- Email: [email protected]
An outbreak due to Echovirus type 30 in a neonatal unit in France in 1997: usefulness of PCR diagnosis
Journal of Hospital Infection (1999) 43: 63–68
An outbreak due to Echovirus type 30 in a neonatal unit in France in 1997: usefulness of PCR diagnosis M. Chambon, J. L. Bailly, A. Béguet, C. Henquell, C. Archimbaud, J. Gaulme, A. Labbé, G. Malpuech and H. Peigue-Lafeuille Departments of Paediatrics and Clinical Virology, Centre Hospitalier Universitaire, Clermont-Ferrand, France Summary: Between February and August 1997, 53 patients with enterovirus meningitis were hospitalized in Clermont-Ferrand, France. All but one were children. Echovirus type 30 was involved in 70% of cases with identified serotype. The outbreak ceased on August 8. Two months later, a neonate was admitted to the neonatal unit with an echovirus type 30 meningitis thought to be acquired at delivery. Twenty days later a nosocomial outbreak of echovirus type 30 involving five neonates occurred. Two of them presented with meningitis and two with febrile seizure; One was asymptomatic. The retrospective examination of the maternal sera in a neutralization test, using the index case strain as a source of antigen, showed that none of the neonates was passively immunized before hospitalization. The use of genome detection in cerebrospinal fluid allowed rapid diagnosis and infection was contained by re-inforcing hygiene measures. Prospective examination of stools in the neonatal and paediatric units showed no further occurrences of the disease. No sporadic case was observed in the general population. Hence, nosocomial infections can occur a long time after an outbreak in the general population; rapid diagnosis with molecular tools is useful both for a definite diagnosis in patients already hospitalized, and to act as a rapid alert, even in intervals between seasonal outbreaks. © 1999 The Hospital Infection Society
Keywords: Enterovirus; echovirus type 30; nosocomial infection; neonate; rapid diagnosis; PCR.
Introduction Human enteroviruses include 66 known serotypes divided into five subgroups: polioviruses, coxsackieviruses A, coxsackieviruses B, echoviruses and non-assigned enteroviruses 68 to Received 16 December 1998; Revised manuscript accepted 16 June 1999 Corresponding author: Hélène Peigue-Lafeuille, Service de Virologie, Faculté de Médecine, 28 Place Henri Dunant, 63001 Clermont-Ferrand Cedex, France.
0195–6701/99/090063 + 06 $12.00
71. Non-polio enteroviruses – mainly coxsackieviruses and echoviruses – are responsible for numerous human illnesses ranging from the insignificant to the severe, especially in children.1 Neonates are at risk of serious and sometimes fatal diseases.2,3 Human enteroviruses have a dangerous potential for nosocomial spread.4 In temperate climates, enterovirus infections occur every year, mainly in the summer and autumn. When outbreaks of non-polio enteroviruses occur in the general population, neonates are especially at risk. © 1999 The Hospital Infection Society
64
We describe an outbreak of echovirus type 30 infections in a neonatal unit in France in October 1997, which occurred two months after a large outbreak of echovirus type 30 meningitis in the population. The rapid diagnosis of the disease by nucleic acid detection of enterovirus genome was useful in limiting nosocomial spread.
Patients and methods All patients admitted to hospital during 1997 with a suspicion of enterovirus meningitis were prospectively investigated. Diagnosis was based on viral examination including polymerase chain reaction (PCR) assay of cerebrospinal fluid (CSF) for the rapid diagnosis of enterovirus meningitis with the AMPLICOR EV test Kit (Roche Molecular Systems, Paris, France). Several regions of the 5′ end of the enterovirus genome are highly conserved among different enterovirus serotypes and so a single pair of oligonucleotide primers can amplify most known enteroviruses with the exception of echoviruses 22 and 23.25 The AMPLICOR EV test kit is a rapid (5 h) PCR-based assay in which reverse transcription and amplification are performed in a one step reaction using rth polymerase. Carryover contamination is prevented by dUTP and AmpErase.6 Detection is performed colorimetrically on a microwell plate. Virus isolation was performed on CSF, throat swab and stool specimens using human oral epidermoid carcinoma (KB) and human lung fibroblast MRC5 cell lines (Bio-Mérieux, Marcy l’Etoile, France). Lim-Benyesh-Melnick antiserum pools were used for enterovirus identification by seroneutralization tests.7 Neutralizing antibody assays (IgG and IgM) were performed by a micromethod with MRC5 cells and the corresponding isolated virus as a source of antigen. The paediatric department consists of six different units that are geographically separated. The neonatal unit is divided into two areas, a paediatric intensive care unit and, immediately adjacent, the neonatal unit itself.
