Discussion
This 10-year retrospective study is the first to provide long-term RSV data in the Philippines. We reported an overall prevalence rate of 11.8% among RSV-related ILI and SARI cases of children (<5y) in the Philippines. This is in agreement with rates reported from other countries including Thailand of 8.4% [15] South Korea of 13.3% [16] and France of 14.0% [17]. Meanwhile, higher prevalence rates which ranged from 19.3 to 40.6% [11, 18-20] were reported from previous studies in the Philippines, but it should be noted that these studies were done with hospitalized cases wherein a higher proportion is expected. These studies were among hospitalized patients, thus proportion of RSV cases were higher as expected. In other instances, active surveillances tend to report higher prevalence rates compared to retrospective studies [21]. Thus, the retrospective aspect of this study could have contributed to the lower reported prevalence rate. Variability among the prevalence rates around the world could be attributed to the difference in case definition, study design, sample size, location and study period.
Genetic variability of RSV viruses in general is due to the diversity of the ectodomain protein. Shifts in the dominance of RSV group can occur at different time intervals and different patterns of subgroup. RSV-A viruses were more frequently detected in epidemic seasons for the period of 1999-2007 in several countries including Germany, Belgium, Argentina and India [22]. On the other hand, the present study reported co-circulation between both RSV subgroups was notable. Increased RSV-A related cases was observed all throughout the study period, except for 2010. Similar occurrence was observed in previous studies in the Philippines (Malasao et al., 2015; Ohno et al., 2013) and other Asian countries (Auksornkitti et al., 2014; Khor, Sam, Hooi, & Chan, 2013; Yoshihara et al., 2016).
Globally, different time intervals for the replacement of RSV genotypes have been observed. A previous study reported genotype replacement every year [23]. Another study in Kenya showed that RSV-A genotype GA2 replaced GA5 in a span of seven years. On the other hand, the newly emerged RSV-A genotype ON1 only took four years to replace NA1. Notably, large epidemics were also attributed to newly emerged genotypes. In Japan, the lack of immunity against the newly emerged RSV-A genotypes (NA1 and NA2) for the periods of 2005-2007 and (ON1) 2014-2015 led to record high reported RSV cases [24, 25].
In this study, we reported the circulation of NA1, GA5 and ON1 in the Philippines between 2006 and 2016. The first RSV-A genotype to be identified was GA2 in 1998, published reports described the genotype to be more geographically widespread and epidemically active [26, 27]. In addition, outbreaks between 1996 and 2006 in Brazil were attributed to both subtypes. Previous studies have reported the global predominance of GA2 which started in 2000 [20, 27-30]. Different variants of GA2 such as NA1-NA2 [24] and NA3-NA4 (Cui et al., 2013) were identified. The present study report the detection of NA1 in the Philippines in 2007 and continued circulation up to the 2015. This indicates that NA1 circulated in some areas in the Philippines as early as 2007 which is a year earlier than the previously reported genotype detection in 2008 [11]. On the other hand, the finding is also consistent with published data from Cambodia [31], South Korea [32] and China [33] which indicated that NA1 predominantly circulated in Asia as early as 2006.
The genotype NA1 continued to be the predominant RSV-A genotype worldwide, until it was subsequently replaced by ON1 between 2009 and 2010. It has been calculated that ON1 branched out from NA1 between 2005 and 2009 [28, 29, 34]. This study reported circulation of ON1 in Pangasinan, Philippines as early as 2010, which is a year earlier than the previous report [19]. The differences between the time of emergence may be due to the variability between the location and period of the studies. Despite the lack of ON1 related cases in 2015, ON1 continued to circulate in 2016. This may be attributed to the limited number of samples tested for this period. It must be noted that similar genotype displacement events between the prevailing RSV-A genotype NA1 and ON1 have been observed and reported globally in a period of three years [35-38].
