This was a parallel group single centre randomised controlled pilot trial with ratio of 1:1 conducted at the Plastic surgical unit of Kirruddu National Referral hospital in Kampala, Uganda from the months of March to July 2021. The trial was approved by the “The AIDS Support Organization (TASO) Research Ethics Committee (TASOREC/060/19-UG-REC-009) and the Kirruddu National Referral Hospital Research Ethics Committee and was also registered in the ClinicalTrials.gov under the registration number NCT04553159. The trial reporting followed the CONSORT Extended checklist for pilot clinical trials and TIDier guidelines.
Participants and eligibility criteria:
Participants who met the eligibility criteria were consecutively selected out of patients seeking care for keloid treatment at the clinic.
Patients were eligible if they were aged between 18 to 65 years and had a single keloid of equal or less than 4cm3 as these are treated non-surgically and hence qualified for non-surgical intervention. Patients who were underweight (BMI less than 18.5) were excluded as they had insufficient fat pads to enable liposuction. Patients who had three months prior received intralesional therapy or cutaneous radiotherapy were excluded as the effect of previous therapy couldn’t be ruled out. Patients with active systemic illnesses demonstrated by symptoms of fever, cough, gastroenteritis, or laboratory diagnosis would be excluded. Similarly, patients with bleeding disorders or anti-coagulation therapy were to be excluded.
Included participants were provided with complete oral and written information ahead of their consent. Informed consent was then obtained following a 48 hours period in which participants were allowed to consult relatives or other stakeholders.
Once participants were included in the study, baseline characteristics were obtained following which they were randomized and allocated into either treatment arms as described below.
The Autologous Adipose-Derived Stromal Vascular Fraction (SVF) group
Participants in this arm received a single dose intra-lesional infiltration of the autologous adipose-derived stromal vascular fraction. The dosage depended on the total number of viable SVF cells in the cellular suspension and was expressed as total viable cells per ml of suspension. Final cell suspension volume was constituted in a 1:1 ratio of the keloid volume with 1ml of cell suspension per cubic centimetre of keloid tissue.
In order to obtain the stromal vascular fraction, the steps described below were followed.
Harvesting adipose tissue:
Tumescence liposuction was performed aseptically on the outer thigh following the infiltration of 300ml of tumescent solution (constituted as 20ml of 2% Lignocaine, 1 ml of 1:1000 epinephrine, 12.5 ml of 8.4% Sodium Bicarbonate in 1 litre of normal saline). Through a 3-4mm skin incision, liposuction into a 10ml Leuer lock syringe attached to a 3 mm Coleman liposuction cannula was performed with 100 to 150ml of lipoaspirate was collected into 50ml sterile falcon centrifuge tubes. The liposuction cannula entry incision was closed with a 6/0 Monocryl absorbable suture and a pressure bandage placed.
Standard operation theatre aseptic protocols were followed and the procedures were conducted in a dedicated plastic surgery operation theatre by qualified plastic surgeons.
Extraction of stromal vascular fraction:
The harvested lipo-aspirate see Fig. 2a, was processed aseptically from a designated sterile unit in the operating theatre. The Lipoaspirate was washed using 1X Dulbecco’s Phosphate Buffered Saline-PBS (Lonza, Walkersville, MD USA) and then subsequently enzymatically digested using 0.075% Type 1A Collagenase (MERCK Millipore, USA). Enzyme stop media comprising of 10% Foetal Bovine Serum-FBS (SIGMA St. Louis, MO USA) in Dulbecco’s Modified Eagle Medium (DMEM)-High glucose (SIGMA St. Louis, MO USA) was used to neutralize the enzymatic process (See Fig. 2b). The Stromal vascular fraction pellet (Fig. 2c) was was subsequently incubated in 10ml of Red Cell Lysis Buffer (SIGMA St. Louis, MO, USA) at room temperature for 10 minutes and later washed in 1XPBS. Through a 100µm nylon cell strainer (BD Falcon, NJ, USA) the mixture was filtered to remove any unwanted tissue debris. Centrifugation at 1200g resulted in the stromal vascular fraction (SVF) pellet which was then resuspended into 1.5mls of Ringer’s Lactate solution.
To determine the cell count and viability, to 10µl of the cell suspension was added an equal volume of 0.4% Trypan blue(SIGMA St. Louis, MO USA) and then mounted into a Neubauer counting chamber as per Strober’s guideline. Cells were counted at X40 Objective magnification as shown in Fig. 3.
