Biological Safety Assessment of Docosahexaenoic Acid Supplementation in a Randomized Clinical Trial for X-Linked Retinitis Pigmentosa

CLINICAL SCIENCES Biological Safety Assessment of Docosahexaenoic Acid Supplementation in a Randomized Clinical Trial for X-Linked Retinitis Pigmentosa

Dianna H. Wheaton, MS; Dennis R. Hoffman, PhD; Kirsten G. Locke, CRA; Reginald B. Watkins, BS; David G. Birch, PhD

Background: In a 4-year placebo-controlled trial to el- Results: Mean plasma docosahexaenoic acid levels evate blood docosahexaenoic acid levels in patients were elevated 2.5-fold by supplementation compared with X-linked retinitis pigmentosa (XLRP), the goal with baseline. Patients receiving placebo capsules was to assess the potential benefit of docosahexaenoic exhibited no change (P=.35) in plasma docosahexae- acid supplementation in altering disease progression. noic acid content. All adverse events reported were However, docosahexaenoic acid (22:6␻3) is a highly minor and equivalently distributed between groups. unsaturated fatty acid and considered a target molecule Plasma vitamin A concentrations remained unchanged for free-radical oxidative damage. Thus, nutritional during the trial. Mean plasma vitamin E concentra- provision of docosahexaenoic acid might lead to an in- tions were correlated with age (P=.005), such that as crease in antioxidant stress. Additional concerns, such patients with XLRP matured, plasma vitamin E con- as decreased platelet aggregation, increased bleeding centrations increased to approach normal values. time, and alterations in lipoprotein cholesterol levels, There was a trend (P=.10) toward lower mean vitamin have been reported in supplementation studies with E concentrations in the docosahexaenoic acid– long-chain polyunsaturates. supplemented group after 4 years. Docosahexaenoic acid supplementation did not compromise plasma antioxi- Objective: To assess the biological safety of long-term dant capacity, platelet aggregation, liver function en- docosahexaenoic acid supplementation. zyme activity, or plasma lipoprotein lipid content in patients with XLRP. Design: Forty-four male patients (mean age, 16 years) enrolled in a randomized, double-masked, clini- Conclusion: Long-term docosahexaenoic acid supple- cal trial and received docosahexaenoic acid, 400 mg/d, mentation to patients with XLRP was associated with no or placebo. Blood samples were collected every 6 identifiable safety risks in this 4-year clinical trial. months. Biological safety analysis included fatty acids, vitamin A and E concentrations, antioxidant capacity, platelet aggregation, alanine aminotransferase activity, and lipoprotein cholesterol and triglyceride profiles. Arch Ophthalmol. 2003;121:1269-1278

ETINITIS PIGMENTOSA (RP) is heterogeneous and can be inherited in an a genetic eye disease char- autosomal dominant, autosomal reces- acterized by progressive sive, X-linked recessive, or sporadic man- retinal degeneration. Clini- ner. To date, no cure for RP has been iden- cal manifestations include tified and clinical interventions to retard disturbancesR of retinal pigment, vessel at- disease progression are limited. tenuation, and thinning of the retinal pig- Numerous studies have reported re- From the Retina Foundation of ment epithelium. Degeneration typically duced blood levels of the ␻3 fatty acid, the Southwest (Mss Wheaton occurs first in the peripheral retina, af- docosahexaenoic acid, in patients with and Locke, Drs Hoffman and fecting the rod photoreceptors and result- RP.1,2 This reduction is particularly strik- Birch, and Mr Watkins); and ing in night blindness. The disease ing in patients with X-linked RP (XLRP).3-5 the Departments of Pediatrics progresses inward toward the macula, lead- After adjusting for the effects of age, con- (Dr Hoffman) and ing to progressive restriction of the pe- centrations of docosahexaenoic acid in red Ophthalmology (Dr Birch), The University of Texas ripheral visual field (ie, tunnel vision). As blood cell (RBC) lipids of patients with Southwestern Medical Center at cone function within the macula be- XLRP were positively correlated with reti- Dallas. The authors have no comes progressively impaired, legal blind- nal function measured by electroretinog- proprietary interest in the ness and loss of ambulatory vision often raphy (ERG), such that higher blood doco- product described in this article. result. Retinitis pigmentosa is genetically sahexaenoic acid concentrations were

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1269

©2003 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 associated with higher (ie, better) rod6 and cone ERG am- lected to monitor the bioavailability and safety of long- plitudes.3 Heterozygous carriers of XLRP have demon- term LCPUFA supplementation. strable reductions in blood docosahexaenoic acid levels and cone ERG function similar to, but less dramatic than, that described in the hemizygous male.7 METHODS Docosahexaenoic acid (22:6␻3), with 6 double bonds, is the most unsaturated membrane fatty acid SUBJECTS present in biological systems, and comprises 30% to 40% Patients diagnosed as having XLRP were recruited from the South- of the total fatty acids in phospholipids of the cerebral west Eye Registry (Retina Foundation of the Southwest)22 for par- 8 9 cortex and retina. Although no functional role for doco- ticipation in this clinical trial. Recruitment was extended na- sahexaenoic acid has been identified in the visual cas- tionally through the referral of clinical centers supported by the cade, research10 implies that this ␻3 fatty acid is biologi- Foundation Fighting Blindness, Owins Mills, Md. Major eligi- cally significant to neural tissue. Within the retina, bility criteria for study participation included a diagnosis of RP docosahexaenoic acid is concentrated in highly special- by an ophthalmologist specializing in retinal disease, a family history consistent with X-linked inheritance, detectable cone ERG ized membranes that make up photoreceptor outer seg- Ͼ ments, and is found in phospholipids that are tightly as- responses to a 31-Hz flicker (ie, 0.34 µV), and willingness to sociated with the visual chromophore rhodopsin.11,12 The return annually for a visual function assessment. Fifty-two patients were screened for inclusion in the trial; dependency on docosahexaenoic acid–containing phos- 4 patients did not meet the enrollment criteria and 4 declined pholipid bilayers has recently been described for the ki- to participate. All participants were male, because the disease netic conversion of metarhodopsin I to metarhodopsin is more severe in hemizygous males than heterozygous carri- II,13 the formation of activated rhodopsin-transducin com- ers. The study was explained in detail, and informed consent plex,14 and activation of phosphodiesterase.15 was obtained from participants and/or parents of minors. This A reduced level of docosahexaenoic acid in blood research protocol observed the tenets of the Declaration of Hel- lipids provides a rationale for a clinical intervention us- sinki, and was approved by the Institutional Review Board of ing docosahexaenoic acid supplementation in patients The University of Texas Southwestern Medical Center at Dal- with RP. The X-linked form of the disease is considered las. In addition, the use of docosahexaenoic acid in this trial one of the most severe forms of RP based on early onset that included minors was conducted under an investigational 16-18 new drug approval (No. 45942) issued by the US Food and Drug of visual loss and functional blindness at a young age. Administration, Rockville, Md. As a group, patients with XLRP have the lowest blood docosahexaenoic acid levels and the earliest onset of dis- STUDY DESIGN ease severity, making this an ideal population for an early treatment intervention. A randomized, placebo- Forty-four patients (mean±SD age, 16±9 years; range, 4-38 controlled, clinical trial was undertaken to assess the po- years) (see Table 1 for baseline demographic characteristics) tential benefit of nutritional docosahexaenoic acid supple- were randomized to parallel arms of the study, and received mentation. The primary goals of this trial were 2-fold: capsules containing either docosahexaenoic acid–enriched oil (1) to determine if daily oral supplementation with doco- or a placebo oil (corn/soy oil) for the duration of the 4-year sahexaenoic acid would elevate blood docosahexaenoic trial. Assignments were made following a 10 per block ran- domization schedule. Close relatives were randomized to- acid levels and (2) to determine if variations in the RBC gether to eliminate a potential for mixing of capsules; there were docosahexaenoic acid levels of patients are related to the 5 sibling pairs in each cohort. rate of disease progression. The primary outcomes have Both oil formulations were gelatin encapsulated to pro- been reported separately.19 Briefly, RBC docosahexa- vide a total fat content of 500 mg per capsule. Each capsule also enoic acid levels were significantly elevated in supple- contained vitamin E, 12.5 IU, and ascorbyl palmitate, 12.5 mg, mented patients; however, the rate of disease progres- as antioxidants to reduce oxidative degradation of LCPUFAs. sion, as measured by cone ERG, was not significantly Placebo and docosahexaenoic acid–enriched capsules were iden- different between the 2 groups using an intent-to-treat tical in appearance, taste, and smell. Patients were instructed analysis. Nevertheless, cone and rod photoreceptor loss to take two 500-mg capsules per day with no recommended correlated with RBC docosahexaenoic acid levels such changes in their normal dietary practices. Oral supplementation with 2 enriched capsules provided 400 mg of docosahexae- that patients with higher docosahexaenoic acid levels had noic acid daily. Thus, patients receiving the enriched capsules reduced ERG loss. averaged about 10 mg of docosahexaenoic acid per kilogram As a long-chain polyunsaturated fatty acid of body weight per day. Seven of the youngest patients were (LCPUFA), docosahexaenoic acid is considered a poten- provided half-sized capsules to ease swallowing; taking 4 of these tial target molecule for free radical oxidative damage. In- capsules provided the full dosage. As the study progressed, all creased LCPUFA intake may lead to an elevated state of patients were switched to full-sized capsules. All supplements oxidative stress, and subsequently result in membrane were labeled either A or B by the manufacturer (Martek Bio- damage.20 Additional findings, such as decreased plate- sciences Corporation, Columbia, Md) to mask study partici- let aggregation, increased bleeding time, and alterations pants and research staff. The manufacturer retained the code in lipoprotein cholesterol level, have also been reported and was willing to divulge group assignment to the Data and 21 Safety Monitoring Committee, the Institutional Review Board, in previous supplementation studies with LCPUFAs. or a patient’s physician in case of a medical emergency. This Therefore, a parallel goal of this treatment trial was to option was not used during the trial. assess the biological safety of long-term docosahexae- The docosahexaenoic acid–enriched oil used for the cap- noic acid supplementation in patients with XLRP. We re- sules was derived from a single-cell algal source and provided port the outcome of biochemical assays that were se- as a highly purified triacylglycerol (DHASCO oil; Martek Bio-

