International Journal of Food Microbiology 79 (2002) 105 112 www.elsevier.

com/locate/ijfoodmicro

Survival of Escherichia coli O157:H7 in traditional African yoghurt fermentation

B.A. Ogwaro, H. Gibson, M. Whitehead, D.J. Hill *
School of Applied Sciences, University of Wolverhampton, Wulfruna Street, Wolverhampton WV1 1SB, United Kingdom Received 15 March 2002; received in revised form 26 April 2002; accepted 29 April 2002

Abstract Growth and survival of a nontoxigenic strain of Escherichia coli O157:H7 (ATCC 43888) was determined in traditionally fermented pasteurized milk. Preheated milk was inoculated with 1% (v/v) of a mixed culture of Lactobacillus delbrueckii ssp. bulgaricus (NCIMB 11778) and Streptococcus salivarius ssp. thermophilus (NCIMB 110368) and incubated at 25, 30, 37 or 43 jC for 24 h. E. coli O157:H7 (105 CFU/ml) were introduced into the milk pre- and post-fermentation. Fermented milk samples were subsequently stored at either 4 jC (refrigerator temperature) or 25 jC (to mimic African ambient temperature) for 5 days. After 24 h of fermentation, the pH of the samples fermented at the higher temperatures of 37 43 jC decreased from 6.8 to 4.4 4.0 ( F 0.2) whereas at the lower temperature of 25 jC, the pH decreased to pH 5.0 F 0.1. During this period, viable counts for E. coli O157:H7 increased from 105 to 108 109 CFU/ml except in milk fermented at 43 jC wherein viability declined to 104 CFU/ml. In fermented (25 30 jC) milk stored at 4 jC for 5 days, E. coli O157:H7 viability decreased from 108 9 to 106 7 CFU/ml whereas milk fermented at 43 jC resulted in loss of detectable cells. In contrast, storage of fermented milk samples at 25 jC for 5 days eventually resulted in complete loss of viability irrespective of fermentation temperature. Stationary phase E. coli O157:H7 inoculated post-fermentation (25 and 43 jC) survived during 4 jC storage, but not 25 jC storage. Fermentation temperature and subsequent storage temperature are critical to the growth and survival of E. coli O157:H7 in traditional fermented products involving yoghurt starter cultures. D 2002 Elsevier Science B.V. All rights reserved.
Keywords: Escherichia coli O157:H7; Lactobacillus delbrueckii ssp. bulgaricus; Streptococcus salivarius ssp. thermophilus; pH; Temperature; Fermentation; Acid

1. Introduction Milk is a major component of the traditional diet among the pastoralist in many regions of Africa. In these communities, most of the milk produced is consumed in the home and is rarely sold. However,

Corresponding author. Tel.: +44-1902-322161; fax: +441902-322680. E-mail address: d.hill@wlv.ac.uk (D.J. Hill).

high temperatures and lack of refrigeration facilities has led to the inability to produce and store fresh milk. Hence, conversion of any surplus liquid milk to stable products such as yoghurt, cheese, acidified milk, butter, and ghee has been done at the household level. Milking is done by hand in an open yard under poor hygienic conditions. The fresh raw milk is allowed to sour in traditional vessels such as animal skins, gourds (Olea africana) or in wooden or earthenware vessels. Acidification develops and is usually complete after 2 3 days at room temperatures, which, for instance,

0168-1605/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 0 5 ( 0 2 ) 0 0 1 8 4 - 8