Total nucleated cells per 10µl were used to determine the final cell dosing while viability was determined using the Trypan blue exclusion test. The total number of viable cells on the Neubauer counting chamber were used to calculate the total number of cells in the 1.5ml suspension factoring in the two times dilution factor during the trypan blue staining.
The stromal vascular fraction dosing was constituted by diluting the SVF suspension to a 1:1 ratio of SVF volume to keloid volume. Hence for a 4cm3 keloid, the 1.5ml of original SVF suspension was constituted to make 4ml by addition of 1XPhosphate Buffered Saline (PBS). The final infiltration cellular dosing calculation per millilitre was computed and medication placed into hypodermic syringes ready for infiltration.
The Triamcinolone Acetonide (TAC) group:
For this arm, one ampoule of Triamcinolone acetonide (TAC) containing 40mg in 1 ml was used for each patient. To each ml of Triamcinolone was added 1ml of 2% lidocaine and 2mls of Water for injection to constitute 4ml of TAC suspension with a concentration of 10mg per ml. The suspension was placed in hypodermic syringes for subsequent infiltration into the keloid tissue. For each patient, a maximum of 40mg of Triamcinolone could be infiltrated for the keloid under study.
The standard dosing for the triamcinolone for keloid volume as described by Rahban and Ganner.
Injection of the SVF and TAC into the keloid:
The selected keloid was prepped with 10% Povidone Iodine and then appropriately draped.
Infiltration into the keloid depended on the keloid morphology, volume and shape.
Each keloid, where applicable was divided into four quadrants and the infiltration dose was equally divided into the four quadrants. Infiltration volume of 0.1ml per cubic millimetre or 1ml for every cubic centimetre was targeted. Hence each keloid received which-ever therapy at a dosing of 1ml for every cubic centimetre.
Where the keloid scar was found to be very hard and difficult to infiltrate, ‘needling’ was performed. ‘Needling’ is when a 22-gauge needle is used to create a mesh/network of interconnecting tunnels 2mm apart into the keloid. Following this, infiltration of the treatment was instituted into the created tunnels.
General trial procedure:
Patients that were included and consented to the trial underwent standard pre-operative care including informed consent for surgery. The research assistants enrolled the participants, obtained baseline characteristics, and then randomized and allocated them into either treatment arm. The intervention was administered by a qualified surgeon who had previously been trained on the standards and practices of the procedures. Following the procedure, post-procedural analgesics were provided to participants based on the hospital guidelines for day-care surgery. At follow-up, patients were reassessed clinically at the plastic surgical clinic. The reviews were scheduled at one week and one month while a final follow-up at the end of three months upon which the participants exited the study. The project principal investigator undertook the follow-up reviews. Participants were provided with a transport refund during the follow-up visits.
The primary outcome variables were safety and feasibility of using autologous adipose derived stromal vascular fraction in the treatment of keloids in comparison to Triamcinolone Acetanoide.
Adverse events were defined based on the Common Terminology Criteria for Adverse Events (CTCAE) v5.0 tool as any un-favourable and un-intended symptom, sign or disease associated with the use of a medical treatment or procedure that may or may-not be directly an intended result of the procedure. These adverse events experienced by the participants were documented based on the CTCAE v5.0  and these were reported on a running daily basis for the entire study period. Specific assessments for adverse events development were conducted on day 0 (immediate post procedure), Day 1, day seven, one and three months. The participants were asked to spontaneously report any adverse events to the study principal investigator through a phone call at any time during the follow-up period and the principal investigator would then follow-up with the necessary intervention.
Adverse events were graded as seen in Table 1 and the adverse event grade at each assessment time was recorded.
Common Terminology Criteria for Adverse Events (CTCAE) v5.0 tool
Mild symptoms, no intervention required.
Minimal or local non-invasive intervention needed.
Severe or medically significant
Severe but not immediately life threatening. Requires hospitalization, disabling
Life threatening consequences
Urgent intervention indicated
The criteria for feasibility were based on four parameters that had to be met for the procedure to be considered feasible. which all had to be feasible for the procedures to be classified as feasible.
Recruitment feasibility: This refers to the enrolment and acceptability to participate among those that were found to be eligible. Obtaining consent in 70% of the first 10 respondents reached out to was described as feasible for recruitment. For those that declined, reasons for the decline were sought out.
