Infection and Cross-infection in a Paediatric Gastro-enteritis Unit

IN TR O D U CTIO N Acute gastro-enteritis is the m ajor cause of death in the infant population of black, coloured and Asiatic communities of South Africa, and ranks second to pneum onia in the white community. Bulletin 60 W H O states the median incidence of diarrhoeal disease as 2,2 episodes per child per year up to the age of 5 years. Repeated attacks of enteritis severely aggravate the already poor nutritional state of black children, with consequent increased susceptibility to a variety of infectious agents, and thus increased morbidity. A two m onth study was conducted on a paediatric gastro-enteritis unit of a black academic hospital. Aims of study were to: 1. Identify bacterial and viral pathogens causing diarrhoea. 2. D ocum ent nosocomially acquired infections. 3. Investigate nursing and other practices which may predispose to infection. Definition of nosocomial infection A nosocomial enteric infection was considered to have occurred when a pathogen was detected for the first time in a third or subsequent stool specimen collected 48 hours or later following admission. Ward Environment The 52 bed paediatric gastro-enteritis ward has a turnover of approxim ately 150 patients per month. Seasonal variations occur with the highest num ber of admissions in the summer months. During busy periods the close proximity of cots and staff shortage increases the risk of cross-infection. M E T H O D S Patients Stool specimens or rectal swabs were collected for bacterial and viral exam ination on admission, the following day and thereafter twice weekly. (Bacteriological testing was supervised by a senior technician to m aintain consistency in results) Details were recorded of the course of illness, including tem perature and frequency of diarrhoea, antibiotics received and note was made of gentamycin resistance. Blood specimens were collected for culture on admission and when clinically indicated. A social history questionnaire was recorded on admission. Staff Stool specimens were collected for culture from nurses and boarder mothers (breast feeding mothers who stayed at hospital to care for their babies). Hand swabs were 30 C urationis Vol.


IN TR O D U CTIO N
Acute gastro-enteritis is the m ajor cause of death in the infant population of black, coloured and Asiatic communities of South Africa, and ranks second to pneum onia in the white community.
Bulletin 60 W H O states the median incidence of diarrhoeal disease as 2,2 episodes per child per year up to the age of 5 years.Repeated attacks of enteritis severely aggravate the already poor nutritional state of black children, with consequent increased susceptibility to a variety of infectious agents, and thus increased morbidity.
A two m onth study was conducted on a paediatric gastro-enteritis unit of a black academic hospital.Aims of study were to: 1. Identify bacterial and viral pathogens causing diarrhoea.

Investigate nursing and other practices
which may predispose to infection.

Definition of nosocomial infection
A nosocomial enteric infection was considered to have occurred when a pathogen was detected for the first time in a third or subsequent stool specimen collected 48 hours or later following admission.

Ward Environment
The 52 bed paediatric gastro-enteritis ward has a turnover of approxim ately 150 patients per month.Seasonal variations occur with the highest num ber of admissions in the summer months.During busy periods the close proximity of cots and staff shortage increases the risk of cross-infection.

M E T H O D S Patients
Stool specimens or rectal swabs were collected for bacterial and viral exam ination on admission, the following day and thereafter twice weekly.(Bacteriological testing was supervised by a senior technician to m aintain consistency in results) Details were recorded of the course of illness, including tem perature and frequency of diarrhoea, antibiotics received and note was made of gentamycin resistance.
Blood specimens were collected for culture on admission and when clinically indicated.A social history questionnaire was recorded on admission.

Staff
Stool specimens were collected for culture from nurses and boarder mothers (breast feeding mothers who stayed at hospital to care for their babies).Hand swabs were collected from nurses, including milkkitchen staff, doctors, boarder mothers and domestic staff at various times throughout the day.Procedures were observed and note taken/specim ens collected if desirable.

Environment
Specimens were collected from the environm ent to try to determine means of spread of pathogens.Specimens of feeds and feeding equipm ent were collected.

