The Spanish influenza surveillance system comprises networks of sentinel physicians and network-affiliated laboratories, including the National Influenza Centre-Madrid (NIC-Madrid, WHO National Influenza Centre, National Centre of Microbiology). In addition to the specimens taken by sentinel physicians, regional laboratories also collect specimens from non-sentinel sources (i.e. hospitals, collaborating laboratories). Influenza detections from non-sentinel sources are notified to the system. In the season 2008-9, SISS included over 524 GPs and 173 PDs from 16 Spanish Autonomous Communities (ACs). The system covered 926,092 inhabitants, representing 2.07% of the total population of the 16 ACs. GPs and PDs report on weekly basis cases of influenza like illness (ILI) detected in their reference populations. In the 2008-9 season, 12 Spanish sentinel networks used the ILI definition of International Classification of Health Problems in Primary Care19: 1) context of influenza epidemic and four of the criteria listed in the point 2; or 2) six of the following criteria: onset within 12hours, cough, fever, chills, prostration and weakness, myalgia or general pain, rhinitis, pharyngitis, contact with a case. Four Spanish sentinel networks used an ILI definition based on the European case definition9: sudden onset of symptoms, and at least one out of four systemic symptoms: fever or feverishness, malaise, headache, myalgia; and at least one out of three respiratory symptoms: cough, sore throat, shortness of breath; and in the absence of other suspected clinical diagnosis.

For virological influenza surveillance, sentinel physicians took nasal or nasopharyngeal swabs, and sent them for influenza virus detection to the regional laboratories and network affiliated laboratories in each Autonomous Communities. Starting the influenza enhanced surveillance, all patients meeting the ILI case definition were swabbed. The swabbing strategy changed along the surveillance period, from all cases to a systematic swabbing, according to network resources and laboratory capacity. If the assay to confirm the influenza pandemic (H1N1) 2009 virus was not available at the regional laboratories, specimens were sent to the the National Reference Laboratory. Two independent assays were used for detection of pandemic influenza (H1N1) 2009 virus: 1) a multiplex reverse transcription (RT)-nested PCR assay for generic detection of Influenza A, B and C designed in the Nucleoprotein gene, and 2) a multiplex RT-nested PCR assay for sub-typing of the Haemagglutinin gene of influenza A viruses. Amplified products from both RT-PCR assays were sequenced and identification of (H1N1) 2009 virus was performed after sequence analysis.20 For molecular characterizations of resistance to antivirals, analysis of Matrix and Neuraminidase genes were also performed in all available isolates.

The data collected in the SISS included: sex, age, sentinel/non sentinel source, date of symptoms onset, date of swabbing, vaccination status, influenza laboratory confirmation, influenza type/subtype and influenza strain. Data was entered weekly, by each regional sentinel network in a web-based application and analysed at central level by the National Centre of Epidemiology, to provide timely information on the evolution of influenza activity at ACs and national level (http://vgripe.isciii.es/gripe).

Based on SISS data, we conducted a case-control study using the laboratory results of swabs taken from week 20 to week 39 (May-September 2009). Cases were medically attended ILI, laboratory confirmed for pandemic influenza (H1N1) 2009. Controls were medically attended ILI that tested negative for any influenza type in a SISS-affiliated laboratory (test-negative controls).

The WHO recommended influenza triple vaccine for Northern hemisphere in the 2008-9 season21 was the exposure for the effectiveness study. Annual vaccination is recommended in Spain for the high risk groups for influenza complications: patients with chronic conditions at any age, health care workers, elderly over 65 years old (over 60 years in some ACs).22 Vaccination is conducted at the GP/PD level during the annual campaign that lasted from October to December 2008, depending on Autonomous Community organization. Data on vaccination is collected by the GP/PD according to the patient clinical history. We excluded from the analysis ILI cases that had unknown vaccination status, no age nor laboratory results reported in the system.

For statistical analysis, we compared the characteristics of cases and controls by sex and age using Pearson's chi-square test. We calculated the crude and adjusted odds ratios (OR), their corresponding 95% confidence intervals (95%CI) using logistic regression and computed the vaccine effectiveness as (1-OR)*100. Variables included in the regression model were: age group, sex, month of swabbing and Autonomous Community. We adjusted for month of swabbing to control for differences in influenza incidence along the period of the study. After checking for clustering by Autonomous Community, we used a two level logistic regression model, introducing AC as random effect.23 We used STATA/IC 10 to carry out the statistical analysis.