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1270

©2003 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 Table 1. Baseline Demographic Characteristics Table 2. Fatty Acid Composition of Intervention Capsules of the Cohorts* Docosahexaenoic Docosahexaenoic Acid−Enriched Placebo Acid−Supplemented Common Capsules Capsules Placebo Group Group Fatty Acid Name (DHASCO Oil)* (Corn/Soy Oil)* Characteristic (n = 21) (n = 23) 10:0 Capric acid 0.6 0 Age, mean ± SD, y 18.0 ± 10.1 14.9 ± 9.0 12:0 Lauric acid 5.0 0 White–African American– 19:0:2:0 19:1:3:0 14:0 Myristic acid 21.6 0 Latin American–other ratio 16:0 Palmitic acid 19.8 11.0 Body weight, mean ± SD, kg 62 ± 29 58 ± 30 16:1 Palmitoleic acid 0.9 0 Body mass index† 21.8 ± 6.5 21.3 ± 6.6 18:0 Stearic acid 0.4 3.1 Mean ± SD 18:1 Oleic acid 10.1 25.0 Ͻ20 9 12 18:2␻6 Linoleic acid 0.6 56.2 20-30 9 9 18:3␻3 ␣-Linolenic acid 0 4.0 Ͼ30 3 2 22:6␻3 Docosahexaenoic acid 40.1 0 Plasma docosahexaenoic acid, 24:1 Nervonic acid 0.5 0 % of total fatty acids Others Not applicable 0.4 0.7 Mean ± SD 1.48 ± 0.33 1.46 ± 0.37 Range 0.94-2.18 0.98-2.19 *Data are given as percentage of total fatty acids. Each placebo and Daily vitamin supplement use docosahexaenoic acid−enriched capsule (500 mg) contained 12.5 mg (12.5 Multivitamin 12 4 IU) of vitamin E and 12.5 mg of ascorbyl palmitate. Fatty acid analysis was Vitamin A 1 3 conducted at the Retina Foundation of the Southwest, Dallas, Tex. Vitamin E 2 0

*Data are given as number of patients unless otherwise indicated. All patients to address the concerns of increased bleeding time and de- were male. Four patients in each group smoked 1 pack of cigarettes per day or creased platelet activation, which had been associated with ␻3 more. No statistically significant group differences were found for age (t = 1.06, fatty acids in some but not all studies that used fish oil as an P = .29), body mass index (t = 0.25, P = .80), or plasma docosahexaenoic acid 21 level (t = 0.19, P = .85). The normative values of the Retina Foundation of the LCPUFA source. At 4 years, plasma alanine aminotransfer- Southwest, Dallas, Tex, are as follows: n = 20; mean ± SD age, 17.2 ± 6.6 years ase (ALT) activity was measured as an index of liver function, (range, 10-30 years); white–African American–Latin American–other ratio, because it was recently reported that the hepatic activity of ⌬5 18:2:0:0; smoking, 0 patients; mean ± 95% confidence interval for body weight, desaturase was reduced in patients with XLRP.24 An overview 66.2 ± 11.6 kg; mean ± 95% confidence interval for body mass index, of the study design is given in Table 3. Of the 44 patients en- 21.9 ± 2.3 (range, 15.6-38.0); mean ± 95% confidence interval for plasma docosahexaenoic acid percentage of total fatty acids, 1.90 ± 0.33; and 2 control rolled in the study, 41, 43, 44, and 41 completed testing at the subjects supplemented with a daily multivitamin. 1-, 2-, 3-, and 4-year annual visits, respectively. †Calculated as weight in kilograms divided by the square of height in meters. Near the conclusion of the study, a group of age- and sex- matched control subjects with healthy vision (mean±SD age, 17±7 years; range, 10-30 years; n=20) provided normative data sciences Corporation). The docosahexaenoic acid content of for each of the biochemical assays. Volunteers with healthy vi- these capsules is standardized at 40% of the total fatty acids sion were subjected to a brief visual function examination (vi- (Table 2); the remaining 60% of the fatty acids are all natu- sual acuity and visual field perimetry measurements), and a 1-time rally occurring and presumably do not compromise the speci- blood sample was drawn as described for study participants. ficity of the study. Extensive toxicology studies23 at elevated doses of this docosahexaenoic acid–enriched oil have been con- BLOOD COLLECTION ducted in animal models to establish safety for human consumption. This docosahexaenoic acid–enriched oil has been used Blood samples were drawn from an antecubital vein of sub- worldwide in infant formulas, and was recently granted “gen- jects who had fasted for at least 8 hours. Typically, 10 to 20 erally regarded as safe” approval from the US Food and Drug mL of blood was collected in tubes containing 15% potassium Administration for inclusion in infant formulas in the United EDTA as an anticoagulant. Beginning at the 3-year point, 2 ad- States. The placebo capsule contained corn/soy oil triglycer- ditional samples (5 mL each) were collected for biological safety ides (TGs) that provided the essential fatty acids, linoleic (18: studies in tubes containing 3.8% sodium citrate or 72 U of so- 2␻6) and ␣-linolenic (18:3␻3). dium heparin as anticoagulant, respectively. Blood samples were collected every 6 months for the duration of the trial. Plasma lipid fatty acid determination was used FATTY ACID ANALYSIS to monitor bioavailability, protocol compliance, and alterations in fatty acid profiles, particularly the ␻3/␻6 fatty acid Details of the lipid analysis have been reported previously.25 ratio. Antioxidant status was monitored to assess the impact Briefly, plasma and RBCs were separated immediately by cen- of LCPUFA supplementation on antioxidant defense mecha- trifugation (3000g for 10 minutes at 4°C), and lipids were ex- nisms. Plasma concentrations of the lipid-soluble antioxidant, tracted using a 2:1 ratio of methanol and chloroform contain- vitamin E, and the retinal chromophore precursor, vitamin A, ing 0.02% butylated hydroxytoluene as an antioxidant. Total were used as indicators of oxidative stress. The effect of low- lipids from plasma were transmethylated (using a combina- dose docosahexaenoic acid supplementation on plasma lipo- tion of 14% boron trifluoride and methanol) and subse- protein cholesterol and TG profiles was also followed through- quently quantified using capillary column gas chromatogra- out the trial. Additional safety assays were introduced at the phy and flame ionization detection. Peak identification was 3-year point; these included plasma antioxidant capacity and confirmed by comparing retention times with those of a stan- platelet aggregation. The total plasma antioxidant capacity as- dard mixture of fatty acid methyl esters. Plasma fatty acid quan- say was used as a supplementary measure of overall antioxi- titation was expressed as relative weight percentage (percent- dant levels. Whole blood platelet aggregation was monitored age of total fatty acid).

