848
AT-111-deficient kindreds. In Italy in 1981 only 0-3% of all deaths were directly attributed to venous thromboembolism. National Bureau of Statistics data, based on death certificates, are probably underestimates: in necropsy studies pulmonary embolism regarded as fatal was found in 9% of cases, but in only 0-8-2-4% of patients was this not associated with some other terminal illness.s,6 Information obtained from published reports is subject to bias because in recent years familial AT III quantitative deficiency will not have had the same attention as formerly and kindreds with more unfavourble clinical courses will have been diagnosed and published more frequently. However, kindreds with severe clinical expression of AT III deficiency do exist and long-term prophylaxis with oral anticoagulants needs to be carefully evaluated in all members of these kindreds, whether they have symptoms or not. The necessity of life-long prophylaxis in those who have had a thrombosis is supported by our observation that a recurrent episode probably accounted for half the deaths from venous thromboembolism. A main target of long-term prophylaxis in AT III deficiency should be the prevention of morbidity from venous thromboembolism and of its sequelae, besides the prevention of mortality. AT-III-deficient individuals have the highest risk of first thrombosis between 15 and 50 years,4 and in our opinion the use of anticoagulants in such symptomless individuals, who as a rule have no indication for prophylaxis, should be reconsidered. A complete list of references available from the authors. Istituto di Semeiotica Medica, Catholic University, 00168 Rome, Italy
on
the AT-III-deficient families
analysed is
VALERIO DE STEFANO GIUSEPPE LEONE
1. Leone G, De Stefano V, Di Donfrancesco A, Ferrelli R, Traisci G, Bizzi B. Antithrombin III Pescara: a defective AT III variant with no alterations of plasma crossed immunoelectrophoresis; but with an abnormal crossed immunoelectrofocusing pattern. Br J Haematol 1987; 65: 187-91. 2. Finazzi G, Caccia R, Barbui T. Different prevalence of thromboembolism in the subtypes of congenital antithrombin III deficiency: review of 404 cases. Thromb Haemostas 1987; 58: 1094. 3. De Stefano V, Leone G. Antithrombin III congenital defects: revising classificaton system. Thromb Haemostas 1989; 62: 820-21. 4. Hirsh J, Piovella F, Pini M. Congenital anrithrombin III deficiency: incidence and clinical features. Am J Med 1989; 87 (suppl 3B): 34S-38S. 5. Macintyre IMC, Ruckley CV. Pulmonary embolism: a clinical and autopsy study. Scot Med J 1974; 19: 20-24. 6. Nielsen HK, Bechgaard P, Nielsen PF, Husted SE, Geday E. 178 fatal cases of pulmonary embolism in a medical department. Acta Med Scand 1981; 209: 351-55.
Helicobacter pylori infection in children
cohabiting
SiR,—The association between Helicobacterpylori in the stomach and histologically proven gastritis is well recognised. However, little is known about the sources, potential reservoirs, and routes of transmission of this bacterium. Since Hpylori has been isolated only from human gastric mucosa and is endemic throughout the world human beings are the most likely natural reservoir.’ Person-toperson transmission would be predicted and H pylori antibodies are found in household contacts of infected patients;2 saliva may be the route of transmission. Identification of the same strain of H pylori in cohabiting subjects would provide further confirmatory evidence of this view. We examined Hpylori isolates from areas of gastritis in 5 mentally handicapped children, who lived in close contact, by restriction endonuclease DNA analysis. All children (3 boys and 2 girls) aged 4-12 years were living in the same long-term care facility for several years. Samples were cultured at 37°C for up to five days under microaerophilic conditions on both selective and unselective media. H pylori colonies were subcultured from unselective medium and DNA was extracted by sodium dodecylsulphate alkaline lysis, as previously described.4 Dual restriction endonuclease analysis was completed with 10 units of Hind III for 1 h and 10 units of Bst EII for 1 h at 37°C and 60°C, respectively, in appropriate buffer as described by the manufacturer (Gibco BRL, Cergy-Pontoise, France). DNA fragments were separated by horizontal 7% agarose gel electrophoresis and photographed with an ’MP-4 Land’ camera
(Polaroid).
