Vaccine Preventable Diseases and Vaccination Coverage in Australia, 2003 to 2005

Influenza

Disclaimer: This is the fourth report on vaccine preventable disease and vaccination coverage in Australia, and is produced by the National Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases and the Australian Institute of Health and Welfare on behalf of the Australian Government Department of Health and Ageing published as a supplement to the Communicable Diseases Intelligence journal Volume 31, June 2007.

Page last updated: 20 July 2007

Influenza causes annual epidemics of respiratory disease often indistinguishable clinically from disease caused by other respiratory viruses. Symptoms include abrupt onset of fever, cough, malaise, myalgia, sore throat, and headache. Influenza epidemics usually occur during the winter months in temperate climates, causing an increase in hospitalisations for pneumonia and exacerbation of chronic diseases and also contributing to increased mortality, particularly among the elderly and those with high risk underlying conditions. In tropical climates, influenza infection may have two annual peaks, as illustrated in the Northern Territory.96 Pandemics of influenza, occurring every 30 years or so, are caused by major antigenic shift. Antigenic drift, however, occurs more regularly, producing smaller epidemics.

Case definitions

Notifications

Laboratory-confirmed influenza is a nationally notifiable disease in all jurisdictions except South Australia. Although influenza is not a notifiable disease in South Australia, laboratory reports are collected and sent to NNDSS. Implementation of influenza notification occurred in all other jurisdictions, except Tasmania, during 2001.

Laboratory-confirmed infections are those in which influenza virus is isolated by cell culture, detected by nucleic acid amplification, or by influenza antigen testing, from appropriate respiratory tract specimens; a recent infection can also be demonstrated by serological methods.11

Hospitalisations

The ICD-10-AM codes used to identify hospitalisations were: J10 (influenza due to identified influenza virus) and J11 (influenza, virus not identified). In this report, we did not make the distinction between admissions where a virus was identified and those where it was not.

Deaths

The ICD-10 codes used to identify deaths were: J10 (influenza due to identified influenza virus) and J11 (influenza, virus not identified).

Top of page

Secular trends

In the period January 2003 to December 2005, there were 10,185 notifications, giving an average annual rate of 16.9 per 100,000. Notifications varied considerably by year, with a high of 4,569 in 2005 and a low of 2,133 in 2004. There was a clear seasonal distribution of notifications in all three years with peaks in August-September in 2003, September in 2004 and June-September in 2005.

Between July 2002 and June 2005, there were 9,116 hospitalisations coded as influenza (an average annual rate of 15.3 per 100,000), with the greatest number of hospitalisations (3,956) recorded in 2003/04. There was a clear seasonal pattern, with dramatic increases over the winter months (Figure 10). The median number of admissions per month was 107 (range 47–1,515) with the highest number (973, 1,515 and 446) occurring in July 2002, August 2003 and September 2004, respectively.

Figure 10. Influenza hospitalisations and notifications,* Australia, July 1993 to December 2005, by month of diagnosis or admission

Figure 10. Influenza  hospitalisations and notifications, Australia, July 1993 to December  2005, by month of diagnosis or admission

* Notifications where the month of diagnosis was between January 2001 and December 2005; hospitalisations where the month of admission was between 1 July 1993 and 30 June 2005. Note that the Northern Territory, Queensland, Tasmania and Victoria did not notify influenza for the complete year in 2001.

Severe morbidity and mortality

A total of 56,955 hospital bed days with an ICD-10-AM code for influenza were recorded over the reporting period. The median length of stay was at least twice as long in the oldest age group (five days among people aged 60 years or over) than in the younger age groups (Table 7). Influenza was the principal diagnosis for 68% of the hospitalisations. Bed days peaked in 2003/2004 (n=24,156).

From 1 January 2003 to 31 December 2004, influenza was recorded on the death certificate as the underlying cause of death in 101 cases. Of these, 85 (84%) were aged 60 years or more, four (4%) were aged 25–59 years and five (5%) were aged 0–4 years. Deaths officially designated as due to influenza are known to be a gross underestimate of the true number of deaths attributable to influenza in the population.97 Although there is no requirement for clinical information to be entered onto NNDSS for influenza cases, 13 deaths were recorded on NNDSS in influenza notifications 2003–2005; two deaths were recorded in children (age range for deaths in notified cases 1–89 years, median 69 years).

