Ensuring the Safety of Yellow Fever Vaccination in Travelers—The Experience at a Large U.S. Academic Medical Center in Colorado

Background: Yellow fever (YF) virus has the potential to cause fatal outcomes among at-risk individuals visiting endemic areas. Vaccinating travelers who are at risk is necessary to prevent virus-related life-threatening complications. We lack data on the clinical features of persons seeking YF vaccination. We aim to describe the characteristics of a cohort of persons receiving the YF vaccine before travel. Methods: A retrospective analysis of 964 travelers receiving the YF vaccine (Stamaril®) from Oct 2016 to Jul 2019 was performed at the University of Colorado Hospital, U.S. Percentages, means, and standard deviations were calculated. A multivariate logistic regression model was built to evaluate the association between receiving YF vaccination less than 10 days before departure and visiting friends and relatives (VFR). Results: The average age of the subjects was 39 ± 18 years with a range of nine months to 83 years. Persons who were 60 years of age and older represented 17%. Women consisted of 52%, and most of the travelers were Caucasians (64%). Travelers reported traveling to Africa (57%) or South America (40%). The primary destinations for travelers overall were Kenya (19%), Uganda (11%), and Tanzania (11%) in Africa; and Peru (14%) and Brazil (13%) in South America. The most common reasons for travel included leisure (44%), VFR (18%), and mission trips (10%). Comorbidities included a history of hematologic disorders (4%), HIV infection (2%), and diabetes mellitus (3%). The average duration between vaccine administration and travel was 43 days. Those VFR were two times more likely to receive the YF vaccination <10 days before departure. Conclusions: Identifying the type of travel, itinerary, and underlying medical conditions allows providers to administer the YF vaccine to travelers safely. There is a need to identify strategies to improve the timing of YF vaccination among VFR travelers.


Introduction
Yellow fever (YF) is a hemorrhagic disease caused by a flavivirus and transmitted by the Aedes aegypti mosquito, which is found in parts of tropical South America and sub-Saharan Africa. Since the early 1990s, the World Health Organization (WHO) estimates that there have been 200,000 cases of YF and 30,000 deaths due to the disease worldwide [1]. An analysis of African data sources in 2013 estimated that the burden of YF was 130,000 severe cases and 78,000 deaths [2].
Treatment for patients with YF is mainly supportive, as there is no specific antiviral therapy available. However, the YF vaccine is widely used for the prevention of YF in travelers, and for people living in endemic areas. There are 20-60 million doses of the vaccine distributed annually [3]. YF-VAX ® and Stamaril ® are live attenuated vaccines prepared by culturing the 17D-204 strain of the virus in chicken embryos, and their efficacy is based upon the development of neutralizing antibodies [4]. The vaccine has been used since the 1930s when it was first developed. WHO modified the length of validation by the vaccine in 2016 from 10 years to lifelong duration for most individuals [5]. In the U.S., the YF vaccine is primarily given as prophylaxis to military personnel and patients at risk due to travel to endemic areas. Sanofi held the manufacturing of YF-VAX ® in 2016 due to factory production issues, and the company made available an alternative vaccine, Stamaril ® , which is administered in the U.S. through an FDA-approved expanded access program. Standard clinical practice of screening patients during the pre-travel encounter is essential to prevent complications associated with YF vaccination. YF is a live attenuated vaccine with known serious adverse events, including vaccine-associated viscerotropic disease (YEL-AVD) and neurotropic disease (YEL-AND). Risk factors include uncontrolled HIV infection, 60 years and older patients [6], and other immunocompromised conditions. The subgroup of patients traveling to visit friends and relatives (VFR) can be a particularly vulnerable population at higher risk for some preventable infections, such as malaria, due to loss of previous immunity, and a lesser likelihood to seek pre-travel advice or take prophylaxis, while going back to their home countries [7,8]. There are limited studies in the U.S. that describe the patient population receiving the YF vaccine. This study aimed to identify important descriptors of patients getting the YF vaccine before travel, including patients traveling to visit friends and relatives. This information can aid public health agencies to enhance strategies to increase immunizations of patients at risk and avoid complications.

Ethics Statement
The present investigation complies with the Health Insurance Portability and Accountability Act (HIPAA) according to the Colorado Multiple Institutional Review Board (COMIRB) at the University of Colorado Denver. Patients received the YF vaccine through an approved Sanofi Pasteur Inc. Protocol Number STA00011, Expanded Access IND Program to Provide Stamaril ® YF Vaccine (17D-204 strain) to Persons in the United States (Quorum Review File #32032). Analysis of clinical data has been performed under an approved protocol (COMIRB Protocol 17-1032).

