Viruses are recognized as important healthcare-associated enteropathogens that spread via person-to-person transmission or point-source infection through food or water. Studies show that enteric viruses have caused 86% of healthcareassociated gastroenteritis in infants and children (10,22, 23 and 24).

Enteric Adenoviruses Enteric adenoviruses consist of two serotypes—40 and 41—which are members of group F adenoviruses (25). These agents primarily infect children younger than 2 years and occur year-round (26,27). Adenoviruses cause a spectrum of conditions ranging from asymptomatic infection in 40% of infected children to diarrhea and vomiting lasting for 7 to 10 days. The incubation period is 3 to 10 days. Illness is frequently associated with fever and respiratory tract symptoms (28). Transmission occurs via the fecal-oral route and readily transfers from person to person (26). Treatment is nonspecific, and fluid
replacement is dictated by the patient’s condition. The frequency of dehydration and fever due to enteric adenovirus gastroenteritis appears to be similar to that of other enteric viruses (29). In several reports, enteric adenovirus was the third most frequent cause of viral gastroenteritis in hospitalized infants and young children (28, 29, 30, 31 and 32). In one study, 54% of 127 enteric adenovirus infections were healthcare associated (31). Adenovirus type F41 caused an outbreak in a pediatric bone marrow transplant unit and was shed in the stool for up to 64 days in these immunocompromised patients (33). These viruses have been shown to be a major cause of morbidity in hospitalized infants who have undergone ileostomy or colostomy procedures for necrotizing enterocolitis (NEC) (34). The HAI rate in these patients was higher than other nursery patients, and infection resulted in a prolonged hospital stay. Diagnosis of enteric adenovirus-associated gastroenteritis can be made by an evaluation of stool specimens using either EM or a commercially available enzyme immunosorbent assay (EIA). Research methods use PCR and gene sequencing for detection, genotyping, and outbreak analysis (33).

Astrovirus Eight antigenic types of human astrovirus have been identified. Gastroenteritis due to astrovirus occurs worldwide and has been associated with outbreaks of mild diarrhea in schools (35,36), childcare centers (37,38), nursing homes (39,40), and pediatric hospital wards (41, 42 and 43). Astroviruses are responsible for approximately 3% to 5% of hospital admissions for gastroenteritis. Illness occurs mainly in children younger than 2 years and frequently causes asymptomatic infection (44). The illness lasts for 1 to 4 days following an incubation period of 24 to 36 hours. Gastrointestinal tract symptoms are nonspecific, consisting of vomiting, diarrhea, fever, and abdominal pain. The mode of transmission is person to person among children. Astrovirus has been reported to be responsible for 5% to 7% of nosocomial gastroenteritis in children’s hospitals (41,43,-45, 46, 47 and 48). An attack rate of between 7% and 62% was reported during an outbreak of nosocomial infection in a children’s ward (43). Astrovirus caused a prolonged outbreak of diarrhea among immunocompromised patients in a pediatric bone marrow transplant unit (49). Astrovirus-associated gastroenteritis is diagnosed by an examination of a stool specimen by EM, EIA, or RT-PCR (37,49,50). Commercial EIAs for the detection of human astroviruses are not available in the United States but may be used in other countries.

Caliciviruses Four genera of the family Caliciviridae have been described including Noroviruses (formerly known as Norwalk-like viruses) and Sapoviruses (formerly known as Sapporo-like viruses) (51). Human calicivirus infections occur year-round, although some studies suggest a seasonal predominance. The incubation period is 12 hours to 4 days, and the clinical symptoms include vomiting and diarrhea, which last for 1 to 4 days. The severity of symptoms caused by caliciviruses is indistinguishable from that of symptoms caused by other enteric viruses (52,53). Persistent excretion may occur in immunocompromised hosts. Caliciviruses have been identified in stools for up to 2 weeks after the onset of symptoms (54). Calicivirus is transmitted by the fecal-oral route through food-borne and waterborne transmission (55). Calicivirus can be detected in stool specimens of 0.2% to 6% of children hospitalized for gastroenteritis. When calicivirus was detected in hospitalized children, it was healthcare-associated in approximately 40% of cases (46,53,56). Caliciviruses have tremendous antigenic and genetic diversity that make detection assays insensitive. Multiple strains have been detected in pediatric hospitals (57). It is apparent that most studies have certainly underreported the significance of calicivirus infections because of these insensitive assays. Caliciviruses can be detected in stool specimens by EM, immune EM, RT-PCR, or EIA, but these tests are available only in research laboratories (25).

