734
toothpastes. More difficult from scientific perspectives are NIDR claims about reduced and more preventive services.
and health sugar
policy consumption
The press release from NIDR attributed the savings from improved oral health in large part to a 33% reduction in refined sugar, and a 200% increase in no-calorie sweeteners over the decade. Although sucrose consumption declined by a third, the use of high fructose syrups and other monosaccharides rose by 50 %; the latter are also refined sugars. The average intake of sugars (as assessed from disapperance rates provided by the US Department of Agriculture, 1990) has not changed over the past 20 years in the USA. The ability of monosaccharides to produce caries might be slightly lower than that of sucrose, but it has not previously been inferred that this equivocal difference is responsible for the substantial overall decline in caries, and savings in dental bills. So the press release statement "much of the savings-$60 billion (of $99 billion) is related to a decline in the per capita consumption of refined sugar" seems incorrect-sucrose consumption has declined, but refined sugar consumption has remained the same. Did the statistical model used take account of total sugar consumption, or only the change in sucrose? If the model was complete, then its interpretation does not seem to accord with knowledge of the ability of sugars to cause caries. If incomplete, it ignored about half the total sugar intake. The other$39 billion in dental savings was attributed to "wide spread use of fluorides, better oral hygiene practices, and a significant increase in preventive services by dental practitioners". Fluoride toothpaste use remained high and there was only slow progress implementing further water fluoridation over this period. Over the decade, effective and efficient school programmes accomplishing these objectives were lost across the USA. In important respects they were even more effective than those services delivered in private offices because they provided access to all children in those programmes and they included preventive sealants, which have been very slowly adopted by the private sector. These programmes especially targetted children from low income families, because dental disease is more prevalent in such children. These are precisely the group of children who lack access to dental services. It is a sad legacy that as effective prevention was being defined through the former National Caries Program of NIDRNIH, fiscal decisions were being taken which denied these benefits to the neediest children. I hope that legislators and health planners will be given clear messages that despite gains in cost containment and caries prevention, dental and oral disease remains a major problem for certain groups in society. The NIDR press release and Greenberg’s article fail to convey this important message, and could lead to a belief that dental health is no longer a public health priority. Dental care must be part of the national health agenda, and public dental programmes are as essential a component as is dental insurance. Furthermore, primary care for all must include preventive and routine dental services. Recent statements by the American Dental Association Task Force on Access to Dental Care, denying that access is a difficulty, are quite unrealistic and complacent. Access, as well as cost, remain drawbacks for many. University of Texas Health Science Center at San Antonio, San Antonio, Texas 78284, USA
JOHN P. BROWN
Igbinedion Hospital SIR,-In your Aug 8 issue (London edition) you published an advertisement for medical posts at the Igbinedion Hospital, Okada Town, Nigeria. Extensive modern facilities were described. This was more an expression of intention than of reality. When I left on July 11, the hospital was still being built and there were no services in any of the departments. These matters apart, any potential recruit is strongly advised to obtain a signed contract, along with any necessary visas, before relinquishing a currently held position. The Cottage, Venns Gate, Cheddar BS27 3LW, UK
ERLE K. HALL
