Large-scale outbreak of Chikungunya virus infection in Thailand, 2018–2019

Between 2018 and 2019, the incidence of chikungunya was approximately 15,000 cases across 60 provinces in Thailand. Here, the clinical presentations in chikungunya, emergent pattern, and genomic diversity of the chikungunya virus (CHIKV) causing this massive outbreak were demonstrated. A total of 1,806 sera samples from suspected cases of chikungunya were collected from 13 provinces in Thailand, and samples were tested for the presence of CHIKV RNA, IgG, and IgM using real-time PCR, enzyme-linked immunoassay (ELISA), commercial immunoassay (rapid test). The phylogenetic tree of CHIKV whole-genome and CHIKV E1 were constructed using the maximum-likelihood method. CHIKV infection was confirmed in 547 (42.2%) male and 748 (57.8%) female patients by positive real-time PCR results and/or CHIKV IgM antibody titers. Unsurprisingly, CHIKV RNA was detected in >80% of confirmed cases between 1 and 5 days after symptom onset, whereas anti-CHIKV IgM was detectable in >90% of cases after day 6. Older age was clearly one of the risk factors for the development of arthralgia in infected patients. Although phylogenetic analysis revealed that the present CHIKV Thailand strain of 2018–2020 belongs to the East, Central, and Southern African (ECSA) genotype similar to the CHIKV strains that caused outbreaks during 2008–2009 and 2013, all present CHIKV Thailand strains were clustered within the recent CHIKV strain that caused an outbreak in South Asia. Interestingly, all present CHIKV Thailand strains possess two mutations, E1-K211E, and E2-V264A, in the background of E1-226A. These mutations are reported to be associated with virus-adapted Aedes aegypti. Taken together, it was likely that the present CHIKV outbreak in Thailand occurred as a result of the importation of the CHIKV strain from South Asia. Understanding with viral genetic diversity is essential for epidemiological study and may contribute to better disease management and preventive measures.


Introduction
symptom onset were analyzed along with the laboratory results, which demonstrated the importance of proper diagnostic tests to confirm CHIKV infection.

Materials and methods
The research proposal was approved by the institutional review board of the Ethics Committee of the Faculty of Medicine, Chulalongkorn University, Thailand (IRB no. 630/61). The IRB waived the need for informed consent from the participants, because the clinical specimens were anonymous. The study adhered to the Declaration of Helsinki and Good Clinical Practice Guidelines (ICH-GCP).

Sample collection
Serum or plasma samples were collected from patients with suspected chikungunya along with demographic and clinical data (age, sex, fever, joint pain, rash, and conjunctivitis). A suspected case of CHIKV was defined as a patient presenting with acute onset of fever (> 38.5˚C) with or without severe joint pain and skin rash, particularly in a person who residing, traveling, and working in an epidemic area with a high risk of CHIKV transmission. Samples were sent for laboratory diagnosis at the Center of Excellence in Clinical Virology, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand. From October 2018 through February 2020, 1806 samples were collected from 13 provinces throughout the Central (Bangkok 1,137 cases, Samut Prakan 173 cases, Samut Sakhon 344 cases, Nonthaburi 1 case); Western (Ratchaburi 1 case, Kanchanaburi 1 case); Northern (Tak 1 case); Northeastern (Maha Sarakham 1 case); Eastern (Chon Buri 34 cases); and Southern (Ranong 1 case, Phuket 108 cases, Songkhla 1 case, Surat Thani 3 cases) regions of Thailand (Fig 1). A confirmed case of chikungunya was defined as a suspected case plus laboratory confirmation either by real-time RT-PCR to detect CHIKV viral RNA or serological test to detect CHIKV-specific IgM in serum. Patients possessing CHIKV-specific IgG without IgM were classified as having had a past CHIKV infection.

Detection of CHIKV by real-time reverse transcription-polymerase chain reaction (RT-PCR)
Total RNA from serum samples was extracted using magLEAD 1 kit and amplified by qualitative real-time RT-PCR based on fluorescent hybridization probe assay using the primer sets and PCR conditions as described in a previous report (1) Detection was performed using LightCycler 1 480 Instrument II (Roche, Basel, Switzerland).

