1043 MEAN URINARY ALBUMIN EXCRETION DURING FOUR
EXPERIMENTAL PERIODS
I
Ualb=
I
I
I
-g albumin/24 h; ACR = mg albumin/mmol creatinine in 24 h sample; OAE
=
)Jg
albumin/min in overnight sample. Analysis of variance used to compare two consecutive periods: *p < 0-05, *p
tp
<
0-02,
Yersinia, Campylobacter, or viral agents, but cultures of stool collected Aug 14 yielded E coli 0157:H7. The second case was a 5-month-old girl who drank raw milk from her grandparents’ dairy farm only once, about ten days before the onset of symptoms. She had bloody diarrhoea on Aug 6, thrombocytopenia, microangiopathic anaemia, and mild renal insufficiency, and was admitted on Aug 11 without requiring dialysis. Stool cultures did not yield Salmonella, Shigella, Yersinia, Campylobacter, or viral agents. A stool specimen collected Aug 16 did not yield E coli 0157:H7. Fecal specimens from animals on both farms were screened for sorbitol-negative colonies which, if present, were screened in 0157 antiserum.7 E coli 0157:H7 was isolated from healthy heifers from both herds. These results support the evidence from epidemiological studies that dairy cattle are a reservoir of E coli 0 157:H7 and that raw milk can be a vehicle of transmission to man. The organism may be more common in diary cattle than in other cattle, since E coli 0157:H7 was not detected in stool samples from 531 healthy beef animals from the Western United States (DJ Fagerberg, Quarles CL, unpublished). In July, 1986, we screened 68 E coli isolates from bovine sources (predominately beef cattle) submitted to veterinary diagnostic laboratories in the Northwestern United States without finding E coli 0157:H7. Much remains to be learned about the relation of this organism to its bovine hosts, the mechanisms for spread among cattle, and its transmission from cattle to man. Centers for Disease
dose used during the first metoprolol period. Although patient E’s albumin excretion rose (non-significantly) during the initial metoprolol period, its failure to fall during drug withdrawal or to worsen significantly during rechallenge suggests that the rise was spontaneous rather than drug induced. Protein intake (as measured by 24 h urinary urea), body weight, and blood glucose control (as measured by serum fructosamine and HbA) remained constant during the trial. This pilot study needs confirmation by larger drug trials of at least one year’s duration to detect any delay in onset of drug response on urinary albumin excretion. The ultimate question remains: will reducing urinary albumin excretion by drugs such as metoprolol prevent overt diabetic nephropathy? Department of Medicine, Diabetes Clinic, and Biochemistry Laboratory,
Waikato Hospital, Hamilton, New Zealand
P. J. FRIEDMAN P. J. DUNN D. R. JURY
1. Mogensen CE, Christensen CK. Predicting diabetic nephropathy in insulindependent patients. N Engl J Med 1984; 311: 89-93. 2. Steno Study Group. Effect of 6 months of strict metabolic control on eye and kidney function in insulin-dependent diabetics with background retinopathy. Lancet 1982; i: 121-23. 3. Christensen CK, Mogensen CE. Effect of antihypertensive treatment on progression of incipient diabetic nephropathy. Hypertension 1985; 7 (suppl II): II109-13.
Control, Atlanta, Georgia 30333, USA
M. LOUISE MARTIN LARRY D. SHIPMAN JOY G. WELLS KATRINA HEDBERG MORRIS E. POTTER I. KAYE WACHSMUTH ROBERT V. TAUXE
Wisconsin State Department of Health and Social Services, Madison
JEFFREY P. DAVIS
Wisconsin State Department of Agriculture, Trade, and Consumer Protection, Madison
JOAN ARNOLDI
Children’s Hospital, St Paul, Minnesota
JOHN TILLELI
1.
Riley LW, Remis RS, Helgerson SD, et al. Hemorrhagic colitis associated with a rare Eschenchia coli serotype. N Engl J Med 1983; 308: 681-85. 2. Remis RS, MacDonald KL, Riley W, et al. Sporadic cases of hemorrhagic colitis associated with Escherichia colt O157:H7. Ann Intern Med 1984; 101: 624-26 3. Karmali MA, Petric M, Lim C, Fleming PC, Arbus GS, Lior H. Association between idiopathic hemolytic uremic syndrome and infection by verotoxin-producing Escherichia coli. J Infect Dis 1985; 151: 775-82. 4. Rodgers MF, Budnick LD, Kirson I, et al. Hemolytic uremic syndrome—an outbreak in Sacramento, California. West J Med 1986; 144: 169-73. 5. Ryan CA, Tauxe RV, Hosek GW, et al. Outbreak of Escherichia coli O157:H7 diarrheal illness in a nursing home: clinical, epidemiologic and pathologic fundings. J Infect Dis (in press) 6. O’Leary MJ, MacDonald KL, Wells JG, Kobayashi JM, Blake PA. Escherichia coli O157:H7 intestinal infections, Seattle. 35th annual EIS conference (Atlanta, Georgia, 1986); abstr 46. 7. Farmer JJ, Davis BR, H7 antiserum-sorbitol fermentation medium: a single tube screening for detecting Escherichia coli O157:H7 associated with hemorrhagic colitis. J Clin Microbiol 1985, 22: 620-25.
