International Journal of Food Microbiology 31 (1996) IL26

Review

article

Microorganisms
Jill A. Snowdon,*,

in honey
Dean 0. Cliverb

Scienti$c Consultant, National Honey Board, 390 Lashley Street, Longmont, CO 80501, USA bPopulation Health and Reproduction, University of California, School of Veterinary Medicine, 1019 Haring Hall, Davis, CA 95616, USA

Received 24 February 1995; revised 6 October 1995; accepted 28 November 1995

Abstract Knowledge of the moisture and temperature conditions influencing growth of microorganisms in honey has long been used to control the spoilage of honey. However, the need for additional microbiological data on honey will increase as new technologies for, and uses of honey develop. Microorganisms in honey may influence quality or safety. Due to the natural properties of honey and control measures in the honey industry, honey is a product with minimal types and levels of microbes. Microbes of concern in post-harvest handling are those that are commonly found in honey (i.e., yeasts and spore-forming bacteria), those that indicate the sanitary or commercial quality of honey (i.e., coliforms and yeasts), and those that under certain conditions could cause human illness. Primary sources of microbial contamination are likely to include pollen, the digestive tracts of honey bees, dust, air, earth and nectar, sources which are very difficult to control. The same secondary (after-harvest) sources that influence any food product are also sources of contamination for honey. These include air, food handlers, cross-contamination, equipment and buildings. Secondary sources of contamination are controlled by good manufacturing practices. The microbes of concern in honey are primarily yeasts and spore-forming bacteria. Total plate counts from honey samples can vary from zero to tens of thousands per gram for no apparent reason. Most samples of honey contain detectable levels of yeasts. Although yeast counts in many honey samples are below 100 colony forming units per gram (cfu/g), yeasts

* Corresponding
fax: +

author, 2623 Poinsettia Drive, White Oak, PA 15131, USA; Tel: + 1 412 672 9706;

1 412 672 9708.

0 1996 Elsevier Science B.V. All rights reserved

0168-1605/96/$15.00

PII SO168-1605(96)00970-l

J.A. Snowdon,

D.O. Cl&r

/ht. J. Food Microbiology

31 (1996) l-26

can grow in honey to very high numbers. Standard industry practices control yeast growth. Bacterial spores, particularly those in the Bacillus genus, are regularly found in honey. The spores of C. botulinurn are found in a fraction of the honey samples tested ~ normally at low levels. No vegetative forms of disease-causing bacterial species have been found in honey. Bacteria do not replicate in honey and as such high numbers of vegetative bacteria could indicate recent contamination from a secondary source. Certain vegetative microbes can survive in honey, at cool temperatures, for several years. However, honey has anti-microbial properties that discourage the growth or persistence of many microorganisms. Typically, honey can be expected to contain low numbers and a limited variety of microbes. A routine microbiological examination of honey might include several different assays. A standard plate count provides general information. Specialized tests, such as a count of yeasts and an assay for bacterial spore-formers, may also be useful. An indicator of sanitary quality as provided by coliform counts might be included. Additional tests, to explain unusually high counts or address a certain problem, may be needed. The use of honey in products that receive no or limited heat treatment may require additional tests. More information on the source and control of microbes in honey is needed to answer the concerns currently facing the industry. Keywords: Bacteria; Yeasts Food safety; Honey; Microorganisms; Molds; Specifications; Spoilage;

1. Microbial

content and the quality and safety of honey

Honey, the nectar and sweet deposits from plants as gathered, modified and stored in the honeycomb by honey bees, is a popular sweetener. New technologies and innovative uses of honey are expanding marketing opportunities. However, new microbiological requirements pertaining to quality and safety may be associated with these opportunities. A more comprehensive understanding of honeys microbiological characteristics is needed as honey is used in new ways. Honey is packaged for retail sale, for food service, and in bulk for use as an ingredient. In the US, retail sales for direct use of honey in the home kitchen, often as an ingredient in sauces and baked goods, total about 100 million pounds of honey per year. Honey is also sold as a food service item to restaurants or directly to consumers at warehouse- or club-style stores. Approximately 46 million pounds of honey are sold this way each year in the US. About 150 million pounds of honey per year are used as a food ingredient. A small amount of honey is used in non-food items such as drugs, cosmetics or pet-food. These figures are estimates provided by the National Honey Board (a research and promotion board administered by the United States Department of Agriculture). Current purchasing specifications for microorganisms in honey are often based on microbial specifications for other foods and ingredients. Due to honeys unique properties, such as its anti-microbial activities, some of those specifications may not be germane to honey. Purchasing specifications currently in use include a standard plate count and tests for coliform bacteria, yeasts, molds and certain pathogenic

J.A. Snowdon,

D.O. Clivrr

1hr. J. Food Microbiology

31 (1996) l-26

bacteria such as Staphylococcus, Salmonella and Clostridium species. Examples of recent microbiological specifications for honey are shown in Table 1. Honey is essentially water (average 17.2%) suspended in fructose (average 38.4%) and glucose (average 30.3%; White et al., 1962b). Honey also contains sucrose (average 1.3%) and other carbohydrates (about 12X), minerals (average 0.169%) and proteins (169 mg/lOO g; White et al., 1962b). The pH of honey ranges from 3.4 to 6.1 with an average of 3.9, while the water activity varies between 0.5 and 0.6 (White et al., 1962b). Honey has distinctive properties that inhibit or kill most microorganisms. Hence, microbes of interest to the honey processing industry are those that withstand the concentrated sugar, acidity and anti-microbial character of honey. These microbes could be put in three categories: (1) microorganisms that are commonly found in honey (certain strains of yeasts and spore-forming bacteria); (2) microorganisms that indicate sanitary or commercial quality (coliforms or yeasts); and (3) microorganisms that, under certain conditions (e.g. germination and growth in a non-heat-treated food product), could cause illness. A fourth category, microorganisms that cause diseases in honey bees, is not considered in this paper.

2. Background

on honey microbiology

2.1. Sources of microbes in honey The following information leads to the conclusion that primary sources of microbial contamination are likely to include pollen, the digestive tracts of honey bees, dust, air, dirt and flowers. Secondary sources of microbes in honey are likely to be the same as for other foods. 2.1. I. Primary sources Microbes found in comb honey are principally bacteria or yeasts and come from the bees, the raw material (nectar) or from external sources. Bacteria and yeasts may have different origins. Larvae may be sterile initially (White, 1921; Gilliam, 1971) but they are fed nectar and pollen by workers and are therefore subject to inoculation by the nectar, pollen and workers flora before pupation. Many microorganisms are associated with specific foods or components of the ecosystem (Jay, 1992). Organisms found in the environment around honey (i.e. bees, hives, pollen, flowers, soil, etc.) are also likely to occur in honey. Actinetobacter, Bacillus, Clostridium, Corynebacterium, Pseudomonas, Psychrobacter and Vagococcus are bacteria commonly found in soil. Air and dust are important sources of Bacillus, Clostridium and Micrococcus species. Bacillus and Clostridium species are important bacterial contaminants of cane and beet sugars, Saccharomyces and Torula yeasts can be found in high-moisture sugars, and Leuconostoc mesenteroides has been found in sugar refineries. Brochothrix, Citrobacter, Enterobacter, Erwinia, Flauobacterium, Lactobacillus, Luctococcus, Leuconostoc, Listeria and Pediococcus are found in plants and plant products.

Table 1 Examples for honey

of microbial

purchasing

specifications

Assay Coliform 10 100 100 100 10 200 100 100 100 10 100 100 100 10 100 100 100 E. coli Yeast Yeast and mold Mold Staphylococcus Salmonella Neg. Neg. Neg. Neg. Neg. Neg. Neg. 50 50 Rope 50 spore (sic)

Company Neg. 100 13 Neg. Neg. Neg. 100 Neg. Neg.

Standard

plate count

A B Cb D E F G H I J :, 10 10 100 10 10 10 10 Neg. (10 (0.3 Neg. Neg. Neg. 10 Neg. Neg. Neg.

1 000 5 000 10 000 10 000 10 000 10 000 1000 10 000 10 000 5000

Colony forming units per gram. bAlso negative for Clostridium botulinurn and C. perfringens. Also, 10 cfu/g of mesophilic and thermophilic aerobes and anaerobes.