isolates, digested bacterial DNA samples were in electrophoresed the same gel with other H pylori strains from unrelated patients. Some had indistinguishable restriction patterns. Two children had identical patterns, as did two others. No overlap with other strains was observed. Our study shows that two individuals with infectious gastritis can be infected by the same strain. Contamination is unlikely because the endoscope was disinfected and the grip changed before each procedure, the linked children were not endoscoped on the same day, and bacterial counts were high. Many epidemiological studies have shown that living conditions may influence the communicability of H pylori. In our study, the living conditions suggest that person-to-person transmission was likely and that there was a common source of infection. The importance of saliva remains uncertain but some studies suggest that the mouth could be a reservoir’ and that oro-oral transmission could take place. To compare
Bacteriology Laboratory A, Faculty of Medicine, 59 045 Lille, France; and Paediatric Service, CHR de Lille, and CPIJ, Vendin-le-Viel
P. VINCENT F. GOTTRAND P. PERNES M. A. HUSSON A. BEJU H. LECLERC J. P. FARRIAUX
1. Graham DY, Adam E, Klein PD, et al. Epidemiology of Campylobacter pylon infection. Gastroenterol Clin Biol 1989; 13: 84B-89B 2. Graham DY, Malatty HM, Klein PD, Evans DG, Evans DJ, Adam E. Helicobacter pylon infection clusters in family. Rev Esp Enferm Apar Dig 1990; 78: 26. 3. Albenque M, Tall F, Dabis F, Megraud F. Epidemiological study of Helicobacter pylori transmission from mother to child in Africa Rev Esp Enferm Apar Dig 1990, 78: 48 4. Beji A, Izard D, Gaviny F, Leclerc H, Leseine-Delstanche M, Krembel J. A rapid chemical procedure for isolation and purification of chromosomal DNA from Gram-negative bacilli. Ann Biochem 1987; 162: 18-23 5. Krajden S, Fuksa M, Anderson J, et al. Examination of human stomach biopsies, saliva and dental plaque for Campylobacter pylori.J Clin Microbiol 1989; 27: 1397-98
Transmission of Brucella melitensis SIR,-Dr Ruben and colleagues (Jan 5, p 14) report person-toperson transmission of Brucella melitensis. The following case provides further evidence of between-person transmission of brucellosis. A 35-year-old Swedish man and his 30-year-old girlfriend spent two weeks in the south of Spain in June, 1989. They shared accommodation and meals with three other people. No-one had symptoms during the stay. The man had respiratory tract symptoms and remittent fever 14 weeks after leaving Spain. 4 weeks later swelling of the right knee developed and 8 weeks after this (December, 1989) he had a right-sided epididymitis. He was given amoxycillin and cefadroxil for airway symptoms, diclofenac for arthritis, and ciprofloxacin for epididymitis. Faecal and blood cultures were negative, but the latter were not taken until after antibiotics were begun. In December he had titres of 640 against Yersinia enterocolitica type 9 and 320 against Brucella abortus. He eventually recovered uneventfully and was believed to have had
yersinia infection. Stored serum samples were sent to the state bacteriological laboratory, Solna, Sweden, for analysis by ELISA. A rising IgG titre (peak 850) and decreasing IgM titre (peak > 1500) compatible with brucella infection were recorded. This patient’s girlfriend became ill in April, 1990-9 months after the stay in Spain-with sudden fever, sweating, headache, and leucopenia. Repeated blood cultures grew B melitensis. Serum agglutinin titre to B abortus was 320 and to Y enterocolitica type 9 160. Peak antibody titres to brucella measured with ELISA were IgM over 1500 and IgG 850. She was given doxycycline and rifampicin and recovered without complications. Although the diagnosis of the man was not verified by culture, the clinical picture as well as the serological findings are, in retrospect, compatible with brucellosis. There is serological crossreactivity between brucella and yersinia antigens, which may account for the serological reactivity that confused the diagnosis. The disease started in the
woman
9 months after
possible
exposure. This exceeds the maximum recorded incubation time of 7 months.1,2 It therefore seems unlikely that both patients were
infected in Spain. The fact that the three individuals with whom they shared accommodation at that time were not infected and had
849
brucella when tested 9 months later is evidence of infection. Brucellosis in domestic animals has not been recorded in Sweden since the 1950s, and human brucellosis acquired in this country has not been reported since then. The possibility of coital transmission of brucella has long been suggested, but only recently has brucellosis been isolated from human spennatozoa.3 We thus conclude that the woman was probably infected by her boyfriend, possibly by sexual intercourse during genital manifestation of the disease. no
antibodies
against
to
a common source
Departments of Infectious Diseases and Clinical Bacteriology, University of Goteborg,
S41685-Goteborg, 1.
