Respiratory Archives 2008

Legionnaires' disease in England and Wales (1999-2005)
FC Naik, KD Ricketts, TG Harrison, CA Joseph

Abstract

The National Surveillance Scheme for Legionnaires' disease in Residents of England and Wales received 1937 reports of Legionnaires' disease between 1999 and 2005, with an overall case fatality rate of 11%. Forty-seven percent of cases were acquired in the community and 42.6% after travelling abroad; the remainder were associated with travel within the UK or had acquired the infection after a stay in hospital. The main method of diagnosis was urinary antigen detection which was used to diagnose a higher proportion of cases year on year. Three large outbreaks were identified during this period; Barrow-in-Furness which occurred between July and October 2002, Cricket St Thomas between February and March 2003 and Hereford between October and November 2003. The significance of these outbreaks on the annual number of cases reported to the national surveillance scheme is considered.

Introduction

In 1976 an outbreak of unknown respiratory illness occurred among a convention of American Legions at a hotel in Philadelphia [1]. The aetiological agent responsible for this outbreak was identified and described in 1977 and the illness became known as "Legionnaires' disease" (LD).

LD is caused by bacteria of the genus Legionella and has a typical incubation period of two to 10 days, although longer periods have been demonstrated in outbreak situations [2]. The disease is characterised by pneumonic illness with an overall case fatality rate in Europe of 6.4% [3]. However, rates can vary considerably depending on the situation and has reached 31.2% in nosocomial outbreaks [4-6]. Men are more at risk than women to LD but the risk in both sexes' increases with age, especially in those over the age of 50 years [7]. Legionellae can also cause non-pneumonic flu-like illnesses, such as Pontiac Fever, which are not associated with any known case fatality.

Legionellae occur naturally in environmental water sources such as freshwater ponds and creeks. Transmission to humans occurs when the bacteria infect the water systems of buildings and industrial premises and become aerosolised via devices such as showers and air conditioning units.

The National Surveillance Scheme for Legionnaires' Disease in Residents of England and Wales was established by the Public Health Laboratory Service (PHLS), Communicable Disease Surveillance Centre (CDSC) in 1979. In 2003, the Health Protection Agency was formed and management of the scheme transferred to the Centre for Infections, Health Protection Agency (CfI, HPA) [8].

Results from this scheme have been published up to (and including) 1998 [8-10]. Annual tables (1980-2007) are available on the HPA website [11] and a short report on activity in 2006 was recently published [12]. This paper presents results for cases of LD in residents of England and Wales for the period 1999 to 2005 and aims to complete this series of reports on this infection.

Methods

Cases of LD in England and Wales are diagnosed by hospital microbiologists, who notify their local Health Protection Unit (HPU) of the case. The local HPUs obtain a full case history from the patient or their representative which normally covers 14 days before onset, in order to identify any risk factors and potential sources of infection. An enhanced national surveillance form is then completed and the information sent to the respective regional office and onto CfI.

CfI classifies cases of LD in accordance with national case definitions that are documented on the HPA website [11]; criteria for determining the category of case are also defined. All travel associated cases are reported to the European Surveillance Scheme for Travel Associated Legionnaires' Disease (EWGLINET) run by the European Working Group for Legionella Infections (EWGLI) [13].

Data are entered onto the national database, which is then searched for other cases that may be linked in time or place, or for premises with which cases have been associated in the past, in order to detect outbreaks and clusters as outlined by the national case definitions [11].

Enhanced surveillance form
The enhanced national surveillance form records demographic details of the patient, the patient's clinical history and outcome, potential exposures during the two weeks prior to onset of illness (including any overnight stays in hospitals or hotels, campsites, cruise ships or other accommodation sites), microbiological diagnoses and results of environmental investigations.

Diagnostic methods
Specimens from locally diagnosed cases are referred to the Respiratory and Systemic Infection Laboratory (RSIL) at CfI for further investigations. Urine samples are examined using a Legionella pneumophila serogroup 1 specific in-house enzyme immunoassay, clinical isolates are subtyped by monoclonal subgrouping and DNA sequence based typing, and sera are examined using an indirect immunofluorescence antibody test (IFAT) in the presence of campylobacter blocking fluid, to eliminate cross reactions [14].

Results

Epidemiology

Cases and outcomes
Between 1999 and 2005, the National Surveillance Scheme for Legionnaires' Disease in Residents of England and Wales received reports of 1937 cases of LD. There has been a notable increase in case numbers from 180 in 2000 to 355 cases reported with onset in 2005 (figure 1).

Figure 1. Case numbers and case fatality rate of Legionnaires' disease in residents of England and Wales, 1980 - 2005

Cases ranged in age from four to 103 years with a median age of 58 years. The most common age group was 50 to 59 years. Age standardised rates using ONS England and Wales mid-year population estimates [15] shows that the rate of infection increases almost proportionally with age, with the highest rate occurring in the 60 to 69 years age group but declining thereafter (figure 2). The male to female ratio was 2.9:1. Two hundred and sixteen deaths were reported, an overall case fatality rate of 11% (range 15% - 8.5%), (figure 1).

Figure 2. Age standardised rates for cases of Legionnaires' disease per 100,000 population, 1999 - 2005.

ONS Annual Population Denominator Estimates, 1999-2005

Cases of LD often show a marked seasonal trend with a peak in August or September (figure 3). In 2002, the peak in July was due to the outbreak in Barrow-in-Furness in Cumbria [16], the largest outbreak of LD detected in England and Wales to date. Similarly, a peak in cases during February and November 2003 marked two outbreaks, one in Somerset [17] the other in Hereford [18].

Figure 3. Number of cases of Legionnaires' disease by month of onset of symptoms, 1999 - 2005

Regional incidence rates
Regional standardised rates using denominators from the ONS mid-year population estimates [15] show that from 1999 through to 2005 the rate fluctuates in all regions but with an overall upward trend, such that the 2005 rate in five out of the 10 regions is double that of 1999 (table 1).

Over the whole period, East Midlands had the highest median incidence rate of 0.7 cases per 100,000; London had the lowest at 0.3. The 2002 peak in the North West is attributable to the Barrow-in-Furness outbreak, similarly the peaks in the West Midlands during 2003 and London in 2005 represent outbreaks involving more than 10 cases.

Table 1. Legionnaires' disease incidence rates by region of residence, per 100,000 of population, 1999 - 2005

Region (by residence)	1999	2000	2001	2002	2003	2004	2005	Total	Mean	Median
East Midlands	0.70	0.38	0.55	0.45	0.75	0.91	0.79	4.53	0.65	0.70
East of England	0.22	0.41	0.26	0.39	0.35	0.54	0.47	2.64	0.38	0.38
London	0.31	0.18	0.26	0.32	0.38	0.39	0.77	2.62	0.37	0.33
North East	0.27	0.39	0.20	0.43	0.55	0.71	0.67	3.22	0.46	0.43
North West	0.31	0.31	0.22	2.54	0.53	0.54	0.51	4.96	0.71	0.51
South East	0.31	0.30	0.22	0.40	0.59	0.50	0.51	2.85	0.41	0.40
South West	0.20	0.24	0.32	0.50	0.54	0.67	0.53	3.02	0.43	0.50
Wales	0.59	0.45	0.41	0.48	0.65	0.41	0.85	3.83	0.55	0.48
West Midlands	0.38	0.40	0.53	0.66	1.22	0.64	0.90	4.73	0.675	0.64
Yorkshire & Humber	0.77	0.56	0.62	0.44	0.46	0.87	0.84	4.56	0.65	0.62

Category of Case
Almost half of all cases reported between 1999 and 2005 were community acquired (47%, 909 cases). Ninety four (10.3%) of these cases died. Nine hundred and seventy one (50%) cases were 'travel associated', of which 828 were categorised as 'travel abroad' (42.6%) of whom 83 (10%) died and 143 as 'travel UK' (7.4%) of whom 14 (9.8%) died. Fifty seven cases were classified as nosocomial (3%) and resulted in 25 (44%) deaths, (figure 4).