106

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112

varies from 22 jC in Kenya to 35 jC in Northern Benin (FAO, 1990). Under these conditions, milk can be contaminated from sources such as diseased animals, the milk handler or contaminated equipment. Although acidic foods such as fermented milk products have been considered intrinsically safe due to their low pH and high acidity, food-borne pathogens such as Escherichia coli, particularly the serotype O157:H7 have been implicated in food poisoning resulting from the consumption of such foods. E. coli O157:H7 was first recognized as a pathogen in 1982 (Riley et al., 1983; Griffin and Tauxe, 1991) and is considered as an important causative agent of diarrhoea, hemorrhagic colitis and hemolytic uremic syndrome (HUS) (Karmali et al., 1985; Griffin and Tauxe, 1991). Browning et al. (1990) reported that E. coli O157 was first isolated from a man in Johannesburg in 1990. However, since then there have been several incidences of E. coli O157 in Africa (Cunin et al., 1999; Effler et al., 2001). Beef, untreated water, and unpasteurized milk were suggested sources of contamination. E. coli O157:H7 has been implicated in the consumption of acidic foods such as apple cider (Besser et al., 1993), yoghurt (Morgan et al., 1993), or mayonnaise (Weagan et al., 1994). These outbreaks suggest that E. coli O157:H7 may be tolerant to the acidic conditions found in yoghurt. It has been observed that E. coli O157:H7 inoculated into yoghurt may remain viable from a few hours at 30 40 jC, and up to 1 8 days while refrigerated (Massa et al., 1997). This implies that even the use of pasteurized milk can pose a health risk if the milk is poorly handled and recontaminated. The aim of this study is to investigate the growth and survival of E. coli O157:H7 during fermentation at different temperatures and subsequent storage conditions typical of African and Western yoghurt production. These comparisons could serve as a basis for recommendations leading to the improvement of traditional African milk fermentation.

salivarius ssp. thermophilus NCIMB 10378 used throughout this project were obtained from the National Culture Collection. L. delbrueckii ssp. bulgaricus was maintained on De Man Rogosa Sharpe Agar (MRSA) (De Man et al., 1960) and S. salivarius ssp. thermophilus was maintained on M17 Agar (Terzaghi and Sandine, 1975) slopes, respectively, at 4 jC. A nontoxigenic strain of E. coli O157:H7 ATCC 43888 was maintained on Trypticase Soy Agar (Difco Laboratories, Detroit, MI, USA) slants at 4 jC and subcultured monthly. 2.2. Culture and enumeration methods Cultures were grown overnight in Trypticase soy broth (TSB) for E. coli O157:H7; De Man Rogosa Sharpe broth (MRSB) for L. delbrueckii ssp. bulgaricus in an anaerobic gas jar and M17 broth for S. salivarius ssp. thermophilus at 37 jC with shaking (150 rpm) and subcultured once before use in the experiments. Growth and survival of E. coli O157:H7 was determined in traditionally fermented pasteurized milk. Duplicate experiments were conducted on pasteurized full cream milk. For each experiment, two samples of preheated milk (85 jC, 30 min) were inoculated, after cooling, with 1% (v/v) of a mixed culture of L. delbrueckii ssp. bulgaricus and S. salivarius ssp. thermophilus. One sample was co-inoculated with 105 CFU/ml of an overnight (stationary phase) culture of E. coli O157:H7. Uninoculated, pasteurized milk served as a control. Each milk treatment was incubated either at 25, 30, 37 or 43 jC for 24 h. After fermentation, the sample containing lactic acid culture only was inoculated with 105 CFU/ml of E. coli O157:H7 (post-fermentation contamination), and then each treatment was stored at either 4 jC (refrigerator temperature) or 25 jC (to mimic African ambient temperature) for 120 h. Viable counts for E. coli O157:H7 were determined by plating on Fluourocult R E. coli O157:H7 Agar (Merck, Darmstadt, Germany) at 0, 2, 4, 6, 8, and 24 h during fermentation and at 0, 24, 48, 72 and 120 h after fermentation. Changes in pH were measured in these samples using a Delta 320 pH meter at each time interval.

2. Materials and methods 2.1. Bacterial strains Lactic starter cultures Lactobacillus delbrueckii ssp. bulgaricus NCIMB 11778 and Streptococcus

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112

107

3. Results 3.1. Growth and survival of E. coli O157:H7 in pasteurized full cream milk The growth of E. coli O157:H7 in pasteurized milk incubated at 25, 30, 37 and 43 jC for 24 h in the absence of lactic starter cultures, or during traditional lactic fermentation are shown in Figs. 1 and 2, respectively. E. coli O157:H7 showed similar growth patterns in response to fermentation temperature in both culture conditions. However, differences were observed in the rate of growth of E. coli O157:H7 between the two cultures. At 25 jC, E. coli O157:H7 grew faster in the absence of lactic starter culture (doubling time c 1.09 h) than during lactic fermentation (doubling time c 1.70 h), to similar cell densities of 108 CFU/ml during the initial 24-h incubation stage. At this temperature, the pH of the milk declined

Fig. 2. Change in numbers of E. coli O157:H7 and pH during fermentation of full cream pasteurized milk by L. delbrueckii ssp. bulgaricus, S. salivarius ssp. thermophilus at 25 jC (y), 30 jC (o), 37 jC (E), or 43 jC (n).