J.A. Snowdon, D.O. Cliver /ht. J. Food Microbiology 31 (1996) l-26

Sackett (1919) summarized research by others which indicated that Bacillus, Micrococcus and Saccharomyces species can be readily isolated from honey combs and adult bees. Comb honey and healthy larvae contained no microbes detectable by methods then available. A number of microbial species were isolated from the feces of bee larvae (Gilliam and Prest, 1987). Bacillus species were most prevalent, followed by gram-variable pleomorphic bacteria. Molds, Actinomycetes, gram-negative rods (probably Enterobacteriaceae) and yeasts were also recovered, while Streptomyces spp. were recovered from one larva. No anaerobic bacteria were recovered. This is in comparison to the intestinal microflora of adult honey bees, which is dominated by gram-variable pleomorphic bacteria of uncertain taxonomic status, Bacillus spp, Enterobacteriaceae, Penicillium spp., Aspergillus spp. and sometimes yeasts (frequently Torulopsis spp) (Gilliam et al., 1988). Pollen may be the original source of microbes in the intestines of honey bees (Gilliam et al., 1983). The honey bees appear to be seeded microbiologically by pollen consumption and by other bees in the colony through food exchange. The scarcity of microbes in nectar suggests that nectar plays a minor, if any, role in this process. Root (1983) suggests that flowers and hives are more important sources of microbes than the soil. Aerobic spore-forming bacilli (which could be Bacillus) were the most frequently encountered microbes on the external surface, crop and intestine of the honey bee (El-Leithy and El-Sibaei, 1972). The guts of larvae, pupae or newly-emerged honey bees are often sterile; fungi and bacteria are regularly found in the intestines of adult worker bees a few days after emergence (Gilliam, 1971). Likewise, no or limited microbes were found in the nectar that was sampled (Gilliam et al., 1983). Bacteria belonging to the genus Bacillus were found in the digestive tract, hemolymph and trachea of healthy honey bees (Gilliam and Valentine, 1976). Gilliam (1978) suggests that pollen may be the main source of the gut microflora in worker bees since there are few or no microbes in nectar. The intestines of bees have been found to contain: 1% yeast-shaped microbes; 29% gram-positive bacteria including Bacillus, Bacteridium (sic), Streptococcus and Clostridium species; and 70% gram-negative or gram-variable bacteria, including Achromobacter, Citrobacter, Enterobacter, Erwin& Escherichia coli, Flavobacterium, Klebsiellrt, Proteus and Pseudomonas (Tysset and Durand, 1968; Tysset et al., 1970a). Nectar, the body of the bee, apiary soil and honey-house air and equipment are all considered possible sources of yeast (Crane, 1979). Troller (1979) suggests that the ultimate reservoir for yeasts in honey may be the hive and that the worker bees distribute the yeast to the nectar as it is collected. Bacteria introduced to honey by the bees are probably members of the Bacillus family (Tysset and Rousseau, 1981). The mold, Aspergillus, may occasionally be present in honey but would be found only in the dormant, spore form (which would preclude the production of mycotoxins such as aflatoxin). The primary sources of sugar-tolerant yeasts are flowers and soil; yeasts may be present in honey combs, air and equipment in the honey house (Graham, 1992). Morse and Hooper (1985) maintain that yeasts are found in nectar.

J.A. Snowdon, D.O. Cliver 1 Imt. J. Food Microbiology 31 (1996) l-26

Nakano et al. (1992) analyzed 273 samples of non-honey sweeteners (e.g., unrefined sugar and corn syrup) destined to become food for bees, for the presence of C. botulinurn. Three out of 56 (5%) of the sweeteners contained botulinal spores, while 8 of 217 (4%) of the sweeteners not intended to feed bees contained spores. The authors suggest that the sweeteners used to feed bees may be a source of spores in honey. Most research on primary sources of microbes in honey is done to understand the microbial ecology of the honey bee. A listing of the microbes reported to be associated with bees can be seen in Table 2. It appears that pollen, not soil, is the proximate source of microbes that seed the intestines of bees. Although not identical, there is a considerable overlap between the microbes that are in bees (Table 2) and those that have been found in honey (Table 3). This suggests that some microbes that are introduced into honey by bees do not survive in honey and that some microbes not associated with bee intestines are introduced into honey. Bacillus, Clostridium, and Micrococcus species are common in air and dust and in honey. This suggests that air and dust are also sources of primary contamination in honey. Soil and flowers may be sources of yeasts in honey. The roles of the hive and nectar as primary sources of microbial contaminants in honey are undetermined, while the sweeteners used to feed bees may be a source of spores in honey. More research is needed to determine the relative importance of these various routes.

Table 2 Microbes associated with honey bees Bacteria Fungi Yeasts

Achromobacter Actinomycetes Bacillus Bacteridium (sic) Citrobacter Clostridium E. co/i Enterobacter Erwinia Flavobacterium Klebsiella Micrococcus Proteus Pseudomonas Streptococcus Streptomyces Saccharomyces Torulopsis

Molds
Aspergillus Penicillium

Does not include microbes that are specifically pathogenic to bees. See Shimanuki and Knox (1991).

J.A.

Snowdon,

D.O.

Clioer

Int.

J. Food

Microbiology

31 (1996)

l-26

Table 3 Microbes reported to be found in honey Bacteria Fungi Yeasts

Alcali~mrs Bucillus Bucteridium Bacterium Clostridium Entrrohncter Flal;ohuc,terium Klebsiellu Micrococcus Neisser ici Proteus Pseudomonas Xunthomonus Ascosphaeru Debaryomyces

Molds
Aspergillus Atichiu Bettsia u&i Cephalosporium Chcretomium Coniothecium Hormiscium Peniciilium Peronsporaceae Peyroneliu Triposporium

(sic) (sic)

Hansen& Lipomyces Nematospora Oosporidium Pichia Rhodotorula Succharomycrs Sc/2irosacclzarumycrs~accl~aromyce.~ Schwanniomyces Trichosporun Torulu Torulopsis Zy~osuc.cliuroniyces

Bre&xwterium

Uredianceae Ustilaginaceae

2.1.2. Secondary sources Data on secondary sources of microbes in honey are relatively sparse in the scientific literature. A comprehensive microbiological analysis of 12 honey samples suggested that microbial contamination is more likely to take place during and after extraction than in the hive (Tysset et al., 1970b). Tysset and Rousseau (1981) point out that secondary sources of contamination in honey are humans, equipment, containers, wind, dust, insects, animals and water. They caution that microbes may survive longer than is generally believed and that good manufacturing practices are important for the control of microbes in all food, including honey. Yeasts have been recovered from equipment in honey houses; contaminated equipment can introduce yeasts into otherwise clean honey (Root, 1983). Secondary sources of contamination are often the same among different foods. Possible routes of transmission into extracted honey would include air (in the honey house or while the honey was being packed), food handlers (from skin infections, sneezing or frank fecal contamination), cross-contamination (largely from animals or animal products), and equipment (including residues of food and water). Floors, walls and ceilings can also be reservoirs of microbes that enter food. Thousands of cfu of vegetative microbes (which are not expected in honey and cannot grow in honey) per gram might indicate recent contamination of the honey via a secondary source. Secondary sources of contamination are controlled by standard sanitation and good manufacturing practices. Numerous texts and manuals have been written on this subject. These control measures have been developed by the processed food products industry and are also applicable to products that receive little or no processing.

J.A. Snowdon,

D.O. Cliver /ht.

J. Food Microbiology

31 (1996) l-26

2.2. Types, levels and persistence of microbes in honey A list of all microbes currently reported to be in honey can be seen in Table 3. No information is available on the presence or persistence of viruses and parasites. 2.2.1. Molds Mold is associated with the intestinal contents of bees, the hive, and the environment in which the bees forage. Aspergillus has been recovered from the intestines of honey bee larvae (Haisig and Kamburov, 1966). A few scientists have described the presence of sooty molds (epiphytic fungi that are osmophilic) in honeydew and honeydew honey (as cited in Crane, 1979). Few of the molds are identified with certainty, but typical genera may include Atichia, Coniothecium, Hormiscium, and Triposporium. Morse and Hooper (1985) report that when combs are stored in damp areas, mold often appears on the comb surface and that honey bees have a remarkable ability to clean and restore moldy comb. Furthermore, they state that stored pollen can be attacked by a fungus called Bettsia alvei. Molds, including Aspergillus, Chaetomium, Penicillium, and Peyronelia, have been isolated from the feces of bee larvae (Gilliam and Prest, 1987) while Tysset et al. (1970b) found Ascosphaera, Aspergillus, Cephalosporium, and Penicillium molds in honey. There are few reports that quantify the levels of mold in honey. Tysset et al. (1970a) found an average of 254/g of honey with a range from 0 to 2500/g. Piana et al. (1991) found mold in all 50 Italian honey samples analyzed. The levels were very low, ranging from l-43 cfu/g. Crane (1979) reported that hyphae and spores of fungi associated with bees (the pollen mold, Bettsia alvei) and plant-pathogenic fungi (Peronsporaceae, Uredinaceae and Ustilaginaceae) can occasionally be found in honey sediments. Troller (1979) lists Chryososporium, Eurotium, Monascus and Xeromyces bisporus as molds that can survive environments in which the water activity ranges from 0.7 down to 0.605, respectively. None of these genera have been recovered from honey or honey bees. This suggests that factors other than a tolerance of osmotic pressure (i.e., low water activity) may be important in determining molds that are likely to appear in honey. The low mold counts reported by Piana et al. (1991) suggest that molds may survive but do not tend to grow in honey, which is confirmed by recent experiences in the industry. One industry report (unpublished data) recorded the absence of mold in 40 out of 50 batches of honey, with the other 10 batches containing no more than 15 cfu/g. High mold counts may be indicative of the recent addition of mold, perhaps from growth in the foraging environment of the bee, in the hive, or on processing equipment. 2.2.2. Yeasts Yeasts can grow under acidic conditions and are not inhibited by sucrose. Osmophilic or sugar tolerant yeasts are a problem in the honey industry, because they can grow even at the limited level of water available in ripe honey. As a result, osmophilic yeasts readily ferment honey.

