Risk of yellow fever virus transmission in the Asia-Pacific region

In 2016, 11 yellow fever (YF)-infected workers returning from Angola to China highlighted the threat of a YF epidemic in Asia. Despite endemicity in Africa and South America, YF has remained absent from Asia for reasons that are not fully understood, including possible mosquito–virus genotype incompatibilities, human viraemia dynamics, lack of a sylvatic cycle, or competition with established flaviviruses such as dengue and Japanese encephalitis. YF can be severe with in-hospital case fatality rates up to 67%. Historically, the understanding of YFV transmission and the development of the 17D vaccine reduced burden, yet substantial annual severe cases and deaths still occur, especially in Africa, due to factors including insecticide resistance and vaccine supply limitations. Transmission requires successful infection of the mosquito midgut, dissemination into the haemocoel, and infection of salivary glands for onward transmission; vector competence metrics (infection, dissemination, transmission) quantify this capability. Ae. aegypti and Ae. albopictus are widely distributed in Asia-Pacific, where ecological changes favor Ae. aegypti in urban areas. Increasing travel between Asia and Africa raises importation risk, as illustrated by laboratory-confirmed YFV cases among travelers to China during 2015–2016. Given immunologically naive populations and suitable vectors in Asia, assessing the vector competence of local Aedes populations for YFV is crucial to evaluate emergence risk. This study investigates 18 populations of Ae. aegypti and Ae. albopictus from the Asia-Pacific region to determine their competence for a West-African YFV genotype and to assess regional transmission risk.

The study builds on limited prior work assessing YFV vector competence in the Asia-Pacific region. Only two earlier studies were noted: (1) Asian Ae. aegypti from Laos transmitted YFV (strain S-79) at least 14 days post-exposure under laboratory conditions; (2) when challenged with an American genotype 1 YFV strain (74018, Brazil), Ae. aegypti from Cambodia and Vietnam were susceptible but showed lower dissemination efficiencies than reported here. Background literature emphasizes Ae. aegypti as the primary urban vector in Africa and South America, the historic spread of YFV and Ae. aegypti via ships, successful early vector control and vaccine development, and the modern resurgence risks due to vector control relaxation, insecticide resistance, and vaccine supply constraints. Global distribution models show extensive Ae. aegypti and Ae. albopictus presence across Asia-Pacific, and travel patterns indicate increased connectivity with YF-endemic regions, heightening introduction risk. Existing modeling studies have assessed international spread via air travel and global infection risk zones, but empirical vector competence data for Asia-Pacific populations remained scarce prior to this work.

Study design: Experimental infections assessed vector competence (infection rate, dissemination rate, transmission rate) at 14 and 21 days post-infection (dpi) for 12 Ae. aegypti and 6 Ae. albopictus field-derived populations from Asia-Pacific. Additional African Ae. aegypti (Cameroon and Congo) were included for comparative viral load analyses. Ethics: Animal use approved by Ethics Committee #89 (APAFIS #6573-2016061412077987 v2), compliant with EU and French regulations. Virus: YFV strain S-79 (West-African genotype; accession MK060080), isolated from a patient returning from Senegal in 1979; passaged twice in mouse brain and twice in C6/36 cells. Stocks produced on C6/36 and stored at -80°C. Mosquito populations: Eggs collected via ovitraps at each locality (12 Ae. aegypti from Cambodia, Vietnam, Laos, Thailand, Singapore, Taiwan, New Caledonia, French Polynesia; 6 Ae. albopictus from China, Japan, Taiwan, Laos, Thailand, Brazil). Eggs shipped to Institut Pasteur, Paris. Larvae reared under controlled conditions to adults. Infectious blood meal: Ten-day-old females starved 24 h and infected in BSL-3. Blood meal: 1.4 mL rabbit erythrocytes + 10 mM ATP + 0.7 mL viral stock to reach 10^7 ffu/mL, delivered via Hemotek at 37°C. Post-feeding, mosquitoes maintained at 28°C, 80% RH, with 10% sucrose until processing at 14 and 21 dpi. Sample processing: Saliva collected by forced salivation for 30 min (proboscis inserted into 5 µL FBS; expelled into 45 µL L-15 and stored at -80°C). Bodies and heads homogenized in 300 µL L-15 + 2% FBS. After centrifugation, supernatants used for virus detection. Virus titration: Serial dilutions inoculated onto C6/36 cells in 96-well plates (50 µL per well, 1 h at 28°C), overlay with CMC + L-15, incubated 5 days at 28°C. Fixation with 3.6% formaldehyde; immunostaining with anti-YFV primary (NB100-64510, 1:200) and fluorescent secondary (A-11029, 1:1000). Foci counted by fluorescence microscopy; titers as ffu/sample. Saliva titers expressed as ffu/saliva. Vector competence metrics: Infection rate (IR) = proportion with infected bodies among blood-fed; Dissemination rate (DR) = proportion with infected heads among infected bodies; Transmission rate (TR) = proportion with infectious saliva among those with infected heads. Transmission efficiency used for risk mapping. Risk mapping: Extracted probabilities of Ae. aegypti occurrence (5 km resolution) from Kraemer et al. at sampling points and averaged locally. Combined with measured transmission efficiencies to estimate regional YFV transmission risk for selected sites in Asia. Statistics: IR, DR, TR compared with Fisher’s exact tests. Viral titers compared with Kruskal–Wallis tests. Correlations between body, head, and saliva titers estimated; linear regression assessed geographic effects on viral loads; logistic regression explored conditioning factors. P<0.05 considered significant with Holm–Bonferroni adjustments as needed. Analyses in Stata.

