(Cite as: 63 FR 20450)


PROPOSED RULES

DEPARTMENT OF HEALTH AND HUMAN SERVICES

Food and Drug Administration

21 CFR Part 120

RIN 0910-AA43

[Docket No. 97N-0511]

Hazard Analysis and Critical Control Point (HACCP); Procedures for the Safe and
Sanitary Processing and Importing of Juice

Friday, April 24, 1998

*20450 AGENCY: Food and Drug Administration, HHS.
ACTION: Proposed rule.

SUMMARY: The Food and Drug Administration (FDA) is proposing to adopt regulations to ensure the safe and sanitary processing of fruit and vegetable juices and juice products. The proposed regulation, if adopted, will mandate the application of Hazard Analysis and Critical Control Point (HACCP) principles to the processing of these foods. HACCP is a preventive system of hazard control. FDA is proposing these regulations because there have been a number of outbreaks of illness, including some directly affecting children, associated with juice products and because a system of preventive control measures is the most effective and efficient way to ensure that these products will be safe. Elsewhere in this issue of the Federal Register, FDA is publishing a warning label proposal for packaged juice.

DATES: Submit written comments by July 8, 1998. For information on the proposed compliance dates for small businesses and very small businesses see the Supplementary Information section of this document.

Submit written comments on the information collection requirements by May 26, 1998.

ADDRESSES: Submit written comments to the Dockets Management Branch (HFA-305), Food and Drug Administration, 12420 Parklawn Dr., rm. 1-23, Rockville, MD 20857. Submit written comments regarding information collection to the Office of Information and Regulatory Affairs, Office of Management and Budget, New Executive Office Bldg., 725 17th St. NW., rm. 10235, Washington, DC 20503, Attention: Desk Officer for FDA.

FOR FURTHER INFORMATION CONTACT: Shellee A. Davis, Center for Food Safety and Applied Nutrition (HFS-306), Food and Drug Administration, 200 C St. SW., Washington, DC 20204, 202-205-4681.

SUPPLEMENTARY INFORMATION:
The agency proposes to make any final rule based upon this proposal effective 1 year after its date of publication in the Federal Register. However, by its terms, the final rule will not be binding on small businesses as defined in proposed s120.1(b)(1) until 2 years after the date of publication of a final rule in the Federal Register; and for very small businesses as defined in proposed s120.1(b)(2), the final rule will not be binding until 3 years after the date of its publication in the Federal Register.

I. Concerns With Juice

A. Microbial Outbreaks

The Seattle-King County Department of Public Health and the Washington State Department of Health reported on October 30, 1996, an outbreak of Escherichia coli O157:H7 infections epidemiologically associated with drinking a particular brand of unpasteurized apple juice, or juice mixtures containing unpasteurized apple juice, purchased from a coffee shop chain, grocery stores, and other locations (Ref. 1). A case was defined as hemolytic uremic syndrome (HUS) or a stool culture yielding E. coli O157:H7 in a person who became ill after September 30, 1996, after drinking the particular brand of juice within 10 days before illness onset. There were at least 66 cases of illness, with 14 cases of HUS and the death of one child, associated with this outbreak (Ref. 2). Cases occurred in British Columbia, California, Colorado, and Washington. E. coli O157:H7 isolates cultured from a previously unopened container of the particular brand of apple juice had a deoxyribonucleic acid (DNA) "fingerprint" pattern (restriction fragment length polymorphism) indistinguishable from case- related isolates (Ref. 1).
Various juices have been documented as vehicles for causing outbreaks from microorganisms. A 1967 outbreak from contaminated water added to orange juice concentrate affected approximately 5,200 persons and was caused by an unidentified virus and possibly other contaminants (Refs. 3 and 4). About 300 people became ill from Salmonella typhimurium in cider made from apples, including some that had been picked up from the ground in an orchard fertilized with manure, in a 1974 outbreak in New Jersey (Ref. 5). A 1991 outbreak of Vibrio cholerae was associated with coconut milk contaminated during manufacturing in Thailand (Ref. 6). There have been two Cryptosporidium outbreaks related to drinking apple cider, the first in Maine in 1993 and the other in New York State in 1996. In the first case, the apples used for cider came from trees near a cow pasture (Ref. 7), and in the second case, water used for rinsing came from a well that tested positive for coliforms (Ref. 8). In 1995 there was an outbreak in Florida that was caused by Salmonella hartford in unpasteurized orange juice (Ref. 9).
E. coli O157:H7 has been recognized relatively recently as a human pathogen and has been a source of a number of outbreaks related to juice. Thirteen and possibly 14 children had bloody diarrhea and developed HUS in Toronto, Canada, between September 15 and 25, 1980. The children's illnesses were associated with drinking fresh apple juice. The children's stools were examined for enteropathogenic E. coli, Campylobacter, Salmonella, Shigella, and Yersinia. None of these organisms were found. E. coli O157:H7 is the suspected causative organism. Conclusive testing for that organism was not done because E. coli O157:H7 was not recognized as a human pathogen before 1982 (Ref. 10).
A 1991 E. coli O157:H7 outbreak in southeast Massachusetts conclusively showed that fresh-pressed unpasteurized apple juice can transmit E. coli O157:H7 bacteria. In this outbreak, 23 individuals had diarrhea, 16 had bloody diarrhea, and 4 developed HUS (Ref. 11).
In Connecticut, a 1996 outbreak of E. coli O157:H7 illness was associated with drinking a particular brand of apple cider. There were 14 cases of illness (including 7 hospitalized), with 3 cases of HUS associated with the outbreak (Ref. 8).
There was a small outbreak of E. coli O157:H7 illness in Washington State in 1996 that was related to apple cider made at a church event. This outbreak occurred during the same time as the unpasteurized apple juice outbreak described in previous paragraphs. The apples were washed in a chlorine solution, but it was not reported how much chlorine was used. Six people became ill, but no estimate was given on how many people may have drunk the apple cider (Ref. 12).
FDA's recall data also provide evidence of microbial hazards in juice. There were 85 cases of illness in 1994 resulting in a recall of orange juice that had fermented and contained Bacillus cereus and yeast (Ref. 13).
State investigations provide additional evidence of microbial hazards in juice. A 1989 outbreak in New York was caused by the presence in orange juice of Salmonella typhi that originated from an infected worker and *20451 resulted in 69 illnesses with 21 individuals hospitalized (Ref. 14). The State of Washington reported that in 1993 one individual was hospitalized from home- made carrot juice found to contain Clostridium botulinum (Ref. 15). A 1993 Ohio outbreak caused by yeast or some other unknown toxicant in orange juice resulted in 23 illnesses (Ref. 16). A home-made watermelon drink contaminated with Salmonella spp. caused illness in 18 individuals in a 1993 Florida outbreak (Ref. 17). The State of Colorado reported two outbreaks of gastrointestinal illness from fresh squeezed orange juice at a mountain resort (Ref. 18). There were food handlers that were ill in both Colorado instances, and a virus was suspected as the causative agent.
The evidence shows that certain juices have been the vehicle for outbreaks of foodborne illnesses. Although fruit juice is acidic, and thus would generally be considered to inhibit the growth of most microorganisms, most juice-related outbreaks have been associated with fruit juices.

B. Illnesses From Nonheat-treatable Hazards

Illnesses that have been caused by hazards that can not be reduced to acceptable levels by heat treatments have also been associated with juice. Tin in canned tomato juice caused illness in 113 individuals in 1969 (Ref. 19). Soil nitrate had resulted in a high nitrate content in the tomatoes, and this high nitrate content accelerated detinning in the cans. In 1984, 11 persons became ill from consuming elderberry juice prepared by staff of a religious/philosophic group that contained poisonous parts of the plant (Ref. 20). A 1990 guanabana juice outbreak was caused by the presence of toxic guanabana seed material and caused illness in nine individuals (Ref. 21). A 1997 outbreak was caused by tin in pineapple juice (Ref. 22).
In 1992 an 18-month-old child with a blood lead level of 36 micrograms per deciliter (MUg/dL) was found in a routine county health department blood lead monitoring program. Investigation of this incident by the county health department revealed that the only significant source of lead exposure for this child was lead in imported fruit juice packed in 12-ounce, lead-soldered cans (Ref. 23). Analysis by the State health department of multiple flavors of the fruit juices in lead-soldered cans available to the child found lead levels ranging from 160 to 810 parts per billion (ppb). An exposure assessment performed by the county health department estimated that the child consumed about three cans of these fruit juices per day and estimated that the child's daily lead intake from these fruit juices was approximately 600 MUg/day (Ref. 23). As a result of this incident, FDA announced an emergency action level of 80 ppb for lead in fruit beverages (such as juices, nectars, and drinks) packed in lead-soldered cans (58 FR 17233, April 1, 1993). The agency subsequently banned the use of lead-soldered cans (60 FR 33106, June 27, 1995).
Recalls also provide evidence of nonheat-treatable hazards in juice. In 1988 a fruit punch drink was recalled because of the presence of tin caused by the acidity of the drink reacting with the tin coating of the cans (Ref. 24). The product had been packaged in the wrong container.
There were 10 recalls between 1990 and 1995 for fruit juice or beverages containing fruit juice because of the presence of food ingredients that were inadvertently added to the product, not declared on the label, or not suitable for the food. Food ingredients involved with these recalls were natamycin (Ref. 25), sulfites (Ref. 26), FD&C yellow No. 5 (Refs. 27 through 33), and salt (Ref. 34).
Five recalls between 1991 and 1997 were caused by improper sanitation procedures or faulty equipment. In 1991 sodium hydroxide from a clean-in-place system contaminated the caps of a citrus punch drink (Ref. 35). In 1992 three persons became ill, with 1 hospitalized, from a sodium hydroxide sanitizing agent that got into fruit drink product containers during cleaning (Ref. 36). In 1993 cracks in a heat exchanger allowed an orange flavored soft drink containing pear juice to come in contact with copper pipe fittings and, thus, to become contaminated with copper (Ref. 37). In 1994 milk was found in orange juice from filler lines that were not cleaned between milk and juice production (Ref. 38). In 1997 the presence of an alkaline cleaning solution in a berry juice caused gastrointestinal distress in several persons (Ref. 39).
Companies have recalled fruit drinks because pieces of glass or plastic were found in their products. The presence of glass in products is typically caused by the use of glass bottles, which can chip or shatter during the production process (Refs. 40 through 42). The plastic was present from the company's practice of draping plastic bags over the side of the bottle loading bin (Ref. 43).
One company recalled apple-prune juice and prune juice in 1996 because of unacceptable levels of lead (Refs. 46 and 47). The cause was contaminated imported prune juice.
In response to the establishment of maximum levels for patulin in apple juice by several foreign governments, FDA initiated a sampling survey to determine the levels commonly found in domestic and imported apple juice. Patulin may be present in juice made from moldy apples. In March 1997 the agency found inordinately high levels of patulin in apple juice from a processor in Washington State (Ref. 48). The level of patulin found in the product was sufficient to pose a health hazard, especially considering the fact that apple juice is commonly used by infants and young children (Ref. 49). All affected products that had left the plant had been used in the manufacture of fermented apple cider. Patulin could not be detected in fermented product, and it was assumed that the patulin was destroyed through the fermentation process.
Therefore, as the foregoing discussion reveals, the evidence demonstrates that juice and juice beverages are susceptible to chemical and physical hazards as well as microbiological hazards.

