Use of an Arginine-Restricted Diet to Slow Progression of Visual Lossin Patients With Gyrate Atrophy

Objective  To quantify the effect of long-term reduction of plasma ornithine levelsthrough adherence to an arginine-restricted diet on visual function in patientsof all ages with gyrate atrophy of the retina and choroid.

Methods  A long-term observational study was conducted on 27 patients with gyrateatrophy, 17 of whom elected to comply with the arginine-restricted diet and10 who were unable to comply. The mean rates of change in the electroretinogramcombined response, electroretinogram flicker response, and kinetic and staticperimetry were determined.

Results  After mean follow-up of 13.9 years for the patients on the diet and14.1 years for those not on the diet, the mean rates of change for the dietgroup compared with those of the no-diet group were statistically significantlyslower for all outcome measures (age-adjusted P<.05)except for static perimetry (P = .06).

Conclusions  Adhering to an arginine-restricted diet so as to lower the plasma ornithinelevel below an average of 5.29 to 6.61 mg/dL (400-500 µmol/L) will slowthe loss of function as measured by sequential electroretinography and visualfield examinations.

In 1991, Kaiser-Kupfer et al¹ reportedthat long-term reduction of plasma ornithine levels slowed the progressionof the chorioretinal degeneration in gyrate atrophy (GA) of the choroid andretina. To minimize variables such as genotype at the disease gene locus,genetic background, compliance with treatment, and age at the time of institutingtreatment, 2 sets of sibling pairs (younger than 10 years) were placed onan arginine-restricted diet. When the diet was begun, the younger siblingswere aged 2 years 10 months and 2 years 8 months; their older siblings wereaged 6 years 4 months and 9 years 4 months, respectively. For each pair, theretina of the younger sibling had minimal changes at the time treatment wasbegun, while the older sibling had already developed extensive chorioretinallesions. By the time the 2 younger siblings had both received the diet formore than half their lives and had reached the age at which the older siblingbegan the diet, the younger siblings had dramatically less ocular involvementthan their older siblings at comparable ages.

We have continued to observe the 2 sets of siblings on an arginine-restricteddiet. After 16 to 17 years on this diet, the younger sibling in each pair,who had received the diet at an earlier age, manifested a slower progressionof chorioretinal lesions than the older sibling. One younger sibling maintainedan average plasma ornithine level of 1.55 mg/dL (117 µmol/L), whilethe other younger sibling maintained an average of 1.84 mg/dL (139 µmol/L)(normal plasma ornithine level is 0.53-1.59 mg/dL [40-120 µmol/L]).Thus, it was concluded that if the diet was started at an early age, long-termreduction in plasma ornithine levels appeared to appreciably slow the progressionof the chorioretinal lesions. With respect to retinal function, a moderategradual decline in sensitivity in the central 30° visual field or a gradualreduction in the amplitude of the electroretinogram combined response wasobserved.²

In an effort to generalize this beneficial treatment effect to a larger,older, and more diverse population of patients with GA, 27 patients who wereenrolled in a long-term observational study were analyzed with respect totheir clinical course. The purpose of this article is to report observationsthat suggest that being on a diet with a sustained decrease in plasma ornithinelevel appears to slow progression of the chorioretinal degeneration in GA.

Twenty-seven patients were enrolled and encouraged to adopt an arginine-restricteddiet. At yearly follow-up visits, diet compliance was monitored by self-reportingand plasma ornithine levels were measured. At the same visits, the primaryoutcome measurements were taken: electroretinogram combined response, electroretinogramflicker response, Goldmann kinetic perimetry, and Humphrey automated staticperimetry. Both eyes were tested, but only data obtained from the right eyewere used for analysis. At the end of follow-up, patients were individuallyidentified as having adhered to the diet to some degree or as not having followedthe diet at all during the study period. Thus, there were 17 patients on thediet who were observed for a total of 236 patient-years and 10 patients noton the diet who were observed for 141 patient-years.