Detection of anti-CHIKV IgM and IgG antibodies by rapid test
In this study, IgM/IgG antibodies against CHIKV in all serum samples were assessed using commercial fluorescence immunoassay on nitrocellulose membranes of a test device (SD BIO-SENSOR, Gyeonggi-do, Korea) according to the manufacturer's instruction. Briefly, serum was added to the membrane of the test device, which consisted of immobilized anti-human IgM or anti-human IgG antibodies on two individual test lines. The assay diluent was then added to release the chikungunya antigen from the antigen pad, and the anti-CHIKV envelope monoclonal antibody conjugated with europium from the conjugation pad. The intensity of the fluorescence signal was evaluated with a STANDARD F200 Analyzer (SD BIOSENSOR, Gyeonggi-do, Korea).

Detection of anti-CHIKV IgM and IgG antibodies by Enzyme-Linked Immunosorbent Assay (ELISA)
In this study, the commercial ELISA kits (EUROIMMUN, Lu¨beck, Germany) [16] were performed to determine anti-CHIKV IgM/IgG antibodies with only serum from the first consecutive 310 CHIKV-infected patients whose sera samples were positive for CHIKV RNA by realtime PCR or positive for anti-CHIKV IgM by rapid test. The assay was performed as recommended by the manufacturer. In brief, patient serum (diluted 1:100) was added to the microplate coated with the recombinant CHIKV protein. After incubation, the wells were washed and incubated with peroxidase-conjugated anti-human IgM/IgG. After removal of the unbound antibodies, the substrate solution was added, and the colorimetric output was measured at 450 nm after the enzymatic reaction was stopped. The OD ratios obtained from each sample were interpreted according to the manufacturer's recommendations.

Genotypic characterization of CHIKV
CHIKV whole-genome and envelope 1 (E1) gene were amplified from the positive real-time RT-PCR samples by using primer sequences as previously reported [17]. The amplified products were excised and purified with the Expin TM Combo GP kit (General Biosystem, South Korea) as per the manufacturer's instructions. The purified PCR products were subjected to sequencing analysis (1st BASE, Singapore). The nucleotide sequences were examined using NCBI BLAST program (http://blast.ncbi.nlm.nih.gov/Blast.cgi) and were edited using CHRO-MAS LITE v. 2.0 program. The edited sequences were assembled by using BioEdit Sequence Alignment Editor v. 7.0.5.3 program. The obtained sequences were aligned with the known CHIKV sequences representing various genotypes from GenBank. The phylogenetic trees of CHIKV whole-genome and E1 gene were constructed using MEGA X program. All sequences were deposited in the GenBank database. The GenBank accession numbers of each sequence are shown in S1 File.

Data analysis
The epidemiological surveillance of patients with CHIKV infection was performed by the Bureau of Epidemiology, MoPH, Thailand, and our recorded data. We retrieved these data, tabulated, and analyzed them using GraphPad Prism8 (GraphPad Software Inc. San Diego, CA) and Microsoft Excel (Microsoft Inc., Redmond, WA). The qualitative data were shown as n (%). The correlation analysis between the date of onset and the presence of IgM and CHIKV RNA was plotted and fit by simple linear regression with correlation coefficient R 2 . The agreement between ELISA assay and rapid test were assessed by Cohen's kappa value. A chi-square test and logistic regression analysis were performed to examine the association between patient demographics including sex, age, and joint pain. Statistical analysis of significance was analyzed by using SPSS software, version 22. P value of less than 0.05 was considered as statistically significant.

Chikungunya virus infection in Thailand
We investigated samples from 1,806 patients with suspected CHIKV infections that were sent to our center from October 2018 to February 2020. Among 1,806 suspected cases, 1,295 (547 male and 748 female) patients were confirmed to have CHIKV infection based on positive results of real-time RT-PCR and/or IgM antibody testing. As shown in Fig 2, the peak of CHIKV-positive cases in 2018 was reached in December, which was similar to the report by MoPH, Thailand. As mentioned above, a large-scale outbreak in Thailand was observed in 2019. The highest number of positive cases detected in our center was 372 cases in November 2019.