M. Chambon et al.
Results Outbreak of enterovirus meningitis in the general population
Between February and August 1997, 59 patients with enterovirus meningitis were admitted to hospital. In 53 of the 59, an enterovirus was isolated and identified. Enterovirus detection by PCR was performed in 48 patients and was positive in 92% (N = 44) of cases. For six patients, PCR in CSF was the only positive result; either cultures were negative (four cases) or specimens not obtained (two cases). The serotypes involved are shown in Table 1. In 37 cases (70%) echovirus type 30 was identified and in eight cases (15%) echovirus type 6. None of these patients was hospitalized in the neonatal unit. The youngest patient was 1 month old. The outbreak of echovirus type 30 occurred suddenly, with a peak in June (23 cases) and a falling off in July (12 cases). The last case was observed on August 8; this patient, a 39-year-old male, was the only adult victim of the enterovirus outbreak. No further echovirus type 30 case was observed until October 2. Nosocomial infections due to echovirus type 30 in the neonatal unit
The index patient was a 5-day-old male, born at term, who was hospitalized on October 2, 1997 because of neonatal meningitis (Figure 1). He was diagnosed as having enterovirus meningitis. Table I Enteroviruses isolated during an outbreak of meningitis in 1997 in the general population. Serotype
n
%
Echovirus 30 Echovirus 6
37 8
70 15
Echovirus 9 Echovirus 11 Echovirus 24
1 1 1
Coxsackie B3 Coxsackie B4 Coxsackie A9
2 1 2 53
] ]
5·6
9·4 100
PCR diagnosis of Echo 30 in a neonatal unit
65
Patient 5
Patient 4
Patient 3
Patient 2
Patient 1
Index patient
15/ 09
01/ 10
15 / 10
31 / 10
15 / 1 /1997
Figure 1 Nosocomial outbreak due to echovirus type 30. Rectangles represent duration of hospital stay(s). Grey triangles represent positive detection of enterovirus genome by PCR in CSF (index patient, patients 1 and 4). Full black circles represent isolation of echovirus type 30 in stools.
PCR in CSF (containing 400 WBC) was positive on admission. Echovirus type 30 was isolated from a stool six days later. His mother had presented with fever and upper respiratory tract infection at delivery. Her stool specimen obtained 14 days after admission of her child was positive for echovirus type 30. The strain was used as an antigen source, and specific antibody titres of IgM (1 : 40) and IgG (1 : 160) were detected in her serum 14 days after birth. Bacteriological cultures were negative in both mother and child. The boy made a complete recovery and was discharged on day 34. Patients 1, 2 and 3 in the nosocomial outbreak were triplets, three females born by caeserean section at 34 weeks of gestation, who had been admitted on September 14, because of prematurity and hypotrophy. Patient 1 was first admitted to the intensive care unit because of deterioration of respiratory function. She was intubated on admission. When her condition improved, on September 18, she was transferred to the same area of the neonatal unit as the index patient and her two sisters. On day 40, October 24, she became febrile. A lumbar
puncture showed pleiocytosis (123 WBC, 68% lymphocytosis). PCR was positive for enterovirus and echovirus type 30 was isolated in a stool specimen taken on the same day. Her condition slowly improved and she left hospital on November 7, after a hospital stay that was two weeks longer than initially predicted. Patients 2 and 3 had similar features. They were admitted to the neonatal unit and discharged on October 24, the day their sister became febrile. They were both readmitted on October 30, because of a 3-day recurrence of fever and seizure to one of the paediatric units. Stool culture on admission was positive with isolation of echovirus type 30. They both improved. Examination of the CSF was normal and they were discharged 8 days later. Patient 4 was a two-day-old male, born at term who was admitted to the neonatal unit on October 8 for neonatal infection due to Proteus mirabilis isolated from both mother (vaginal secretions) and child (meconium, navel, gastric fluid specimens). Cefotaxime, amoxycillin and amikacin were given for 10 days and then discontinued. One day later, he was febrile
66
and became lethargic with loss of appetite. Recurrence of his Proteus mirabilis infection was suspected. Bacterial examinations remained negative. On October 21, a lumbar puncture showed pleiocytosis (65 WBC), and PCR was positive for enterovirus. Stool specimens obtained the same day and eight days later were positive for echovirus type 30. The enterovirus meningitis prolonged the hospital stay by two weeks. Patient 5 was a four-day-old male, born at term, who was admitted for neonatal meningitis due to E. coli, isolated from his urine and from maternal specimens. After three weeks of therapy (cefotaxime, amoxycillin and amikacin), he improved. Before he was discharged, systematic bacteriological examination of stools, urine, CSF and throat swabs were negative but echovirus type 30 was isolated from stools on October 24. He was discharged on October 31. Systematic stool examinations during the same period of eight patients in the neonatal area where the other cases had been observed, and 88 in the paediatric unit to which patients 2 and 3 had been readmitted, revealed no further occurrence of echovirus type 30 infection. Hospital staff were not examined. Serum specimens from the mothers of patients 1–5 were analysed for the presence of echovirus type 30 antibodies, with virus from the index patient as a source of antigen. They showed low and non-significant titres of 1: 20, 1: 20 and 1: 10 respectively. The total duration of the nosocomial outbreak was 10 days, from October 21 to October 30. Because of this outbreak, the hospital stay of the neonates was prolonged by about 6 weeks. Many additional laboratory examinations were necessary and the five patients were hospitalized for a total of about 42 days. The extra cost of the outbreak was estimated at more than £16 000.