Figure 4. Category of cases of Legionnaires' disease in residents of England and Wales, 1999 - 2005

The travel associated cases visited a total of 75 different countries, (table 2). Spain was associated with the highest number of cases, followed by UK (England, Wales and Scotland) and Italy. Eight hundred and twenty-five cases (85%) visited European countries including 89 cases who visited more than one European country in the two to 10 days before onset. Of the remaining 146 cases, 26 visited more than one non-European country.

Table 2: Countries visited by travel-associated cases of Legionnaires' disease in residents of England and Wales, 1999-2005

COUNTRY OF TRAVEL	Number of cases 1999-2005	COUNTRY OF TRAVEL	Number of cases 1999-2005
More than one European country	74	INDIA	7
More than one non-European country	31	IRELAND	3
ANTIGUA	2	ITALY	85
ARCTIC	1	JAMAICA	2
ARGENTINA	1	KENYA	2
AUSTRAILIA	4	KUWAIT	1
AUSTRIA	4	LATVIA	2
BAHAMAS	1	MALAYSIA	5
BARBADOS	1	MALTA	25
BELGIUM	4	MEXICO	17
BELIZE	1	MONTENEGRO	1
BERMUDA	2	MOROCCO	1
BULGARIA	10	NETHERLANDS	4
CANADA	3	NORWAY	1
CARIBBEAN ISLANDS	1	PHILLIPINES	1
CHINA	3	PORTUGAL	13
CROATIA	7	ROMANIA	1
CRUISE	6	RUSSIA	2
CUBA	3	SINGAPORE	1
CYPRUS	16	SLOVENIA	1
CZECH REPUBLIC	3	SOUTH AFRICA	1
DENMARK	1	SPAIN	204
DOMINICAN REPUBLIC	4	SRI LANKA	6
EAST TIMOR	1	SWITZERLAND	3
EGYPT	8	THAILAND	11
ENGLAND , WALES & SCOTL'D	146	TUNISIA	8
FRANCE	63	TURKEY	57
GERMANY	5	UNITED ARAB EMIRATES	5
GREECE	59	USA	34
HONG KONG	1	UZBEKISTAN	1

Microbiology

A diagnosis was established by culture for 210 cases (10.8%), by urinary antigen detection for 1490 cases (76.9%), by a four-fold rise in antibody levels for 107(5.5%) and by a single or standing high serology titre for 126 cases (6.5%). Four cases were diagnosed by 'other' methods, e.g. Polymerase Chain Reaction, Immunofluorescence (0.2%) (figure 5).

Figure 5. Primary method of diagnosis of cases of Legionnaires' disease in residents of England and Wales, 1999 - 2005

Between 1999 and 2005, the overall number of clinical isolates increased resulting in between eight and 13.5% of diagnoses each year. Urinary antigen detection has increased as the primary method of diagnosis from 53% of cases in 1999 to 85.6% in 2005. In contrast, the number of cases diagnosed solely by serology has fallen from 33% to 5.7%.

An isolate was obtained more frequently for nosocomial cases than for other cases; 29.8% compared with 11.3% for community acquired cases, 8% for travel-abroad and 16.8% for travel UK cases.

Clusters and Outbreaks

Between 1999 and 2005, 59 outbreaks or clusters occurred in England and Wales involving 364 cases of LD and 45 cases of non-pneumonic legionellosis; 35 of the pneumonic cases died. Between six and seven clusters/outbreaks occurred each year, with the exception of 2002 (12 clusters/outbreaks) and 2004 (15 clusters/outbreaks) (figure 6).

Figure 6. Number of outbreaks in England and Wales by category, 1999 - 2005

Thirty-eight clusters/outbreaks were linked to community-acquired infection and investigations showed the source of eight to be caused by cooling towers; three on industrial sites and five within the community. Two clusters/outbreaks were caused by a spa pool, one by the hot and cold water system and one by a combination of sources (eg hot water system and spa pool); the remaining 26 (44%) community clusters/outbreaks had no source identified. Of the eight nosocomial outbreaks identified in the years studied, hot and cold water systems were found to be the source in four, three were caused by the hot water system alone and one had no source identified.

Thirteen clusters were associated with hotels in the UK; two were due to the hot and cold water systems, one by the hot water system alone, one by only the cold water system, one by a spa pool, one a humidifier and one by a combination of sources. No source was identified in the remaining six clusters.

Categorisation of the clusters/outbreaks by the strength of evidence towards a source shows that 32 (54.2%) of the 59 clusters/outbreaks (27 community and five travel UK outbreaks/clusters) were epidemiologically linked (no clinical or environmental isolate obtained, cases linked by time and place); 14 (23.7%) outbreaks/clusters (eight community, four travel UK and two nosocomial clusters/outbreaks) had investigations leading to a probable source (either a clinical or an environmental isolate was obtained). The remaining 13 (22%) clusters/outbreaks (six nosocomial, four travel UK and three community clusters/outbreaks) were investigated and identified a source with a strong link between the epidemiology and microbiology (matched clinical and environmental isolates).

In July 2002, the largest outbreak of LD in England and Wales to date occurred in Barrow-in-Furness [16]. One hundred and forty-six cases of LD were diagnosed in UK residents including two Scottish nationals. Seven of these cases died and a further 35 cases of non-pneumonic legionellosis were also identified. Extensive investigations determined that the source of the outbreak was a cooling tower situated in the council-owned "Forum 28" theatre and arts complex in the town centre. Barrow Borough Council pleaded guilty to breaching health and safety laws, and an employee was found guilty of the same charge; both were cleared of manslaughter charges [16].

In February 2003 a spa pool contaminated with legionellae at the leisure centre of a hotel in Cricket St Thomas, Somerset led to two deaths amongst 20 cases of LD and two cases of non-pneumonic legionellosis [17]. In November 2003 an outbreak occurred at a cider plant in Hereford [18] where 27 cases of LD were diagnosed along with one non-pneumonic case of legionellosis; two deaths were reported [19]. The source of this outbreak was found to be a cooling tower at the manufacturing plant. Subsequently the owners of the cider plant were fined £300,000 for their part in the outbreak, and the contracted water treatment company, Nalco, were also fined £300,000 for failing to effectively manage legionella risk at the plant [21].

