Early Phase Clinical Trials with Human Immunodeficiency Virus-1 and Malaria Vectored Vaccines in The Gambia: Frontline Challenges in Study Design and Implementation

Recruitment of key study staff.

Most early phase clinical trials require testing of the vaccines or investigational products in a small number of study participants ranging from 20 to 50 individuals. The small sample size is usually selected for assessment of safety, tolerability, and immunogenicity of the investigational vaccines5; this sample size balances the need to avoid exposing a large number of study participants to unknown risks with the need for data from an adequate sample. Although there is no doubt about the significance of safety and immunogenicity data at this vital stage of vaccine development, there is usually a tendency to think that these trials do not require a full complement of staff because the sample size is relatively small compared with efficacy trials. The general assumption is that the size of a clinical trial should dictate the number of staff. Our experience showed that the magnitude of workload in early phase trials is not proportional to sample size. This is because the same degree and quality of preparation, planning, and implementation strategies required for clinical trials with large sample size are also needed to successfully conduct early phase trials. We suggest that an appropriate number of clinical, field, and laboratory staff is used to execute early phase trials.

Furthermore, participation in early phase clinical trials represents excellent training opportunities for these cadres of staff. However, because of a low yield of scientific publications from these trials, it is important to identify additional research questions within the overall trial, which the team could address in collaboration with the main trial's Principal Investigator. This could maintain satisfaction within the team and prepare the staff for higher responsibilities in future trials.

Host community engagement.

Unlike other African communities,18,19 a Community Advisory Board (CAB) does not exist in The Gambia, but the roles of participants' advocates are informally and efficiently played by the village heads called “Alkalos.” Before commencement of the clinical trials, the host communities were sensitized by first introducing the studies to the village heads. This was accompanied by presentation of traditional kola nuts, which signified the symbolic introduction of the studies and to solicit support from the host community for smooth conduct of the studies. The study rationale and its justifications were subsequently explained to the “Alkalo” and his chiefs. Concerns like the issues of blood collections and required laboratory assays were explained to the community leaders in lay language. Subsequently, the “Alkalo” passed the information about the studies to potential participants through household heads and religious leaders. Apart from these initial visits, community sensitization meetings in the form of annual Open Days were organized to provide feedback on findings of previous studies and the introduction of the new trials. This provided ample opportunities for researchers and members of host communities to engage in mutual partnership that could contribute to success in the vaccine trials.

Recruitment of study participants.

The excellent working relationship between the Medical Research Council (MRC) and host communities makes it easier to recruit study participants. The participants are usually keen to join MRC studies because of perceived medical benefits.13,20 However, this phenomenon changed during the HIV vaccine trial and it was difficult to recruit study mothers because of the belief that the vaccine might cause HIV infection in the infants. This was further complicated by the likelihood of a “false HIV antibody positivity” elicited by the vaccine. As the majority of intending mothers were illiterate, it was practically challenging to explain this technical adverse event to them. We communicated this to the study mothers by using an illustration of a “real car” and a “toy car.” Just as the two cars look alike, a real car's mobility is powered by an engine, whereas a toy car, which does not have an automobile engine has limited mobility; so also is the laboratory-synthesized Modified Vaccinia Ankara expressing HIVA immunogen (MVA.HIVA) vaccine, which could elicit the HIV antibody that is not truly positive. Although no study infants had false positive antibody results, communicating this technical information was a significant challenge.

Similar challenges were faced in translating the informed consent documents to local languages as the languages are only spoken and do not have standardized writing formats. After several attempts at forward and back translations of informed consent documents, very few literate participants could read and understand it. This persuaded our local ethics committee to recommend the use of trained field assistants in interpreting the contents of the informed consent documents to the potential participants.13 To ensure an objective assessment of understanding of participants before enrollment into the study, a study quiz covering the domains of “study purpose,” “study procedures,” “randomization,” “placebo,” “adverse events,” “confidentiality,” “compensation,” and “rights of withdrawal” was developed and administered to the participants. Satisfactory performance assessed by success in giving seven correct answers out of eight questions was used as an indicator of “good understanding” of the study. Participants who failed the first attempt received further education on the areas of deficiency and were allowed one more attempt at the quiz. This effort increased the participation rate and 100% retention of participants was recorded throughout the study period. This study quiz has been further developed into a digitized format. Finally, a multimedia informed consent tool has been developed in Gambian local languages to deliver study information to low literacy participants.