C. Underreporting

There is wide agreement that the laboratory-confirmed cases from outbreaks and recalls understate the actual number of juice-related cases, but no consensus exists on the size of the understatement. Individuals may not manifest all symptoms or have severe enough symptoms to necessitate medical attention. Medical personnel may simply treat an individual's symptoms without determining the underlying cause. The laboratory-confirmed cases only represent those cases where individuals sought medical attention, and where medical personnel performed the necessary testing and reported the case to a government agency.
While the actual number of juice-related illnesses is unknown, FDA has derived an estimate of the total number by multiplying the average number of laboratory-confirmed cases by factors that account for under-reporting. The factors are based on the relationships between annual outbreak cases and published estimates of the number of foodborne illnesses. For example, using these adjustment factors, it is estimated that the average 16 annual laboratory-confirmed cases of Salmonella represents 4,900 to 7,600 actual cases (Ref. 50). For E. coli O157:H7, an average 22 laboratory-confirmed cases per year may actually represent 2,200 to 4,300 total juice-related cases (Ref. 50). Therefore, the agency assumes that the *20452 actual number of illnesses from the outbreaks described in sections I.A and I.B of this document is much greater than the confirmed number of illnesses. (For a more complete discussion of these estimates, see the agency's preliminary regulatory impact analysis and Ref. 50)

D. Pesticides

Pesticides are usually applied to plants to combat insects, plant diseases, and weed growth to assist in the growth of the fruit or vegetable. A food is considered adulterated under section 402(a)(2)(B) of the Federal Food, Drug, and Cosmetic Act (the act) (21 U.S.C. 342(a)(2)(B) if pesticide residues are present above the Environmental Protection Agency (EPA) established tolerances, or if EPA has not established a tolerance for use of the pesticide on the particular plant. FDA annually monitors a wide variety of foods for pesticide residues.
In 1994 FDA sampled 1,411 domestic fruits and fruit products, including apple juice and other fruit juices, for pesticide residues and found that less than 1 percent were violative for being over tolerance and less than 1 percent were violative for having no tolerance (Ref. 51). None of the 122 samples of apple juice or 44 samples of other fruit juices were violative. Out of 1,795 samples of domestic vegetables and vegetable products tested, FDA found that less than 1 percent of samples were over tolerance, and that 2 percent were violative for having no tolerance.
FDA also tested 1,940 imported fruits and fruit products in its 1994 pesticide residue monitoring program. Less than 1 percent of the items tested were over tolerance and 3 percent were violative for having no tolerance. None of the 110 fruit juices sampled were violative. The agency sampled 2,460 imported vegetables and vegetable products and found that less than 1 percent were violative for being over tolerance and 4 percent for having no tolerance.
In its 1995 pesticide monitoring program FDA found less than 1 percent of 1,437 samples of domestic fruits and fruit products to be violative for being over tolerance and 1 percent to be violative for having no tolerance (Ref. 52). Of the 110 apple juices and 22 other fruit juices sampled, only a single apple juice sample was found to be violative, because of the presence of a pesticide with no established tolerance. Analysis of 1,585 samples of domestic vegetable and vegetable product produced results similar to the results found in 1994, i.e., less than 1 percent of samples were over tolerance, and approximately 2 percent were violative because there were no tolerances for the pesticide residues that FDA found.
The agency sampled 1,757 imported fruits and fruit products for pesticides in 1995 and found that less than 1 percent were violative for being over tolerance, and that 3 percent were violative for having no tolerance. Of the 19 apple juices and 52 other fruit juices tested, 2 apple juice samples were violative because they contained pesticides for which there were no established tolerances. The agency sampled 2,535 imported vegetables and vegetable products and found that 1 percent were violative for being over tolerance, and that 3 percent were violative for having pesticide residues for which there was no tolerance. Some of these samples contained both residues over tolerance and residues with no tolerance.
Although there are no documented outbreaks caused by unlawful pesticide residues, chronic exposure to pesticide residues that do not conform to EPA tolerances increase risks to the public health. Therefore, juice processors must determine whether the possible presence of unlawful pesticide residues is a hazard that is reasonably likely to occur.

E. FDA's Public Meeting

As a result of the October 1996 apple juice outbreak from E. coli O157:H7, FDA held a public meeting on December 16 and 17, 1996 (hereafter referred to as the juice meeting) (see notice of meeting (61 FR 60290, November 27, 1996)), to review the current science, including technological and safety factors, relating to fresh juices and to consider measures necessary to provide safe fruit juices to the public. Interested persons were given until January 3, 1997, to submit written comments on the notice. On January 2, 1997 (62 FR 102), FDA extended the comment period to February 3, 1997, in response to several requests for an extension.
The purpose of the juice meeting was to provide a forum for an information exchange on current industry practices for the production of juice products and on developments in the science underlying the production of safe juices. Experts from industry, academia, and the regulatory and consumer sectors presented information on illnesses and the epidemiology of outbreaks arising from contaminated juices; concerns about emerging pathogens; the E. coli O157:H7 outbreak in October 1996 caused by contaminated apple juice; procedures for processing juices; and new and existing technology to remove or decrease the number of pathogens or other contaminating microorganisms. Time was available for questions and comments from all attendees.
The meeting provided an opportunity to: (1) Consider how FDA's regulatory program for fresh juice and juice products should be revised,(2) discuss and exchange information on relevant safety issues, (3) to identify research needs where appropriate, (4) consider whether additional consumer education is necessary, and (5) consider whether other measures are needed to reduce the risk of future outbreaks of illness from juice.
FDA received over 180 comments from industry (with a number of these describing themselves as small businesses), consumers, consumer organizations, trade organizations, scientific/technical companies, academic institutions or organizations, State agencies, a local government agency, and members of Congress. Although most of the comments concerned apple juice specifically, many comments pertained to juices in general, and some referred only to citrus juices. Most comments were concerned with changes in processing to improve the safety of juices. Among the changes suggested were requiring pasteurization of juices, requiring HACCP, or establishing current good manufacturing practices (CGMP's) in juice processing. The agency has attempted to address the comments made at the meeting or submitted in response to the Federal Register notice in this proposal. If there are any significant concerns that the agency has not addressed, these concerns should be brought to the agency's attention in comments on this proposal.
The Fresh Produce Subcommittee (FPS) of the National Advisory Committee on Microbiological Criteria for Foods (NACMCF) attended the public meeting. The FPS met after the public meeting and made recommendations to the NACMCF. The NACMCF subsequently met to discuss the issues that were raised at the meeting. Based on information that was presented at the meeting and on the FPS's expertise, the full NACMCF made several recommendations (Ref. 53). The NACMCF stated that there are many aspects that affect pathogen control, such as agricultural practices; product handling; equipment used; growing location, including produce obtained from below ground (carrots), on ground (e.g., tree drops), or picked from trees; pH; acidulants; method of processing; degree of animal contact; refrigeration; packaging; and the distribution system. It stated that, in determining the best control mechanisms, it is important to remember that the conditions for *20453 microbial survival differ from those for growth. The NACMCF recognized that, while the risks associated with specific juices vary, there are safety concerns associated with juices, especially unpasteurized juices.
The NACMCF concluded that: (1) The history of public health problems associated with fresh juices indicates a need for active safety interventions, and (2) for some fruit (e.g., oranges), the need for intervention may be limited to surface treatment, but for others, additional interventions may be required (e.g., pasteurization of the juice).
The NACMCF recommended to FDA the use of safety performance criteria instead of mandating the use of a specific intervention technology. In the absence of known specific pathogen-product associations, the NACMCF recommended the use of E. coli O157:H7 or Listeria monocytogenes as the target organism, as appropriate. This recommendation was based on the premise that these organisms are two of the most difficult to control (i.e., by juice acidity or heat lethality), and that, by controlling them, other pathogenic organisms will likely be controlled. The NACMCF suggested that a tolerable level of risk may be achieved by requiring interventions that have been validated to achieve a cumulative 5 log reduction in the target pathogen or a reduction in yearly risk of illness to less than 10 super-5 , assuming consumption of 100 milliliters (mL) of juice daily.
In addition, the NACMCF stated that HACCP and safety performance criteria should form the general conceptual framework to ensure the safety of juices, and that control measures should be based on a thorough hazard analysis. The NACMCF also stated that validation of the process must be an integral part of this framework. The NACMCF recommended mandatory HACCP for all juice products, and that processors should implement and strictly adhere to industry CGMP's. The NACMCF also recommended industry education programs addressing basic food microbiology, the principles of cleaning and sanitizing equipment, CGMP's, and HACCP.
The NACMCF recommended further study in several areas:
(1) The efficacy of new technologies and intervention strategies for safety;
(2) The contamination, survival, and growth of pathogens on produce with or without breaks in skin, with or without areas of rot, and within the core;
(3) How produce becomes contaminated with human pathogens, including the relevant microbial ecology during production and processing of juice. In particular, the NACMCF stated that there is an urgent need for these types of studies on E. coli O157:H7 in apple juice;
(4) The baseline incidence of human pathogens on fruits and vegetables, particularly on those used in juice processing; and
(5) Labeling information needed for consumer understanding and choice of safer juices and juice products.
On the basis of all the testimony presented at the December 16 and 17, 1996, meeting, the NACMCF agreed that there is a need to understand the differences among all juice and juice products (e.g., citrus versus other). A significant problem identified by the NACMCF is that consumers presently do not have a means to clearly differentiate between unpasteurized and pasteurized products, and that terms used to refer to juice products do not always have universal meanings. For example, the term "cider" is perceived to be an unpasteurized product whereas the term "juice" is often perceived to be pasteurized.
The NACMCF also stated that traditional heat treatments given to juices and juice products have been designed to achieve shelf stability, to remove water (i.e., concentration), or to affect other quality-related factors, and that these treatments, commonly referred to as "pasteurization," are greatly in excess of a process needed to inactivate foodborne pathogens.
Because of the lack of sufficient data to evaluate the effectiveness of labeling statements as safety interventions or to inform consumer choice, the NACMCF stated that it could not strongly endorse labeling as an interim safety measure.
Although the NACMCF did not endorse labeling as an interim safety measure, elsewhere in this issue of the Federal Register FDA is proposing interim labeling measures for packaged juice. The agency sets forth its reasons for proposing to adopt these measures in that proposal.

II. Consideration of How to Address Problems

A. Current Regulation of Juice

FDA has established labeling regulations and standards of identity for a number of juices. 21 CFR 101.30 pertains to percentage juice declaration for beverages that contain fruit or vegetable juice. Common or usual name regulations for nonstandardized beverages that contain fruit or vegetable juice are found in 21 CFR 102.33. Standards of identity are found in part 146 (21 CFR part 146) for a number of fruit juices and beverages and in part 156 (21 CFR part 156) for tomato juice. The standard of identity for pasteurized orange juice (s146.140) states that "The orange juice is so treated by heat as to reduce substantially the enzymatic activity and the number of viable microorganisms." Pasteurized orange juice must be labeled as such.
In the 1997 Food Code, FDA articulated its policy regarding unpasteurized apple juice (Ref. 54). The code states that food establishments (e.g., nursing homes) that serve apple juice, apple cider, or other beverages that contain apple juice to segments of the population that are highly susceptible to disease (e.g., the elderly) should serve juice that has been pasteurized or that is in a commercially sterile, shelf-stable form, in a hermetically sealed container.

B. The Current Inspection System

Juice processors, like other food processors, are subject to periodic unannounced, mandatory inspection by FDA. This inspection system provides the agency with a picture of conditions at a facility at the time of the inspection. However, assumptions must be made about conditions at the facility before and after that inspection, as well as about important factors beyond the facility that have a bearing on the safety of the finished product. The reliability of these assumptions over the intervals between inspections can create questions about the adequacy of the system.
FDA's inspections are based, in part, upon its regulations on CGMP in the manufacturing, packing, or holding human food in part 110 (21 CFR part 110). For the most part, these regulations set out broad statements of general applicability to all food processing on matters such as sanitation, facilities, equipment and utensils, processes, and controls. HACCP-type controls are listed as one of several options available to prevent food contamination (s110.80(b)(13)(i)), but they are not integral to the controls outlined in the regulations.
The inspection and surveillance strategies that FDA uses ascertain a manufacturer's knowledge of hazards and preventive control measures largely by inference (i.e., based on whether a company's products are in fact adulterated, or whether conditions in a plant are consistent with CGMP). It is the manufacturer's responsibility to ensure that its products are in compliance with the act. However, in the face of new pathogens, such as E. coli O157:H7, and the risk of illness associated with these pathogens, especially for children, the elderly, and *20454 the immunocompromised, FDA tentatively concludes that, at least for juices, new measures to control microbial, chemical, and physical hazards are necessary to ensure that finished products comply with the act's standards.

