EFFICACY AND SAFETY OF CENTELLA ASIATICA 441 reports and similar studies from their inception to May 2019. Search keywords were “C. asiatica,” “Gotu kola,” “Asiatic pennywort,” and “Indian pennywort.” We included studies that assessed cosmetic effi cacy and safety by (a) RCTs that compared C. asiatica with no C. asiatica (b) RCTs having suffi cient information about cosmetic outcomes classifi ed as (i) skin wrinkling, (ii) AEs, or (iii) participant opinion/satisfaction and (c) being published in any language. Two discovered studies were unpublished masters thesis lodged in the Thai thesis data- base. We also searched national databases in India and Malaysia, but no study fi tted our criteria. DATA EXTRACTION Two investigators (N. C. and P. M.) independently screened the titles and abstracts of relevant studies and made full-text assessments of those written by the same authors fol- lowing our eligibility criteria. We extracted by study design, area of use, participant characteristics (such as gender, mean age, and number of participants), details of the in- tervention (such as dosage regimens, dose, and durations), and cosmetic outcomes. All extracted data were independently checked by C. K. and C. N. S. QUALITY ASSESSMENT Each included study was assessed by three investigators (C. K., P. M., and C. N. S.) using the Cochrane Risk of bias 2.0 tool for RCTs (15). DATA ANALYSIS For each cosmetic outcome, the mean differences (MD) between C. asiatica and comparator groups [and 95% confi dence intervals (CIs)] after 0, 4, 8, and 12 w using C. asiatica. Inter- group comparisons were performed using the DerSimonian–Laird random effects model (16). Heterogeneity in each comparison was assessed using the I2 statistic (17). Statistical analysis was performed using STATA/SE, v. 14 (StataCorp LLC, College Station, TX). RESULT SEARCH RESULTS Six hundred seventy-one articles were identifi ed. After screening, 654 articles were ex- cluded because they did not fi t the eligibility criteria, and 12 articles described the same studies, leaving fi ve included articles (Figure 1). STUDY CHARACTERISTICS Two studies were from South Korea and two from Thailand (in Thai) and one from China (C. asiatica has been tested in many Italian trials but excluded because the intervention

JOURNAL OF COSMETIC SCIENCE 442 contained additional active ingredients). The published results considered in our systematic review included 172 participants in the fi ve studies (18–22) (Table I). All the studies were RCTs that were double-blinded to treatment. Four studies were single groups using “split” protocols for both test and comparator treatments in the same participants. All partici- pants were female of mean age 33.4 years. Three dosage forms were used [two studies used a cream (20–22), two used gels (18,19), and one study used a lipstick (19)]. Two cosmetic outcomes were identifi ed: four studied periorbital wrinkles (three beneath the eyes and four lateral angle) and one lip wrinkles (19). Two studies were unpublished masters thesis lodged in the Thai thesis database, whereas the remainder were full journal articles. QUALITY OF INCLUDED STUDIES For study quality, the risk of biases was mostly low (Table II), but two studies had “some concerns” due to potential bias from unclear randomization. In addition, Lee et al. (18) and Zhang et al. (22) had incomplete data reporting, making it diffi cult to quantitate conclusions. Figure 1. Flow of inclu d ed studies.