Influence of cross-sex transmission on measles mortality in rural Senegal SIR,-Between September, 1987, and January, 1989,have been involved in paediatric consultations on the primary health care level in Ziguinchor, Senegal. In my experience, there is a tendency to neglect female children in Senegal, by contrast with what I have seen in Zaire. The chance that one of two successive children will be neglected, by preference for the other sex, is obviously greater if the two children are of opposite sex than if they are of the same sex. Consequently, child neglect is more frequent in families with siblings of opposite sex. Since child neglect can increase risks and severity of infectious disease in families with siblings of opposite sex, infection spread will probably be greater than in other families, leading to an increased exposure in both siblings and a more severe outcome for the neglected child, mostly girls. Mr Aaby’s (Aug 15, p 388) finding that measles infection contracted from a person of the opposite sex is more severe can thus be explained by child neglect. The fact that in Niakhar, Senegal, the same proportions of boys and girls are taken to health centres does in no way discount child neglect, which can also occur through other behaviour such as feeding practices, time interval before consultation, or compliance with treatments. Centre for Human Genetics, Herestraat 49,
J. VAN DEN BROECK
Leuven, Belgium
&bgr;-glucuronidase antibodies
in ulcerative
colitis SiR,—Autoantibodies against neutrophil granulocytes
are
valuable diagnostic tools in systemic vasculitis. On ethanol-fixed cells they appear with a cytoplasmic antineutrophil cytoplasmic antibody (c-ANCA) or a perinuclear (p-ANCA) fluorescence pattern. The p-ANCA pattern has revealed antibodies against myeloperoxidase (MPO), elastase, lactoferrin because of a redistribution of these granular proteins during fixation. In general, it has been assumed that p-ANCA usually corresponds to myeloperoxidase antibodies. A high frequency of the p-ANCA pattern has been reported in patients with ulcerative colitis.l-4 Antibodies to MPO, elastase 2,3cathepsin G,3 and lactoferrin constituted 10-20% of p-ANCA, which leaves 80-90% of antibodies unidentified. Occurrence of p-ANCA has also been reported to correlate with clinical activity of ulcerative colitis but not with that of Crohn’s disease.’ We have studied 54 patients during their first attack of ulcerative colitis. Serum from 77 patients showed neither cytoplasmic nor nuclear reactivity. Of the 43 positive serum samples, 42 had a p-ANCA and 1 a c-ANCA pattern. 24 patients with infectious colitis (or other non-relapsing colitis) were controls. None of these had p-ANCA, but 1 had weak c-ANCA reactivity. Serum was tested on human lymphocytes, and endothelial cells and on rat liver slices, and was negative. The serum samples were tested for autoantibodies against myeloperoxidase and elastase (purified according to Olsson et aI,s and Baugh and Travisrespectively), cathepsin G (Calbiochem), lactoferrin, and P-glucuronidase (Sigma) by ELISA/ and to lysozyme (Calbiochem). The figure shows the distribution of different autoantibodies with p-ANCA fluorescence. In about 10% of these, anti-MPO was the only antibody detected. Antibodies to elastase, cathepsin G, and lactoferrin occurred at low frequency (3%). 57% of the samples captured circulating antibodies to p-glucuronidase. No antibodies against lysozyme were detected. In 20% of patients the autoantigen remains unknown. The 2 sera with c-ANCA had no reactivity. defend Neutrophil granulocytes against invading microorganisms and release lysomal enzymes. P-glucuronidase is localised in azurophil granules and is released in parallel with MPO. Released neutrophil components may undergo antigen processing in antigen-presenting cells to induce antibody production. There might be a subgroup of ulcerative colitis patients characterised by a specific antigen-presentation pathway for 0-glucuronidase. At first attack of ulcerative colitis it is important to
735
Thus, a strain of C difficile that did not produce toxin B was associated with severe PMC. Possibly the toxin is not associated with PMC at all. The enzyme immunoassay for toxin A has been said to be less sensitive than the cytotoxic tests for toxin B in the diagnosis of PMC.3° We feel that the relative insensitivity of the current enzyme immunoassay for PMC diagnosis should not prevent the development of more sensitive toxin A tests, especially because toxin A is more closely related to the disease. When severe antibiotic-associated diarrhoea is clinically diagnosed, it is unwise to rely on the toxin B test in stool as an indicator of PMC, and additional assay of toxin A may be more valid. GRAHAM SMITH STEPHEN DEALLER JAMES ANSON DAVID TOMPKINS
Microbiology Department, Bradford Royal Infirmary, Bradford BD9 6RJ, UK
Distribution
of
autoantibodies against neutrophil in active ulcerative colitis.
