Occurrence of AH1N1 viral infection and clinical features in symptomatic patients who received medical care during the 2009 influenza pandemic in Central Mexico

Influenza A virus is a main cause of winter epidemics that results in increments in respiratory morbidity. It constitutes a public health concern due to the burden that it represents for the health system and labor market, and for its potential to evolve into a pandemic [1–3]. The virus is sub-classified based on its surface glycoproteins, hemaglutinin and neuroaminidase, which confer the pathogenicity [1, 3–6]. Its high antigenic mutation capability is responsible for the cyclic outbreaks observed annually [1, 5].

Transmission is mainly person to person through aerosol particles expelled by ill individuals when coughing or sneezing, but also through contact with hands or contaminated fomites [5]. The incubation period ranges from hours to days [5], with an average of two days [6]. Although some patients might be asymptomatic, most present signs and symptoms of varying severity after a few days of contracting the disease [6, 7], but the majority usually recover within one or two weeks [1].

The first flu pandemic was recorded in 1538 [5]. In the last century three pandemics reached the equivalent of Category 2 on the CDC pandemic severity index (case-fatality ratio >0.1%), and all were attributed to the subtype A [1]: The Spanish flu (AH1N1) from 1918–1920 disseminated in Europe and the US causing about 50 million deaths [3, 5, 6, 8]; the Asian flu (AH2N2) from 1957–1958, responsible for 1–4 million deaths, was caused by a viral mutation that resulted in a combination of avian and human strains; and the Hong Kong flu (AH3N2) from 1968–1969 that resulted from an antigenic shift caused nearly 1 million deaths [5, 8].

The last pandemic started in Mexico in March 2009, when the Ministry of Health recognized an unexpected outbreak of respiratory disease [9] that would later be identified as AH1N1, a new viral strain resulting from a genetic combination of bird, pig and human influenza viruses [10]. The mutation detected in the nuclear M protein, responsible for the resistance to amantadine compounds, similar to that found in the AH5N1 Hong Kong virus from 1997, caused concern due to the high lethality seen with this kind of strain, and because of the high infectivity of this new virus [8].

By May, many countries started reporting cases, and as a result WHO defined the event of international importance [9], and governments implemented an emergency response plan to limit the viral spread and consequences [11]. In Mexico, educational and other non-essential activities were suspended for weeks to prevent people from getting infected [12]. By June 11, the WHO officially declared a pandemic [13].

Fortunately, this pandemic was less lethal than expected, resulting in about 18 500 deaths worldwide in 2009 [13]. In Mexico around 73 000 cases with nearly 1 350 deaths were confirmed, resulting in a lethality rate of ~1.8% [14, 15]. San Luis Potosi (SLP) state, where this study was conducted, accounted for 6.1% of the cases registered in the country, in spite of having only 2% of the Mexican population (2.5 million) [14]. It was in fact the state with the third most cases reported [15].

The limited knowledge about the symptomatology associated with the presence of AH1N1 made it very difficult for physicians to clinically distinguish this flu virus from other respiratory infections. A better understanding of the association between symptomatology and the presence of the virus could assist in future diagnostic efforts.

This study estimated the AH1N1 infection, hospitalization and mortality rates in SLP during the 2009 pandemic, and aimed at identifying clinical features associated with AH1N1 infection in individuals with flu-like illness who sought medical care.

An objective of this study was to estimate the AH1N1 infection, hospitalization and mortality rates in this representative population sample. We observed that 62.7% of the AH1N1 positive cases occurred in persons aged 0–18 years, which was very similar to the 60% reported from April 15 to May 5, 2009 in the US [10], and to the 61.8% national Mexican figure registered from March 11 to May 27, 2009 in those aged 0–19 years [2] (64.2% in this study). Based on a longer observation period, our study reinforces the observation that the AH1H1 virus affected mostly infant and adolescent populations.

The high infection rate seen indicated a rapid transmission of this virus, since nearly one out of two persons presenting respiratory symptoms such as fever, headache and cough had a confirmed laboratory diagnosis, and this supports the value of the operational definition used to screen individuals.

The hospitalization rate is more problematic to interpret and compare, as no data regarding the criteria used for hospitalization was available. While different viral pathogenicity and/or clinical severity cannot be completely ruled out, it seems that hospitalization criteria varied considerably across geographic areas, as indicated by the relatively large variation observed (5.8% in SLP, 1.5% in Mexico, and 14-15% in the US and Canada) [23].

The infection, hospitalization, and mortality rates can be compared with those of seasonal influenza in SLP during the same period, which were 5.3% (0-14y 3.7%, ≥15y 4.8%), 9.6% (0-14y 4.9%, ≥15y 12.5%), and 1.8% (0-14y 0.9%, ≥15y 2.4%), respectively [26]. While the seasonal infection rate contrasts with the much higher rate seen for AH1N1 (44.9%), the mortality rate is nearly 2.5 times higher (1.7 vs. 0.7%), pointing to the high infectivity, but low lethality of this virus, as previously observed [2, 6].

The mortality rate was also similar to that of the Mexican (0.6%) and Canadian (0.8%) national estimates [23, 24] though clearly lower than the US figure (1.9%). This difference is also difficult to explain, but it is unlikely to be due to underreporting in SLP, as health authorities mobilized most resources to track down any potential case of infection during the pandemic.

This study also looked at the association between clinical features and the presence of AH1N1 infection in persons who received medical care during the influenza pandemic in central Mexico, and the role played by age for the magnitude and direction of the association.