Demography -Patient population
One hundred and thirty black children, 75 male and 55 female, were adm itted over the 8 week period from 25 July -17 Septem ber 1988, suffering from severe gastro-enteritis necessitating intravenous rehydration.Their ages ranged from 2 weeks -5 years 8 months.Seventy four per cent of patients were under 1 year of age.
• See Table I.They gave a history of diarrhoea and vomiting from 1-21 days (mean of 3 days).A history of previous diarrhoeal disease was obtained in 33 (25%) children.Enemas had been adm inistered to 32 (25%) patients; these included soap, milk, hebs, ash, alum and castor oil.In many cases it was not known whether diarrhoea had preceded or followed enem a adm inistration.Total Laboratory Findings Enteric pathogens, bacterial or viral, were found on admission in 61 of 129 (47,2%) patients.A total of 81 pathogens were recorded from these 61 patients; 56 bacterial and 25 viral.In 10 of these 61 (6%) patients a com bination of bacterial and viral pathogens were isolated.A further 8 patients w ithout dem onstrable enteric pathogens on admission subsequently acquired nosocomial pathogens, bringing the total with enteric pathogens to 69 (53,4%).A total of 105 pathogens were recorded from the 69 patients by the end of the study.In 43 patients only one enteric pathogen was dem onstrated, 22 patients had two pathogens, 5 patients dem onstrated three pathogens, whilst 1 patient had four pathogens.See Table III.

Bacteriological Findings
Campylobacter jeju n i was present in the stools of 28 (22%) patients on admission; 18 of these had at least one other enteric pathogen.Only 2 patients nosocomially acquired Campylobacter jeju n i, and that occurred 7 days after admission.Campylobacter jeju n i was observed to be shed in the stool for up to 32 days.Enteropathogenic serotypes of E. coli (EPEC) were found in 14 (10,8%) patients on admission; another 8 patients nosocomially acquired EPEC during the duration of their illness; only 4 EPEC were not associated with other pathogens.Two patients dem onstrated 2 different EPEC serotypes.EPEC was nosocomially acquired between 5 and 43 days following admission.
Shigella flexn eri was isolated from the stools of 6 patients on admission and was acquired nosocomially in one.Shigella sonnei was present on admission in 3 patients and acquired nosocomially in two.The nosocomial infections occurred between 5 and 9 days following admission.
Salmonella reading, seftenberg and typhim urium were each dem onstrated in one patient respectively on admission.A further 6 acquired a nosocomial Salmonella typhim urium between 6 and 18 days following admission.In one patient it was shed in the stool for 15 days.
Cryptosporidium was present on admission in 2 patients and Aerom onas hydrophilia was dem onstrated in one, and acquired nosocomially in another.

Viral Findings
Rotavirus was isolated from the stools of 25 of 129 patients (19%) on admission (21 of these were under 6 m onths of age), and acquired nosocomially by a further 3 patients between 9 and 11 days following admission.It was shed for up to 10 days.Bacterial pathogens were associated with 10 of the 28 patients with Rotavirus.

Nosocomial Enteric Pathogens
Nosocomially acquired enteric pathogens occurred in 22 (17,1%) patients.Of these, 14,8% were bacterial and 2,3% viral.See Table III.The time of acquisition ranged from 3-43 days (mean of 12,5).Eight of these patients did not have enteric pathogens dem onstrated on admission, whereas the other 14 acquired additional nosocomial pathogens.The length of stay in hospital ranged from 2-48 days (mean 7,2) but for those with a nosocomially acquired enteric infection the mean increased to 20,2 days.O ther non-enteric nosocomial infections which occurred in the study were 1 pneum onia, 1 bronchitis and 1 abscess on the scalp.

Deaths due to gastro-enteritis
D uring the study period 5 (3,8%) patients died.Three patients were under one year of age.None of the deaths was due to nosocomial infection.

