養殖物種水生病毒傳輸動態之模擬及感染控制策略

Modeling the waterborne virus transmission dynamics in aquaculture species with infection control measure implications

呂恬萱(Tien-Hsuan Lu)

指導教授 ： 廖中明

國立臺灣大學/生物資源暨農學院/生物環境系統工程學研究所/碩士(2019年)

https://doi.org/10.6342/NTU201901544

摘要

雙核糖核酸病毒(Birnavirus)、吳郭魚湖泊病毒(Tilapia lake virus)及神經壞死病毒(Nervous necrosis virus)分別為造成臺灣養殖文蛤、吳郭魚以及石斑魚幼苗死亡之水生病原。這些重要養殖物種大量死亡會對養殖產業造成重大經濟損失。傳染性疾病之傳輸動態模擬已被認定可為疾病管理政策提供見解，因此，本論文之目的為：(1)評估病毒濃度對累積死亡率之影響，(2)建構傳輸動態模式以推估關鍵流行病學參數，(3)驗證建構之傳輸動態模式以模擬族群動態，並進行敏感度分析，及(4)建構疾病控制模式以提供遏止病毒性疾病傳播之策略。
本研究以希爾模式描述雙核糖核酸病毒‐文蛤系統與吳郭魚湖泊病毒‐吳郭魚系統中病毒濃度與累積死亡率之關係。另一方面，為由累積死亡率數據取得關鍵疾病訊息，本論文以基本易感-感染-復原架構為基礎，建構雙核糖核酸病毒‐文蛤系統、吳郭魚湖泊病毒‐吳郭魚系統及石斑魚神經壞死病毒‐石斑魚幼苗系統之傳輸動態模式，以推估流行病學參數與模擬族群動態，接著，基於模擬結果進行累積死亡率對流行病學參數之敏感度分析。此外，為了解近年溫度對雙核糖核酸病毒傳輸之影響，本研究連結傳輸動態模式與回歸模式預測文蛤養殖池之每月傳輸動態。本研究進一步以可得之疾病相關知識為基礎，分別針對雙核糖核酸病毒‐文蛤系統、吳郭魚湖泊病毒‐吳郭魚系統及石斑魚神經壞死病毒‐石斑魚幼苗系統設計相應之溫度相關控制模式、流行病學控制模式及疫苗接種控制模式，以評估疾病控制措施之成效。
由劑量‐反應評估結果指出，希爾模式可良好描述文蛤與吳郭魚分別暴露於雙核糖核酸病毒與吳郭魚湖泊病毒之累積死亡率與病毒濃度之關係。另一方面，本研究提出之傳輸動態模式能夠描述所有系統之累積死亡率數據。參數推估結果指出，所有系統之基本再生數(R0)皆大於1。為達到控制再生數(RC)小於1與減低累積死亡率之目標，本研究針對各個系統藉由敏感度分析執行、族群動態模擬及疾病控制模式應用提出不同之控制策略。
於雙核糖核酸病毒‐文蛤系統中，R0,h於水溫低於26°C時皆為1.29，而當水溫升高至33°C時，R0,h急遽上升至15.11，顯示雙核糖核酸病毒於較高之水溫下會加速傳播。本研究亦發現疾病爆發季節之氣溫呈現逐年上升趨勢，可導致雙核糖核酸病毒爆發更具嚴重之程度。疾病控制模式模擬結果顯示，文蛤應於低溫下進行移除，以提升控制疾病之成效。於吳郭魚湖泊病毒‐吳郭魚系統中，在同居感染情境下，R0,t中位數估計值為2.59。敏感度分析結果指出，累積死亡率對傳輸率較敏感。疾病控制模擬結果指出，透過降低傳輸率、感染期及族群大小皆有助控制吳郭魚湖泊病毒疾病之傳播。於神經壞死病毒‐石斑魚幼苗系統中，未接種疫苗時之R0,g估計值為2.44。參數推估結果顯示，於浸泡疫苗20、60及120分鐘條件下，能成功產生免疫能力之石斑魚幼苗比例分別為0.42、0.94及0.52。結果亦指出，浸泡疫苗60與120分鐘會延遲發病時間，然浸泡疫苗20分鐘則提前發病時間。疾病控制模擬結果指出，為有效藉由疫苗接種控制神經壞死病毒之傳輸，至少需0.6比例之族群其浸泡疫苗時間為70至80分鐘。
本論文提出之傳輸動態模式，不僅用以推估流行病學參數，亦可模擬養殖物種暴露於病毒之族群動態，而建構之疾病控制模式更可用以評估措施介入對疾病傳輸之影響。深入了解病毒性疾病之流行有助於更有效之疾病控制策略之決定。本研究期望所提出之以流行病學為基礎之架構於未來能實際應用於養殖場，以改善或建立病毒性疾病之控制方案。