On 1 July 2002 EWGLI introduced European guidelines for the control and prevention of travel associated LD. Ninety six EWGLI clusters were identified between July 2002 and 2005 that involved two or more residents of England and Wales who travelled abroad in the two to 10 days before onset of infection. These were all investigated according to the procedures outlined in the guidelines [22]. Fourteen clusters were associated with travel to Spain and 11 each to France and Italy. These clusters contribute to an overall total of 338 clusters detected in European residents between July 2002 and 2005.

Discussion

The number of cases of LD in residents of England and Wales reported to the National Surveillance Scheme has shown a notable increase over the period 1999 to 2005. There are a number of factors that may have contributed to this rise in cases.

Prior to the outbreak in Barrow-in-Furness in 2002, a range between 111 and 281 cases was reported to the surveillance scheme each year; this increased to 314 to 355 cases a year between 2003 and 2005. LD suffers from severe under-diagnosis [20]; recent Hospital Episode Statistics for England and Wales show that approximately 97% of all community acquired pneumonias admitted to hospital during 2005 and 2006 were not further classified: a proportion of these can be attributed to legionella [20]. There is also the potential for an increase in awareness of the disease to lead to a real increase in reported numbers. The outbreak in Barrow-in-Furness received considerable media attention, increasing the profile of the disease amongst both clinicians and members of the public, and may have contributed to the subsequent higher rate of diagnosis.

The risk to an individual of contracting LD has been shown to increase with age [3] and the 1999 to 2005 figures for England and Wales support these findings. It is therefore possible that our gradually aging population is contributing to the upward pressure on case numbers. This has obvious resource implications for future levels of a disease which can cause serious long-term morbidity.

A further factor that may have contributed to the rise in cases is climate change. Very warm temperatures followed by a period of high relative humidity are thought to create ideal conditions for the proliferation and survival of the bacteria; investigations into this theory are ongoing [23]. The rise may also be attributed to the dominance and increase of certain strains of legionella circulating in the community.

The proportion of cases diagnosed primarily by urinary antigen has increased over this period, whilst the proportion of cases primarily diagnosed serologically has fallen dramatically. This is largely due to the speed of the results that can be obtained, less than a day, from the urinary antigen test in comparison to those from serological tests which require paired sera no less than seven days apart. The widespread uptake of the urinary antigen test in England and Wales may also be facilitating detection of less seriously ill cases which would previously have gone undiagnosed; this is supported by an overall fall in the case fatality rate, from 15% in 1999 to 8.5% in 2005.

Although the urinary antigen test has given undoubted public health benefits, public health authorities need to encourage clinicians to obtain respiratory samples for culture so that the subtype of the infection can be determined. Clinical isolates can then be matched to environmental isolates to assist in identifying sources of infection. Between 1999 and 2005, 146 cases were diagnosed by both culture and urinary antigen detection and 64 by culture alone. The bias in obtaining cultures for nosocomial patients clearly contributes to determining sources of infection in hospital settings but a higher proportion of culture positive cases in other settings would also contribute to improved linking of clinical and environmental legionella strains.

Between 2003 and 2005 an average of nine clusters/outbreaks were identified annually in England and Wales with a further 46 clusters/outbreaks identified in cases who travelled abroad during the same period. Regardless of the category, site or size of a cluster, there continue to be difficulties in determining sources. No source was identified for 68.4% of community clusters/outbreaks over the period under study, compared with just 12.5% of nosocomial clusters/outbreaks. This is largely because the number of potential sources present in a community setting vastly outnumbers the potential sources in a hospital setting. Testing of environmental samples in conjunction with clinical isolates in a hospital frequently yields results that identify the source of infection.

Investigations into travel associated cases of LD in Europe showed that 52% of cluster accommodation sites were positive for legionella in years 2003 to 2005 but without clinical isolates it cannot be stated categorically that these results have identified the source of infection. Nevertheless, control and prevention measures taken in response to these findings contribute to minimising risk at these accommodation sites.

The number of cases of LD diagnosed in residents of England and Wales has been increasing due to an interaction of improved awareness and diagnostics, environmental factors and other pressures affecting the at-risk population. It is possible that the increased awareness that has resulted from the rise in cases and the large outbreaks from recent years may lead to better control and prevention measures, which in turn may lead to a decrease in cases. However the pressure on case numbers is complex and it is likely that there will continue to be an increase in case numbers before there is any sign of a decrease.

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.

Declaration of Interests
There are no conflicts of interest to declare.

References

1. Sharrar RG. Respiratory Infection - Pennsylvania. Mortality and Morbidity Weekly Report, 1976; 25: 244

2. Den Boer JW, Yzerman EP, Schellekens J, Lettinga KD, Boshuizen HC, Van Steenbergen JE, et al. A large outbreak of Legionnaires' disease at a flower show, the Netherlands, 1999. Emerg Infect Dis. 2002 Jan; 8(1): 37-43. Erratum in: Emerg Infect Dis 2002 Feb; 8(2):180.

3. Ricketts KD, Joseph CA. Legionnaires' disease in Europe 2005-2006. Euro Surveill, 12(12) (2007). Available at: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=680.

4. Ricketts KD, Joseph CA. Legionnaires' disease in Europe 2003-2004. Euro Surveill, 10(12) (2005). Available at: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=588.

5. Joseph CA and Ricketts KD.  The Epidemiology of Legionnaires' Disease.  Chap. 3 in Legionella: Molecular Microbiology, eds. Klaus Heuner and Michelle Swanson. Horizon Press, Norwich, England. 2008 Mar.

6. Joseph CA, Watson JM, Harrison TG, Bartlett CLR. Nosocomial Legionnaires' disease in England and Wales, 1980-92. Epidemiol Infect. 1994, 112: 329-345.

7. Joseph CA, et al. Legionnaires' disease in residents of England and Wales: 1996. Communicable Disease and Public Health, Vol 7: 153 - 159 (1997).

8. http://www.hpa.org.uk/ (accessed 12.11.2008)

9. Joseph CA, et al. Legionnaires' disease in residents of England and Wales: 1997. Communicable Disease and Public Health, Vol 1: 252-258 (1998).

10. Joseph CA, et al. Legionnaires' disease in residents of England and Wales: 1998. Communicable Disease and Public Health, Vol 2: 280-284 (1999).

11. http://www.hpa.org.uk/webw/HPAweb&Page&HPAwebAutoListName/Page/1191942128217?p=
1191942128217 (accessed 26.06.2008)

12. HPA. Legionnaires' disease in residents of England and Wales: 2006 Health Protection Report 2008: 2(5)

13. www.ewgli.org (accessed 12.11.2008)

14. http://www.hpa.org.uk/webw/HPAweb&HPAwebStandard/HPAweb
_C/1195733768734?p=1200577814562 (accessed 12.11.2008).

15. Office for National Statistics. Post Census(2001).

16. Health and Safety Executive. Report of the public meetings into the legionella outbreak in Barrow-in-Furness, August 2002

17. HPA Update on Legionnaires' disease associated with a hotel and leisure centre in Somerset . Commun Dis Rep CDR Wkly , 2003 13(12)

18. Kirrage D, ReynoldsG, Smith GE, Olowokure B. Investigation of an out break of Legionnaires' disease: Hereford, UK 2003. Respir Med. 2007 Aug; 101(8):1639-44.