C. Alternatives

Comments from the juice meeting suggested several alternatives to ensure that juice products are safe. These alternatives are discussed in sections II.C.1 through II.C.6 of this document along with their impact on the current situation with juice.

1. Increased Inspection

Continuous visual inspection of juice production is not a viable alternative because few hazards associated with juice are detectable through visual inspection.
Another possibility is to direct significant additional resources toward increasing the frequency of FDA's inspection of juice manufacturers, as well as increasing the agency's sampling, laboratory analysis, and related regulatory activities with respect to these products. While many samples of domestic and imported juice products are collected each year for analysis in FDA laboratories, and this sampling is designed to represent a broad range of products and to target known problems, the product sampled represents only a small fraction of the total poundage of the juice products consumed in this country. Substantially more expenditures would be needed to increase laboratory analyses to statistically significant levels.
Even if the funds for increased FDA inspection and increased sampling and analysis were available, this approach alone would not likely be the best way for the agency to spend its limited resources to protect the public health. Reliance on end-product testing involves a certain amount of inefficiency and enormous sample sizes and testing on a lot-by-lot basis are necessary to overcome that inefficiency. Therefore, this option has significant limitations.
Some comments from the juice meeting stated that juice safety would be improved through more local/State inspection rather than Federal inspection.
FDA agrees that more local/State inspection would help to ensure the safety of juices, particularly where because FDA lacks jurisdiction, there is no connection between the juice products and interstate commerce. However, FDA is not in a position to mandate that State and local regulatory agencies conduct additional inspections with their limited resources. Further, FDA cannot mandate that a State ensure that a firm is complying with FDA's regulations. Therefore, increased local/State inspection for juice is not an option upon which FDA can rely.

2. CGMP's

Many comments from the juice meeting urged the implementation of industry CGMP's or sanitation standards to increase the safety of juices. Some comments provided State rules, model CGMP's, or sanitation guidelines for FDA's consideration. Other comments stated that there is a need for more industry education on sanitation and hygiene.
CGMP regulations have a twofold purpose: (1) To provide guidance on how to reduce insanitary manufacturing practices and on how to protect against food becoming contaminated, and (2) to set out objective requirements that enable industry to know what FDA expects an investigator to find when he or she visits a food plant (51 FR 22458 at 22459, June 19, 1986). CGMP's consist generally of broad statements on sanitation, facilities, equipment and utensils, processes, and controls that are of general applicability to food processing. Therefore, FDA issuance of CGMP's for juice would be an approach that could assist manufacturers in the production of safe juices.
FDA encourages the juice industry to use CGMP's to help ensure the safety of their juices. As stated previously, the NACMCF recommended that processors implement and strictly adhere to industry CGMP's. However, the use of CGMP's alone may not be adequate to ensure that juices are safe because of the broad based nature of CGMP's. CGMP's are directed at plantwide operating procedures and do not concentrate on the identification and prevention of food hazards. Therefore, the agency tentatively concludes that CGMP's, although useful, will not be adequate, without additional measures, to ensure the safety of juices.

3. Mandatory Pasteurization
The majority of the comments from the juice public meeting pertained to pasteurization of juice. A number of comments urged FDA to mandate pasteurization or other equivalent treatment of fruit juice to ensure its safety. One person who commented reported that customers of his apple cider had not complained about a difference in flavor when he implemented pasteurization. One comment requested a 2-year grace period for small businesses before implementation if pasteurization were to be required. Another suggested that pasteurization be required for apple cider only if CGMP's and HACCP fail. One comment suggested that pasteurization be required only for apple juice, because of the difficulty in cleaning apples as compared to other fruits.
However, most comments opposed mandatory pasteurization of juices because of: (1) The expense of pasteurization equipment, (2) preference by some consumers for the flavor of unpasteurized over pasteurized juice, (3) the safety record of juices, and (4) degradation of nutritional value from heat treatment. Many comments from small businesses claimed that they would be forced to close their operations if pasteurization were required. Some comments also stated an economic need for the use of dropped apples ("drops"), with one recommending the use of only hand-picked (rather than machine-picked) drops. Other comments stated that the use of drops should be prohibited, at least in unpasteurized juices.
FDA is aware of the significant safety advantages of pasteurizing juice as well as of the reasons that some processors choose not to pasteurize their products. Pasteurization is a heat treatment used to kill the vegetative forms of specific bacteria in liquid or semi-liquid food products. Pasteurization is an effective and proven technology to ensure that juice does not contain pathogens. However, there may be other methods that are equally effective. Thus, the NACMCF recommended the establishment of safety performance criteria for appropriate target organisms rather than the establishment of a specific intervention technology. The NACMCF stated that safety performance criteria would be most effective.
For example, whole oranges with an intact skin may be processed so that pathogens on the surface of the fruit are destroyed. Because pathogens are not reasonably likely to be present in the interior of an orange, surface treatment could be adequate to ensure the safety of the juice. This example illustrates that if FDA were to mandate pasteurization, such action could have the effect of limiting the development of new technologies that are as effective as pasteurization in particular circumstances but less intrusive and less expensive.
Therefore, the agency tentatively concludes that relying on safety performance criteria, as recommended by the NACMCF, is an approach preferable to pasteurization. However, if the use of safety performance criteria *20455 does not significantly decrease the number of microbial outbreaks caused by juice, the agency may consider adopting a regulation that mandates pasteurization.
The agency disagrees with the comments that stated that it should require that apple juice be pasteurized because apples can be difficult to clean. FDA recognizes that pastuerization is a process that has been validated to meet NACMCF's recommendations. Manufacturers may be able to use other technologies and practices provided that their process is validated to achieve a 5-log reduction in the target pathogen. Therefore, reliance on safety performance criteria is a better long-term approach because it provides for the development of new technologies.
A number of comments at the juice meeting urged FDA to consider alternatives to pasteurization to increase the safety of juices. Alternatives suggested by the comments included extreme isostatic pressure, high pressure sterilization, ultra short time-heat exchanger processing, ohmic heating, aseptic packaging, modified atmosphere packaging, ultrafiltration, high temperature and high pH adjustment of wash-water, ultrahigh hydrostatic pressure, electric pulses, electromagnetic field, pulsed light, ultraviolet (UV) water treatment, UV treatment with photoreactivation, electron beam sterilization, irradiation, ozonated water treatment, microbiocidal additives (benzoate, sorbate), and pH adjustment. The comments recommended that sanitizers or ingredients for washes include use of chlorine, chlorous acid, chlorine with emulsifiers, trisodium phosphate, peroxyacetic acid, peracetic acid, or dimethyl dicarbonate.
The agency agrees that there may be a number of agents that can reduce the number of microorganisms present in juice. As the NACMCF recommended, a tolerable level of risk may be achieved by interventions that have been validated to achieve a cumulative 5 log reduction in the target pathogens or a reduction in yearly risk of illness to less than 10 super-5 , assuming consumption of 100 mL of juice daily. However, the NACMCF did not specify the manner in which this risk reduction should be accomplished, only the target that must be reached. In section IV.M of this document the agency will discuss its proposed approach as to how this performance standard will apply to juice.

4. Labeling

A number of comments suggested that labeling to distinguish pasteurized from unpasteurized juice would enable consumers to make an informed choice. One of the comments requested warnings to those "at-risk," one urged the publication of warnings in the newspaper, and another wanted labeling with no warning. Rather than labeling, one comment suggested point of sale information. One comment urged FDA not to require labeling to distinguish pasteurized from unpasteurized juices.
The NACMCF recommended research on labeling information needed for consumer understanding and choice of safer juice products. The NACMCF concluded that, while the risks associated with specific juices vary, there are safety concerns associated with juices generally, especially unpasteurized juices.
Labeling whether a product is pasteurized or unpasteurized is useful information that the agency encourages processors to place on labels. However, such labeling would not inform purchasers of unpasteurized product that children, the elderly, and the immunocompromised are "at-risk" from consuming the product. Without effective consumer education, the label statements "pasteurized" and "unpasteurized" are likely to have relatively little meaning to consumers and could even cause confusion because some consumers might select unpasteurized juice, considering it more "healthy" because it is less processed. Finally, a labeling requirement that focuses only on whether a product is pasteurized or unpasteurized does not take into account technologies other than pasteurization that are adequate to control pathogens, and, thus, such a requirement could be viewed as restricting the development of new technologies.
The agency outlined interim measures in a notice published August 28, 1997 (62 FR 45593), and elsewhere in this issue of the Federal Register, FDA is issuing a proposal on labeling for packaged juice. These labeling measures attempt to provide information on the risks that juice that has not been processed to control for pathogens poses to children, the elderly, and the immunocompromised. The agency is proposing that the labeling measures be superseded when these juice products are processed under adequate HACCP programs or are otherwise processed to destroy pathogens (e.g., pasteurization).
It is possible for firms that manufacture juice to control for pathogens. Labeling a product to alert consumers to possible harmful effects from its consumption must not substitute for a manufacturer adequately addressing those concerns during processing. FDA is reluctant to rely on labeling as a safety measure and does so only when its analysis of the countervailing factors reveals that, on balance, labeling provides the most reasonable approach to protecting the public health. Juice is a product that is typically consumed by children, as well as adults. Therefore, FDA tentatively concludes that, for juice, manufacturers need to implement controls for pathogens to ensure that their products are safe and not rely solely on labeling, except as an interim measure. FDA requests comment on this tentative conclusion.

5. Education

Other comments from the juice meeting suggested that education would increase the awareness associated with the safety of juices and of all foods. Some comments suggested that more industry education or training was needed. Other comments wanted more consumer education, especially for those at highest risk from foodborne disease.
The NACMCF recommended that the industry be educated on basic food microbiology, the principles of cleaning and sanitizing equipment, CGMP's, and HACCP. FDA agrees that industry education can serve a valuable role in controlling potential food hazards and encourages the industry to take an active part in educating its employees and utilizing up-to-date technologies. The agency will assist the industry in its education effort.
Concerning consumer education, the agency has launched several initiatives to inform consumers about the potential hazards presented by juice to at-risk individuals (see 62 FR 45593, August 28, 1997). However, no matter how extensive a consumer education initiative the agency undertakes, it is doubtful that consumer education will reach all at-risk consumers. Therefore, consumer education alone will not be adequate to inform the at-risk population of the potential hazards of consumption of juice that has not been processed to control pathogens. Given that effective processing methods are available, primary reliance needs to be placed on them to ensure the safety of juice.

6. The HACCP Option
Many of the attendees at the juice meeting urged FDA to mandate HACCP for juice processors, whereas others were opposed. A number of the attendees urged use of CGMP's together with HACCP. Some attendees at the juice meeting recommended that microbiological criteria or performance *20456 standards be used in addition to HACCP, with two suggesting a 5 log reduction for E. coli O157:H7.
The NACMCF concluded that HACCP and safety performance criteria can provide the general conceptual framework needed to ensure the safety of juices, and that validation of the HACCP plan for the juice process (i.e., ensuring that the process is adequate to control hazards) must be an integral part of this framework. The NACMCF stated that processors should establish HACCP control measures based on a thorough hazard analysis.
HACCP is a preventive system of hazard control that places the responsibility for identifying safety problems with the manufacturer. Use of the HACCP system means that a firm is engaged in continuous problem prevention and problem solving, rather than relying on facility inspections by regulatory agencies or consumer complaints to detect a loss of control. HACCP provides for real time monitoring to assess the effectiveness of control. A HACCP system put in place by a manufacturer for a particular facility is unique and must reflect the type of juice, its method of processing, its packaging, the facility in which it is prepared, and the intended consumers.
As discussed previously, there is sufficient evidence to demonstrate that there are significant problems with the presence of pathogens in some juice products. Pathogens in juice can be controlled by heat treatment. However, there may be other treatments that meet the same performance standard that are equally effective (e.g., multiple barriers, surface treatment of intact fruit). The use of a HACCP system provides flexibility to a processor to use alternative pathogen control methods and, thus, encourages the development of new technologies but does not dictate either their development or use. Moreover, not only is HACCP effective in controlling microbiological hazards, it also is effective in preventing chemical and physical hazards. Thus, HACCP is particularly well-suited for the juice industry given, as discussed previously, the range of hazards that must be addressed in processing juice.
The agency agrees with the comments that urged use of CGMP's together with HACCP. CGMP's form the foundation upon which a HACCP system is built. Therefore, CGMP's are integral to the HACCP approach.
Because there are significant concerns with the microbial safety of juices, HACCP systems must control pathogens. As will be discussed in section IV.M of this document, FDA is proposing a 5 log reduction in target pathogens, as the NACMCF recommended, as a necessary step in a HACCP plan for juice. Validation of a HACCP system must ensure that the process that is employed is adequate to control the relevant pathogens, in addition to chemical and physical hazards. Validation of performance standards consists of determining the ability of the pathogens in question to resist acid and other chemical or heat treatment and the ability of the process applied to overcome that resistance. The agency requests comment on this approach to safety performance criteria. FDA also requests comment on the benefits of requiring a general HACCP approach as opposed to those of specifically requiring pasteurization.