Cultural practices.

Understanding and responding appropriately to the cultural beliefs and practices of the host communities play significant roles in the successful conduct of clinical trials.16,21,22 In many communities in The Gambia, newborns and the mothers should not be seen outside their homes until 8 days after birth. This caused delay in taking BCG vaccination for the neonates and may affect clinical trials that require crucial clinic visits within the first week of life.23 Similarly, vaccination and follow-up visits during the malaria vaccine trial in adult men fell within the month of fasting in which all adult Muslims are expected to abstain from eating and drinking from dawn to dusk. All participants in the trial were Muslims and of fasting age. We communicated the requirements of the clinical trial to vaccinate the study participants with prime and boost vaccines and collect 30 mL of blood during each study visit during the informed consent process. We also informed them that these procedures might take place during the fasting period. All participants agreed with the arrangement but later refused to be bled or vaccinated during the fasting period. The clinic visits were subsequently rescheduled to evenings to allow vaccination and bleeding of study participants after the fast had ended for the day. This development underscores the need for researchers to have back-up plans for executing every stage of clinical trials.

Study design.

The clinical trial was based on the hypothesis that protective T cell-mediated responses could be elicited by vaccines in early life, and BCG as a vaccine vector might be an appropriate prime. Thus, all potential study infants were recruited at birth or within the first week of life to ensure the study team vaccinated them with BCG. As BCG vaccination was central to the success of the experimental vaccine, the presence of a BCG scar was actively sought on the left deltoid area of the infants, which was also the site of vaccination with the experimental HIV vaccine.

Although BCG has shown great promise as an HIV vaccine vector, some antigens in MVA.HIVA, such as HIV env, are poorly expressed in BCG and this may result in sub-optimal immune responses to the HIV antigen.24 More important, infants have immature innate and adaptive immune responses that result in a diminished capacity to activate other immune cells which plays a major role in induction of the downstream adaptive immune responses.25 In addition, infants' T cells show diminished capacity to express Th-1 effector function caused by hyper-methylation of the proximal promoter of the interferon-gamma (IFN-γ) gene, which results in restricted IFN-γ responses to most stimulants.26,27

In view of the aforementioned, we suggest that vaccine developers targeting infants in infectious disease-endemic countries should address the issues of immature infant immune system in early stages of the vaccine development. For example, the use of an improved strategy on BCG vector such as codon optimization and combination with viral vector boost24,28 have proved more useful in HIV vaccine development. Similarly, induction of broadly neutralizing antibodies remains critical toward successful HIV-1 vaccine development.29,30

HIV antibody testing in study mothers and infants.

One of the major inclusion criteria for the study was HIV antibody testing of study mothers at 6 weeks post-partum. Of the 65 mothers approached, 61 (93.8%) agreed to undergo voluntary counseling and HIV antibody testing by enzyme-linked immunosorbent assay (ELISA). The mothers attended immunization clinics with their infants until 19 weeks when the infants were enrolled in the study. Although the prevalence of HIV-1 was low in The Gambia,31 there are a number of traditional practices that could increase risk of HIV transmission during infancy. These include breast-feeding of infants by female relations whose HIV status are unknown and use of unsterilized instruments for male and female circumcisions.32 We recommend HIV antibody retesting of the study mothers within 6 weeks after the first test, according to international guidelines,33 to avoid enrolling infants whose mothers were in a serological “window” period.

The possibility of a false “positive” HIV antibody test was documented as an adverse event related to vaccination with MVA.HIVA caused by cross-reactivity with the “gag” protein of the HIVA immunogen.34 To detect this, all study infants had ELISA tests at 28 weeks of age and showed negative antibody results. Maternal antibodies are known to persist beyond 28 weeks, hence the World Health Organization (WHO) recommendation of HIV antibody testing only in children 18 months of age and older. To conform to this guideline, we advise similar vaccine trials to use WHO recommended technology like the PCR test for this age group.35

Absence of local hematology and biochemical references.

