolones was noted. This troubling trend has also been reported worldwide (2, 6, 25, 26). In a previous study in Cambodia with samples collected from 2004 to 2006 (7), only 1% of the nontyphoidal Salmonella isolates were resistant to CIP, compared to 10% in this study. Moreover, 56% of Salmonella isolates (361/642) were categorized by the Clinical and Laboratory Standards Institute (CLSI) as having intermediate susceptibility, with a zone diameter of 21 to 30 mm. Continued use of CIP for the treatment of infections caused by intermediate-susceptible organisms may drive an emergence of high-level resistant strains (25) and result in prolonged clinical illness or higher rate of treatment failures (16). Fortunately, the isolates with concurrent resistances to both fluoroquinolones and third-generation cephalosporins are uncommon (3/642 [0.5%]) and similar to what has been reported in the United States and Europe (2, 26).

Many travelers to Cambodia and elsewhere rely on AZM as empirical treatment for their diarrheal illness (3). Ongoing monitoring of drug susceptibility is needed due to frequent use of antibiotics in Cambodia (27). The common practice of self-prescribing of antibiotics is likely contributing to the rise of drug resistance, but antimicrobial usage practices in Southeast Asia are not well characterized and require further study.

Conclusions.

The increasing rate of antimicrobial resistance to fluoroquinolones, macrolides, and third-generation cephalosporins is of concern. Recent travel to Southeast Asia should be considered in the clinical management of infectious diarrhea. In this report, the MDR and XDR of Shigella and decreased susceptibility of nontyphoidal Salmonella isolates to various antibiotic classes have been described. These results can be used to aid policy makers and local health care providers to develop best clinical practice guidelines for empirical therapy of acute infectious diarrhea. Future assessments should focus on delineating the risk factors for contracting a diarrheal illness with the resistant strains and how to limit its spread. In addition, antimicrobial stewardship programs at hospitals and clinics can help curb the current trends. It is unclear how current prescribing practices in Cambodia affect the disease course of diarrheal illness in the context of rising drug resistance, and further study may be helpful. The selection of preferred antibiotic treatments should be region and country specific based on observed differences in drug resistance. Patients with diarrheal illness that require antibiotic treatment should have stool culture and drug susceptibility testing performed due to the high rate of MDR in Cambodia. When susceptibility testing is not available, the choice of antimicrobial agents should be guided by local resistance patterns and clinical follow-up. The observations made in this study underscore the importance of judicious use of antimicrobial agents in clinical practice to help curb the worrisome trends in antimicrobial drug resistance.

MATERIALS AND METHODS

The human use protocol was reviewed and approved by the Cambodia National Ethics Committee for Health Research (NECHR) and the Walter Reed Army Institute of Research (WRAIR) Institutional Review Board. Subjects or subjects’ guardians provided written informed consent prior to participation.

Source of bacterial isolates.

Bacterial isolates were obtained from a prospective surveillance study in which stool samples were collected from Cambodian children aged 3 months to 5 years and adult military personnel aged 18 to 60 years with and without acute diarrhea as cases and controls, respectively. Stool samples were submitted for culture, identification, and antimicrobial susceptibility testing. The study was conducted at the Battambang Referral Hospital (BRH), Svay Por Health Center (SPHC), and Military Regional 5 Hospital (MR5H) in Battambang, Cambodia, from July 2014 to June 2019.

Antimicrobial susceptibility.

All stool samples were initially examined for enteric pathogens by standard bacteriological methods as previously described (7). Antimicrobial susceptibility testing was performed on all bacterial species isolated from collected specimens. Standard disk diffusion technique was applied to nine antimicrobial agents (Becton, Dickinson, Sparks, MD), including ampicillin (AMP), AZM, cefotaxime (CTX), ceftazidime (CAZ), ceftriaxone (CRO), ciprofloxacin (CIP), nalidixic acid (NAL), trimethoprim-sulfamethoxazole (SXT), and tetracycline (TET). Results were interpreted following Clinical and Laboratory Standards Institute (CLSI) guidelines as summarized in Table S1 (28). Zone diameter interpretive standards for Enterobacterales were used for all antimicrobials tested when available. Except for AZM, CLSI clinical breakpoints are established only for Salmonella enterica serovar Typhi and thus were applied for all Salmonella isolates in this study. The CLSI AZM epidemiological cutoff values (ECVs) were used to determine S. sonnei and S. flexneri as non-wild type (NWT) or wild type (WT) because clinical breakpoints for AZM have not been established for Shigella. Escherichia coli ATCC 25922 and Staphylococcus aureus ATCC 25923 were used for quality control. Additionally, a subset of isolates were tested for MIC using Etest (bioMérieux SA, France). In this case, Staphylococcus aureus ATCC 29213 was used for quality control. Multidrug resistance (MDR) was defined as resistance to at least one agent in three or more classes of antimicrobial agents as previously described (18). Extreme drug resistance (XDR) in Shigella was defined as resistance to AMP, SXT, and CIP and decreased susceptibility to AZM (19).

Testing for ESBL production.

Phenotypic detection of extended-spectrum beta-lactamases (ESBLs) was performed by using the standard methodology from CLSI guidelines (28). ESBL phenotypic confirmatory testing by a standard disk diffusion procedure using beta-lactam combination agents (both cefotaxime and ceftazidime, alone and in combination with clavulanate) was conducted in parallel with the commercial Microscan Walkaway Plus system (Beckman Coulter, West Sacramento, CA). Klebsiella pneumoniae ATCC 700603 and Escherichia coli ATCC 25922 were used for quality control.

Statistical analysis.

The chi-square and Fisher’s exact tests were used to analyze the differences between cases and controls or children and adults groups. Differences in antimicrobial resistance rate between serogroups were analyzed using SPSS 25; a P value of <0.05 was considered statistically significant.