J.A. Snowdon,

D.O. Cliver / Int. J. Food Microbiology

31 (1996) I --26

Conditions that encourage fermentation in honey include increased moisture, moderate temperatures, granulation, a high yeast count and the presence of ash and nitrogen (Crane, 1979). During fermentation, the yeast acts upon the sugars, producing alcohol and carbon dioxide. In the presence of oxygen, the alcohol may be converted into acetic acid. Fermentation usually happens in micro-environments (such as the top of a barrel of honey) where the water content has increased. This subject is reviewed comprehensively in beekeeping textbooks (Crane, 1979; Root, 1983) and scientific journals (Tysset and de Rautlin de la Roy, 1974). Succhuromyces spp. represents the dominant yeast found in honey (Tysset and Rousseau, 1981) but other genera have been reported. Tysset and de Rautlin de la Roy, 1974 recovered Rhodotorula as well as Succharomyces from honey. Furuta and Okimoto (1978) list Debnryon?yces, Hansenula, Lipomyces, Oosporidium, Pichiu, Succharomyces, Torulopsis, and Trichosporan as genera of yeast recovered from Japanese honey. Studies since the turn of the century, assembled by Crane (1979), report the following yeasts: Nematospora, Saccharomyces, Schizosctccharomyces, Schwanniomyces, Torulu and Zygosacchuromyces. The concentration of yeasts is proportional to the availability of moisture (Crane, 1979; Quilez and Barrado, 1976). Honey made from flowers in humid regions has more yeast and can spoil in the comb (Tysset and Rousseau, 1981). Piana et al. (1991) assayed 50 samples of Italian honey and primarily found osmophilic yeasts in the range of l-3500 cfu/g, only 34 of the samples contained osmophilic yeasts. An average of 254 cfu/g of yeasts and molds, with a range of 0 to 2500, were found by Tysset et al., 1970b in honey. However, there were fewer than 10 cfu/g in 64% of the samples. Combined mold and yeast counts averaging 90/g (range O&2500) and osmophilic yeast counts averaging 102/g (range 0- 10 500) were reported in a survey of 175 samples of honey (Tysset and Rousseau, 1981). Yeast counts averaging 9 cfu/g (ranging from O&300) were observed by Nakano and Sakaguchi (1991) during the analysis of 270 retail honey samples. A range from 0 to 10000 (with the majority between 10 and 100) was attributed to Italian scientists in the report of Tysset et al. (1970b). It is generally accepted that the number of yeast spores in various honeys can vary a million-fold from 1 in 10 g to 100 000/g (Graham, 1992). In one study, osmophilic yeasts were found only in samples of honey with a water activity higher than 0.65 (Piana et al., 1991). Assays of 320 samples of Canadian honey demonstrated sugar-tolerant yeasts ranging from 0.1 to 1000000 cfu/g (Root, 1983). As fermentation is proportional to the concentration of yeast, honey with a very high yeast count is not likely to be palatable or marketable. Out of 50 samples of finished honey assayed by a packer in the honey industry (unpublished data), only 10 contained yeasts. Although the highest count was 315 cfu/g, the median was zero, and four of these samples contained osmophilic yeasts, ranging up to 250 cfu (average; 6 cfu/g). Because of industry control efforts, yeast counts in finished honey are not likely to exceed a few hundred cfu/g.

10

J.A. Snowdon, D.O. Cliwr 1 ht. J. Food Microbiology 31 (1996) l-26

2.2.3. Bacteria

2.2.3.1. Species, quantity and frequency of isolation. With the exception of research on C. botulinum, very little quantitative measurement of bacteria in honey is reported in the scientific literature (Table 4). Nakano and Sakaguchi (1991) tested 270 honey samples (from retail supplies in Japan) and recorded a mean aerobic plate count of 83 cfu/g. Tysset et al. (1970b) tested 14 samples of freshly harvested French honey and found total plate counts of less than 100 cfu/g. They were unable to detect E. coli, Streptococcus D (now Enterococcus), sulfite-reducers or Staphylococcus. In 198 1, Tysset and Rousseau examined 175 samples of commercial honey from different geographical regions of France and found the average total plate count to be 227 cfu/g. These figures are within the range of current industry experience where the bacterial levels of finished honey tends to range from 1 to 5000 cfu/g, with lower numbers possible with additional treatment. Variation in bacterial numbers may be due to the type of sample (raw, finished or retail), the age of the honey, the time of harvest and the analytical technique. Qualitative examination of the organisms recovered on total plate counts from 12 samples of French honey (Tysset et al., 1970b) found the following organisms: Bacteridium (sic), Bacterium (sic), Bacillus, Brevibacterium, Enterobacter, Flavobacterium, Micrococcus, Neisseria, Pseudomonas, and Xanthomonas. The most numerous isolates were members of the genus Bacillus, specifically B. cereus and B. pujilus. Thirteen percent of the samples contained Micrococcus and Pseudomonas (which the authors attribute to the intestines of bees). Micrococcus, Pseudomonas, and Staphylococcus have recently been found in industrial honey in the USA (unpublished data). Staphylococcus was found in 3 of 25 samples. Micrococcus was found in 4 of 12 samples and Pseudomonas was found in 1 of 8 samples. Tysset et al., 1970b also found Flavobacterium lactis in honey and suggested that it was probably introduced by man via contaminated water. Vegetative microbes such as these are likely to be introduced into honey via secondary contamination and hence may be expected to appear in honey on a sporadic basis. The authors concluded that natural honey contains very few types of microbes and is devoid of vegetative, non-spore-forming bacteria. The incidence of Bacillus and Clostridium spores in honey from processing plants and retailers has been studied (Kokubo et al., 1984). Spores were found in 67 of the 71 (94%) of the honey samples tested at levels of lo-100 spores per gram. Most of the spores were facultative and in the genus Bacillus. The predominant species was B. cereus followed by B. coagulans, B. megaterium, and B. alvei. Fifty- six of 71 (79%) of the samples contained clostridial spores, with 6 samples containing C. perfiingens. None of the honey samples in this study contained C. botulinum. B. cereus spores were found in 24 of the 50 samples tested by Piana et al. (1991) with counts ranging from 0.1~ 1 cfu/g. C. perfringens or C. botulinum was not recovered, although unidentified anaerobic spores were found in 22 of the 50 samples, at levels between 0.1 - 1 cfu/g. Spores of aerobic species were found in all samples, at levels ranging from l-67 per gram.

Table 4 Levels of microbes

in honey

Study E. coli Yeast Mold Osmophilic Yeasts and molds yeasts

Assay Clostridium ianaerobes Bacillus

Standard count O-300 I-43 0.1-106 O-2500 O-2500 <3 < 10 < IO I-3500

plate

O-260 l-55

l-132 0.1-I 0

O-180

25- 1600 3-9500

<IO0 O-IO 500

Nakano et al. (1989) Piana et al. (1991) Root (1983) Tysset et al. (1970b)b Tysset and Rousseau (1981) Industry data (1990S)

170-8500

Colony forming units per gram. hStreptococcu.s group D = 0; Fla~ohacterium lactis = 250. Coliforms < 3: Salmonella and Streptococcus = 0.

12

J.A. Snowdon,

D.O. Cliuer /Ini.