7. Alternative Approach

An alternative approach to mandating HACCP would be to draw a distinction between untreated apple cider and all other juices. Manufacturers of apple cider would be provided a permanent option choosing between labeling or implementing a HACCP program with a 5-log pathogen reduction. All juices other than untreated apple cider would be provided a permanent option of choosing between labeling, implementing a HACCP system, or achieving a 5-log pathogen reduction as discussed in section M of this document, entitled "Pathogen Reduction." The agency requests comments on this alternative approach to a mandatory HACCP program.

D. Decision to Propose HACCP

The evidence discussed in section I.A of this document shows that juices have been a vehicle for pathogens that have caused a number of foodborne illnesses. Pathogens can be controlled through heat treatment. Information set forth in sections I.B and I.D of this document, however, demonstrates that there are many hazards that can occur with juice and juice beverages that cannot be controlled through heat treatment. Although not all of the problems discussed in section I of this document are caused by hazards that could be considered reasonably likely to occur in many juice operations, through the use of HACCP programs, a firm can evaluate its process to determine if the problem could have been controlled.
As discussed in section I.E of this document, the NACMCF stated that HACCP and safety performance criteria can form the general conceptual framework needed to ensure the safety of juices. FDA has evaluated each of the seven alternatives that have been suggested for dealing with the problems with juice. While the agency finds that these alternatives are by no means mutually exclusive, FDA has tentatively concluded that a preventive system, such as HACCP, appears to offer the most effective way to control the significant microbial hazards, along with other hazards, that have become a problem with juice.
Increased inspection, while having some beneficial impact on the safety of juices, is resource intensive to the agency. Even if funds were available to the agency for this purpose, increased inspection would likely not be the best way for the agency to utilize its resources to protect the public health. It is ultimately the responsibility of manufacturers to ensure that their products are safe. A preventive approach, such as HACCP, on the other hand, enhances a processor's ability to make safe products because HACCP concentrates on examining all aspects of production, identifying hazards that are reasonably likely to occur in that production process, and establishing measures that will control or minimize those hazards. HACCP also enhances FDA's inspections because it allows the agency to inspect the production facility more efficiently and then to verify that the firm is operating in accordance with the firm's HACCP plan, and it provides some assurance that any problems that have occurred have been identified and appropriately addressed.
CGMP's, the second alternative to HACCP, are plantwide operating procedures. Although FDA supports the use of CGMP's, it tentatively concludes that use of CGMP's alone would not be sufficient to control the problems with juices because CGMP's do not concentrate on the identification and prevention of food hazards. Nonetheless, CGMP's are necessary to provide the foundation on which a HACCP system is built. Therefore, the agency tentatively concludes that, while CGMP's are important to a HACCP system, they are not an adequate alternative to HACCP.
Mandating pasteurization, the third suggested alternative to HACCP, would reduce many microbial hazards in juices but would eliminate the incentive to develop alternative methods (e.g., use of multiple barriers, surface treatment of fruit) that can accomplish the same purpose. FDA does not want to limit innovative approaches to achieving food safety. HACCP, on the other hand, allows and encourages firms to explore more technologically efficient and more cost-efficient ways of managing all of the hazards that they face. Moreover, pasteurization only controls microbial hazards. HACCP systems can control all *20457 food hazards that are reasonably likely to occur.
Labeling was also suggested as an alternative. FDA acknowledges that, from a public health protection standpoint, there are certain advantages to labeling. Elsewhere in this issue of the Federal Register, FDA is proposing to require certain labeling, in the form of a warning statement, for packaged juice products that have not been processed to control, reduce, or eliminate pathogenic microorganisms that may be present in such juices. Such labeling will serve to reduce the risk of foodborne illness. However, such reduction will occur only to the extent that consumers read and understand the labeling. Accordingly, the agency has tentatively concluded that mandating HACCP for most juice products will provide more comprehensive public health protection by greatly reducing the number of juice products that contain dangerous pathogens.
Importantly, manufacturers do have the ability to process juice to control pathogens. Labeling a product to alert consumers to possible harmful effects from its consumption is not a substitute for a manufacturer adequately addressing those concerns during processing. Juice is a product consumed by children, as well as by adults. FDA is reluctant to rely on labeling as a safety measure and does so only when its analysis of the countervailing factors reveals that, on balance, labeling provides the most reasonable approach to protect the public health. Here, a situation in which HACCP offers a real long-term solution to controlling, if not eliminating, hazards in juice, the agency tentatively believes that labeling is not a reasonable long-term approach. The agency is soliciting comment on the appropriateness of this tentative conclusion.
The fifth alternative to HACCP that was suggested is education. Industry education can play a valuable role in the production of safe juices. Consumer education can play an important part in consumer purchasing choices. However, education is only effective if people understand and use the information conveyed. Moreover, even an extensive education program may not reach all consumers. Conversely, mandatory HACCP would ensure that industry produces safe juice, and that the product that reaches consumers is safe.
For the foregoing reasons, FDA has tentatively concluded that HACCP represents the appropriate system of controls that is necessary for producing safe juice products. Therefore, FDA is proposing to add part 120 to its regulations to establish procedures for implementing HACCP systems for fruit and vegetable juices. As the agency did with seafood, it is proposing to issue these HACCP regulations under various sections of the the act, including, most significantly, sections 402(a)(1) and (a)(4) and 701(a) of the act (21 U.S.C. 371(a)).
Section 402(a)(1) of the act states that a food is adulterated if it bears or contains any poisonous or deleterious substance that may render the food injurious to health. Section 402(a)(4) of the act states that a food is adulterated if it has been prepared, packed, or held under insanitary conditions whereby it may have been contaminated with filth, or whereby it may have been rendered injurious to health. It is important to recognize that section 402(a)(4) of the act addresses conditions that may render a food injurious to health, rather than conditions that have actually caused the food to be injurious (see United States v. 1,200 Cans, Pasteurized Whole Eggs, etc., 339 F. Supp. 131, 141 (N.D. Ga. 1972)). The question is whether the conditions under which the food is processed and held are insanitary and may render the food injurious to health. The agency tentatively finds that, if a processor of juice products does not incorporate certain basic controls into its procedures for preparing, packing, and holding food, it is operating under insanitary conditions that may render the juice that is produced injurious to health and, therefore, adulterated under the act. Section 701(a) of the act authorizes the agency to adopt regulations for the efficient enforcement of the act.
The legal basis for mandating HACCP systems for juice processors is the same as that for seafood. Additional discussion of the legal basis is set out in the proposed rule (59 FR 4142 at 4150, January 28, 1994) and final rule (60 FR 65096 at 65098) for fish and fishery products.

E. Notice of Intent

FDA published a notice of intent on August 28, 1997 (62 FR 45593), that announced a comprehensive program to address the incidence of foodborne illness related to consumption of fresh juice and to address ultimately the safety aspects of all juice products. The agency invited comment on the appropriateness of its strategy to: (1) Initiate rulemaking on a mandatory HACCP program for some or all juice products, (2) propose that the labels and labeling of some or all juice products not specifically processed to prevent or eliminate the presence of harmful bacteria bear a warning statement informing consumers of the risk of illness associated with consumption of the product, and (3) initiate several educational programs to minimize the hazards associated with fresh juice. The agency stated that it would consider comments received within 15 days of publication of the notice prior to publication of any proposed rule.
Some comments on the notice suggested that FDA mandate HACCP only for fresh juice processors. One comment stated that HACCP should be mandated only for firms that process large quantities of fresh juice. Other comments supported mandatory pasteurization or equivalent treatment of juice, especially apple cider. One comment added that pasteurization and use of CGMP would preclude the need for the mandatory use of HACCP.
In section II.D of this document the agency has already discussed its reasons for proposing HACCP. The illnesses discussed in sections I.A and I.B of this document did not pinpoint problems related solely to fresh juice processors or to the amount of fresh juice that a firm produced. The comments have not provided any new information to alter the agency's tentative conclusion that HACCP is necessary to ensure the safe production of juice. However, FDA requests information on whether there are categories of juice that should be excluded from the proposed regulation.
FDA has reviewed all of the comments received within 15 days of publication of the notice and has determined that the comments provided no information that would cause the agency to conclude that this proposal is inappropriate. The agency has attempted to address these comments to the extent that they are relevant to this proposal. All comments received in response to the notice that address the issues in this proposal will be considered either in this proposal or in any final rule published in response to this proposal.

F. Fresh Produce Guidance

FDA, working with the U.S. Department of Agriculture (USDA) and the agricultural community, has developed voluntary good agricultural practice (GAP) and GMP guidance for fruits and vegetables that has been issued in draft for comment. The guidance, which is a science-based evaluation of risks, will address potential food safety problems throughout the food production and distribution system such as sanitation, worker health, and water quality. This voluntary guidance can be used by both domestic and foreign fresh fruit and vegetable producers to help ensure the safety of their produce.*20458

III. The HACCP System

The HACCP concept is a systematic approach to the identification and assessment of the risk (likelihood of occurrence and severity) of biological, chemical, and physical hazards from a particular food production process or practice and the control of those hazards. HACCP is a preventive strategy for food safety. Under it, the food producer develops a plan that anticipates and identifies the points in the production process where a failure would likely result in a food hazard being created or allowed to persist. These points are referred to as critical control points (CCP's). Under HACCP, identified CCP's are systematically monitored to ensure that critical limits (CL's) are not exceeded, and records are kept of that monitoring. Corrective actions are taken when control of a CCP is lost, including proper disposition of the food produced during that period, and these actions are documented. The effectiveness of HACCP is also systematically verified by the processor.
HACCP has been endorsed by the NACMCF as an effective and rational means of ensuring food safety. HACCP also is recognized in the international food safety community as the state-of-the-art means to ensure the safety and integrity of food. In particular, the Committee on Food Hygiene of the United Nations' Codex Alimentarius Commission (Codex) has endorsed the HACCP concept as a worldwide guideline. The European Union (EU) and other countries around the world have begun to require that foods produced within their borders be processed in a HACCP system. HACCP also is required for shipment of some foods (e.g., seafood) into EU countries.

A. Five Preliminary Steps of HACCP

The NACMCF recommends a process for developing a HACCP system that includes: (1) Assembling a HACCP team, (2) describing the food and its distribution, (3) identifying the intended use and consumers of the food, (4) developing a flow diagram, and (5) verifying the flow diagram (Ref. 55). These steps have been identified by the NACMCF as the "five preliminary steps" of HACCP. Although the agency is not proposing to mandate that processors use these preliminary steps, processors will greatly benefit from using these preliminary steps in developing their HACCP systems. The NACMCF advises that the preliminary tasks should be accomplished before the application of HACCP principles to a specific process (Ref. 55).