19. Hunt D, Kirrage D . Outbreak of Legionnaires ' disease in the west of England. Euro Surveill, 7(46) (2003). Available at: Available at: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2325

20. British Thoracic Society. Community-acquired pneumonia in adults in British hospitals in 1982 - 1983: a survey of aetiology, mortality, prognostic factors and outcome. The British Thoracic Society and the Public Health Laboratory Service. Q J Med. 1987 Mar; 62 (239): 195-220.

21. http://news.bbc.co.uk/1/hi/england/hereford/worcs/7483569.stm (accessed 12.11.2008)

22. European Working Group for Legionella Infections. European Guidelines for control and Prevention of Travel Associated Legionnaires' disease, 2002: PHLS London.

23. Ricketts KD, Charlett A, Gleb D, Lane C, Lee JV and Joseph CA. Weather Patterns and Legionnaires' disease: a meteorological study Epidemiol Infect (in press).

Surveillance of influenza and other respiratory viruses in the United Kingdom: October 2007 to May 2008
P Mook, R Pebody, H Zhao, J Ellis, M Zambon, DM Fleming, JM Watson

Summary

 

 

 

 

 

 

 

 

 

 

Influenza activity in the United Kingdom (UK) remained at low levels during the 2007/08 season. Clinical indices of activity remained at the lower end of the range of 'normal seasonal activity' and peaked in early-January 2008 in England and Scotland. Activity remained within baseline levels in Wales for the duration of the season. Virological detection also remained at low levels in England and Wales with influenza A/Solomon Island/3/2006 (H1N1)-like virus identified as the dominant circulating strain although there was some influenza B activity later in the season. During the season, outbreaks, mainly caused by influenza B, were reported in care homes, schools, hospitals and a prison.

During the 2007/08 season, 11% of the influenza A(H1) viruses circulating in the UK were found to be resistant to oseltamivir along with 23 other countries in Europe with oseltamivir resistance rates ranging from 1% to 67%. Oseltamivir resistant influenza viruses were also found outside Europe.

Internationally, the H5N1 avian influenza virus continued to spread in poultry and cause sporadic cases in humans. An H5N1 outbreak in wild birds was reported during January 2008 in Dorset, UK. There were no associated human cases. In June 2008, there was an outbreak of H7N7 avian influenza in poultry in Oxford, UK which also had no associated human cases.

Introduction

In the UK, the activity of influenza and other respiratory viruses in humans is monitored throughout the year, but there is a particular focus on the winter season between October (week 40) and May (week 20). Data is collated from a variety of sources to provide information on circulating influenza strains for early detection of strains with epidemic potential and to contribute towards the decision on influenza vaccine composition for the following year. Surveillance activities also produce timely reports on influenza activity and burden of disease for health professionals, the media and public.

Methods

Influenza surveillance is dependent on both clinical and virological data collected from across the UK. The information sources have been previously described [1, 2]. This season, reporting of surveillance information by these sources continued to be tested via the Department of Health funded Health Protection Informatics (HPI) website in parallel with established methods of reporting. The aim of submitting influenza surveillance data via the HPI is to make the data available to relevant groups in a more timely way.

Results

Clinical

RCGP Weekly Returns Service
The weekly General Practitioner (GP) incidence rate for ILI remained at or near baseline levels (<30 new episodes per 100,000 population) for the duration of the season. (It should be noted that RCGP incidence rates in this report only refer to first or new episodes of infection diagnosed by a GP.)

ILI levels increased to the baseline threshold in week 01/08 at 29.7/100,000 before peaking in week 02/08 at 30.5/100,000 population, which is at the low end of the range of 'normal seasonal activity'. The rate declined back to baseline levels in week 03/08. Activity peaked earlier than in the 2006/07 season (43.7/100,000 in week 07/07) (Figure 1).

Figure 1. RCGP incidence rates for influenza-like illness: England 2007/08 and recent years

The highest ILI incidence rates were among those aged 15-44 years (41.5/100,000 in week 01/08) and those age 45-64 years (37.4/100,000 in week 02/08) (Figure 2). Rates were highest in central England (39.5/100,000 in week 02/08), followed by the south (29.4/100,000 in week 02/08) and the north (15.9/100,000 in week 02/08).

RCGP rates for acute bronchitis peaked in week 01/08 (203.1/100,000). This rate peaked higher than last season (184.5/100,000 in week 01/07). The rates were highest among children aged 0 to 4 years in week 48/07 (514.3/100,000), followed by people age 65 years and over (467.7/100,000) in week 01/08.

Figure 2. RCGP incidence rates for influenza-like illness by age group, 2007/08

Q FLU - HPA and Nottingham University Division of Primary Care
The consultation rates for ILI peaked at 22.1/100,000 in week 01/01. The age specific rates were highest among those aged 15-44 years at 28.7/100,000 (week 01/0) and 45-64 years at 26.7/100,000 (week 01/08), mirroring the RCGP data. The trends were also similar to those indicated by the RCGP data but the overall reported rates were lower. No thresholds have yet been set for Q FLU.

Wales - National Public Health Service (NPHS Wales)
Consultation rates in the sentinel GP scheme in Wales remained within baseline levels (<25/100,000) for the duration of the season. The rate peaked in week 01/08 at 8.5/100,000 (Figure 3) which was lower and earlier than the peak in the 2006/07 season of 17.8/100,000 in week 07/07.

Scotland - Health Protection Scotland
GP consultation rates in the flu spotter scheme for Scotland (Figure 3) remained within baseline levels (<50/100,000) until week 02/08 when the rate peaked at 50/100,000. This was considerably lower than the peak in the 2006/07 season of 158/100,000 in week 02/07.

Northern Ireland - Communicable Disease Surveillance Centre (CDSC Northern Ireland)
This is the eighth year of the enhanced surveillance scheme in Northern Ireland and baseline levels are yet to be established. The combined consultation rate for influenza and ILI peaked in week 05/08 at 64.5/100,000. This peak was lower than the previous season (204/100,000 in week 05/07) (Figure 3).

Figure 3. GP consultation rates for influenza-like illness for England, Wales, Scotland and Northern Ireland, 2007/08

NHS Direct for England and Wales
The national proportion of NHS Direct 'cold/flu' calls for all ages peaked during weeks 52/07 at 1.4% (Figure 4), just above the baseline threshold of 1.2% of total calls [3]. The highest proportion of 'cold/flu' calls was among those aged 15 to 44 years at 2.1% in week 52/07; this value was above the baseline threshold of 1.4% and was consistent with the age distribution for ILI incidence rates in the RCGP scheme. The national proportion of 'fever' calls for all ages peaked during week 52/07 at 4.1% and was highest among those aged 0-4 years at 14.8%, above the baseline threshold of 12.3% of total calls.

Figure 4. NHS Direct proportion of 'cold/flu' and fever calls: 2005/06 and 2007/08

Mortality in England and Wales ( Office for National Statistics)

The weekly total number of estimated deaths due to all causes peaked at 11,954 in week 03/08 (Figure 5). The pronounced troughs in Figure 5 during the 2007/08 season and other seasons are likely to be as a result of public holidays in which registry offices are shut. Increases the following week are likely to be in part due to registering "rebound". This peak was higher and earlier than for the previous season (11,322 in week 08/07). The weekly number of total respiratory deaths peaked at 2388 (week 01/07). The annual estimate of excess mortality due to influenza is calculated using a time series model [4]. Each year it is revised to incorporate data from the current season and the fitted model provides a figure for the most recent season and readjusts the previous years' figures. (Table 1). Only 219 extra deaths due to influenza were estimated for 2007/08.