Hematological and biochemical references are essential indices used in establishing safety parameters, especially in early phase intervention trials. The accuracy and appropriateness of these laboratory parameters are key to understanding the results from such trials. At the time of conduct of this trial, there were no local reference values for hematology and biochemical parameters for the pediatric age group. Previous studies conducted within the institution made use of values obtained from a pediatric population from an East African country. In view of differences in environmental, genetic, and nutritional factors that may influence these laboratory indices,36,37 we developed “in-house” laboratory references from previous infant studies, which compared favorably with similar references from neighboring West African countries. This challenge stimulated a comprehensive study, which is currently underway to determine the laboratory reference values for all age groups in different geographic regions of The Gambia.

Administration of supplementary EPI vaccines.

In line with international recommendations to facilitate eradication of childhood vaccine preventable diseases, the Gambian Health Ministry organizes periodic house-to-house campaigns to immunize infants < 5 years of age, irrespective of whether they have received the Expanded Program on Immunization (EPI) vaccines before. Among the study infants, these additional immunizations could potentially act as confounding variables because of variations in age at vaccination and batches of vaccines administered. More important, these variations may distort findings about interference of the EPI and study vaccines. All Gambian children presenting for EPI vaccinations are issued an infant welfare card, which the mothers keep for the infants' immunization record. We designed and placed an informative sticker on the health card indicating that the child was enrolled in an important vaccine trial, and should only receive EPI vaccines from the study team at the trial center. This greatly minimized the number of study infants who received the supplementary immunizations. The stickers were removed when the child exited the study, i.e., end of the study period, not eligible for randomization or dropped out. The study team ensured all infants received all recommended EPI vaccines according to Gambian immunization schedule. To further minimize potential disruptions in clinical trial schedules, we suggest that the clinical trial team should work with the local EPI team to plan and identify favorable dates for the supplementary immunizations.

Use of analgesics for post-vaccination fever.

Fever is a common adverse event after infant vaccinations. Studies have shown the efficacy of acetaminophen prophylaxis after whole cell pertussis vaccination and also in treating febrile episodes after routine immunizations.38,39 Because of a study report that prophylactic or therapeutic use of acetaminophen may reduce immunogenicity to vaccines,40 we encouraged mothers not to give the drug except when infants became irritable after vaccination. Study infants were observed for immediate local and systemic adverse events at the trial clinic for 1 hour after administration of the study vaccine. This was followed by post-vaccination home visits by trained field staff that collected further adverse events on purpose-designed diary cards. Although most mothers reported fever in vaccinated and non-vaccinated infants within 6–24 hours after vaccination, the temperature records taken by trained field staff did not support the episodes of reported fever. Furthermore, the infants did not develop any clinical symptoms and signs necessitating indiscriminate use of analgesics. This observation lends credence to rational use of acetaminophen after infant vaccinations.

Infants' weight faltering during follow-up.

Weight faltering, a common growth pattern in developing countries, is characterized by undernutrition when complementary feedings are introduced to growing infants.41 Almost equal numbers of study infants in the vaccine and control groups had weight faltering after study intervention. Although these were documented as adverse events, they were considered not related to the study vaccine. The mothers were educated to use locally available food items as complementary diets. However, there was no appreciable improvement in weight of the study infants until locally produced food supplements were procured by the study team and distributed to the mothers of affected infants. Sustainability of such intervention was a major financial challenge as it was not factored into the clinical trial budgets. We suggest that clinical trials conducted on children in low-income countries should make appropriate budgetary provisions for malnutrition-related events that commonly occur in this age group.

Malaria antigen testing.