CHLOROQUINE-RESISTANT FALCIPARUM ISOLATION OF ESCHERICHIA COLI O157:H7 FROM DAIRY CATTLE ASSOCIATED WITH TWO CASES OF HAEMOLYTIC URAEMIC SYNDROME
S1R, Escherichia coli 0157:H7,
an aetiological agent of has also been associated with haemolytic uraemic syndrome (HUS). 1-4 Although the reservoir of this bacteria is unknown, epidemiological studies have demonstrated an association between infection by E coli 0157:H7 and consumption of undercooked ground beef or raw milk. 1,5,6 Because dairy cattle are a general source of both milk and meat for hamburgers, we suspected that they might be a reservoir for E coli 0157:H7. We have lately investigated two cases of HUS. A 13-month-old boy began drinking raw milk from his family’s dairy farm during the month before the onset of clinical signs. He had bloody diarrhoea on Aug 8, 1986, and was admittted to hospital because of seizures on Aug 12; thrombocytopenia, microangiopathic anaemia, and renal failure requiring haemodialysis developed. Stool cultures during the diarrhoeal prodrome did not yield Salmonella, Shigella,
haemorrhagic colitis,
MALARIA IN BENIN
SIR,-Chloroquine-resistant falciparum malaria has not yet been reliably reported in West Africa except for two therapeutic failures in Nigeria and Ghana.1,2 Since 1984, 59 Plasmodium falciparum strains imported into France from West Africa (west Nigeria/Chad) and studied at the national reference centre in Paris were all sensitive in vitro to chloroquine (semimicrotest, median inhibitory concentrations 8-84 nmol/1). In 1985 chloroquine resistance suddenly emerged in urban areas of Limbe (Cameroon) and Brazzaville and Pointe Noire (Congo) where, respectively, 27/58, 13/22, and 36/44 P falciparum strains from resident children were found to be resistant in vitro to chloroquine a few months after communication of index cases.3-5 This has also been observed in Cotonou, Benin, in the middle of 1986. Over the 2 weeks Sept 2-17 eleven P falciparum malaria cases of prophylactic or therapeutic chloroquine failure were admitted to one of eight hospitals in France. The seven strains studied were resistant in vitro to chloroquine; one showed much decreased sensitivity to quinine
1044 IN-VIVO AND IN-VITRO DRUG SENSITIVITY CHARACTERISTICS OF SEVEN P FALCIPARUM CASES FROM BENIN
CQ = chloroqume; AQ =amodlaqume; Q = quinine; MQ=mefloqume. Cut-off points for resistance in vitro are CI50 *Late recrudescence; t3 weeks’ follow-up; #then MQ 25 mg/kg daily for 1 day; then MQ 17 mg/kg daily for I day.
(table). Cases 1 and 5 were French expatriates and the others were Africans residing in France, all had returned to France after a few months’ stay in Benin. All but case 5 were probably infected in Cotonou or the suburbs. The malaria attacks were typical and were not severe clinically, and occurred despite regular excess chloroquine prophylaxis. The failure of quinine treatment in three cases underlines the need for adequate doses and duration when using this drug alone. As in Gabon in the past year,6 the decrease in sensitivity to quinine suggests that it may be worth combining a cycline with quinine for the treatment of severe malaria from areas of West Africa where chloroquine resistance of P falciparum has been recorded. The need for epidemiological surveillance is obvious in such highly endemic malarial areas which rely mostly on chloroquine for therapeutic control of the disease. National Reference Centre for Chemoresistance of Malaria, Hôpital Claude Bernard; 75944 Paris, France
JACQUES LE BRAS
Parasitic Diseases Service, Centre Hospitalier de Bicètre Kremlin Bicètre
PATRICE BOUREE OMER COCO-CIANCI
Parasitic Diseases Service, Hôpital de la Croix-Rousse,
JEAN-PAUL GARIN
Lyon
MICHEL REY
Infectious Diseases Service, Hôpital Pasteur, Paris
GUY CHARMOT
Infectious Diseases Service, Hôpital des Armées Bégin, Saint-Mande
100, 500, and up to 60 nmo]/l for CQ, Q, and MQ, respectively.