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1271

Annual Visit

Analysis 01234 Visual function 21/23 18/23 20/23 21/23 19/22 Fatty acids 21/23 18/23 20/23 21/23 19/22 Vitamins 21/23 18/23 20/23 21/23 19/22 Antioxidants 0/0 0/0 0/0 21/23 19/22 Aggregation 0/0 0/0 0/0 21/23 19/22 Lipoproteins 21/23 18/23 20/23 21/23 19/22 Alanine aminotransferase 0/0 0/0 0/0 0/0 16/14

*Data are given as the number of patients tested at each annual visit for the placebo/docosahexaenoic acid−supplemented groups. An intent-to-treat protocol was used for data analysis at each time point.

ANTIOXIDANT ANALYSIS ALANINE AMINOTRANSFERASE

Measurement of Vitamins A and E Liver function was indexed by assaying ALT enzyme activity with an automated analyzer (Cholestech L·D·X; Cholestech The rhodopsin chromophore precursor, vitamin A (retinol), and Corp, Hayward, Calif). This analyzer measures ALT level by the antioxidant, vitamin E (␣-tocopherol), were measured on an enzymatic method and solid-phase technology using a cas- a high-pressure liquid chromatograph (Beckman Instru- sette-based assay (No. 11-772; Cholestech Corp). The addi- ments, Fullerton, Calif). The fat-soluble vitamins were sol- tion of heparin-anticoagulated whole blood (50 µL) to the cas- vent extracted from EDTA-anticoagulated plasma and fraction- sette initiates the first of 3 enzyme-catalyzed reactions and ated on a C18 reversed-phase analytical column (Alltech terminates with the formation of a blue color at a rate propor- Associates, Deerfield, Ill) by methods previously described by tional to the ALT activity of the sample. The resulting color is Bui.26 Analytes were resolved with an isocratic flow of 1 mL/ measured by reflectance photometry, with enzyme activity re- min of mobile phase (acetonitrile–tetrahydrofuran– ported as units per liter at 37°C. methanol–1% ammonium acetate, 68:22:7:3 vol/vol/vol/vol) and detected at 325 nm for vitamin A and 290 nm for vitamin E. LIPOPROTEIN LIPID PROFILES Quantitation was achieved by comparison with calibration curves derived from standards for vitamins A and E (Sigma-Aldrich The analyzer (Cholestech L·D·X)was also used to determine Corp, St Louis, Mo), and expressed as microgram per deciliter plasma lipoprotein lipid profiles. EDTA-anticoagulated plasma and milligrams per deciliter, respectively. (50 µL) was added to a lipid profile cassette (No. 10-989; Cho- lestech Corp) that allowed the simultaneous measurement of Total Antioxidant Capacity total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and TGs via a series of multistep enzymatic reac- The total antioxidant capacity of plasma was measured using tions. Values for low-density lipoprotein cholesterol (LDL-C) an enhanced chemiluminescence modification of the total per- were calculated by the analyzer using the measured values. Con- oxyl radical trapping parameter.27 The antioxidant capacity of centrations were recorded as milligrams per deciliter . an aliquot (20 µL) of citrate anticoagulated plasma diluted 1:10 with isotonic sodium chloride solution was established by its ADVERSE EVENTS ability to quench a horseradish peroxidase–catalyzed reaction generated by a chemiluminescence kit (catalog number RPN Adverse events were self-reported on a per-incident basis to the 190; Ortho-Clinical Diagnostics, Buckinghamshire, England). study coordinator (K.G.L.). Postevent follow-up was con- Quenching capacity was assayed on a luminometer (Turner De- ducted by telephone interview, and outcome was recorded ac- signs, Sunnyvale, Calif) with subsequent quantification by com- cordingly. Patients were encouraged to inform their primary parison with a standard curve (10-100 µmol/L) of the syn- care physician of their participation in the supplementation trial. thetic vitamin E derivative (Trolox [6-hydroxy-2,5,7,8- A letter was prepared and given to each patient or parent on tetramethylchroman-2-carboxylic acid]; Aldrich, Milwaukee, enrollment to provide to their physician in the event that spe- Wis). Antioxidant activity was expressed as micromolar equiva- cific questions pertaining to the trial or supplements were raised. lents of this vitamin E derivative. STATISTICAL ANALYSIS WHOLE BLOOD PLATELET AGGREGATION All patient data were used as dictated by an intent-to-treat pro- Platelet aggregation was measured using an aggregometer tocol. Data from previous measurements were used to replace miss- (Chrono-Log Corp, Havertown, Pa). Citrate-anticoagulated ing values. A 2-tailed t test was used to compare fatty acid and blood (200 µL) was diluted (1:4) with Tyrode buffered saline biosafety outcome variables between the treatment groups at each solution (sodium chloride, 120 mmol/L; potassium chloride, time point in the trial. In addition, repeated-measures analyses 5 mmol/L; sodium bicarbonate, 10 mmol/L; calcium chloride, of variance were used for longitudinal analysis of vitamins A and 2 mmol/L; magnesium chloride, 2 mmol/L; sodium phos- E. Statistical significance was considered at PϽ.05 for antioxi- phate, 4 mmol/L; glucose, 10 mmol/L; and heparin, 2 U/mL), dant, platelet aggregation, ALT, and lipoprotein lipid measures. and aggregation was induced with 10 µg of Type I collagen However, for most fatty acid values, a more stringent assessment (No. 385; Chrono-Log Corp). Results are expressed as imped- (PϽ.003, Bonferroni adjustment) was used. Values for patients ance (ohms) to aggregation of the sample vs a Tyrode blank with XLRP are given as mean±SD, and those for age- and sex- control. matched controls as mean±95% confidence interval (CI).

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1272

©2003 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 Table 4. Fatty Acid Profiles in Total Plasma Lipids Placebo Group Docosahexaenoic Acid – Supplemented Group of Trial Subjects With XLRP* ARA Docosahexaenoic Acid Docosahexaenoic Placebo Acid−Supplemented Group Group P 10 Fatty Acid Variable (n = 21) (n = 23) Value