Figure 5. Deaths by all causes, 2007/08

Table 1 Estimated excess mortality due to influenza in England and Wales

Influenza Season	Number of excess deaths
88/89	1061
89/90	27675
90/91	8426
91/92	5951
92/93	1590
93/94	14079
94/95	2012
95/96	15428
96/97	21170
97/98	-
98/99	17386
99/00	21497
00/01	841
01/02	6913
02/03	6472
03/04	5215
04/05	1914
05/06	-
06/07	-
07/08	219

 

Virological

Virological Surveillance in General Practice
There are two complementary surveillance schemes that use sentinel GPs to collect samples for virological analysis in England. These schemes are the RCGP/HPA Community Surveillance Scheme and the HPA Virological Surveillance Scheme. A subset of the GPs involved in the RCGP sentinel clinical surveillance scheme, also participates in virological surveillance (RCGP/HPA Community Surveillance Scheme), by sending respiratory specimens directly to the Respiratory Virus Unit (RVU), at the Centre for Infections, for testing (Figure 6). Virological data are obtained by analysing influenza strains using antigenic and molecular methods. Table 2 shows the surveillance data obtained via these two surveillance schemes and similar data provided by Northern Ireland and Scotland . Virus isolates are derived from respiratory specimens provided by these sentinel GP schemes or by hospital diagnostic laboratories.

Table 2 Sentinel GP Surveillance Data

Location	Predominant influenza virus type	No. of Samples	No. of influenza positives (%)	Peak week of positives for influenza			No. of positives in peak week (%)
				Flu A		Flu B	Flu A	Flu B
England (RCGP/HPA Community Surveillance Scheme)	A(H1) early B late	1219	294 (24%)	52/07		10/08	34 (52%)	9 (26%)
England (HPA Virological Surveillance Scheme)	A(H1) early B late	509	154 (30%)	01/08	08/08		17 (37%)	7 (50%)
Northern Ireland	B	149	38 (26%)	03/08	10/08		2 (17%)	11 (55%)
Scotland	A(H1) early B late	692	131 (19%)	02/08	10/08		14 (48%)	12 (43%)

Influenza A(H1) and influenza B were the predominant virus types isolated from community samples throughout the UK. All the sentinel GP schemes, with the exception of that in Northern Ireland, indicated that influenza A(H1) dominated in the early part of the season (until approximately week 05/08) and influenza B dominated there after. The peak week of positives for influenza virus in England correlated well with the peak week of clinical activity but the peak weeks in Northern Ireland and Scotland were later than the clinical indicators in those countries.

In total, the RVU identified 387 positive community samples for influenza between week 40/07 and week 20/08. This figure is based on those respiratory specimens submitted directly to the RVU from the RCGP/HPA Community Surveillance Scheme, the HPA Virological Surveillance Scheme, and outbreaks. 73.1% were identified as influenza A viruses while 26.9% were influenza B viruses. These detections were predominantly from those aged 15-44 years.

Figure 6. RCGP clinical and virological influenza surveillance, 2007/08

Reports of Influenza Infection from Hospital Laboratory Reporting
Laboratory reports (NHS and HPA) by week of specimen indicated there were 540 confirmed influenza A infections between week 40/06 and 20/07, peaking in week 01/07. One hundred and fifty six of these were detected by serological test methods. There were 576 confirmed influenza B infections during this same period, peaking in week 14/08. One hundred and seventy three of these were detected by serological test methods.

Between week 40/07 and week 20/08, the RVU identified 501 hospital samples as positive for influenza. 64.9% were identified as influenza A viruses while 35.1% were influenza B viruses. These detections were mainly from children aged less than five years.

The peak week for positive samples for influenza A (combined figures from community and hospital sources submitted to the RVU), was week 01/08, with 87 positives (84 of these were A(H1). This correlates well with the peak week of clinical activity. The peak week for influenza B was later, in week 10/08 with 31 positives. 

Virus Characterisation
Influenza A viruses were the dominant strain this season with 60.2% (327) isolated as A/Solomon Island/3/2006 (H1N1)-like and 4.1% (22) A/Wisconsin/67/05 (H3N2)-like. The remaining 35.7% (194) were influenza B viruses, all of which antigenically similar to the B/Florida/4/2006-like virus (B/Yamagata/16/88 lineage) (Figure 7).

Figure 7. Influenza strains characterised by RVU/VRD from community and hospital sources

 

Antiviral Resistance
Surveillance data of the antiviral drug susceptibility of influenza viruses circulating in the UK during the 2007/08 season showed that 11% (38/347) of A(H1N1) tested viruses were resistant to oseltamivir [6]. These viruses retain sensitivity to zanamivir, amantadine and rimantadine. To date, resistant viruses have been detected in 24 European countries, including the UK with the oseltamivir resistance rates ranging from 1% to 67% [6]. They have also been detected elsewhere outside of Europe [7].

Vaccine Match
Of the influenza viruses characterised during the 2007/08 season, all the influenza A strains showed a good match with the corresponding 2007/08 influenza vaccine components - an influenza A/Solomon Island/3/2006 (H1N1)-like virus and an influenza A (H3) A/Wisconsin/67/05 (H3N2) - like virus. All the influenza B viruses analysed this season belonged to the B/Yamagata lineage (B/Florida/4/2006-like viruses). These viruses were antigenically and genetically distinct from the B/Victoria lineage (B/Malaysia-like virus) which was included in the 2007/08 northern hemisphere influenza vaccine.

Vaccine Uptake Monitoring on Behalf of the Department of Health
Influenza vaccine uptake for the 2007/08 season was 73.5% for those aged 65 years and over in England [5]. All 152 Primary Care Trusts (PCTs) participated in the vaccination programme and 146 PCTs achieved uptake rates of 70% or more.

Outbreaks
The Centre for Infections received 26 reports of acute respiratory illness (ARI) outbreaks from across the UK. Most (69%) were from elderly care homes (affecting those aged over 65 years) with the remainder from schools, hospital wards and a prison. Of the 23 outbreaks with laboratory confirmation, 18 were confirmed as due to influenza B infection (including 14 of the care home outbreaks), two as influenza A(H3), one as RSV, one as rhinovirus and one as parainfluenza type 3. Three of these outbreaks were reported to have associated deaths. These outbreaks were caused by influenza B, parainfluenza and unknown causes.

Influenza activity elsewhere in Europe
Influenza activity first increased above baseline levels in Europe between weeks 48/07 to 51/07 and peaked in most countries between weeks 04/2008 and 08/2008. Clinical consultation rates have been lower than during the 2006-2007 season for the majority of countries. The highest consultation rates have generally been reported in the 0-4 and 5-14 age groups. However, Ireland , Norway and Switzerland also reported elevated consultation rates in the 15-64 age group. During the season, 16,763 influenza viruses were detected in Europe. Of these, 61% were influenza A viruses and 39% influenza B viruses. Of the 3,504 viruses further characterised, 10 were A/New Caledonia/20/99 (H1N1)-like, 2141 were A/Solomon Island/3/2006 (H1N1)-like, 21 were A/Wisconsin/67/2005 (H3N2)-like, 23 were A/Brisbane/10/2007 (H3N2)-like, 1293 were B/Florida/4/2006-like (B/Yamagata/16/88 lineage) and 16 were B/Malaysia/2506/2004-like (B/Victoria/2/87 lineage).

