One hundred and sixty cases of legionnaires’ disease were reported to the PHLS Communicable Disease Surveillance Centre in 1995. Twenty cases (13%) were known to have died. Ninety cases (56%) were associated with travel (in the United Kingdom or abroad), four were associated with a stay in hospital, and the remaining 66 were presumed to have acquired infection in the community. One hundredand twenty-three cases (77%) occurred sporadically. Three community outbreaks and one outbreak at an industrial site were detected in England and Wales. One outbreak and five clusters were detected among visitors to Turkey, Spain, and Italy. Seven cases and one outbreak of non-pneumonic legionellosis were also reported. Cases of travel associated legionnaires’ disease continue to account for the largest proportion of the total reported in 1995 and the number of hospital acquired cases continues to decline.
A cause for concern in 1995 was a fall in the proportion of cases diagnosed by culture of the organism (from 16% in 1994 to only 9% in 1995). This corresponded with a small increase in the proportion of cases diagnosed solely by detection of antigen toL. pneumophila serogroup 1 in urine.

Introduction

Since 1979, a total of 2800 cases of legionnaires’ disease in residents of England and Wales have been reported to

the surveillance scheme coordinated by the PHLS Communicable Disease Surveillance Centre (CDSC). Among these, 1281 cases (46%) were associated with travel, 235 (8%) were acquired in hospital, and the remaining 46% were presumed to have been acquired in the community.

The numbers of cases reported fell between 1989 and 1991 following a peak of 279 cases in 1988 (figure). After 1991 the numbers of reports rose a little but have remained stable in the past two years. The increase has been associated with a rise in the proportion of cases associated with travel reported to the scheme.

Methods

The National Surveillance Scheme for Legionnaires’ Disease in residents of England and Wales was established at CDSC in 1979. Cases are ascertained through voluntary reports from microbiologists in PHLS or other hospital laboratories, consultants in communicable disease control (CCDCs), and other health care workers. Clinical, epidemiological, and microbiological information is obtained from a questionnaire sent to reporting doctors, and from the PHLS Legionella Reference Unit. The case definition for legionnaires’ disease used by CDSC requires clinical evidence of pneumonia and laboratory confirmation of legionella infection 1 : the main methods of diagnosis are demonstration of the presence of antibody against legionella in serum; culture of the organism from sputum, lung tissue, or blood; and detection of legionella antigen in urine. Information about cases that meet the surveillance definition are entered onto a national database, which is then searched for clusters in time and place and premises with which cases have been associated before. Definitions for classifying cases of legionella infection have been described 1 .
Cases associated with travel are reported to the European Surveillance Scheme for Travel Associated Legionnaires’ Disease run by the European Working Group on Legionella Infections (EWGLI) and coordinated at CDSC since July 1993. On learning of each new case, the coordinator searches the European dataset for other cases linked to the same place of accommodation. CDSC informs the EWGLI collaborator in the presumed country of infection of each new case and alerts all European collaborators and the World Health Organisation (WHO) of clusters of cases associated with the same hotel or resort. WHO in turn is asked to inform the ministry of health in the country concerned. A cluster is defined as two or more cases who stayed at the same accommodation and became ill within six months of each other.
The term ‘outbreak’ is reserved for travel associated clusters in which a common source of infection is suspected or identified and whose local investigators report the results of their investigations and describe their control measures to the surveillance scheme.

Results

Epidemiological features

One hundred and eighteen (74%) of the 160 cases in 1995 were male with an age range of 19 to 87 years; 42 (26%) were female (age range of 24 to 75 years). The median age of both males and females was 52 years , as it was in 1994 2 . The highest rate of infection (7 cases per million population) was reported in the 10 year age band of cases aged 60 to 69 years and compares with an overall rate of 3.1/1 000 000 for all ages. Twenty cases (13%) were reported to have died, the lowest case fatality rate since 1988. Most cases acquired infection during the summer months, which correlates with the large number of travel associated cases reported to the surveillance scheme in recent years.

Microbiological diagnosis

Clinical isolates were obtained from only 14 cases (9%), 13 of which were identified as L. pneumophila serogroup (sg)i1 and one L. bozemanii sg 1. Seventy-three cases were confirmed by seroconversion by L. pneumophila sg 1 yolk sac antigen (fourfold or greater rise in antibody titre), 45 by single high antibody titre, and 27 were presumptively diagnosed by the detection of L.ipneumophila sg 1 antigen in urine using an enzyme linked immunosorbent assay (ELISA). One case was diagnosed by direct immunofluorescence. Thirty-three cases were diagnosed by more than one method. In total, 50 cases had a positive urinary antigen detection test (with or without an additional method of diagnosis).