8 % of total fatty acids ∗ ␻3 ␣-LNA 0.44 ± 0.08 0.46 ± 0.13 .55 6 EPA 0.40 ± 0.11 0.40 ± 0.08 1.0 ␻3 DPA 0.73 ± 0.11 0.46 ± 0.09† Ͻ.001 4 † Docosahexaenoic acid 1.57 ± 0.32 3.91 ± 1.24† Ͻ.001 % of Fatty Acids ␻6 2 LA 26.60 ± 2.67 26.60 ± 1.67 1.0 ARA 9.35 ± 0.75 7.51 ± 1.10† Ͻ.001 0 ␻6 DPA 0.41 ± 0.12 0.20 ± 0.06† Ͻ.001 0 12243648 Time Taking Supplement, mo Totals Saturates‡ 35.00 ± 1.21 35.20 ± 1.28 .60 Figure 1. Relative mean±SD percentage of total fatty acids of plasma Monounsaturates§ 22.20 ± 1.18 22.00 ± 1.92 .68 docosahexaenoic acid and arachidonic acid (ARA) in the placebo group LCPUFAs (n=21) and the docosahexaenoic acid–supplemented group (n=23) as ␻3 2.66 ± 0.43 4.75 ± 1.16† Ͻ.001 function of time taking supplements. Normative data for age- and ␻6 12.90 ± 0.94 10.50 ± 1.19† Ͻ.001 sex-matched control subjects (n=20) obtained from a single blood sample Ratios are represented by the brackets, illustrating the mean±95% confidence ␻6 PUFA/␻3 PUFA 12.80 ± 2.20 7.56 ± 1.63† Ͻ.001 interval (1.9%±0.3% and 8.8%±2.5% for docosahexaenoic acid and ARA, ␻6 DPA/docosahexaenoic 0.27 ± 0.06 0.07 ± 0.04† Ͻ.001 respectively). All values for the docosahexaenoic acid–supplemented group acid from 6 months to 4 years are significantly different from those of the placebo Ͻ group (the asterisk indicates PϽ.05; and the dagger, PϽ.003). ARA/docosahexaenoic acid 6.20 ± 0.96 2.36 ± 1.14† .001

Abbreviations: ARA, arachidonic acid; DPA, docosapentaenoic acid; EPA, RESULTS eicosapentaenoic acid; LA, linoleic acid; LCPUFA, long-chain polyunsaturated fatty acid; LNA, linolenic acid; XLRP, X-linked retinitis pigmentosa. *Data are given as mean±1SDforyears 0.5 to 4. FATTY ACIDS †The difference between the placebo group and the docosahexaenoic acid−supplemented group is significant. There was no significant difference at baseline between ‡Includes 14:0, 16:0, 17:0, 18:0, 20:0, 22:0, and 24:0. mean plasma docosahexaenoic acid levels of patients ran- §Includes 16:1, 18:1, 20:1, 22:1, and 24:1. domized to either the placebo or the docosahexaenoic acid–supplemented group (1.48% vs 1.46%) (Table 1); was reduced by 50% in the docosahexaenoic acid– values for the entire cohort ranged from 0.94% to 2.19% supplemented group (0.20% vs 0.41% of total fatty acids). of total fatty acids. A group of age- and sex-matched con- These alterations were also reflected in the ratios of ARA/ trols with healthy vision had a mean±SD plasma doco- docosahexaenoic acid, ␻6 docosapentaenoic acid/ sahexaenoic acid level of 1.90%±0.33% (range, docosahexaenoic acid, and ␻6 PUFA/␻3 PUFA, which 1.06%-3.68%). Thus, the study patients exhibited plasma were reduced by 62%, 75%, and 41%, respectively, in pa- docosahexaenoic acid levels that were approximately 75% tients receiving docosahexaenoic acid (Table 4). The doco- of normal before randomized capsule assignment. sahexaenoic acid–supplemented group exhibited an over- The docosahexaenoic acid–supplemented group ex- all elevation in total ␻3 LCPUFAs and a decrease in total hibited an increased plasma docosahexaenoic acid con- ␻6 LCPUFAs; however, the total sum of saturated and tent within the first 6 months of supplementation monounsaturated fatty acids was not different between (Figure 1). During the 4-year trial, the plasma docosa- the 2 groups (PϾ.60). Although the placebo capsule pro- hexaenoic acid level of the docosahexaenoic acid– vided the essential fatty acids, ␣-linolenic and linoleic ac- supplemented group (mean±SD for years 0.5-4, ids (18:3␻3 and 18:2␻6, respectively), no significant 3.91%±1.24%; range, 2.12%-6.46%) was elevated an av- group differences were observed for these fatty acids erage of 2.5-fold compared with baseline levels. Signifi- (PՆ.55). The plasma level of the ␻3 intermediate, eicosa- cant group differences (PϽ.003) were maintained pentaenoic acid (EPA) (20:5␻3), was not different be- throughout the trial. In the placebo group, plasma doco- tween the 2 groups; however, ␻3 docosapentaenoic acid sahexaenoic acid content at each annual visit remained (22:5␻3) was reduced in the docosahexaenoic acid– unchanged from the baseline value (PϾ.20). supplemented group. A corresponding reduction in arachidonic acid (ARA) (20:4␻6) also was observed in the docosahexaenoic acid– ANTIOXIDANT ANALYSIS supplemented cohort after the first 6 months of supplementation. The plasma ARA levels remained signifi- Measurement of Vitamins A and E cantly reduced (PϽ.001), compared with the placebo group, for the duration of the trial (0.5 to 4 years) (Fig- Plasma concentrations of vitamin A did not differ be- ure 1 and Table 4). ␻6 Docosapentaenoic acid (22: tween the placebo and docosahexaenoic acid– 5␻6), the end product of the ␻6 biosynthetic pathway, supplemented groups at baseline (mean±SD, 77±24 µg/dL

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1273

©2003 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 42 to 125 µg/dL (1.46-4.36 µmol/L); and vitamin E con- Placebo Group centrations, from 0.7 to 2.7 mg/dL (16.3-62.9 µmol/L) A 172 Docosahexaenoic Acid – for the combined cohorts with XLRP. Vitamin A and E Supplemented Group concentrations were significantly correlated with age 143 (Pearson product moment correlation; r=0.56[PϽ.001] Ͻ 115 and r=0.38 [P .02], respectively), such that vitamin lev- g/dL

µ els increased with increasing age in patients with XLRP. 86

57 Total Antioxidant Capacity Plasma Vitamin A Concentration, 29 Total plasma antioxidant capacity was not significantly

0 different between the 2 groups after 4 years of supple- B mentation (mean±SD, 366±80 vs 355±80 µmol/L of syn- 2.15 thetic vitamin E derivative equivalents; P =.67) (Figure 3A). Both groups exhibited plasma antioxi- 1.72 dant capacities within the 95% CIs of the age- and sex- 1.29 matched controls (mean±95% CI, 352±29 µmol/L of synthetic vitamin E derivative equivalents). 0.86 PLATELET AGGREGATION