J. Food Microbiology

31 (1996) l-26

There are over a dozen reports in the scientific literature on the incidence of botulinal spores in honey (Sugiyama et al., 1978; Midura et al., 1979; Flemming and Stojanowic, 1980; Hartgen, 1980; Huhtanen et al., 1981; Kautter et al., 1982; Stier et al., 1982; Guilfoyle and Yager, 1983; Kokubo et al., 1984; Aureli et al., 1985; Sakaguchi et al., 1987; Hauschild et al., 1988; Nakano and Sakaguchi, 1991; Nakano et al., 1989; and Du et al., 1991). These reports represent information on honey from all over the world. Of the 2033 total samples tested in the studies listed above, 104 (5.11%) contained detectable levels of botulinum spores. Many of the studies detected no botulinum spores at all. One research group, while evaluating the efficacy of various techniques to detect botulinum spores in honey, recorded that 62% (15 out of 24) of their samples contained botulinum spores (Stier et al., 1982); their work highlights variation in test results between samples and analytical techniques. Most of the studies found botulinum spores in 5 to 15% of samples. Typically, botulinum spores were found at levels below l/g of honey (Sugiyama et al., 1978; Stier et al., 1982; and Nakano et al., 1989). Midura et al. (1979) assayed honey suspected of transmitting infant botulism, and estimated (sic) the concentration of botulinum spores as being from 5 to 80 spores per gram. Nakano et al. (1989) found that 1 of their 270 honey samples contained high levels of spores (36660/g) with the remainder containing < l/g. The reason occasional samples of honey contain higher than usual levels of spores is not known. It can be concluded that botulinum spores are present in honey some of the time (about 5%) and typically at low numbers ( < l/g). Nonetheless, honey is an inappropriate food for infants because infants have a limited intestinal microflora and are susceptible to developing infant botulism (Snowdon, 1991). From 0 to 60 spores of C. botulinurn were detected, per gram of honey, in 8.5% of the 270 samples tested by Nakano et al. (1989). Spore counts were most common in samples collected from apiaries (23%) followed by samples from drums (18%), followed by samples from retail packages (5%). This corresponds with reports from Sugiyama et al. (1978) who reported a higher incidence in samples taken from apiaries than from retail supplies. Nakano et al. (1989) also observed a decrease in viable spores after honey was stored for a year at 25C and concluded that some processes during purification and prolonged storage contributed to the lower et al. (1981) sought to determine why incidences in marketing honey. Huhtanen numbers of botulinal spores vary so greatly among honey samples. They found spores in 3 of 20 (15%) of the bulk samples and only 4 of 60 (6.6%) of the samples from apiaries. Counts for Bacillus spp. varied from 0 to 180 cfu/g and counts of anaerobic microbes from 1 to 132 cfu/g. They also detected Alcaligenes, Pseudomonas, Flavobacterium and gram-positive asporogenic bacteria. Huhtanen et al. (1981) inoculated bees with C. botulinurn spores and recorded the appearance of botulinal spores in the honey in their hives. They found no evidence that C. botuknum multiplied during honey-making and did not detect any anaerobic organisms, such as C. botulinurn, in the digestive tracts of the bees. Hartgen (1980), assayed 210 commercial samples of honey from retail markets in South Bavaria and found no botulinal toxin. C. botulinurn must multiply for toxin to be produced. Growth of C. botulinurn in ripe honey has never been demonstrated and is unlikely.

J.A. Snowdon, D.O. Clivrr / hi. J. Food Microbiology 31 (1996) l-26

13

Over 100 strains of yeasts and bacteria isolated from honey were studied by El-Leithy and El-Sibaei (1972). Six strains of Bacillus spp. were studied further via inoculation into different solutions of honey and water; none could multiply in solutions containing over 50% honey or more acidic than pH 4. Only moderate growth was reported at sugar concentrations as low as 30%. Though bacteria may be unable to multiply in mature honey, perhaps, as has been suggested for Clostridia, bacterial growth can occur while the nectar is ripening into honey.

2.2.3.2. Survival. Most studies on bacterial survival in honey involve the introduction of vegetative pathogenic organisms, which are not normally present in honey (Table 5). Sackett (1919) hypothesized that bees could carry pathogenic microbes from human excrement to honey. He demonstrated that 10 species of bacteria (non-spore-forming intestinal bacteria) inoculated into pure honey survived only a few hours or a few days. Solutions of less than 50% honey in water sustained bacterial life for longer periods, but never exceeding 40 days. Sackett concluded that the probability of honey acting as a carrier of typhoid fever, dysentery and various diarrhoeal affections is very slight. The survival of some gram-negative bacteria in commercial honey was studied by Tysset and Durand (1973). These bacteria were similar to Sacketts species and are of interest because they may cause human illness. The investigators inoculated aseptically-collected honey with various bacteria, held it at 20C (68F) and monitored the change in bacterial numbers. Loss of bacterial viability was observed within 8 to 34 days (dependent upon the species). In another study, Tysset and Durand (1976) investigated the survival of similar bacteria at 10C (50F) in freshly harvested honey of mixed flower sources. The bacteria survived from 6 months to almost 2.5 years. The authors emphasized the need for vigilance not to introduce these microbes into honey and cautioned against overestimating the antibacterial properties of honey. In another experiment Tysset et al. (1979) reported survival times, at 20C (68F), ranging from 26 to 77 days for various species of M~~obucterium, including the species responsible for tuberculosis. Molan, 1992b suggests that bacterial survival differs among different types of honey. Staphylococcus was recovered from 3 of 12 samples of raw honey being prepared by a US honey packer (unpublished data). No staphylococci were recovered from 13 finished samples, and in no instance was a pathogenic species (e.g. Staphylococcus aureus) recovered. Staphylococcus appears to be present in honey only rarely, appears not to survive processing, and is not likely to grow. The survival of spore-forming bacteria found in commercial honey was studied by Kokubo et al. (1984). B. cereus, C. perfringens, and C. botulinurn spores were inoculated into honey and stored at 25C (77F) for 4 months. The spore counts remained the same. The C. botulinum spore population in honey did not change in over a year when stored at 4C (39F; Nakano et al., 1989). At 25C (77F) however, the number of spores began to decrease after 100 days. No spores were detected after 5 days of storage at 65C (117F). These findings indicate that spores may persist in honey, but do not germinate to vegetative cells that might multiply.

14

/.A. Snowdon, D.O. Cliver / Int. J. Food Microbiology 31 (1996) l-26

In conclusion, spore-forming bacteria, molds and yeasts are typically found in honey, often at low numbers, while spores can persist indefinitely. Vegetative forms of pathogenic bacteria have never been found in honey; if introduced, they can survive in honey for extended periods of time, particularly at cool temperatures. Microbial survival may be influenced by the type of honey and its moisture content.
Table 5 Survival of vegetative Species Non-spore-forming intestinal bacteria Edwardsiella tarda Erwinia amylovora Escherichia coli Mycobacterium chelonie phlei tuberculosis tuberculosis bovis tuberculosis avium Proteus vulgaris Pseudomonas aeruginosa Salmonella derby dublin enteritidis enteritidis typhi typhi typhimurium typhimurium Serratia marcescens Shigella

bacteria

in honey Tempterature Room 20 4 20 (C) Survival time Reference Sackett (1919)

temperature

A few days

i 10 days 8 weeks
< 10 days

Tysset and Durand (1973) De Wael et al. (1990) Tysset and Durand (1973) Tysset (1973) Tysset (1973) Tysset Tysset Tysset Tysset (1973) Tysset (1973) 13 days 12 Tysset (1976) Tysset (1976) Tysset (1973) Tysset (1976) Tysset (1973) Tysset (1976) Tysset (1973) Tysset (1976) Tysset (1973) Tysset (1976) and Durand and Durand et al. (1979) et al. (1979) et al. (1979) and Durand and Durand

20 20 20 20 20 20 20

26 days 17 days 67 days 77 days 71 days < 10 days 8h

10 10 20 10 20 10 20 10 20 10

6 months,

and Durand and Durand and Durand and Durand and Durand and Durand and Durand and Durand and Durand and Durand

2 years, 4 months days 34 days t year, 1 t months, days 26 days 4 months, 30 days 2 years, 4 months, days 18 days 2 months, 22 days 21 days

12

J.A. Snorwlon, Table 6 Factors that might

D.O. Cliuer / Ini. J. Food Microbiology

31 (1996) I-26

15

contribute

to the anti-microbial

nature

of honey

High osmotic pressure, low water activity (A,) Low pH-acidic environment Glucose oxidase system ~~ forms hydrogen peroxide Low protein content High carbon to nitrogen ratio Low redox potential (Eh), due to high content of reducing sugars Viscosity opposes convection currents and limits dissolved oxygen Chemical agents Pinocembrin Lysozyme Acids (phenolic) Terpenes Benzyl alcohol Volatile substances (perhaps phytochemicals influenced by bee enzymes) (Molan, 1992a; and Tysset and de Rautlin de la Roy, 1974).

2.2.4. Other microbes Microbes other than yeasts, molds and bacteria could be present in honey. Algae may be present in honeydew honey and in honey sediment under certain climatic factors such as when the relative humidity is high (Crane, 1979). The skeletons of dinoflagellates could be introduced into honey from diatomaceous-earth filters, but are non-viable. Some human enteric viruses, such as hepatitis A, sustain dry conditions. Human enteric viruses could be expected to persist in honey, their numbers decreasing at rates dependent upon temperature and time. As human feces are the exclusive source of human enteric viruses, control measures that keep human feces out of food will also preclude the presence of human enteric viruses. Protozoa and multicellular parasites are not likely to be transmitted through honey. These microbes are given no other consideration in this paper due to lack of information in the scientific literature. 2.2.5. Summury Molds and yeasts are the only microbes that have been reported to grow in honey. Certain bacteria will survive in honey but growth is unlikely. Vegetative bacteria inoculated into ripe honey for experimental purposes die off within a few weeks, unless the honey is stored at cool temperatures. The hostile conditions found in ripe honey increases the demise of most bacteria. In practice, the spores of Bacillus, molds and yeasts tend to be present in honey on a regular basis. Clostridial spores (typically of species other than C. botulinum) may also be present. Reports from industry (unpublished data) indicate that Pseudomona and Micrococcus might also be found and that total numbers vary unpredictably. No information is available on viruses or parasites.