Discussion Echovirus type 30 is one of the most frequently isolated echoviruses. Numerous seasonal outbreaks of aseptic meningitis involving echovirus
M. Chambon et al.
type 30 have been reported in the last 20 years in Cuba,8 New York,9 Taïpei,10 Japan11 and Belgium12 indicating active circulation of the serotype worldwide. Several outbreaks were reported recently in Europe: in 1991 and 1994 in Norway,13 1994–1995 in Spain,14 1995 and 1996 in Poland15 and in 1996 in Switzerland.16 In 1996 the most frequently isolated enterovirus serotypes in France were echovirus types 11, 30 and 6.17 In 1997 outbreaks of echovirus type 30 aseptic meningitis were observed in many areas throughout the country,18 but none were nosocomial outbreaks. Nosocomial enterovirus outbreaks are generally observed at the peak of seasonal outbreaks in the general population,3 especially those involving nurseries. Our report is novel because of the time that elapsed between the end of the outbreak (beginning of August) and hospital admission of the index patient (October 2). In that interval no case was observed, although the virus probably continued to circulate in the general population. The mother of the index patient excreted echovirus type 30 for at least 19 days after delivery. This was not surprising as enteroviruses are known to be shed in the upper respiratory tract for 1–3 weeks and in the faeces for up to 8 weeks after primary infection.3 The nosocomial outbreak occurred two and a half weeks after admission of the index patient. Both the mother and child were still excreting the virus. Hence they may both have been the source of the nosocomial outbreak and responsible for contaminating neonates in the same area of the unit. Neonates are especially at risk of enterovirus nosocomial infections. Direct contact with faeces through activities such as nappy changing, the resistance of enteroviruses in the environment, the high viral load in patient faeces, which can reach 106 infectious particles per gram,19 have led to a strict hygiene policy that combines rigorous hand washing by staff, and virological surveillance of babies to monitor any cross infection.20 It is important that surfaces and equipment should be cleaned and disinfected with effective virucidal products. We
PCR diagnosis of Echo 30 in a neonatal unit
have shown that about 30% of children admitted to hospital shed viruses in stools on the first day of admission,21 and that about 10% of them shed enteroviruses. Similarly 3% of pregnant women may excrete enteroviruses at term during epidemics.22 The outcome of neonatal infection is strongly influenced by the presence or absence of passively acquired maternal antibody specific for the infecting enterovirus serotype.22 Three of the five patients in our study were preterm, and serological results showed that none of the five was protected by passive immunization from their mother. A protective titre has been recently evaluated by Abzuz et al. to be at least 1 : 800.23 Clinical findings in nosocomial enterovirus outbreaks in neonates present different clinical patterns, sometimes linked to particular serotypes.2 Briefly, coxsackie viruses B1, B2, B3 have both neurotropism and cardiotropism, and myocarditis has been responsible for deaths reported in neonates. Coxsackie viruses A (A5, A14), B5 and echovirus types 11, 17, 30, 31 cause meningitis, sometimes severe. Echovirus types 18, 19, 33 cause fever without apparent focus. Echovirus types 22 and 23 are responsible for respiratory symptoms.24 Despite its frequent isolation, echovirus type 30 is not often involved in nosocomial outbreaks in neonates. The most frequently incriminated of all serotypes is echovirus type 11, with eleven outbreaks reported. There is not therefore a strict correlation between the frequency of isolation in the population of a particular serotype and its involvement in nosocomial outbreaks. For example, echovirus type 9 is the most prevalent of all enteroviruses, but is rarely involved in nursery outbreaks.2 Whatever the serotype, the diagnosis of an enterovirus infection in neonates is sometimes overlooked: infections may be under-recognized or under-investigated. In published reports of nosocomial outbreaks, especially of meningitis, there have been up to 20 children involved. The reports have mainly been from the 1960s and 1970s, when rapid diagnosis was not available and diagnosis of enterovirus infections relied on culture of virus from the
67
cerebrospinal fluid or from a peripheral site. This technique is not very sensitive, especially for CSF specimens, and takes 6 to 7 days to detect viral growth. It is laborious and time consuming. Hence diagnosis has often been delayed or not performed and, as a result, hygiene measures to limit the number of affected children not always implemented early enough. By contrast, the use of standardized molecular tools such as PCR as a routine allowed a rapid diagnosis in our index case and a definite diagnosis in patients 1 and 4, ruling out the possibility of recurrence of bacterial infection. As it is not always possible to have sufficient single rooms to isolate neonates, swift diagnosis is important. It allows rapid alert and early implementation of hygiene measures to prevent nosocomial spread. We stress the usefulness of such molecular tools irrespective of the seasonal periods of enterovirus infections. Sporadic cases have been described for all known serotypes and nosocomial infections are always a risk, especially in susceptible patients, even in intermediate periods. Antigenic and genetic variations in enteroviruses have to be taken into consideration and newly isolated strains have to be analysed at a molecular level to determine their evolution and to identify possible new virulence determinants. The molecular analysis of these outbreak strains is currently being performed in our laboratory.