Travel associated infection

Ninety cases (57%) gave a history of travel in the ten days before developing symptoms of legionnaires’ disease, 86iof whom were associated with travel abroad (table), and the remaining four with overnight travel within the United Kingdom. The commonest destinations of travel associated cases were Spain and the Spanish islands (37 cases), Turkey (15), Italy (6), and Greece and the Greek islands (4). Ten of the 86 cases who had travelled abroad were reported to have died. One outbreak and five clusters were linked to travel abroad. An outbreak of 11 cases of legionnaires’ disease at a hotel in Kusadasi, Turkey between May and September 1995 included seven cases from England and Wales, two from Scotland, one from the Netherlands, and one from New Zealand. No deaths occurred. Cases were ascertained through the European Surveillance Scheme for Travel Associated Legionnaires’ Disease. At the end of September all visitors from the United Kingdom were withdrawn from the hotel and, in October 1995, two scientists from the PHLS were invited to help local staff obtain epidemiological and environmental information from the hotel, take water samples from several sites for microbiological analysis, and advise the hotel management on control measures. L. pneumophila sg 1 was obtained from one clinical isolate, a sample from the hotel’s water system and from the pond water of the hotel’s cooling tower. Further characterisation carried out by the PHLS Legionella Reference Unit showed that isolates from the clinical specimen and the environmental sample from the hotel’s water system were indistinguishable by monoclonal antibody (mAb) subgrouping and restriction fragment length polymorphism (RFLP) subtyping. These results, combined with the epidemiological information, provided strong evidence that the hotel’s water system was the source of infection.

Sixteen cases occurred in the five clusters identified in 1995. Two separate clusters were associated with the Spanish resort of Salou. Three cases arose in the first cluster, in May 1995, which was linked to an apartment block where two previous cases of legionnaires’ disease, (an English and a Scottish holidaymaker) were reported to have stayed in 1994. The second, a cluster of seven cases, was linked to a group of apartments where four women and three men were affected in September and October and where two Dutch cases had stayed in 1993. Two clusters, each of two cases, were associated with hotels in Cambrils and Benidorm in Spain. Both hotels had been linked with previous cases: in one an outbreak of 26 cases arose in 1980 3 . The fifth cluster comprised two men who had shared the same holiday itinerary and had stayed at hotels in France and Italy.

Community acquired infection

Three community outbreaks and one outbreak at an industrial site were reported in 1995. An outbreak of four cases (three men and one woman aged 36 to 68 years) occurred between January and March in the Reading area. L. pneumophila sg 1 was isolated from one patient, who died. Epidemiological and environmental investigations were carried out to determine the extent of the outbreak and possible sources of infection. No further cases were identified, despite the help of local hospitals and general practitioners. All environmental samples from cooling towers and other water systems were negative for legionella, apart from one cooling tower, which yielded L.ipneumophila sg 6. No links were identified between the four cases and no source of infection was found. An outbreak in 1983 and 1984 affected 16 people in Reading and was associated with a cooling tower in the town centre 4 .
Two clusters comprising a total of eight cases occurred in the community in London between February and May 1995. Five cases were associated with an area of south London bounded by the Elephant and Castle, Waterloo, London Bridge, and Lewisham, two were associated with an area within the City of London, and one was linked to both areas. Detailed interviews with cases and mapping of their movements in these parts of London failed to identify any places or journeys common to the cases. Environmental investigations of cooling towers in the immediate areas revealed no poorly maintained systems. No definite links could be identified within or between the two clusters and no common sources of infection were detected.

Two cases arose in an outbreak at an industrial site in the West Midlands in November 1995. Both cases, one of whom died, were employed at the same workplace. Water samples were collected from the site and examined. One cooling tower on premises adjacent to the site was found to be in poor condition and subsequently closed, but no Legionella species were isolated. An outbreak of non-pneumonic legionellosis due to L. micdadei occurred between January and March 1995. It was identified when eight children aged 12 to 13 years developed a flu-like illness after visiting a leisure complex in the north west of England. A case finding exercise identified a further 32 possible infections. All were investigated for evidence of legionella infection and water samples were taken from the leisure complex’s swimming pool and whirlpool systems. All environmental samples were negative for Legionella species by conventional culture techniques. The PHLS Legionella Reference Unit showed serological evidence of L. micdadei infection in 13 cases (eight with a fourfold rise in titre). Subsequently, L. micdadei was isolated from stored whirlpool samples by amoebic co-cultivation 5 and culture at Leeds Public Health Laboratory. The outbreak control team and Carlisle Public Health Laboratory identified deficiencies in the maintenance of the whirlpool bath. Control measures were implemented and recommendations were made to the leisure complex staff in line with guidelines issued for the maintenance of spa pools (Marsh J et al. Proceedings of EWGLI, 11th meeting, Oslo, 1996).