16

J.A. Snowdon,

D.O. Cliver / Int. J. Food Microbiology

31 (1996) I-26

2.3. Anti-microbial properties of honey Honey has unique properties that render it bacteriostatic and bactericidal (Table 6). The subject of the antibacterial properties of honey has been recently and extensively reviewed (Molan, 1992a). Limited information is available on other anti-microbial properties (e.g., honeys effect on viruses, fungi or parasites). The inhibitory properties of honey were noted early in the century by Sackett (1919) and others. The responsible agent was called inhibine, but White et al. (1962a) later identified this factor as hydrogen peroxide. The peroxide is formed by the glucose oxidase system, which is an enzymatic process that is active only in unripe or diluted honey. The enzyme (glucose oxidase) is heat-labile and is destroyed by pasteurization. The system breaks down a small amount of glucose, producing gluconic acid and hydrogen peroxide. Molds and yeasts, do not appear to be as sensitive to peroxide as bacteria. (White et al., 1962a). In a study of honey from eight different floral sources, Smith et al. (1969) showed bacteriostatic effects (at 100% concentration) against B. cereus, Micrococcus fluuus, and Surcina lutea. It was found that not all honey is equally effective against all microbes all the time; one of the eight types of honey under scrutiny, for example, did not inhibit B. subtilis. Gram-negative bacteria inoculated into honey were found to decline rapidly Tysset and Durand, 1973. Tysset and de Rautlin de la Roy (1974) suggested that a variety of factors may be responsible for the antibacterial nature of honey. The low protein content and high carbon-to-nitrogen ratio of honey are not conducive to microbial growth, nor is the acidity of honey. The low redox potential of honey (which is due to its high content of reducing sugars) discourages growth of molds and aerobic bacteria, while the viscosity of honey opposes convection currents and limits the entry of dissolved oxygen. As the osmotic pressure is high, the microbes shrivel as water flows out of their cells into the surrounding honey. Since honey has antibacterial activity against B. cereus, the low incidence of this organism in foraging worker bees might be due to honey consumption by the bees (Gilliam, 1978). Honey contains unknown factors active mainly against gram-negative bacteria and higher fungi such as Aspergillus (Radwan et al., 1984). The antibacterial activities of a variety of honeys from New Zealand were tested against Staphylococcus (Allen et al., 1991). Experimentation with 345 unpasteurized, monofloral samples (26 floral sources) ascertained that the floral source made a difference in the antibacterial activity of the honey. Honey derived from manuka and vipers bugloss flowers had antibacterial activity due to a non-peroxide component. Willix et al. (1992) determined that honey with non-peroxide antibacterial activity and honey with antibacterial activity due to hydrogen peroxide are both effective in preventing the growth of bacteria that infect wounds. However, the order of efficacy among the seven bacterial strains was dependent upon the nature of the antibacterial activity in the honey. Molan (1992~) studied the same samples of honey and showed that manuka honey has the strongest antibacterial action against E. coli, Helicobacter pylori, Proteus mirabilis, Pseudomonas aeruginosa, Salmonella typhimurium, Serratia marcescens, S. aureus, and Streptococcus pyogenes.

J.A. Snowdon, D.O. Cliver / ht. J. Food Microbiology 31 (1996) 1-26

11

An extensive review of the antibacterial properties in honey was published by Molan (1992a). He proposed that the antibacterial properties of honey are due to acidity; osmolarity; the conversion from glucose to hydrogen peroxide via glucose oxidase upon dilution of the honey; and other less clearly defined factors such as pinocembrin (an antibacterial component of honey), lysozyme, acids (phenolic and others), perhaps terpenes and benzyl alcohol and volatile substances (perhaps phytochemicals influenced by bee enzymes). Dozens of species of bacteria have been found to be susceptible to honey at a wide range of concentrations. Honeys antibacterial factors are reported also to be effective against many fungal species, including Aspergillus, Candida, Penicillium, and Saccharomyces. Growth of bacteria and fungi from sewage, soil, air, and tap water has been prevented by solutions containing as little as 25% honey, while solutions containing as little as 200/o honey prevented the growth of airborne contaminants. 2.4. Control of microbes There are intrinsic properties of foods that affect microbial growth. These include pH, moisture content, oxidation-reduction potential, nutrient content and anti-microbial constituents. As discussed above, honey has a number of inherent qualities that make it bacteriostatic or bactericidal. Extrinsic factors that have the greatest influence in determining the presence of foodborne organisms include storage temperature, relative humidity of the environment and presence and concentration of gases in the environment (Jay, 1992). Due to the natural properties of honey and control measures in the honey industry, honey is a product with minimal types and levels of microbes. Honey as stored in the comb serves as an excellent model of how to preclude microbial growth, although dormant forms of microbes may persist in the honey. Bees produce a food for themselves that is resistant to microbial degradation, by removing water and creating a barrier to oxygen. Similar principles can be followed even after the honey is harvested. Packaging the product so as to exclude air and to prevent cycles of water vaporization and condensation that can cause local dilution of the sugar will prevent microbial growth. Controlling the moisture content of the honey and the temperature at which honey is stored are two examples of how industry controls the growth of microbes in honey. These protective features vanish when honey is diluted or used as an ingredient. Indeed, the dormant forms of microbes normally present in honey or hitch-hiking microbes from post-harvest contamination can grow in the formulated food product. There is little or no control over the exposure of honey to microbes before harvest. It is not practical to try to control the quality of air, dust, earth and flowers or the materials carried in by the bees. The conditions in the hive will have some influence on the microbial quality of the honey. The practices that keep bees healthy will also influence the type and number of microbes to which honey is exposed. Moldy combs, for example, could be a source of mold spores in honey. American foulbrood disease (a common disease of honey bees in North America) is caused by bacteria of the genus Bacillus and could contribute to the presence of

18

J.A. Snowdon, D.O. Cliver /ht. J. Food Microbiology 31 (1996) l-26

Bacillus spores in honey. Good beekeeping practices can generally be expected to keep the number of microbes in honey low; but it has been suggested that using Bacillus thuringiensis (which is harmless to honey bees) to control wax moth could result in Bacillus spores in honey. Good manufacturing practices for beekeeping and for packing are described thoroughly elsewhere (Crane, 1979; Graham, 1992). Hence, this paper will make only brief reference to a few recent reports on the post-harvest control of microbes in honey. It is important to (1) avoid inoculation of honey with undesirable microbes and (2) avoid handling or storing the honey in such a way as to encourage microbial growth. Additionally, it is possible to treat foods such as honey so as to minimize or eliminate microbes. Certain microbes that originate in mammals, especially man, could cause illness if swallowed with honey or other foods. These microbes are carried with skin infections, in nasal passages and in feces. Soil contaminated with these types of microbes could transmit disease indirectly. Pathogenic microbes have never been found to occur naturally in honey and they do not survive in honey for very long at 20C (68F); however, lengthy survival (Table 5) is possible if honey is stored below 10C (50F). Control of these microbes can be achieved by using sanitary practices (hand washing, avoidance of sneezing or coughing into food, etc.). Controlling microbes from sources other than man (air, equipment, etc.) is equally important, but often more difficult. Troller (1979) suggests limiting the exposure of honey to the atmosphere and observing good sanitary practices in the apiary and in the honey processing facility. He also suggests drying honey processing equipment thoroughly after washing so the water activity of the honey doesnt increase and allow microbes to grow. Root (1983) urges strict cleanliness at the time of extraction, and further recommends the use of vessels that are essentially sterile. He suggests removing any traces of honey or nectar from equipment when processing is finished, using fastidious practices during the ripening of honey to avoid a high moisture content, and maintaining the storage temperature below 50F (lOC), or heating to 140-145F (60&63C) for 30 min to control the growth of undesirable spoilage microbes. Fermentation of honey can be prevented by storage at 10C (50F) or below with a relative humidity below 50% or by pasteurization (Tysset and Rousseau, 1981). Sugar-tolerant yeasts will not grow below 11C (52F) or above 38C (100F). Honey that contains more than 17% water is susceptible to fermentation and honey over 19% moisture is very likely to ferment. Heating honey to 63C (145F) for 30 min will destroy most of the yeasts responsible for fermentation (Crane, 1979, Graham, 1992). However, temperatures that are cool enough to deter spoilage are, unfortunately, cool enough to prolong bacterial survival. Current recommendations for heat-treating honey are probably based on inactivating yeasts and preventing fermentation. Different time and temperature combinations may be necessary to inactivate other types of microbes. Spores are particularly difficult to inactivate in honey because extreme heat treatment destroys the organoleptic qualities of honey. Holding food for 30 min at 63C (145F) is generally considered sufficient for the removal of vegetative microbes. Commer-

J.A. Snowdon,

D.O. Cliver /ht.

