Legionnaires' disease in England and Wales in 2007
Four hundred and forty-two cases of Legionnaires' disease in residents of England and Wales, with onset of symptoms in 2007, have been reported to the national surveillance scheme. Over half of these cases (231) were acquired in the community, 12 were hospital acquired and the remaining 199 cases were associated with travel, either abroad or in the United Kingdom. The highest proportions of cases were diagnosis by urinary antigen detection. Twenty-four outbreaks/clusters were detected: 10 associated with community acquired infection, one associated with a hospital and 13 with clusters/outbreaks abroad. An overall case fatality rate of 12% was reported.
Cases of Legionnaires' disease, a bacterial infection characterised by pneumonia, have been reported to the National Surveillance Scheme for Legionnaires' disease in residents of England and Wales since its establishment at the PHLS Communicable Disease Surveillance Centre (CDSC) in 1979. In 2003 the scheme was transferred to the Health Protection Agency, Centre for Infections (HPA, CfI) from where it is now administered. From 1980 to 1999 between 111 and 280 cases were reported each year. There has since been a notable rise in case numbers, reaching a high of 551 cases in 2006. Between 1980 and 2007 a total of 6391 cases have been reported to the surveillance scheme. Forty-eight percent of these cases originate from within the community, 41% are associated with travel abroad, 6% travelled within the UK and the remaining 5% are considered hospital acquired. This paper describes the epidemiology of cases with onset of symptoms in 2007.
Diagnosed cases of Legionnaires' disease are reported to local Health Protection Units (HPUs). Healthcare workers are then asked to obtain a 14 day history of the case's movements, prior to onset of symptoms, to identify risk factors and possible sources of infection and then to complete a national surveillance scheme reporting form which provides clinical, epidemiological and microbiological details for each case.
The Atypical Pneumonia Unit within the HPA CfI Respiratory and Systemic Infections Laboratory (RSIL) provides confirmatory diagnostic testing of reported cases. It uses an in-house enzyme immuno-assay (EIA) test on urine samples, and carries out immunofluorescence antibody testing (IFAT) on serum samples (in the presence of campylobacter blocking fluid to eliminate cross reaction). Positive isolates can be sub-typed, using sequence based typing, to further differentiate the Legionella strain. In appropriate circumstances Polymerase Chain Reaction (PCR) can be carried out for the detection of L. pneumophila DNA .
Reported cases that meet the case definitions (table 1) are entered onto the national database where they are checked for links in time and/or place with previous cases. Travel associated cases are reported to the European Surveillance Scheme for Travel Associated Legionnaires' disease (EWGLINET) .
Table 1: HPA case definitions for legionella infections * When submitted to the Respiratory and Systemic Infections Laboratory, all positive serum specimens are examined by the IFAT test in the presence of campylobacter blocking fluid, to eliminate cross reactions.
** This is a working definition: the decision to follow up cases is made locally. Results
The National Surveillance Scheme for Legionnaires' disease in residents of England and Wales received 442 completed surveillance forms for cases with onset of symptoms in 2007 (figure 1). A further 14 cases, (seven females and seven males), were lost to follow-up. Three hundred and twenty-eight (74%) cases were males aged 19 to 94 years and 114 (26%) were females aged 25 to 91 years. Collectively the median age of the cases was 59 years with a male to female ratio of 2.9:1.
Figure 1: Cases of Legionnaires' disease in residents of England and Wales by case category, 1998-2007
During 2007 the main method of diagnosis for 370 (83.7%) of the 442 cases was by urinary antigen detection. Culture was the main method of diagnosis in 65 (14.7%) cases of which 59 were L. pneumophila serogroup 1, three as serogroup 5, one as serogroup 3, one L. pneumophila and one L. longbeachae. In 56 of the culture confirmed cases urinary antigen detection was also confirmed. RSIL confirmed one case as L. pneumophila sg 1 by four fold rise (FFR) in serum antibody levels. The remaining six cases of L. pneumophila were presumptive cases; four diagnosed by single high titre (SHT) and two by polymerase chain reaction (PCR) (table 2) . The Atypical Pneumonia Unit in RSIL received and confirmed 76% of all reported cases.Table 2: Method of diagnosis for cases of Legionnaires' disease in residents of England and Wales 2007
Case distribution and outcomes
The distribution of cases by month of onset in 2007 saw a gradual rise beginning in April and peaking at 90 cases with onset in July (figure 2). This is an earlier peak than that observed in 2006, where cases peaked at 118 in August and at 66 cases in each of August and September in 2005 .