granulocyte components
The ordinate represents percentages of
patient serum samples with antibodies against myeloperoxidase (mpo), elastase, cathepsin G, and
P-glucuronidase exclude infectious colitis. Serological markers would be valuable in this diagnosis. Our data indicate that p-ANCA is an early marker for inflammatory bowel disease. Whether patients with p-ANCA reactivity and the distinct subset of P-glucuronidase antibodies differ from patients without p-ANCA reactivity from a prognostic or pathogenic point of view is unknown. Department of Medical Microbiology, University of Lund, S-223 62 Lund, Sweden
Department of Gastroenterology. Danderyd University Hospital, Danderyd, Sweden
SP, Honour P Detection, isolation and identification of Clostridium difficile. In. Borriello SP, ed. Antibiotic associated diarrhoea and colitis. Boston: Martinus Nijhoff, 1984: 37-47. 2. Boriello SP, Honour P. Detection of Clostridium difficile toxins. In: Boriello SP, ed. Antibiotic associated diarrhoea and colitis. Boston: Martinus Nijhoff, 1984: 49-56. 3. Borriello SP, Vale T, Brazier JS, et al. Evaluation of a commercial enzyme immunoassay kit for the detection of Clostridium difficile toxin A. Eur J Clin Microbiol Infect Dhs 1992; 11: 360-63. 4. De Girolami PC, Hanff PA, Eichelberger K, et al. Multicenter evaluation of a new enzyme immunoassay for detection of Clostridium difficile enterotoxin A. J Clin Microbiol 1992; 30: 1085-88. 1. Boriello
HLA typing by amplification-created restriction site
LENNART NÄSSBERGER ASA LJUNGH
GUDRUN SCHUMACHER BO KOLLBERG
1. Snook
JA, Chapman RW, Fleming K, Jewell DP. Antineutrophil clear antibody in ulcerative colitis, Crohn’s disease and primary sclerosing cholangitis. Clin Exp Immunol 1989; 36: 30-33. 2. Falk RJ, Sartor RB, Jones DA, Jeffries BD, Jennette JC. Antineutrophil cytoplasmic antibodies (ANCA) in ulcerative colitis (UC). Clin Res 1990; 38: 387A. 3. Duerr H, Targan SR, Landers CJ, et al. Antineutrophil cytoplasmic antibodies: a link between primary sclerosing cholangitis and ulcerative colitis. Gastroenterology 1991; 100: 1385-91. 4 Scibold F, Weber P, Klein R, Berg PA, Wiedmann KH. Clinical significance of antibodies against neutrophils in patients with inflammatory bowel disease and primary sclerosing cholangitis Gut 1992, 33: 657-62 5 Olsson I, Olofsson T, Odeberg H. Myeloperoxidase-mediated iodination of granulocytes Scand J Haematol 1972; 9: 483-89. 6. Baugh RJ, Travis J. Human leucocyte granule elastase rapid isolation and characterization. Biochemistry 1976; 15: 836-41 7 Nassberger L, Sjoholm AG, Jonsson H, Sturfelt G, Akesson A Autoantibodies against neutrophil cytoplasma components in systemic lupus erythematosus and in hydralazine-induced lupus. Clin Exp Immunol 1990; 81: 380-83. 8. Edwards SW, Nurcombe HL, Hart CA Oxidative inactivation of myeloperoxidase released from human neutrophils. Biochem J 1987; 245: 925-28.
Diagnosis of pseudomembranous colitis Sin,—The diagnosis of pseudomembranous colitis (PMC) depends on rectal biopsy but tests in the stool for cytotoxin (toxin B) produced by Clostridium difficile are often used as an indicator of the condition. We report PMC that was not associated with production of toxin B by the bacterium. A 56-year-old woman was admitted with a diagnosis of gastroenteritis. She had received ampicillin and had profuse diarrhoea. Her condition worsened and a perforated caecum was discovered at laparotomy; a subtotal colectomy was done with formation of an ileostomy. Despite intensive care, including metronidazole, she died 12 days later. The bowel condition was histologically diagnosed as PMC. Stool was cultured for C difficile with cycloserine cefoxitin agar, and a pure, profuse growth of the bacterium was isolated. The isolate and a prelaparotomy stool sample were tested for toxin B.12 These tests were negative and the Anaerobic Reference Laboratory confirmed that the isolate did not produce toxin B. However, the C dlfficile isolate was subsequently found to be positive for enterotoxin (toxin A) by an enzyme immunoassay (Meridian).