Demographic and clinical presentation
CHIKV-infected patients were further classified according to age groups and symptoms ( Table 1). The age ranges of patients, including in this study, were 3 to 97 years old. We found that chikungunya fever occurred in all ages. According to this study, the percentages of CHIKV-confirmed cases was higher in females than males (57.8 vs. 42.2%). The presence of joint pain was slightly lower in males (70.2%) than in females (74.6%), but these differences did not reach statistical significance (p = 0.079 for sex and p = 0.089 for the presence of joint pain). The most affected age range was between 31 and 40 years, with 299 cases (23.1%). Patients infected with CHIKV commonly presented with fever (96.6%), rash (50.2%), and  joint pain (72.7%). Joint pain is a common presenting symptom similarly observed in patients with rheumatoid arthritis and is a significant cause of distress in patients with chikungunya. In our study, 72.7% (942/1295) patients complained of joint pain. CHIKV-infected patients with joint pain were further analyzed according to age groups. As shown in Table 1, the presence of joint pain in CHIKV-confirmed cases was associated with age (p < 0.001). The highest percentage (81.5%) of patients with joint pain was observed in patients over 60. The prevalence of joint pain in elderly patients, particularly in the age group 51-60 (p = 0.021) and above 60 years old (p = 0.010), was significantly higher than in the young age group (11-20 years; reference group).

Comparison of real-time PCR, rapid IgM/IgG, and ELISA IgM/IgG for the diagnosis of CHIKV
In this study, real-time PCR and chikungunya IgM rapid tests were used to screen 1,806 suspected cases. CHIKV infection was confirmed in 1,295 samples by real-time RT-PCR, and/or rapid IgM assay was shown concerning the timing of sampling after the onset of disease ( Fig  3). The samples positive by RT-PCR were higher than that of the rapid IgM assay if the sample was obtained during the first five days of disease onset. In contrast, the number of positive samples by using a rapid IgM assay was higher than that of RT-PCR if the sample was obtained after at least six days of disease. Our large sample size was advantageous, as we could perform  Table 2.
Among the 310 consecutive samples tested, 71.29%, 38.39%, and 38.06% were positive by RT-PCR, rapid IgM, and ELISA IgM, respectively. Although RT-PCR is considered a sensitive assay during the first few days after infection, 93.92% of the confirmed cases were positive at days 1-3. Rapid IgM and ELISA IgM were positive in 8.11% and 6.08% patients, respectively, during days 1-3. RT-PCR was positive in <40% patients from day 7 onwards following the onset of fever and was positive <10% from day 13 onwards following the onset of fever. Both IgM assays were positive in more than 90% of patients from day 7 onward following the onset of fever. Notably, an excellent agreement (Cohen's kappa value of 0.97) was observed between the rapid and ELISA IgM as demonstrated in S2 File. The percentages of real-time RT-PCR positive samples were <50% after 5 days, whereas the positive results by IgM antibody detection were >50% from day 6 after fever onset.
As shown in Table 2, the percentages of IgG-positive cases increased slower than those of IgM-positive cases. Among 310 consecutive samples tested, 12.90%and 18.06% were positive by rapid IgG and ELISA IgG, respectively. ELISA IgG was positive in 91.18% of patients after day 13, while rapid IgG was positive in 64.71% of patients after day 13. These results indicate that IgG detection by ELISA showed higher sensitivity than IgG detection by the rapid test. Nevertheless, the agreement between rapid and ELISA IgG showed Cohen's kappa value of 0.76, which represented the substantial agreement between rapid and ELISA IgG.

Genotypic characterization of CHIKV detected in Thailand in 2018-2020
Phylogenetic tree analysis of CHIKV whole genomes was carried out from CHIKV isolated from six patients from the Central (Samut Sakhon, Bangkok, Samut Prakan); Eastern (Chonburi): Lower Northern (Tak); and Southern (Ranong) parts of Thailand with the reference CHIKV published previously in the GenBank database (Fig 4). Moreover, partial E1 phylogeny was also generated from 251 CHIKV isolated from various regions of the country, including the Central (Bangkok, Samut Sakhon, Samut Prakan, Nonthaburi); Eastern (Chonburi); and Southern (Phuket) regions of the country from October 2018 to February 2020 (Fig 5). From phylogenetic tree analysis, CHIKV isolates during the outbreaks in late 2018 to early 2020 belong to the Indian Ocean lineage within the ECSA genotype. However, these virus strains did not cluster with CHIKV strains previously circulating from the first massive outbreak in Thailand during 2008 to 2009. The current viruses were grouped with Bangladeshi strain of    (Figs 4 and 5). We further investigated the similarity at nucleotide levels of the complete CHIKV genome among Thai strains in this present study (2018-2020) and the past outbreaks (1958,1988,2008,2009 The changes of amino acid residues in the E2 envelope protein, including G205S, K252Q, and V264A were observed. Additionally, the distinct amino acid mutations in the capsid and nonstructural proteins (nsPs) between the present Thai strains and the past were located in the capsid (K73R), nsP2 (H130Y, E145D, N495S, S539L, and V793A), nsP3 (D372E), and nsP4 (S55N and R85G) (S4 File).