Plasma Vitamin E 0.43 Concentration, mg/dL After the fourth year of nutritional intervention, whole 0 012243648 blood platelet aggregation was not significantly differ- Time Taking Supplement, mo ent between the 2 groups (mean±SD, 16.8±5.7 vs ⍀ Figure 2. Mean±SD plasma concentrations of vitamin A (A) and vitamin E 14.9±7.1 ; P=.35). The placebo and docosahexaenoic (B) at each annual visit of the trial for the placebo group (n=21) and the acid–supplemented groups demonstrated less collagen- docosahexaenoic acid–supplemented group (n=23). Brackets represent the induced platelet aggregation than controls (mean±95% mean±95% confidence interval for the age- and sex-matched control ⍀ ⍀ subjects (n=20) (vitamin A, 90±8 µg/dL; and vitamin E, 1.5±0.2 mg/dL). CI, 22±2 ; range, 13-29 ) (Figure 3B). Although re- A trend (P=.10 and P=.11) toward reduced levels of vitamin E in the duced platelet aggregation was a concern of long-term docosahexaenoic acid–supplemented group compared with the placebo supplementation with LCPUFAs, reduced aggregation was group was observed at the 3- and 4-year intervals, respectively. To convert noted in both study groups and, therefore, was not as- vitamin A to micromoles per liter, multiply by 0.0349; and to convert vitamin E to micromoles per liter, multiply by 23.33. sociated with docosahexaenoic acid supplementation per se. As an alternative, vitamin E in placebo and docosahexaenoic acid–enriched capsules (12.5 IU of vitamin E [2.69±0.84 µmol/L] vs 76±24 µg/dL [2.64±0.83 µmol/ per capsule) may accumulate in vivo to approach a level L]) or at any subsequent annual visit through the re- capable of inhibiting platelet aggregation.28 However, no mainder of the trial (repeated-measures analysis of vari- correlation was found between plasma vitamin E con- ance; F=0.05, P=.82) (Figure 2A). At year 4, the plasma tent and whole blood platelet aggregation at the 4-year vitamin A concentration of the study cohort was not sig- point for the combined cohorts with XLRP (PϾ.90). nificantly different (P=.07) from that of the age- and sex- matched controls (mean±95% CI, 90±8 µg/dL [3.15±0.28 ALANINE AMINOTRANSFERASE µmol/L]; range, 57-137 µg/dL [1.99-4.79 µmol/L]). No significant group differences (P=.38) for mean±SD plasma A subset of patients from the placebo group (n=16) and vitamin E concentrations were recorded at baseline (pla- the docosahexaenoic acid–supplemented group (n=14) cebo group, 0.82±0.63 mg/dL [19.1±14.6 µmol/L]; and were randomly selected for assessment of ALT activity docosahexaenoic acid–supplemented group, 0.70±0.23 after 4 years of supplementation. The mean±SD ALT ac- mg/dL [16.2±5.3 µmol/L]) (Figure 2B). There was no sig- tivity for the docosahexaenoic acid–supplemented group nificant group effect on vitamin E level during the 4-year was 18.1±7.1 U/L, and was not significantly different from trial (repeated-measures analysis of variance; F=1.2, the placebo group (24.8±15.8 U/L) (P=.15) (Figure 3C). P=.27). However, a trend for a reduced plasma vitamin When compared with normative data (mean±95% CI, E level was observed in the docosahexaenoic acid– 20.6±3.1 U/L; range, 12-40 U/L), the placebo group ex- supplemented group for years 3 and 4 (P=.10 and P=.11, hibited ALT activity slightly greater than the upper 95% respectively; t test). In addition, both groups exhibited CI, whereas the docosahexaenoic acid–supplemented vitamin E levels below normal (mean±95% CI, 1.5±0.2 group exhibited activity near the lower 95% CI. mg/dL [35.6±4.8 µmol/L]; range, 0.9-2.8 mg/dL [21.3- 64.0 µmol/L]) throughout the 4-year trial. Although nu- LIPOPROTEIN LIPID PROFILES merous patients from both cohorts reported taking vitamin supplements during the trial (Table 1), no No significant differences were found at year 4 between individuals showed plasma levels for vitamin A or vita- measures of plasma lipoprotein lipid profiles in the pla- min E elevated above the normative range; only 2 refer- cebo and docosahexaenoic acid–supplemented groups ence controls reported daily supplementation with a mul- (PϾ.70) (Figure 4A-D) for TC (mean±SD, 170±34 vs tivitamin. At year 4, vitamin A concentrations ranged from 160±40 mg/dL [4.40±0.88 vs 4.14±1.03 mmol/L]), TGs

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1274

500 30 40 400

Ω 25 mol/L

µ 30 300 20

200 15 20 Impedance,

10 U/L Activity, ALT Equivalents, 100 10 5 Synthetic Vitamin E Derivative 0 0 0 Placebo Docosahexaenoic Placebo Docosahexaenoic Placebo Docosahexaenoic Group Acid – Supplemented Group Acid – Supplemented Group Acid – Supplemented Group Group Group

Figure 3. Mean±SD year 4 values for total antioxidant capacity of plasma (A), whole blood platelet aggregation (B), and plasma alanine aminotransferase (ALT) activity (C) in the placebo group and the docosahexaenoic acid–supplemented group. Brackets represent the mean±95% confidence interval for the age- and sex-matched control subjects (n=20) for antioxidant capacity (352±29 µmol/L of synthetic vitamin E derivative [Trolox] equivalents), platelet aggregation (22±2 ⍀ of impedance), and ALT activity (20.6±3.1 U/L), respectively. There were no significant differences between the placebo and docosahexaenoic acid–supplemented groups for antioxidant capacity (P=.67), platelet aggregation (P=.35), or ALT activity (P=.15). The data for ALT were analyzed for subgroups of 16 and 14 patients receiving placebo or docosahexaenoic acid supplements, respectively.

(mean±SD, 110±78 vs 117±55 mg/dL [1.24±0.88 vs 13-32 years) and 4 subjects from the docosahexaenoic 1.32±0.62 mmol/L]), HDL-C (mean±SD, 47±15 vs acid–supplemented group (age range, 9-19 years). Those 45±14 mg/dL [1.22±0.39 vs 1.16±0.36 mmol/L]), and symptoms thought to be potentially associated with cap- LDL-C (mean±SD, 98±27 vs 98±39 mg/dL [2.53±0.70 sule/LCPUFA consumption were as follows: bruising, pro- vs 2.53±1.01 mmol/L]), respectively. In addition, no sig- longed bleeding, and gastrointestinal symptoms. nificant group differences (PϾ.32) were found at base- Adverse events reported by patients in the placebo line or years 1, 2, or 3 for any lipoprotein lipids (data group included bruising (n=2), sinusitis (n=1), pro- not shown). Concentrations of TC, TGs, and LDL-C for longed epistaxis (n=1), eructation (burp back) (n=1), both groups were within the 95% CI of controls nausea (n=1), and flatulence (n=1). Postevent fol- (mean±95% CI for controls, 163±15, 95±25, and 94±14 low-up disclosed that the 2 cases of bruising were due mg/dL [4.22±0.39, 1.07±0.28, and 2.43±0.36 mmol/ to physical contact sports. The prolonged epistaxis was L], respectively). Both groups exhibited HDL-C levels be- preceded by a case of sinusitis in 1 individual. A bleed- low normal (mean±95% CI, 51±5 mg/dL [1.32±0.13 ing time test conducted by the patient’s family physi- mmol/L]). Higher HDL-C levels are thought to be a health cian subsequently determined bleeding time to be within benefit, whereas an HDL-C level below 35 mg/dL (0.91 normal limits. The case of nausea was short-term (Ͻ1 mmol/L) is associated with an increased risk for coro- week), and eructation and flatulence were related to cap- nary heart disease (CHD).29 Because both groups exhib- sule consumption. One additional case each of apnea, ited mean HDL-C levels greater than this threshold rec- fainting, and a “weird” feeling among patients in the pla- ommendation, no CHD risk was associated with this cebo group were judged “not study related” by the par- finding. ent or family physician. Mean±SD ratios for TC/HDL-C were 3.86±1.20 for The docosahexaenoic acid–supplemented group re- the placebo group compared with 3.91±1.50 for the doco- ported 1 case of eructation that was related to capsule sahexaenoic acid–supplemented group. Maintaining a TC/ consumption. Additional reported events included acne HDL-C ratio of greater than 6.4 is a risk factor for coro- (n=2), headache (n=2), ear infection (n=1), fatigue nary disease.30 Patients whose lipoprotein lipid profile (n=1), and weight gain (n=1). These events were judged indicated numerous increased risk factors for CHD (TC/ by the parent or family physician to be consistent with HDL-C ratio Ͼ6.4, LDL-C level Ͼ160 mg/dL [Ͼ4.14 mmol/ normal events of adolescence. L], and TG level Ͼ250 mg/dL [Ͼ2.82 mmol/L]29-31) were informed of these findings and advised to consult their fam- COMMENT ily physician for follow-up testing. An elevated TC/ HDL-C ratio was used as the primary risk indicator, plus Supplementation with a low dose of docosahexaenoic acid, an elevated LDL-C level and/or an elevated TG level. By 400 mg/d, for a 4-year duration was not associated with using these criteria, 4 patients (2 in each group) were iden- any identifiable safety risk. Plasma docosahexaenoic acid tified for physician follow-up. levels were successfully elevated by 2.5-fold in the docosahexaenoic acid–supplemented group, and remained at ADVERSE EVENTS baseline levels for the placebo group. Both groups reported minor adverse events judged by the parent or phy- Patients were requested to report all adverse events sician not to be study related as well as minor gastroin- throughout the trial; subsequently, several individuals testinal symptoms (eructation and flatulence) associated (n=5) reported more than 1 event. Adverse events were with capsule consumption. No group differences were recorded for 6 subjects from the placebo group (age range, observed for plasma antioxidant capacity, vitamin A con-