Patients and Data Collection
Data from patients receiving the Stamaril ® vaccine at the University of Colorado Hospital clinic from 31 October 2016 to 7 July 2019, were submitted for data extraction. Electronic medical records (EPIC) were automatically interrogated for the cohort of travelers through a software supported by Health Data Compass Data Warehouse project (healthdatacompass.org). Study data were collected and managed using REDCap electronic data capture tools hosted at the University of Colorado Denver. The following variables were automatically collected: gender, age, race, state of residency, date of YF vaccine, and the following comorbidities based on International Classification of Diseases (ICD) codes: diabetes mellitus, neoplasms, HIV infection, history of hematology-immune disorders, and pregnancy (see Appendix A for full ICD-9, and ICD-10 code definitions). VFR was defined as a form of travel wherein the purpose of the trip or the type of accommodation was visiting friends and/or relatives as consigned in the patient's history. The following variables were manually collected through chart review: verification of vaccine date if it was unavailable per the automatic search, pregnancy at the time of vaccine administration, the reason for travel, the continent of travel, and destination countries. Some travelers had missing information on key variables (Table 1).

Statistical Analysis
The means and standard deviations for continuous variables were calculated. For categorical variables, frequencies and percentages were calculated. Patient characteristics were compared between those reporting to visit friends and relatives versus other reasons of travel using chi-squared, students t-tests, or Fisher exact tests. A multivariate logistic regression model was built to evaluate the association between those receiving YF vaccination less than 10 days before departure and those VFR, after controlling for confounders of age, sex, race, and the continent of travel. All analyses were conducted in SAS 9.4. We selected the 10 days based on WHO recommendation of the optimal time of immunization against YF before traveling to endemic areas [9].
Data access: The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit the manuscript for publication. The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.
Since the vaccine is licensed only to infants older than nine months of age, we only had four infants less than one year of age in our cohort, representing 0.4%. We did not have any reported side effects or complications in this group of travelers.

Travelers Visiting Friends and Relatives
Travelers visiting their friends and relatives were more predominantly men (56%), younger, and identified as African Americans for their primary ethnicity. Women in this subgroup were more likely to be pregnant and more likely to be HIV positive, but less likely to have a history of cancer. They were more likely to visit Africa as opposed to South America, and they had less time between vaccine administration and travel departure. Common destinations were Kenya, Ethiopia, and less commonly Brazil ( Figure 1C). VFR travelers were more likely to receive the vaccine less than 10 days before departure compared to other reasons for travel (18% vs. 10%, p = 0.005). Those VFR were 2.2 times (OR 2.2 (1.3-3.7), p = 0.003) more likely to receive the YF vaccination <10 days before departure, after controlling for confounders of age, sex, race, and destination of Africa vs. South America.

Discussion
We describe the clinical characteristics of a cohort of travelers seeking YF vaccination at a U.S medical center. Some of those travelers presented themselves as family groups. Most were young adults, but the age varied widely from infants to seniors. In Colorado, most travelers were Caucasian, had a few comorbidities, and traveled to Africa most often. Travelers sought YF vaccination on an Travelers who were 60 years of age and older represented 17% of the total study population. They were 53% women, predominantly white (79%), and mostly traveling for leisure (63%). They visited Africa (52%) and South America (47%) more often and had plenty of time to receive the vaccine before departure with a mean of 50 days. Kenya (23%), Tanzania (22%), Brazil (24%), and Peru (16%) were popular destinations among these travelers. Only 7% received the vaccine less than 10 days from departure. Sixty years or older travelers had higher rates history of heme-immune conditions (8%), diabetes mellitus (8%), and neoplasms (22%).
Since the vaccine is licensed only to infants older than nine months of age, we only had four infants less than one year of age in our cohort, representing 0.4%. We did not have any reported side effects or complications in this group of travelers.

Travelers Visiting Friends and Relatives
Travelers visiting their friends and relatives were more predominantly men (56%), younger, and identified as African Americans for their primary ethnicity. Women in this subgroup were more likely to be pregnant and more likely to be HIV positive, but less likely to have a history of cancer. They were more likely to visit Africa as opposed to South America, and they had less time between vaccine administration and travel departure. Common destinations were Kenya, Ethiopia, and less commonly Brazil ( Figure 1C). VFR travelers were more likely to receive the vaccine less than 10 days before departure compared to other reasons for travel (18% vs. 10%, p = 0.005). Those VFR were 2.2 times (OR 2.2 (1.3-3.7), p = 0.003) more likely to receive the YF vaccination <10 days before departure, after controlling for confounders of age, sex, race, and destination of Africa vs. South America.