JOHAN LINDBERG PETER LARSSON
Sweden
To assess the minimum number of inclusions detectable by these techniques, we inoculated serial dilutions of C pneumoniae elementary bodies onto McCoy cell monolayers (in triplicate) and incubated them for three days. One monolayer at each dilution was then stained with chlamydial fluorescein-conjugated lipopolysaccharide antibody and inclusions were counted. A second monolayer was swabbed and tested with a commercially available EIA (’STD-EZE’, Abbott). The third monolayer was swabbed, rolled onto a glass slide, and tested by a commercially available DFA test (’Microtrak’, Syva, USA) with lipopolysaccharide rather than major outer-membrane protein fluorescein-conjugated antibody. C pneumoniae inclusion counts with EIA and DFA for serial dilutions are given below:
Spink WW. The nature of brucellosis. Minneapolis: University of Minnesota Press,
SIR,-Dr Gallen and Dr Ispahani (Feb 2, p 308) report a case of fulminant Capnocytophaga canimorsus septicaemia. We have seen another case of severe C canimorsus disease, in which immediate suspicion of the infection made prompt and adequate treatment
possible. A 74-year-old
with cardiac failure (New York Heart was bitten by a dog. Signs of local infection then developed, and he was started on roxithromycin 150 mg twice daily (allergy to penicillins was suspected). The local reaction subsided, but treatment was discontinued after three doses because of diarrhoea. Fever, lethargy, and increasing confusion developed over the next two days. On admission several days after the bite the patient was unresponsive to verbal stimuli, but reacted with agitation to painful stimuli. Temperature was 37’3°C, pulse 108min, and blood pressure 110/70 mm Hg. Lumbar puncture showed 240 leucocytes/ml (80% granulocytes) and few gramnegative rods, whose morphology and gliding motility raised the possibility of C canimorsus. Treatment with rapidly increasing intravenous doses of penicillin (the drug of choice1) was instituted. The diagnosis was confirmed by positive cultures from CSF. A lumbar puncture 20 h after admission showed more than 2000 leucocytes/ml and was strongly positive for endotoxin by the limulus test; blood culture and this test on serum were negative. The patient was kept on a ventilator for 5 days and slowly improved. Penicillin 5 MU three times daily was given for 15 days, but heart failure and urinary tract infections delayed discharge until 67 days after admission. This case confirms that C canimorsus infections typically follow dog-bites, and that patients with no obvious immunoincompetence can be infected. In our patient early antibiotic treatment was discontinued, and this may have contributed to the dissemination of the disease. We fully agree with Gallen and Ispahani that early antibiotic treatment and debridement are imoortant. KJELD S. KRISTENSEN Department of Internal Medicine, MORTEN WINTHEREIK Central County Hospital, MORTEN LYSDAHL RASMUSSEN Naestved, Denmark man
Association class III-IV)
1
McCarthy M, Zumla A. DF-2 infection. Br Med J 1988; 297:
Detection of
1355-56.
Chlamydia pneumoniae
SIR; Chlamydia pneumoniae
is
a common cause
Inclusion count
-2 -3 -4 -5 -6 -7
1956. 2. Porter IA. Pathogenesis of brucellosis. Scot Med J 1976; 21: 126-27. 3. Vandercam B, Zech F, de Cooman S, Bughin C, Gigi J, Wauters G. Isolate of Brucella melitensis from human sperm. Eur J Clin Microbiol Infect Dis 1990; 9: 303-04.
Capnocytophaga canimorsus infection after dog-bite
Log 10 dilution
>
675 70 15 0 0 0
EIA results
DFA results (cut-off 0,234) > 3 (reactive) Many (reticular bodies) 0-779 (reactive) Few 0-254 (reactive) Rare 0-012 (NR) Negative 0-082 (NR) Negative 0-084 (NR) Negative
NR= non-reactive
There
good specificity for both EIA and DFA tests with inclusion counting, and both techniques could detect about twelve inclusions per specimen. These preliminary results suggest that chlamydial EIA and DFA tests should be studied further. was
compared
Arcand Park Clinic, Dean Medical Center, Madison 53704, USA
Wisconsin State Laboratory of Madison, Wisconsin
DAVID L. HAHN Hygiene,
Grayston JT, Campbell LA, Kuo C-C, et al. A new respiratory tract pathogen: Chlamydia pneumoniae strain TWAR.J Infect Dis 1990; 161: 618-25. 2. Sillis M, White P. Rapid identification of Chlamydia psittaci and TWAR (C pneumoniae) in sputum samples using an amplified enzyme immunoassay. J Clin Pathol 1990; 43: 260. 1.