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1275

140 200 180 60 120 180 130 50 100 160

80 TC Level, mg/dL

TG Level, mg/dL 40 140 80 HDL-C Level, mg/dL LDL-C Level, mg/dL

120 30 30 60 Placebo Docosahexaenoic Placebo Docosahexaenoic Placebo Docosahexaenoic Placebo Docosahexaenoic Group Acid – Supplemented Group Acid – Supplemented Group Acid – Supplemented Group Acid – Supplemented Group Group Group Group

Figure 4. Mean±SD year 4 values for plasma lipoprotein total cholesterol (TC) (A), triglycerides (TGs) (B), high-density lipoprotein cholesterol (HDL-C) (C), and low-density lipoprotein cholesterol (LDL-C) (D) in the placebo group (n=21) and the docosahexaenoic acid–supplemented group (n=23). Brackets represent the mean±95% confidence interval for the age- and sex-matched control subjects (n=20) (TC, 163±15 mg/dL; TGs, 95±25 mg/dL; HDL-C, 51±5 mg/dL; and LDL-C, 94±14 mg/dL). There were no significant differences between the placebo and the docosahexaenoic acid–supplemented groups (PϾ.70). To convert HDL-C, LDL-C, and TC to micromoles per liter, multiply by 0.02586; and to convert TGs to micromoles per liter, multiply by 0.01129.

centration, whole blood platelet aggregation, liver func- jor ␻6 pathway intermediate, were reduced by 20% in tion enzyme activity, and plasma lipoprotein lipid pro- supplemented patients compared with the placebo group. files, although both study groups had lower platelet Although ARA is an essential substrate for prostanoid pro- aggregation and vitamin E values than the laboratory ref- duction and integral in the regulation of immune re- erence group. Plasma vitamin E analysis exhibited a trend sponses and inflammation, reductions in this pathway toward lower mean concentrations of this antioxidant vi- are generally associated with benefits to biological sys- tamin in the docosahexaenoic acid–supplemented group tems.35 Whereas ␻6-derived eicosanoids are involved in at the 3- and 4-year intervals. Although these reduc- promoting inflammatory and allergic responses, ␻3- tions never reached statistical significance, future supple- derived eicosanoids are antagonistic because of less po- mentation trials, especially those of long duration and/or tent metabolic properties.35,36 Therefore, increasing the elevated dosages of LCPUFAs, should consider contin- amount of available ␻3 fatty acids not only results in the ued surveillance of vitamin E concentrations. competitive inhibition of ␻6 fatty acid metabolism but The docosahexaenoic acid–supplemented cohort also reduces the ␻6 substrate available for prostanoid pro- maintained mean plasma docosahexaenoic acid levels of duction. Metabolic properties associated with fatty ac- 3.91% of total fatty acids during the 4 years of supple- ids and eicosanoids from both pathways are essential for mentation. This amount of plasma docosahexaenoic acid maintenance of health; therefore, adopting a balanced in- is well below physiological plasma levels of populations take of ␻3 and ␻6 fatty acids seems desirable. However, consuming large amounts of fish, particularly cold- because Western diets typically have an abundance of ␻6 water fish (salmon, tuna, mackerel, and sardines). For series fats, increased dietary intake of ␻3 series fats (eg, example, Alaskan Eskimos who consume food enriched cold-water fish) or nutritional supplementation with ␻3 in ␻3 PUFAs had plasma docosahexaenoic acid levels of series fatty acids (eg, ␣-linolenic acid, EPA, and doco- 5.22%.32 Clinical trials of healthy adults supplemented sahexaenoic acid) may be needed to attain a balance of with relatively high dosages of docosahexaenoic acid these fats. (DHASCO oil) have not been associated with adverse The mean total ␻6 PUFA/␻3 PUFA blood lipid bal- safety concerns. Innis and Hansen33 used a dosage of 7.5 ance achieved in the docosahexaenoic acid– g/d of DHASCO (2.9 g/d of docosahexaenoic acid) for supplemented group in this trial was 8:1, compared with 14 days to increase docosahexaenoic acid levels greater 13:1 in the placebo group and 12:1 in the age- and sex- than 2.5-fold in plasma phospholipids (from 2.1% to matched controls. The consumption of PUFAs has been 6.1%). Biosafety assays, including liver enzymes, lipo- persistently indicated as beneficial to reducing the risk protein cholesterol, and bleeding time tests, did not iden- of CHD. Although a specific ratio of ␻6/␻3 has not been tify adverse effects associated with this dosage. A study34 established to obtain this benefit, a low ratio is desir- of blood clotting variables and platelet aggregation in men able.37 The Japan Society for Lipid Nutrition recom- supplemented with 15 g/d of DHASCO (6 g/d of doco- mended a dietary ␻6/␻3 ratio of 2:1 or less for preven- sahexaenoic acid; 76 mg/kg of body weight of docosa- tion of chronic disease in an elderly population (ie, those hexaenoic acid per day) for 90 days found no identifi- with CHD and cerebrovascular disease).38 Furthermore, able changes in these properties. maintenance of a balanced ␻6/␻3 intake of 2:1 or 1:1 is The total amount of LCPUFA and the ratios of the considered a target goal in the infant nutrition field.39 The major ␻6 and ␻3 LCPUFAs in the blood are key indexes dietary ratio of ␻6/␻3 consumed in the United States is in assessing fatty acid status. Patients with XLRP receiv- 10.6:1; while this reflects a reduction from previous es- ing docosahexaenoic acid exhibited a corresponding in- timates (12.4:1), the ratio is still much higher than decrease in total ␻3 series LCPUFAs and a reduction in the sirable.40 Alaskan Eskimos, whose diet is rich in fish and ␻6 series LCPUFAs. Mean plasma levels of ARA, a ma- marine mammals, have a plasma ␻6 PUFA/␻3 PUFA ra-