The methods of the primary outcome measures were presented in detailpreviously.² Briefly, the exponential modelof change proposed by Caruso et al³ impliesthat the log-transformed measures change linearly with time, and the resultantslope is the natural index of rate of change. Slopes for each patient weretherefore estimated from the simple linear regression of log (variable) overtime in years. The 2-sided hypothesis that there was no mean difference inslopes between these 2 groups (on and off diet) was evaluated by analysisof covariance to adjust for differences in age.

Table 1 summarizes patientcharacteristics. Mean follow-up time was comparable between the groups, butthe median age of the patients on the diet was younger at the end of the studyby about 15 years. The serial plasma ornithine measurements were averagedfor each patient, and the resulting means were averaged over the diet groups,weighted by each patient's follow-up time. By this index, ornithine levelsin the off-diet group were twice as high as those in the on-diet group. Thefemale-male ratio was essentially reversed: 14 (82%) of patients on the dietand 2 (20%) not on the diet were female.

The comparisons of primary outcomes are summarized in Table 2. Because the diet groups were discrepant in age, the meanslopes (ie, mean rates of change) of the outcome measures were adjusted forage by analysis of covariance. After adjustment, all were statistically significantlygreater for the off-diet group (P<.05) exceptfor static perimetry (P = .06).

In previous articles^1,2 ourgroup reported the beneficial effect of an arginine-restricted diet in 2 pairsof siblings younger than 10 years with GA. Those studies concluded that ifthe diet was started at an early age, long-term reduction in plasma ornithinelevels, especially in younger patients, appeared to appreciably slow the progressionof the chorioretinal lesions. With respect to retinal function, progressiveloss was slowed but to a lesser extent. These results are consistent withthose in a mouse model of GA in which long-term reduction in plasma ornithinelevel with an arginine-restricted diet prevented retinal degeneration.⁴

Because many patients with GA do not come to medical attention withthe correct diagnosis until sometime in the second or third decade of life,it is also important to determine whether sustained reduction of plasma ornithinelevel at this stage of life has a beneficial effect. To answer this question,we compared the progress of the disease in 2 groups of patients with GA ofall ages: those who were able to comply with the diet over several years,with average plasma ornithine levels weighted by length of follow-up of 4.47mg/dL (338 µmol/L) or about 6 times normal; and those who elected notto go on the diet or were poor compliers, with an average plasma ornithinelevel weighted by length of follow-up of 9.28 mg/dL (702 µmol/L) orabout 10 times normal. As judged by 3 of the outcome measures, the differencebetween those on the diet and those not on the diet was statistically significant.For the fourth measure, static perimetry, the estimated effect of diet wasconsistent in magnitude and direction with the other measures; its lack ofstatistical significance may be due to lack of statistical power.

A limitation of the observational study is the self-selection of thepatients into the diet groups. Consequently, there may be other factors thatcould play a role as confounding variables. Age was one such possible factor,but significant differences between the 2 groups mostly persisted after theeffect of age was removed. Although the sex distributions between the groupswere markedly disparate, there is no evidence to suggest that the course orseverity of GA is influenced by sex. Therefore, the difference in the proportionof women in the group is not likely to have played a role in the outcome ofthe study. The sex distribution in the 2 groups may merely reflect the factthat in our sample, women were more successful in complying with the dietthan men. The ornithine-δ-aminotransferase genotypes of the patientsin each group were analyzed and, other than in sibling pairs, no pattern wasobserved. Finally, the amount of residual ornithine-δ-aminotransferaseactivity in patients in each group was analyzed and appeared to be randomlydistributed in each group. We concluded that adhering to an arginine-restricteddiet so that the plasma ornithine level falls below an average of 5.29 to6.61 mg/dL (400-500 µmol/L) will slow the loss of function as measuredby sequential electroretinography and visual field examinations.

Correspondence: Muriel I. Kaiser-Kupfer, MD, 6016 Neilwood Dr, Rockville,MD 20852 (isolde@comcast.net).

Submitted for publication August 28, 2003; accepted January 7, 2004.