INTRODUCTION
Pandemic A(H1N1) 2009 influenza, caused by a quadruple reassortant virus, emerged in late March 2009, in Mexico, and quickly spread to other countries worldwide [Reference Neumann, Noda and Kawaoka1, 2]. The first pandemic wave in the Northern Hemisphere coincided with the end of the annual influenza season, in contrast with the Southern Hemisphere.
Respiratory infections, have a higher incidence and impact during the cold winter months [Reference Falagas3]. The interplay between environmental factors (exposure to cold and dry weather), host factors (alterations in local and systemic immune responses), and social factors (indoor crowding) could account for this phenomenon [Reference Falagas3, Reference Mourtzoukou and Falagas4].
In this regard, the experience gained during the 2009 winter season in the Southern Hemisphere could best reflect the true impact of the 2009 influenza pandemic. This experience could be useful for evaluating the course of the pandemic in the Northern Hemisphere during the 2009–2010 winter season, and particularly for predicting the nature of future influenza pandemics. We systematically reviewed the published evidence originating from the Southern Hemisphere regarding the epidemiological characteristics of pandemic A(H1N1) 2009 influenza, with a focus on specific risk groups.
METHODS
Data sources
We searched the collection ‘Latest H1N1 citations in PubMed’, as retrieved on 4 November 2009. The collection of these articles was retrieved using the following combined search terms: ‘(swine or H1N1) and (flu or influenza or virus or outbreak or pandemic) and ‘last 6 months’[edat]’. The bibliographies of relevant articles of interest were also hand-searched. We selected for inclusion in our review studies that provided original data, obtained in countries of the Southern Hemisphere that referred to the epidemiological characteristics of groups of patients of any age with clinically suspected or laboratory-confirmed pandemic A(H1N1) 2009 influenza. Articles published in languages other than English, Spanish, French, German, or Italian, as well as studies published only as abstracts in scientific conferences, were excluded.
Data extraction
Two author reviewers (P.K.K and E.K.B.) independently performed the literature search, study selection, and data extraction. Data extracted from each of the included studies consisted of the definition of study population, the location and period of study, the criteria used for diagnosis of influenza, as well as the number of total study population and the characteristics and number of different epidemiological subgroups, that were determined according to age, sex, or presence of comorbid conditions, or risk factors for influenza. We extracted specific relevant data for groups of patients with influenza who received different levels of healthcare [e.g. outpatient care, hospitalization, intensive care unit (ICU) admission] and for those with fatal outcome. We excluded information that did not derive from original sources of each study.
RESULTS
Characteristics of the included studies
Among the 1289 citations initially retrieved from PubMed, we finally identified 15 studies as eligible for inclusion in our review [Reference Raffo5–Reference Baker19]. The flow chart for the detailed process of selection of the studies to be included in the review is shown in Figure 1.
Flow diagram of the process of selection of articles for inclusion in the review.
Table 1 (see also expanded online version of Table 1) presents the data extracted from each of the included studies. Specifically, seven of the 15 studies referred to South American countries [Reference Raffo5–Reference Munayco11], six to Australia or New Zealand [Reference Webb14–Reference Baker19], and two to Africa [Reference Archer12, Reference Thouillot13]. The data reported in the included studies were obtained through various types of influenza surveillance systems in nine of the studies [Reference Oliveira8, Reference Gianella10–Reference Thouillot13, 16–Reference Baker19], at multiple hospitals (specifically ICUs) in two studies [Reference Webb14, Reference Davies15], and at single hospitals in the remaining four studies [Reference Raffo5–Reference Bantar7, Reference Schout9].
Data regarding the epidemiological characteristics of cases of pandemic A(H1N1) 2009 influenza extracted from studies originating in the Southern Hemisphere that were included in our review (only data for the total study population are presented in the printed table below; an expanded version of Table 1 with additional data for specific epidemiological/risk factors groups is available as Supplementary online material)
CDC, United States Centers for Disease Control and Prevention; CI, confidence interval; d, days; ECMO, extracorporeal membrane oxygenation; ILI, influenza-like illness; IQR; interquartile range; M, male; MV, mechanical ventilation; (P)ICU, (paediatric) intensive care unit; pts, patients; RSV, respiratory syncytial virus; RT–PCR, reverse transcriptase–polymerase chain reaction; SARI, severe acute respiratory infection.
* Denominators refer to the cases for which specific relevant data were available.