SIR,-HLA-DR DNA typing is usually done by hybridising sequence-specific oligonucleotides (SSO) to dot blots of PCR
products. But PCR-SSO has technical difficulties. Other approaches include testing polymorphism by restriction endonuc1eases.2-4 The advantages are a simple DNA digestion plus rapid and easy agarose gel electrophoresis and the absence of false positives. But a restriction enzyme for a particular polymorphism is not
always available. We propose
a new
strategy for HLA class II
genotyping: restriction sites can be artificially created within the HLA DNA sequence by site-directed mutagenesis in the amplification step, so-called amplification-created restriction site (ACRS).’ We have tested this idea designing an amplification primer that, introducing mutations in the amplified products, generates a restriction site specific for the desired HLA-DR sequence or set of sequences. The electrophoretic analysis detects the digested DR sequences and permits the isolation of these sequences in the same step: grouping HLA-DR sequences by means of restriction endonuclease digestion is a much more secure method than specific amplification, and this is the other novelty of our method. After a new PCR amplification, this separated set of sequences can be analysed independently by means of, again, restriction endonucleases. We have designed a system to type unequivocally the four variants of the HLA-DRB3 locus. These variants are not detected serologically but, since most are recognised by T cells (Dw24,
Creation of restriction site
specific for DRB3.
(a) DRB3 sequences have T (unique to them) needed in the BglIl recognition sequence (b) Mutated pnmer with two mismatches (arrows) creates sequence that can be cut by Bglll. (c) Same pnmer when used to amplify sequences other than DRB3 does not create a Bg/ll recognition site
toothpastes. More difficult from scientific perspectives are NIDR claims about reduced and more preventive services.
and health sugar
policy consumption
The press release from NIDR attributed the savings from improved oral health in large part to a 33% reduction in refined sugar, and a 200% increase in no-calorie sweeteners over the decade. Although sucrose consumption declined by a third, the use of high fructose syrups and other monosaccharides rose by 50 %; the latter are also refined sugars. The average intake of sugars (as assessed from disapperance rates provided by the US Department of Agriculture, 1990) has not changed over the past 20 years in the USA. The ability of monosaccharides to produce caries might be slightly lower than that of sucrose, but it has not previously been inferred that this equivocal difference is responsible for the substantial overall decline in caries, and savings in dental bills. So the press release statement "much of the savings-$60 billion (of $99 billion) is related to a decline in the per capita consumption of refined sugar" seems incorrect-sucrose consumption has declined, but refined sugar consumption has remained the same. Did the statistical model used take account of total sugar consumption, or only the change in sucrose? If the model was complete, then its interpretation does not seem to accord with knowledge of the ability of sugars to cause caries. If incomplete, it ignored about half the total sugar intake. The other$39 billion in dental savings was attributed to "wide spread use of fluorides, better oral hygiene practices, and a significant increase in preventive services by dental practitioners". Fluoride toothpaste use remained high and there was only slow progress implementing further water fluoridation over this period. Over the decade, effective and efficient school programmes accomplishing these objectives were lost across the USA. In important respects they were even more effective than those services delivered in private offices because they provided access to all children in those programmes and they included preventive sealants, which have been very slowly adopted by the private sector. These programmes especially targetted children from low income families, because dental disease is more prevalent in such children. These are precisely the group of children who lack access to