Figure 2: Cases of Legionnaires' disease in residents of England and Wales by month of onset of symptoms in 2007
The overall case fatality rate has increased in recent years from 8.5% in 2005, 9.6% in 2006 to 12% in 2007 (table 3), however the increase in absolute number of deaths is not significant (Chi squared = 2.182, 1 d.f., P = 0.1396). Of the 53 reported deaths in 2007, 33 (62.3%) were in males aged between 36 and 89 years and 20 (37.7%) in females aged between 44 and 85 years. The highest proportions of deaths were reported in nosocomial cases (25%) and the lowest in travel UK cases at 3%. Regionally, the North West region experienced the highest proportion of deaths (17.8%) whilst Wales reported the lowest proportion of deaths at 7.4%.Table 3: Cases and (number of deaths) by category of exposure Case category and clusters/outbreaks
NosocomialTwelve cases (2.7%) were categorised as hospital acquired: 10 males, three of whom died and two females. Eight cases were classified as 'definite' nosocomial cases, in accordance with the national surveillance scheme's case definitions, one as 'probable' and three as 'possible' nosocomial cases (table 1). The proportion of nosocomial cases in 2007 was similar to 2005 where 2% of cases were reported as hospital acquired.
Three of the nosocomial cases, one of whom died, were associated with an outbreak at a hospital in the East of England . Two of the three cases were confirmed by RSIL as L. pneumophila serogroup 1, subgroup Philadelphia and were indistinguishable to the strains obtained from the hospital's hot and cold water system. The same hospital was involved in an outbreak between 2002 and 2004 involving three patients, two of whom died.Community
Two hundred and thirty-one cases (52.3%) in 2007 acquired infection in the community, 176 males and 55 females, with a sum of 39 deaths. A quarter of these cases had onset of symptoms in July when the number of community cases reached its peak (figure 2). The proportion of community cases is slightly down on 2006 figures which were exceptionally high but figures were up by about 5% on the 2005 community cases.
Ten community clusters ranging in size from two to eight cases (and five deaths) were identified in 2007, including one associated with an industrial site in the North East. The industrial cluster involved two cases where the place of work was the only common site linking them together but despite extensive investigations, no definitive source was found for the cluster. The largest urban area in England and Wales is London [6,7] which experienced four community clusters, the highest number among the 10 regions studied.
The largest of these clusters involved five boroughs in southeast London during July, when a total of eight cases were reported to the national surveillance scheme with no deaths. No links were identified between the cases other than time and place. A cluster of five cases in east London were identified with onset of symptoms spanning June and July, along with two further London clusters involving three cases each; the first in northwest London where onset of symptoms for the cases ranged from June to August, resulting in one death and the second in north London where all cases had onset of symptoms during July. Investigations for these three clusters were unable to determine a definitive source apart from an association in time and space.
The West Midlands and South East regions also have large urban areas and experienced two clusters each. The two West Midlands clusters began in June as a single cluster of cases in the Black Country but as case numbers increased it became apparent that there was a group of five cases clustered together in the southwest area of the Black Country (cluster BC1), whilst the remaining five cases were randomly spread across the north of the Black Country (cluster BC2). One death was reported in cluster BC1 and none in cluster BC2. A number of sites were sampled: only one cooling tower in the area of the BC1 cluster tested positive for L. pneumophila serogroup 1 but was found to be a different strain to the clinical isolate obtained from one of the cases . Thus no source was identified for either of the clusters.
