Study design and setting
We conducted a double-blind randomized controlled trial at the HIV clinic of Mbarara Regional Referral Hospital (MRRH), a publicly funded teaching hospital in southwestern Uganda serving ten districts, with a population of over 5 million people. The study was conducted between December 2017 and August 2020.
Recruitment Of The Study Participants
With the introduction of test and treat protocols in Uganda, our study population consisted of PLWH and on HAART for at least one year. A trained Research assistant screened all HIV-positive participants on HAART attending the MRRH HIV Clinic to identify those that were; 1) 18 years or older, 2) had a CD4 count ≤ 350cells/µl, 3) were living within a 60 km radius from the clinic for ease of transport, 4) had normal haematological, liver and renal function tests, and 5) were of sound mind and able to sign the informed consent. We excluded individuals who had existing opportunistic infections, used other herbal/complementary medicines, or were pregnant. The research assistants obtained informed consent from all participants. A study nurse trained in human participant research obtained informed consent in the local language in a private area of the hospital. All consenting participants gave written informed consent to read and sign in the presence of their witnesses, and for those who could not write, a thumbprint was made on the consent form.
Randomization, Blinding, And Preparation Of Treatments For The Study Participants
Participants were first stratified according to baseline CD4 levels: 350–250, 249–150, and 149 and below before randomisation to the three groups. Computer-generated numbers, delivered by an independent biostatistician were used to assign eligible participants to the three study groups using a simple randomization method in blocks of 10 to either the control group with HAART only, HAART with A. annua group, or HAART with A. annua + M. oleifera group using a ratio of 1:1:1.
Enrollment of the participants to the study arms/groups was carried out by the research assistants led by the principal investigator according to the computer-generated numbers. The two treatments; A. annua and A. annua + M. oleifera were separately prepared, packaged, and clearly labeled by the study pharmacist from a pharmacy before dispatch to the study clinic. Participants and research assistants were blinded to the specific name of the treatment they were getting as these were labeled as A (for A. annua + M. oleifera) or B (for A. annua). Two research assistants were recruited and trained as observers and data collectors to increase accuracy and consistency in documenting the needed data. They were assisted by the HIV clinic staff and these research assistants were blinded to the hypotheses of this study to minimize observer bias. A senior clinician at the HIV clinic supervised the trial and administration of treatments for HAART participants attending the MRRH clinic. Both herbal treatments were procured from Jenna Herbals Ltd (Kampala, Uganda). Before the use of the medications, we performed quality tests for both intervention herbal materials at MUST Pharmaceutical Laboratory. To blind data collectors, treatments were packaged, labeled, and delivered in opaque parcels from the pharmacist at Devine Hospital in Mbarara City. Trained research assistants were responsible for distributing A. annua and M. oleifera powders. Healthcare providers, including nurses, pharmacists, and doctors, as well as phlebotomists, were blinded to the allocation of participants to the study groups.
Study Procedures
Research assistants received opaque parcels already fixed with study codes. These parcels contained either 4g of A. annua and 10g of M. oleifera or its placebo (starch powder) which were to be taken at least 8 hours apart from each of the daily ARVs dosing. All participants were encouraged to take note and report all side effects.
Preparation of study materials
A. annua and M. oleifera leaf powders were derived from A.annua and M. oleifera plants grown in Uganda. They were authenticated by a botanist and specimens were stored in a herbarium at MUST. The leaf powders of A. annua and M. oleifera were processed and packaged in packets of 4g and 10g respectively by a pharmacist in a setting of good manufacturing practices at Devine Hospital. The concentration of artemisinin in our plant was 0.5 to 0.6% of the dry weight sourced from farmers in Kabale District in South Western Uganda.
Moringa oleifera Lam. is a herb commonly consumed by HIV-infected people on antiretroviral therapy mainly as a nutritional supplement [10]. M. oleifera has also been reported to be used in the management of HIV infections in up to 80% of HIV patients in Africa. According to Asare and colleagues [12], the intake of M. Oleifera is very safe at levels ≤ 1000 mg/kg body weight. Monera and colleagues also found out that co-administration of M.oleifera Lam. leaf powder at the traditional dose did not alter the steady-state pharmacokinetics of nevirapine in HIV-infected adults [10].
Administration Of Treatments
Participants in the A. annua study arm self-administered 4g of A. annua leaf powder per day pre-packaged in a single dose packet that was emptied in the routinely consumed porridge of choice (cassava, millet, or posho porridges) or water, and taken every morning at 8 am for 12 months. These participants did not receive M. oleifera but instead, they were given 10 g of starch powder to take with any porridge of choice at 8 am, or at least 8 hours apart from the routine ARVs dosing.
Participants in the A. annua + M. oleifera study arm were given 10 g and 4g of M. oleifera and A. annua powders respectively to take in their drink of choice (cassava, millet, or posho porridges) or water every morning at 8 am for 12 months. The porridge was preferred for uniformity, convenience, and masking the bitter taste of A.annua when taken in plain water.
Each parcel supplied to the participant contained herbal treatments to last one month and the parcels had participants' study numbers and dosing instructions in addition to other relevant information on storage and safety. To ensure adherence to the treatments, every participant received a reminder SMS every morning between 7 and 8 am. Participants were requested to respond with the message 'Taken'' or a call prompt as a proxy to confirm adherence/ compliance to herbal medicine use. The study team worked closely with the HIV clinic to ensure that participants were reviewed and accessed routine HIV care as prescribed, including HAART. During the consenting process, we emphasized to all participants that these herbal medicines are not replacements for HAART in the treatment of HIV, and as such the study participants remained in the study only while enrolled on HAART as prescribed. This phrase was also added in the consent form, "Please note that these herbal medicines are not being given to you for treatment of HIV and as such you MUST keep taking your prescribed ARVs as usual and correctly".