Discussion
We describe the clinical characteristics of a cohort of travelers seeking YF vaccination at a U.S medical center. Some of those travelers presented themselves as family groups. Most were young adults, but the age varied widely from infants to seniors. In Colorado, most travelers were Caucasian, had a few comorbidities, and traveled to Africa most often. Travelers sought YF vaccination on an average of about a month and a half before their departure date. The subgroup of VFR travelers was younger, of African American descent, traveling to the African continent, more often pregnant, and had a higher likelihood of having an HIV infection.
We also showed that VFR travelers were more likely to receive the YF vaccine at a suboptimal time before travel. WHO recommends immunization against YF at least 10 days before travel to endemic areas [9]. Studies in travel clinics have shown inadequate timing of the YF vaccine before travel in children [10]. Those VFR may have a harder time making travel clinic appointments and may present just before travel and may be less prepared to take appropriate preventative measures [11].
Receiving the vaccine less than 10 days prior can also have implications for possible denial of entry or increased paperwork at the country of destination. Those VFR may also not recognize the specific country requirement of YF vaccination until just shortly before departure. VFR travelers carry a higher risk of acquired travel-related illnesses such as Hepatitis A, typhoid, malaria, soil-transmitted helminths, and influenza [12]. Specific risk factors associated with the increased threat of illness among VFR include longer stays, decreased pre-travel health plans, sick contacts while abroad, and poorer sanitary conditions during their stay. Public health interventions can aim to increase rates and enhance the optimal timing of YF vaccination among those VFR.
We have shown a large cohort of travelers who safely received the YF vaccine before travel. Since the rate of adverse events with the Yellow vaccine is of about three events per 100,000 doses [13], with our relatively small representative sample we cannot extrapolate a different safety profile. Nevertheless, through standardized travel advice encounters, we safely delivered the vaccine to more than 150 travelers aged 60 years or older, travelers with controlled HIV, and history of cancer, pregnancy, or heme-immune disorders not listed as absolute contraindications. Although travel clinic providers screened travelers for contraindications to receive the YF vaccine, the more comprehensive interrogation of our electronic medical records found a small rate of non-prohibitive relative contraindications in some travelers. We still encourage the avoidance of vaccination in travelers with relative contraindications if the risk of YF acquisition during travel is deemed low, but our findings suggest vaccination is safe in this relatively small cohort among travelers older than 60 years of age. Our cohort delivered some safety evidence of YF vaccine administration among travelers with those listed conditions. This data can reassure clinicians and travelers with a history of those conditions to make pre-travel decisions where the YF vaccine administration is mandatory.
Previous reports in Nigeria have documented the safe administration of the YF vaccine during pregnancy [14]. Administration of the vaccine to HIV-infected individuals with CD4 counts greater than 500 cells/mm 3 is safe [15]. YF vaccine has been also administered safely in immunocompromised patients after the withdrawal of their immunosuppressive therapy [16,17].
Our population of travelers visiting family and friends reflects the diversity of African immigrants in Aurora, Colorado. A significant number of travelers seeking pre-travel advice will benefit from a continued comprehensive pre-travel screening of immunocompromised conditions. Providing the YF vaccine remains a critical public health strategy to decrease transmission and disease. Although overall coverage for the YF vaccine has increased in endemic countries [18], travelers are an important target for this preventive strategy as well. Data from the recent Brazilian outbreak found a case fatality of up to 40% among unimmunized travelers [19].
Models incorporating clinical features have been important to showcase disease burden and to enhance vaccination strategies [20]. The study of high-risk populations can inform the best vaccine policies [21]. VFR travelers are considered a high-risk population. We recommend public health policies to enhance the inclusion of vulnerable populations such as people VFR for YF prophylaxis. We should explore policies such as outreach community messages on the importance of pre-travel health care among foreign-born populations in the US. Additional considerations include community health workers reaching VFR communities to explain the importance of pre-travel vaccination and assessing individual risks. There are a few limitations to this study. The retrospective selection of data limits the reliability and number of variables analyzed. Misdiagnosis or irrelevant past medical history could have been selected through the automatic ICD screening of the previous diagnosis. However, selection bias was decreased through the automatic collection of some key risk factor variables. Missing data occurred in some medical records as well. We did not have data on foreign-born status among the VFR travelers, which can also account for different clinical characteristics or outcomes.
YF vaccine remains a priority for decreasing disease burden. Coloradans seeking the vaccine represent the current demographics of our community. Despite the history of uncommon well-known comorbidities, YF vaccination was effective and safe. YF disease remains a potentially lethal complication during travel. The current outbreak in Nigeria and the 2018 outbreak in Brazil, both with high case fatality ratios, highlight that prevention strategies are a priority. Efforts should be enhanced to continue YF disease prevention strategies in travel clinics throughout the United States.