關鍵字

雙核糖核酸病毒 ； 吳郭魚湖泊病毒 ； 神經壞死病毒 ； 模擬 ； 傳輸動態 ； 基本再生數 ； 控制策略

並列摘要

In Taiwan, birnavirus (BV), tilapia lake virus (TiLV), and nervous necrosis virus (NNV) are waterborne pathogens that have been reported to cause mortality in farmed hard clams, tilapia, and grouper larvae, respectively. Outbreaks of mortality in these important farmed species could lead to great economic losses for aquaculture industry. Modeling the transmission dynamics of infectious diseases has been considered to provide insight into policies for managing diseases. Therefore, the purposes of this thesis were: (i) to assess the effects of virus concentration on cumulative mortality, (ii) to construct the transmission dynamic models to estimate key epidemiological parameters, (iii) to calibrate the constructed transmission dynamic models to simulate the population dynamics and to perform sensitivity analyses, and (iv) to develop the disease control models to provide implications for containment of the viral disease spreads.
This thesis applied the Hill model to describe the relationships between virus concentration and cumulative mortality in BV-hard clams and TiLV-tilapia systems. On the other hand, to derive key disease information appraised with published cumulative mortality data, the transmission dynamic models were developed based on a basic susceptible-infected-recovered (SIR) structure among BV-hard clams, TiLV-tilapia, and NNV-grouper larvae systems to estimate epidemiological parameters and to simulate the population dynamics. Then, the sensitivity analyses of cumulative mortality to epidemiological parameters were performed based on simulation results. Moreover, to understand the effects of temperature on BV transmission in recent years, the monthly transmission dynamics in real clam farms were predicted by linking the transmission dynamic and regression models. This thesis further designed the corresponding temperature-dependent, epidemiology-based, and vaccination-based control models for BV-hard clams, TiLV-tilapia, and NNV-grouper larvae systems, respectively, to assess the effectiveness of disease control measures based on the available knowledge among these diseases.
Results of the dose-response assessment indicated that the relationships between virus concentration and cumulative mortality of hard clams and tilapias exposed to BV and TiLV, respectively, were well described by the Hill model. On the other hand, the results of model calibrations demonstrated that the developed transmission dynamic models were capable of describing the cumulative mortality data in all systems. Results of parameter estimation indicated that basic reproduction numbers (R0s) were higher than unit in all systems. To attain the goals of control reproduction number (RC) < 1 and reduction of cumulative mortality, different control measures were provided for each system by performing sensitivity analysis, population dynamics simulation, and disease control model applications.
For BV-hard clams system, R0,h was 1.29 at water temperature ≤ 26.2°C, whereas R0,h dramatically increased to 15.11 at 33°C, implying that BV fast spread at higher water temperatures. This study also found that there were yearly increased trends of ambient temperature in outbreak seasons, potentially causing higher severity of BV outbreaks. Simulation results from the disease control models indicated that hard clams should be removed at lower temperatures to enhance the effectiveness of containing BV outbreaks. For TiLV-tilapia system, median R0,t estimate was 2.59 in the cohabitation scenario. Results of sensitivity analysis indicated that cumulative mortality was more sensitive to the transmission rate. Simulation results from the disease control models indicated that reduction of the transmission rate, infectious period, and host population size would be help contain the outbreaks of TiLV disease. For NNV-grouper larvae system, R0,g was estimated to be 2.44 without vaccination. Results of parameter estimation showed that the proportions of successful immunization were 0.42, 0.94, and 0.52 in the conditions of 20, 60, and 120 min bath immunization, respectively. Results also indicated that 60- and 120-min bath immunization would prolong the disease onset, whereas 20-min bath immunization would ahead of the time for disease onset. Simulation results from the disease control models indicated that, to effective control NNV transmission by vaccination, at least 0.6 of proportion of population needed to be immunized for 70 ‒ 80 min.
This thesis proposed the transmission dynamic models not only to estimate epidemiological parameters but also to simulate the population dynamics of aquaculture species exposed to viruses. The developed disease control models can be used to evaluate the effects of interventions on disease transmission. A better understanding of the epidemics of a viral disease is helpful for determining more effective disease control strategies. Hoping the epidemiology-based framework proposed in this thesis can be applied in real farms to improve or establish control programmes for viral diseases in the future.

並列關鍵字

Birnavirus ； Tilapia lake virus ； Nervous necrosis virus ； Modeling ； Transmission dynamics ； Basic reproduction number ； Control measure

參考文獻

Alaliyat S, Yndestad H. 2015. An agent-based model to simulate contagious disease dynamics in fish populations. International Journal of Simulation: Systems, Science & Technology 6: 8.1‒8.10.

Aldrin M, Huseby RB, Jansen PA. 2015. Space-time modelling of the spread of pancreas disease (PD) within and between Norwegian marine salmonid farms. Preventive Veterinary Medicine 121: 132‒141.

Anderson RM, May RM. 1991. Infectious diseases of humans: dynamics and control. Oxford, UK: Oxford University Press.

Arechavala-Lopez P, Sanchez-Jerez P, Bayle-Sempere JT, Uglem I, Mladineo I. 2013. Reared fish, farmed escapees and wild fish stocks―a triangle of pathogen transmission of concern to Mediterranean aquaculture management. Aquaculture Environment Interactions 3: 153‒161.

Arimoto M, Mori K, Nakai T, Muroga K, Furusawa I. 1993. Pathogenicity of the causative agent of viral nervous necrosis disease in striped jack, Pseudocaranx dentex (Bloch & Schneider). Journal of Fish Diseases 16: 461‒469.