March 4, 1999

Dr. F. Owen Fields
Wyeth Nutritionals International
145 King of Prussia Road
Radnor, PA 19087

Re: GRAS Notice No. GRN 000007

Dear Dr. Fields:

The Food and Drug Administration (FDA) is responding to your notice, dated August 27, 1998, that you submitted in accordance with the agency's proposed regulation, proposed 21 CFR 170.36 (62 FR 18938; April 17, 1997; Substances Generally Recognized as Safe (GRAS)). FDA received your notice on September 1, 1998, and designated your notice as GRAS Notice No. GRN 000007.

The subject of your notice is ARASCO (arachidonic acid-rich single-cell oil derived from a nontoxigenic strain of the soil fungus Mortierella alpina) and DHASCO (docosahexaenoic acid-rich single-cell oil derived from a nontoxigenic strain of the microalgal species Crypthecodinium cohnii) oils. Your notice informs FDA of the view of Wyeth Nutritionals International (WNI) that certain conditions of use of ARASCO and DHASCO are GRAS. Specifically, your notice states that WNI has determined, through scientific procedures, that ARASCO and DHASCO are GRAS when used in combination with each other or with other safe and suitable sources of arachidonic acid (ARA) or docosahexaenoic acid (DHA) for use in infant formulas that would be consumed by all formula-fed infants, whether of low birth weight or born at term. The final ratio of ARA (provided by ARASCO) to DHA (provided by DHASCO) in the infant formula would range between 1.0 and 2.0. The combined maximal level of the fatty acids ARA and DHA (i.e., the total of the amount of ARA provided by ARASCO and the amount of DHA provided by DHASCO) would be no more than 70 mg per 100 kcal, equivalent to 1.4 percent of total fatty acids in the infant formula. The combined "target level" of the fatty acids ARA and DHA is 50 mg per 100 kcal, equivalent to 1.0 percent of total fatty acids in the infant formula. According to your notice, the maximal level of use is approximately 1.4-fold higher than the "target level" to allow for analytical and manufacturing variance under commercial production conditions.

Because the subject of your notice is ARASCO and DHASCO as sources of ARA and DHA, FDA did not evaluate "other safe and suitable sources" of ARA and DHA that WNI would consider combining with ARASCO or DHASCO for use in infant formula. Therefore, the agency's comments are restricted to ARASCO and DHASCO as sources of ARA and DHA, respectively. FDA has evaluated the information in GRAS Notice No. GRN 000007 as well as other available data and information and has identified certain questions, discussed in more detail below, regarding the safety of ARASCO and DHASCO for their intended use in infant formula. In light of these questions, your notice does not provide a sufficient basis for a determination that the subject use of ARASCO and DHASCO is GRAS.

Data and information that WNI presents to support WNI's GRAS determination

You calculate that the estimated combined intake of ARASCO and DHASCO from an infant formula that contains ARASCO and DHASCO at the combined "target level" would be 146 mg/kg bw/day for preterm infants and infants up to one month old; 90 mg/kg bw/day for infants who are approximately 6 months old; and 56 mg/kg bw/day for infants who are approximately 10 months old.

Your notice describes detailed information about the identity of ARASCO and DHASCO, including published information about ARA and DHA as they occur in ARASCO and DHASCO; published and unpublished information about the source microorganisms for ARASCO and DHASCO; the method of manufacture of ARASCO and DHASCO, including published fermentation procedures and procedures for post-fermentation processing of material derived from the fermentations; and specifications for ARASCO and DHASCO. In addition, your notice cites references to support your understanding of the physiological role of the long chain polyunsaturated fatty acids (LCPUFAs) ARA and DHA, which are present as components of human milk. Your notice further states that a number of expert groups have recommended that these LCPUFAs be added to preterm infant formulas, term infant formulas, or both, to provide these physiologically active components of human milk to formula-fed infants. Your notice explains the basis for your conclusion that published and unpublished preclinical absorption studies indicate that the fatty acids ARA and DHA are efficiently absorbed from ARASCO and DHASCO and accreted into selected tissue sites.