Fig. 1. Change in numbers of E. coli O157:H7 and pH during growth in full cream pasteurized milk at 25 jC (y), 30 jC (o), 37 jC (E), or 43 jC (n).

from an initial 6.8 ( F 0.1) to 6.2 ( F 0.2) in the absence of lactic starter culture and to 5.1 ( F 0.2) during lactic fermentation. In the absence of lactic starter culture at 30 or 37 jC, E. coli O157:H7 grew from 105 CFU/ml to approximately 1010 CFU/ml during the initial 12 h of incubation (doubling times of 0.78 and 0.56 h, respectively) and subsequently maintained this population size. During this period, the pH of the milk declined from 6.8 ( F 0.2) to 5.2 and 4.9, respectively (Fig. 1). When cultured with lactic acid bacteria (Fig. 2), E. coli O157:H7 grew more slowly (doubling times of 1.35 and 1.17 h, respectively) and achieved lower cell yields (approximately 109 CFU/ ml) in the first 12 h of fermentation after, which the numbers slightly declined by 0.5 log units. The pH of the milk declined from 6.8 ( F 0.2) to 4.6 (30 jC) and 4.4 (37 jC).

108

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112

3.2. Survival of E. coli O157:H7 in fermented milk during storage at 4 and 25 jC Subsequent storage of the E. coli O157:H7 coinoculated, lactic fermented milk at 4 jC (Fig. 3) showed that over the 4-day period, the number of viable E. coli O157:H7 cells declined slightly (approximately 1 2 log units) in samples that had been fermented at 25 to 37 jC. In the milk fermented at 43 jC prior to storage, the E. coli O157:H7 viable cell population continued to decline to undetectable levels after just 3 days. In contrast, storage of the E. coli O157:H7 coinoculated lactic fermented milk at 25 jC (Fig. 4) resulted in a decrease of viable E. coli O157:H7 cells to undetectable levels over the 5-day storage period. The rate of this viable population decline was directly

Fig. 3. Change in numbers of E. coli O157:H7 and pH during fermentation of full cream pasteurized milk by L. delbrueckii ssp. bulgaricus, S. salivarius ssp. thermophilus at 25 jC (y), 30 jC (o), 37 jC (E), or 43 jC (n) and subsequent storage at 4 jC. The arrow shows the point of storage.

In the absence of lactic starter culture at 43 jC, only 2 log units of E. coli O157:H7 growth were observed during the first 12 h of incubation and thereafter the number of viable cells subsequently declined by 1 log unit after 24 h of incubation. In the presence of lactic acid bacteria, no initial growth was observed and the E. coli O157:H7 population steadily declined from 105 to 103 4 CFU/ml after 24 h of incubation. The pH of the milk during this period decreased from pH 6.8 to approximately pH 4.0. The reduction in pH in the absence of lactic starter culture (Fig. 1) was solely attributed to E. coli O157:H7 as incubation of the milk without inoculation of cultures only resulted in an insignificant reduction in pH of 0.2 units.

Fig. 4. Change in numbers of E. coli O157:H7 and pH during fermentation of full cream pasteurized milk by L. delbrueckii ssp. bulgaricus, S. salivarius ssp. thermophilus at 25 jC (y), 30 jC (o), 37 jC (E), or 43 jC (n) and subsequent storage at 25 jC. The arrow shows the point of storage.

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112

109

related to the initial fermentation temperature. Storage of the fermented milk samples at 25 jC was also associated with a further increase in acidity of all the fermented milk samples resulting in final pH values in the range of 3.57 2.97 for milk originally fermented at 25 to 43 jC, respectively. 3.3. Survival of E. coli O157:H7 inoculated into fermented milk stored at 4 or 25 jC Milk was fermented at 25 or 43 jC and then inoculated with E. coli O157:H7 and stored at 4 jC for 5 days (Fig. 5). After 5 days of incubation, a slight reduction of 0.5 (25 jC fermentation) and 1.0 (43 jC fermentation) log reduction in viable E. coli O157:H7 cells occurred. During this period, the pH of the samples declined only by approximately 0.9 and by 0.3 units in samples fermented at 25 and 43 jC, respectively.