dental services. It is a sad legacy that as effective prevention was being defined through the former National Caries Program of NIDRNIH, fiscal decisions were being taken which denied these benefits to the neediest children. I hope that legislators and health planners will be given clear messages that despite gains in cost containment and caries prevention, dental and oral disease remains a major problem for certain groups in society. The NIDR press release and Greenberg’s article fail to convey this important message, and could lead to a belief that dental health is no longer a public health priority. Dental care must be part of the national health agenda, and public dental programmes are as essential a component as is dental insurance. Furthermore, primary care for all must include preventive and routine dental services. Recent statements by the American Dental Association Task Force on Access to Dental Care, denying that access is a difficulty, are quite unrealistic and complacent. Access, as well as cost, remain drawbacks for many. University of Texas Health Science Center at San Antonio, San Antonio, Texas 78284, USA
JOHN P. BROWN
Igbinedion Hospital SIR,-In your Aug 8 issue (London edition) you published an advertisement for medical posts at the Igbinedion Hospital, Okada Town, Nigeria. Extensive modern facilities were described. This was more an expression of intention than of reality. When I left on July 11, the hospital was still being built and there were no services in any of the departments. These matters apart, any potential recruit is strongly advised to obtain a signed contract, along with any necessary visas, before relinquishing a currently held position. The Cottage, Venns Gate, Cheddar BS27 3LW, UK
ERLE K. HALL
Influence of cross-sex transmission on measles mortality in rural Senegal SIR,-Between September, 1987, and January, 1989,have been involved in paediatric consultations on the primary health care level in Ziguinchor, Senegal. In my experience, there is a tendency to neglect female children in Senegal, by contrast with what I have seen in Zaire. The chance that one of two successive children will be neglected, by preference for the other sex, is obviously greater if the two children are of opposite sex than if they are of the same sex. Consequently, child neglect is more frequent in families with siblings of opposite sex. Since child neglect can increase risks and severity of infectious disease in families with siblings of opposite sex, infection spread will probably be greater than in other families, leading to an increased exposure in both siblings and a more severe outcome for the neglected child, mostly girls. Mr Aaby’s (Aug 15, p 388) finding that measles infection contracted from a person of the opposite sex is more severe can thus be explained by child neglect. The fact that in Niakhar, Senegal, the same proportions of boys and girls are taken to health centres does in no way discount child neglect, which can also occur through other behaviour such as feeding practices, time interval before consultation, or compliance with treatments. Centre for Human Genetics, Herestraat 49,
J. VAN DEN BROECK
Leuven, Belgium
&bgr;-glucuronidase antibodies
in ulcerative
colitis SiR,—Autoantibodies against neutrophil granulocytes
are
valuable diagnostic tools in systemic vasculitis. On ethanol-fixed cells they appear with a cytoplasmic antineutrophil cytoplasmic antibody (c-ANCA) or a perinuclear (p-ANCA) fluorescence pattern. The p-ANCA pattern has revealed antibodies against myeloperoxidase (MPO), elastase, lactoferrin because of a redistribution of these granular proteins during fixation. In general, it has been assumed that p-ANCA usually corresponds to myeloperoxidase antibodies. A high frequency of the p-ANCA pattern has been reported in patients with ulcerative colitis.l-4 Antibodies to MPO, elastase 2,3cathepsin G,3 and lactoferrin constituted 10-20% of p-ANCA, which leaves 80-90% of antibodies unidentified. Occurrence of p-ANCA has also been reported to correlate with clinical activity of ulcerative colitis but not with that of Crohn’s disease.’ We have studied 54 patients during their first attack of ulcerative colitis. Serum from 77 patients showed neither cytoplasmic nor nuclear reactivity. Of the 43 positive serum samples, 42 had a p-ANCA and 1 a c-ANCA pattern. 24 patients with infectious colitis (or other non-relapsing colitis) were controls. None of these had p-ANCA, but 1 had weak c-ANCA reactivity. Serum was tested on human lymphocytes, and endothelial cells and on rat liver slices, and was negative. The serum samples were tested for autoantibodies against myeloperoxidase and elastase (purified according to Olsson et aI,s and Baugh and Travisrespectively), cathepsin G (Calbiochem), lactoferrin, and P-glucuronidase (Sigma) by ELISA/ and to lysozyme (Calbiochem). The figure shows the distribution of different autoantibodies with p-ANCA fluorescence. In about 10% of these, anti-MPO was the only antibody detected. Antibodies to elastase, cathepsin G, and lactoferrin occurred at low frequency (3%). 57% of the samples captured circulating antibodies to p-glucuronidase. No antibodies against lysozyme were detected. In 20% of patients the autoantigen remains unknown. The 2 sera with c-ANCA had no reactivity. defend Neutrophil granulocytes against invading microorganisms and release lysomal enzymes. P-glucuronidase is localised in azurophil granules and is released in parallel with MPO. Released neutrophil components may undergo antigen processing in antigen-presenting cells to induce antibody production. There might be a subgroup of ulcerative colitis patients characterised by a specific antigen-presentation pathway for 0-glucuronidase. At first attack of ulcerative colitis it is important to
735
Thus, a strain of C difficile that did not produce toxin B was associated with severe PMC. Possibly the toxin is not associated with PMC at all. The enzyme immunoassay for toxin A has been said to be less sensitive than the cytotoxic tests for toxin B in the diagnosis of PMC.3° We feel that the relative insensitivity of the current enzyme immunoassay for PMC diagnosis should not prevent the development of more sensitive toxin A tests, especially because toxin A is more closely related to the disease. When severe antibiotic-associated diarrhoea is clinically diagnosed, it is unwise to rely on the toxin B test in stool as an indicator of PMC, and additional assay of toxin A may be more valid. GRAHAM SMITH STEPHEN DEALLER JAMES ANSON DAVID TOMPKINS
Microbiology Department, Bradford Royal Infirmary, Bradford BD9 6RJ, UK
Distribution
of
autoantibodies against neutrophil in active ulcerative colitis.
granulocyte components
The ordinate represents percentages of
patient serum samples with antibodies against myeloperoxidase (mpo), elastase, cathepsin G, and
P-glucuronidase exclude infectious colitis. Serological markers would be valuable in this diagnosis. Our data indicate that p-ANCA is an early marker for inflammatory bowel disease. Whether patients with p-ANCA reactivity and the distinct subset of P-glucuronidase antibodies differ from patients without p-ANCA reactivity from a prognostic or pathogenic point of view is unknown. Department of Medical Microbiology, University of Lund, S-223 62 Lund, Sweden
Department of Gastroenterology. Danderyd University Hospital, Danderyd, Sweden
SP, Honour P Detection, isolation and identification of Clostridium difficile. In. Borriello SP, ed. Antibiotic associated diarrhoea and colitis. Boston: Martinus Nijhoff, 1984: 37-47. 2. Boriello SP, Honour P. Detection of Clostridium difficile toxins. In: Boriello SP, ed. Antibiotic associated diarrhoea and colitis. Boston: Martinus Nijhoff, 1984: 49-56. 3. Borriello SP, Vale T, Brazier JS, et al. Evaluation of a commercial enzyme immunoassay kit for the detection of Clostridium difficile toxin A. Eur J Clin Microbiol Infect Dhs 1992; 11: 360-63. 4. De Girolami PC, Hanff PA, Eichelberger K, et al. Multicenter evaluation of a new enzyme immunoassay for detection of Clostridium difficile enterotoxin A. J Clin Microbiol 1992; 30: 1085-88. 1. Boriello
HLA typing by amplification-created restriction site
LENNART NÄSSBERGER ASA LJUNGH
GUDRUN SCHUMACHER BO KOLLBERG
1. Snook