† Suspected case: sudden onset of fever (⩾38°C) and respiratory symptoms.
‡ Coexisting conditions: for patients aged ⩾16 years: any condition defined within the Chronic Health Evaluation component of the APACHE III score; for patients aged <16 years: prematurity, immune deficiency, cystic fibrosis, congenital heart disease, neuromuscular disorder, or chronic neurological impairment.
§ The population presented in this study may be a subgroup of the population of the Fielding et al. [Reference Fielding17] study.
All but one of the included studies [Reference Bantar7], provided data on the characteristics of patients with laboratory-confirmed pandemic A(H1N1) 2009 influenza; the number of this category of patients evaluated in each of the included studies varied widely (median 856, range 20–34 506). It should be mentioned that three of the 14 studies that reported data on laboratory-confirmed cases did not provide adequate data regarding the exact laboratory methodology used [Reference Thouillot13, 16, Reference Baker19]. Four of the 15 included studies additionally provided data on patients with clinically suspected influenza [severe acute respiratory infection or influenza like illness (ILI)] [Reference Raffo5, Reference Bantar7, Reference Oliveira8, 16]. Seven studies reported data for hospitalized cases of pandemic A(H1N1) 2009 influenza [Reference Raffo5, Reference Bantar7, Reference Schout9, Reference Thouillot13, 16, Reference Fielding17, Reference Baker19], while nine reported data specifically for patients admitted to ICUs [Reference Raffo5–Reference Bantar7, Reference Thouillot13–Reference Fielding17, Reference Baker19].
Laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza
The rate of positivity for pandemic A(H1N1) 2009 influenza in cases tested with RT–PCR was 31·5–53·0% in four studies from Brazil, Argentina, and Australia [Reference Raffo5, Reference Oliveira8, Reference Schout9, Reference Kelly and Grant18], whereas a lower rate of 12·7% was recorded in a reference laboratory in Bolivia [Reference Gianella10]. The percentage of confirmed cases of seasonal influenza in all cases tested for pandemic A(H1N1) 2009 influenza was 8·1% and 8·4% in two studies from Brazil [Reference Oliveira8, Reference Schout9], while the percentage of seasonal influenza A, in particular, was 3·3% in cases tested at a general hospital in Argentina [Reference Raffo5].
Hospitalized cases
The percentage of cases admitted to hospital was 44·5% and 45·3% in cases of pandemic A(H1N1) 2009 influenza identified in two hospitals in Brazil [Reference Schout9] and Argentina [Reference Raffo5], respectively. Among cases of pandemic A(H1N1) 2009 influenza identified through surveillance systems, 16·8–30·6% were admitted to hospital and 3·5–4·4% were admitted to ICUs, as reported by three studies in Australia and New Zealand [16, Reference Fielding17, Reference Baker19]. The rate of ICU admission in cases hospitalized for pandemic A(H1N1) 2009 influenza varied between 7·5% and 26·0% in five studies that reported specific relevant data [Reference Bantar7, Reference Thouillot13, 16, Reference Fielding17, Reference Baker19].
Fatality
The fatality rate among the laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza was between 0·5% and 0·9% in 5/8 studies that reported specific relevant data [Reference Munayco11, Reference Archer12, 16, Reference Fielding17, Reference Baker19]. A rate of 1·5% was reported for cases identified at a reference laboratory in Bolivia [Reference Gianella10], while a higher rate of 3·4% was reported for cases identified at a reference hospital in Brazil [Reference Schout9], and an even higher rate of 11·2% was recorded for cases identified through Brazil's case notification system [Reference Oliveira8]. Notably, a case-fatality rate of 4·5% was recorded for all cases of severe acute respiratory infection identified through the latter system.
Among the cases of pandemic A(H1N1) 2009 influenza that were admitted to hospital, the fatality rate varied between 2·4% and 7·6% in the five studies that reported specific relevant data [Reference Bantar7, Reference Schout9, Reference Thouillot13, 16, Reference Fielding17]. The fatality rate for the patients admitted to ICUs, in particular, varied between 14·3% and 22·2% in 5/7 studies that reported specific relevant data [Reference Bantar7, Reference Webb14–Reference Fielding17]. An additional study from Argentina reported a fatality rate of 50·0% and 22·2% for cases admitted to adult and paediatric ICUs, respectively [Reference Raffo5]. The remaining study reported a fatality rate of 50% for cases admitted to a paediatric ICU in Argentina [Reference Caprotta6].