Effect of massive dose vitamin A on morbidity and mortality in Indian children SIR,-We wish to draw K. Vijayaraghavan and colleagues’ (Dec 1, p 1342) attention to one or two points. In their double-blind trial, after the administration of placebo, the prevalence of night blindness and Bitot spots decreased significantly from 6% to 2-9%. The reason for this reduction was not explained. On the basis of Vijayaraghavan and colleagues’ table ill (morbidity cases) and the percentages of patients who received zero, one, or two doses of vitamin A or placebo given in their results section, we have calculated the morbidity rate according to dose (table). Morbidity at zero dose for the placebo group is greater than that for the treatment group (p < 0-005). The baseline morbidity in the placebo group was probably high. After one dose of vitamin A morbidity increased in the treatment group to 87% but did not change in the placebo group. Why should morbidity increase drastically after one dose of vitamin A? The treatment group might not have been homogeneous. We also find perplexing the fact that morbidity in the placebo group increased by nearly 26% from one to two doses. Your editorial in the same issue (p 1350) states that the field workers who were engaged in the distribution of vitamin A or placebo provided information about management of illness with special reference to diarrhoea, respiratory infection, and measles. =
MORBIDITY STATUS ACCORDING TO DOSE OF VITAMIN A OR PLACEBO*
of bronchitis
and pneumonia,l but
diagnosis is difficult because of non-specific clinical presentation, poorly defmed serodiagnostic criteria, and lack of availability of both culture and serodiagnostic tests. Preliminary evidence suggests that enzyme immunoassay (EIA) and direct fluorescent antibody (DFA) techniques for identification of
Cpneumoniae may be useful.2
RuTH W. DODGE
*Cases/reclplents (aged 1-5 yr).
AT-111-deficient kindreds. In Italy in 1981 only 0-3% of all deaths were directly attributed to venous thromboembolism. National Bureau of Statistics data, based on death certificates, are probably underestimates: in necropsy studies pulmonary embolism regarded as fatal was found in 9% of cases, but in only 0-8-2-4% of patients was this not associated with some other terminal illness.s,6 Information obtained from published reports is subject to bias because in recent years familial AT III quantitative deficiency will not have had the same attention as formerly and kindreds with more unfavourble clinical courses will have been diagnosed and published more frequently. However, kindreds with severe clinical expression of AT III deficiency do exist and long-term prophylaxis with oral anticoagulants needs to be carefully evaluated in all members of these kindreds, whether they have symptoms or not. The necessity of life-long prophylaxis in those who have had a thrombosis is supported by our observation that a recurrent episode probably accounted for half the deaths from venous thromboembolism. A main target of long-term prophylaxis in AT III deficiency should be the prevention of morbidity from venous thromboembolism and of its sequelae, besides the prevention of mortality. AT-III-deficient individuals have the highest risk of first thrombosis between 15 and 50 years,4 and in our opinion the use of anticoagulants in such symptomless individuals, who as a rule have no indication for prophylaxis, should be reconsidered. A complete list of references available from the authors. Istituto di Semeiotica Medica, Catholic University, 00168 Rome, Italy
on
the AT-III-deficient families
analysed is
VALERIO DE STEFANO GIUSEPPE LEONE
1. Leone G, De Stefano V, Di Donfrancesco A, Ferrelli R, Traisci G, Bizzi B. Antithrombin III Pescara: a defective AT III variant with no alterations of plasma crossed immunoelectrophoresis; but with an abnormal crossed immunoelectrofocusing pattern. Br J Haematol 1987; 65: 187-91. 2. Finazzi G, Caccia R, Barbui T. Different prevalence of thromboembolism in the subtypes of congenital antithrombin III deficiency: review of 404 cases. Thromb Haemostas 1987; 58: 1094. 3. De Stefano V, Leone G. Antithrombin III congenital defects: revising classificaton system. Thromb Haemostas 1989; 62: 820-21. 4. Hirsh J, Piovella F, Pini M. Congenital anrithrombin III deficiency: incidence and clinical features. Am J Med 1989; 87 (suppl 3B): 34S-38S. 5. Macintyre IMC, Ruckley CV. Pulmonary embolism: a clinical and autopsy study. Scot Med J 1974; 19: 20-24. 6. Nielsen HK, Bechgaard P, Nielsen PF, Husted SE, Geday E. 178 fatal cases of pulmonary embolism in a medical department. Acta Med Scand 1981; 209: 351-55.