Discussion
Although CHIKV was identified in 1952, there is still no specific antiviral drug or vaccine against CHIKV. CHIKV infection is not life-threatening, but infected patients commonly suffer from severe joint pain resulting in prolonged disability in some patients [18][19][20]. Arthralgia is a common clinical manifestation of patients hospitalized for CHIKV infection. Our study showed that older age is a significant risk factor for arthralgia associated with CHIKV infection. Accordingly, many studies have shown that increasing age is a predictive factor for the development of rheumatic sequelae in patients with chikungunya fever [21][22][23]. The use of proper laboratory tests for the diagnosis of CHIKV infection is essential for patient management and outbreak surveillance. We showed the usefulness of the detections of viral RNA and IgM specific to CHIKV antigen for diagnosis. Information on the duration of patient symptoms is crucial for the proper selection of laboratory testing. During the first week following infection, RT-PCR is the most appropriate test. However, patients who visit the doctor later after disease onset typically show negative RT-PCR results. The detection of the IgM antibody could be more useful in these cases. The IgG antibody is usually detectable later than the IgM antibody and is not commonly used for early diagnosis. The presence of IgG could be from past infection. Our study included a large group of patients whose samples were collected at different time points after disease onset. We concluded that for the diagnosis of chikungunya, qualitative real-time RT-PCR was highly sensitive between days 1-5 after the onset date, and IgM testing was highly sensitive after day 7. These findings corroborate with previous studies [24][25][26] as well as World Health Organization (WHO) guidelines for laboratory diagnosis of chikungunya [27]. Besides the laboratory diagnosis, the continuous surveys and studies of genetic variation of CHIKV are crucial for infection control and understanding outbreaks for further development of the vaccine, specific antiviral drugs or prevention of future CHIKV outbreaks. Phylogenetic tree analysis revealed that CHIKV Thai strains isolated from patients in 2018-2020 belong to the ECSA genotype, the same genotype as CHIKV responsible for the massive outbreak in 2008-2009. However, genetic analysis of the present Thai strains of 2018-2020 showed marked differences compared to CHIKV reported in 2008-2009. CHIKV isolated from the massive outbreak in 2008-2009 in Thailand carrying E1-A226V mutation [28], whereas the causative agents for outbreaks before 2008 was caused by Asian genotype [17,29].
We have previously reported that the CHIKV isolated from the massive outbreak in 2008-2009 clustered evolutionally with the CHIKV strain of Sri Lanka, Singapore, and Malaysia [12]. Several studies have shown that E1-A226V mutation became more adapted for the CHIKV transmission by Aedes albopictus [30][31][32]. However, this variant would not sufficiently affect the evolution of the virus transmitted by Aedes aegypti. The mutation of E1-A226V enhances virus infectivity in the midgut cells of Aedes albopictus, resulting in increasing CHIKV dissemination and transmission by Aedes albopictus [33]. Remarkably, we found that none of the CHIKV ECSA strains isolated during the second massive outbreak in Thailand in late 2018 to early 2020 possess E1-A226V mutation. The present Thai strains of 2018-2020 harboring E1-226A without valine substitution at position 226 of E1 envelope glycoprotein were similarities to the first reported Asian genotype outbreak in 1958 in Thailand [11]. Interestingly, our study revealed that all present Thai strains of 2018-2020 possess two mutations, E1-K211E and E2-V264A, in conjunction with E1-226A. Our previous study also showed that CHIKV isolated from Aedes aegypti mosquito in Bangkok in August 2018 harboring these two mutations [15]. One document showing that E1: K211E variant was observed earlier in the CHIKV Asian genotype isolated from India in 1963 [34]. Double mutant virus containing E1-K211E and E2-V264A mutations in the background of E1-226A in the ECSA genotype was first observed in India in the year 2010 [35]. This mutated virus was also found