Epidemiological characteristics of cases
Table 2 presents the percentage of specific epidemiological groups (defined according to age, sex, race, or presence of comorbid conditions) among the cases evaluated in each of the included studies (stratified according to type of diagnosis, level of healthcare received, and outcome).
Summary of data from different studies in the Southern Hemisphere regarding the representativeness of different epidemiological groups according to age and underlying condition among the cases of pandemic A(H1N1) 2009 influenza
COPD, Chronic obstructive pulmonary disease; ILI, influenza-like illness; SARI, severe acute respiratory infection.
Age
Among the laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza, the majority were young and middle-aged adults in 4/5 studies reporting specific relevant data. Specifically, the percentage of this age group was between 53·6% and 60% in these four studies [Reference Raffo5, Reference Schout9, 16, Reference Kelly and Grant18]. In the remaining study, which reported national surveillance data from Peru, young and middle-aged adults represented 32·7% of the cases, while young children and adolescents accounted for 58·9% of the cases [Reference Munayco11]. Seniors aged >60–65 years represented a minority of cases, varying between 0% and 6·4%, in the four studies providing specific relevant data [Reference Schout9, Reference Munayco11, 16, Reference Kelly and Grant18]. However, older adults constituted a substantial percentage of the cases admitted to hospital (9·5–16·6%) [Reference Bantar7, 16], ICU (13·3–28·5%) [Reference Bantar7, 16], or fatal cases (29·2%) [16].
Comorbidity
Cases without any underlying comorbidity constituted the majority of those with severe acute respiratory infection due to ILI (66·7%) and those with laboratory-confirmed pandemic A(H1N1) 2009 influenza (65·4%) that were identified through Brazil's case notification system [Reference Oliveira8]. However, only 15·5% of the laboratory-confirmed cases evaluated at a Brazilian hospital were considered to have no comorbidity [Reference Schout9]. Among the cases of pandemic A(H1N1) 2009 influenza that were admitted to hospital, three studies reported that 30·6–53·3% were considered previously healthy [Reference Raffo5, Reference Bantar7, Reference Thouillot13].
Specifically regarding cases admitted to ICUs, three studies from Argentina (of which two referred to paediatric patients) reported that 48·1–70·0% of the cases were previously healthy [Reference Raffo5–Reference Bantar7]. In a larger study from Australia and New Zealand the same figure was 31·7% [Reference Webb14], while in another study from Australia 10·2% of the ICU-admitted cases had no underlying risk factor, including obesity or pregnancy [16]. The percentage of cases without any comorbidity was similar for those who died and those who were admitted to ICUs, in two studies that provided specific relevant data [Reference Caprotta6, 16]. Studies from South America reported a higher percentage (50·1–60·0%) for cases without comorbidity among the fatal cases of pandemic A(H1N1) 2009 influenza [Reference Caprotta6, Reference Oliveira8, Reference Gianella10, Reference Archer12, Reference Thouillot13, 16], compared to studies from Africa or Australia (10·4–21·0%) [Reference Archer12, Reference Thouillot13, 16].
Common comorbid conditions in the patients evaluated in the studies included in our review were respiratory disease (asthma, chronic obstructive pulmonary disease), cardiovascular disease, chronic renal insufficiency, diabetes mellitus and other metabolic disorders, immune suppression, and haematological disease (Table 2).
Pregnancy
Across the different studies reporting specific relevant data, pregnant women constituted 5·4–8·1% of the cases with severe acute respiratory infection [Reference Raffo5, Reference Oliveira8], 7·4–9·1% of the unselected laboratory-confirmed cases with pandemic A(H1N1) 2009 influenza [Reference Oliveira8, Reference Schout9], and 0–13% of hospitalized cases [Reference Bantar7, Reference Thouillot13, Reference Fielding17]. They also constituted 7·1–9·1% of ICU admissions with pandemic A(H1N1) 2009 influenza, in three of the included studies [Reference Webb14, 16, Reference Fielding17], but this figure rose to 11·3% and 17·8%, respectively, in two other studies [Reference Raffo5, Reference Davies15]. Among the fatal laboratory-confirmed cases, pregnant women represented 2·1–4·2% of cases in two studies from Australia [16, Reference Fielding17] but this figure rose to 12·5% and 28·4% in two studies from Brazil [Reference Schout9] and South Africa [Reference Archer12], respectively. In the latter study, many pregnant women were co-infected with HIV or tuberculosis. Regarding women of reproductive age, pregnant women comprised 20·8% of cases with severe acute respiratory infection due to ILI, 23·3% of laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza [Reference Oliveira8], 28·6% of ICU admissions [16], and 55·6% of those that died [Reference Archer12].