The first of the two clusters in the South East region was detected in Berkshire and involved three cases with onset of symptoms during June and resulted in one death. With no clinical isolates, a definitive source could not be determined from any of the sites sampled. The second South East region community cluster involved five cases in Surrey with onset between June and July; no deaths were reported . Again, numerous sites were sampled but no source was identified.
The last of the 10 community clusters was detected in the Yorkshire and Humber region involving two cases with onset of symptoms in June, resulting in one death. Although environmental investigations were conducted on the case homes, workplace, and local cooling towers, an associated source was not identified.Incidence rates by region
With the exception of London, all regions in England and Wales experienced a fall in incidence rates when compared with 2006 figures. East Midlands saw the greatest fall of 0.93 cases per 100,000 population. The London rate was relatively stable with only a 0.03 increase in cases per 100,000 population. A comparison of 2007 figures with 2005 showed that all but the North East and South West regions increased their incidence rate with the South East showing the greatest increase at a rate of 0.43 cases per 100,000 population.Figure 3: Incidence rates of Legionnaires' disease by region of residence per 100,000 population 2007*
*Denominators taken from ONS 2001 census data. . Map excludes five cases from 'unknown' or 'other' regionsStrength of evidence for defining clusters/outbreaks
A total of 11 clusters/outbreaks were identified in England and Wales during 2007 compared with six and seven outbreaks/clusters identified in 2005 and 2006, respectively. When these are categorised by the strength of evidence towards a source; four of the community clusters (two of the large London clusters and the two West Midlands clusters) were found to be epidemiologically linked (no clinical or environmental isolates were obtained, cases only linked by time and place). Three of the community clusters had investigations leading to a probable source; either a clinical isolate or an environmental isolate from the most likely source was obtained but not both, which would have enabled matching of strains to be carried out. Three clusters had both clinical and environmental isolates obtained as part of the investigations but in each incident the patient isolate was distinguishable from the most likely environmental source(s). It was only in the nosocomial outbreak that investigations identified a source with a strong link between the epidemiology and microbiology (indistinguishable clinical and environmental isolates).Travel abroad
One hundred and sixty-six cases (37.6%) were reported to have travelled abroad during the 2-10 day period before onset of symptoms. One hundred and seventeen of the travel cases were males, eight of whom died and 49 were females, one of whom died. Thirty-three percent of travel abroad cases had onset in June and September. The June peak was a month earlier than in previous years. Overall the proportion of cases who travelled abroad increased by almost 9% compared with 2006, but decreased in excess of 4% compared with 2005.
All travel associated cases were reported to the surveillance scheme EWGLINET run by the European Working Group for Legionella Infections (EWGLI). The scheme identified and investigated, using the European Guidelines , 13 clusters each involving at least two cases from England and Wales linked to hotels and other accommodation sites. Ten of the clusters occurred in six European countries: Bulgaria, France, Greece, Italy, Malta and Turkey and three in non-European countries: China, Tunisia and United States of America. In June 2007, three British nationals, one of whom died, were among 18 cases of Legionnaires' disease associated with a community outbreak in Spain. The source of this outbreak was a cooling tower linked to an ice-rink at a sports centre [12,13].
Outbreaks involving UK residents were also reported on two cruise ships. The first in a ship that sailed around the Baltic islands resulting in nine cases and one death associated with a cruise that sailed between 15 and 30 July 2007. Five passengers were hospitalised in Sweden, four were hospitalised in Kent, where the cruise was prematurely terminated . Further passengers with symptoms were admitted at various hospitals around England. Three cases were confirmed by culture as L. pneumophila, three as L. pneumophila by PCR and three as L. pneumophila by single high titre. Environmental sampling showed the presence of L. pneumophila in the ship's water system.