Acknowledgements We would like to thank Mr Jeffrey Watts for his revision of the English manuscript. This work was supported in part by a grant from Ministère de I’Education Nationale, de la Recherche et de la Technologie to Hélène Peigue-Lafeuille.
References 1. Moore M. Enteroviral disease in the United States 1970–1979. J Infect Dis 1982; 146: 103–108.
68
2. Cherry JD. Enteroviruses. In: Remington JS, Kline JO, eds. Infectious diseases of the fetus and newborn infant. 4th ed. Philadelphia: WB Saunders 1995; 404–446. 3. Modlin JF. Update on Enterovirus infections in infants and children. Adv Pediatr Infect Dis 1997; 12: 155–180. 4. Dowsett EG. Human enteroviral infections. J Hosp Infect 1988; 11: 103–115. 5. Rotbart HA, Sawyer MH, Fast S, et al. Diagnosis of enteroviral meningitis using the polymerase chain reaction with a colorimetric microwell detection essay. J Clin Microbiol 1994; 32: 2590–2592. 6. Rotbart HA. Reproducibility of Amplicor Enterovirus PCR test results. J Clin Microbiol 1997; 35: 3301–3302. 7. Lim KA, Benyesh-Melnick M. Typing of viruses by combinations of antiserum pools. Application to typing of enteroviruses (coxsackie and ECHO). J Immunol 1960; 84: 309–317. 8. Bello M. Viral meningoencephalitis caused by enterovirus in Cuba from 1990–1995. Rev Argent Microbiol 1997; 29: 176–183. 9. Leonardi GP, Greenberg AJ, Costello P, Szabo K. Echovirus type 30 infection associated with aseptic meningitis in Nassau County, New York, USA. Intervirology 1993; 36: 53–56. 10. Hsu CM, Chen JM, Huang LM, Lee PI, Kao CL, Lee CY. Outbreak of aseptic meningitis in Taipei in spring 1993. J Formos Med Assoc 1995; 94: 14–18. 11. Yamashita K, Miyamura K, Yamadera S, et al. Enteroviral aseptic meningitis in Japan, 1981–1991. Jpn J Med Sci Biol 1992; 45: 151–161. 12. Druyts-Voets E. Epidemiological features of entero non-poliovirus isolations in Belgium 1980–94. Epidemiol Infect 1997; 119: 71–77. 13. Gjoen K, Bruu AL, Orstavik I. Intratypic variability of echovirus types 30 in part of the VP4/VP2 coding region. Arch Virol 1996; 141: 901–908.
M. Chambon et al.
14. Lopez Alcala MI, Rodriguez Priego M, de la Cruz Morgado D, Barcia Ruiz J M. Outbreak of meningitis caused by echovirus type 30. An Esp Pediatr 1997; 46: 237–240. 15. Majda-Stanislawska E, Kuydowicz J, Barczak D. Epidemic of echo type 30 virus meningitis in children from the Lodz region in the years 1995 and 1996. Przegl Epidemiol 1997; 51: 267–274. 16. Gorgievski-Hrisoho M, Schumacher JD, Vilimonovic N, Gemann D, Matter L. Detection by PCR of enteroviruses in cerebrospinal fluid during a summer outbreak of aseptic meningitis in Switzerland. J Clin Microbiol 1998; 36: 2408–2412. 17. Aymard M, Thouvenot D, Foray S, Lina B. National Reference Center For Enteroviruses and Hepatite A. Lyon France. Annual Report 1996. 18. Aymard M, Chomel JJ, Lina B, Thouvenot D. National Reference Center For Enteroviruses and Hepatite A. Lyon France. Annual Report 1997. 19. Melnick JL, Rennick V, Infectivity titers of enterovirus as found in human stools. J Med Virol 1980; 5: 205–220. 20. Breuer J, Jeffries DJ. Control of viral infections in hospitals. J Hosp Infect 1990; 16: 191–221. 21. Peigue-Lafeuille H, De Champs C, Laveran H et al. Survey of viruses of the gut in 10477 children admitted to general paediatric wards from 1981 to 1986. J Hosp Infect 1989; 14: 375–378. 22. Modlin JF, Polk BF, Horton P et al. Perinatal echovirus 11 infection: risk of transmission during a community outbreak. N Engl J Med 1981; 305: 368–371. 23. Abzug MJ, Keyserling HL, Lee ML, et al. Neonatal enterovirus infection: Virology, serology, and effects of intravenous immune globulin. Clin Infect Dis 1995; 20: 1201–1206. 24. Ehrnst A, Eriksson M. Echovirus type 23 observed as a nosocomial infection in infants. Scand J Infect 1996; 28: 205–206.
