140

steroid-sparing agent. Longer-term asthmatic patients are required. only

as

a

trials in

We thank the pharmacy department at the Brompton Hospital, particularly Ms K. Youssef and Dr Bryn Hughes: Glaxo Laboratories for supplying the diary cards and peak flow meters; and Lederele Laboratories for the methotrexate and matched placebo tablets. This study was supported by a grant from the Asthma Research Council. Participating centres and physicians: Brompton Hospital (D. Geddes, D. Barnes, M. Turner-Warwick, M. Green, J. Collins, T. Evans); Royal Free Hospital (S. Clarke, M. Johnson); Central Middlesex Hospital and Willesden General Hospital (M. McNicol, J. Riordan); Ealing (P. Ind, M. Rudolf); Croydon (R. Courtney-Evans, A. Miller); Kingston (G. Knowles); Frimley Park (R. Knight); Slough (M. Smith); St Albans (I. Williams); Hertford (R. Dent).

REFERENCES

1. Hargreave FE, O’Byrne PM, Ramsdale EH. Mediators, airway responsiveness and asthma. J Allergy Clin Immunol 1985; 76: 272-76. 2. Schleimer RP. The mechanisms of anti-inflammatory steroid action in allergic disease. Annu Rev Pharmacol Toxicol 1985; 25: 381-412. 3. Laitinin LA, Heino M, Laitinen A. Kava T, Haahtela T. Damage of the airway epithelium and bronchial rectivity in patients with asthma. Am Rev Respir Dis 1985; 131: 599-606. 4. Adinoff AD, Hollister JR. Steroid-induced fractures and bone loss in patients with asthma. N Engl J Med 1983; 309: 265-68. 5. Zeiger RS, Schatz M, Sperling W, Simon RA, Stevenson DD. Efficacy of troleandomycin in outpatients with severe corticosteroid-dependent asthma. J Allergy Clin Immunol 1980; 66: 438-46. 6. Bernstein EI, Bernstein IL, Bodenheimer SS, Pietruski RG. An open

study of auranofin in the treatment of steroid dependent asthma. J Allergy Clin Immunol 1988; 81: 6-16. 7. Hodges NG, Brewis RAL, Howell JBL. An evaluation of azathioprine in severe chronic asthma. Thorax 1971; 26: 734-39. 8. Mullarkey MF, Blumenstein BA, Andrade WP, Bailey GA, Olason I, Wetzel CE. Methotrexate in the treatment of corticosteroid-dependent asthma. N Engl J Med 1988; 318: 603-07. 9. American Thoracic Society. Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease (COPD) and asthma. Am Rev Respir Dis 1987; 136: 225-44. 10. Rees RB, Bennett JH, Maibach HJ. Methotrexate for psoriasis. Arch Dermatol 1967; 95: 2-11. 11. Wilke WS, MacKenzie AH. Methotrexate in rheumatoid arthritis. Current status. Drugs 1986; 32: 103-13. 12. Weinstein GD, Frost P. Methotrexate for psoriasis. A new therapeutic schedule. Arch Dermatol 1971; 103: 33-38. 13. Andersen PA, West SG, O’Dell JR, Via CS, Claypool RG, Kotzin BL. Weekly pulse methotrexate in rheumatoid arthritis. Clinical and immunological effects in a randomised double-blind study. Ann Intern Med 1985; 103: 489-96. 14. Lewis JH, Schiff E. Methotrexate-induced chronic liver injury: guidelines for detection and prevention. Am J Gastroenterol 1988; 88: 1337-45. 15. Health and Public Policy Committee, American College of Physicians. Methotrexate in rheumatoid arthritis. Ann Intern Med 1987; 107: 418-19. 16. Louie S, Lillington GH. Low dose methotrexate pneumonitis in rheumatoid arthritis. Thorax 1986; 41: 703-04. 17. Cannon GW, Ward JR, Clegg DD, Samuelson CO, Abbot TM. Acute . lung disease associated with low dose pulse methotrexate therapy in patients with rheumatoid arthritis. Arthritis Rheum 1983; 26: 1269-74. 18. Engelbrecht JA, Calhoon SL, Scherrer JJ. Methotrexate pneumonitis after low dose therapy for rheumatoid arthritis. Arthritis Rheum 1983; 26: 1275-78. 19. Sostman HD, Mathay RA, Putman CE, Smith CJ. Methotrexateinduced pneumonitis. Medicine (Baltimore) 1976; 55: 371-88.