- Ae. aegypti competence at 14 dpi: IR ranged from 41.7% (CAMB, Cambodia) to 95.8% (TRUNG, Vietnam; CSP, Thailand; TAINAN, Taiwan; NOUMEA, New Caledonia); significant differences across populations (Fisher’s exact P<1e-4). DR ranged 42.8% (FENG, Taiwan) to 86.9% (CSP, Thailand; NOUMEA); some variation (P=0.06 overall; Asian populations P=0.13; Pacific P=0.04). TR detected in 5/12 populations (CSP, SING, ANNAN, FENG, NOUMEA), ranging 12.5% (SING) to 45% (CSP); significant differences (P<1e-4). - Ae. aegypti competence at 21 dpi: IR reached 100% in BLX (Laos), CSP (Thailand), SING (Singapore), NOUMEA (New Caledonia); differences across populations (P<1e-4). DR ranged 47.6% (NANZI, Taiwan) to 95.8% (CSP, Thailand) (P<1e-4). TR ranged 10% (VIET, Vietnam) to 56.5% (CSP, Thailand); no significant difference across populations (P=0.10). All 12 populations transmitted by 21 dpi (12/12), while 5/12 transmitted at 14 dpi (42%). - Ae. albopictus competence: At 14 dpi, IR 4.2% (THAI, Thailand) to 62.5% (FOSHAN, China) (P<1e-4); DR 0% (THAI) to 85.7% (LINGYA, Taiwan) (P=0.41). Transmission observed only in FOSHAN (TR=22.2%). At 21 dpi, IR 8.3% (THAI) to 54.2% (FOSHAN) (P=0.003); DR 0% (XKH, Laos; THAI) to 100% (LINGYA) (P=0.038). TR up to 66.7% (LINGYA). Four populations (YYG, Japan; XKH, Laos; THAI, Thailand; PMNI, Brazil) did not transmit at 14 or 21 dpi. Overall, Ae. albopictus were less competent than Ae. aegypti. - Saliva viral loads: Ae. aegypti at 14 dpi: 10^(1.6±1.5) ffu/saliva (NOUMEA) to 10^3 (FENG). At 21 dpi: from 5 ffu (VIET) to 10^(3.7±4) (NANZI), min–max 10–23,000 particles. Ae. albopictus at 14 dpi: 10^(1.7±1.7) (FOSHAN); at 21 dpi: 10^(2.2±1.4) (LINGYA), min–max 133–167. - Comparative viral loads (Ae. aegypti transmitters, 21 dpi): Body and saliva loads showed no significant regional differences (Kruskal–Wallis P>0.05; regression P=0.11 for body, P=0.54 for saliva). Head loads significantly higher in African mosquitoes (10^(4.6±3.7)) vs Asia (10^(3.9±3.9)) and Pacific (10^(3.7±3.6)); Kruskal–Wallis P=0.0095; regression coefficients: Asia −0.73 log (95% CI −1.21 to −0.26), Pacific −0.82 (−1.53 to −0.12); P=0.01. - Correlations of viral load: Body vs head: correlation coefficient p=0.31, P=0.012; body vs saliva: p=0.22, P=0.11; head vs saliva: p=0.04, P=0.77. - Modeled transmission risk (Ae. aegypti, Asia): Highest estimated risks at CSP (Thailand) 54% (95% interval 32.8–74.4%), TRUNG (Vietnam) 25% (9.8–48.7%), and NANZI (Taiwan) 21% (7.1–42.2%), combining local vector occurrence probabilities with measured transmission efficiencies.