An outbreak due to Echovirus type 30 in a neonatal unit in France in 1997: usefulness of PCR diagnosis M. Chambon, J. L. Bailly, A. Béguet, C. Henquell, C. Archimbaud, J. Gaulme, A. Labbé, G. Malpuech and H. Peigue-Lafeuille Departments of Paediatrics and Clinical Virology, Centre Hospitalier Universitaire, Clermont-Ferrand, France Summary: Between February and August 1997, 53 patients with enterovirus meningitis were hospitalized in Clermont-Ferrand, France. All but one were children. Echovirus type 30 was involved in 70% of cases with identified serotype. The outbreak ceased on August 8. Two months later, a neonate was admitted to the neonatal unit with an echovirus type 30 meningitis thought to be acquired at delivery. Twenty days later a nosocomial outbreak of echovirus type 30 involving five neonates occurred. Two of them presented with meningitis and two with febrile seizure; One was asymptomatic. The retrospective examination of the maternal sera in a neutralization test, using the index case strain as a source of antigen, showed that none of the neonates was passively immunized before hospitalization. The use of genome detection in cerebrospinal fluid allowed rapid diagnosis and infection was contained by re-inforcing hygiene measures. Prospective examination of stools in the neonatal and paediatric units showed no further occurrences of the disease. No sporadic case was observed in the general population. Hence, nosocomial infections can occur a long time after an outbreak in the general population; rapid diagnosis with molecular tools is useful both for a definite diagnosis in patients already hospitalized, and to act as a rapid alert, even in intervals between seasonal outbreaks. © 1999 The Hospital Infection Society
Keywords: Enterovirus; echovirus type 30; nosocomial infection; neonate; rapid diagnosis; PCR.
Introduction Human enteroviruses include 66 known serotypes divided into five subgroups: polioviruses, coxsackieviruses A, coxsackieviruses B, echoviruses and non-assigned enteroviruses 68 to Received 16 December 1998; Revised manuscript accepted 16 June 1999 Corresponding author: Hélène Peigue-Lafeuille, Service de Virologie, Faculté de Médecine, 28 Place Henri Dunant, 63001 Clermont-Ferrand Cedex, France.
0195–6701/99/090063 + 06 $12.00
71. Non-polio enteroviruses – mainly coxsackieviruses and echoviruses – are responsible for numerous human illnesses ranging from the insignificant to the severe, especially in children.1 Neonates are at risk of serious and sometimes fatal diseases.2,3 Human enteroviruses have a dangerous potential for nosocomial spread.4 In temperate climates, enterovirus infections occur every year, mainly in the summer and autumn. When outbreaks of non-polio enteroviruses occur in the general population, neonates are especially at risk. © 1999 The Hospital Infection Society
64
We describe an outbreak of echovirus type 30 infections in a neonatal unit in France in October 1997, which occurred two months after a large outbreak of echovirus type 30 meningitis in the population. The rapid diagnosis of the disease by nucleic acid detection of enterovirus genome was useful in limiting nosocomial spread.
Patients and methods All patients admitted to hospital during 1997 with a suspicion of enterovirus meningitis were prospectively investigated. Diagnosis was based on viral examination including polymerase chain reaction (PCR) assay of cerebrospinal fluid (CSF) for the rapid diagnosis of enterovirus meningitis with the AMPLICOR EV test Kit (Roche Molecular Systems, Paris, France). Several regions of the 5′ end of the enterovirus genome are highly conserved among different enterovirus serotypes and so a single pair of oligonucleotide primers can amplify most known enteroviruses with the exception of echoviruses 22 and 23.25 The AMPLICOR EV test kit is a rapid (5 h) PCR-based assay in which reverse transcription and amplification are performed in a one step reaction using rth polymerase. Carryover contamination is prevented by dUTP and AmpErase.6 Detection is performed colorimetrically on a microwell plate. Virus isolation was performed on CSF, throat swab and stool specimens using human oral epidermoid carcinoma (KB) and human lung fibroblast MRC5 cell lines (Bio-Mérieux, Marcy l’Etoile, France). Lim-Benyesh-Melnick antiserum pools were used for enterovirus identification by seroneutralization tests.7 Neutralizing antibody assays (IgG and IgM) were performed by a micromethod with MRC5 cells and the corresponding isolated virus as a source of antigen. The paediatric department consists of six different units that are geographically separated. The neonatal unit is divided into two areas, a paediatric intensive care unit and, immediately adjacent, the neonatal unit itself.