Obesity
In the studies included in our review, obesity was recorded in 1·5% of patients with severe acute respiratory infection due to ILI [Reference Oliveira8], 1·8% of unselected laboratory-confirmed cases [Reference Oliveira8], 1·6–13·3% of hospitalized cases [Reference Raffo5, Reference Bantar7, Reference Thouillot13], and 28·5–44% of cases admitted to ICUs [Reference Bantar7, Reference Webb14, 16]. Among the fatal cases, obesity was noted in 14·5–21·9% [Reference Archer12, 16]. With regard to different categories of patients evaluated in a single study, obesity appeared to be more frequent in unselected laboratory-confirmed pandemic A(H1N1) 2009 influenza cases compared to cases of severe acute respiratory infection due to ILI [Reference Oliveira8], in ICU admissions compared to hospitalized cases [Reference Raffo5, Reference Bantar7], although it appeared to be less frequent in fatal cases compared to ICU admissions [16]. Morbidly obese represented 13·3% of the obese cases with severe acute respiratory infection due to ILI [Reference Oliveira8], 16·7% of unselected laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza [Reference Oliveira8], 33·0% ICU admissions [16], and 57·2% of those that died [16].
DISCUSSION
The evaluation of the published literature included in our review suggests that during the first wave of the pandemic A(H1N1) 2009 influenza in the Southern Hemisphere, the virus was frequently identified in the cases specifically tested with real time RT–PCR, compared to seasonal influenza viruses. Among the laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza, a substantial proportion was admitted to hospital, while many of those hospitalized received ICU care. The fatality rate for the severely ill patients of pandemic A(H1N1) 2009 influenza was considerable.
The age group that accounted for the majority of cases was young and middle-aged adults, followed by older children and adolescents. However, the percentage of elderly individuals increased considerably in the most seriously afflicted and diseased cases. Although relevant findings varied between studies, a considerable proportion of affected cases, even some admitted to ICUs or that died, did not have any prior comorbidity. Pregnancy and obesity represented particular risk factors for infection with pandemic A(H1N1) 2009 influenza.
Pandemic A(H1N1) 2009 influenza has shown a reversal of the pattern of the age distribution of cases compared to seasonal influenza, which mainly affects elderly individuals and young children [Reference Thompson20]. Similar observations have also been made for other recent influenza pandemics [Reference Simonsen21]. Elderly individuals may have substantial rates of protective antibodies against the novel pandemic virus, presumably due to previous immunological encounters with influenza viruses having antigenic epitopes with high homology to those of the novel strain [Reference Hancock22]. Still, elderly individuals may have worse outcome when infected with pandemic A(H1N1) 2009 influenza, because of the frequent presence of underlying comorbidity [Reference Jain23]. It should also be mentioned that young children appear to be the age group primarily affected at the start of a typical influenza outbreak, while elderly individuals may follow in this regard [Reference Glezen and Couch24].
Pregnant women represent a considerable proportion of the cases affected by pandemic A(H1N1) 2009 influenza. This could partly be attributed to the fact that the majority of the cases were noted in young and middle-aged adults (without major differences between the two sexes). Another explanation for the high incidence of pandemic influenza in pregnant women is that they often have relatively close contact with young children and therefore increased transmission opportunities for influenza. Pregnant women have been identified as a particular risk group for adverse influenza outcome during previous pandemics, and also during seasonal influenza [Reference Dodds25]. Immune system alterations, other physiological adaptations of pregnancy, and underlying comorbidity could account for this association [Reference Poole and Claman26, Reference Hartert27].
Nevertheless, across the different studies included in our review, the percentage of laboratory-confirmed cases of pandemic A(H1N1) 2009 influenza that were pregnant women, did not appear to substantially increase for the higher level of healthcare that the cases received (i.e. simple laboratory confirmation, hospitalization, ICU admission) or even for fatal cases. This could indicate that the relative risk of requirement for higher level of healthcare or of a fatal outcome might not be considerably different for pregnant women compared to the average cases of pandemic A(H1N1) 2009 influenza receiving the same level of healthcare. In other words, pregnancy might be mostly a risk factor for acquisition of infection with pandemic A(H1N1) 2009 influenza virus, rather than for an adverse outcome.