J. Food Microhio1og.t~ 31 (1996) l-26

19

cially prepared honey is often heated to 71C (160F) for about 30 min; yet viable microbes can be recovered from finished product. A given combination of time and temperature may be less effective in honey than other foods (such as milk) due to the absence of water. The number of effective combinations of time and temperature to remove undesirable organisms is very large. Inactivation of microbes is often spoken of in terms of thermal death time. Thermal death time is dependent upon: time, temperature, concentration of microbes, form (spore or vegetative) of microbe, stage in the microbial life cycle and health (stressed or robust) of the microbe. The nature of the food product must also be considered, e.g., sugars might protect microbes from inactivation. Effective thermal processes can be calculated for various situations; the assistance of a food microbiologist or food engineer is recommended. Aseptic packaging may avert microbiological problems after the honey has been properly processed. Other physical processes can kill microbes in honey. Shimanuki et al. (1984) eliminated viable Bacillus spores from honey via two 3 Mrad doses of high velocity electrons. Huhtanen (1991) determined the irradiation levels necessary for the inactivation of C. botulinum and B. subtilis spores in honey; the D values ranged from 1.91 to 12.8 kGy. Spoilage yeasts in honey are very sensitive to ultraviolet rays (Tysset and de Rautlin de la Roy, 1974). Ten minutes at 20 cm under a germicidal tube will inactivate microbes in the laboratory; a similar technique could be developed for use in industry for disinfection of a thin top layer of honey. Ultrafiltration will remove all bacteria. Nakano et al. (1989) tried to reduce or inactivate the spores of C. botulinum in honey. They studied heat shock, sonication, detergents, enzymes, alcohol, acids and bases; none had a significant effect. They suggest long term storage (presumably about 400 days) at 25C (77F) or mild heating at 65C (149F) for 5 days to eliminate or reduce C. botulinum spores in honey. Chemical disinfectants such as copper, iodine, mercury, quaternary ammonium salts or sodium hypochlorite may be used for cleaning food processing equipment (Tysset and de Rautlin de la Roy, 1974). Certain antibiotics - nystatin, actidione or cycloheximide, and amphotericin B could be used to prevent microbial growth. Preservatives such as benzoic acid, sorbic acid, sulfites, and CO, are listed as agents that control microbes (Jay, 1992). Preservatives such as sorbic acid or potassium sorbate and sodium or potassium benzoate are used in pie fillings, jams and jellies and may have use in controlling microbes in honey. More research, is definitely needed in this area. There is increasing interest and activity in the honey industry regarding the microbial content of honey. In the absence of information from the scientific literature, the honey industry is starting to generate its own data. Precise or comprehensive information is lacking, but general questions are being posed and introductory information is being developed. As a consequence control methods are being developed to reduce the levels of microbes in the finished product.

20

J.A. Snowdon, D.O. Cliver /ht. J. Food Microbiology

31 (1996) l-26

3. Analyzing

honey for microbes

No special techniques for the sampling or analysis of honey, for microbes of significance to human health or honey quality, have been developed and standardized. Standard methods for microbiological analysis can be adapted for honey. Details of these techniques are available in standard microbiology or food microbiology texts such as Jay (1992) or references such as the Bacteriological Analytical Manual (1984); Official Methods of Analysis (1990); the Codex Alimentarius (1992); Microorganisms in Foods (1982) and the Compendium of Methods for the Microbiological Examination of Foods (Vanderzant and Splittstoesser, 1992). There is little information in the scientific literature about detecting microbes in honey. More information can be expected if research is conducted and the results are made available. References in the scientific literature for the detection of microbes in honey (other than those pathogenic to bees) are primarily for the recovery of botulinal spores (Sugiyama et al., 1978; Midura et al., 1979; Flemming and Stojanowic, 1980; Hartgen, 1980; Huhtanen et al., 1981; Kautter et al., 1982; Stier et al., 1982; Guilfoyle and Yager, 1983; Hauschild and Hilsheimer, 1983; Kokubo et al., 1984; Aureli et al., 1985; Sakaguchi et al., 1987; Nakano and Sakaguchi, 1991; Nakano et al., 1989; and Du et al., 1991). Others who have reported microbial levels in honey include Tysset et al. (1970b) and Tysset and Rousseau, 1981. Bonvehi and Jorda (1993) examined a few different methods of determining aerobic colony counts. They maintain that a membrane filter method eliminates swarming and is more accurate. Standard techniques need to be adapted for use in honey. For example, traditional media for detecting foodborne fungi may underestimate fungi capable of growing at reduced water activity; some media for moderately osmophilic yeasts and molds may also detect other types of yeasts and molds (personal communication, Dr. Larry Beuchat, University of Georgia). Research to identify optimal analytical techniques for the recovery of yeasts and molds from honey is warranted.

4. Types and levels of microbes expected in honey 4.1.

Data

currently available

Information on the levels of microbes in honey, as reported in the scientific literature, is summarized in Table 4. The median value of 14 honey samples from France, surveyed by Tysset et al. (1970b) was less than 100 cfu/g. Tysset and Rousseau, 198 1 examined 175 samples of honey from different regions of France and found a mean total plate count of 227 cfu/g. Thirty-five retail samples of honey of international origin were found to contain a range of O-72 and a mean of 24 cfu/g (Nakano and Sakaguchi, 1991). An additional 18 samples from 5 lots of Argentinean honey (receiving additional scrutiny for high clostridial spore

J.A. Snowdon, D.O. Cliver /ht. J. Food Microbiology 31 (1996) 1-26

21

counts) ranged from 21-260 cfu/g. If these data, and the data in Table 4, are representative of current product in the US, one might generally conclude that total plate counts will vary from 10 to 10000 cfu/g of honey with low numbers possible in the finished product. In addition to the data on total plate counts presented above, some of the reports provide information on selected microbial species. Root (1983) recounts a Canadian study on yeasts and molds where as few as 1 in 10 g and as many as a million spores per gram were found among the 320 samples tested (the fermentation status was not mentioned). In a survey of 14 samples of French honey Tysset et al. (1970b) found an average yeast and mold count of 254 cfu/g (range o-2500). French researchers studying 175 honey samples found an average of 90 yeast and mold cfu/g with a range of 0 to 2500 (Tysset and Rousseau, 1981). Counts for osmophilic yeasts had a were similar, but ranged up to 10500. An average of 9 yeast colonies per gram of honey was found in 35 retail samples tested by Nakano and Sakaguchi (1991). It appears that while yeasts and molds can grow to high numbers in honey, microbial levels are controlled by standard industry practices that prevent fermentation. Bacillus spore counts reported by two research groups suggest that levels of BacilZus spores in honey are likely to be below 200 per gram. Tysset et al. (1970b) found less than 100 spores, and Nakano and Sakaguchi (1991) from 0 to 180 cfu/g. This is in contrast to recent industry reports (unpublished data) which include at least one sample where counts of BacilluS were closer to 10000/g; these high counts are thought to be the exception, not the norm. Great variation in levels of Bacillus spores among samples has also been observed. Tysset et al. (1970b) also looked for spores of Clostridium, but found none in any of their 14 samples. From 1 to 132 anaerobic microbes (likely to be Clostridium) were found in 35 retail samples of honey (Nakano and Sakaguchi, 1991). Typically, C. botulinum spores are found at levels ( l/g of honey (Sugiyama et al., 1978; Stier et al., 1982; and Nakano et al., 1989). In rare instances, a sample of honey has been found to contain from 36660 C. botulinurn spores/g (Midura et al., 1979; Nakano et al., 1989). Information from industrial experience is increasing but is not widely available. A compilation of six different reports on total plate counts in honey from laboratories in the US showing an overall mean value of 2270 cfu/g is summarized in Table 4. No coliforms, E. coli, yeasts, molds, Staphylococcus or Salmonellu were found in these six examples. Total plate counts of finished products observed in informal industry reports tend to range in 104-lo5 cfu/g of honey. Total plate counts in raw product are more variable and can reach 104-lo5 cfu/g. Additionally, specific microbes can appear in raw honey in high numbers (104-lo5 cfu/g) on an occasional and erratic basis. Bacillus species and Pseudomonas have both been found at high levels (thousands per gram) in raw honey. It is unknown if the microbes originate from primary or secondary sources.

22

J.A. Snowdon,

D.O. Ciiver 1 ht.