Table 7. Influenza notifications, hospitalisations and deaths, Australia, 2002 to 2005,* by age group

Age group
(years)
Notifications
3 years
(2003–2005)
Hospitalisations
3 years
(July 2002–June 2005)
LOS per admission
(days)
Deaths
2 years
(2003–2004)
  n Rate n (§) Rate (§) Median(§) n Rate
0–4
3,173
83.7
3,116
(2,458)
82.1
(64.7)
2.0 (2.0)
5
0.20
5–14
1,125
13.8
731
(544)
9.0
(6.7)
2.0 (2.0)
5
0.09
15–24
1,157
13.9
727
(488)
8.9
(6.0)
1.0 (1.0)
2
0.04
25–59
2,994
10.2
2,499
(1,542)
8.6
(5.3)
2.0 (2.0)
4
0.02
60+
1,724
16.4
2,043
(1,264)
19.9
(12.3)
5.0 (4.0)
85
1.23
All ages||
10,185
16.9
9,116
(6,296)
15.3
(10.6)
2.0 (2.0)
101
0.25

* Notifications where the month of diagnosis was between January 2003 and December 2005; hospitalisations where the month of separation was between 1 July 2002 and 30 June 2005; deaths where the date of death was recorded in 2003 or 2004.

† LOS = length of stay in hospital.

‡ Average annual age-specific rate per 100,000 population.

§ Principal diagnosis (hospitalisations).

|| Includes cases with unknown ages.

Top of page

Age and sex distribution

Influenza notification rates vary markedly by age, with peak rates in those under five years of age (Figure 11). In this age group, the highest rates of notifications are in those under one year of age and the rate declines progressively thereafter. The overall male to female ratio was 1.08:1.

Among the age groups specified in Table 7, hospitalisation rates were highest in children aged under five years (82.1 per 100,000). Although overall hospitalisation rates were lower than those in 0–4 year olds, among people aged 60 years and over, the rates increased with increasing age, ranging from 12.2 per 100,000 for those aged 60–64 years to 37.3 per 100,000 for those aged 85 years or more (data not shown). Among children aged less than five years, the hospitalisation rates were highest among infants (annual average of 173 per 100,000 population aged less than one year, ranging from 97.2 per 100,000 in 2004/2005 to 261.8 per 100,000 in 2003/2004).

The overall male to female hospitalisation ratio was 0.92:1; however, this was not consistent across all age groups. In children under 10 years of age, male hospitalisations were more common (1.2:1).

Figure 11. Influenza notification rates 2003 to 2005 and hospitalisation rates 2002/2003 to 2004/2005, Australia,* by age group

Figure 11. Influenza  notification rates 2003 to 2005 and hospitalisation rates 2002/2003 to  2004/2005, Australia,  by age group

* Hospitalisations where the month of separation was between 1 July 2002 and 30 June 2005.

Geographical distribution

There was a wide variation by state or territory in the average crude hospitalisation rate recorded for the three year review period, ranging from 3.5 per 100,000 (n = 34) in the Australian Capital Territory to 30.0 per 100,000 in the Northern Territory (n = 179) (Appendix 3). Notification rates were similarly varied, with the highest rates reported in the Northern Territory and Queensland (41.8 and 27.5 per 100,000 population, respectively) and the lowest in Tasmania and the Australian Capital Territory (2.0 per 100,000 and 4.8 per 100,000, respectively) (Appendix 2).

In regard to annual hospitalisation rates, the distribution by state or territory was generally consistent with the national pattern, with the highest rates occurring in 2003/2004 in all states except Queensland, while the lowest rates occurred in 2004/2005 in all states except the Australian Capital Territory. Historically, the winter of 1997 remains the period between 1993/1994 and 2004/2005 during which most states and territories recorded the highest number of hospitalisations.

Influenza type and subtyping

Almost all of the influenza notifications reported to NNDSS (98%) recorded influenza type. Most of these were type A (n=8,284; 83%). A further 1,586 cases (16%) were recorded as type B and both A and B were documented in 91 notifications (1%). Subtype or strain information was recorded for 12.4% of notifications (n=1,267). Ninety-five per cent of subtyped influenza A was H3 and 5% was H1. Subtype information was only recorded on NNDSS from some notifications of influenza isolates received in Victoria, New South Wales, Queensland and the Northern Territory, with Queensland and Victoria reporting most of the notifications where subtyping was documented.

Top of page

Comment

Both laboratory notifications and hospitalisations show a similar seasonal pattern. There is a strikingly higher notification and hospitalisation rate in children under one year or age, compared with even the extremely elderly. This is likely to reflect patterns of health care use, and, in particular, diagnostic testing for respiratory viruses, in this age group.