JA, Chapman RW, Fleming K, Jewell DP. Antineutrophil clear antibody in ulcerative colitis, Crohn’s disease and primary sclerosing cholangitis. Clin Exp Immunol 1989; 36: 30-33. 2. Falk RJ, Sartor RB, Jones DA, Jeffries BD, Jennette JC. Antineutrophil cytoplasmic antibodies (ANCA) in ulcerative colitis (UC). Clin Res 1990; 38: 387A. 3. Duerr H, Targan SR, Landers CJ, et al. Antineutrophil cytoplasmic antibodies: a link between primary sclerosing cholangitis and ulcerative colitis. Gastroenterology 1991; 100: 1385-91. 4 Scibold F, Weber P, Klein R, Berg PA, Wiedmann KH. Clinical significance of antibodies against neutrophils in patients with inflammatory bowel disease and primary sclerosing cholangitis Gut 1992, 33: 657-62 5 Olsson I, Olofsson T, Odeberg H. Myeloperoxidase-mediated iodination of granulocytes Scand J Haematol 1972; 9: 483-89. 6. Baugh RJ, Travis J. Human leucocyte granule elastase rapid isolation and characterization. Biochemistry 1976; 15: 836-41 7 Nassberger L, Sjoholm AG, Jonsson H, Sturfelt G, Akesson A Autoantibodies against neutrophil cytoplasma components in systemic lupus erythematosus and in hydralazine-induced lupus. Clin Exp Immunol 1990; 81: 380-83. 8. Edwards SW, Nurcombe HL, Hart CA Oxidative inactivation of myeloperoxidase released from human neutrophils. Biochem J 1987; 245: 925-28.
Diagnosis of pseudomembranous colitis Sin,—The diagnosis of pseudomembranous colitis (PMC) depends on rectal biopsy but tests in the stool for cytotoxin (toxin B) produced by Clostridium difficile are often used as an indicator of the condition. We report PMC that was not associated with production of toxin B by the bacterium. A 56-year-old woman was admitted with a diagnosis of gastroenteritis. She had received ampicillin and had profuse diarrhoea. Her condition worsened and a perforated caecum was discovered at laparotomy; a subtotal colectomy was done with formation of an ileostomy. Despite intensive care, including metronidazole, she died 12 days later. The bowel condition was histologically diagnosed as PMC. Stool was cultured for C difficile with cycloserine cefoxitin agar, and a pure, profuse growth of the bacterium was isolated. The isolate and a prelaparotomy stool sample were tested for toxin B.12 These tests were negative and the Anaerobic Reference Laboratory confirmed that the isolate did not produce toxin B. However, the C dlfficile isolate was subsequently found to be positive for enterotoxin (toxin A) by an enzyme immunoassay (Meridian).
SIR,-HLA-DR DNA typing is usually done by hybridising sequence-specific oligonucleotides (SSO) to dot blots of PCR
products. But PCR-SSO has technical difficulties. Other approaches include testing polymorphism by restriction endonuc1eases.2-4 The advantages are a simple DNA digestion plus rapid and easy agarose gel electrophoresis and the absence of false positives. But a restriction enzyme for a particular polymorphism is not
always available. We propose
a new
strategy for HLA class II
genotyping: restriction sites can be artificially created within the HLA DNA sequence by site-directed mutagenesis in the amplification step, so-called amplification-created restriction site (ACRS).’ We have tested this idea designing an amplification primer that, introducing mutations in the amplified products, generates a restriction site specific for the desired HLA-DR sequence or set of sequences. The electrophoretic analysis detects the digested DR sequences and permits the isolation of these sequences in the same step: grouping HLA-DR sequences by means of restriction endonuclease digestion is a much more secure method than specific amplification, and this is the other novelty of our method. After a new PCR amplification, this separated set of sequences can be analysed independently by means of, again, restriction endonucleases. We have designed a system to type unequivocally the four variants of the HLA-DRB3 locus. These variants are not detected serologically but, since most are recognised by T cells (Dw24,
Creation of restriction site
specific for DRB3.
(a) DRB3 sequences have T (unique to them) needed in the BglIl recognition sequence (b) Mutated pnmer with two mismatches (arrows) creates sequence that can be cut by Bglll. (c) Same pnmer when used to amplify sequences other than DRB3 does not create a Bg/ll recognition site