J. Food Microbiology

31 (1996) l-26

4.2. Expected impact of processing upon microbes in honey Honey prepared for table use is subjected to minimal processing. It is typically heated to as much as 70C (160F) for 30 min. Strained honey is passed through a 1.50pm screen, while filtered honey is often passed through a filter with 1 pm pores. The process of heating and filtering may reduce or eliminate many microbes. Honey used as an ingredient is subjected to several different processes. Sold in bulk, it has had similar treatment as honey for table use. Some honey is dried to crystalline form before being sold as an ingredient. Additional treatment is dependent upon the food with which the honey is combined. Most of the processes will inactivate any microbes present, usually due to high heat. New products and new processes all need to be evaluated from the view of microbiological safety. The cereal and bakery industries are the two largest consumers of honey. Honey is also used in condiments, salad dressing, barbecue sauce and peanut butter. Dairy, meat, beverage, snack and candy manufacturers also use honey as an ingredient. Short summaries of the processes associated with foods where honey is used as an ingredient are presented in Table 7. Honey is a product that is free of most microbes, and those microbes that may be present are likely to be in very low
Table 7 Processing

of foods

containing

honey

l BREAD: In bread, honey is added to the dough as part of the fermentation process. The dough is held at 36C (96F) at 80&85%1 humidity for IO-25 min. This is followed by baking at 199216C (390&420F) for 20 min; the internal temperature of the bread reaches 96C (204F). 0 CANDIES: The manufacture of honey candies usually involves blending and heat treatment. Heat treatment may include holding at lO4- I 16C (220&240F) for 30-40 min. These tempera tures will inactivate many microbes. 0 CEREAL: In some types of cereal production, honey is incorporated into the cereal and baked at temperatures exceeding 149C (300F). In other types of cereal, honey is blended into a slurry which is used to coat the cereal. The temperature of the slurry is between 82288C (180-190F). The coated cereal enters a dryer for lo-30 min at 999104C (2lO-220F). 0 GRANOLA: In making granola, honey is mixed with butter and sweeteners and heated to 70C (IWF). The mixture is cooled to 45550C (I l3- 122F) mixed with the granola, cooled to 4.441OC (40&50F) and packed at about 15C (59F). 0 HAM: In making a honey-cured ham, the meat is injected with a solution that includes honey. The meat is held at approximately 73C (164F) for 8 h. The meat is then chilled and prepared for sale. Other meat products, such as a honey beef stick, are made with an emulsion which contains honey. The emulsion is heated to 60C (140F) before it is packed into a casing. These processes will inactivate many microbes. Fresh sausage with honey is not heated at all; it remains refrigerated at about 3.3C (38F) and is intended to be cooked thoroughly before being eaten. 0 PEANUT BUTTER: Honey and other dry ingredients are added to ground peanuts in making peanut butter. The mixture is ground, oxygen is removed and the product is heated to 88C (190F). It is then cooled and packed. A number of conditions (absence of oxygen, low water activity, heat) make microbial persistence unlikely. 0 SALAD DRESSING AND MUSTARD are generally not heat processed. Honey would be blended in with agitation. The factors that ensure microbial stability of the dressing or mustard (e.g. preservatives, low water activity) are not likely to be affected by the addition of honey.

(Personal

communication,

Jill Clark,

Dutch

Gold

Honey,

Lancaster,

PA.).

J.A. Snowdon,

D.O. Clicer i hr. J. Food Microbiology

31 (1996) l-26

23

numbers. Most of the processes involve heat treatments that would inactivate any microbes contributed by honey. Control of microbial growth in products that are not heated depends on other anti-microbial procedures or agents. 4.3. Comments on microbial speciJications Purchase specifications for microbes in honey vary widely (Table 1). They are often based on information pertinent to other foods. Tysset et al. (1970b) suggested that honey is no longer marketable if it contains over 1000 yeast spores/g. The authors do suggest that such honey could be used in products requiring high heat (e.g. making candy). Tysset et al. (1970b) recommended measuring total counts, fecal contamination (via testing for E. coli or Streptococcus D (sic)), and anaerobic sulfite reducers (as an indicator of C. perf%zgens). The authors of this paper comment on current specifications below. (1) The standard plate count provides very general information and is useful as a point of comparison to other data and as a general indicator of the microbial quality of honey. Additional testing to identify the type of microbe present is often needed in conjunction with high standard plate counts. Honey with fairly high standard plate counts (10 000/g) could be acceptable if other microbial criteria (e.g., indicating presence of yeast or freedom from fecal contamination) were satisfied. (2) Yeasts and molds can be expected to be found in honey on a regular basis, but levels can be controlled by standard industry practices. A routine count for yeasts and molds combined would be an economical way to indicate the quality and condition of honey, as well as predict shelf life and spoilage potential. As mold counts are typically low, their inclusion may not be necessary; however, some commercial laboratories only provide the combined tests, making it more convenient to count both types of microorganisms. (3) Coliform counts are an indicator of sanitary practices. Since fecal contamination of honey has not been reported, and growth of any associated microbes does not appear possible, an assay for coliforms could also be used as a general indicator of fecal contamination as well as sanitation. As such, it could replace tests for specific pathogens of fecal origin (such as Salmonella). (4) E. coli counts are a specific indicator of fecal contamination and, consequently, sanitation relative to fecal contamination. As fecal contamination is not regularly associated with honey, this assay may not be necessary. (5) Although vegetative bacterial cells such as Salmonella may survive in honey, they will not multiply, and there have been no reports of their presence in honey. Vegetative bacterial cells can be controlled by thermal processing. Testing for Salmonella could be applicable for those that are producing products with no heat treatment. A general indicator of sanitation might make this assay unnecessary. (6) Staphylococcus is unable to reproduce in honey; hence, its toxin will not be produced in honey. The chances of a food handler introducing Stuphylococcus into honey are no greater than for any other food and are best controlled by

24

J.A. Snowdon,

D.O. Cliuer 1 ht.

J. Food Microbiology

31 (1996) 1-26

good manufacturing practices. No vegetative pathogenic species of bacteria have ever been recovered from honey. Hence, testing for this microbe or its toxin is not indicated. (7) The spores of bacteria in the genera Bacillus and Clostridium can be expected to be found in honey. Although typically found below about 200 cfu/g, reasonably high levels (thousands per gram) of Bacillus could be present with impurity depending upon the species and the nature of the product. If present, C. botulinum spores are typically found at levels of < 1 cfu/g, but are infrequently found as high as 60 cfu/g. (8) When a particular microbe becomes a problem, additional special testing for identification and confirmation may be necessary. In short, one scheme for testing would be a total plate count, a count for yeasts and molds, a coliform count and a count for total spore-forming bacteria. Additional tests ~ to explain unusually high counts, measure osmophilic microbes, or address a certain problem - may be needed. The use of honey in products that receive no or limited heat treatment may require additional tests. Since the information reported in the scientific literature is limited, these ideas should be revised after industry data are collected. Research in the areas of microbial quality in honey is needed to answer more of these questions with accuracy and precision.

Acknowledgements

The authors would like to thank L. Beuchat and H. Shimanuki for their review of this manuscript.

References
Allen, K.L., Molan, P.C. and Reid, G.M. (1991) A survey of the antibacterial activity of some New Zealand honeys. J. Pharm. Pharmacol. 43, 817-822. Aureli, P., Ferrini, A.M. and Negri, S. (1985) Ciostridiztm botulinurn spores in honey. Riv. Sot. Ital. Sci. Aliment. 12, 457-466. Bacteriological Analytical Manual, 6th edn. (1984) Assoc. Off. Anal. Chem., Washington, D.C. Bonvehi, J.S. and Jorda, R.E. (1993) The microbiological quality of honey as determined by aerobic colony counts. J. Food Protect. 56, 336-337. Codex Alimentarius, 2nd edn. (1992) United Nations, Rome. Crane, E. (1979) Honey: A Comprehensive Survey, Heinemann, London. De Wael, L., De Greef, M. and Van Laere, 0. (1990) The honeybee as a possible vector of Erwinia amylovora. Acta Hort. 273, 107-l 13. Du, S., Cheng, C., Lai, H. and Chen, L. (1991) Combined methods of dialysis, cooked meat medium enrichment and laboratory animal toxicity for screening Clostridium botulinurn spores in honey and infant food. Chin. J. Microbial. Immunol. 24, 240-247. El-Leithy, M.A. and El-Sibaei, K.B. (1972) Role of microorganisms isolated from honey-bees (Apis mellifera) in ripening and fermentation of honey. Egypt. J. Microbial. 7, 89-95. Esteban Quilez, M.A. and Marcos Barrado, A. (1976) Water activity of honey and the growth of osmotolerant yeasts. Anal. Bromatol. 28, 33-44.