Two of these genera—the Noroviruses and Sapoviruses —infect humans. Many of the Noroviruses are known only from a single outbreak and have been named after the sites at which the outbreaks occurred. They include Norwalk; Hawaii; Snow Mountain, MX (Mexico); and Lordsdale (51,58,59) Norwalk virus is the best-studied member of the genus. Norwalk virus illness follows an incubation period of 18 to 48 hours and is characterized by vomiting, diarrhea, abdominal pain, and low-grade fever lasting 1 to 2 days (60). Epidemics of Noroviruses have been reported in nursing homes, schools, recreational areas, cruise ships, and hospitals (61, 62, 63, 64, 65, 66 and 67). Waterborne (55), food-borne (68,69), and person-to-person transmission have all been implicated in epidemics (58,70,71), and the results of volunteer studies suggest the possibility of fecal-oral transmission. There is evidence of the survival of noroviruses in environmental reservoirs (72). Aerosolization of vomitus has also been implicated as a mode of transmission. In one hospital outbreak, 55% of elderly patients and 61% of the healthcare workers on one floor became ill (73). The healthcare workers most likely spread the infection from patient to patient. Another reported outbreak affected 57 patients and 69 staff members over a 26-day period. The index case was a patient hospitalized with acute abdominal pain and diarrhea 2 days prior to the outbreak. The epidemic curve indicated person-to-person transmission (74). In another report of an outbreak in a children’s ward, 15 children had the Norwalk virus in stool specimens, and the ward had to be closed to control the outbreak (61).

Norovirus infections in enclosed settings were studied for the effectiveness of control measures. One review identified norovirus outbreaks as the most common cause to result in the closure of a medical department (75). The median outbreak duration was 16 days. Other studies have not shown interventions that successfully shortened the outbreaks (76).

The Sapoviruses have also been associated with sporadic outbreaks and have been named after the location of the outbreak. They include Sapporo, Houston, London, Manchester, and Parkville (77, 78 and 79). Illness due to Sapoviruses is similar to that associated with the Noroviruses.

Rotavirus Rotavirus is the most thoroughly investigated and described etiology of healthcare-associated viral gastroenteritis and is one of the most important enteric pathogens worldwide. There are six distinct rotavirus groups, three of which infect humans. Group A rotavirus is the most common cause of diarrhea in infants and children throughout the world, including the United States. Groups B and C cause human disease in the Far East (25,27,80).

Rotavirus has an incubation period of 1 to 3 days. Excretion of rotavirus in stool can precede the onset of illness by several days and can persist for 8 to 10 days after symptoms of illness have abated (81). The illness usually has an abrupt onset characterized by explosive, watery diarrhea and is often associated with vomiting either before or after the onset of diarrhea. Dehydration occurs in 40% to 80% of patients and is usually mild, but severe dehydration and death have been reported in children and adults (82). Rotaviruses are transmitted principally by the fecal-oral route. They are found in nearly 50% of stool specimens from children admitted to the hospital with gastroenteritis. The majority of patients with rotavirus infection are between 6 and 24 months old (83). In North America, the annual rotavirus season begins in late fall in Mexico and moves across the continental United States from the southwest to the northeast, resulting in a peak of rotavirus activity in March and April in eastern Canada and the northeastern United States (84).

A 5-year retrospective study of 577 children with confirmed rotavirus gastroenteritis revealed that 121 (20%) were healthcare associated. The incidence was 4.2 cases per 10,000 patient days. The median age was 11 months. All healthcare-associated rotaviruses were type G1 with four different PCR subtypes described. The long-term and intensive care units had a considerably higher proportion of nosocomial infection than the more acute care medical beds where the majority of community-acquired rotavirus infections were treated (85).

