Baguio City is an urbanized city located in the highland of northern Luzon, Philippines with over 298,500 inhabitants. It has a population density of 5,190 persons/km². Children aged <5 years and teenagers comprised 10.1% and 22.2%, respectively, of the total population, while the elderly (age, >60 years) comprised only 5.0% [1]. Thus, the population structure is largely clustered in the working age group.

In early April 2009, the first human case of influenza A(H1N1)pdm09 [A(H1)pdm09] was identified in Mexico and the United States [2]. This virus then spread rapidly to other countries. The Philippines identified its earliest case on May 22, 2009, and Baguio City reported its first A(H1)pdm09 cases 3 days later. The timing of the initial introduction in mid-May was similar to that of other neighboring Asian countries [3-8]. A total of 159 cases and 8 fatalities was officially reported in the Cordillera Administrative Region, where Baguio City is located, by 31 August, 2009 [9].

Prior to the emergence and spread of A(H1)pdm09, enhanced surveillance of influenza-like illness (ILI) and severe acute respiratory infection (SARI) was conducted as part of the Influenza Disease Burden study (BOD study) in the Baguio City. The study aimed to estimate influenza disease burden through surveillance in 16 city district health centers and 5 hospitals in the city. In brief, this system detected overall 508 outpatients and 98 inpatients who were positive for A(H1)pdm09 from May 2009 to Dec 2010.Numerous reports have described the epidemiological and clinical features of A(H1)pdm09 cases; however, they mostly covered the pandemic periods and few covered the post-pandemic period [10]. Moreover, there was no study about A(H1)pdm09 epidemiology in the Philippines, where year-round influenza activity was observed. This present study therefore aimed to describe the epidemiological and clinical features of A(H1)pdm09 both during the pandemic and post-pandemic periods.

The number of A(H1)pdm09 outpatients was 2.3-fold higher in 2009 than in 2010, and the number of inpatients was 1.8-fold higher in 2009 (Table 1). For the 2-year period, the outpatient incidence was highest among children aged <5 years, followed by the individuals aged 5–14 years, and lowest among those aged ≥50 years. The distribution by sex and age group was similar between outpatients and inpatients for both periods, except among those aged ≥50 years, who accounted for 11.5% among inpatients and 1.1% in outpatients. Among cases with the data about their occupations (22%), students were most frequently recorded. Elementary school students accounted for 66% and 71% of all students in 2009 and 2010, respectively. Housewives and unemployed individuals were significantly higher among both outpatients and inpatients in 2009 than in 2010 (p < 0.001) (Table 2).

A(H1)pdm09 outpatients generally had ILI signs and symptoms without any symptoms and comorbidities; a few had diarrhea, vomiting, abdominal pain, loss of appetite and 2 cases reported pregnancy (Table 3). Clinical manifestation and physical examination findings among confirmed inpatients between 2009 and 2010 were similar. Almost half of the outpatients (49%) were prescribed antibiotics, while only 1 was prescribed antivirals. There observed no difference of median duration of illness prior to consultation between 2009 and 2010 (Table 4).

Cases were admitted within 3–4 days of onset of illness, with the majority presenting with signs of SARI, including difficulty in breathing, chest indrawing, and wheezing on physical examination. Asthma was the predominant comorbidity in almost a third of the cases for both periods, while a few had diabetes mellitus (DM) (5.8% vs. 0%), cardiovascular disease (CVD) (7.2% vs. 10.3%), and chronic obstructive pulmonary disease (COPD) (2.9% vs. 13.8%) as underlying conditions (Table 3). Abnormal radiological findings (49.3% vs. 41.4%) of pneumonia were also similar between the pandemic and post-pandemic years. The majority of cases during the two periods were treated with antibiotics (88.4% vs. 93.1%), while only 36.2% were treated with antivirals in 2009. There were 33.3% and 24.1% patients who received supplemental oxygen in 2009 and 2010, respectively, and 5 were intubated while in the wards in 2009. The majority, except for 2 fatalities in 2009, were discharged recovered after 3 days of confinement. The first fatality was of a 56-year-old male with several underlying conditions, such as COPD, hypertensive CVD, myocardial infarction, and DM, who was admitted 6 days after the onset of illness. He was treated with antivirals on the 2nd day of confinement and intubated, but he died after 7 days. The other fatality was of a 40-year-old male with DM and chronic renal disease who was admitted 6 days after the onset of illness. The patient was treated with antivirals after 6 days of confinement and intubated, but he died after 30 days of hospitalization.

Figure 1 shows the weekly number of A(H1)pdm09 cases among outpatients and inpatients. The outbreaks were observed during morbidity weeks 25–35 in 2009 and weeks 24–37 in 2010. There were fewer confirmed cases sporadically distributed between these two epidemics. On the basis of the number of cases during the growth phase, we estimated the basic reproduction number of 1.14 (95% confidence interval (CI) 1.13–1.15) in 2009 (from June 18 to July 6) and 1.09 (95% CI 1.08–1.10) in 2010 (from June 18 to July 11), with a mean generation time (MGT) of 2.6 days (Table 5). These numbers were altered to 1.09 and 1.06 with MGT of 4 days in 2009 and 2010, respectively (Figure S1).

Figure 2A and B shows the weekly ratio of students to non-students together with the proportion of age groups during outbreaks of 2009–2010. We observed over 1-fold ratio in the growth phase of peaks in both outbreaks. The ratio increased a week prior to the peak of confirmed cases. This peak ratio was obviously higher in 2009 than in 2010 (3.7 and 1.9, respectively). After the growth phase, the ratio scaled down to near 1.0. The highest proportion of cases was observed among children aged <5 years.