Fig. 6. Change in numbers of E. coli O157:H7 and pH of the sample during storage at 25 jC when inoculated after fermentation at 25 jC (y), or 43 jC (n). The arrow shows the point of inoculation.

Milk fermented at 25 or 43 jC and then inoculated with E. coli O157:H7 was also stored at 25 jC for 5 days (Fig. 6). In this environment, the pathogen failed to maintain its contamination level and was not recovered after 48 h (43 jC fermentation) or 120 h (25 jC fermentation) of storage. The pH of the fermented milk samples decreased further during storage at 25 jC to pH 3.8 and 3.0 ( F 0.2) for milk originally fermented at 25 and 43 jC, respectively.

4. Discussion In this study, we report the growth and survival of E. coli O157:H7 in a traditionally fermented yoghurt. Growth and survival of E. coli O157:H7 in fermented milk products have been reported in the cottage-cheese and cheddar-cheese manufacturing

Fig. 5. Change in numbers of E. coli O157:H7 and pH of the sample during storage at 4 jC when inoculated after fermentation at 25 jC (y), or 43 jC (n). The arrow shows the point of inoculation.

110

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112

processes (Arocha et al., 1992; Reitsma and Henning, 1996) and in studies of fermented milk drinks at 37 jC (Chang et al., 2000). In this study, we also observed growth of E. coli O157:H7 in full cream pasteurized milk during incubation at 25 to 37 jC during lactic fermentation. The failure of E. coli O157:H7 to grow in milk fermented at 43 jC is of significance, an effect also observed by Massa et al. (1997) when inoculating 103 or 107 E. coli O157:H7 into a similar yoghurt-type fermentation. When E. coli O157:H7 was grown as a monoculture, the growth in milk was substantial, but was significantly reduced when grown in the presence of a lactic starter culture of S. salivarius ssp. thermophilus and L. delbrueckii ssp. bulgaricus. These reduced growth rates in fermenting milk are attributed to acidification of the milk, but this does not explain the lower growth rates observed during the initial 6 h of fermentation and implies other factors to be involved. Other mechanisms including the production of bacteriocin, hydrogen peroxide, and ethanol by lactic acid bacteria might also contribute to the decline in the population of E. coli O157:H7 (Frank and Marth, 1988). The lactic fermentation temperature significantly affects the rate of acidification of full cream pasteurized milk and the subsequent survival of E. coli O157:H7. E. coli O157:H7 only substantially declined to undetectable levels when the pH of the milk decreased to less than pH 4.4, an effect only apparent during storage at 25 jC for 4 5 days or during fermentation at 43 jC. Thus, the decline in the viable population of this pathogen is wholly or partially attributed to the detrimental effect of the low pH environment. Similar findings for E. coli O157:H7 survival during the fermentation of diluted cultured milk drink have been reported (Chang et al., 2000). Contamination of the milk can occur during collection and preparation of milk, during fermentation and after fermentation. The results of this study show that the decline in E. coli O157:H7 viability during storage of fermented milk at 25 jC was not apparently influenced by the timing of contamination of the milk with this pathogen. Thus, E. coli O157:H7 inoculated before or after fermentation (Figs. 4 and 6) exhibited similar death curves. Although stationary phase cells were introduced into the milk in both