Healthcare-associated transmission of rotavirus has been well documented on pediatric hospital wards; 2% to 24% of children admitted to the hospital with other diagnoses acquire rotavirus in the hospital (21,30,86, 87, 88, 89, 90, 91, 92 and 93). In several reports, up to 70% of infants in a nursery have healthcare-associated rotavirus (94, 95, 96 and 97). Hospital surveillance for rotavirus using molecular methods has detected newly emerging strains and the nosocomial transmission of those new strains (96, 97, 98, 99 and 100). Immunocompromised children acquire rotavirus in the hospital with an infection rate of 12% to 25%. Immunocompromised children have an extended period of virus excretion and may be a source of virus for transmission to others (101,102). Communityacquired symptomatic rotaviral infection in children admitted to the hospital and asymptomatic rotavirus shedding by neonates and other hospitalized infants appear to be the primary reservoirs for nosocomial rotavirus infection in susceptible children (103, 104 and 105). Fomites may play a role in rotavirus transmission. Rotavirus contamination was detected by PCR in 19% of inanimate objects in a childcare center during an outbreak (106,107). Nosocomial rotavirus infection may cause both outbreaks and endemic diarrheal disease in newborn nurseries; however, infection is usually asymptomatic (94,105,108, 109, 110, 111, 112, 113, 114 and 115). Asymptomatic rotavirus infection also occurs in older children, with evidence of asymptomatic rotavirus infection occurring in 24% to 50% of the infections in infants younger than 2 years during the rotavirus season (116, 117 and 118). Asymptomatic excretion of rotavirus by healthcare personnel has been proven (119). The incidence of confirmed nosocomial rotavirus diarrhea in a large pediatric hospital was found to be 0.5 to 0.9 per 1,000 admissions (120,121).

HAIs in developing countries occurred in 12% to 22% of all rotavirus cases with an incidence of 0.024 to 1.6 cases per 1,000 child-days. No HAIs occurred in children older than 4 years (122,123,124).

Prospective surveillance for rotavirus gastroenteritis in three Spanish hospitals in the winter of 2006 to 2007 identified nosocomial rotavirus infection in 2.8 cases per 100 inpatients and 0.48 cases per 100 patient days. Rotavirus was the etiology of 59% of healthcare-associated gastroenteritis. G9P[8] and G1P[8] were the predominant serotypes (15).

A review of healthcare-associated rotavirus gastroenteritis in European hospitals made several conclusions. Healthcare-associated rotavirus infections have an incidence from 1.6 to 6.3 cases per 1,000 children younger than 5 years; these infections represent 1.6 to 15.8 cases per 1,000 hospital days and account for 3,000 to 20,000 rotavirus infection cases in children younger than 5 years. Rotavirus gastroenteritis lengthens hospital stay by 1.8 to 5.0 days. Hand washing, and specifically the use of alcohol-based hand sanitizers, is very effective in reducing the number of HAIs (125,126). Unique serotypes that occur in a community will be seen as a cause of HAIs during that season (127).

Breast feeding was protective against both infection and symptoms, whereas 32% of formula-fed and 11% of breastfed infants acquired rotavirus in the hospital (p < .005). No breast-fed infant had symptomatic infection (128).

Rotavirus is detected in stool specimens by EIA, latex agglutination, EM, polyacrylamide gel electrophoresis, and PCR (129, 130, 131 and 132). Several EIA and latex agglutination assays are commercially available. Detection rates depend on which assay is employed because of their varying sensitivities.

Universal mass vaccination against rotavirus in Austria resulted in a considerable decrease in hospitalization for rotavirus gastroenteritis and a considerable decrease in the number and percentage of healthcare-associated rotavirus infections (133, 134 and 135). Rotavirus vaccination in the United States has lowered the frequency of severe diarrhea and, therefore, hospitalizations. The impact on HAIs remains to be seen.

Human Parechoviruses Human parechoviruses belong to the family Picornaviridae. A retrospective evaluation identified human parechoviruses type 4 to 6 in association with gastrointestinal tract symptoms in addition to respiratory tract symptoms. Twenty of 277 children with the virus detected in fecal samples had gastrointestinal symptoms. The authors describe at least one child who acquired the virus nosocomially, because it was a 3-month-old infant who had never left the hospital. The virus was detected by RT-PCR in nasopharyngeal aspirates, feces, and plasma. It could be detected in fecal samples for up to 40 days after initial detection (136).