Other viruses
One thousand two hundred and sixteen community samples were submitted to the RCGP/HPA Community Surveillance Scheme, of which 28 (3%) were positive for Respiratory Syncytial Virus (RSV). Five hundred and nine community samples were submitted to the HPA Virological Surveillance Scheme, of which 26 (5%) were positive for RSV, four for adenovirus, one for coronavirus, 10 for human metapneumovirus, one for mycoplasma, two for parainfluenza, 14 for rhinovirus and one sample tested positive for both parainfluenza and rhinovirus.

The total number of NHS and HPA laboratory detected RSV infections was 5439 between weeks 40/07 and 20/08, peaking in week 49/08 at 572 reports. Of these, 4453 (81.9%) were specimens taken from infants aged less than one year. This was an increase in the number of infections reported during the 2006/07 season (3495). Detections of parainfluenza by these laboratories were also monitored. The predominant parainfluenza serotype throughout the 2007/08 season was serotype 3, peaking in week 19/08. The majority of specimens were from those aged four years and under for all serotypes.

Discussion

The season was characterised by the circulation of influenza A, predominantly influenza A/Solomon Island/3/2006 (H1N1)-like , early in the season and influenza B/Florida/4/2006-like virus in the latter part. This was a trend reflected throughout Europe. The last season with substantial influenza B activity was 2005/06 which is in keeping with the concept that influenza B activity tends to occur every two to three seasons.

Confirmed infections of RSV followed the established seasonal pattern during the 2007/08 season, which is characterised by a gradual increase in detections from mid-October, reaching a peak in December/early January. The peak number and total number of detections this year increased compared with the previous two seasons but were still within normal limits.

Influenza activity was low in the UK during the 2007/08 season, making it the eighth consecutive season of low activity. While virological data from sentinel GP surveillance schemes were consistent with the peak weeks of clinical activity in England, those in Northern Ireland and Scotland were some weeks later, perhaps as a result of influenza B circulating later in the season. In England, RCGP incidence rates remained within baseline levels for the majority of the season and normal seasonal activity was only observed in week 02/07. NHS Direct cold/flu calls for the 15 to 44 years age group were a better indicator of influenza activity than fever calls this season: cold/flu calls rose above baseline levels in this age group and the peak of calls correlated with RCGP incidence rates for the same age group. Activity in Wales remained at baseline levels for the duration of the season, as it had in 2006/07, and it peaked one week before that observed in England (week 01/08) while in Scotland, normal seasonal activity was observed for only one week, the same week as in England (week 02/08). Clinical activity in Northern Ireland in 2007/08 peaked in week 05/08 and the updated consultation rate was higher than elsewhere in the UK. Without validated thresholds it is difficult to interpret the significance of the finding.

The most important feature of this season is the finding of the wide-spread oseltamivir resistance of influenza A(H1N1). Throughout Europe and elsewhere, of those influenza A(H1N1) viruses tested for antiviral drug susceptibility during the 2007/08 season, some were found to be resistant to oseltamivir. According to WHO [7], preliminary test results show that the viruses carry the specific neuraminidase mutation (H274Y) that confers oseltamivir resistance in N1, while no sign of adamantane resistance by genetic analysis have been found. The HPA will continue to test the antiviral drug susceptibility of influenza viruses circulating in the UK during 2008/09.

Seasonal Influenza Outbreaks

The majority of laboratory confirmed ARI outbreaks during the 2007/08 season were due to Influenza B, particularly in care homes for those aged 65 years and over. Confirmed influenza B outbreaks in these closed settings (one of which had associated deaths) continued to be reported until relatively late into the season (week 17/08). These outbreaks along with the age specific incidence data for this season challenges the widely held view that influenza B causes only mild disease and predominantly affect school aged children. The age group involved with the majority of outbreaks (those aged 65 and over) did not correspond with the age groups with highest influenza activity, as identified by the RCGP Weekly Returns Service and QFLU (15-44 years) emphasising the importance of undertaking virological sampling of ILI cases in those over 65 years of age in primary care surveillance schemes.

Avian influenza (H5N1) in humans and animals

During 2008, the number of new human cases of avian influenza A(H5N1) virus infection continued to increase as did the number of associated deaths globally. Since the beginning of the outbreak in 2003, 15 countries had confirmed cases: Azerbaijan, Bangladesh, Cambodia, China, Djibouti, Egypt, Indonesia, Iraq, Lao People's Democratic Republic, Myanmar, Nigeria, Pakistan, Thailand, Turkey and Viet Nam. As of 19 June 2008, 385 human cases were reported; of which 243 (63%) were fatal [8]. Bangladesh reported its first human case in 2008. Four other countries (Indonesia, Egypt, Vietnam and China) continued to report new cases and deaths in 2008. The World Health Organization (WHO) continue to categorise the current threat of pandemic influenza as Phase 3: there are human infection(s) with a new subtype, but no new human to human spread or, at most, very limited transmission to a close contact [9].

During the 2007/08 season, outbreaks of avian influenza A(H5N1) in animals continued to be reported from Asia, Middle East, Africa and Europe. Twenty countries reported new animal cases in 2008 (figures as of 29 July 2008), of which five were in Europe (Russia, Switzerland, Turkey, Ukraine and the United Kingdom) [10].

On 10 January 2008, the Department for the Environment, Food and Rural Affairs (Defra) confirmed an outbreak of the highly pathogenic H5N1 avian influenza strain amongst mute swans in the Chesil Beach area of Dorset [11].

Avian influenza (H7N7) in poultry in the UK

On 3 June 2008, Defra confirmed an outbreak of avian Influenza in chickens on premises near Banbury in Oxfordshire, England after preliminary tests were positive for the H7 strain. This was later confirmed as a highly pathogenic H7N7 strain [12, 13]. All birds on the premises were slaughtered as a precautionary measure. There were no human cases of H7N7 associated with these outbreaks, but antiviral drugs were offered to those exposed to the birds without protection and those involved in the incident response.  Seasonal influenza vaccine was offered where appropriate [14].

Vaccine recommendations

The WHO recommended that the components for the 2008-09 vaccine for the northern hemisphere should contain the following:

• an A/Brisbane/59/2007 (H1N1)-like virus
• an A/Brisbane/10/2007 (H3N2)-like virus*
• a B/Florida/4/2006-like virus#

* A/Brisbane/10/2007 is a current southern hemisphere vaccine virus.
# B/Florida/4/2006 and B/Brisbane/3/2007 (a B/Florida/4/2006-like virus) are current southern hemisphere vaccine viruses.

Acknowledgements

The authors are grateful to all microbiologists, consultants in communicable disease control, infection control nurses, and GPs, especially those within the primary care networks, for their contribution to the surveillance schemes. We are appreciative to Joy Field and Praveen Sebastian Pillai at the CFI for their administrative and technical support. Thanks are also due to colleagues at the HPA Primary Care Unit West Midlands, Health Protection Scotland, NPHS Wales and CDSC Northern Ireland for contributing data to this report.