Fermentation of cereals for reduction of bacterial contamination of weaning foods in Ghana

Unfermented and fermented maize dough weaning foods prepared by mothers in a Ghanaian village were examined for gram-negative bacilli (GNB) immediately after preparation and during to assess the antimicrobial effect of fermentation. GNB were cultured from all samples of unfermented dough (51) and from 16 of 51 samples of fermented dough. The extent of contamination was significantly higher in the unfermented dough than in fermented dough (5·9

storage

[SEM 0·1 ] vs 4·0 [0·4] log10 colony forming units/ g). After 6 h and 12 h storage, a higher proportion of samples of porridge made from unfermented dough contained GNB than did those made with fermented dough (45/51 vs 22/55; 49/51 vs 20/51, respectively) and levels of GNB were significantly higher in the porridge made from unfermented dough after 6 h (4·2 [0·2] vs 3·8 [0·2]). Fermentation of maize dough is an effective method to reduce contamination of maize dough weaning foods with GNB.

Introduction In many developing countries infant food is commonly stored for prolonged periods, when it often becomes contaminated with unacceptable levels of bacteria.1-5 The increase in diarrhoeal morbidity and mortality after the start of weaning has been attributed to the consumption of contaminated weaning foods.6,7 Improved hygiene during the preparation of weaning food has lately been

emphasised.8 Prevention of diarrhoeal illness has centred

on

advice to

cook fresh infant food at every meal, but this is practicable only in rich communities. In rural communities where food and fuel are short and where there is little time for cooking several meals a day, a method to inhibit bacterial growth during storage of infant food could be of considerable benefit to infant health and nutrition.

mothers

to

ADDRESSES Noguchi Memorial Institute, PO Box 25, Legon, Ghana (P. P A. Mensah, PhD), and Department of Clinical Sciences, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1 E 7HT, UK (A. M. Tomkins, FRCP, B. S. Drasar, DSc, T. J. Harrison, PhD). Correspondence to Dr A. M. Tomkins.

141

Among the many standard food preparation technologies, fermentation of milk produces microbiologically safer foods.9 However, this method has become less popular because of "modem" emphasis on consumption of freshly prepared food, and the use of fermented food has been discouraged by some health workers.10 Research into the

microbial safety of fermented cereal-based weaning foods has received little attention. We have previously simulated in the laboratory the unhygienic conditions of a traditional African village by inoculation of Shigella flexneri into fermented Ghanaian maize dough." Other investigators have shown that fermentation of cereals can inhibit diarrhoeal pathogens. 12,13 We here investigate the microbial quality of fermented and unfermented maize dough weaning foods prepared and stored by mothers in a rural Ghanaian community.

Subjects and methods Study area and subjects The study was carried out between January and October, 1988, in Ghanaian villages, Ashongman and Agbogba, which are 2 km apart. They are situated about 10 km from the Noguchi Memorial Institute, University of Ghana, where the laboratory tests were done. There was no organised sewage system, and water for drinking and cooking was obtained from unprotected wells and ponds. The population of the study area was 604 of whom about 120 were children under 4 years of age. The people were mostly subsistence farmers but some mothers were traders and some fathers were craftsmen. 63 mothers with children who ate fermented maize dough porridge as part of their normal diet were invited to take part in the study, which was approved by the Ghana National Ethical Committee. two

Maize dough porridge To prepare maize dough porridge, whole maize is first soaked for up to 24 h after which the water is drained. The maize is then milled into a flour which is used to prepare the maize dough (1part water to 3 parts maize flour). This dough is fermented for 24-72 h during which portions are removed and cooked as porridge (1 part fermented maize dough to 6 parts water). Food is stored in cooking pots or plastic containers. Enough food for the family is cooked in the morning and portions are eaten for up to 12 h after cooking. This porridge is first given to infants aged between 1 and 3 months.

Sample collection Food samples were collected between June and October, 1988. This coincided with the maize harvest during which there were abundant supplies of maize in every home. An average of 5 mothers took part in the study every week. All mothers obtained their own maize for grinding into meal. Field personnel (including P.M.) arrived at the village by 0600h on days designated for sample collection. Mothers ground maize into meal which they used to prepare fresh maize dough: we describe this as "unfermented maize dough", from which they then cooked a porridge. Enough porridge was cooked for infant feeding and sample collection. 500 g samples were taken immediately after cooking (0 h) and then after 6 and 12 h storage. Samples were collected into sterile containers on each occasion. Uncooked maize dough samples (500 g) were also collected. Similar samples were taken from maize dough that the mothers had allowed to ferment for 48 h, which we describe as "fermented maize dough". Each mother therefore provided two food samples which had each been prepared under the same conditions. The food samples were transported to the laboratory within 30 min of collection, and were inoculated onto bacteriological media immediately on arrival.