M. Chambon et al.
Results Outbreak of enterovirus meningitis in the general population
Between February and August 1997, 59 patients with enterovirus meningitis were admitted to hospital. In 53 of the 59, an enterovirus was isolated and identified. Enterovirus detection by PCR was performed in 48 patients and was positive in 92% (N = 44) of cases. For six patients, PCR in CSF was the only positive result; either cultures were negative (four cases) or specimens not obtained (two cases). The serotypes involved are shown in Table 1. In 37 cases (70%) echovirus type 30 was identified and in eight cases (15%) echovirus type 6. None of these patients was hospitalized in the neonatal unit. The youngest patient was 1 month old. The outbreak of echovirus type 30 occurred suddenly, with a peak in June (23 cases) and a falling off in July (12 cases). The last case was observed on August 8; this patient, a 39-year-old male, was the only adult victim of the enterovirus outbreak. No further echovirus type 30 case was observed until October 2. Nosocomial infections due to echovirus type 30 in the neonatal unit
The index patient was a 5-day-old male, born at term, who was hospitalized on October 2, 1997 because of neonatal meningitis (Figure 1). He was diagnosed as having enterovirus meningitis. Table I Enteroviruses isolated during an outbreak of meningitis in 1997 in the general population. Serotype
n
%
Echovirus 30 Echovirus 6
37 8
70 15
Echovirus 9 Echovirus 11 Echovirus 24
1 1 1
Coxsackie B3 Coxsackie B4 Coxsackie A9
2 1 2 53
] ]
5·6
9·4 100
PCR diagnosis of Echo 30 in a neonatal unit
65
Patient 5
Patient 4
Patient 3
Patient 2
Patient 1
Index patient
15/ 09
01/ 10
15 / 10
31 / 10
15 / 1 /1997
Figure 1 Nosocomial outbreak due to echovirus type 30. Rectangles represent duration of hospital stay(s). Grey triangles represent positive detection of enterovirus genome by PCR in CSF (index patient, patients 1 and 4). Full black circles represent isolation of echovirus type 30 in stools.
PCR in CSF (containing 400 WBC) was positive on admission. Echovirus type 30 was isolated from a stool six days later. His mother had presented with fever and upper respiratory tract infection at delivery. Her stool specimen obtained 14 days after admission of her child was positive for echovirus type 30. The strain was used as an antigen source, and specific antibody titres of IgM (1 : 40) and IgG (1 : 160) were detected in her serum 14 days after birth. Bacteriological cultures were negative in both mother and child. The boy made a complete recovery and was discharged on day 34. Patients 1, 2 and 3 in the nosocomial outbreak were triplets, three females born by caeserean section at 34 weeks of gestation, who had been admitted on September 14, because of prematurity and hypotrophy. Patient 1 was first admitted to the intensive care unit because of deterioration of respiratory function. She was intubated on admission. When her condition improved, on September 18, she was transferred to the same area of the neonatal unit as the index patient and her two sisters. On day 40, October 24, she became febrile. A lumbar
puncture showed pleiocytosis (123 WBC, 68% lymphocytosis). PCR was positive for enterovirus and echovirus type 30 was isolated in a stool specimen taken on the same day. Her condition slowly improved and she left hospital on November 7, after a hospital stay that was two weeks longer than initially predicted. Patients 2 and 3 had similar features. They were admitted to the neonatal unit and discharged on October 24, the day their sister became febrile. They were both readmitted on October 30, because of a 3-day recurrence of fever and seizure to one of the paediatric units. Stool culture on admission was positive with isolation of echovirus type 30. They both improved. Examination of the CSF was normal and they were discharged 8 days later. Patient 4 was a two-day-old male, born at term who was admitted to the neonatal unit on October 8 for neonatal infection due to Proteus mirabilis isolated from both mother (vaginal secretions) and child (meconium, navel, gastric fluid specimens). Cefotaxime, amoxycillin and amikacin were given for 10 days and then discontinued. One day later, he was febrile
66
and became lethargic with loss of appetite. Recurrence of his Proteus mirabilis infection was suspected. Bacterial examinations remained negative. On October 21, a lumbar puncture showed pleiocytosis (65 WBC), and PCR was positive for enterovirus. Stool specimens obtained the same day and eight days later were positive for echovirus type 30. The enterovirus meningitis prolonged the hospital stay by two weeks. Patient 5 was a four-day-old male, born at term, who was admitted for neonatal meningitis due to E. coli, isolated from his urine and from maternal specimens. After three weeks of therapy (cefotaxime, amoxycillin and amikacin), he improved. Before he was discharged, systematic bacteriological examination of stools, urine, CSF and throat swabs were negative but echovirus type 30 was isolated from stools on October 24. He was discharged on October 31. Systematic stool examinations during the same period of eight patients in the neonatal area where the other cases had been observed, and 88 in the paediatric unit to which patients 2 and 3 had been readmitted, revealed no further occurrence of echovirus type 30 infection. Hospital staff were not examined. Serum specimens from the mothers of patients 1–5 were analysed for the presence of echovirus type 30 antibodies, with virus from the index patient as a source of antigen. They showed low and non-significant titres of 1: 20, 1: 20 and 1: 10 respectively. The total duration of the nosocomial outbreak was 10 days, from October 21 to October 30. Because of this outbreak, the hospital stay of the neonates was prolonged by about 6 weeks. Many additional laboratory examinations were necessary and the five patients were hospitalized for a total of about 42 days. The extra cost of the outbreak was estimated at more than £16 000.