B. The Seven Principles of HACCP

The NACMCF has developed the following seven principles that describe the HACCP concept:

1. Conduct a Hazard Analysis

The first step in the establishment of a HACCP system for a food production process or practice is the identification of the hazards associated with the product. The NACMCF defines a hazard as a biological, chemical, or physical factor that may cause a food to be unsafe for consumption. The hazard analysis step should include not only a written identification of the hazard but a written assessment of the likelihood that the hazard will occur and its severity if it does occur. This analysis should also involve the identification of CCP's along with control measures for each identified hazard.

2. Determine the CCP's

A CCP is a point, step, or procedure at which control can be applied, so that a potential food hazard can be prevented, eliminated, or reduced to acceptable levels. Points in the manufacturing process that may be CCP's include heat treatment, chilling, specific sanitation procedures, product formulation control, prevention of cross contamination, and certain aspects of employee and environmental hygiene.

3. Establish Critical Limits

This step involves establishing parameters that must not be exceeded for each control measure associated with a CCP. Critical limits (CL's) can be thought of as boundaries of safety for each CCP and may be set for control measures such as temperature, time, physical dimensions, moisture level, water activity, pH, and available chlorine. A CL is used to distinguish between safe and unsafe operating conditions at a CCP. For example, the minimum temperature and time combination that will kill pathogens in a heat treatment step is the CL for that CCP.

4. Establish Monitoring Procedures

Monitoring is a planned sequence of observations or measurements to assess whether a CCP is under control (i.e., operating within its CL) and to produce an accurate record of the monitoring for use in future verification procedures. An unsafe food may result if a process is not properly controlled and a deviation occurs. Because of the potentially serious consequences of a CL deviation, monitoring procedures must be effective. Continuous monitoring is possible with many types of physical and chemical methods. When it is not possible to monitor a CL on a continuous basis, monitoring intervals must be established that are frequent enough to permit the manufacturer to determine whether the step/process/procedure designed to control the hazard is working.

5. Establish Corrective Actions

While the HACCP system is intended to prevent deviations in a planned process from occurring, total prevention can rarely, if ever, be achieved. Therefore, there needs to be a corrective action plan in place to fix or correct the cause of the deviation to ensure that the CCP is brought under control, to ensure that there is appropriate disposition of any food produced during a deviation, and to ensure that records are made of the corrective actions taken. Out of control situations should be used to identify opportunities for improvement of the process to prevent future occurrences.

6. Establish Verification Procedures

This process involves the application of methods, procedures, tests, and evaluations, other than monitoring, to determine the adequacy of, and compliance with, the HACCP system. The major infusion of science in a HACCP system centers on proper identification of the hazards, CCP's, and CL's and the institution of proper verification procedures.

7. Establish Recordkeeping and Documentation Procedures

This principle requires the preparation and maintenance of written HACCP records that list the hazards, CCP's, and CL's identified by the firm, as well as the monitoring, recordkeeping, and other procedures that the firm intends to use to implement the system. This principle also requires the maintenance of records generated during the operation of the HACCP system.

C. History of the Use of HACCP

1. HACCP for Fish and Fishery Products

On December 18, 1995, FDA published a final rule in the Federal Register (60 FR 65096) on procedures for the safe and sanitary processing and importing of fish and fishery products (part 123 (21 CFR part 123)) (seafood final rule). The regulations require that seafood processors develop, implement, and document sanitation control procedures and mandate the application of HACCP principles to the processing of seafood. The effective date for the seafood final rule was December 18, 1997.*20459
The regulations proposed herein are based on the seafood final rule with some modification to reflect the differences between seafood and juice products and to reflect recent developments in the application of HACCP. An extensive administrative record was compiled in the seafood proceeding. FDA is incorporating that record as support for the current proposal. Although the regulations proposed herein differ in some aspects from part 123, they are not intended to supersede or otherwise alter the seafood final rule.

2. Advance Notice of Proposed Rulemaking for the Development of HACCP for the Food Industry

In the Federal Register of August 4, 1994 (59 FR 39888), FDA published an advance notice of proposed rulemaking (ANPRM) requesting public comment about whether and how the agency should develop regulations that would establish requirements for a new comprehensive food safety assurance program, based on HACCP, for both domestically produced and imported foods. The agency stated its tentative view that, if such regulations were issued, they would enhance FDA's ability to ensure the safety of the U.S. food supply. FDA requested comments on a number of specific issues, as well as on all aspects of such a food safety program.

3. HACCP Pilot Programs

In addition to the ANPRM, FDA also published in the Federal Register on August 4, 1994 (59 FR 39771), a notice announcing that it intended to conduct a pilot program in which volunteers from the food manufacturing industry would use a HACCP system that FDA would audit. The pilot program was intended to provide information that FDA could use in deciding whether to propose to adopt regulations and in developing and implementing a regulatory system in which food manufacturers are required to perform the food safety aspects of their operations based on HACCP principles. In the notice, FDA invited individual firms that wished to participate in the program to submit letters of interest. Approximately 50 firms expressed initial interest in participating in the pilot program, and 11 firms were selected to participate. In 1997 FDA completed the pilot program at six firms and published a second interim report.

4. HACCP for Meat and Poultry

On July 25, 1996, USDA published a final rule (61 FR 38806) that, among other things, required that each meat and poultry establishment develop and implement written sanitation standard operating procedures (Sanitation SOP's) and a system of HACCP controls designed to improve the safety of their products. The effective date for the Sanitation SOP's was January 27, 1997, and for the HACCP regulations was January 26, 1998. FDA has reviewed the meat and poultry HACCP regulations and has incorporated portions of them as appropriate in the proposed HACCP regulations for juice.