Hospital acquired infection

Four sporadic cases of hospital acquired legionellosis, two of whom died, were reported in 1995. In one fatal case, reported from Northern and Yorkshire region, L.ibozemaniiesg 1 was isolated from the sputum of a patient admitted with a cerebral tumour. The second case was a renal transplant recipient with impaired immunity and the third was a patient with endocarditis who had been admitted to a cardiac unit. The fourth case, an elderly man who had had a total hip replacement and died, was linked to a Welsh hospital previously associated with cases of legionnaires’ disease.

Discussion

The annual reported incidence of legionnaires’ disease in residents of England and Wales in 1995 is the same as was reported in 1994 2 . On the whole, the epidemiological features in 1995 are similar to those reported in recent years 2,6 except that a greater proportion of the travel related cases were associated with outbreaks or clusters abroad in 1995.

The rise in the proportion of cases associated with travel seen in 1994 has been maintained in 1995; 56% of cases in 1994 and 1995 had travelled in the 10 days before developing symptoms of legionnaires’ disease compared with 51% in 1993 2 . A gradual increase in the number of cases associated with travel to Turkey also continued in 1995. The annual rate of infection per million travellers to Turkey from England and Wales was 17.3 in 1995 compared with 9.7 in 1994; this was mainly attributable to the outbreak in Kusadasi, which was the largest travel associated outbreak in Europe since 1987. This outbreak highlighted the importance of rapid and regular communication between the country of infection and the countries of residence of cases about the control measures and investigations being carried out at the hotel. It also emphasised the difficulty tour operators may have in making informed decisions on whether to withdraw clients from hotels suspected to be associated with one or more cases of legionnaires’ disease.

The European Council Directive for Package Travel, which was implemented in 1996, now makes tour operators responsible for the acts and omissions of their suppliers and they can be held liable if they knowingly place clients in hotels that may be associated with legionella infection. New procedures for informing tour operators about travel related cases of legionnaires’ disease have now been introduced in the UK 7 and discussed by EWGLI collaborators in the European surveillance scheme with a view to introducing similar procedures in their own countries (Joseph CA, et al. Proceedings of EWGLI, 11th meeting, Oslo, 1996).

The number of cases associated with travel to Spain and the Spanish islands, which had fallen in 1994, rose again in 1995. Thirty-seven cases were reported in 1995 and, although Spain is the commonest destination of British tourists who holiday abroad (about seven million in 1995), the rate of infection per million travellers to Spain rose from less than three per million in 1994 to over four in 1995. Twenty-three of the 37 cases were linked to clusters or to accommodation with which cases of legionnaires’ disease had been associated previously, suggesting that breakdown of prevention measures may have occurred and highlighting the importance of regular checks and maintenance of hotel water systems. Recurrent contamination of hotel water systems has been reported 8 .

Two encouraging trends were observed in 1995. The first was the continuing decline in the reported number of cases of legionnaires’ disease associated with a stay in hospital. No hospital acquired outbreaks were reported for the second consecutive year and the number of sporadic cases acquired in hospital was the lowest since 1984.
Effective design, monitoring, and maintenance of hospital water systems, including cooling towers, in accordance with official guidelines may have contributed to this decline 9,10 , but continued reporting remains essential.

The second encouraging trend was a fall in the case fatality rate from 17% in 1993 and 1994 to 13% in 1995. Earlier detection of legionella infection through greater use of rapid diagnostic tests and treatment with appropriate antibiotics may have contributed to this fall.

The decreasing proportion of cases acquired in the community, including a fall in the number of large community outbreaks which was first observed in 1993 6 , continued in 1995. The introduction of statutory notification of cooling towers and evaporative condensers 11 has resulted in more effective monitoring of their maintenance records by environmental health departments and earlier recognition of potential sources of infection when community cases have been reported. These measures may have contributed to the fall in the number of community acquired cases reported.