The main findings were that fever was associated, but only in those aged ≥15 years, while headache only in the 0–14 year group. Clear rhinorrhea and cough were positively related in both groups. Arthralgia, dyspnea and vaccination history were related to lesser risk in those aged ≥15 years, as dyspnea, purulent rhinorrhea and leukocytosis were in children.

The frequency of signs and symptoms found among infected individuals can be compared with that from other published studies. Fever, cough and sore throat were present in 96, 94, and 54%, respectively, compared to 94, 92 and 66% in the American Investigation Team [10]. According to the CDC morbidity and mortality weekly report (May 6, 2009) the most frequent clinical features reported in Mexico were fever (98%), cough (94%), rhinorrhea (83%), and headache (80%) [27]; these proportions were relatively similar to those found in SLP (96, 94, 75, 93%, respectively). However, comparisons must be treated with caution, especially with clinical signs, as diagnostic criteria used by physicians could have varied across studies. The results of this study are also in line with previous ones showing that AH1N1 infection was not sex specific [28–30].

Although several attempts have been made to relate clinical characteristics with the presence of the AH1N1 influenza virus, this study has three major strengths. A first merit lies in the size of the study population, which comprises the largest population-based sample (more than 6000 individuals) reported thus far. A second strength relates to the inclusion of both hospitalized subjects and outpatients of all ages. The third strength is related to the data treatment itself where multivariate regression analyses were used stratified by age group (0–14 and ≥15 years) defined based on the cut-off seen for the observed distribution of individuals with confirmed AH1N1 PCR-diagnosis.

The protective odds for infection found with seasonal vaccination against influenza in the adult population add to the controversy on this topic, as previous studies have reported protection [31–35], no effect [36–40], and even increased risk of infection [41]. While the potential cross-reactive protection of seasonal influenza vaccines through humoral and cell-mediated immune responses [42, 43] needs further investigation, an upcoming review to assess the protection offered by influenza vaccines against circulating influenza A or B viruses that are not antigenically well-matched to vaccine strains will help elucidating this issue [44].

An additional finding is that clinical manifestations commonly related to bacterial respiratory infections, such as purulent rhinorrhea or leukocytosis, were negatively associated with the likelihood of positivity. This finding is similar to that reported by other authors. In a sample of 362 patients presenting with flu syndrome to an emergency unit in Spain, nasopharyngeal swabs were taken for AH1N1 detection with PCR; results showed that positive cases had significantly lower mean leucocyte counts, and the association remained in the multivariate logistic regression analysis when the lymphocyte count was used [45]. However, it is worth noting that purulent rhinorrhea could also result from a superimposed bacterial respiratory infection [6].

Similarly, the symptomatology associated with lower respiratory tract infections and its severity (e.g. cyanosis, dyspnea and tachypnea) was also related to protective odds. This finding might have to do with the time period between the onset of the symptoms and the visit to the health unit, as an indicator of the progression of the infection. However, no differences were seen in the mean number of days elapsed from the reported start of the symptoms to the health visit (~3 days) between those who had and did not have these symptoms (p > 0.05).

Very few clinical features might be relatively specific to flu viral infections [27, 46, 47], such as the presence of clear rhinorrhea that relates to the hemmagglutinin inhibiting effect on the surface glycoproteins that protect the respiratory tract cilia [4]. Respiratory diseases caused by influenza and parainfluenza viruses, adenoviruses or syncytial respiratory virus usually produce similar unspecific clinical symptomatology, characterized by fever, chills, malaise, headache, myalgia or cough [48]. In fact, a recent review of multivariate models devised to clinically diagnose influenza reported that only the combination of fever, cough and acute onset has a modest accuracy [49].

Compared with confirmed cases of seasonal influenza in SLP during the same period, those with AH1N1 infection aged ≥15 years showed significantly (p < 0.05) higher proportions of cough, clear rhinorrhea, nasal congestion, sore throat, malaise, chills, and myalgia (ranging from 8.3 to 16.4% higher), but had lower proportion of dyspnea (5.3% less); however, among those aged 0–14 years, the proportions were significantly higher only for cough (8.3%) and clear rhinorrhea (8.7%) [26].

For the AH1N1 influenza virus infection, various studies have not been able to identify symptoms that could predict the presence of AH1N1 infection using various designs and analytical procedures [50–52]. However, others have identified signs and symptoms associated to positive status. A retrospective study with 117 adult cases and 236 matched controls presenting with respiratory symptoms to hospital emergency departments in Toronto, Canada, showed higher associations for AH1N1 when various combinations of signs, symptoms and laboratory indicators were used, though only age, cough and fever remained associated in the multivariate analyses [53]. Another study with military personnel who visited primary health care clinics for febrile respiratory illness (i.e. fever, cough and sore throat) compared the signs and symptoms between those with positive and negative AH1N1 influenza virus using real-time PCR. Of the 2858 subjects recruited 821 were influenza cases, of which 434 were 2009 pandemic influenza AH1N1. The comparison of clinical features using multivariate logistic regression showed that sore throat, photophobia, injected pharynx and nausea/vomiting were negatively associated, while running nose, chills, fever and eye symptoms were positively related [54].

This study estimated the frequency and progression on the AH1N1 influenza infection in symptomatic individuals, and was able to identify various associated clinical features that revealed the age specificity of several of them, indicating the importance of this factor when establishing the presumptive clinical diagnosis.