Intervention feasibility: As a day-care procedure:
Procedure time: The total procedure time in hours was determined. Durations of equal or less than 5 hours were described as feasible.
Procedure completion rates and general anaesthesia conversion:
The procedure was intended to be conducted under local anaesthesia with optional opioid analgesia addition. Completion, as described above without the need for conversion to general anaesthesia or procedural abandonment, was described as feasible.
Keloid thickness measurement: The keloid height/thickness was determined in millimetres using a high-frequency Ultrasound (Healcerion SONON portable ultrasound model 300L). This ultrasound scan device has a capacity of length measurement with up to two decimal point accuracy. Suitability was described as the inter-observer ability to consistently reproduce three ultrasound thickness measurements at the same site with a variation of ≤ ± 0.03.
The feasibility of assessing the primary outcome measures of the main trial of Patients and Observer Scar Assessment Score (POSAS), Keloid volume calculation were also assessed.
Completion rates and challenges in assessing each of these were evaluated.
Participant utilization of any other treatment during the follow-up period, the timing of use, and the reason were determined.
The study follow-up was set up for three months however the appropriateness of this follow-up was evaluated to establish the optimal timing before subsequent follow-up therapy. This was based on the duration of symptom remission or symptom-free duration.
Clinical outcomes were the mean change in the Patient and Observer Scar Assessment Scores (POSAS) from baseline to one month and three months.
Monitoring treatment fidelity:
Because it was the first time this kind of study was being conducted, it was important that treatment fidelity was monitored. Specifically, fidelity adherence checklists were developed and each key step in the study implementation was compared to the described protocol to evaluate the degree of deviation. Deviations from the protocol were identified and the reasons reported. For each procedural step, the investigator was asked to rate how compliant they were in keeping to the trial protocols. Secondly, standardization training of the participants and research assistants was performed to ensure consistency in trial procedures
Sample size estimation:
This was a phase one study intended to explore the safety, feasibility and refine the trial process and not establish effect. Therefore, hypothesis testing to establish a sample size was not necessary. This was based on the recommendations of Arain.
In order to establish the sample size, we used the formular by Lackey et al where they propose that pilot study the sample size comprise of 10% of the main trial size. Based on this recommendation, a pilot sample size of 8 participants which is 10% of the main clinical trial sample size was established with 4 participants allocated in each arm.
Randomization and Allocation concealment:
Randomization was performed using STATA command ralloc by a statistician who was not directly involved in the study. Block randomization using block size 2 was used to allocate patients following a 1:1 ratio.
Sequentially numbered opaque envelopes were used to conceal the allocation sequences and these were prepared by the same statistician that generated the randomization. The envelopes were stapled and handed to the recruitment nurse who only interfaced with the patients during the allocation of treatment the day of the procedure and not interacted with the enrolling research assistant.
At the time of allocation into the different study arms, the recruiting nurse would detach the sequential envelope to identify the allocation and subsequently keep the envelope.
The nature of the study procedure made it impossible to blind the patients and the surgeons as liposuction instantly revealed which arm the patient was allocated. The Outcome assessors on the other hand who followed up patients for review were blinded as they didn’t know what intervention the participants had received.
Data analysis for the participant demographics, baseline characteristics, safety, feasibility, took on a descriptive approach with all data being exported and analysed in STATA 15.0.
Continuous data was reported as means (± SD) while categorical data was reported as proportions with their percentages.
Categorical data analysis was performed for the first primary end-point of safety. Proportions with the percentages of the different Common Terminology Criteria for Adverse Events (CTCAE) v5.0 grades were obtained. Broadly two categories of “adverse event” and “no adverse events” were analysed and compared between the two treatment arms using chi square tests.
For the second primary endpoint of feasibility, variables were categorized into feasible or not feasible and the proportions determined. Comparison between the two treatment arms was performed using chi square tests.
For the secondary outcomes, efficacy endpoints were merely exploratory and therefore analytical assessments were not performed. These were continuous variables and were reported as means with standard deviations. We evaluated for the presence of mean differences in the two treatment arms at one week, one month and three months and baseline.
Patient & Public involvement:
Patients and participants were not be involved in the development of the research question, choice of outcome measures, design of the trial, recruitment of participants and conduct of the trial. Results of the trial were disseminated to the study participants through direct consultation.