Shigella spp. Antimicrobial Drug Resistance, Papua New Guinea, 2000–2009

To the Editor: Approximately half the Shigella spp. infections in developing countries are caused by endemic shigellae (1), which in these countries are responsible for ≈10% of all episodes of diarrhea among children <5 years of age and up to 75% of deaths from diarrhea (2). Deaths from epidemic Shigella spp. in the community are estimated to outnumber deaths within the healthcare setting. In Papua New Guinea, diarrhea is a major cause of hospital admission and death (3); Shigella spp. are among the most common causes of enteric bacterial infection (4,5), and S. flexneri is the most common serotype (3,6). Outbreaks of bloody diarrhea are frequently reported; however, diagnosis in remote settings is challenging, partly because the storage requirements for the organism are difficult to meet. 
 
Multidrug resistance of shigellae is not new (1); many countries have reported resistance to amoxicillin, co-trimoxazole, and chloramphenicol. For this reason, the World Health Organization recommends that all patients with bloody diarrhea be treated with either ciprofloxacin or 1 of the 3 second-line drugs: pivmecillinam, azithromycin, and ceftriaxone (7). The antimicrobial drug currently recommended for patients with bloody diarrhea in primary healthcare settings in Papua New Guinea is co-trimoxazole (8); ciprofloxacin is available only in hospitals. 
 
In August 2009, an epidemic of multidrug-resistant S. flexneri infection associated with widespread illness and death across 4 provinces of Papua New Guinea was reported to health authorities. To understand the trends and to inform antimicrobial drug policy makers, we reviewed retrospective microbiological data for 2000–2009. With the exception of 3 isolates collected during an outbreak in the border regions of the 4 provinces during 2009 (excluded from analysis), all isolates in our study were obtained as part of routine surveillance. Fecal samples were collected by clinicians from any patient seeking care for severe diarrhea at Port Moresby General Hospital. 
 
Before serologic testing was conducted, samples were spread directly on desoxycholate citrate agar and MacConkey agar plates for culture. Antimicrobial drug resistance testing was performed by using the Kirby-Bauer method. 
 
From a total of 3,419 fecal samples cultured, 136 (4.0%) were positive for Shigella spp. The most commonly isolated species was S. flexneri (90.4%); less frequently isolated were S. boydii (3.7 %), S. dysenteriae (2.9%), and S. sonnei (1.5%). Of the 123 S. flexneri isolates, 20 (16%) were further characterized; the most frequent serovars were serovar 2 (40%) and serovar 3 (30%). Many (48%) Shigella spp.–positive isolates were from children <5 years of age. The highest rates of antimicrobial drug resistance of all Shigella spp. were to amoxicillin (96%), co-trimoxazole (86%), and chloramphenicol (60%); no resistance to ciprofloxacin and cephalexin was found (Table). 
 
 
 
Table 
 
Antimicrobial drug resistance of Shigella spp., Papua New Guinea, 2000–2009* 
 
 
 
Current evidence supports the use of ciprofloxacin, ceftriaxone, and pivmecillinam for treatment of bloody diarrhea (9). It also suggests that dysentery rarely relapses if an infected child has received a full course of treatment with 1 of these drugs and the causative pathogen is sensitive to the drug. Reducing the risk for relapse of bacterial infections among children is beneficial because it reduces the likelihood of subsequent episodes of dysentery occurring in that child and of transmission to others (9). In our study, most isolates were resistant to co-trimoxazole and the other available antimicrobial drugs, indicating that their use would not have reduced illness and subsequent transmission in this setting. The lack of resistance to ciprofloxacin and cephalexin indicates that these drugs may be more effective; however, they are neither available at the primary healthcare level nor recommended in Papua New Guinea, which is cause for concern. 
 
Surveillance for antimicrobial drug resistance is essential for the containment of antimicrobial drug resistance globally. However, international surveillance depends on strong national surveillance systems. Despite the existence of a network of subnational laboratories where fecal sample cultures had been performed, these laboratories no longer perform these cultures. In 1964, the laboratory in 1 provincial hospital analyzed and subtyped 1,000 stool samples over a 15-month period (6). In our study, conducted at the national referral hospital (which limits the representativeness), we analyzed 3,419 fecal samples over a 10-year period. 
 
Outbreaks of bloody diarrhea are common in remote settings in Papua New Guinea, yet with the exception of the 3 isolates from 2009 that were excluded from analysis, no Shigella spp.–positive samples have been identified during outbreaks. Molecular methods may serve as an adjunct to traditional laboratory methods by improving sensitivity and also enabling diagnosis of Shigella spp. outbreaks among remote populations where specimen storage and transport requirements may be challenging (10). 
 
We describe extremely high rates of resistance of Shigella spp. to co-trimoxazole, the recommended treatment for bloody diarrhea in Papua New Guinea. Strengthening national surveillance for antimicrobial drug resistance would provide the evidence to better inform policy decision makers. A review of the national antimicrobial drug policy for management of bloody diarrhea is urgently needed.