Discussion Echovirus type 30 is one of the most frequently isolated echoviruses. Numerous seasonal outbreaks of aseptic meningitis involving echovirus
M. Chambon et al.
type 30 have been reported in the last 20 years in Cuba,8 New York,9 Taïpei,10 Japan11 and Belgium12 indicating active circulation of the serotype worldwide. Several outbreaks were reported recently in Europe: in 1991 and 1994 in Norway,13 1994–1995 in Spain,14 1995 and 1996 in Poland15 and in 1996 in Switzerland.16 In 1996 the most frequently isolated enterovirus serotypes in France were echovirus types 11, 30 and 6.17 In 1997 outbreaks of echovirus type 30 aseptic meningitis were observed in many areas throughout the country,18 but none were nosocomial outbreaks. Nosocomial enterovirus outbreaks are generally observed at the peak of seasonal outbreaks in the general population,3 especially those involving nurseries. Our report is novel because of the time that elapsed between the end of the outbreak (beginning of August) and hospital admission of the index patient (October 2). In that interval no case was observed, although the virus probably continued to circulate in the general population. The mother of the index patient excreted echovirus type 30 for at least 19 days after delivery. This was not surprising as enteroviruses are known to be shed in the upper respiratory tract for 1–3 weeks and in the faeces for up to 8 weeks after primary infection.3 The nosocomial outbreak occurred two and a half weeks after admission of the index patient. Both the mother and child were still excreting the virus. Hence they may both have been the source of the nosocomial outbreak and responsible for contaminating neonates in the same area of the unit. Neonates are especially at risk of enterovirus nosocomial infections. Direct contact with faeces through activities such as nappy changing, the resistance of enteroviruses in the environment, the high viral load in patient faeces, which can reach 106 infectious particles per gram,19 have led to a strict hygiene policy that combines rigorous hand washing by staff, and virological surveillance of babies to monitor any cross infection.20 It is important that surfaces and equipment should be cleaned and disinfected with effective virucidal products. We
PCR diagnosis of Echo 30 in a neonatal unit
have shown that about 30% of children admitted to hospital shed viruses in stools on the first day of admission,21 and that about 10% of them shed enteroviruses. Similarly 3% of pregnant women may excrete enteroviruses at term during epidemics.22 The outcome of neonatal infection is strongly influenced by the presence or absence of passively acquired maternal antibody specific for the infecting enterovirus serotype.22 Three of the five patients in our study were preterm, and serological results showed that none of the five was protected by passive immunization from their mother. A protective titre has been recently evaluated by Abzuz et al. to be at least 1 : 800.23 Clinical findings in nosocomial enterovirus outbreaks in neonates present different clinical patterns, sometimes linked to particular serotypes.2 Briefly, coxsackie viruses B1, B2, B3 have both neurotropism and cardiotropism, and myocarditis has been responsible for deaths reported in neonates. Coxsackie viruses A (A5, A14), B5 and echovirus types 11, 17, 30, 31 cause meningitis, sometimes severe. Echovirus types 18, 19, 33 cause fever without apparent focus. Echovirus types 22 and 23 are responsible for respiratory symptoms.24 Despite its frequent isolation, echovirus type 30 is not often involved in nosocomial outbreaks in neonates. The most frequently incriminated of all serotypes is echovirus type 11, with eleven outbreaks reported. There is not therefore a strict correlation between the frequency of isolation in the population of a particular serotype and its involvement in nosocomial outbreaks. For example, echovirus type 9 is the most prevalent of all enteroviruses, but is rarely involved in nursery outbreaks.2 Whatever the serotype, the diagnosis of an enterovirus infection in neonates is sometimes overlooked: infections may be under-recognized or under-investigated. In published reports of nosocomial outbreaks, especially of meningitis, there have been up to 20 children involved. The reports have mainly been from the 1960s and 1970s, when rapid diagnosis was not available and diagnosis of enterovirus infections relied on culture of virus from the
67
cerebrospinal fluid or from a peripheral site. This technique is not very sensitive, especially for CSF specimens, and takes 6 to 7 days to detect viral growth. It is laborious and time consuming. Hence diagnosis has often been delayed or not performed and, as a result, hygiene measures to limit the number of affected children not always implemented early enough. By contrast, the use of standardized molecular tools such as PCR as a routine allowed a rapid diagnosis in our index case and a definite diagnosis in patients 1 and 4, ruling out the possibility of recurrence of bacterial infection. As it is not always possible to have sufficient single rooms to isolate neonates, swift diagnosis is important. It allows rapid alert and early implementation of hygiene measures to prevent nosocomial spread. We stress the usefulness of such molecular tools irrespective of the seasonal periods of enterovirus infections. Sporadic cases have been described for all known serotypes and nosocomial infections are always a risk, especially in susceptible patients, even in intermediate periods. Antigenic and genetic variations in enteroviruses have to be taken into consideration and newly isolated strains have to be analysed at a molecular level to determine their evolution and to identify possible new virulence determinants. The molecular analysis of these outbreak strains is currently being performed in our laboratory.