Each study participant was followed up for 12 months. All participants had their files with study numbers for easy identification. The study participants were reviewed once a month by the study clinician, except during emergencies where the participant could call on the study clinician as the need arose. Physical examination of the participants was done by the study clinician and documented in addition to self-report by the study participants. Performance status and quality of life were also evaluated and documented. Case report forms were used to capture participants' vital outcome data in addition to laboratory report forms.
Study Measures
A blood sample for the measurement of CD4 count, viral load, complete blood count, and liver and renal function tests was drawn immediately after enrolment but before initiation of herbal treatments, and at six and twelve months following initiation of treatment. An additional blood sample was drawn at 1- and 2 weeks following treatment initiation to assess ARV plasma levels. Research assistants interviewed all participants at baseline, 6 and 12 months to assess the quality of life (QoL) using the WHOQOL-BREF questionnaire [13]. Research assistants also recorded all reported side effects using a standardized data collection form.
Study Outcomes
Our primary outcome measure was a change in absolute and relative CD4 counts after twelve months of follow-up. The CD4 count difference was calculated by comparing the mean CD4 count at the exit with the mean CD4 count at baseline comparing intervention groups with control. Our secondary outcomes included; a change in viral load, complete blood count (CBC), liver function tests (ALT & AST), kidney function tests (creatinine and urea), QoL, antiretroviral plasma levels and reported side effects. Change in viral load, CBC, LFTs, RFTs, QoL, and side effects were determined after twelve months of follow-up. Changes in efavirenz and nevirapine plasma levels were determined after one and two weeks after herbal treatment initiation. The laboratory technicians were blinded by coding the different samples with participants' unique identification numbers before delivery to the lab. The samples did not indicate any study arm allocated.
Quantitative questionnaire data on socio-demographics, prior herbal usage, depression, health [14] food insecurity[15], alcohol use [16], HIV stigma [17] and social support [18] and quality of life [13] were collected from study participants at enrolment.
Study Sample Size And Statistical Analysis
To compare the mean outcomes between each intervention group with the control group (using the t-test statistic) of a superiority trial, we considered our threshold probability for rejecting the null hypothesis α (two-tailed) to be 0.05. With power at 80% and a 1:1:1 ratio of subjects in each group. The mean CD4 among active participants attending the HIV clinic at MRRH, who have been on ART for more than 1 year and are 18 years or more, with a CD4 ≤ 350 after at least 1 year of initiation, was found to be at 160 (SD = 110). With that mean, we would require a sample size of 282 participants to detect a 30% CD4 count increase, at 80% power, 0.05 alpha, while assuming a 10% loss to follow-up. We would require a total of 94 participants in each of the 3 arms.
All data were cross-checked for completeness before entry. Data were collected using an online questionnaire developed in CommCare application by Dimagi. Data were downloaded in excel and exported into STATA Version 12 for statistical analysis. Data analysis was both by intention-to-treat (ITT) consisting of all participants randomized and thus supposed to be treated [19] and per protocol. Descriptive statistics were used to describe key characteristics of study participants. Different baseline variables were explored for normal distribution and comparison. Independent t-tests were used to compare the distributions and the means for each intervention group and the control. The means and mean differences were presented with their standard deviations. In the per-protocol analysis, the follow-up time for each participant was considered in the analysis, and the mean changes in CD4 count by follow-up time of measurements were compared. In both analyses, a difference was considered statistically different if p < 0.05. Secondary analysis compared the distribution of side effects such as nausea, dizziness, abnormal CBC, and kidney or liver function tests.
The WHOQOL-BREF questionnaire was used to collect data [13]. This questionnaire contains 26 questions, and 4 domains; physical, psychological, environmental, and social questions. Questions were presented on a linkert scale of 1–5, with 1 as poor and 5 as excellent. We transformed the raw linket scores of 1–5 into 1-100 using STATA V12 to enable us make uniform comparisons, especially for the domains that have unequal number of items. We compared the mean QoL scores across all domains using independent t-test.
Ethical approval
The study was approved by the MUST Research Ethics Committee with registration number 27/05-17 and Uganda National Council for Science and Technology (UNCST). The study was also registered with ClinicalTrials.gov NCT03366922. Written informed consent was obtained from each participant and all the procedures used were per the ethical standards of the responsible committee on human experimentation according to the Helsinki Declaration. Participants' identities were kept anonymous throughout the study process.
All treatments administered to participants have previously been documented to be safe when used for prophylaxis or as a supplement [8, 20, 21]. Monthly reviews by a senior clinician were provided in this study, samples and results of ART plasma levels, CD4 cell count, viral loads, liver function, and kidney function tests were provided to the HIV clinic physicians for monitoring. Although severe side effects were not expected, proper management and routine care protocols were adhered to following any such reports. All adverse events were brought to the attention of the Data and Safety Monitoring Committee and Institutional Review Committee in case of safety issues.
Data safety and monitoring
The data and safety monitoring committee was constituted and included senior medical scientists to ensure the safety of participating individuals. Three safety checks were done at 1 month, at 50%, and 75% of recruitment. The data safety management board members were routinely given laboratory and clinical data of all the participants to review and advise the MUST Ethics Committee and the UNCST in Kampala, Uganda where this study was also registered.