the predictive value of FAB classification and weight loss, but none of the other seventy-one clinical and pathological variables studied were significant in predicting survival. Varela et aF studied data on 56 patients and did a multivariate analysis (Cox regression proportional hazards) of the prognostic importance of several variables, including the scoring system proposed by two working conferences of the FAB Cooperative Group (including quantitative and qualitative variables-neutrophil and platelet counts, megakaryocytes, dysgranulopoiesis, and dysmegakaryopoiesis). The scoring system was the only covariate with predictive value and no other factors, not even the FAB classification, provided more information to predict survival. We have done a multivariate survival analysis of the prognostic significance of 36 clinical, biological, and morphological characteristics at diagnosis, including FAB classification, in a series of 107 cases with MDS. The best combination of characteristics,
ISABELLE HATIN
Infectious Diseases Service, Hôtel Dieu, Clermont-Ferrand
=
SUMMARY OF MULTIVARIATE ANALYSIS ON SURVIVAL IN MDS
_ i !— —
RENÉ ROUE
OJ P falciparum infection not responding to chloroquine in Nigeria. Trans Roy Soc Trop Med Hyg 1985; 79: 141. 2. Neequaye J. In vivo chloroquine-resistant falciparum malaria in western Africa. Lancet 1986; i: 153 3. Sansonetti PJ, Le Bras J, Verdier F, Charmot G, Dupont B, Lapresle C. Chloroquine resistant Plasinodium falciparum in Cameroon. Lancet 1985; i: 1154-55. 4. Le Bras J, Coulaud JP, Bricaire F, Le Bras M, Roue R, Grenier B, Fournier M. Chloroquine resistant falciparum malaria in the Congo. Lancet 1985; ii: 1071. 5 Le Bras J, Simon F, Foulon G, Lambert F Evolution de la chimiorésistance du paludisme en 1985 Bull Epidemiol Hebdom OMS 1986, 26: 1-3. 6. Simon F, Le Bras J, Charmot G, et al. Severe chloroquine resistant falciparum malaria in Gabon with decreased sensitivity to quinine. Trans Roy Soc Trop Med Hyg 1986; 1. Ekanem
80: 5.
selected by the Cox method, is given in the table, variables being listed in the order entered by the forward stepwise modelling procedure. Once again, the FAB classification was the first variable to enter the regression model. MIGUEL A. SANZ Clinical
Haematology Section, Haematology Service, Hospital La Fe, 46009 Valencia, Spain
SONIA GARCIA
J. IGNACIO LORENZO VICTORIA AMIGO
1. Foucar 2.
K, Langdon RM, Armitage JO, Olson DB, Carroll TJ. Myelodysplastic syndromes: a clinical and pathologic analysis of 109 cases. Cancer 1985; 56: 553-61. Varela BL, Chuang C, Woll JE, Bennett JM. Modification in the classification of primary myelodysplastic syndromes: the addition of a scoring system. Hematol Oncol 1985; 3: 55-63
PROGNOSTIC FACTORS IN MYELODYSPLASTIC SYNDROMES
ASPIRIN-LIKE ANALGESICS AND CATARACT
S:R,—Your Aug 23 editorial discusses the prediction of outcome impact of this in planning therapy in myelodysplastic syndromes (MDS). We have observed an uncritical attitude to prognostic factors mentioned in reports cited in your editorial. That the FAB classification is the most useful prognostic marker in MDS seems to have almost universal agreement. However, the prognostic value of peripheral blood findings, multilineage defects, and other disease characteristics, must be questioned because there are controversial results and the statistical approach has usually been inadequate. Most analyses were done by univariate analysis, and no account was taken of correlation between covariates. The only report cited in your editorial that incorporated multivariate analysis1 confirmed
SIR,-Professor Kewitz and colleagues (Sept 20, p 689) were unable to confirm our fmding1 that aspirin-like analgesics are associated with protection against cataract. They suggest that this is because our controls were from a different hospital or general practice. Although taking controls from the same hospital confers advantages, it may lead to problems in surveys of major eye diseases in an eye hospital. Kewitz et al do not state if their controls had the same age-sex distribution as the cases; presumably they did not. If we had adopted their approach but retained the age-sex distribution match we would have found that most of the controls had glaucoma or diabetic retinopathy. Both glaucoma and diabetes are powerful risk factors for cataract.’ One of us saw the results of a case-control study of glaucoma presented where diabetes was identified as a
and the