Acknowledgements We would like to thank Mr Jeffrey Watts for his revision of the English manuscript. This work was supported in part by a grant from Ministère de I’Education Nationale, de la Recherche et de la Technologie to Hélène Peigue-Lafeuille.
References 1. Moore M. Enteroviral disease in the United States 1970–1979. J Infect Dis 1982; 146: 103–108.
68
2. Cherry JD. Enteroviruses. In: Remington JS, Kline JO, eds. Infectious diseases of the fetus and newborn infant. 4th ed. Philadelphia: WB Saunders 1995; 404–446. 3. Modlin JF. Update on Enterovirus infections in infants and children. Adv Pediatr Infect Dis 1997; 12: 155–180. 4. Dowsett EG. Human enteroviral infections. J Hosp Infect 1988; 11: 103–115. 5. Rotbart HA, Sawyer MH, Fast S, et al. Diagnosis of enteroviral meningitis using the polymerase chain reaction with a colorimetric microwell detection essay. J Clin Microbiol 1994; 32: 2590–2592. 6. Rotbart HA. Reproducibility of Amplicor Enterovirus PCR test results. J Clin Microbiol 1997; 35: 3301–3302. 7. Lim KA, Benyesh-Melnick M. Typing of viruses by combinations of antiserum pools. Application to typing of enteroviruses (coxsackie and ECHO). J Immunol 1960; 84: 309–317. 8. Bello M. Viral meningoencephalitis caused by enterovirus in Cuba from 1990–1995. Rev Argent Microbiol 1997; 29: 176–183. 9. Leonardi GP, Greenberg AJ, Costello P, Szabo K. Echovirus type 30 infection associated with aseptic meningitis in Nassau County, New York, USA. Intervirology 1993; 36: 53–56. 10. Hsu CM, Chen JM, Huang LM, Lee PI, Kao CL, Lee CY. Outbreak of aseptic meningitis in Taipei in spring 1993. J Formos Med Assoc 1995; 94: 14–18. 11. Yamashita K, Miyamura K, Yamadera S, et al. Enteroviral aseptic meningitis in Japan, 1981–1991. Jpn J Med Sci Biol 1992; 45: 151–161. 12. Druyts-Voets E. Epidemiological features of entero non-poliovirus isolations in Belgium 1980–94. Epidemiol Infect 1997; 119: 71–77. 13. Gjoen K, Bruu AL, Orstavik I. Intratypic variability of echovirus types 30 in part of the VP4/VP2 coding region. Arch Virol 1996; 141: 901–908.
M. Chambon et al.
14. Lopez Alcala MI, Rodriguez Priego M, de la Cruz Morgado D, Barcia Ruiz J M. Outbreak of meningitis caused by echovirus type 30. An Esp Pediatr 1997; 46: 237–240. 15. Majda-Stanislawska E, Kuydowicz J, Barczak D. Epidemic of echo type 30 virus meningitis in children from the Lodz region in the years 1995 and 1996. Przegl Epidemiol 1997; 51: 267–274. 16. Gorgievski-Hrisoho M, Schumacher JD, Vilimonovic N, Gemann D, Matter L. Detection by PCR of enteroviruses in cerebrospinal fluid during a summer outbreak of aseptic meningitis in Switzerland. J Clin Microbiol 1998; 36: 2408–2412. 17. Aymard M, Thouvenot D, Foray S, Lina B. National Reference Center For Enteroviruses and Hepatite A. Lyon France. Annual Report 1996. 18. Aymard M, Chomel JJ, Lina B, Thouvenot D. National Reference Center For Enteroviruses and Hepatite A. Lyon France. Annual Report 1997. 19. Melnick JL, Rennick V, Infectivity titers of enterovirus as found in human stools. J Med Virol 1980; 5: 205–220. 20. Breuer J, Jeffries DJ. Control of viral infections in hospitals. J Hosp Infect 1990; 16: 191–221. 21. Peigue-Lafeuille H, De Champs C, Laveran H et al. Survey of viruses of the gut in 10477 children admitted to general paediatric wards from 1981 to 1986. J Hosp Infect 1989; 14: 375–378. 22. Modlin JF, Polk BF, Horton P et al. Perinatal echovirus 11 infection: risk of transmission during a community outbreak. N Engl J Med 1981; 305: 368–371. 23. Abzug MJ, Keyserling HL, Lee ML, et al. Neonatal enterovirus infection: Virology, serology, and effects of intravenous immune globulin. Clin Infect Dis 1995; 20: 1201–1206. 24. Ehrnst A, Eriksson M. Echovirus type 23 observed as a nosocomial infection in infants. Scand J Infect 1996; 28: 205–206.