Helicobacter pylori infection in children
cohabiting
SiR,—The association between Helicobacterpylori in the stomach and histologically proven gastritis is well recognised. However, little is known about the sources, potential reservoirs, and routes of transmission of this bacterium. Since Hpylori has been isolated only from human gastric mucosa and is endemic throughout the world human beings are the most likely natural reservoir.’ Person-toperson transmission would be predicted and H pylori antibodies are found in household contacts of infected patients;2 saliva may be the route of transmission. Identification of the same strain of H pylori in cohabiting subjects would provide further confirmatory evidence of this view. We examined Hpylori isolates from areas of gastritis in 5 mentally handicapped children, who lived in close contact, by restriction endonuclease DNA analysis. All children (3 boys and 2 girls) aged 4-12 years were living in the same long-term care facility for several years. Samples were cultured at 37°C for up to five days under microaerophilic conditions on both selective and unselective media. H pylori colonies were subcultured from unselective medium and DNA was extracted by sodium dodecylsulphate alkaline lysis, as previously described.4 Dual restriction endonuclease analysis was completed with 10 units of Hind III for 1 h and 10 units of Bst EII for 1 h at 37°C and 60°C, respectively, in appropriate buffer as described by the manufacturer (Gibco BRL, Cergy-Pontoise, France). DNA fragments were separated by horizontal 7% agarose gel electrophoresis and photographed with an ’MP-4 Land’ camera
(Polaroid).
isolates, digested bacterial DNA samples were in electrophoresed the same gel with other H pylori strains from unrelated patients. Some had indistinguishable restriction patterns. Two children had identical patterns, as did two others. No overlap with other strains was observed. Our study shows that two individuals with infectious gastritis can be infected by the same strain. Contamination is unlikely because the endoscope was disinfected and the grip changed before each procedure, the linked children were not endoscoped on the same day, and bacterial counts were high. Many epidemiological studies have shown that living conditions may influence the communicability of H pylori. In our study, the living conditions suggest that person-to-person transmission was likely and that there was a common source of infection. The importance of saliva remains uncertain but some studies suggest that the mouth could be a reservoir’ and that oro-oral transmission could take place. To compare
Bacteriology Laboratory A, Faculty of Medicine, 59 045 Lille, France; and Paediatric Service, CHR de Lille, and CPIJ, Vendin-le-Viel
P. VINCENT F. GOTTRAND P. PERNES M. A. HUSSON A. BEJU H. LECLERC J. P. FARRIAUX
1. Graham DY, Adam E, Klein PD, et al. Epidemiology of Campylobacter pylon infection. Gastroenterol Clin Biol 1989; 13: 84B-89B 2. Graham DY, Malatty HM, Klein PD, Evans DG, Evans DJ, Adam E. Helicobacter pylon infection clusters in family. Rev Esp Enferm Apar Dig 1990; 78: 26. 3. Albenque M, Tall F, Dabis F, Megraud F. Epidemiological study of Helicobacter pylori transmission from mother to child in Africa Rev Esp Enferm Apar Dig 1990, 78: 48 4. Beji A, Izard D, Gaviny F, Leclerc H, Leseine-Delstanche M, Krembel J. A rapid chemical procedure for isolation and purification of chromosomal DNA from Gram-negative bacilli. Ann Biochem 1987; 162: 18-23 5. Krajden S, Fuksa M, Anderson J, et al. Examination of human stomach biopsies, saliva and dental plaque for Campylobacter pylori.J Clin Microbiol 1989; 27: 1397-98
Transmission of Brucella melitensis SIR,-Dr Ruben and colleagues (Jan 5, p 14) report person-toperson transmission of Brucella melitensis. The following case provides further evidence of between-person transmission of brucellosis. A 35-year-old Swedish man and his 30-year-old girlfriend spent two weeks in the south of Spain in June, 1989. They shared accommodation and meals with three other people. No-one had symptoms during the stay. The man had respiratory tract symptoms and remittent fever 14 weeks after leaving Spain. 4 weeks later swelling of the right knee developed and 8 weeks after this (December, 1989) he had a right-sided epididymitis. He was given amoxycillin and cefadroxil for airway symptoms, diclofenac for arthritis, and ciprofloxacin for epididymitis. Faecal and blood cultures were negative, but the latter were not taken until after antibiotics were begun. In December he had titres of 640 against Yersinia enterocolitica type 9 and 320 against Brucella abortus. He eventually recovered uneventfully and was believed to have had