For example, in three of the studies, the percentage of pregnant women among the fatal cases of pandemic A(H1N1) 2009 influenza was lower than the percentage of pregnant women among ICU-admitted cases (Table 2) [Reference Raffo5, 16, Reference Fielding17]. This could be interpreted as a lower relative risk for death for pregnant women admitted to ICUs compared to the average cases of pandemic A(H1N1) 2009 influenza admitted to ICUs. Contrary to the above observations, the percentage of pregnant women among women of reproductive age appeared to increase with higher levels of healthcare.
A substantial proportion of the cases of influenza reported in the studies were obese. Moreover, obese individuals represented an increasing percentage of the cases with higher levels of healthcare than the cases were receiving. This trend was also observed for the percentage of morbidly obese between the obese cases. The above observations could indicate that obesity and particularly morbid obesity is an adverse prognostic factor for pandemic A(H1N1) 2009 influenza.
Although obesity has not generally been listed among the risk factors for seasonal influenza [Reference Harper28], there are studies to suggest that the outcome of bacterial infections is worse for obese patients [Reference Falagas29]. Immunological alterations associated with obesity, impaired respiratory function, and associated comorbidity (such as diabetes), along with inadequate antimicrobial drug dosing could be contributory factors to the above association [Reference Falagas29, Reference Falagas and Karageorgopoulos30]. A question that has not been adequately answered is whether the dosage of oseltamivir must be adjusted for body weight in obese patients.
The statements made above on the potential role of pregnancy and obesity as prognostic factors for infection with pandemic A(H1N1) 2009 influenza should be considered simply as observations made by the authors of this review. Formal statistical analysis to support the above statements cannot be made because the data referring to the cases of influenza receiving different levels of healthcare come from different studies and are not directly comparable. Moreover, an accurate elucidation of the value of the various potential risk factors for pandemic A(H1N1) 2009 influenza would require a multivariate analysis to be made for adjustment for potential confounders.
For example, the use of antiviral agents, including the timing of administration and dosage, as well as the availability of ICU beds, and supportive modalities like extracorporeal membrane oxygenation, might be important in the determination of the outcome of severe influenza. In this review, we primarily focused on the epidemiology of pandemic A(H1N1) 2009 influenza, particularly with regard to potential differences related to the various baseline characteristics of cases. Other reviews in the literature have addressed the value of the various treatment options [Reference Falagas31].
It also should be mentioned that most cases of pandemic influenza have a mild uncomplicated illness and remain unrecognized in the community [Reference Reed32]. Thus, the true impact and characteristics of the 2009 influenza pandemic in the Southern Hemisphere cannot be accurately reflected from the published evidence retrieved herein, as most of the included studies focused on laboratory-confirmed cases. Still, the available data provide useful information for the most seriously affected patients.
An additional limitation in the synthesis of data provided from different studies of the type evaluated in our review, is the heterogeneity regarding the methodology, setting, and period of the study, along with the characteristics of the reference population, healthcare infrastructure, and medical attitudes, worldwide. The above factors, coupled with potential changes in the response measures to the pandemic, the increasing levels of immunity in the population, and, even, the chance for evolution of the virus itself, might preclude the use of the 2009 Southern Hemisphere experience for predicting the course of the current and future influenza pandemics. In addition, the data analysed in the published studies are only a fraction of the information collected for the 2009 influenza pandemic through various relevant surveillance systems, worldwide. Further evaluation of the latter type of data could provide more meaningful conclusions.
CONCLUSION
The synthesis of the available published literature regarding the Southern Hemisphere 2009 experience for the first wave of pandemic A(H1N1) 2009 influenza suggests that pandemic influenza was a common cause of acute respiratory infection, affecting younger patients compared with seasonal influenza, and causing considerable morbidity in those presenting with serious illness, even those without any of the traditional risk factors for seasonal influenza. Pregnant women and obese individuals represented epidemiological groups in which pandemic influenza was relatively frequent and virulent, respectively.
NOTE
Supplementary material accompanies this paper on the Journal's website (http://journals.cambridge.org/hyg).
DECLARATION OF INTEREST
None.