A report of attenuated immunogenicity in malaria vaccine volunteers from The Gambia and Kenya42 was considered to be caused by previous exposure to malaria. This was expected as the volunteers lived in malaria-endemic countries with high transmission rates. After this 2007 report, malaria transmission declined drastically in The Gambia resulting in low malaria parasitemia.43,44 Despite this development and to further ensure that immunogenicity to study vaccines is not distorted by level of parasitemia, we included a negative malaria antigen test as one of the eligibility criteria. The First Response histidine-rich protein-2 rapid diagnostic test was used to screen potential participants for presence of malaria antigens. The diagnostic accuracy of this test-kit was comparable with malaria microscopy.45,46 Of 140 children and infants screened over a period of 2 years, only 4 (2.9%) tested positive for malaria antigens. These children were treated with a full course of antimalarial drugs and excluded from participating in the trial. We considered enrolling them in the trial after treatment but decided against this because of evidence that malaria antigens could persist after treatment.46 Nevertheless, none of the children/infants who had negative antigen tests developed clinical malaria throughout the study period, proving the value of testing before enrollment.

Narrow recruitment age.

After showing satisfactory safety and immunogenicity profiles in adults and children trials; further studies were conducted in 5–12-month-old infants, closely followed by 10-week-old infants. To select optimal dosage of ChAd63 ME-TRAP for this age group, the study was further divided into two stages with a low dose ChAd63 group paired with a no-treatment control group, after which infants in high dose ChAd63 group were recruited and screened for eligibility. Because the infants had to receive the prime vaccine at exactly 10 weeks of age, recruitment, screening, and randomization of eligible infants had to take place about 2 weeks before this period. This implied that 8-week-old infants presenting at the trial center for routine EPI vaccines were the target population for recruitment. Unfortunately, most mothers did not bring their infants to the clinic until they were about 10–12 weeks of age making pretrial procedures logistically challenging. We therefore recommend that enrollment age of participants in similar clinical trials should be more flexible to accommodate predominant health-seeking behavior and practices of potential participants, unless there is a scientific rationale to be very restrictive.

Required volume of blood.

The adult trial had a total of eight study visits and the volume of blood required from each volunteer at each visit was 30 mL, whereas during the pediatric trials 5 mL of blood was collected at each visit. Some participants had concerns and pre-conceived ideas that MRC sold the blood collected from study participants abroad and used the proceeds to procure vehicles. We made use of purpose-designed visual aids to illustrate the relatively small volume of blood collected from each participant when compared with the total body volume of individual participants. The participants were invited to our laboratories where we demonstrated how the collected blood samples were processed. This practice allayed fear of uncertainties about volume of blood collected and improved participants' retention and adherence to study protocols.

Use of comparator vaccine as control.

In the pediatric trial, human diploid cell rabies vaccine was used as a comparator with the malaria vaccines. The reconstituted anti-rabies vaccine was a pink liquid, whereas the study vaccines were colorless. The comparator and study vaccines were prepared separately using identical 27G 1 mL syringes that were covered with an opaque label to maintain blinding of study mothers and field staff who were involved in assessment of primary safety outcomes. In the control group, two doses of anti-rabies vaccines were designed to be given to the randomized participants. This raised an important ethical issue because of the concern that two doses of anti-rabies vaccines may confer limited protection against rabies. However, it was established that the two doses could confer optimal protection provided they are given 28 days apart, which suited the study design. Furthermore, because of a relatively high rabies risk in The Gambia and the prohibitive cost of general prophylactic vaccination, it was decided that participants randomized to receive malaria vaccines should also receive anti-rabies vaccine at the end of the study. Although this decision led to an increase in research cost, it was ethically acceptable as a form of compensation to participants involved in early phase trials that in most instances do not confer any immediate benefit to the participants. We suggest the inclusion of similar health benefits in future vaccine trials conducted in developing countries.

Dry ice supply.

Study vaccines were required to be transported at −20°C or below to the trial site, which is located about 20 km away from the storage facilities. Unfortunately, dry ice is in limited supply in most developing countries. For instance, only one company produces dry ice in commercial quantity in The Gambia and this supply is often unreliable. Attempts to produce dry ice within our host institution were not sustainable. On many occasions, we had to import dry ice from Europe. We suggest development of a sustainable “in-house” dry ice production in research institutions in developing countries. We also recommend development of cold-chain free vaccines for resource-constrained settings.