References

1. Goddard NL, Joseph CA, Zambon M, Nunn M, Fleming D, Watson JM. Influenza surveillance in the United Kingdom: October 2000 to May 2001. CDR Wkly 2001; 11 (CDR Suppl):1-7.

2. Mook P, Joseph CA, Ellis J, Zambon M, Fleming DM, Watson JM.  Surveillance of influenza and other respiratory viruses in the United Kingdom: October 2006 to May 2007. Health Protection Report 2008: supplement (see below).

3. Cooper DL, Verlander N, Joseph C, Elliot AJ, Smith GE. Can syndromic thresholds provide early warning of national influenza outbreaks? J Public Health. Advance Access published on November 20, 2007. Available at: http://jpubhealth.oxfordjournals.org/cgi/content/full/fdm068v1

4. Serfling RE. Methods for current statistical analysis of excess pneumonia-influenza deaths. Public Health Reports 1963; 6:494-506.

5. HPA [online]. Vaccination uptake among the 65 years and over and under 65 years at risk in England 2007-08. Available at: http://www.hpa.org.uk/web/HPAweb&HPAwebStandard/HPAweb_C/1195733756886

5. ECDC [online]. Resistance to oseltamivir (Tamiflu) found in some European influenza virus samples. Available at: http://ecdc.europa.eu/Health_topics/influenza/antivirals.html

7. WHO [online]. Influenza A(H1N1) virus resistance to oseltamivir. Available at http://www.who.int/csr/disease/influenza/h1n1_table/en/index.html

8. WHO [online]. Confirmed Human Cases of Avian Influenza A (H5N1) - 19 June 2008. Available at: http://www.who.int/csr/disease/avian_influenza/country/en/

9. WHO global influenza preparedness plan [online]. The role of WHO and recommendations for national measures before and during pandemics. 2005. Available at: http://www.who.int/csr/resources/publications/influenza/GIP_2005_5Eweb.pdf

10. World Organisation for Animal Health (OIE) [online]. Update on avian influenza in animals (type H5). Available at: http://www.oie.int/downld/AVIAN%20INFLUENZA/A_AI-Asia.htm .

11. HPA [online]. Confirmed H5N1 avian influenza in wild swans in Dorset. HPR 2008; 2(2). Available at: http://www.hpa.org.uk/hpr/archives/2008/hpr0208.pdf

12. HPA [online]. UK outbreak of H7N7 avian influenza. HPR 2008; 2(23). Available at: http://www.hpa.org.uk/hpr/archives/2008/news2308.htm

13. Defra [online]. Avian Influenza in Oxfordshire. http://www.defra.gov.uk/animalh/diseases/notifiable/disease/ai/index.htm

15. HPA [online]. Confirmed H7N7 avian influenza in Oxfordshire poultry farm. http://www.hpa.org.uk/webw/HPAweb&HPAwebStandard/HPAweb_C/1212650893247?p=1204186170287

 

Surveillance of influenza and other respiratory viruses in the United Kingdom: October 2006 to May 2007
P Mook, CA Joseph, J Ellis, M Zambon, DM Fleming, JM Watson

Summary

 

 

 

 

 

 

 

 

 

 

 

Influenza activity began late in the United Kingdom and remained at low levels during the 2006/2007 season. Clinical indices of activity remained at the lower end of the range of 'normal seasonal activity' and peaked in mid-February 2007 in England and mid-January in Scotland. Activity remained within baseline levels in Wales for the duration of the season. Virological activity also remained at low levels in England and Wales with influenza A/Wisconsin/67/05 (H3N2)-like virus identified as the dominant circulating strain. During the season, outbreaks, mainly caused by influenza A, were reported in schools, care facilities and military barracks.

Internationally, the H5N1 avian influenza virus continued to spread in poultry and cause sporadic cases in humans. An H5N1 outbreak in poultry was reported during February 2007 in a farm in Suffolk but there were no laboratory confirmed human cases. There was also an outbreak of H7N2 avian influenza in poultry in North West England and North Wales in May 2007 that was associated with four confirmed human cases: two in Wales and two in England.

 

Introduction

In the United Kingdom (UK), the activity of influenza and other respiratory viruses in humans is monitored throughout the year, but there is a particular focus on the winter season between October (week 40) and May (week 20). Information is collated from a variety of sources to provide information on circulating influenza strains for early detection of strains with epidemic potential and to contribute towards the vaccine composition for the following year. Surveillance activities also produce timely reports on the burden of disease for health professionals, the media and public.

Methods

Influenza surveillance is dependent on both clinical and virological data collected from across the UK. The information sources have been previously described [1]. This season, reporting of surveillance information by these sources via the Department of Health-funded Health Protection Informatics (HPI) website was piloted in parallel with established methods of reporting. It is proposed that in future all influenza surveillance data will be submitted by this route with the aim of making the data available to relevant groups in a more timely way.

This new means of submitting data was exploited during the 2006/07 season in an investigation of the usefulness of school absenteeism data as an early warning system for influenza activity in the community [2]. From January to April 2007, 11 schools in five Health Protection Agency (HPA) regions in England submitted weekly returns, via the website, of absenteeism data. These data were then compared with established indices of influenza activity, Royal College of General Practitioners (RCGP) episode incidence rate of influenza-like illness (ILI) and NHS Direct call rate for fever.

NHS Direct established formal syndromic thresholds which were used during the 2006/07 season, based on data from previous seasons [3].

Information was also collected on influenza activity from QFLU, a national primary care surveillance system which provides general practice derived data from over 3300 practices spread throughout the UK , covering a total population of about 22 million patients. It was set up by the University of Nottingham Division of Primary Care and EMIS (a supplier of general practice computer systems) in collaboration with the HPA ( West Midlands ). QFLU, which is part of QResearch, can provide daily influenza data and is specifically set up to support pandemic influenza planning. Although included in weekly reports towards the end of the 2005/06 season, influenza data collected by QFLU were documented by the HPA for the duration of the 2006/07 season.

Results

Clinical

RCGP Weekly Returns Service
The weekly GP incidence rate for ILI remained near baseline levels (< 30 new episodes per 100,000 population) during the season. (It should be noted that RCGP incidence rates in this report only refer to first or new episodes of infection diagnosed by a GP for an illness such as ILI or acute bronchitis.)

ILI levels began to increase over the baseline threshold in week 06/07 before peaking in week 07/07 at 43.7 per 100,000 population, which is at the low end of the range of 'normal seasonal activity'. Rates remained within this range for four weeks before declining back to baseline levels in week 10/07. Activity peaked at much the same time as in the 2005/06 season although these are both relatively late compared to previous years. The peak rate was also similar to that in the 2005/06 season (42.2 per 100,000 in week 07/06) (figure 1).

Figure 1 RCGP incidence rates for influenza-like illness: England 2006/07 and recent years

The highest incidence rates were among those aged 15-44 years (53.1/100,000 in week 07/07) (figure 2). Rates were highest in central England (53.8/100,000 in week 07/07), followed by southern England (43.4/100,000 in week 07/07). In contrast, northern England peaked last and with the lowest rate (32.9/100,000 in week 09/07).