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1276

©2003 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 tio of 3.5:1.32 Interestingly, lower rates of cardiovascu- sual function equivalent to that of breastfed infants.52-54 lar disease and atherosclerosis have been well docu- Dietary recommendations for docosahexaenoic acid (spe- mented in this native population.41,42 cifically EPA plus docosahexaenoic acid) intake for healthy The importance of docosahexaenoic acid in modu- adults have been set at a minimum of 650 mg/d by the lating visual function and in maintaining a healthy retina International Society for the Study of Fatty Acids and Lip- remains undetermined. Suh et al43 demonstrated that ids.55 A maximum tolerated dose of EPA plus docosa- supplementing rats with ARA plus docosahexaenoic acid hexaenoic acid of 300 mg/kg of body weight per day has altered the fatty acid composition of developing photo- been determined for humans.56 Recently, combined doses receptors, elevated the rhodopsin content of the retina, of 875 mg/d of EPA plus docosahexaenoic acid were ben- and altered the postbleaching kinetics of rhodopsin. These eficial in reducing myocardial infarctions, stroke, and mor- results indicated that structural changes in components tality in adult patients with cardiovascular disease,57 and of the retina (ie, increased membrane unsaturation in rod dosages of 40 mg of EPA plus docosahexaenoic acid per outer segments) are associated with an increased effi- kilogram of body weight per day have been reported to ciency of rhodopsin function. However, the susceptibil- result in improvements in patients with rheumatoid ity to light-induced oxidative damage of a docosahexae- arthritis.58 noic acid–laden retina has been demonstrated. Organisciak The results of this randomized, placebo-controlled, et al44 reported that rats receiving an ␻3 fatty acid supple- clinical trial demonstrate that long-term supplementation ment (linseed oil containing the docosahexaenoic acid with docosahexaenoic acid, 400 mg/d, increased blood doco- precursor, ␣-linolenic acid) had elevated levels of docosahexaenoic acid levels by 2.5-fold in patients with XLRP sahexaenoic acid in their rod outer segments and were and was not associated with any identifiable safety risks. more susceptible to light-induced retinal damage com- Monitoring biosafety, particularly plasma vitamin E lev- pared with animals raised on ␻3 fatty acid–deficient chow. els, at increased docosahexaenoic acid dosages and/or for Subsequently, Anderson et al20 hypothesized that ani- a longer duration would be prudent. mals and humans undergo an adaptive response to metabolic oxidative stress to protect the retina from further Submitted for publication July 2, 2002; final revision re- damage. Such actions include up-regulating antioxi- ceived April 3, 2003; accepted April 17, 2003. dant defense mechanisms and metabolically minimiz- This study was supported by grant FD-R-001232 from ing the tissue’s susceptibility to lipid peroxidation by re- the Orphan Products Development Program of the US Food ducing retinal docosahexaenoic acid levels.45 and Drug Administration; grant EY05235 from the Na- It remains to be determined whether docosahexae- tional Institutes of Health, Bethesda, Md; and the Founda- noic acid supplementation in humans leads to lipid per- tion Fighting Blindness. The study capsules were donated oxidation and subsequent membrane damage or if by Martek Biosciences Corporation. docosahexaenoic acid confers a protective effect.46 We thank the patients and families; the retinal spe- Docosahexaenoic acid has been shown to specifically cialists affiliated with the Texas Retina Associates, Dallas delay the onset and slow progression of apoptotic pho- (Rand Spencer, MD, and Gary Fish, MD) and the Founda- toreceptor cell death in vitro.47 Reinboth et al48 demon- tion Fighting Blindness (Gerald Fishman, MD, Chicago, Ill; strated that light elicits the in vitro release of docosa- John Heckenlively, MD, Los Angeles, Calif; Samuel Jacob- hexaenoic acid from photoreceptor phospholipids in son, MD, PhD, Philadelphia, Pa; Richard Lewis, MD, Hous- rats, and proposed that this release may serve a protec- ton, Tex; Paul Sieving, MD, PhD, Ann Arbor, Mich; and tive role in the retina by suppressing ARA-derived eico- Richard Weleber, MD, Portland, Ore) for referring pa- sanoids associated with inflammatory responses. In a tients for participation in this study; Argye Hillis, PhD, for dietary supplementation study49 of young rats fed high statistical advice during the trial; Katherine Franke, MS, levels of docosahexaenoic acid (ethyl ester; 9.69% of Maia Lapus, BS, and Scott Urban, BS, for their technical total calories), there was no increased tendency for in vivo assistance in conducting the biochemistry safety assays; and hydroperoxide formation in the rod outer segment mem- members of the Data and Safety Monitoring Committee branes of supplemented animals. These and other inves- (Gerald Fishman, MD [chair]; Norman Salem, Jr, PhD; Jo- tigations20,50 suggest that susceptibility to light damage hanna Seddon, MD; and Barbara Philippon, RN, MS) for of retinal tissues is multifactorial, and may be related more their time and effort invested throughout the conduct of this to antioxidant potential than LCPUFA status. study. The efficacy of this long-term supplementation trial Corresponding author and reprints: Dianna H. Whea- with docosahexaenoic acid, 400 mg/d, to retard disease ton, MS, Retina Foundation of the Southwest, 9900 N progression was inconclusive as measured by cone ERG Central Expressway, Dallas, TX 75231 (e-mail: dwheaton functional loss.19 Nevertheless, patients who attained the @retinafoundation.org). highest docosahexaenoic acid dosage (10-23 mg/kg of body weight per day) demonstrated reduced disease progression. The dose used in this trial was comparable to REFERENCES that received by breastfed infants (12-48 mg of docosahexaenoic acid per kilogram of body weight per day; cal- 1. McColl AJ, Converse CA. Lipid studies in retinitis pigmentosa. Prog Lipid Res. culated from milk compositional data51). Infants receiv- 1995;34:1-16. 2. Hoffman DR. Fatty acids and visual dysfunction. In: Chow CK, ed. Fatty Acids in ing 23 mg of docosahexaenoic acid per kilogram of body Foods and Their Health Implications. New York, NY: Marcel Dekker Inc; 2000: weight per day from docosahexaenoic acid and ARA– 817-842. enriched formula had no adverse events and had a vi- 3. Hoffman DR, Birch DG. Docosahexaenoic acid in red blood cells of patients with