In this study, we investigated influenza A(H1)pdm09 epidemiology from 2009 to 2010 in Baguio City. The total number of cases reported in 2009 was much higher than in 2010, and the highest number of cases was observed in the 5 to 14-year-old group, most of which were school-age children. This is consistent with findings from a study of 8 countries that estimated the highest cumulative incidences of A(H1)pdm09 among school-age children [17]. Ten of the 13 inpatients belonged to the >50-year-old group were aged <65 years. This younger age shift of inpatients could explain the lower mortality observed in this study. The ratio of deaths to hospitalizations due to A(H1)pdm09 was the highest among the >65-year-old group [18] partly because those who are aged >80 years have pre-existing immunity against A(H1)pdm09 [19,20].

There were two epidemics in June–August of each year, while cases were sparsely distributed in between (Figure 1).We did not observe a wave during the cooler months of September–November in 2009, which was widely observed in neighboring countries and cities[21-24] but not in Malaysia [25]. The increasing number of cases occurred mainly in June when schools in the country were opened. This coincidence between the resumption of classes and expanding transmission in the community has been described in other studies [14,26]. With a higher ratio of students to non-students, especially in the growth phase of epidemics, A(H1)pdm09 transmission was multiplied among students who play a role in expanding transmission of influenza in the community [27]. Moreover, housewives and unemployed individuals were significantly higher among both outpatients and inpatients in 2009 than in 2010. This may be because housewives are more likely to take care of patients at home. Household transmission could be one of the important opportunities for A(H1)pdm09 to spread, as reported in other studies [28,29], but its qualitative risk is difficult to measure.

We estimated R0 as 1.14 in the growth phase of the 2009 wave. A review of transmission parameters for the (H1N1)2009 pandemic indicated the median reproduction number as 1.6 (1.1–3.1) [16], and a higher value was mostly obtained from outbreak settings in schools [30,31]. This variability could be explained as a reflection of differing estimation methods and assumptions regarding generation time and the under-reporting rate. Heterogeneity resulting from control measures undertaken as well as a wave period over the entire pandemic period could affect the estimation of reproduction number. A study that investigated the reproduction number in the southern hemisphere demonstrated a positive correlation between the proportion of children aged <20 years and reproduction number [32]. The fact that 45% of A(H1)pdm09 cases were children aged <20 years may have led to a potentially higher reproduction number in the city. The reasons why we had a lower basic reproduction number remain unclear. Baguio City is located in the highland and has a mean temperature of 19.6°C and mean relative humidity of 88.1%. This could have decreased the effective transmissibility of influenza virus [33]. Moreover, a social contact study in Vietnam revealed the concentration of close contact in home settings, although the physical contact levels were low [34]. This pattern of close contact in home settings would be similar in the Philippines. In this regard, additional study focusing on social contacts and the dynamics of influenza in the community is necessary.

The clinical condition of outpatients was generally mild and self-limiting. Some 34% of inpatients had developed wheezing or chest indrawing, and 12 of the 21 cases who developed wheezing did not have asthma as their comorbidity. A study in Japan indicated that about a third of the cases who had asthma attacks during A(H1)pdm09 infection were not diagnosed to have asthma [35]. A(H1)pdm09 infection may lead to an aberrant asthmatic response although this mechanism remains unclear. No fatal cases were reported although asthma was the most common comorbidity among inpatients. This observation was similar to other studies in that asthma is observed as one of the risk factors for hospitalization but is not directly correlated with fatal severity [18,35-37]. In terms of intensive clinical severity, 33% of inpatients were administered oxygen in total and 7.2% were intubated with 2 deaths in 2009. None was intubated and expired in 2010.

In the Philippines, antivirals are not commonly used for ILI treatment. Only the laboratory-confirmed cases were administered antivirals, subject to the time and condition of patients when the laboratory results arrived. During the (H1N1)2009 pandemic, interim guidelines on use of antivirals were issued by the national government, particularly the Department of Health, which encouraged physicians to prescribe antivirals. Therefore, 25 inpatients were administered Oseltamivir in 2009. We could not see a linear correlation between intervals from onset of medication and duration of hospitalization (p = 0.64). Another study demonstrated that early treatment with antivirals leads to better clinical outcomes [38], but this study did not show similar results. This is partly because 75% of cases were administered antivirals after admission, which suggests that cases already developed SARI conditions, and because of the small number of cases.

There were some limitations in this study. First, we detected influenza A(H1)pdm09 among outpatients at the city health centers and in one of outpatients’ department of a tertiary hospital. Thus, we may have potentially missed cases that did not seek consultation or consulted at other health facilities such as private clinics. Second, our reproduction estimation was truly based on the exponential growth rate and known generation times studied in other countries. In spite of this limitation, this method has been widely used [12-14]. Third, because the updated census (2010 population census) data were not available as age-stratified figures, we may be overestimating rates because of a lower population in the 2007 census.

In conclusion, the demographic and clinical characterization for A(H1N1)pdm09 cases shown in this study were similar to that of other countries. We observed epidemics due to A(H1)pdm09 in July–August of both years but not during the cooler months of September–December of 2009. School-age children played a relatively significant role in expanding transmission in the city, while lower a reproduction number was estimated in the growth phase. During the 2-year study period, only 2 fatalities were recorded. There was relatively quick medical attention for both outpatients and inpatients during the pandemic period although the frequency of antiviral prescription was significantly low. This reinforces the importance of the growing capability to provide appropriate medical consultation in the community. Finally, it is important to establish surveillance systems to monitor influenza activity, which will provide helpful information to both stakeholders and the public for implementing a pandemic influenza response.