Risk Factors Associated With Dengue Fever Outbreak in Taiz Governorate, Yemen, 2018: Case-control Study

Background: Dengue Fever (DF) is a signicant health problem in Yemen especially in the coastal areas. On November 6, 2018, Taiz governorates surveillance ocer notied the Ministry of Public Health and Population on an increase in the number of suspected DF in Al Qahirah and Al Mudhaffar districts, Taiz governorate. On November 7, 2018, Field Epidemiology Training Program sent a team to perform an investigation. The aims were to conrm and describe the outbreak by person, place and time in Taiz governorate, and identify its risk factors. Methodology: Descriptive and case-control study (1:2 ratio) were conducted. WHO case denition was used to identify cases in Al Qahirah or Al Mudhaffar districts during August-November 2018. Control was selected from the same districts who did not suffer from DF. Predesigned questionnaire was used to collect data related to sociodemographic, behavioral and environmental characteristics. Bivariate and multivariate backward stepwise analyses were used. The adjusted odds ratios (aOR) and 95% condence intervals (95%CI) were calculated. A P value < 0.05 was considered as the cut point for statistically signicant. Epi info version 7.2 was used. Results: A total of 50 DF cases were found. Almost 52% were males and 76% were <30 years of age. The overall attack rate was 1/10,000 of the population. Case fatality rate was 4%. In multivariate analysis, not working (aOR = 26.6, 95% CI: 6.8–104.7), not using mosquito repellent (aOR = 13.9, 95% CI:1.4–136.8), wearing short sleeves/pants (aOR = 27.3, 95% CI: 4.8–156.8), poor sanitation (aOR = 5.4, 95% CI: 1.4– 20.3), presence of outdoor trees (aOR = 13.2, 95% CI: 2.8–63.0) and houses without window nets (aOR = 15.7, 95% CI: 3.9–63.4) were statistically signicant risk factors associated with DF outbreak. Eleven 11 (58%) of blood samples were positive for DF IgM. Conclusions: DF outbreak in Al Qahirah and Al Mudhaffar districts, Taiz governorate was conrmed. This study provides evidence-based information regarding the identied risk factors that contributed to the occurrence of this outbreak. Raising community awareness on the importance of personal protection measures and improving the sanitation services are strongly recommended.


Introduction
Dengue fever (DF) is a mosquito-borne disease with signi cant morbidity and mortality [1].
It's caused by the dengue virus that is transmitted by the Aedes aegypti mosquito [1,2]. Severe cases are causing serious illness and death among children. Some patients may develop dengue haemorrhagic fever and dengue shock syndrome [3]. There is no speci c treatment but the early detection and good clinical management can reduce the case fatality rate (CFR) to less than 1% [3].
Globally, DF incidence has increased 30-fold over the last 50 years, with increasing geographic expansion to new countries [3,4]. DF is endemic throughout the tropics and sub-tropics with 3.5-4.1 billion people (approximately 53% of the global population) is currently at risk of dengue virus infection [4]. Nearly Africa and the Americas [4]. In 2020, dengue affected several countries, with an increased number of cases in Bangladesh, Brazil, Ecuador, India, Indonesia, Maldives, Mauritania, Mayotte (Fr), Nepal, Singapore, Sri Lanka, Sudan, Thailand, Timor-Leste and Yemen [6].
Dengue is currently the most widely spread mosquito-borne disease in the Eastern Mediterranean Region.
Reports of dengue and severe dengue epidemics in the Region started in 1998 and have increased in frequency and spread ever since, with outbreaks occurring in Djibouti, Egypt, Oman, Pakistan, Saudi Arabia, Somalia, Sudan, and Yemen [7].
In Yemen, dengue is a signi cant health problem especially in the coastal areas. The rst case of dengue recorded was between 1870 and 1873. In 1954, DF occurred in Al Hodeidah governorate which affected 98% of the population. In 1984 travelers returned from Yemen were con rmed to have dengue antibodies.
Recently, the rst reported and con rmed dengue outbreak was in Shabwah governorate in 2002 and it distributed to others coastal governorates (Hadramout, Al Hodeidah, Taiz, Aden, Hajjah, Al Dhale'e, Lahj and Al Mahrah governorates) [8]. Moreover, the number of DF cases reported by electronic Disease Early Warning System (eDEWS) has escalated from 28,000 cases including 46 deaths in 2018 to more than 65,000 cases including 170 deaths in 2020 [9]. Similarly, in Taiz governorate the number of DF cases has increased from 3,000 cases with 17 deaths in 2018 to more than 5,000 cases w 4 deaths in 2020 [9].  [10]. It has a humid subtropical climate. It is rainy and hot in summer and moderate in winter. Al Qahirah and Al Mudhaffar were the most affected districts with more than half of the DF cases reported from Taiz governorate [9].