Healthcare-associated bacterial gastroenteritis is less common than viral gastroenteritis. In the United States, the majority of information about healthcare-associated bacterial gastroenteritis consists of reports of outbreaks, many of which are food-borne or waterborne. Several prospective studies of healthcare-associated diarrhea in children failed to demonstrate any bacterial etiologies (14,22), although C. difficile is common in adults. Bacterial infections are much more common in less-developed countries. A patient with poor nutritional status or an immunocompromised
patient is at particularly high risk for healthcare-associated bacterial gastroenteritis.

Campylobacter Species Twenty-one species have been identified in the family Campylobacteraceae, but only 12 cause disease in humans. Campylobacter jejuni, Campylobacter coli, Campylobacter upsaliensis, and C. jejuni subspecies doylei are the most common species isolated from children. Campylobacter fetus is a rare cause of bloodstream and systemic infections occurring mostly in immunocompromised and debilitated hosts, as well as a cause of perinatal infection and abortion. Since C. jejuni is the species that usually causes intestinal illness, many laboratories place stool specimens on selective media with incubation temperatures to isolate this species. With this method, several other Campylobacter species will be missed as a cause of diarrhea. Therefore, the extent of HAIs involving many of the Campylobacter species is not known. Isolation of Campylobacter spp. from blood and other sterile body sites does not present the same isolation problem as isolation from feces (137).

Predominant symptoms are diarrhea, abdominal pain, malaise, and fever. Stools may contain blood. C. jejuni has been reported as a cause of severe infection in neonates following vertical transmission. Vertical transmission of a microorganism is considered a nosocomial infection in the nursery for purposes of surveillance reporting. Postnatal person-to-person transmission has also been documented with reports of nursery epidemics of Campylobacter spp. diarrhea and meningitis (138). The incubation period is 1 to 7 days.

Healthcare-associated transmission of Campylobacter spp. is not common. The gastrointestinal tract of domestic and wild birds and animals is the reservoir of infection. A Finnish hospital reported a waterborne nosocomial outbreak of C. jejuni gastroenteritis in both patients and hospital staff (139). Several community-acquired outbreaks caused by C. jejuni have been reported, usually due to ingestion of contaminated raw milk, water, or food. In addition, infection can occur through person-to-person transmission or contact with infected animals (140). For example, an outbreak involving two very closely related strains of C. upsaliensis in four childcare centers implicated person-to-person transmission (141).

Clostridium Difficile Healthcare-associated gastrointestinal tract infection caused by C. difficile in adults and children is discussed in Chapter 37. The role of C. difficile in antibioticassociated diarrhea has been more difficult to establish in infants and young children than in adults, since C. difficile is commonly recovered from the stools of asymptomatic infants and young children.

The reported incidence of neonatal colonization varies, with isolation rates as high as 90% in neonatal intensive care units (NICUs) and between 2% and 30% in healthy newborn infants. The C. difficile toxin has been detected in up to 36% of sick neonates without gastrointestinal tract symptoms. The incidence of C. difficile toxin detection in stool specimens declines with age and approaches 1% to 3% in healthy adults (142). Pseudomembranous colitis has been reported in infants and children, but the incidence is difficult to assess. C. difficile was described as the etiology of 13% to 16% of healthcare-associated gastrointestinal tract infections (9,143). Outbreaks of diarrhea associated with C. difficile have been reported in childcare centers (144). The incubation period is unknown. The virulence and mortality of C. difficile-associated disease are increasing (145). This may be associated with hypertoxigenic strains.

C. difficile may be isolated from stool using a selective cycloserine cefoxitin fructose agar in an anaerobic environment. C. difficile produces two toxins. The C. difficile cytotoxin B may be detected by cell culture cytotoxicity assay or EIA. Some commercially available EIAs will detect both toxins A and B (146). There have been reports of toxin A negative, toxin B positive C. difficile antibiotic-associated diarrhea in adults (147). Arbitrarily primed PCR has been used for genotypic differentiation of strains in hospital outbreaks (148).