The Influenza 2008-9 season (from week 40/2008 to week 19/2009) was moderate, associated to a mixed circulation of influenza A(H3N2) during the epidemic wave, and a further predominant circulation of B virus.24 The number of sentinel physicians participating in SISS increased from 525 GPs and 173 PDs in week 20/2009 to 647 GPs and 220 PDs in week 39/2009. The population under surveillance increased with 22%, up to 2.53% in week 39/2009.

From week 20/2009 to week 25/2009 (first weeks of the extended surveillance period), the influenza activity was low (ILI rate below 12 cases/100,000). The ILI rate started to increase steadily from week 26 (28 June – 4 July 2009), although influenza incidence rates maintained below baseline levels, overpassing it in week 38 (19-25 September 2009). Figure 1 shows the evolution of weekly ILI rates during 2008-9 season, during the usual surveillance period (week 40/2008 to week 20/2009) and during the extended surveillance period (May-September 2009).

Figure 1.

Weekly influenza incidence rate and proportion of influenza positive specimens in the sentinel system, 2008-2009 season, Spain.

(0.25MB).

From May to September 2009, the proportion of influenza sentinel positive specimens increased from 0 to 31%. From week 28, this proportion was higher than 30% (fig. 1).

The highest ILI rate was in the age group 5-14 years old (233 cases/100,000 population), followed by the 0-4 age group (153 cases/100,000 population) in week 39/2009.

The first influenza pandemic (H1N1) 2009 virus was detected in week 20, from a non-sentinel source. In week 23/2009 the sentinel system detected the first (H1N1) 2009 confirmed case. Among the total influenza virus detected by the sentinel and non-sentinel sources, the proportion of pandemic influenza (H1N1) 2009 was 14% in week 20/2009, 67% in week 21/2009 and more than 90% from week 34/2009.

In the study period, 3245 sentinel ILI cases were swabbed (3220 cases with age reported) representing 70% of total ILI medically attended by sentinel GPs/PDs, ranging from 85% in 0-4 aged group to 66% in over 64 years old (table 1). Out of 919 sentinel confirmed influenza, 92% were pandemic influenza (H1N1) 2009 (table 1).

Molecular characterization of the pandemic influenza viruses detected by RT-PCR assays and subsenquentely sequenced, showed that the circulating viruses were similar to A/California/07/2009(H1N1) strain in both Nucleoprotein and Haemagglutinine genes. Moreover, all pandemic viruses tested were resistant to M2 inhibitors (Adamantanes) and none of them showed mutation in the NA gene which conferred resistance to neuraminidase inhibitors (oseltamivir).

After the exclusion of influenza types B and C (n=24), no subtyped influenza A cases (n=53), ILI patients with unknown vaccination status (n=148) or no age reported (n=25), we included in the seasonal vaccine effectiveness study 2707 sentinel reported ILI patients: 844 influenza pandemic (H1N1) 2009 cases and 1863 test-negative ILI controls. The weekly number of pandemic (H1N1) 2009 cases and test-negative controls increased by week of swabbing (fig. 2).

Figure 2.

A (H1N1) 2009 laboratory confirmed cases and test-negative controls by week of swabbing, May-September 2009, Spain. ILI: influenza like illness.

(0.14MB).

The cases were significantly younger than controls (p<0.001): mean age of cases: 20.8 years (standard deviation: 14.4), compared with controls mean age: 25.1 years (standard deviation: 20.1). No difference regarding gender was registered between cases and controls (p=0.202).

A total of 196 (7.24%) ILI patients included in the study were vaccinated with the 2008-9 season trivalent vaccine: 42 (4.98%) cases, and 154 (8.27%) controls. The crude OR was 0.58 (95%CI: 0.39; 0.83). Stratifying by age group (table 1), we found no difference between strata (homogeneity test p=.61) and a Mantel Haenzel combined OR of 0.79 (IC 95%: 0.54; 1.16). The OR adjusted for sex, age group, mounth of swabbing and Autonomous Community, was 0.88 (95%CI: 0.59; 1.30), resulting in an IVE of 12% (95% CI: -30; 41).