Laboratory methods The pH of each food sample was measured with a pH meter (TD Tokal Electronics, Japan) after mixing 10 g food with 10 ml distilled water. The pH of the sample was used to indicate whether the maize dough was fermented. Bacteriology. 0,001 ml of a suspension of maize dough and porridge (10 g food in 100 ml buffered peptone water) was inoculated into ’Selenite F’ broth for selective enrichment of Shigella and Salmonella species. These broths were subcultured onto ’Salmonella-shigella’ agar, ’MacConkey’ agar, and ’Xylose Lysine Deoxycholate’ agar after incubation for up to 18 h at 37°C. All gram-negative bacilli (GNB) were characterised with ’API 20E’ kits (API System S. A., La Balme les Grottes, Montalieu Vercieu, France). Two 25 g samples of each food were examined by the spread plate method .14,15 Ten-fold serial dilutions (25 g of food in 225 ml of sterile phosphate buffered saline [Oxoid, UK]) were inoculated onto MacConkey agar (Oxoid). Agar plates were incubated at 37°C for 18 to 24 h. Plates yielding no bacterial growth were incubated for another 24 h. MacConkey agar plates with 300 or less colonies were selected to calculate total GNB. In addition, GNB were selected and characterised with API 20E kits. 10 strains identified as E coli were emulsified in trypticase soy broth (Oxoid) and were transferred onto microtitre plates. Sterile glycerol (BDH chemicals, UK) was added after overnight incubation. These strains were stored at -20°C and tested for the presence of genes encoding heat labile (LT) as well as two heat stable (ST) enterotoxins (STal and STa2) of enterotoxigenic E coli (ETEC).

Detection of ETEC A DNA-DNA hybridisation assay was done with a modification16 of the method of Grunstein and Hogness.17 Briefly, stored E coli strains were transferred onto ’Hybond-N’ filters (82 mm diameter, Amersham International, UK) and placed on nutrient agar (Oxoid) by means of a template.18 After overnight incubation, the colonies were lysed to release the DNA which was denatured with 05 M sodium hydroxide and 15 M sodium chloride. After neutralisation, the single stranded DNA was covalently bonded onto the filters with an ultraviolet transilluminator. These procedures were carried out in Ghana. The filters were stored under vacuum until required for DNA-DNA hybridisation with radiolabelled probes in London. Before hybridisation, the filters were pre-wetted in 6 x standard saline citrate (SSC). They were then prehybridised at 37°C overnight with calf thymus DNA to block non-specific binding of the probe. Hybridisation of filters was done in 6 x SSC at 65°C. The LT DNA probe derived from plasmid pA153.H6, which consisted of an 0.8 kb pair HindIII-generated fragment, was used.19 The STal probe was derived from STal recombinant plasmid which consisted of a 157 bp Hinfl-generated fragment20 while the STa2 probe is a HaeIII-generated fragment derived from STa2 recombinant plasmid pRITIO.250.’ The LT, STal, and STa2 coding fragments were labelled in vitro with the random oligonucleotideprimed DNA radiolabelling method.22 The specific activities of the LT, STal, and STa2 probes were 2-0 x lOS, 18 x 107, and 1-6 x 106 cpm/pg DNA, respectively. After washing the filters in 2 x SSC at 65°C, bound probe was detected by autoradiography (xar 5,

Kodak) at -70°C. Statistical methods The

pH data of fermented and unfermented food samples were analysed with the paired Student’s t-test. The number of contaminated food samples was analysed with McNemar’s chi squared test. In addition, levels of bacteria in unfermented and fermented food samples were compared by means of the unpaired Student’s

t-test.

Results We obtained data from 51 mothers. The remaining 12 were excluded from analysis because they had moved out of the village or all the food had been eaten before the final 12-h collection.

142

TABLE I-GRAM-NEGATIVE BACILLI ISOLATED FROM MAIZE DOUGH

pH of maize dough foods The mean pH (SEM) of unfermented maize dough samples was 5.9 (0-005) and that of fermented maize dough was 3-94 (0-02) (p < 0-001).

Bacteriology All unfermented maize with GNB compared contaminated dough samples with only 16 of 51 of the fermented maize dough samples (p < 0-001). Unfermented maize dough samples yielded more GNB than did fermented samples (5-9 logio colony forming units/g [0’1] vs 4-0 [0-4], p