More effort is needed in Africa to enable reliable and standardized laboratory diagnoses of Salmonella infections and to sustain TF surveillance and drug sensitivity surveys. Moreover, introduction of a vaccination program should be discussed after more data are obtained from other areas in Ghana and West Africa. Such data currently are collected in an extensive standardized surveillance program across the continent performed by our group and others. In parallel, trials should be conducted to assess the effectiveness and cost-effectiveness of currently available and newly developed TF vaccines.  (1), which in these countries are responsible for ≈10% of all episodes of diarrhea among children <5 years of age and up to 75% of deaths from diarrhea (2). Deaths from epidemic Shigella spp. in the community are estimated to outnumber deaths within the healthcare setting. In Papua New Guinea, diarrhea is a major cause of hospital admission and death (3); Shigella spp. are among the most common causes of enteric bacterial infection (4,5), and S. fl exneri is the most common serotype (3,6). Outbreaks of bloody diarrhea are frequently reported; however, diagnosis in remote settings is challenging, partly because the storage requirements for the organism are diffi cult to meet.
Multidrug resistance of shigellae is not new (1); many countries have LETTERS reported resistance to amoxicillin, cotrimoxazole, and chloramphenicol. For this reason, the World Health Organization recommends that all patients with bloody diarrhea be treated with either ciprofl oxacin or 1 of the 3 second-line drugs: pivmecillinam, azithromycin, and ceftriaxone (7). The antimicrobial drug currently recommended for patients with bloody diarrhea in primary healthcare settings in Papua New Guinea is co-trimoxazole (8); ciprofl oxacin is available only in hospitals.
In August 2009, an epidemic of multidrug-resistant S. fl exneri infection associated with widespread illness and death across 4 provinces of Papua New Guinea was reported to health authorities. To understand the trends and to inform antimicrobial drug policy makers, we reviewed retrospective microbiological data for 2000-2009. With the exception of 3 isolates collected during an outbreak in the border regions of the 4 provinces during 2009 (excluded from analysis), all isolates in our study were obtained as part of routine surveillance. Fecal samples were collected by clinicians from any patient seeking care for severe diarrhea at Port Moresby General Hospital.
Before serologic testing was conducted, samples were spread directly on desoxycholate citrate agar and MacConkey agar plates for culture.
Antimicrobial drug resistance testing was performed by using the Kirby-Bauer method.
Current evidence supports the use of ciprofl oxacin, ceftriaxone, and pivmecillinam for treatment of bloody diarrhea (9). It also suggests that dysentery rarely relapses if an infected child has received a full course of treatment with 1 of these drugs and the causative pathogen is sensitive to the drug. Reducing the risk for relapse of bacterial infections among children is benefi cial because it reduces the likelihood of subsequent episodes of dysentery occurring in that child and of transmission to others (9). In our study, most isolates were resistant to co-trimoxazole and the other available antimicrobial drugs, indicating that their use would not have reduced illness and subsequent transmission in this setting. The lack of resistance to ciprofl oxacin and cephalexin indicates that these drugs may be more effective; however, they are neither available at the primary healthcare level nor recommended in Papua New Guinea, which is cause for concern.
Surveillance for antimicrobial drug resistance is essential for the containment of antimicrobial drug resistance globally. However, international surveillance depends on strong national surveillance systems. Despite the existence of a network of subnational laboratories where fecal sample cultures had been performed, these laboratories no longer perform these cultures. In 1964, the laboratory in 1 provincial hospital analyzed and subtyped 1,000 stool samples over a 15month period (6). In our study, conducted at the national referral hospital (which limits the representativeness), we analyzed 3,419 fecal samples over a 10-year period.
Outbreaks of bloody diarrhea are common in remote settings in Papua New Guinea, yet with the exception of the 3 isolates from 2009 that were excluded from analysis, no Shigella spp.-  positive samples have been identifi ed during outbreaks. Molecular methods may serve as an adjunct to traditional laboratory methods by improving sensitivity and also enabling diagnosis of Shigella spp. outbreaks among remote populations where specimen storage and transport requirements may be challenging (10). We describe extremely high rates of resistance of Shigella spp. to cotrimoxazole, the recommended treatment for bloody diarrhea in Papua New Guinea. Strengthening national surveillance for antimicrobial drug resistance would provide the evidence to better inform policy decision makers. A review of the national antimicrobial drug policy for management of bloody diarrhea is urgently needed.

Fatal Avian Infl uenza (H5N1) Infection in Human, China
To the Editor: Since the fi rst avian infl uenza virus (H5N1) was isolated from a goose in the southern region of the People's Republic of China a decade ago (1), no poultry outbreak has been reported in Shandong Province in eastern China, although adjacent provinces have experienced an avian infl uenza epidemic (2). In fall 2008, several rounds of investigation of poultry farms and markets were conducted in Jinan, Shandong Province, and no infl uenza virus (H5N1) was isolated by reverse transcription-PCR (RT-PCR) from 19,340 poultry oropharyngeal, cloacal, and cage specimens.
However, a fatal infl uenza (H5N1) infection in a human was identifi ed on January 17, 2009 (3). The patient was a 27-year-old woman from Jinan. Infl uenza-like illness (ILI) developed on January 5, and the patient received intravenous ribavirin and cephalosporins on January 9. On January 11, she was hospitalized for fever (41°C) and respiratory symptoms. On January 15, extensive infi ltration in both lungs developed; the diagnosis was pneumonia of unknown etiology. Early on January 17, she underwent endotracheal intubation. She died of acute respiratory distress syndrome and multiple organ failure later that day.
Two endotracheal aspirates collected on January 17 were positive for infl uenza virus (H5N1) and for genes encoding matrix protein by real-time PCR and RT-PCR. However, throat swabs collected on January 15 and 16 had been negative even after repeated testing (Table). The infl uenza virus (H5N1) was isolated on January 22 after 48-hour culture and named A/Shandong/1/2009(H5N1). Wholegenome sequencing showed that all segments were of avian origin. The