Unlike previously mentioned RSV-A genotypes (GA2, NA1 and ON1), circulating patterns of non-predominant genotypes such as GA1, GA4 and GA5 may differ temporally or geographically. In European countries including Germany and Sweden, GA5 was the predominant genotype circulating from 1999 to 2007. The prolonged circulation of GA5 in these countries could be attributed to positive selection pressure or modulation of glycosylation sites among genotypes [39]. On the contrary, other studies have described GA5 to be an endemic genotype and have limited epidemic activities [26, 27]. Occasional detection of GA5 were reported from Netherlands [40], Spain [26] Vietnam [41] and South Africa [42]. A long term study in Malaysia reported the co-circulation of GA5 with NA1, NA2 and GA2 from 2003 until 2010 [43]. In this study, a sporadic case of GA5 strain was detected in Cagayan, Philippines in 2011. The sequence of Philippine GA5 strain is highly similar (x%) with GA5 strains from neighboring Asian countries including Japan, China and Malaysia. One explanation for the high similarity GA5 strains could be due to the availability of transportation in between these Asian countries as well as influx of tourists from these countries. Interestingly, it was observed that the predominance of GA2 in 2005 inhibited the endemic genotypes (reference or data mo ba ito). It was speculated that antibodies against GA2 have broad reactivity against GA3, GA5 and GA7 [26]. Thus, possible immune cross-reaction among antibodies produced during the GA2 infection could have prevented infection of other rare genotypes [27].
A nucleotide insertion similar to ON1 was also identified in the BA genotype. The genotype BA was first identified in Buenos Aires, Argentina in 1999 and different genotypes (BA1-BA10) have diverged over a period of 20 years [37, 44-46]. This study only detected RSV-B genotypes BA2 and BA9 within the study period. Although few RSV-B sequences were included, it was apparent that most RSV-B genotype were identified as BA9. This study reported the circulation of BA9 in the Philippines in 2009. This finding is concordant with previous reports which described the emergence and predominance of BA9 during epidemics among children in the Philippines from 2009 to 2011[19, 37]. The global transmission of BA9 could be attributed to travelers or selective fitness advantage due to the 72-nucleotide duplication [47]. Taken together with other data from different countries worldwide, BA9 continued circulate globally.
Gender was not identified as a factor of among RSV-related cases in this study. Meanwhile, the study showed majority of RSV-related cases were identified among infants below two years of age. This is consistent with reports from Mexico [48] and Kenya [49]. However, some studies reported increased detection of RSV among infants below one year of age [49]. Factors such as differing airway anatomy, immaturity of the immune system and waning levels of maternal antibodies may have affected the vulnerability of younger children to severe infections [50]. Nucleolin, which is a nucleolar protein, has been reported as a receptor for RSV that binds with the F protein (Tayyari et al., 2011). Interestingly, increased quantities of nucleolin can be found in surfaces of actively dividing cells such as the alveoli which continues to grow until about two years of age. This may potentially play a role in the preferential infection of the lower respiratory tract in young children.
Although the prevalence and variability of the circulating RSV genotypes in the Philippines were reported in this study, several limitations have been identified. Primarily, due to the retrospective nature of the study which used archived ILI and SARI surveillance samples from different PNIC sentinel sites and some of the provinces were not consistently represented in the 10-year surveillance scope. Also the use of ILI and SARI case definitions may have underestimated the detection of RSV cases in this study. ILI and SARI case definition required suspect cases to present fever prior to inclusion in the influenza surveillance. Contrary to the WHO Global RSV Surveillance case definition which defined RSV cases with at least one of the following: cough, shortness of breath, sore throat, and coryza. Suspect RSV cases did not require fever presentation to be admitted in the RSV surveillance. It should be noted that there is a considerable fraction (>50%) of RSV-infected young children and elderly patients present without fever in the RSV surveillance. Another limitation is the low RNA concentration of the some samples. Several factors may affect the quality of viral RNA including: specimen collection procedure, timing of collection and specimen transportation conditions. Since this study included archived specimens in the biobank, frequent freeze thawing cycle could have contributed to the low viral RNA concentration. The distribution of RSV genotypes could have been underestimated due to the large number of RSV samples that were not amplified during heminested PCR and classified as untypable.
In conclusion, this study showed that RSV is more prevalent among younger children (<2y) in the Philippines. The study identified co-circulation of both RSV-A (GA2, GA5, NA1 and ON1) and RSV-B (BA2 and BA9) subgroups. Runny nose, bronchitis and pneumonia are clinical manifestations related to RSV infections among Filipino children. This work is the first study to report on a nationwide scale, the prevalence of RSV and genotype displacement over the period of 2006-2016.