On the other hand, the role of influenza in exacerbating chronic cardiac and respiratory diseases in the elderly may not be adequately reflected in these surveillance data. Deaths and hospitalisations coded as influenza are widely acknowledged to underestimate deaths and hospitalisations due to influenza.97–100 The proportion of deaths due to influenza occurring in people aged 60 years and over (84%) is lower than most other published studies.101,102 This may be due to other contributing causes being preferentially coded, as deaths reported here are only those where influenza was recorded as the principal cause, or to lower rates of virological testing in the elderly age group in Australia in comparison to elsewhere.

It should be noted that there is no specialised diagnostic influenza laboratory in Tasmania or the Northern Territory, with specimens positive on direct fluorescent antibody testing referred interstate. Hospitalisation data referred to in this report are based on discharge coding and it is possible that some of those with less specific influenza codes (e.g. J11) may be due to other respiratory pathogens such as respiratory syncytial virus (RSV),103 coronavirus104 or picornavirus.105,106 The apparent differences in hospitalisation rates between states and territories should be treated with caution as they may reflect differences in coding practices or rates of virological testing of inpatients between jurisdictions.

Both influenza A and B are well known to cause major epidemics of respiratory disease resulting in severe morbidity and mortality. In 2003, the predominant influenza isolate was influenza A (94%), with a majority of the A (99%) being H3 subtypes. Ninety-eight per cent of these were A/Fujian/411/2002 (H3N2) which had shown a significant antigenic drift. The 2003 Australian influenza vaccine contained A/Panama/2007/99, which induced two– to fourfold lower antibody response against the drifted strain.107 An A/Fujian/411/2002 (H3N2)-like virus was incorporated in the Australian influenza vaccine for 2004.

In 2004, influenza A was again predominant (76.9%) but with increased influenza B activity compared to the previous season. Of the record low number of viable influenza isolates received (n=454), 98% of type A isolates were H3N2, which showed further antigenic drift, with one third of the isolates distinguishable from the A/Fujian/411/2002-like reference strain.108

In 2005, influenza A remained the predominant type notified (73%), with further increased influenza B activity. Antigenic analysis of 1,174 samples demonstrated that 689 (58.7%) were H3N2, 210 were A H1N1 (17%), and 275 (23.4%) were B viruses.109

Annual influenza vaccination is currently recommended as the primary method of influenza prevention in people aged 65 years and over, all Indigenous people aged 50 years and over, and all individuals aged 6 months and over with chronic medical conditions likely to be exacerbated by influenza or its complications, e.g. chronic pulmonary or cardiovascular disease.76 Vaccination uptake in Australians aged 65 years and over was estimated at 76.9% and 79.1% for 2003110 and 2004,111 respectively. Extension of influenza vaccination to all adult Indigenous people should be considered as both influenza hospitalisation rates and mortality related to influenza and pneumonia are twice that documented in persons of non-Indigenous background across all age groups.77 Influenza vaccination is also encouraged in health care workers and carers of people with high-risk conditions76 as carer vaccination has been shown to reduce morbidity and mortality in those with high-risk conditions.112

As for the paediatric population, in 2003 the USA Advisory Committee on Immunization Practices (ACIP), recommended routine influenza vaccination113,114 of healthy American children aged 6–23 months based on the high burden of illness.99,115,116 More recently, in 2006, the ACIP extended its recommendation to include children up to the age of five years.117 Whilst available Australian data suggest a similar significant burden of illness in young children,118,119 examination of cost-effectiveness, efficacy and feasibility of universal immunisation of healthy children is required before implementing such a population level strategy.120

There is increasing concern regarding the likelihood of an influenza pandemic with the ongoing presence of H5N1 avian influenza in Asia and more recent spread to Europe and Africa.121 Studies of H5N1 vaccines are beginning to show reassuring immunogenicity even in children and the elderly. Other control measures, such as enhanced infection control (handwashing, quarantine, cohorting), will be especially important early on should a pandemic arise as vaccine production will take some months. Furthermore, achieving high coverage now against seasonal influenza could usefully increase local production capacity of influenza vaccine through increased demand which could speed a switch of production to a pandemic vaccine.122–125

Document download

This publication is available as a downloadable document.

Vaccine Preventable Diseases and Vaccination Coverage in Australia, 2003 to 2005(PDF 1622 KB)