Escherichia Coli Escherichia coli strains that cause acute diarrheal disease may be classified into five groups: enterotoxigenic (ETEC), enteroaggregative (EAEC), enteroinvasive (EIEC), enteropathogenic (EPEC), and enterohemorrhagic (EHEC) (149). ETEC usually infects infants and children in developing countries or adults following travel to developing countries. EAEC produces acute or chronic diarrhea in all age groups, but predominantly infants, by attachment to and effacement of the intestinal mucosa. EIEC infects all ages and causes diarrhea containing blood and mucus as a result of tissue invasion. These infections may occasionally be food-borne or occur as the result of travel to developing countries. EPEC produces acute and chronic diarrhea, generally in infants under 2 years old in developing countries (150). EHEC causes abdominal pain and bloody diarrhea in children and adults, mostly in developed countries. The illness may be complicated by hemolytic uremic syndrome (HUS) in children or thrombotic thrombocytopenic purpura in adults. It is most frequently spread by undercooked contaminated meat, but many other vehicles of transmission have been described.

The reported incidence of healthcare-associated gastroenteritis caused by these five groups of E. coli is low. This may be a reflection of the unavailability of detection methods in most clinical microbiology laboratories. The incubation periods range from 10 hours to 6 days.

Reports of healthcare-associated ETEC-associated diarrheal outbreaks in special care nurseries due to heatstable enterotoxin-producing strains include one report from Spain with six ETEC-associated neonatal diarrheal outbreaks (151). In another report, ETEC was cultured from infants, nurses, family members, infant formula, and surfaces in the nursery (152). This report implicated person-to-person transmission and food-borne transmission by formula. In another report of a hospital outbreak, a rare phage type further differentiated the infected strain (153). E. coli has also been reported as a contaminant of expressed human milk, which caused both asymptomatic infections and gastroenteritis in a nursery (154).

A single report described person-to-person transmission of EIEC in students and staff of a school for mentally retarded adults and children: 48% of the students and 28% of the staff were ill. Control of the outbreak was achieved by cohorting and an emphasis on hand washing (155).

Community-acquired food-borne outbreaks of EHEC serotype O157:H7 are well documented. E. coli O157:H
phage type 8 caused a hospital outbreak with food brought to a party from the outside (156). Many outbreaks have occurred in nursing homes, and a report of an outbreak in an institution for mentally retarded children and adults demonstrated the devastating effects of this microorganism in an outbreak. Eight of 20 infected residents developed HUS, and four died of complications (157). Twenty-nine children with E. coli O157:H7 in nine childcare centers were reported. There was evidence of person-to-person transmission in all nine facilities (158). Spread of E. coli O157:H7 from a patient to a nurse in the hospital setting has been reported (159).

EPEC is the strain most commonly associated with HAIs. It was the etiology of 1.3% of gastrointestinal tract infections in a children’s hospital in the United States (10) and of 6 of 10 nosocomial bacterial gastrointestinal tract infections in a South African hospital (9). Many studies of outbreaks of diarrhea in NICUs have demonstrated person-to-person transmission by the hands of hospital personnel. Premature infants are the most susceptible to severe morbidity and to mortality resulting from these infections (160, 161, 162, 163, 164 and 165). Detection of EPEC requires a high level of suspicion. Colonies of E. coli from a routine bacterial culture must be screened by type-specific antisera. Research methods for identifying related serotypes include adherence of microorganisms in HEp-2 cells, DNA probes, and PCR to detect EPEC strains with the enteroadherence plasmid.

Leuconostoc Species Leuconostoc species are members of the family Streptococcaceae and commonly found in dairy products, vegetable matter, and in the soil. They are not considered part of the normal gastrointestinal flora. These microorganisms have been described as the etiology of bacteremia and central catheter-associated infections in children with underlying gastrointestinal disease. The most common reports include children with short bowel syndrome receiving enteral feeds and/or total parenteral nutrition. Several of the children had concomitant gastrointestinal symptoms, thereby implicating the gastrointestinal tract as the source of infection (166).