D. Issues from the ANPRM

FDA received approximately 150 comments in response to the August 4, 1994, ANPRM. The comments represented the views of consumers, consumer organizations, health professionals, academicians, food industry officials, trade associations, and foreign, State, and local government agencies. The agency has attempted to address these comments to the extent that they are relevant to this proposal.
1. The agency asked in the ANPRM how the responsibility for food safety should be shared between the food industry and government. Comments generally agreed that the food industry is responsible for producing safe food products. All respondents on this issue recognized that the Government's role is to verify industry compliance with any applicable safety regulations.
FDA agrees that it is the manufacturer's responsibility to ensure that the food that it produces is safe, and that it is the Government's role to verify that manufacturers are fulfilling their responsibility. Through use of a HACCP system, both the firm and FDA are able to better fulfill their roles. The proposed regulation in part 120 underscores the division of roles. Under the proposed regulation, industry is charged with examining all aspects of production, identifying hazards that are reasonably likely to occur, and establishing measures that will control or minimize those hazards. HACCP records enable the agency to inspect the production facility more efficiently and to verify that the firm is operating in accordance with its HACCP plan. They also give the agency insight into whether any problems that have occurred have been identified and appropriately addressed.
It is important that the juice industry focus on its responsibility to produce safe food. Recent outbreaks evidence that some members of the industry have not kept up with the need to evaluate the hazards presented by juice and to design processes to address those hazards. Firms need to be aware of the emerging problems presented by their raw materials and to decide whether, and if so what, steps are necessary to address these problems. Firms may decide that it is necessary to incorporate a step designed to kill bacteria into their process (e.g., pasteurization), that there are alternative steps that they can take to ensure the safety of their products, or that, given the nature of the raw materials, no steps are necessary. Firms also need to monitor the process that they decide to employ to ensure that it is functioning adequately and appropriately. FDA notes that some firms have already addressed food safety concerns and have implemented HACCP systems.
Moreover, given the heightened concerns about these products, Government needs to be in a position to fulfill its role of verifying that industry is doing its job. Given the sporadic and variable way in which the problems that have been associated with juice arise, sampling and end-product testing of juice products will not enable it to do so. Other steps that will give Government insights into the production itself appear to be in order.
2. FDA requested comment in the ANPRM about the likelihood of occurrence of a hazard that would warrant HACCP-type control. Generally, the comments consistently identified two features that would characterize a properly formulated definition of likelihood: Processing conditions and nature of hazard. The majority of comments offered by the food industry stipulated that the necessary condition for likelihood of occurrence of the hazard appropriate to trigger HACCP control must not be speculative, as in worst-case scenarios, but be real, practical, and intrinsic to the processing or hazards demonstrably present for specific commodities. Several responses recommended that the question be referred to broadly based expert panels to establish the likelihood of risk.
According to the NACMCF, each potential hazard is evaluated based on the severity of the potential hazard and its likely occurrence (Ref. 55). Severity is the seriousness of the consequences of exposure to the hazard. Considerations of severity (e.g., impact of sequelae and magnitude and duration of illness or injury) can be helpful in understanding the public health impact of the hazard. Likelihood of occurrence of a hazard is generally judged based on processing experience, epidemiological data, and information in the technical literature.
The agency agrees with the comments that stated that the processing conditions and the nature of the hazard are key elements in assessing the *20460 likelihood of a hazard occurring. It would be futile for processors to attempt to control for every theoretical hazard because doing so would entail assessing hazards that the processor could not reasonably anticipate would actually occur. The assessment of the likelihood of risk of illness or injury to consumers should be practical for the specific commodity and not be speculative. For example, use of pesticides on fruits and vegetables is a common practice while these foods grow. The presence of pesticides on fruits or vegetables used to make juice is considered a hazard if: (1) The pesticide is not approved for use on the fruit or vegetable, or (2) it is found in amounts above its EPA established tolerance. If a pesticide is applied to fruits or vegetables in conformance with EPA regulations, and the appropriate period of time has elapsed between application and harvest, the presence of the pesticide is not considered to present a hazard that is reasonably likely to occur.
The agency disagrees that it should rely on broadly based expert panels to establish likelihood of occurrence of a hazard. Although such committees could provide insight into the issue, on balance, the insights that they would be likely to provide would not justify the expenditure of resources that convening such committees would require. However, interested persons are welcome to consider voluntarily the question and to submit the results of their consideration to the agency.
3. Comments on the ANPRM stated that because epidemiological studies consistently show that microbial pathogens are the most significant source of food hazards, issues such as pesticides, heavy metals, filth, physical contaminants, and others pale by comparison with the immediate health consequences of foodborne microbial pathogens. They stated that HACCP is best suited for preventing microbial hazards rather than physical or chemical hazards because CCP monitoring can be readily established in a timely fashion for pathogens and, particularly, for the unsanitary conditions that promote their growth.
The comments added that effects that result from events that occur after the food has left the processor's HACCP system are not controllable by the processor. The comments said that this fact is significant because food service establishments and the lack of consumer education have contributed to the majority of incidences of foodborne illness reported in current epidemiological data. They stated that HACCP systems are essentially localized management tools that will not permit any measurable improvement in national or international food safety effectiveness and have been implemented voluntarily solely as a corporate practice to provide strategic business advantages in increasingly competitive markets.
The comments stated that regulation may be premature because of the adequacy and feasibility of presently available analytical tests to control all hazards. They stated that, consequently, HACCP is an excellent tool but only in the very specific case of high-risk food processing that is focused on controlling microbiological risks. The comments stated that, instead of misdirecting its efforts, FDA needs to look to itself to reinforce food preparation safety awareness at food service establishments and to pursue vigorously an enhanced consumer education policy on unsafe food practices as the best preventative food risk control program.
FDA agrees that microbial hazards are a significant source of food hazards. FDA also agrees that HACCP is an ideal mechanism to deal with microbial hazards because it is a system of prevention. Prevention makes up for the inadequacies of end-product testing. For example, for maximum quality, nonshelf stable juice must be distributed quickly, and end-product testing usually takes at least several days to obtain results. If pathogens are discovered in the juice after distribution, the product must be recalled, and consumers may have already ingested product. Finally, the particular samples taken in end product testing may not contain pathogens because the pathogens may not be ubiquitous in the lot (i.e., there may be low level or sporadic contamination) and thus produce false negatives.
A system of preventive controls, like HACCP, on the other hand, is designed to identify and manage conditions where pathogens could be present in juice while it is still being processed. HACCP is designed to ensure that there is early discovery, and timely correction, of any problems that may develop. Although HACCP is well suited for preventing microbial hazards, this does not mean, as some of the comments asserted, that it is not useful for other types of hazards. As the NACMCF has recognized, it is well suited for preventing chemical and physical hazards. For example, processors can establish CCP's to prevent pieces of glass from contaminating a product when glass bottles are used.
The NACMCF endorses HACCP as an effective and rational means of assuring food safety (Ref. 55). According to the NACMCF, its use will likely result in measurable improvement in food safety. Under HACCP, processors view the processing plant from a prevention perspective and thus are in a position to react appropriately to new hazards if they arise. In preparing this proposal, FDA has reviewed the history of juice related outbreaks. All of these outbreaks might have been prevented if a HACCP system of the type that FDA is proposing herein had been in use.
The agency agrees that there are hazards that can occur after food has left the processing plant that the processor cannot control. The agency has established the Food Code to assist State agencies and food workers in retail food establishments and has addressed handling of high risk foods in the Food Code. FDA also provides consumer information on food safety through a consumer hotline, public affairs specialists in FDA's district offices, and various brochures and other publications. These efforts are intended to educate consumers on safe handling of foods at home. In addition, as described in the interim notice, the agency has initiated a consumer education program concerning juice that is not treated to prevent or eliminate the presence of harmful bacteria.
4. The agency requested information in the ANPRM on its possible role in assisting the food industry in the development of HACCP plans. Comments stated that FDA preparation of general background materials on HACCP would be beneficial in establishing a common approach to plan development, in assisting hazard identification analysis, and in using consistent language. They stated that FDA could provide informational resources such as examples of HACCP plans adaptable to the individual circumstances of a business' operations or consultative documents that could serve to guide plan development.
However, some comments urged that FDA avoid over-regulation. They stated that an excessively ambitious regulatory approach will limit the effectiveness of any HACCP program.
The agency agrees that it should avoid over-regulation because such an approach can inhibit future developments and new technology in HACCP systems and in safe food processing. FDA is proposing a HACCP regulation that, if adopted, will be mandatory for juice processors (as defined at proposed s 120.3(i)) but that can be used as a model for other foods in that it outlines the minimum essential components of a HACCP system. To the extent possible, the proposed regulation is in harmony with *20461 the existing HACCP regulations for seafood and meat and poultry.
FDA has developed the "Fish & Fisheries Products Hazards & Controls Guide" to assist manufacturers in the implementation of HACCP for seafood. The Federal Safety and Inspection Service (FSIS) has developed, in conjunction with the International Meat and Poultry HACCP Alliance, 13 HACCP models for meat and poultry products, a "Guidebook for the Preparation of HACCP Plans," and the "Meat and Poultry Products Hazards and Control Guide." However, it is not clear whether FDA will be able to provide such detailed information for juice. Therefore, in this rulemaking, the agency will attempt to provide guidance, to the extent possible, concerning the application of the regulation to juice.
5. Some comments on the ANPRM stated that, if EPA tolerances for pesticides in agricultural commodities become HACCP-focused safety issues in food processing and service industries, then explicit coordination by FDA with EPA is needed to define truly significant hazards. They stated that this effort would greatly assist HACCP development in such circumstances, so that duplication of effort would be avoided, consistency among regulatory requirements would be achieved, and impediments to international commerce would be removed.
FDA has attempted to harmonize its regulations with those of other Federal agencies and with Codex. EPA establishes regulations for pesticide use and tolerances for pesticide residues, and FDA and USDA enforce those tolerances on foods.
Under section 402(a)(2)(B) of the act, a food is deemed to be adulterated if it bears or contains a pesticide chemical residue unless a tolerance or an exemption for such pesticide has been established, and the quantity of such pesticide on the commodity is within the tolerance limits. Pesticide chemical residues for which there is no tolerance or exemption are deemed to be unsafe as a matter of law. HACCP is intended to protect against unsafe products. Thus, there is no reason why pesticide residues and similar types of food safety measures should be outside the scope of HACCP.
6. In the ANPRM, the agency asked if there was a need for microbiological criteria in HACCP regulations. Some comments favored inclusion of microbiological criteria for known high risk foods because such criteria are practical, efficient, and cost effective. However, most comments maintained that microbiological criteria, set as national standards, are not warranted because: (1) Criteria are discordant with HACCP purposes because they depend on end product testing, (2) criteria possess inadequate scientific basis, and (3) criteria are preemptive of localized development of HACCP systems.
The agency tentatively agrees with those comments that stated that microbiological criteria in HACCP regulations are warranted for some foods. Contrary to what many of the comments asserted, effective microbial controls depend not on end product testing but on processing controls and the establishment of CL's. For example, juice made from apples that have fallen on the ground must be processed in some manner to destroy pathogens because pathogens are likely to be present and, as discussed previously, end product testing may produce false negatives. If a regulation is flexible, it should not "preempt" the processor's development of HACCP, but it can provide the CL's needed for the safe processing of food under a HACCP system. However, the agency agrees that the decision on which processing controls are to be used must have a valid scientific basis.
Microbial pathogens have emerged as a significant problem in unpasteurized juice in recent years. The NACMCF recommended that safety performance criteria, rather than a specific intervention technology, be mandated for juice (Ref. 53). The safety performance criteria recommended by the NACMCF is whether the measures that a juice processor employs have been validated to achieve a cumulative 5 log reduction in the target organisms or a reduction in yearly risk of illness to less than 10 super-5 , assuming consumption of 100 mL of juice daily. As will be discussed in section IV.M of this document, FDA is proposing to require that firms include in their HACCP plans measures that will produce, at a minimum, a 5 log reduction in target pathogens.
7. Comments on the ANPRM stated that FDA should require end product testing records to provide information as to the effectiveness of a HACCP program. These comments stated that end product testing was practical because mandated testing was a necessary, continuing, and recordable validation of the completeness of a HACCP system, thereby ensuring that 100 percent control is manifested.
Comments from the juice meeting also supported the use of end product testing. One of the these comments proposed using testing to decide whether to pasteurize each lot. Several comments pointed to new rapid testing technologies and testing kits for pathogens.
However, other comments maintained that information generated from end product tests would not be useful. One comment stated that end product testing activities were counterproductive to a well-planned HACCP system. Furthermore, these comments added, any requirements that FDA puts forward must be practical, and no process can be regulated into 100 percent certainty.
The agency is not proposing to require end product testing. End product testing is most useful where there are high levels of the substance being tested, and there is uniformity throughout the lot being sampled. Product sampled for testing for microbial hazards, where a pathogen (e.g., E. coli O157:H7) is hazardous even at very low levels, or for physical hazards (e.g., glass), where the hazard is the presence of a discrete unit, may not contain the hazard even under the best sampling procedure. In these cases end product testing is likely to produce false negatives and, thus, to provide scant protection. It is prohibitive to use end product testing adequately in these situations because of the amount of testing that is necessary for a statistically valid test, and because it would be necessary to channel a significant portion of the product for that testing. Therefore, the agency has tentatively concluded that use of control measures under a HACCP system to prevent hazards from occurring, with subsequent monitoring, verification, validation, and recordkeeping, is more effective than end product testing in ensuring that food is safe. Thus, FDA has not included a requirement for end product testing in this proposed rule on juice products.
8. The agency asked in the ANPRM whether it should mandate HACCP for all segments of the food industry. Many comments stated that mandatory HACCP regulations for low-risk foods would be inappropriate because trying to manage low risk hazards through HACCP would dilute agency resources and therefore the effectiveness of HACCP. The comments stated that FDA could utilize its resources most efficiently by focusing on those high-risk food processing operations identified in its 1993 model Food Code as "Potentially Hazardous." They stated that the U.S. food supply is already demonstrably the world's safest, so that there is no valid reason for requiring HACCP plans of the entire industry. The comments stated that enforcement mechanisms in the act are, and will continue to be, sufficient without adding to the regulatory burden on *20462 industry. They added that incorporation of HACCP into food industry operations should be permitted to proceed on a voluntary basis, unless a well-defined need requires implementation through specific authority provisions of the act into specific high-risk segments of the food industry.
However, some comments stated that unless all segments of the food chain are mandatorily included, adoption of HACCP is unlikely to result in measurable enhancement of the safety of the food supply. They stated that less than universal coverage would create confusion about what should be excluded. The comments stated that any attempt to limit HACCP to identified "high-risk" processors would hinder efforts to address significant public health problems that may arise in the future. They concluded that it is not unduly burdensome to mandate HACCP for all. The comments maintained that HACCP regulations should be as comprehensive as practicable and applied throughout the food chain to the fullest extent possible and reasonable, and that HACCP principles must be applied from farm to fork.
FDA disagrees with the comments that stated that HACCP is inappropriate for low-risk foods. Both food processors and government regulatory agencies would benefit from the use of HACCP systems. The U.S.'s excellent record for having a safe food supply does not mean that this country should not consider ways of improving on that record. In the face of emerging pathogens and other new food hazards, HACCP provides a flexible system in which processors reassess their procedures on an on-going basis. HACCP also enables processors to meet future demands.
The use of HACCP allows food processors to concentrate their efforts on the aspects of the processes that they use where risks are highest and provides regulatory agencies with assurance that processors are observing prudent processing practices. HACCP also provides assurance that problems in the process are likely to be discovered, and that unsafe product is unlikely to leave the firm. The complexity of HACCP is a function of the number of hazards that must be controlled and the nature of the controls for each hazard. Foods that involve few hazards will tend to have fewer CCP's, and, conversely, those that have multiple hazards will tend to have more complex HACCP plans and monitoring requirements.
FDA is proposing a regulation that will mandate HACCP for juices. The agency has tentatively concluded that there is a safety basis to require that processors use HACCP systems in the processing of juice. As the agency gains experience and additional information from the pilot program and from seafood HACCP implementation, it will examine the appropriateness of expanding the scope of proposed part 120 (if the agency adopts it) to include other foods. Clearly, the agency will consider HACCP's use with foods that it has identified as presenting likely hazards, as it is doing in this proposal.
In developing the proposed regulations for juice, FDA came to recognize that the elements of a HACCP regulation for juice are really no different than those for seafood. This insight suggests that part 120 can act as a model for HACCP for other parts of the food industry should the agency become aware of facts that would justify extending the coverage of the regulation. Firms that are interested in voluntarily instituting HACCP can use the regulations in part 120 as a guide for doing so.
9. The ANPRM requested information on the criteria that FDA should use in deciding whether to cover some or all segments of the food industry with a mandatory HACCP rule. Some of the comments stated that exclusions cannot be justified on the basis of business size because about 75 percent of the food industry would be considered to be small businesses. The comments asserted that exclusions can only be judged with respect to properly defined risks for the food hazards involved in producing the end-product.
FDA agrees that exemptions from HACCP regulations cannot be justified on the basis that a business is small because food hazards that are reasonably likely to occur in the production of most foods occur regardless of the size of the firm. The agency also agrees that any exceptions to mandatory HACCP systems must be based on instances in which risks are not reasonably likely to occur. However, FDA is required by law to consider ways to assist small businesses when it implements regulations. While FDA does not propose to exempt any small businesses from the food safety requirements in this proposed rule, FDA is considering ways to provide regulatory options that will serve to reduce the burden of compliance on such small businesses.