This study directly addresses whether Asia-Pacific Aedes populations can transmit YFV by experimentally demonstrating that multiple Ae. aegypti populations from the region are highly competent, with all tested populations transmitting by 21 dpi and several transmitting by 14 dpi. In contrast, regional Ae. albopictus populations generally showed lower infection, dissemination, and transmission capacities, indicating a lesser role in potential urban YFV transmission relative to Ae. aegypti. Viral loads excreted in Ae. aegypti saliva in Asia-Pacific were comparable to those from African vectors once transmission occurred, implying similar potential for human infection. However, Asia-Pacific Ae. aegypti showed significantly lower head viral loads than African counterparts, suggesting reduced dissemination efficiency past the midgut in Asian-Pacific mosquitoes; the salivary gland barrier appears comparatively more limiting than the midgut. These findings, together with high modeled Ae. aegypti occurrence across Southeast Asia and documented travel-linked YFV importations, imply a substantial risk of YFV emergence in the Asia-Pacific. The discussion highlights that vector competence alone does not define field transmission; environmental and ecological factors (e.g., mosquito lifespan, feeding behavior, population density) modulate vector capacity. Potential sylvatic components could involve Asian Macaca spp. and zoophilic mosquitoes, while interactions with endemic flaviviruses such as JEV may influence YFV emergence dynamics. Preparedness should pair surveillance for imported cases with vaccination strategies and innovative vector control (e.g., Wolbachia), acknowledging current vaccine supply constraints.

Ae. aegypti populations from across the Asia-Pacific region are competent YFV vectors, with multiple populations capable of transmission within 14 days and all tested populations by 21 days post-infection. Ae. albopictus populations are generally less competent. Salivary viral loads from transmitting Ae. aegypti are comparable to African vectors, though Asia-Pacific mosquitoes show lower head viral loads, indicating reduced dissemination efficiency. Combined with high Ae. aegypti occurrence in Southeast Asia, these results suggest that vector populations are unlikely to prevent YF emergence in the region. Future research should examine viral evolution and bottlenecks during mosquito dissemination (especially post-midgut), the role of mosquito immune responses (e.g., RNA interference), potential interactions with endemic flaviviruses, and the existence or feasibility of sylvatic cycles in Asia. Public health preparedness should prioritize early detection of imported cases, vaccination policies (including fractional dosing during shortages), securing vaccine stockpiles, and implementing effective vector control strategies such as Wolbachia-based interventions.

- Laboratory conditions may not reflect field realities; environmental factors could shorten mosquito lifespan and reduce the probability of surviving the extrinsic incubation period to transmit. - The infectious blood meal titer (10^7 ffu/mL) may differ from typical patient viremia levels (approximately 4.98 [3.50–5.79] log10 copies/mL), potentially affecting observed competence. - Forced salivation may not represent the physiological dose delivered during actual mosquito biting. - Vector competence does not encompass all components of vector capacity, which also depends on vector density, feeding behavior, host availability, and environmental conditions. - A single YFV strain (West-African genotype, S-79, 1979) was used; results may vary with other genotypes/strains. - Some Ae. albopictus populations were long-term lab colonies, which could influence competence relative to contemporary field populations.