cases, inoculation prior to fermentation eventually lead to growth of these cells in an ever-increasing acidic environment. Arnold and Kaspar (1995) reported that stationary-phase cells of E. coli O157:H7 strains are significantly more acid resistant than log-phase cells, a response regulated by the alternate sigma factor r38 (McCann et al., 1991; Hengge-Aronis, 1993a,b; Cheville et al., 1996). In addition, Cheng and Kaspar (1998) reported that anaerobic growth of E. coli O157:H7 in an acid medium (artificial rumen fluid) resulted in the development of acid tolerance earlier than cells grown aerobically in this medium. Auger et al. (1989) reported that, this may be a result of the production or rpoS-regulated proteins or arginine decarboxylase, which is also induced during anaerobic growth at acidic pH. Lin et al. (1996) has since described the role of arginine decarboxylase in the development of acid tolerance in E. coli O157. Since fermenting milk becomes anaerobic and acidic during milk fermentation, the development of a similar acid tolerance effect would be expected to also occur in this study. However, since this would occur in both cells added before fermentation and in stationary phase cells added after fermentation (prior to storage), no apparent development of acid tolerance would be observed. In contrast, in milk fermented at 43 jC survival of E. coli O157:H7 at 4 jC was dependent upon the time of inoculation. Cells inoculated prior to fermentation declined more rapidly than cells inoculated postfermentation (Figs. 3 and 5), despite the reduction in pH being similar. It is concluded that exposure of E. coli O157:H7 cells to the high fermentation temperature exerts an additional stress upon the cells from which they are unable recover when maintained in an acidic environment. It is recognized, however, that since different strains of E. coli O157:H7 exhibit slightly different growth temperature optima (from 38.5 to 41.5 jC) (Gonthier et al., 2001) that strain differences may occur in their response to acidification at 43 jC.

5. Conclusions The study has shown that E. coli O157:H7 can substantially grow during a yoghurt fermentation

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112

111

when the fermentation temperature is between 25 and 37 jC, but not when at 43 jC. In relation to traditional African fermentation (fermentation temperature of 20 to 35 jC), the levels of fermentation acids produced are unlikely to be effective against E. coli O157:H7. However, when these products are stored at temperatures typical of traditional African production (around 25 jC), this is more likely to promote a safer product than that stored at 4 jC. It is concluded that the fermentation temperature is a critical control point in the production of yoghurt that is safe from microbiological hazard. This is especially important when the product is stored at 4 jC, a temperature that will control growth of E. coli O157:H7, but will facilitate survival of existing pathogens. Thus, the contaminating level at storage, and post fermentation contamination are important risk factors regarding the safety of the product. This is of particular importance when the higher fermentation temperature cannot be easily achieved or controlled as in traditional African fermentation.

References
Arnold, K.W., Kaspar, C.W., 1995. Starvation and stationary-phase induced acid tolerance in Escherichia coli O157:H7. Appl. Environ. Microbiol. 61, 2037 2039. Arocha, M.M., McVey, M., Loder, S.D., Rupnow, J.H., Bullerman, L., 1992. Behaviour of hemorrhagic Escherichia coli O157:H7 during the manufacture of cottage cheese. J. Food Prot. 55 (5), 379 381. Auger, E.A., Redding, K.E., Plumb, T., Childs, L.C., Meng, S.Y., Bennett, G.N., 1989. Construction of lac fusions to the inducible arginine and lysine decarboxylase genes of Escherichia coli K12. Mol. Microbiol. 3, 609 620. Besser, R.E., Lett, S.M., Weber, J.T., Doyle, M.P., Barrett, T.J., Wells, J.G., Griffin, P.M., 1993. An out break of diarrhea and haemolytic uremic syndrome from Escherichia coli O157:H7 in fresh pressed apple cider. J. Am. Med. Assoc. 2269, 2217 2220. Browning, N.G., Botha, J.R., Sacho, H., Moore, P.J., 1990. Escherichia coli O157:H7 haemorrhagic colitis report of the South African case. S. Afr. Med. J. 28, 28 29. Chang, J.-H., Chou, C.-C., Li, C.-F., 2000. Growth and survival of Escherichia coli O157:H7 during the fermentation and storage of diluted cultured milk drink. Food Microbiol. 17, 579 587. Cheng, C.-M., Kaspar, C.W., 1998. Growth and processing condi-