IV. FDA's Proposal

A. Applicability

1. Scope

The agency tentatively concludes that HACCP is necessary for the safe and sanitary production of fruit and vegetable juices to address the special concerns discussed previously. Therefore, FDA is proposing new s120.1(a), which states that part 120 applies to juice and defines what juice means for purposes of this regulation.
Fruit and vegetable juices may be used as ingredients in other beverages (e.g., flavored bottled waters; juice beverages and cocktails). These products often resemble juices, are processed in a manner that is similar to juices, and handled by consumers similarly to juices. Thus, they can present the same food hazards as juices. Therefore, FDA is proposing to require that any juice sold as such or used as an ingredient in beverages be processed in accordance with the requirements of part 120.
As stated in section II of this document, FDA has established standards of identity for a number of fruit juices in part 146 and for tomato juice in s 156.145. These standardized juices are generally described as the liquid extracted or expressed from a fruit or vegetable. However, prune juice (s146.187) is prepared from a water extract of dried prunes.
A typical dictionary definition of the term "juice" is a fluid naturally contained in plant or animal tissue (Ref. 56). As described above, the present situation has demonstrated a need to control food hazards associated with fruit and vegetable juices. The present situation does not include oil extracts of fruits and vegetables (e.g., olive oil) because these are not traditionally considered juice. Some juices (e.g., banana juice) and fruit nectars, when purees of the fruit used, need to be included in any definition FDA proposes because such purees are often blended with other juices. If there are food hazards associated with extractives of a fruit or vegetable, those food hazards will be present in purees of that fruit or vegetable. Concentrates of juice and purees also need to be included in the definition because, if a hazard is present in the juice or puree, it could also likely be present in the juice concentrate. Therefore, the agency is tentatively defining "juice" as the aqueous liquid expressed or extracted from a fruit or vegetable, purees of the edible portions of a fruit or vegetable, or any concentrates of such liquid or puree.
The agency requests comments on the definition of "juice." FDA also requests comments on the scope of the regulation and on whether it should mandate HACCP for all types of juices, or whether it would be sufficient to mandate HACCP for certain types of juices.*20463

2. Effective Date

The seafood final rule provided processors 2 years to implement HACCP. This was done to: (1) Allow time for training of industry personnel and regulatory personnel; (2) provide the States with the time to have a full opportunity to understand and respond to the effects of these regulations; (3) increase the likelihood that more agreements with other countries will exist; (4) increase the opportunity for processors to engage in "voluntary" HACCP inspections in advance of the effective date to obtain preliminary, informal feedback from the agency on their progress; and (5) allow incorporation of modifications made in the final rule and publication of FDA assistance materials for the seafood industry (60 FR 65096 at 65169).
The period of time between publication of the final rule and the effective dates of the HACCP regulations for meat and poultry issued by FSIS are: (1) Eighteen months for large establishments with 500 or more employees, (2) Thirty months for smaller establishments with 10 or more employees but fewer than 500, and (3) Forty-two months for very small establishments with fewer than 10 employees or annual sales of less than $2.5 million (61 FR 38806).
A comment from a fresh juice trade association submitted to the agency in response to the NACMCF recommendations to FDA on the safety of juices, requested that FDA mandate HACCP for all juice products and phase this requirement in over a 3-year period from the publication of the final rule in a manner similar to the FSIS HACCP regulation. The comment requested that FDA consider annual inspections of fresh juice firms until the regulation is effective. It stated that the delay in implementing HACCP requirements would allow FDA and juice processors the ability to review conclusions of specific research and establish performance standards based on this research.
Comments on FDA's notice of intent (62 FR 45593) generally supported a phased- in approach for small firms taking 3 to 4 years. However, one comment expected that a phase-in approach would take no more than 2 years.
The agency is considering the significant issues surrounding orderly implementation of HACCP. FDA must balance the need for immediate implementation of HACCP, because of its associated food safety benefits, against the costs of implementation and consider options to minimize the burden to small businesses. The proposed timeframe for implementation of these regulations attempts to balance these competing concerns. The implementation of HACCP may be more burdensome for small firms than for large firms. Large firms tend to have quality control personnel already in place. In addition, many regulatory requirements are less burdensome for a large firm in proportion to output than they are for a small firm.
FDA recognizes that HACCP systems cannot be developed and implemented overnight. The HACCP system of controls can involve new ways of thinking and performing on a routine basis.
The agency issued a notice on August 28, 1997 (62 FR 45593), that provided interim measures, and elsewhere in this issue of the Federal Register, FDA is proposing to require labeling for juice to address the agency's immediate public health concerns. If finalized, these measures will require labeling on juice to provide information that juice unprocessed to control pathogens poses risks to children, the elderly, and the immunocompromised. The agency is proposing that the labeling measures be superseded once packaged juice products are processed under adequate HACCP programs, or are otherwise processed in a manner to destroy pathogens (e.g., pasteurization). Therefore, as proposed, before the applicable effective date, juice will be processed to control for pathogens or, if not, will bear labeling to alert consumers that such processing has not occurred. After any applicable effective date, processors will use HACCP systems in the production of juice.
The agency has considered the precedents established by other HACCP regulations and the comments submitted on juice. There are two significant differences between the HACCP regulation that FDA is proposing for juice and the HACCP regulations for seafood and for meat and poultry. First, FDA has issued interim guidance suggesting that juice that has not been processed to control pathogens be labeled accordingly. Elsewhere in this issue of the Federal Register, the agency is proposing to require such labeling. Second, at the present time, FDA's available resources would make it very difficult, if not impossible, to implement a comprehensive inspection program for the entire juice industry. A phased in approach for compliance will thus ease the burden not only on small businesses but also on the agency itself. Accordingly, FDA is proposing that the regulations proposed herein generally be effective 1 year after the date of publication of the final rule, with special provisions that will extend the phase-in to up to 3 years after publication of the final rule. This proposed phase-in approach will permit the regulated industry time to accomplish the training of personnel and adjust its activities to include necessary HACCP activities and takes into account the needs of smaller businesses.
The agency proposes to establish a timetable for phasing in HACCP based on business size. FDA proposes in s120.1(b) that the effective date be 1 year following publication of the final rule. The agency is proposing that, by its terms, the regulation will not be binding until 2 years following the date of publication of the final rule for small businesses employing fewer than 500 persons (s120.1(b)(1)). This is based on the definition of a small business used by the Small Business Administration. In addition, the agency is proposing that, by its terms, the regulation will not be binding until 3 years following the date of publication of the final rule for very small businesses that have either total annual sales of less than $500,000, or that have total annual sales that are greater than $500,000 but total annual food sales of less than $50,000, or that employ fewer than an average of 100 full-time equivalent employees and that sold fewer than 100,000 units of the product in the United States (s120.1(b)(2)). These criteria are consistent with those that the agency has used in its regulation on small firms and compliance with the nutrition labeling rules that implement the Nutrition Labeling and Education Act (the 1990 amendments) (61 FR 40963) (see s101.9(j)(1) and (j)(18)) (21 CFR 101.9(j)(1) and (j)(18)). In the 1990 amendments context, these criteria represent the outcome of three hearings in different parts of the country, an act of Congress, and informal rulemaking by FDA. Thus, FDA tentatively concludes that food manufacturers agree with and understand the definition of very small businesses. As discussed in the next section of this document, for purposes of this proposed rule, the agency has tentatively decided that a retail establishment as set out in proposed s120.3(h)(2)(iii) includes a very small processor that makes juice on its premises and directly sells this juice both to consumers and other retailers provided that total juice sales do not exceed 40,000 gallons per year.
In implementing proposed s120.1(b)(2), FDA intends to use the definitions for the terms "unit," "food product," "person," and "full-time equivalent employee" in *20464 s101.9(j)(18)(vi). These definitions are as follows: (1) "Unit" means the packaging or, if there is no packaging, the form in which a food product is offered for sale to consumers; (2) "food product " means food in any size package that is manufactured by a single manufacturer or that bears the same brand name, that bears the same statement of identity, and that has similar preparation methods; (3) "person" means all domestic and foreign affiliates, as defined in 13 CFR 121.401, of the corporation, in the case of a corporation, and all affiliates, as defined in 13 CFR 121.401, of a firm or other entity, when referring to a firm or other entity that is not a corporation; and (4) "full-time equivalent employee" means all individuals employed by the person claiming the exemption. The number of full-time equivalent employees is determined by dividing the total number of hours of salary or wages paid directly to employees of the person and of all of its affiliates by the number of hours of work in 1 year, 2,080 hours (i.e., 40 hours x 52 weeks).
FDA is committed to its mission of ensuring that food is safe and not misbranded. This commitment is the basis for proposing interim labeling measures. The agency tentatively finds that a phase-in HACCP implementation is necessary because of the logistical effort required to manage a fundamental change in work processes, roles, and responsibilities for smaller processors. The proposed implementation schedule reflects the abilities of processors of varying sizes to implement HACCP, and the time needed by industry to develop HACCP plans and train employees.
Upon the proposed implementation date, processors must be ready to operate their HACCP system, and FDA will conduct inspection activities according to HACCP principles to ensure that the HACCP system is operating acceptably. FDA requests comment on its proposed phased-in implementation of HACCP.

B. Definitions

FDA is proposing in the introductory paragraph of s120.3 that the definitions and interpretations of terms in section 201 of the act (21 U.S.C. 321), in s 101.9(j)(18)(vi), and in part 110 be applicable to such terms when used in part 120, except where they are redefined in s120.3.
The agency is proposing to include in s120.3 all definitions applicable to juice that are in the seafood HACCP regulation. The following terms have proposed definitions that are the same as their definitions in s123.3: "critical limit" (s120.3(d)), "food hazard" (s120.3(e)), "importer" (s120.3(f)), "shall" (s120.3(j)), and "should" s120.3(k)).
However, FDA is proposing to modify the term "preventive measure" to "control measure" (s120.3(b)) and to modify its definition from that used in the seafood HACCP regulation (s123.3(i)) to conform with recent NACMCF changes in terminology (Ref. 55). The term "control measure" is used because not all hazards can be prevented, but virtually all can be controlled to some degree. The new NACMCF definition describes the control measures as actions or activities rather than as chemical, physical, or other factors. Further, the term "control" is clarified to mean prevention, elimination, or reduction of hazards. The agency tentatively concludes that the recent NACMCF definition better describes the measures that processors must take. Therefore, FDA is proposing that "control measure" means any action or activity that can be used to prevent, eliminate, or reduce a hazard.
The NACMCF also recently modified its definition for "critical control point" (Ref. 55). The modified definition incorporates the new definition of "control measure" and emphasizes the essential or critical nature of the step. Thus, FDA tentatively concludes that the recent NACMCF definition better characterizes the term. Therefore, the agency is proposing in s120.3(c) that "critical control point" means a point, step, or procedure in a food process at which a control measure can be applied and at which control is essential to reduce an identified food hazard to an acceptable level.
The seafood HACCP regulation defines "processing" in s123.3(k) with specific product application. To apply these definitions to juice and to avoid listing specific processes, the agency is proposing in s120.3(h)(1) to define "processing" as activities that are conducted by a processor that are directly related to the production of juice products.
As with the seafood HACCP regulation, there are certain handlers of juice products that are not covered by the proposed definition. FDA has tentatively concluded that harvesting, picking, or transporting raw agricultural ingredients of juice products, without otherwise engaging in processing, should not be included in the term "processing" (s120.3(h)(2)(i)). FDA has developed voluntary GAP guidance that has been issued in draft for comment and will apply to these activities. The agency believes that growers will find GAP's useful and that the regulations that it is proposing in this rulemaking will, if adopted, reinforce use of both FDA and specific industry GAP's, thus affecting harvesting, picking, or transporting indirectly through processor and importer controls over raw materials and imported shipments (e.g., preventive controls such as the purchasing of raw materials only from farms that engage in proper handling of produce).
The agency notes that, with FSIS, it published an ANPRM (61 FR 59372, November 22, 1996) concerning transportation and storage requirements for potentially hazardous foods. In that ANPRM, FDA and FSIS requested information and comments on approaches that the two agencies should take to foster food safety improvements in the transportation and storage of potentially hazardous foods. While juice has not historically been considered a potentially hazardous food, recent illnesses associated with juice necessitate reconsideration of whether this food should not be included in that category. FSIS and FDA are reviewing the comments received in response to the joint transportation notice and will decide whether rulemaking is warranted. FDA invites comment on whether its approach to transportation is adequate.
The agency has also tentatively decided to exclude the operation of a retail establishment from the definition of "processing" (s120.3(g)(2)(ii)). For purposes of this rule, the agency has tentatively decided that a retail establishment as set out in proposed s120.3(h)(2)(iii) includes a very small processor that makes juice on its premises and directly sells juice to consumers and other retailers provided that total juice sales do not exceed 40,000 gallons per year.
FDA has traditionally refrained from directly regulating retail establishments, although it has authority to do so. FDA provides training and other forms of technical assistance to States and local governments who inspect retail food establishments through the agency's retail Federal/State cooperative program. A major part of that cooperative program involves the development of model codes, some of which have been widely adopted by States and local governments. FDA has consolidated those model codes into a single, updated food code for the retail sector. Appropriate controls are included in the food code that can be applied to address juice hazards at retail. FDA will continue to operate through the Federal/State cooperative mechanism and, consequently, has not proposed to regulate juice retailers in *20465 this proposal. However, elsewhere in this issue of the Federal Register, the agency is proposing to require labeling statements for packaged juice products including those sold by retailers that have not been pasteurized or otherwise processed to reduce, eliminate, or control pathogens. The proposed labeling requirement would apply to packaged untreated juice products produced in retail establishments for immediate consumption (such as grocery stores and very small processors) and would serve to inform consumers of the risk of untreated juices. (Retail processors selling unpackaged juice on-site for immediate consumption, such as restaurants and juice bars, would be exempt from both HACCP and labeling.) FDA notes that 2 of the outbreaks associated with apple cider (an outbreak of E. Coli. 0157:H7 infection and an outbreak of cryptosporidosis involving very small apple cider mills, refs. 8, 8A, and 11) would have fallen under the retail exclusion. Under the proposed labeling rule, the cider mills would have been required to label their apple cider. FDA seeks comment on whether the provisions of the food code in combination with the labeling statements will provide adequate public health protection. In addition, in formulating its proposal to include in the definition of retailer a processor that sells less than 40,000 gallons per year, the agency considered two other alternatives on which it requests comments. The first alternative would be to subject these establishments to the HACCP requirements and to provide a 3-year effective date. The second alternative would be to subject these establishments to the HACCP requirements and to provide a 5-year effective date. The agency is also soliciting comment on the appropriateness of including these establishments in the retail exemption as well as the appropriateness of the other two options considered.
The agency is proposing to define the term "control," even though it was not included in s123.3. FDA is proposing in s120.3(a), that "control" means to prevent, eliminate, or reduce. This definition is consistent with the use of the term "control" in the definition for "control measure" (s120.3(b)) and describes more specifically what is to be accomplished in the control of food hazards.
FDA is also proposing to define the term "monitor," even though it was not included in s123.3. FDA is proposing in s120.3(g) to define "monitor" as conducting a planned sequence of observations or measurements to assess whether a process, point, or procedure is under control and producing an accurate record of those observations or measurements for use in verification. This definition is identical with that of the NACMCF (Ref. 55). The agency tentatively concludes that defining this term will assist juice processors to be aware of what activities constitute monitoring of the various components of the HACCP system and prerequisite programs; and comply with the monitoring and recordkeeping requirements necessary for acceptable verification of HACCP.

