CANTHAXANTHIN DEPOSITION IN THE RETINA - A BIOSTATISTICAL EVALUATION OF 411 PATIENTS

Canthaxanthin (4,4'-diketo-beta-carotene) is a naturally occurring car otenoid that has been used in both medical applications and for cosmet ic skin coloration. In some individuals who have histories of long-ter m, high-dosage canthaxanthin intake, there has been reported an associ ated asymptomatic appearance of reversible golden crystals in the reti na. Because of the potential for continued medical use of this agent, it is important to ascertain whether a dose-response relationship can be described for canthaxanthin-related crystal formation. In addition, the delineation of a dose-response relationshnip is important in asse ssing the safety of the much lower-dose food coloring applications of canthaxanthin, and would provide a statistical basis for the establish ment of a ''no-effect level.'' The possibility of a dose-response rela tionship between crystal formation in the retina and high-dosage canth axanthin use was investigated biostatistically by analyzing the dosage information from 691 cases, as described by 25 different research gro ups in both published and unpublished reports. Of these 691 cases, 131 showed retinal crystals, however, only 411 cases yielded information that could be used for a detailed biostatistical evaluation. Of these 411, 95 demonstrated crystals in the retina. In this population of 411 cases that could be statistically evaluated, daily dosages varied fro m 15 to 240 mg, yearly dosages ranged from 420 to 50,400 mg, and total dosages were from 630 to 201,600 mg within a duration of treatment th at varied from 1 to 14 years. In reviewing the data of this study, a n umber of significant statistical relationships emerge. The increasing percentage of cases with crystals as a function of both total and dail y (< 30 mg = 0%, 30 mg = 9.6%, 45 mg = 20.3%, 60 mg = 23.4%, 75-105 mg = 43.1%) dosage indicates clearly that a highly significant (p < 0.00 01) dose-response relationship exists. The daily, yearly, and total do sage parameters, along with duration of dosage and age, were also foun d to be significantly higher (p < 0.0001) for those cases with crystal s compared with those with no crystals. These dose-response relationsh ips demonstrate statistically for the first time that the phenomenon o f crystallization in the retina after intake of canthaxanthin is assoc iated only with high-dose and long-term treatment. Furthermore, the mi nimum dosage at which crystals were seen in any reported case was at l east 30 mg per day, with no crystals appearing in patients ingesting l evels below this dosage.