in India in 2015-2017, Pakistan in 2016, Italy in 2017, and Bangladesh in 2017 [36][37][38][39]. There have been reported that positive selection had a dramatic effect on the alteration of amino acid residue from lysine (K) to glutamic acid (E) at position 221 of CHIKV E1 protein (E1: K211E) [40]. Previous studies reported that mutations on the E1 and E2 of CHIKV envelope glycoproteins impacted on vector competence, transmission efficiency, and virus pathogenicity [30,32,[41][42][43]. Agarwal et al. demonstrated that combined E1-K211E and E2-V264A in the background of E1-226A was responsible for enhancing virus infection in Aedes aegypti, which led to more efficient CHIKV dissemination and facilitated the transmission of CHIKV in Aedes aegypti compared to wildtype CHIKV E1-226A. Nevertheless, these adaptive mutations in the background of E1:226A showed no significant impact on CHIKV fitness for Aedes albopictus mosquito [44]. Our study characterized the CHIKV isolated from various parts of Thailand, mostly from the patients in urban areas, where Aedes aegypti could be the most likely responsible vector. Therefore, present Thai strains harboring E1: K211E and E2: V264A in the background of E1:226A, responsible for the higher fitness for the Aedes aegypti vector [44], may facilitate viral circulation and persistence in the urban area of Thailand.
Notably, our study reveals that CHIKV Thai strains of 2018-2020 harboring E2-Q252K, while this mutation was not found in CHIKV responsible for the large scale outbreak in Thailand during 2008-2009. CHIKV Thai strains of 2008-2009 possessed K252Q substitution, associated with Aedes albopictus fitness [45,46]. These results indicated that the present Thai strains might not harbor one of the Aedes albopictus-adaptive mutations. Moreover, substitutions in other amino acid residues such as E1-I317V and E2-G205S were observed in our study (S4 File). The E2-G205S detected in our isolates was also reported in the Italian isolates from the outbreak in Italy in 2017 [38]. Notably, the E1-I317V mutation, which was recently found in India, Bangladesh, and Italy [38,[47][48][49], was detected in our study. Nevertheless, the significant impact of E1-I317V, E2-G205S, and E2-Q252K mutations on CHIKV evolution is still not known. Therefore, further investigation is needed to evaluate the role of these mutations.
A comparison of phylogenetic relationship of 6 whole-genome sequences and 251 partial E1 from CHIKV isolated during the 2018-2020 outbreaks in Thailand with other sequences available in the GenBank showed that all the sequences clustered with representative South Asian isolates of 2016-2017 rather than the CHIKV Thailand isolated from the previous outbreak in 2008-2009, indicating that the recent outbreak did not originate from the circulating strain in Thailand. Interestingly, the Bangladeshi strain of 2017 shared the highest nucleotide identity with the present Thai strains of 2018-2020. Overall, our results suggested that the recent outbreak of CHIKV in Thailand during 2018-2020 belongs to the ECSA genotype originating from other countries in South Asia, mostly Bangladesh.

Conclusions
This study investigated a large population of CHIKV-infected patients to demonstrate their clinical profiles and the diagnostic approaches of CHIKV infection. We found that joint pain in chikungunya was widespread in infected patients after middle age and showed an increase in older patients. Regarding laboratory diagnoses of CHIKV infection, real-time PCR detection of CHIKV RNA was more sensitive for samples obtained within five days of disease onset, whereas IgM assay would be more sensitive for samples obtained after six days of fever onset. Moreover, our genetic analysis revealed the changes of CHIKV from the first massive outbreak during 2008-2009 in the country. The present CHIKV Thai strains are closely related to the viral strains from South Asia, especially Bangladesh. Here, our findings revealed that the second massive outbreak of CHIKV virus in Thailand occurred due to the importation of the viral strain from Bangladesh. The continuous surveys and studies of genetic variation of CHIKV are crucial for infection control and further development of specific drugs or prevention of CHIKV outbreaks.