Salmonella Species The genus Salmonella is now considered to comprise a single species named Salmonella enterica based on DNA structure and biochemical properties. Within this species are seven subspecies with almost all serotypes pathogenic for humans classified into subgroup I (S. enterica subspecies enterica). The subspecies can be divided into serotypes based upon their O (somatic) and H (flagellar) antigens. Two main clinical syndromes are associated with Salmonella. The first is the protracted bacteremia of typhoid (Salmonella typhi) and paratyphoid (S. paratyphi) fevers. The second is the predominantly gastrointestinal tract illness caused by animal adapted Salmonella strains. S. typhimurium is the serotype most commonly reported as the cause of Salmonella infections in humans in the United States. Many outbreaks of Salmonella gastroenteritis in hospitalized patients due to a variety of serotypes have occurred through various methods of transmission. Person-to-person transmission may occur among patients or from healthcare personnel. Commonsource outbreaks have also been traced to diagnostic agents and medications (167, 168 and 169). Salmonella infections have been acquired from reptiles, highlighting the importance of avoiding exposure to pet reptiles in a hospital setting (170). The incubation period for gastroenteritis is from 6 to 12 hours. For enteric fever, the incubation period is from 3 to 60 days but is usually 7 to 14 days.

Food-borne Salmonella outbreaks simultaneously may affect patients in multiple hospitals. In 1962 and 1963, a large outbreak of healthcare-associated gastroenteritis caused by S. derby occurred among patients, medical staff, and employees of 53 hospitals in 13 states (171). Contaminated eggs that were eaten raw or undercooked were responsible for this and many other outbreaks (172, 173, 174 and 175). Person-to-person transmission to hospital staff and to other patients has been documented (174,176,177). Foodborne outbreaks have also occurred following the ingestion of improperly cooked and stored poultry (178,179). Foodborne outbreaks may originate in hospital personnel or in patients (180, 181, 182 and 183). An epidemic caused by S. kottbus was traced to contaminated pooled human milk (184). S. poona mastitis was the source of infection for a 5-week-old infant. This case was not healthcare associated but indicated the need to consider pumped breast milk as a potential source of infection (185).

Common-source Salmonella outbreaks have also been traced to contaminated diagnostic reagents and medications. These types of outbreaks generally do not present as typical common-source outbreaks and, therefore, may be difficult to recognize. An interstate outbreak of S. cubana occurred in 1966 due to contaminated carmine dye used as a marker of gastrointestinal tract transit (167, 168 and 169). Healthcare-associated outbreaks of salmonellosis have also been traced to bile salts, gelatin, pancreatin, pepsin, vitamins, and extracts of various endocrine glands (186,187). Eight cases of S. enteritidis occurred in hospitalized patients receiving enteral nutrition containing lyophilized egg albumin (188). These outbreaks appeared to be sporadic and, therefore, required a high index of suspicion to document their association with a common vehicle.

Outbreaks of Salmonella gastroenteritis have also been associated with a variety of medical instruments or procedures, including upper gastrointestinal tract endoscopy (189), fiberoptic colonoscopy (190), rubber tubing attached to a suction apparatus (191,192), rectal thermometers (193), and contaminated mattresses (194). Healthcare-associated S. hadar infection occurred in laundry personnel at a nursing home following a food-borne outbreak in the nursing home residents. This report implicated the handling of soiled laundry in the absence of person-to-person contact (195).

Outbreaks of Salmonella infection have been reported in nurseries. The microorganism is generally introduced to the nursery by an infant recently born to a mother with clinical or asymptomatic salmonellosis (196) or a child with community-acquired Salmonella infection. In another report, a 12-day-old infant acquired S. brandenburg from his visiting mother (197). Salmonella is transmitted among the staff and patients through person-to-person contact. The acquisition of multiple resistant microorganisms by premature infants in special care nurseries results in increased rates of morbidity and mortality (198,199). A case-control study of an outbreak of Salmonella infantis in a neonatal care unit in Brazil demonstrated protection
by increased birth weight, and peripheral IV catheter use was a risk factor. Overcrowding and understaffing were associated with the outbreak (200).