Methods
The patients came from two general dermatology clinics. Because there were no previous topical VCO studies in children, pediatric subjects were not allowed by the institutional review board of the Skin and Cancer Foundation. Adult AD diagnoses were based on the modified Hanifin major criteria of a history of a chronic and relapsing course; pruritus; a pattern of facial and extensor eczema and xerosis at a younger age, becoming flexural at adult age; and frequent association with a family history of AD[1] ( Table 1 ).
Those included were newly diagnosed patients aged 18 to 40 years and patients who were previously documented and managed as having AD. Patients with new and old cases of AD had low to high moderate O-SSI scores and had not taken topical steroids or topical or oral antibiotics for at least 2 weeks before enrollment. Excluded were those with (1) grossly infected lesions needing oral or intravenous antibiotics and ancillary therapy; (2) dermatologic diagnoses other than AD; (3) previous hypersensitivity to coconut or olive oil, known diabetes mellitus, or compromised immune states. All patients who met the criteria were oriented on the study's objectives, procedures, and expected outcomes. Informed consent for the study and for photographs was obtained.
Prior to randomization, the participants were stratified on the basis of age and O-SSI score, to control potential confounding variables. The O-SSI score, at baseline and at the end of intervention, was calculated with the formula A/5 + 7 × B/2 (range of 0-83), where "A" represents extent (graded 0-100, based on the rule of nines on a front/back drawing of the patient's inflammatory lesions) and "B" represents intensity (graded 0-18, based on a 0-3 rating of erythema/darkening, edema/papulation, oozing/crust, excoriation, lichenification/prurigo, and dryness); the cut-off points were mild (score < 15), moderate (15-40), and severe (> 40).[11]
The preparation of test bottles, the randomization key, and the codes were carried out by the pharmacist of Skin Sciences Laboratory, Inc., and was disclosed to the investigators only at the end of the study. Subjects previously matched by age and O-SSI score underwent simple concealed random allocation (by drawing rolled pieces of paper labeled "A" or "B") to control or treatment arms by the two dermatology residents, both of whom were blind to the codes and who also dispensed the packaged bottles according to a random listing.
High-quality pure oils were sourced and repackaged in uniform medicinal opaque plastic bottles with a small opening to mask the color and scent of both oils. Unlike the scent of ordinary CO (which is prepared with heat), that of VCO (prepared without any heat) is like that of VOO (ie, botanical and musty). Upon application of either oil, the scent is notable but disappears within just a few minutes. Hours later, neither the patients nor the investigators who see them can identify either oil on their skin by scent. VCO is clear as water and colorless; VOO is also clear but is light yellow green. As either oil is poured onto the hand and applied to the treatment sites, the skin's color makes the VCO and VOO look similarly brownish and indistinguishable from each other.
For the control arm, a commercial VOO was chosen on the basis of its package literature and perceived market value; it was then subjected to a series of microbiologic tests to ensure that it contained no unwanted pathogens.
The VCO used by Agero and Verallo-Rowell was processed without heat but with water-soluble food-derived lipases. The VCO used for the intervention arm of the present study was manufactured without heat under sterile laboratory conditions that followed standard GMP, to avoid the use of chemicals. The temperature was kept at about 33ºC and no higher than 39ºC, temperatures similar to those experienced under tropical sunlight. The resulting VCO passed a series of microbiologic tests and was found to have no unwanted pathogens.
The instructions for applying both EVCO and EVOO were as follows: "On the affected areas that include the test sites, apply 5 mL of EVCO two times a day and massage gently but thoroughly into the skin for several seconds. Do not apply other emollients, creams, or oil-based products that can mask the effect of the oil."
The test bottles were brought in and replaced with new ones at each visit.
All patients were advised to practice good skin hygiene, and all were given a bar of white baby soap to use.
Collection of Skin Swabs. The two resident-investigators together selected two clinically uninfected AD-affected sites, each with an easy-to-identify anatomic landmark. These sites and their identification landmarks were described on the data collection sheet, and photographs were taken. Starting at the center, cotton swabs soaked with sterile normal saline solution were swept over each site (one swab for each site); the cotton swabs were then submitted to the medical technologist at the microbiology laboratory of the Quirino Memorial Medical Center, Pasig, Philippines. After 4 weeks of treatment, cotton swab samples were again taken from the identical sites of the first cultures and sent to the microbiology laboratory.
Colony Growth and Growth Effectiveness. With standard laboratory technique, SA was identified as gram-positive cocci and not resident skin flora or spurious contamination. Colony growth yield of microorganisms considered as significant was also based on standard microbiologic criteria.[12]
At baseline and at the end of intervention, the cultures were examined by a medical technologist who was blind to the treatment arms of the study. The presence of colony counts from one or both sites was reported as "positive"; the absence of colony counts in both sites was reported as "negative." All colony count reports were given to the investigators after the second set of cultures was done.
SA colony growth effectiveness was based on the absence of SA colonies in two separate cultures from two separate sites of clinically noninfected atopic skin after administration of the corresponding intervention.
Descriptive Statistics. Descriptive statistics included (1) means and their standard deviations for categorical variables and (2) percentage frequency distribution for categorical data. Testing of homogeneity of samples was done with the chi-square test for categorical data and an independent t-test for continuous numerical data.
The proportion of significant colonies was compared before and after the trial with nonparametric chi-square tests. Precision estimates were pegged at 95% confidence limits. All tests of significance were carried out with NCSS-PASS software (NCSS, East Kaysville, UT).
Dropouts and Noncompilers. Patients were taught to assess their symptoms and their skin's appearance daily. They were advised to contact the investigators for further treatment when their condition was rated as worse. If topical antibiotics were dispensed, the subject was included in analysis as a dropout; patients who failed to comply with the regimen were classed as protocol violators.