RCGP rates for acute bronchitis peaked in week 01/07 (184.5 per 100,000). This rate peaked higher and later than last season (169/100,000 in week 52/05). The rates were highest among children aged 0 to 4 years in week 51/06 (524.0/100,000), followed by people age 65 years and over (446.9/100,000) in week 01/07. Compared to the 2005/06 season, the rate for the 0 to 4 years age group was slightly lower while that for the 65 years and over age groups was higher (598.5/100,000 and 374.7/100,000 in 2005/06 respectively).

Figure 2 RCGP incidence rates for influenza-like illness by age group, 2006/07

Q FLU - HPA and Nottingham University Division of Primary Care
The consultation rates for ILI peaked at 28.7/100,000 in week 07/07, the same week as reported by the RCGP. Rates were highest among those aged 15-44 years at 36.3/100,000 (week 07/07). The trends were similar to those indicated by the RCGP data but the overall reported rates were lower. No thresholds have yet been set for Q FLU.

Wales - National Public Health Service (NPHS Wales)
Consultation rates in the sentinel GP scheme in Wales remained within baseline levels (< 25/100,000) for the duration of the 2006/2007 season. The rate peaked in week 07/07 at 17.7/100,000 (figure 3).

Scotland - Health Protection Scotland
GP consultation rates in the flu spotter scheme for Scotland (figure 3) remained within baseline levels (< 50/100,000) until week 52/06. Activity then increased to normal seasonal levels for five weeks before returning to baseline levels in week 06/07. The rate peaked in week 02/07 at 158/100,000, which was higher than the peak of the previous season (33/100,000 in week 01/06) when activity remained within baseline levels.

Northern Ireland - Communicable Disease Surveillance Centre (CDSC Northern Ireland)
This is the seventh year of the enhanced surveillance scheme in Northern Ireland and baseline levels are yet to be established. The combined consultation rate for influenza and ILI peaked in week 05/07 at 204/100,000. This peak rate occurred earlier and was higher than the previous season (101/100,000 in week 08/06) (figure 3).

Figure 3 GP consultation rates for influenza-like illness for England, Wales, Scotland and Northern Ireland, 2006/07

NHS Direct for England and Wales
The proportion of NHS Direct 'cold/flu' calls for all ages peaked during weeks 06/07 and 07/07 at 1.1% (figure 4), just below the baseline threshold of 1.2%. The highest proportion of 'cold/flu' calls was among those aged 15 to 44 years at 1.6% in week 06/07; this value was above the baseline threshold of 1.4% and was consistent with the age distribution for ILI incidence rates in the RCGP scheme (figure 5). The proportion of 'fever' calls peaked during week 07/07 at 4.8% and was highest among those aged 0-4 years, at 15.8%, also during week 07/07; this value was above the baseline threshold of 15%. The trends in age groups most affected have changed since the last season: during the 2005/06 season, cold/flu calls were highest among those aged between 5-14 years at 1.8% in week 06/06 and most 'fever' calls were recorded for children aged between 5-14 years 12.8% in week 05/06.

Figure 4 NHS Direct proportion of 'cold/flu' and fever calls: 2005/06 and 2006/07

 

Figure 5 GP episode incidence rates and NHS Direct calls among the peak age groups during weeks 40-20

 

Mortality

The Office for National Statistics temporarily suspended reporting of rapid weekly death estimates before the end of the influenza season because of problems following the introduction of a new electronic registration system at local registrar offices. As a result, mortality data for the 2006/07 season was based on data from week 40/06 to 13/07. The weekly total number of estimated deaths due to all causes peaked at 11,322 in week 08/07. This number was lower and later than for the previous year (11,851 in week 02/06). The weekly number of actual respiratory deaths peaked at 1746 (week 01/07). Twenty deaths were specifically attributed to influenza during the season. The annual estimate of excess mortality due to influenza is calculated using a time series model [4]. Due to the missing data, no estimated excess could be calculated for the 2006/07 season but estimates for previous years have been recalculated (table 1).

Table 1 Estimated excess mortality due to influenza in England and Wales*

Influenza Season	Number of excess deaths
88/89	358
89/90	26945
90/91	8125
91/92	5967
92/93	1687
93/94	14544
94/95	2480
95/96	16241
96/97	21770
97/98	0
98/99	17982
99/00	22040
00/01	1067
01/02	7078
02/03	6559
03/04	5207
04/05	1795
05/06	0
06/07	No Estimate

* Each year the model is revised to incorporate data from the current season. The fitted model provides a figure for the most recent season and readjusts the previous years' figures.

Virological

Respiratory Virus Unit (RVU), HPA
A subset of the general practitioners (GPs), involved in the RCGP sentinel clinical surveillance scheme, also participates in virological surveillance (RCGP/HPA Community Surveillance Scheme), by sending respiratory specimens directly to the RVU for testing. A complementary GP based virological surveillance scheme (HPA Virological Surveillance Scheme) utilizes HPA Regional Microbiology Network (RMN) laboratories for the majority of testing but also provides some additional specimens directly to the RVU. Virus strains are either obtained by isolation from respiratory specimens provided by these sentinel GP schemes, or are submitted by hospital diagnostic laboratories. Virological data are obtained by analysing influenza strains using antigenic and molecular methods.

Table 2 shows the surveillance data obtained via HPA surveillance schemes, compared to data provided by Northern Ireland and Scotland.

Table 2 Sentinel GP surveillance data

Location	No. of Samples	No. of influenza positives by any method	Peak week of positives for influenza	Predominant influenza virus type
England (RCGP/HPA Community Surveill. Scheme)	1538	432	07/07	A(H3)
England (HPA Virological Surveill. Scheme)	489	170	06/07	A(H3)
Northern Ireland	111	47	05/07	A(H3)
Scotland	616	110	52/07	A(H3)

According to these schemes, influenza A (H3) was the predominant virus type isolated from community samples throughout the UK. The peak week of positives for influenza in England, Northern Ireland and Scotland correlated well with the peak week of clinical activity in those countries (within two weeks) but the timing of these peaks varied between countries.

In total, the RVU identified 467 positive community samples for influenza between week 40/06 and week 20/07. This figure is based on those respiratory specimens submitted directly to the RVU from both the RCGP/HPA Community Surveillance Scheme and the HPA Virological Surveillance Scheme. The majority (461(98.7%)) were identified as influenza A viruses. During this same period, the laboratory also identified 427 hospital samples positive for influenza with similar proportions of influenza A viruses. Community detections were predominantly from those aged 15-44 years, while hospital detections were mainly from children aged less than five years.

Five hundred and seventy one influenza viruses were further characterised (figure 6). Influenza A viruses were the dominant strain this season with 90.5% (517) isolated as influenza A(H3) A/Wisconsin/67/05 (H3N2)-like, 4.9% (28) as influenza A/NewCaledonia/20/99 (H1N1)-like and 3.5% (20) as influenza A/California/7/2004 (H3N2)-like. 0.4% (2) were influenza B viruses, antigenically similar to B/Malaysia/2506/2004-like and 0.7% [5] B/Jiangsu/10/2003-like viruses.

Figure 6 Influenza strains characterised by RVU/VRD from community and hospital sources