yersinia infection. Stored serum samples were sent to the state bacteriological laboratory, Solna, Sweden, for analysis by ELISA. A rising IgG titre (peak 850) and decreasing IgM titre (peak > 1500) compatible with brucella infection were recorded. This patient’s girlfriend became ill in April, 1990-9 months after the stay in Spain-with sudden fever, sweating, headache, and leucopenia. Repeated blood cultures grew B melitensis. Serum agglutinin titre to B abortus was 320 and to Y enterocolitica type 9 160. Peak antibody titres to brucella measured with ELISA were IgM over 1500 and IgG 850. She was given doxycycline and rifampicin and recovered without complications. Although the diagnosis of the man was not verified by culture, the clinical picture as well as the serological findings are, in retrospect, compatible with brucellosis. There is serological crossreactivity between brucella and yersinia antigens, which may account for the serological reactivity that confused the diagnosis. The disease started in the
woman
9 months after
possible
exposure. This exceeds the maximum recorded incubation time of 7 months.1,2 It therefore seems unlikely that both patients were
infected in Spain. The fact that the three individuals with whom they shared accommodation at that time were not infected and had
849
brucella when tested 9 months later is evidence of infection. Brucellosis in domestic animals has not been recorded in Sweden since the 1950s, and human brucellosis acquired in this country has not been reported since then. The possibility of coital transmission of brucella has long been suggested, but only recently has brucellosis been isolated from human spennatozoa.3 We thus conclude that the woman was probably infected by her boyfriend, possibly by sexual intercourse during genital manifestation of the disease. no
antibodies
against
to
a common source
Departments of Infectious Diseases and Clinical Bacteriology, University of Goteborg,
S41685-Goteborg, 1.
JOHAN LINDBERG PETER LARSSON
Sweden
To assess the minimum number of inclusions detectable by these techniques, we inoculated serial dilutions of C pneumoniae elementary bodies onto McCoy cell monolayers (in triplicate) and incubated them for three days. One monolayer at each dilution was then stained with chlamydial fluorescein-conjugated lipopolysaccharide antibody and inclusions were counted. A second monolayer was swabbed and tested with a commercially available EIA (’STD-EZE’, Abbott). The third monolayer was swabbed, rolled onto a glass slide, and tested by a commercially available DFA test (’Microtrak’, Syva, USA) with lipopolysaccharide rather than major outer-membrane protein fluorescein-conjugated antibody. C pneumoniae inclusion counts with EIA and DFA for serial dilutions are given below:
Spink WW. The nature of brucellosis. Minneapolis: University of Minnesota Press,
SIR,-Dr Gallen and Dr Ispahani (Feb 2, p 308) report a case of fulminant Capnocytophaga canimorsus septicaemia. We have seen another case of severe C canimorsus disease, in which immediate suspicion of the infection made prompt and adequate treatment
possible. A 74-year-old
with cardiac failure (New York Heart was bitten by a dog. Signs of local infection then developed, and he was started on roxithromycin 150 mg twice daily (allergy to penicillins was suspected). The local reaction subsided, but treatment was discontinued after three doses because of diarrhoea. Fever, lethargy, and increasing confusion developed over the next two days. On admission several days after the bite the patient was unresponsive to verbal stimuli, but reacted with agitation to painful stimuli. Temperature was 37’3°C, pulse 108min, and blood pressure 110/70 mm Hg. Lumbar puncture showed 240 leucocytes/ml (80% granulocytes) and few gramnegative rods, whose morphology and gliding motility raised the possibility of C canimorsus. Treatment with rapidly increasing intravenous doses of penicillin (the drug of choice1) was instituted. The diagnosis was confirmed by positive cultures from CSF. A lumbar puncture 20 h after admission showed more than 2000 leucocytes/ml and was strongly positive for endotoxin by the limulus test; blood culture and this test on serum were negative. The patient was kept on a ventilator for 5 days and slowly improved. Penicillin 5 MU three times daily was given for 15 days, but heart failure and urinary tract infections delayed discharge until 67 days after admission. This case confirms that C canimorsus infections typically follow dog-bites, and that patients with no obvious immunoincompetence can be infected. In our patient early antibiotic treatment was discontinued, and this may have contributed to the dissemination of the disease. We fully agree with Gallen and Ispahani that early antibiotic treatment and debridement are imoortant. KJELD S. KRISTENSEN Department of Internal Medicine, MORTEN WINTHEREIK Central County Hospital, MORTEN LYSDAHL RASMUSSEN Naestved, Denmark man
Association class III-IV)
1
McCarthy M, Zumla A. DF-2 infection. Br Med J 1988; 297:
Detection of
1355-56.