Your notice contends that the presence of the LCPUFAs ARA and DHA in human milk establishes the safety of these fatty acid components of ARASCO and DHASCO. To address the safety of components of ARASCO and DHASCO that derive from their respective microbial sources, which have no prior use in food, your notice describes a series of published and unpublished preclinical safety studies of ARASCO and DHASCO. Based on unpublished genotoxicity studies conducted by the manufacturer of ARASCO and DHASCO, you conclude that ARASCO and DHASCO do not exhibit mutagenic potential. Based on your analysis of published and unpublished toxicological studies that tested ARASCO and DHASCO at doses that were as high as could be delivered without potentially disrupting the diets of the animals, you conclude that these oils, either singly or in combination, have no toxicologic potential under the conditions of the studies. As a confirmatory step, you used analytical and biochemical techniques and a standard bioassay to assess the potential of C. cohnii to produce known dinoflagellate toxins or other toxins and detected no such toxins.

In addition to preclinical animal studies, your notice summarizes three clinical studies that you conducted in preterm and term infants.

FDA's evaluation of the available data and information regarding ARASCO and DHASCO

FDA has evaluated the information that you discuss in your notice as well as other data and information that are available to the agency.

Data and information that are used to determine whether the use of ARASCO and DHASCO in infant formula is GRAS should be appropriate to establish that these ingredients are safe at the maximum level of their intended use, even if that maximum level is above the customary level of use. Accordingly, using your estimate of dietary exposure, FDA calculated that the estimated combined intake of ARASCO and DHASCO from an infant formula that contains ARASCO and DHASCO at the maximum level would be about 200 mg/kg bw/day for preterm infants and infants who are up to one month old, about 130 mg/kg bw/day for infants who are approximately 6 months old, and about 80 mg/kg bw/day for infants who are approximately 10 months old.

In FDA's view, multiple studies, including three of the published studies (which your notice identifies as references 51, 97, and 108) that you rely on to establish the safety of ARASCO and DHASCO, show treatment related effects that raise questions about the safety of ARASCO and DHASCO. In particular, these studies consistently show treatment related effects on the liver (mean relative liver weight and liver function tests), mean relative spleen weight, and hematological parameters in rats who consume ARASCO or a blend of ARASCO/DHASCO compared to rats who consume a high-fat control diet. In addition, in one unpublished study that is not included in your notice, oil droplets are reported in the mesenteric lymph nodes and the intestinal villi, and lipogranulomas are reported in the mesenteric lymph nodes in treated rats who consumed ARASCO/DHASCO, or high-dose level of ARASCO alone. In its evaluation of another unpublished 90-day study that is not described in your notice, FDA considered that a no observed effect level (NOEL) in the rat for combined exposure to ARASCO and DHASCO is 1100 mg/kg bw/day, based on increased mean relative liver weights, changes in liver function tests, mean relative spleen weights, and changes in hematological parameters of the mid- and high-dose ARASCO/DHASCO treated females and males. Given this NOEL, the estimated dietary intake by infants who would consume infant formula that contains the blend of ARASCO and DHASCO at your "target level" is 8- to 20-fold less than the NOEL seen in the rat, on a weight basis. At the maximum levels of use, this difference between the NOEL in the rat and the estimated dietary intake by infants is 5- to 14-fold.

In FDA's view, these effects in rats raise concerns because of the potential that such effects would occur in human infants. The authors of all three published studies that you rely on (i.e., your references 51, 97, and 108) have dismissed these effects without sufficient explanation. Likewise, your notice provides no explanation for why the effects observed in published studies in rats are not toxicologically significant in that species. Importantly, your notice provides no explanation for why such effects in rats are not of concern in human infants, particularly in preterm and young infants.