The second cruise ship sailed around the Spanish islands. Two passengers from England and Wales and one from Scotland became ill between June and November 2007. A further case was identified in a British resident with onset of symptoms in January 2008. No source has been identified.Travel UK
The remaining 33 cases (7.5%) travelled within the UK during the 2-10 days prior to onset of symptoms: 25 males, one of whom died, and eight females. Seven of these cases occurred in July producing a peak in UK travel associated cases. No outbreaks or clusters were found to be associated with travel UK cases during 2007.Discussion
The number of cases reported to the National Surveillance Scheme for Legionnaires' disease in residents of England and Wales with onset of symptoms in 2007 was 442, a fall of 19.8% compared with 551 cases in 2006 but an increase of 24.5% from 2005 (figure 1). 2006 was an exceptional year where the annual number of cases was much higher than those typically observed, the greatest proportion of which were community acquired cases. Unusual meteorological conditions, where a period of sustained high temperature was followed by intense rainfall and humidity, may have contributed to the rise in community infections . These specific weather patterns were not observed in 2007 [16,17] and may have contributed to the 30.8% fall of community acquired infections relative to the previous year.
In 2007 the peak month for onset of cases acquired in the UK was July rather than August or September as typically observed in other years. Again, meteorological conditions may have influenced this finding since a warm spring was followed by periods of high rainfall, especially in those regions where high rainfall and high case numbers coincided [16,17].Despite the fall in case numbers compared with 2006, the number of deaths remained the same (53 deaths), resulting in a rise in the case fatality rate from 9.6% in 2006 to 12% in 2007. No hypothesis is currently available to explain this rise in deaths, particularly in the community acquired cases where 39 (73.6%) of the 53 deaths were reported, an increase from the 58.5% observed in 2006. Conversely, deaths associated with travel UK cases fell from seven (13.2%) in 2006 to one case (1.9%) in 2007. As in previous years the proportion of deaths by age group increased with increasing age.
The proportion of cases identified using culture has increased from 8.5% in 2005 to 14.7% in 2007, whilst serological testing has almost ceased to be used as a primary method of diagnosis and urinary antigen testing has remained static. The rise in obtaining cultures is encouraging, particularly because of the possibility to identify typing information of strains and the ability to exploit this information to identify possible sources of infection. Cultures are needed when clusters or outbreaks are detected in order to support environmental investigations but only a third of all cases involved in such incidents had samples taken for culture of the organism.
The number of cases of Legionnaires' disease reported to the national surveillance scheme may continue to rise in the future, as evidenced by trends in recent years. It is also possible that as a consequence of climate change we may see more frequent years similar to 2006 where sudden unexpectedly high numbers of cases are generated. Such increases will undoubtedly have an impact on local resources for following up each case and investigating potential sources of infection. Policies and procedures for prevention of legionella infections will be ever more important if the impact of meteorological changes is to be minimised.Acknowledgements
The authors would like to thank all microbiologists, CCDC's, infection control nurses, and others for providing epidemiological and microbiological data on individual cases and for their continued support of the surveillance scheme.References
1. Health Protection Agency. Atypical Pneumonia Unit Atypical Pneumonia Unit
2. European Working Group for Legionella Infections EWGLI
3. Health Protection Agency Epidemiological data
4. Health Protection Agency Cases by Month of Onset 1980-2007
5. BBC News Legionnaires' Disease at Basildon Hospital, Essex
6. Department for Environment, Food and Rural Affairs Rural Definition and Local Authority Classification
7. Department for Environment, Food and Rural Affairs Local Authority Classification Dataset
8. BBC News One dead in Legionnaires outbreak - Dudley
9. BBC News Legionnaires' probe at Pirbright
10. Office of National Statistics. Post Census (2001)
11. EWGLI European Guidelines for Control and Prevention of Travel Associated Legionnaires' disease.
12. HC Info - Outbreak in Spain Legionnaires' Disease Outbreaks 2007-2008
13. Manchester Evening News ex-journalist hit by deadly bug
14. Euro surveillance Outbreak of respiratory infection on a cruise ship. 09 Aug 2007 Vol 12(32)
15. Weather patterns and Legionnaires' disease: a meteorological study Epidemiol. Infect. 2008 Nov 19:1-10.
16. Met Office June-July 2007 and May-July 2007 record rainfall
17. Met Office 2007 Monthly weather summary.