We are concerned, however, by a gradual decline in the proportion of cases diagnosed by culture of the organism from 22% in 1992, through 17% and 16% in 1993 and 1994, to 9% of cases in 1995. This trend may reflect the recent reported decline in the numbers of both hospital acquired cases and outbreaks in England and Wales and the increase in cases associated with travel abroad. Diagnosis by culture of the organism is more likely to occur in hospital acquired cases and in outbreaks and less likely among travel associated cases. In addition, the proportion of cases diagnosed by detection of antigen in urine has increased. In 1994, 24% of cases had a urine ELISA test and this increased to 33% in 1995. This new test is a valuable addition to existing diagnostic methods. It provides rapid diagnosis, allowing appropriate antibiotic treatment to be given earlier, and sample collection is non-invasive. Rapid diagnosis allows earlier investigation of outbreaks and implementation of control measures. The test is cheaper than other methods. Clinical isolates, however, are required for mAb subgrouping and RFLP subtyping of the organism and for comparison with environmental isolates when investigating sources of infection. This cannot be achieved when clinical cases are diagnosed using serological or antigen detection methods. Prompt completion and return of surveillance questionnaires is essential for CDSC’s follow up of legionella infections reported to the national surveillance scheme, in order to identify cases linked in time or place to the same possible source of infection, and to quickly feed back information to local public and environmental health investigators. Dates of onset of infection and dates and places of travel, including hotel names and room numbers if available, are needed to alert the presumed country of infection when travel associated cases of legionnaires’ disease are reported and to detect clusters of cases associated with the same hotel or accommodation. The patient’s occupation and place of work may be relevant for the detection of community acquired legionnaires’ disease and information about coexisting diseases is important when hospital acquired infections are reported.

Acknowledgements

The Respiratory Diseases Section at CDSC is grateful to all microbiologists and consultants in communicable disease control and others for their continued support of the surveillance scheme and for obtaining the epidemiological and microbiological information necessary for further action. We also thank Dunja Car for her invaluable administrative support.

References

1. Saunders CJP, Joseph CA, Watson JM.
Investigating a single case of legionnaires’ disease: guidance for consultants in communicable disease control. Commun
Dis Rep CDR Rev 1994; 4: R112-4.

2. Joseph CA, Hutchinson EJ, Dedman D, Birtles RJ,Watson JM, Bartlett CLR.
Legionnaires’ disease surveillance: England and Wales 1994.
Commun Dis Rep CDR Rev 1995; 5: R180-3.

3. Bartlett CLR, Swann RA, Casal J, Canada Royo L, Taylor AG, et al.
Legionella: proceedings of the 2nd International Symposium. Washington DC:
American Society of Microbiology, 1984: 237-9.

4. Anderson P, Bartlett C, Cook G, Woodward M.
Legionnaires’ disease in Reading – possible association with a cooling tower.
Community Med 1985; 7: 202-7.

5. Fallon RJ, Rowbotham TJ.
Microbiological investigations into an outbreak of Pontiac fever due to Legionella micdadei associated with the use of a whirlpool.
J Clin Pathol 1990; 43: 479-83.

6. Joseph CA, Dedman D, Birtles R, Watson JM, BartlettiCLR.
Legionnaires’ disease surveillance: England and Wales, 1993.
Commun Dis Rep CDR Rev 1994; 10: R109-11.

7. CDSC. Legionella and the European Council Directive for Package Travel.
Commun Dis Rep CDR Wkly 1996; 6: 9.

8. Bartlett CLR, Watson JM, Joseph CA, Bezzant M.
Travel-associated legionnaires’ disease. Proceedings of the second conference on international travel medicine,
Atlanta, Georgia, 1991.

9 NHS Estates.
Health Technical Memorandum 2040: the control of legionellae in healthcare premises – a code of practice (volume 1 management policy; volume 2 design considerations; volume 3 validation and verification; volume 4 operational management; volume 5 good practice guide).
London: HMSO, 1993.

10. Health and Safety Executive.
The control of legionellosis including legionnaires’ disease. London: HMSO, 1993.

11. Health and Safety Executive.
The notification of cooling towers and evaporative condensers regulations 1992.
Statutory Instrument, 1992 No 2225, Health and Safety.
London: HMSO, 1992.

L H Newton BSc, C A Joseph MSc, E J Hutchinson MSc,J M Watson FRCP, C L R Bartlett FFPHM
PHLS Communicable Disease Surveillance Centre

T G Harrison PhD
PHLS Central Public Health Laboratory

Article written by
L H Newton, C A Joseph, E J Hutchinson, T G Harrison, J M Watson,
C L R Bartlett
CDR Review 1996