Chlamydia pneumoniae
SIR; Chlamydia pneumoniae
is
a common cause
Inclusion count
-2 -3 -4 -5 -6 -7
1956. 2. Porter IA. Pathogenesis of brucellosis. Scot Med J 1976; 21: 126-27. 3. Vandercam B, Zech F, de Cooman S, Bughin C, Gigi J, Wauters G. Isolate of Brucella melitensis from human sperm. Eur J Clin Microbiol Infect Dis 1990; 9: 303-04.
Capnocytophaga canimorsus infection after dog-bite
Log 10 dilution
>
675 70 15 0 0 0
EIA results
DFA results (cut-off 0,234) > 3 (reactive) Many (reticular bodies) 0-779 (reactive) Few 0-254 (reactive) Rare 0-012 (NR) Negative 0-082 (NR) Negative 0-084 (NR) Negative
NR= non-reactive
There
good specificity for both EIA and DFA tests with inclusion counting, and both techniques could detect about twelve inclusions per specimen. These preliminary results suggest that chlamydial EIA and DFA tests should be studied further. was
compared
Arcand Park Clinic, Dean Medical Center, Madison 53704, USA
Wisconsin State Laboratory of Madison, Wisconsin
DAVID L. HAHN Hygiene,
Grayston JT, Campbell LA, Kuo C-C, et al. A new respiratory tract pathogen: Chlamydia pneumoniae strain TWAR.J Infect Dis 1990; 161: 618-25. 2. Sillis M, White P. Rapid identification of Chlamydia psittaci and TWAR (C pneumoniae) in sputum samples using an amplified enzyme immunoassay. J Clin Pathol 1990; 43: 260. 1.
Effect of massive dose vitamin A on morbidity and mortality in Indian children SIR,-We wish to draw K. Vijayaraghavan and colleagues’ (Dec 1, p 1342) attention to one or two points. In their double-blind trial, after the administration of placebo, the prevalence of night blindness and Bitot spots decreased significantly from 6% to 2-9%. The reason for this reduction was not explained. On the basis of Vijayaraghavan and colleagues’ table ill (morbidity cases) and the percentages of patients who received zero, one, or two doses of vitamin A or placebo given in their results section, we have calculated the morbidity rate according to dose (table). Morbidity at zero dose for the placebo group is greater than that for the treatment group (p < 0-005). The baseline morbidity in the placebo group was probably high. After one dose of vitamin A morbidity increased in the treatment group to 87% but did not change in the placebo group. Why should morbidity increase drastically after one dose of vitamin A? The treatment group might not have been homogeneous. We also find perplexing the fact that morbidity in the placebo group increased by nearly 26% from one to two doses. Your editorial in the same issue (p 1350) states that the field workers who were engaged in the distribution of vitamin A or placebo provided information about management of illness with special reference to diarrhoea, respiratory infection, and measles. =
MORBIDITY STATUS ACCORDING TO DOSE OF VITAMIN A OR PLACEBO*
of bronchitis
and pneumonia,l but
diagnosis is difficult because of non-specific clinical presentation, poorly defmed serodiagnostic criteria, and lack of availability of both culture and serodiagnostic tests. Preliminary evidence suggests that enzyme immunoassay (EIA) and direct fluorescent antibody (DFA) techniques for identification of
Cpneumoniae may be useful.2
RuTH W. DODGE
*Cases/reclplents (aged 1-5 yr).