Sample size
Sample size of 150, with ratio of 1:2 (50 cases and 100 controls) were calculated by Epi Info program using a con dence level of 95%, power of 80%, an odds ratio of 2.9 and 20% of exposed control and 42% cases.

Case de nitions
Case de nition of the World Health Organization (WHO) was used. Suspected DF; any person has sudden fever more than 38°C for 2-10 days with two of these symptoms; severe headache, retro-orbital pain, myalgia, arthralgia, rash, hemorrhagic manifestations, leucopenia. Con rmed DF; a suspected case with laboratory con rmation.

Selection criteria of cases and controls
The line list of the DF cases reported by eDEWS was used for the selection of cases. A person who met the suspected or con rmed case de nition of WHO and lived in Al Qahirah or Al Mudhaffar districts, Taiz governorate, during August to November 2018 was selected as a case. Cases that had a disease history before August 2018 were excluded to avoid recall bias. While any person who lived in the same area of the case and who did not develop signs and symptoms of DF was included as a control. Two controls were selected from the same area of the case.

Data and laboratory sample collection
Active search from house to house was performed to identify the cases. A predesigned questioner was used to collect data that related to the following variables: sociodemographic (age, gender, education and occupation), date of onset and clinical symptoms. As well it included data related to behavioral characteristics, such as the status of water containers, using mosquito repellents and nets, and type of clothes (sleeves/pants). Environmental characteristics such as houses surrounded by garbage, sanitation (presence of stagnant water or sewage), presence of outdoor trees and windows of the house were collected. For con rmation of the outbreak, blood samples were collected randomly from 19 cases and sent to National Center for Public Health laboratories in Taiz governorate for laboratory con rmation.

Data processing and Analysis
Data were entered and analyzed by Epi info version 7.2. Bivariate and multivariate backward stepwise analyses were used to calculate the crude Odds Ratio (cOR) and the adjusted Odds Ratio (aOR) with 95% Con dence interval (CI). A P value < 0.05 was considered as the cut point for statistically signi cant.

Patient Characteristics
A total of 50 DF cases were found in both Al Qahirah and Al Mudhaffar districts, during August to November 2018. The rst case was in week 31, reaching 11 cases in week 40 (Fig. 1). Table 1 shows that 76% of cases were among age group less than 30 years and 52% were males. About 86% of cases were illiterate-basic education and 78% were not working. Figure 2 shows that all cases suffered from fever, followed by headache (94%), arthralgia (90%) and retro orbital pain (88%). The overall attack rate (AR) was 1/10,000 of the population and the CFR was 4% of DF cases. Out of 19 blood samples, 11 (58%) were positive for DF IgM. The remaining DF cases were epidemiologically linked by place with the cases. Risk factors associated with dengue fever