C. CGMP's

Section 120.5 of the proposed regulations references the umbrella CGMP regulations in part 110 as providing general guidance to such matters as facility design, materials, personnel practices, and cleaning and sanitation procedures. Because part 110 provides guidance of general applicability to all foods, including juice, the agency intends that this guidance will continue to apply to juice processors even if FDA adopts the proposed regulations in part 120.

D. Prerequisite Program Standard Operating Procedures

The available evidence, including FDA's experience with the HACCP pilot programs, points to the effectiveness of two programs that do not fall within the parameters of traditional HACCP. FDA will refer to these programs in this document as "prerequisite programs." The first of these programs is that the firm have in place SOP's designed to ensure plant sanitation.
The seafood final rule requires in s123.11 that the processor monitor certain sanitation measures and document both the monitoring activities and any corrective actions taken when such monitoring finds an insanitary condition that may contribute to the likelihood of product becoming hazardous. While seafood processors are not required under s123.11(a) to develop and implement written sanitation or prerequisite program SOP's, processors must maintain sanitation control records that, at a minimum, document that certain monitoring requirements have been met, and that corrective actions are taken when necessary (s123.11(c)). Section 123.11(b) sets forth requirements for sanitation monitoring.
FSIS's regulations for meat and poultry require that official establishments develop, implement, and maintain written SOP's for sanitation (9 CFR 416.11). Each official establishment must take appropriate corrective action when it or FSIS determines that the SOP's have failed to prevent direct contamination or adulteration of product (9 CFR 416.15). Each establishment must maintain daily records that are initialed and dated to document the implementation and monitoring of the SOP's and any corrective actions taken (9 CFR 416.16). Finally, FSIS verifies the adequacy and effectiveness of the SOP's (9 CFR 416.17).
Insanitary facilities or equipment, poor food handling, improper personal hygiene, and similar insanitary conditions create an environment in which products may become contaminated with microorganisms, including pathogens. However, sanitation controls may be difficult to fit into HACCP plans. Sanitation covers the whole processing environment, not just CCP's. A prerequisite program is an appropriate mechanism for a situation, such as sanitation, that does not lend itself well to HACCP controls. Therefore, sanitation SOP's are a type of prerequisite program that is essential to provide a solid foundation for HACCP systems. The agency tentatively concludes that sanitation SOP's are an essential foundation for HACCP systems for juice.
The second prerequisite program is one that provides control over materials that are entering the plant. The SOP requirements of both the seafood and FSIS regulations are limited to sanitation. However, the pilot program experience has suggested the utility of controls on incoming material. A processor could use incoming material prerequisite program SOP's, in a manner similar to the sanitation SOP's, i.e., to cover a range of processing factors, not just CCP's. Although use of incoming material SOP's may not obviate the need for some CCP's in a HACCP plan, FDA anticipates that their use could help to ensure the safety of the food produced.
Incoming material controls for raw produce could be invaluable in establishing the conditions under which produce needs to be grown (including pesticide application) and harvested to provide assurance to the processor that the raw produce will not present hazards that the processor will otherwise need to control. For example, the processor's incoming material SOP's could specify that the processor will only purchase carrots that have not been fertilized with manure during growth. Another example is that the incoming material control could specify that the processor will only accept apples that have been picked from the tree, and that dropped apples are unacceptable. A simple solution to control the possible *20466 presence of unlawful pesticide residues on fruits and vegetables is to establish SOP's for incoming material control that ensure that any pesticides that have been used on the produce are approved for that use, are used at the appropriate level, and that appropriate time has elapsed between application and harvest.
As discussed previously, FDA is developing GAP and GMP guidance that has been issued in draft for comment. The guidance will address potential food safety problems throughout the food production and distribution system such as sanitation, worker health, and water quality.
A manufacturer also could use controls on the packaging materials that it receives. Proper packaging is essential if a processor is to minimize the possibility of the occurrence of hazards after juice has been processed. Juice that is not packed in hermetically sealed containers may be subject to contamination from a number of sources. The processor also needs to ensure that the container coating that it uses will not deteriorate through reasonable storage. Evidence in section I.B of this document showed examples where the acid content of some juices corroded the tin lining of the container, and the tin was present in sufficient concentration to be toxic. Incoming material controls will mean that the processor will act to ensure that packaging materials are safe and suitable before accepting them.
Incoming material controls for ingredients that a processor may add to juice can also be helpful. For example, if a processor is purchasing juice or juice concentrate from a supplier for use in a multi-juice beverage, it is essential that that juice have been processed under an adequate HACCP system and have not been contaminated during transportation. Thus, incoming material SOP's will lead the processor to establish controls on ingredients as criteria for acceptance in the plant.
However, the agency is not proposing to provide for the use of incoming materials SOP's in part 120 at this time and requests comment on this issue. FDA is seeking comment on whether incoming material SOP's can be utilized in a similar relationship to the HACCP system as the sanitation SOP's. Do interested persons see value in FDA requiring that these SOP's be written, monitored, and verified? How do these SOP's relate to FDA's draft guidance on fresh produce? What are reasonable procedures for acceptance of incoming materials that could be incorporated into SOP's?

1. Sanitation SOP's

FDA is proposing in s120.6(a)(1) to require that processors have and implement SOP's that address sanitary conditions and practices before, during, and after processing. Good sanitation practices are critical to the prevention of microbiologically related foodborne illnesses. FDA's CGMP regulations for food in part 110 set out general principles of sanitation that should be followed in plants that manufacture, package, label, or hold human food. They address such matters as personal hygiene and cleanliness among workers who handle food, the suitability of the plant design to sanitary operations, and the cleaning of food-contact surfaces. The proposed sanitation SOP's relate to the entire facility, not just to a limited number of CCP's. FDA tentatively concludes that this step is necessary to fully implement section 402(a)(4) of the act and yet at the same time not overload the HACCP system. FDA invites comments on this approach.
FDA did not elect to make the development of a written sanitation SOP mandatory for seafood because it recognized that some processors may be able to achieve satisfactory sanitary conditions and practices without having to commit their sanitary control procedures to writing (60 FR 65096 at 65149). In the seafood final rule, FDA concluded that as long as there were records demonstrating that the plant was being kept in sanitary condition, it was not necessary to require written sanitation SOP's, even though the agency strongly recommended that a processor have them. The agency requests comment on whether it should require for juice HACCP that sanitation SOP's be written.
In the evidence discussed in section I.A of this document, there were several instances where contaminated water was the cause of the outbreak. The water that the processor used was contaminated and when produce was washed with it before juicemaking, the water contaminated the produce, resulting in contaminated juice. Therefore, the safety of the water that comes into contact with food or food contact surfaces is an important factor that a processor must consider to maintain proper sanitation and prevent contamination of the product and plant. The seafood HACCP regulation in s123.11(b) lists eight sanitary conditions and practices that processors must monitor, and monitoring the safety of the water that comes into the plant is one of them (s123.11(b)(1)). Based on the foregoing, FDA is proposing a similar requirement in s120.6(a)(1).
In section I.B of this document, FDA recounted the evidence demonstrating, that several outbreaks were caused by cleaning solution directly contaminating the juice. Sanitation SOP's for seafood in s123.11(b)(5) require that processors protect food from adulteration with cleaning compounds. Given that cleaning compounds, sanitizing agents, pesticides, and other materials can pose a similar threat if not properly used in a juice processing facility, FDA is proposing a parallel requirement in s120.6(a)(5).
The other provisions of s123.11(b) are based on CGMP and encompass basic sanitation principles. Based on its consideration of the factors that it cited in arriving at s123.11(b), the agency tentatively concludes that it is appropriate to require in s120.6(a) that juice processors address the same sanitary conditions and practices in their SOP that must be monitored by seafood processors. FDA requests comment on the proposed matters that must be addressed in the sanitation SOP, and whether others are necessary for juice.

2. Other Requirements for Prerequisite Program SOP's

FDA is proposing in s120.6(b) that processors monitor sanitation conditions and practices during processing with sufficient frequency to ensure, at a minimum, conformance with those conditions and practices specified in part 110 that are appropriate both to the plant and to the food being processed. The seafood HACCP regulation requires sanitation monitoring (s123.11(b)). Because prerequisite programs potentially include facility-wide control points and provide a foundation for HACCP systems, processors need to monitor the performance of the SOP's to ensure that they are functioning as designed, and that they are corrected if there is a problem.
The agency is proposing in s120.6(c) that processors maintain records that document the monitoring that they do under the prerequisite program SOP's and any corrections to those SOP's that they make. Monitoring and recording of conditions and practices under the prerequisite program SOP's are as much keys to the success in improving those conditions as is the development by a processor of the SOP's. As in the case of HACCP records, FDA is proposing to require that processors engage in systematic monitoring of their own sanitation practices and conditions. This proposed requirement is similar to what is required for sanitation SOP's for seafood (s123.11(c)). Monitoring to *20467 ensure that sanitation is under control is the responsibility of all processors. Monitoring records help processors to see trends, and also allow the regulator to assess a processor's compliance over a period of time, not ju