We acknowledge that dietary manipulation may cause some physiological effects or adaptation in some cases. However, it is possible that some of the observed effects are toxicological effects. We are particularly concerned about toxicological effects that could derive from constituents of the source microorganisms, which have no prior use as a source of food ingredients. As is the case with commonly consumed vegetable oils, these single cell oils have an unsaponifiable fraction that likely contains sterols. Some of these sterols may be novel sterols that are not present in commonly consumed vegetable oils. For example, a recent report (Ref. Shimizu et al.) identifies a novel sterol in M. alpina.

In addition to the treatment-related effects that FDA has seen consistently in multiple studies in rats who consume ARASCO or a blend of ARASCO/DHASCO compared to rats who consume a high-fat control diet (i.e., treatment-related effects on the liver, mean relative spleen weight, and hematological parameters), data and information in some unpublished studies raise further concerns. An unpublished subchronic study with an in utero phase indicated that there was a dose related increase in the number of litters with dead pups in the treated groups. Some studies conducted with ARASCO/DHASCO have reported changes attributed to possible effects on the kidney, although the reported changes are not consistent across different studies. We believe that these observations merit consideration, particularly because the sources of ARASCO and DHASCO have no prior use in food.

As mentioned, your notice summarizes three clinical studies that you conducted in preterm and term infants. We limited our evaluation of the summarized clinical studies to a consideration of whether the studies appeared to measure any clinical endpoints that bear on the question of the toxicological significance, in humans, of the effects that FDA has consistently seen in rat studies conducted with ARASCO or a blend of ARASCO and DHASCO. It is not apparent to us that these studies did in fact measure such endpoints. Therefore, we believe that the clinical studies already conducted do not address, and thus cannot resolve, the questions raised by the observed effects in the rat of ARASCO or a blend of ARASCO and DHASCO.

Conclusions

FDA has evaluated the information in GRAS Notice No. GRN 000007 as well as other available data and information and has identified certain questions regarding the safety of ARASCO and DHASCO for their intended use in infant formula. In light of these questions, it is our view that your notice does not provide a sufficient basis for a determination that the subject use of ARASCO and DHASCO is GRAS.

In light of the program responsibilities of the Office of Special Nutritionals (OSN) under section 412 of the Federal Food, Drug, and Cosmetic Act, the Office of Premarket Approval consulted with OSN during its evaluation of your notice and has shared with OSN its view that your notice does not provide a sufficient basis for a determination that the subject use of ARASCO and DHASCO is GRAS. If you would find it useful to discuss these issues with agency scientists, we are willing to meet with you.

In accordance with proposed 21 CFR 170.36(f), we have placed a copy of this letter, as well as a copy of the information in your notice that conforms to the information in proposed § 170.36(c)(1), in a file that is available for public review and copying in the public reading room of the agency's Freedom of Information Staff.

References Cited by Wyeth Nutritionals International

WNI Reference 51.

Wibert GJ. Burns RA. Diersen-Schade DA. Kelly CM. "Evaluation of single cell sources of docosahexaenoic acid and arachidonic acid: a 4-week oral safety study in rats." Food & Chemical Toxicology. 35(10-11):967-74, 1997

WNI Reference 97.

Boswell K, Koskelo EK, Carl L, Glaza S, Hensen DJ, Williams KD and Kyle DJ (1996) "Preclinical Evaluation of Single-cell Oils that are Highly Enriched with Arachidonic Acid and Docosahexaenoic Acid" Food and Chemical Toxicology 34:585-593.

WNI Reference 108.

Koskelo EK, Boswell K, Carl L, Lanoue S, Kelly C and Kyle D (1997) "High Levels of Dietary Arachidonic Acid Triglyceride Exhibit No subchronic Toxicity in Rats" Lipids 32(4):397-405.

Additional Reference

Shimizu, S., Kawashima, H., Wada, M., and Yamada, H., Occurrence of a novel sterol, 24,25-methylenecholest-5-en-3b-ol, in Mortierella alpina 1S-4. Lipids 27:481-483, 1992.

Content last updated by lsk on 1999-MAR-24
Hypertext last updated by rdb on 1999-JUN-04