J.A. Snowdon,

D.O. Cliver / Int. J. Food Microbiology

31 (1996) l-26

25

Flemming, R. and Stojanowic, V. (1980) Untersuchungen von Bienenhonig auf Ciostridium hotulinum Sporen. Arch. Lebensmittelhyg. 31, 179- 180. Furuta. T. and Okimoto, Y. (1978) Further investigations on honey yeasts. Bull. Fat. Agr., Tamagawa Univ. 18, 32-38. Gilliam, J. (1971) Microbial sterility of the intestinal content of the immature honey bee, Apis me/l@a. Ann. Entomol. Sot. Am. 64, 315-316. Gilliam, M. (I 978) Bacteria belonging to the genus Bacillus isolated from selected organs of queen honey bees, Apis mellifera. J. Invert. Pathol. 31, 3899391. Gilliam, M. and Prest, D.B. (1987) Microbiology of feces of the larval honey bee, Apis mell~fhra. J. Invert. Pathol. 49, 70-75. Gilliam, M. and Valentine, D.K. (1976) Bacteria isolated from the intestinal contents of foraging worker honey bees, Apis mellifera: the genus Bacillus. J. Invert. Pathol. 28, 2755276. Gilliam, M., Moffett, J.O. and Kauffeld, N.M. (1983) Examination of floral nectar of citrus, cotton and Arizona desert plants for microbes. Apidologie 14, 299-302. Gilliam, M., Lorenz, B.J. and Richardson, G.V. (1988) Digestive enzymes and micro-organisms in honey bees, Apis mellifrra: influence of streptomycin, age, season and pollen, Microbios 55. 95- 114. Graham, J.M. (Ed.) (1992) The Hive and the Honey Bee, Dadant and Sons, Hamilton, Illinois, Guilfoyle, D.E. and Yager, J.F. (1983) Survey of infant foods for Ciostridium botulinurn spores. J. Assoc. Off. Anal. Chem. 66, 1302- 1304. Haisig, M. and Kamburov, G. (1966) Yeast from the larvae of healthy honey bee colonies. Vet. Arh. 36, 66669. Hartgen, V.H. (1980) Untersuchungen von Honigproben auf Botulinustoxin. Arch. Lebensmittelhyg. 31, 1777178. Hauschild, A. and Hilsheimer, R. (1983) Detection of Clostridium botulinurn in honey by a procedure involving membrane filtration. J. Can. Inst. Food Sci. Technol. 16, 2566258. Hauschild, A.H.W., Hilsheimer, R., Weiss, K.F. and Burke, R.B. (1988) Clostridium botulinurn in honey, syrups and dry infant cereals. J. Food Protect. 51, 8922894. Huhtanen, C.N., Knox, D. and Shimanuki, H. (1981) Incidence and origin of Clostridium botulinurn spores in honey. J. Food Protect. 44, 812-814. Huhtanen, C.N. (1991) Gamma radiation resistance of Clostridium botulinurn 62A and Bacihs sub/i/is spores in honey. J. Food Protect. 54, 8944896. Jay, J.M. (1992) Modern Food Microbiology, 4th edn. Van Nostrand Reinhold, New York, New York. Kautter, D.A., Lilly, T., Jr., Solomon, H.M. and Lynt. R.K. (1982) Clostridium botulinurn spores in infant foods: a survey. J. Food Protect. 45, 102881029. Kokubo, Y., Jinbo, K., Kaneko, S. and Matsumoto, M. (1984) Prevalence of spore-forming bacteria in commercial honey. Ann. Rep. Tokyo Metr. Res. Lab. Pub. Health. 35, 192-196. Microorganisms in Foods. (1982). vol. I Their Significance and Methods of Enumeration, International Commission on Microbiological Specifications for Foods. University of Toronto Press, Toronto. Midura, T.F., Snowden, S., Wood, R.M. and Arnon, S.S. (1979) Isolation of Clostridium botulinum from honey. J. Clin. Microbial. 9, 282-283. Molan, P. (1992a) The antibacterial activity of honey. 1. The nature of the antibacterial activity. Bee World 73, 5-28. Molan, P. (1992b) The antibacterial activity of honey. 2. Variation in the potency of the antibacterial activity. Bee World 73, 59976. Molan, P. (1992~) Medicinal uses for honey. Beekeepers Quarterly. No. 26. Waikato University, New Zealand. Morse, R. and Hooper, T. (1985) The Illustrated Encyclopedia of Beekeeping. E.P. Dutton, Inc. NY, NY. Nakano, H. and Sakaguchi, G. (1991) An unusually heavy contamination of honey products by Clostridium botulinurn type F and Bacillus alvei. FEMS Microbial. Lett. 79, 171-178. Nakano. H., Okabe, T., Hashimoto, H. and Sakaguchi, G. (1989) Incidence of Clostridium botulinurn in honey of various origins. Jpn. J. Med. Sci. Biol. 43, 183-195. Nakano, H., Yoshikuni, Y., Hashimoto, H. and Sakaguchi, G. (1992) Detection of Clostridium botulinurn in natural sweetening. Int. J. Food Microbial. 16, 117-121. Official Methods of Analysis, 15th edn. (1990) Assoc. Off. Anal. Chem., Washington, D.C.

26

/.A. Snowdon, D.O. Cliaer / ht. J. Food Microbiology 31 (1996) 1-26

Piana, M.L., Poda, G., Cesaroni, D., Cuetti, L., Bucci, M.A. and Gotti, P. (1991) Research on microbial characteristics of honey samples of Udine province. Riv. Sot. Ital. Sci. Aliment. 20, 293-301. Radwan, S.S., El-Essawy, A.A. and Sarhan, M.M. (1984) Experimental evidence for the occurrence in honey of specific substances active against microorganisms. Zbl. Mikrobiol. 139, 249-255. Root, A.I. (Ed.). (1983) The ABC and XYZ of Bee Culture. The A. I. Root Co., Medina, OH. Sackett, W.G. (1919) Honey as a carrier of intestinal diseases. Col. St. Univ. Agr. Expt. Sta., Ft. Collins, Colorado, pp. 18. Sakaguchi, G., Sakaguchi, S., Kamata, Y., Tabita, K., Asao, T. and Kozaki, S. (1987) Distinct characters of Clostridium botulinum type A strains and their toxin associated with infant botulism in Japan. Int. J. Food Microbial. 11, 231-241. Shimanuki, H. and Knox, D.A. (1991) Diagnosis of Honey Bee Diseases. U.S.D.A., Agriculture Handbook No. AH-690, pp. 53. Shimanuki, H., Herbert, E.W., Jr. and Knox, D.A. (1984) High velocity electron beams for bee disease control. Apicultural Res. 124(12), 865-867. Smith, M.R., McCaughey, W.F. and Kemmerer, A.R. (1969) Biological effects of honey. .I. Api. Res. 8, 99%110. Snowdon, J.A. (1991) Infant Botulism, National Honey Board, Longmont, CO. Stier, R.F., Ito, K.A. and Stevenson, K.E. (1982) Methods for determining Clostridium botulinurn spores in honey. Co-operative Agreement No. 58-3244-9-94 for U.S. Department of Agriculture, SEA-AR, NER, Eastern Regional Research Center, Philadelphia. Sugiyama, H., Mills, D.C. and Kuo, L-J.C. (1978) Number of CIostridium botulinurn spores in honey. J. Food Protect. 41, 848-850. Troller, J.A. (1979) Food spoilage by microorganisms tolerating low-A, environments. Food Technol. 33, 72-75. Tysset, C. and Durand, C. (1968) Contribution to the study of the intestinal microbism of healthy foraging bees: inventory of the gram-negative bacterial populations, fourth report. Bull. Apicole I I, 107-118. Tysset, C. and Durand, C. (1973) On the survival of some gram negative, non-sporulated bacteria in commercial honey. Bull. Acad. Vet. Fr. 46, 191~196. Tysset, C. and Durand, C. (1976) Survival of enterobacteria in honey stored at 10C. Bull. Acdd. Vet. Fr. 49, 417-422. Tysset, C. and de Rautlin de la Roy, Y. (1974) Assays on the study of osmophilic yeasts, organisms causing fermentations of honey collected in France. Ass. Diplom. Microbial. Fat. Pharm. Univ. Nancy Bull. 134, l-26. Tysset, C. and Rousseau, M. 1981, Problem of microbes and hygiene of commercial honey. Rev. Med. Vet. 132, 591-600. Tysset, C., Brisou, J., Durand, C. and Malaussene, J. (1970a) Contribution to the study of intestinal microbial infection of healthy honey bees (Apis mellijkra): inventory of bacterial populations by negative gram. Ass. Diplom. Microbial. Fat. Pharm. Univ. Nancy Bull. 116, 41-53. Tysset, C., Durand, C. and Taliergio, Y.P. (1970b) Contribution to the study of the microbial contamination and the hygiene of commercial honey. Rec. Med. Vet. 146, 1471-1492. Tysset, C., Haas, P. and Durand, C. (1979) Survival of some mycobacteria in honey stored at room temperature. Bull. Acad. Vet. Fr. 52, 447-452. Vanderzant, C. and Splittstoesser, D.F. (Eds.) (1992) Compendium of Methods for the Microbiological Examination of Foods, 3rd edn. American Public Health Association, Washington, D.C. White, J.W., Jr., Subers, M.H. and Schepartz, A.T. (1962a) The identification of inhibine the antibacterial factor in honey, as hydrogen peroxide and its origin in a honey glucose-oxidase system. Biochim. Biophys. Acta. 73, 57-79. White, J.W., Jr., Riethof, M.L., Subers, M.H. and Kushnir, I. (1962b) Composition of American Honeys. Tech, Bull 1261, Agricultural Research Service, U.S. Department of Agriculture, Washington, DC. White, P.B. (1921) The normal bacteria1 flora of the bee. J. Pathol. Bacterial. 24, 64-78. Willix, D.J., Molan, PC. and Harfoot, C.G. (1992) A comparison of the sensitivity of wound-infecting species of bacteria to the antibacterial activity of manuka honey and other honey. J. AppI. Bact. 73, 388-394.