Antibiotics and Gentamicin Resistance
One hundred of 129 (77,5%) patients received antibiotics.These included Bactrim, Neomycin, Amoxil, Penicillin, Ampicillin, Flucloxacillin, Chloram phenicol, Erythrom ycin, Amikacin and Vancomycin.Gentamicin resistant bacteria were recovered from the stools of 110 (85%) patients.In 8 (6%) patients gentamicin resistance was noted in the first specimen collected on the day of admission.It is not known whether they had been previously hospitalised or received antibiotics.In 84 (77%) patients gentamicin resistance occurred within 1-4 days (1st and 2nd specimen).O f the 19 The study conducted over 8 weeks during a non-busy period dem onstrates a nosocomial enteric infection rate of 17,1% ^ with increased morbidity, and prolonged ^ hospitalisation.In a busier period with a higher bed occupancy rate the risk of nosocomial infection would probably increase.
The high incidence of m alnutrition, concom itant illness, age of patients and selective effect of antibiotic therapy probably contributed to the increased risk of nosocomial infection in the compromised host.The pathogen most frequently isolated as an agent of nosocomial infection was EPEC which occurred in 6,2% patients.D upont et al.9 state EPEC as third to Clostridium difficle and Salmonella as a reported cause of nosocomial infection in NNIS data.
The mode of spread may be from the hands of unit personnel or fomites.Pharyngeal colonisation of infants occurs and may be an interm ediate step in the spread of infection.The organism may also be detected in dust and fomites in the ward environm ent.Salmonella typhim urium occurred as a nosocomial infection in 4.6% patients.This is confirmed as a frequent agent in other studies8,9 and is typically A W H O consultant wrote that such infections could be airborne and evidence suggests that vacuum cleaners could disseminate Salm onella typhim urium .10Shigella spp.which occurred as a nosocomial infection in 3 patients was high, as the initial patients adm itted with Shigella spp.am ounted to only 9, indicating a high potential for spreading.Other studies8, indicate the cross infection rate as low, but the high incidence in this study may be due to the very low dose of organisms (less than 103) necessary to produce disease.Rotavirus occurred as a nosocomial infection in 2,3% patients, which is a low rate considering there were 25 initial patients and the virus was shown to be shed for to 10 days.Cam pylo bacter jejuni likewise occurred rarely as a nosocomial agent in 2 patients, although 28 had Campylobacter isolated on admission, and the organism was evidenced to be shed in the stool for up to 32 days.This indicates the low incidence of • horizontal spread.
It may be seen from results of hand swab cultures that hand disinfection by staff with Chlorhexidine 0,5% and alcohol 70% handspray was strictly observed between patient contact.Even domestic staff were observed to be fastidious in this m atter.It would therefore appear that hands of unit personnel were not im portant factors in the spread of cross infection.
Two main areas of weakness were observed in the nursing procedures which were probable factors in the spread of cross infection a) the bath routine and b) the method of naso-gastric feeding.D uring the daily bath routine babies and young children werebathed in succession in the ward sinks which were not adequately disinfected between each bathing (see Table IV).After being bathed patients were laid on a shared blanket covered with a towel.

•
As the towel became wet, pathogens ere able to be transferred to the blanket below, and could subsequently be transferred to the next patient to be dried.Enteric pathogems resistant to Gentamicin were isolated from both sinks used for bathing and from the bath mattress (towel and blanket) on which patients were dried.
A nother im portant contributory factor to the spread of nosocomial pathogens was the method of giving naso-gastric feeds, and the feeds themselves, of which 56% contained bacteria 104 -105.See Table V. Babies who were unable to take feeds orally were fed via a naso-gastric tube.The naso-gastric tube became colonised, within 24 to 48 hours of insertion, with the patients' enteric flora (50% of which were resistant to Gentamicin and Amikacin).The means of giving the naso-gastric feed was via the barrel of a 20 ml syringe connected to a rubber tube with a plastic connection to attach it to the patients' naso-gastric tube.A com mon syringe barrel, tubing and connection was used consecutively for each patient without any attem pt to sterilise it between patients.It was rinsed through with cold water only.Seventy two per cent of cultures from feed equipment proved contam inated with enteric pathogens.
It would appear therefore, that patients receiving 3 hourly naso-gastric feeds were exposed to high risks of acquiring a nosocomial infection from contam inated feed equipment.The close proxim ity of cots and the constantly high bed occupancy rate in the acute admission ward made nursing procedures difficult to perform w ithout contam ination of nurses' hands or equipment.Although nurses were observed to w ash/disinfect hands before performing a "procedure" casual contact by touching bedlinen or a patient could result in pathogens being spread (See Table IV).It must be remembered that the colonised patient is a silent source of bacteria.
It is possible that the nosocomial infection rate was higher than reported, and the acquisition of the cross infection could have occurred within 48 hours of admission whilst in the acute admission ward.If pathogens found in a second stool specimen within 48 hours of admission (following a previous negative specimen) are considered to be nosocomial the incidence of cross infection increases to involve 34 (26,3%) patients.
Klebsiella spp.and Acinetobacter spp.were resistant to Gentamicin and Amikacin.One pathogenic organism, Yersinia Enterolytica was identified in the stool of a boarder m other who was asymptomatic.

-
Klebsiella spp., E.coli and Acinetobacter spread by hands, fomites or in dust.