tions affecting acid tolerance in Escherichia coli O157:H7. Food Microbiol. 15, 157 166. Cheville, A.M., Arnold, K.W., Buchrieser, C., Chen, C.-M., Kaspar, C.W., 1996. rpoS regulation of acid, heat and salt tolerance in Escherichia coli. Appl. Environ. Microbiol. 62, 1822 1824. Cunin, P., Tedjouka, E., Germani, Y., Ncharre, C., Bercion, R., Morvan, J., Martin, P.M.V., 1999. An epidemic of bloody diarrhea in Escherichia coli O157 emerging in Cameroon? Emerg. Inf. Dis. 5 (2). De Man, J.C., Rogosa, M., Sharpe, M.E., 1960. A medium for the cultivation of Lactobacilli. J. Appl. Bacteriol. 23, 130 135. Effler, P., Isaacson, M., Arntzen, L., Heenan, R., Canter, P., Barrett, T., Lee, L., Mambo, C., Levine, L., Zaida, A., Griffin, M., 2001. Factors contributing to the emergence of Escherichia coli O157 in Africa. Emerg. Inf. Dis. 7 (5), 812 819. FAO, 1990. The technology of traditional milk products in developing countries. FAO Animal Production and Health Paper 85. FAO Publication, Rome. Frank, J.F., Marth, E.H., 1988. Fermentation. In: Wong, N.P., Jenness, R., Keeney, M., Marth, E.H. (Eds.), Fundamentals of Dairy Chemistry. Van Nostrand-Reinhold, New York, pp. 655 738. Gonthier, A., Guerin-Faublee, V., Tilly, B., Delignette-Muller, M.-L., 2001. Optimal growth temperature of O157 and non-O157 Escherichia coli strains. Lett. Appl. Microbiol. 33, 352 356. Griffin, P.M., Tauxe, R.V., 1991. The epidemiology of infections caused by Escherichia coli O157:H7, other enterohaemorrhagic E. coli and the associated hemolytic uremic syndrome. Epidemiol. Rev. 13, 60 98. Hengge-Aronis, R., 1993a. Osmotic regulation of rpoS-dependent genes in Escherichia coli. J. Bacteriol. 175 (1), 259 265. Hengge-Aronis, R., 1993b. Survival of hunger and stress: the role of rpoS in early stationary phase gene regulation in Escherichia coli. Cell 72, 165 168. Karmali, M.A., Perric, M., Lim, C., Fleming, P.C., Arbus, G.S., Lior, H., 1985. The association between idiopathic haemolytic uremic syndrome and infection by verotoxin-producing Escherichia coli. J. Infect. Dis. 151, 775 781. Lin, J., Smith, M.P., Chapin, K.C., Baik, H.S., Benneth, G.N., Foster, J.W., 1996. Mechanisms of acid resistance in enterohemorrhagic Escherichia coli. Appl. Environ. Microbiol. 62, 3094 3100. Massa, S., Altier, C., Quaranta, V., De Pace, R., 1997. Survival of Escherichia coli O157:H7 in yoghurt during preparation and storage at 4 jC. Lett. Appl. Microbiol. 24, 342 350. McCann, M.P., Kidwell, J.P., Martin, A., 1991. The putative y factor KatF has a central role in development of starvation mediated general resistance in Escherichia coli. J. Bacteriol. 173, 4188 4194. Morgan, D., Newman, C.P., Hutchinson, D.N., Walker, A.M., Rowe, B., Majid, F., 1993. Verotoxin-producing Escherichia coli O157:H7 infections associated with the consumption of yoghurt. Epidemiol. Infect. 111, 181 187. Reitsma, C.J., Henning, D.R., 1996. Survival of enterohemorrhagic Escherichia coli O157:H7 during the manufacture and curing of cheddar cheese. J. Food Prot. 59, 460 464.

112

B.A. Ogwaro et al. / International Journal of Food Microbiology 79 (2002) 105112 streptococci and their bacteriophages. Appl. Environ. Microbiol. 29, 807 813. Weagan, S.D., Bryant, J.L., Bark, D.H., 1994. Survival of Escherichia coli O157:H7 in mayonnaise and mayonnaise-based sauce at room and refrigerated temperatures. J. Food Prot. 57, 629 631.

Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B., Wells, J.G., Devis, B.R., Herbert, R.J., Olcott, E.S., Johnson, L.M., Hargrett, N.T., Blake, F.A., Cohen, M.L., 1983. Hemorrhagic colitis associated with a rare Escherichia coli serotype. New Engl. J. Med. 308, 681 685. Terzaghi, B.E., Sandine, W.E., 1975. Improved medium for lactic