Discussion
Dengue is a signi cant health issue especially in the coastal areas in Yemen. This study showed that not working, not using mosquito repellent, wearing short sleeves/pants, poor sanitation, outdoor trees and houses without window nets were the potential risk factors for DF outbreak in Taiz governorate.
Our result indicates that the highest peak of DF cases occurs in week 40, this possibly attributed to accumulation of water in the watercourse during the rainy season, that creates vector breeding sites. This nding shows that males were slightly more affected than females. This might be due to higher exposure of male as a result of spending more time outdoors than females and the females have protective clothes traditionally. Our result is similar to results of outbreaks investigation conducted in some Yemeni governorates (Al Hodeidah, Taiz and Hadramout), and in Pakistan [1,11,12,13]. People with an age group < 30 years were more affected. This result might be due to people at this age being more active that makes them at risk of DF infections. These results are consistent with studies conducted in Al Hodeidah governorate, Vietnam, Hadramout governorate and India [1,2,12,14].
Our result indicated that there isn't a signi cant association between illiterate-basic educational level and getting DF infection. This result agrees with studies in India and Sudan [5,15] and disagrees with previous study in Al Hodeidah governorate [1]. Not working people have more than twenty times the odds of getting DF infection compared with working people. This nding is similar to a study in Sudan that attributed to the fact that unemployment may pose a lifestyle behavior that provides a suitable habitat for the breeding of the mosquito vector, thus increasing the risk of mosquito contact [15]. Conversely, this nding is dissimilar to studies in India and Malaysia [5,16].
Our ndings revealed that there isn't a signi cant association between not covering water containers and getting DF infection. Similar ndings were reported in studies from Ethiopia and India [4,17], but dissimilar to previous studies in Al Hodeidah governorate and Vietnam [1,2].
Our result is in agreement with the fact that mosquito repellent is a protective measure against DF infection and showed that people who do not use mosquito repellents are at higher risk of getting infection with DF. These ndings are consistent with studies in India, Kenya and Pakistan [17,18,19] and inconsistent with two studies in Ethiopia and others studies in Malaysia and China [3,4,16,20].
Our result is similar to studies in Sudan, Malaysia and China [15,16,20] that indicate there isn't a signi cant association between not using mosquito nets and the chance of getting DF infection. This nding might be due to the fact that mosquito nets are usually used at night, while Aedes aegypti mosquito is active during the day, especially in the early morning and late afternoon. However, this result is dissimilar to two studies in Ethiopia and one in Kenya [3,4,18]. Furthermore, our nding is similar to studies in Ethiopia [4] and Malaysia [16] reported that wearing long clothes is found to be protective and reduces the risk of DF infection, but dissimilar to study in Ethiopia [3].
The current investigation indicated that there is signi cant association between poor sanitation and DF infection. The people who live with poor sanitation have ve times the odds of getting DF infection compared with people who have good sanitation. Our result is consistent with study in India [5], but inconsistent with studies in China and Brazil that reported that the lack of statistical signi cance attributable to improving the residential living environment, reducing mosquito breeding and coverage of sanitation services [20,21].
Also there is a signi cant association between the presence of outdoor trees and getting DF infection. This might be due to the presence of outdoor trees that attract mosquitoes into human settlement. Conversely, a study in Vietnam reported that there isn't a signi cant association between vegetation around living space and DF infection [2].
Additionally, there is signi cant association between those who are living in houses without window nets and getting DF infection. The individuals who live in houses without window nets are at risk of getting DF infection fteen times the odds of getting DF infection compared with those who live in houses with window nets. This is possible to increase the exposure to mosquito biting. Our nding is consistent with studies in Malaysia and Pakistan [16,19], but inconsistent with studies in Ethiopia and Kenya [4,18].
As a result of limited nancial resources, laboratory tests were not performed to ensure selection of the eligible controls. The wide con dence intervals in some variables is possibly attributed to small sample size of this study. However, these limitations should be taken into account when a similar study is conducted in the future.

Conclusion
DF outbreak in Al Qahirah and Al Mudhaffar districts, Taiz governorate was con rmed. This study provides evidence-based information regarding the behavioral and environmental factors that potentially contributed to the occurrence of this outbreak, such as not working, not using mosquito repellents, wearing short sleeves/pants, poor sanitation, presence of outdoor trees and houses without window net.
This study recommends the implementation of prevention and control measures focused on modifying the risk factors to avoid the future outbreaks. Raising community awareness on the importance of personal protection measures, such as using mosquito repellents, wearing long sleeves/pants, and screening windows. As well improving the sanitation services and covering water containers are recommended.
Abbreviations AR: Attack rate; aOR: adjusted Odds Ratio; CFR: Case Fatality Rate; CI: Con dence Interval; cOR: crude Odds Ratio; DF: Dengue fever; eDEWS: electronic Disease Early Warning System; WHO: World Health Organization Declarations Figure 2 Distribution of dengue fever cases by signs and symptoms, Taiz governorate, August-November 2018e