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1277

©2003 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/29/2021 X-linked retinitis pigmentosa. Invest Ophthalmol Vis Sci. 1995;36:1009-1018. 30. Kinosian B, Glick H, Garland G. Cholesterol and coronary heart disease: predict- 4. Gong J, Rosner B, Rees DG, Berson EL, Weigel-DiFranco CA, Schaefer EJ. Plasma ing risks by levels and ratios. Ann Intern Med. 1994;121:641-647. docosahexaenoic acid levels in various genetic forms of retinitis pigmentosa. In- 31. NIH Consensus Development Panel on Triglyceride, High-Density Lipoprotein, vest Ophthalmol Vis Sci. 1992;33:2596-2602. and Coronary Heart Disease. NIH consensus conference: triglyceride, high- 5. Schaefer EJ, Robins SJ, Patton GM, et al. Red blood cell membrane phosphati- density lipoprotein, and coronary heart disease. JAMA. 1993;269:505-510. dylethanolamine fatty acid content in various forms of retinitis pigmentosa. J Lipid 32. Parkinson AJ, Cruz AL, Heyward WL, et al. Elevated concentrations of plasma Res. 1995;36:1427-1433. ␻-3 polyunsaturated fatty acids among Alaskan Eskimos. Am J Clin Nutr. 1994; 6. Hoffman DR, Wheaton DH, Locke K, Birch DG. Rod ERG function and erythro- 59:384-388. cyte docosahexaenoic (DHA) in X-linked retinitis pigmentosa (XLRP) [abstract]. 33. Innis SM, Hansen JW. Plasma fatty acid responses, metabolic effects, and safety Invest Ophthalmol Vis Sci. 1997;38:S797. of microalgal and fungal oils rich in arachidonic and docosahexaenoic acids in 7. Hoffman DR, Wheaton DH, Locke K, Birch DG. Docosahexaenoic acid (DHA) in healthy adults. Am J Clin Nutr. 1996;64:159-167. red blood cells and electroretinographic (ERG) function in carrier heterozygotes 34. Nelson GJ, Schmidt PS, Bartolini GL, Kelley DS, Kyle DJ. The effect of dietary of X-linked retinitis pigmentosa (XLRP) [abstract]. Invest Ophthalmol Vis Sci. docosahexaenoic acid on platelet function, platelet fatty acid composition, and 1998;39:S725. blood coagulation in humans. Lipids. 1997;32:1129-1136. 8. Martinez M. Tissue levels of polyunsaturated fatty acids during early human de- 35. Okuyama H, Kobayashi T, Watanabe S. Dietary fatty acids—the N-6/N-3 bal- velopment. J Pediatr. 1992;120(pt 2):S129-S138. ance and chronic elderly diseases. Prog Lipid Res. 1996;35:409-457. 9. Fliesler SJ, Anderson RE. Chemistry and metabolism of lipids in the vertebrate 36. Lands WEM, Morris A, Libelt B. Quantitative effects of dietary polyunsaturated retina. Prog Lipid Res. 1983;22:79-131. fats on the composition of fatty acids in rat tissues. Lipids. 1990;25:505-516. 10. Salem NJ, Litman B, Kim H-Y, Gawrisch K. Mechanisms of action of docosa- 37. Hu FB. The balance between ␻-6 and ␻-3 fatty acids and the risk of coronary hexaenoic acid in the nervous system. Lipids. 2001;36:945-959. heart disease. Nutrition. 2001;17:741-742. 11. Deese AJ, Dratz EA, Dahlquist FW, Paddy MR. Interaction of rhodopsin with two 38. Okuyama H. Recommended LNA/LA ratio for the prevention of chronic elderly unsaturated phosphatidylcholines. Biochemistry. 1981;20:6420-6427. diseases [abstract]. Paper presented at: 88th American Oil Chemist’s Society An- 12. de Turco EB, Jackson FR, Parkins N, Gordon WC. Strong association of unes- nual Meeting and Expo; May 12, 1997; Seattle, Wash. terified [3H] docosahexaenoic acid and [3H-docosahexaenoyl] phosphatidate to 39. Koletzko B, Agostoni C, Carlson SE, et al. Long chain polyunsaturated fatty ac- rhodopsin during in vivo labeling of frog retinal rod outer segments. Neuro- ids (LC-PUFA) and perinatal development. Acta Paediatr. 2001;90:460-464. chem Res. 2000;25:695-703. 40. Kris-Etherton PM, Shaffer Taylor D, Yu-Poth S, et al. Polyunsaturated fatty acids in 13. Niu SL, Mitchell DC, Litman BJ. Optimization of receptor-G protein coupling by the food chain in the United States. Am J Clin Nutr. 2000;71(suppl):179S-188S. bilayer lipid composition, II. J Biol Chem. 2001;276:42807-42811. 41. Middaugh JP. Cardiovascular disease among Alaskan Natives, 1980-89. Am J 14. Mitchell DC, Niu SL, Litman BJ. Optimization of receptor-G protein coupling by Public Health. 1990;80:282-285. bilayer lipid composition, I. J Biol Chem. 2001;276:42801-42806. 42. Davidson M, Bulkow LR, Gellin BG. Cardiac mortality in Alaska’s indigenous and 15. Litman BJ, Mitchell DC. A role for phospholipid polyunsaturation in modulating non-Native residents. Int J Epidemiol. 1993;22:62-67. membrane protein function. Lipids. 1996;31(suppl):S193-S197. 43. Suh M, Wierzbicvi AA, Lien EL, Clandinin MT. Dietary 20:4n-6 and 22:6n-3 modu- 16. Fishman GA, Farber MD, Derlacki DJ. X-linked retinitis pigmentosa: profile of clini- lates the profile of long- and very-long-chain fatty acids, rhodopsin content, and cal findings. Arch Ophthalmol. 1988;106:369-375. kinetics in developing photoreceptor cells. Pediatr Res. 2000;48:524-530. 17. Birch DG, Anderson JL, Birch EE. Early abnormalities of rod function in children 44. Organisciak DT, Darrow RM, Jiang YL, Blanks JC. Retinal light damage in rats with X-linked retinitis pigmentosa. Clin Vision Sci. 1993;8:329-335. with altered levels of rod outer segment docosahexaenoate. Invest Ophthalmol 18. Birch DG, Anderson JL, Fish GE. Yearly rates of rod and cone functional loss in reti- Vis Sci. 1996;37:2243-2257. nitis pigmentosa and cone-rod dystrophy. Ophthalmology. 1999;106:258-268. 45. Penn JS, Anderson RE. Effect of light history on rod outer-segment membrane 19. Hoffman DR, Wheaton DH, Locke KG, Fish GE, Birch DG. Four-year outcomes composition in the rat. Exp Eye Res. 1987;44:767-768. from a randomized clinical trial of docosahexaenoic acid (DHA) supplementa- 46. Budowski P. The omega-3 fatty acid peroxidation paradox. Redox Rep. 1996;2: tion in X-linked retinitis pigmentosa (XLRP) [abstract]. Invest Ophthalmol Vis 75-77. Sci. 2003;44:E-abstract 4851. Available at: http://abstracts.iovs.org/cgi/content 47. Rotstein NP, Aveldano MI, Barrantes FJ, Roccamo AM, Politi LE. Apoptosis of /abstract/44/5/4851. retinal photoreceptors during development in vitro: protective effect of docosa- 20. Anderson RE, Maude MB, Alvarez RA, Acland G, Aguirre GD. A hypothesis to hexaenoic acid. J Neurochem. 1997;69:504-513. explain the reduced blood levels of docosahexaenoic acid in inherited retinal de- 48. Reinboth JJ, Clausen M, Reme CE. Light elicits the release of docosahexaenoic generations caused by mutations in genes encoding retina-specific proteins. Lip- acid from membrane phospholipids in the rat retina in vitro. Exp Eye Res. 1996; ids. 1999;34(suppl):S235-S237. 63:277-284. 21. Knapp HR. Dietary fatty acids in human thrombosis and hemostasis. Am J Clin 49. Wang J-Y, Saito M. Dietary supplementation of N-3 fatty acids and hydroperox- Nutr. 1997;65(suppl):1687S-1698S. ide levels in rat retinas. Free Radic Res. 2001;35:367-375. 22. Wheaton DH, Daiger SP, Birch DG. The Southwest Eye Registry: distribution of 50. Koutz CA, Wiegand RD, Rapp LM, Anderson RE. Effect of dietary fat on the re- disease types and mutations. In: Anderson RE, LaVail MM, Hollyfield JG, eds. sponse of the rat retina to chronic and acute light stress. Exp Eye Res. 1995;60: New Insights Into Retinal Degenerative Diseases. New York, NY: Kluwer- 307-316. Plenum; 2001:339-345. 51. Jensen RG. Lipids in human milk. Lipids. 1999;34:1243-1271. 23. Kyle DJ, Arterburn LM. Single cell oil sources of docosahexaenoic acid: clinical 52. Birch EE, Hoffman DR, Uauy R, Birch DG, Prestidge C. Visual acuity and the es- studies. World Rev Nutr Diet. 1998;83:116-131. sentiality of docosahexaenoic acid and arachidonic acid in the diet of term in- 24. Hoffman DR, DeMar JC, Heird WC, Birch DG, Anderson RE. Impaired synthesis fants. Pediatr Res. 1998;44:201-209. of DHA in patients with X-linked retinitis pigmentosa. J Lipid Res. 2001;42:1395- 53. Birch EE, Garfield S, Hoffman DR, Uauy R, Birch DG. A randomized controlled 1401. trial of early dietary supply of long-chain polyunsaturated fatty acids and mental 25. Hoffman DR, Uauy R, Birch DG. Metabolism of omega-3 fatty acids in patients development in term infants. Dev Med Child Neurol. 2000;42:174-181. with autosomal dominant retinitis pigmentosa. Exp Eye Res. 1995;60:279-289. 54. Hoffman DR, Birch EB, Birch DG, et al. Impact of early dietary intake and blood 26. Bui MH. Simple determination of retinol, ␣-tocopherol and carotenoids (lutein, lipid composition of long-chain polyunsaturated fatty acids on later visual de- all-trans-lycopene, ␣- and ␤-carotenes) in human plasma by isocratic liquid chro- velopment. J Pediatr Gastroenterol Nutr. 2000;31:540-543. matography. J Chromatogr B Biomed Appl. 1994;654:129-133. 55. Simopoulos AP, Leaf A, Salem NJ. Conference report: workshop on the essen- 27. Whitehead TP, Thorpe GHG, Maxwell SRJ. Enhanced chemiluminescent assay tiality of and recommended dietary intakes for omega-6 and omega-3 fatty ac- for antioxidant capacity in biological fluids. Anal Chim Acta. 1992;266:265-277. ids. J Am Coll Nutr. 1999;18:487-489. 28. Williams JC, Forster LA, Tull SP, Wong M, Bevan RJ, Ferns GAA. Dietary vitamin 56. Burns CP, Halabi S, Clamon GH, et al. Phase I clinical study of fish oil fatty acid E supplementation inhibits thrombin-induced platelet aggregation, but not mono- capsules for patients with cancer cachexia. Clin Cancer Res. 1999;5:3942-3947. cyte adhesiveness, in patients with hypercholesterolaemia. Int J Exp Pathol. 1997; 57. Stone NJ. The Gruppo Italiano per lo Studio della Sopravvivenze nell’Infarto Mio- 78:259-266. cardio (GISSI)-Prevenzione Trial on fish oil and vitamin E supplementation in 29. Summary of the second report of the National Cholesterol Education Program myocardial infarction survivors. Curr Cardiol Rep. 2000;2:445-451. (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cho- 58. Volker D, Fitzgerald P, Major G, Garg M. Efficacy of fish oil concentrate in the lesterol in Adults (Adult Treatment Panel II). JAMA. 1993;269:3015-3023. treatment of rheumatoid arthritis. J Rheumatol. 2000;27:2343-2346.

(REPRINTED) ARCH OPHTHALMOL / VOL 121, SEP 2003 WWW.ARCHOPHTHALMOL.COM 1278