36 Continuous metiamide therapy, in a dose of 200 mg. four times daily, produced a dramatic response, with disappearance of symptoms in 9 days. Repeat endoscopy demonstrated healing of the jejunal ulcer after one month, and the stomal ulcer after two months. This was associated with a fall in the serum-gastrin to 235 pg. per ml. after two months, which rose again to 825 pg. per ml. after four months.
The non-invasive diagnosis of the Zollinger-Ellison syndrome hinges on the secretin test: a marked rise in a serum-gastrin, which is almost all big " gastrin, is very suggestive of a tumour. This test does require further evaluation before its routine use can be accepted. The therapeutic effect of metiamide in this case was dramatic, suggesting a major therapeutic role for metiamide in patients with refractory peptic ulceration, and possibly in the Zollinger-Ellison syndrome. We are at a loss to explain the striking changes in the fasting serum-gastrin during therapy, but they were greater than any of the day-to-day variations we have seen in this patient. Another two similar patients have been treated with
not contribute to cerebral symptoms and it is serious problem in acute falciparum malarias treated adequately in the endemic area.
tion does not a
Ramathibodi
Hospital, Bangkok 4, and Phrabudhabat Hospital, Saraburi, Thailand.
"
metiamide over a shorter period. The initial response assessed symptomatically and endoscopically was as good as in the above patient. Department of Surgery, University of Newcastle upon Tyne, Newcastle upon Tyne NE1 4LP. Sunderland Royal Infirmary, Sunderland, Co. Durham.
Department of Physiology, University of Newcastle upon Tyne.
M. H. THOMPSON C. W. VENABLES. I. T. J. D. D. J. E. R. E. L.
MILLER.
REED SANDERS GRUND BLAIR.
INTRAVASCULAR COAGULATION AND MALARIA
conflicting reports 14 of the incidence of increased intravascular coagulation in malaria. Most reports of increased intravascular coagulation came from selected groups of hospital inpatients in non-endemic We have found evidence contrary to the majority areas. opinion. We studied 24 cases of acute falciparum malaria first seen and admitted in a hospital in the endemic area. All were treated promptly with quinine. 6 were complicated by clinical jaundice, 2 had cerebral symptoms, and 1 developed fatal pulmonary oedema. The others recovered. SIR,-There
are
Parasite counts and platelet counts were made on admission. Fibrinogen levels were determined by radial immunodiffusion. Serum fibrinogen degradation products !F.D.P.) were determined according to Das,,’ and F.D.P. and fibrinogen values on the third day after admission were available in 16 cases. In 8 cases there was a slight increase in serum-F.D.P. (above 10 tLg. per ml.) and hypofibrinogenmmia (below 200 mg. per 100 ml.), singly or together. In these 8 cases, all except 1 had heavy parasitaEmia above 100,000 per c.mm. and clinical complications. Those which did not have increased intravascular coagulation were 2 patients with cerebral malaria, 1 with jaundice, and 1 with no complications and a parasite count above 100,000 per c.mm. Thrombocytopenia was found in 23 cases. Thrombocytopenia in malaria could be caused by other mechanismsand it should not be taken as evidence of intravascular coagulation. Highest serum-F.D.p. levels in these patients are still much lower than the value found in classical disseminated intravascular coagulation. In our patients a mildly increased intravascular coagulation could be detected in a third and it was associated closely with heavy parasitoemia.
Our conclusion is that increased intravascular
cbagula-
Reid, H. A., Nkrumah, R. K. Lancet, 1972, i, 218. Jaroonvesama, N. ibid. p. 221. 3. Neva, F. A., Sheagren, J. N., Shulman, N. R., Canfield, C. J. Ann. intern Med. 1970, 73, 295. 4. Punyagupta, S., Srichaikul, R., Nitiyanant, P., Petchclai, B. Am. J. trop. Med. Hyg. 1974, 23, 551. 5. Das, P. C. J. clin. Path. 1970, 23, 299.
1. 2.
BENCHA PETCHCLAI SOPAPORN PRASONGSOM WISITH BENJAPONGS.
ANTICOAGULANT CONTROL, IMMUNOSUPPRESSION, AND GLOMERULONEPHRITIS
SIR,-In a paper (Nov. 16, p. 1166) on anticoagulant immunosuppressant treatment of rapidly progressive glomerulonephritis, the data on the degree of anticoagula-
and
tion and its control
were
incorrect.
coagulant regimen was carried hoematology, as follows:
out
Control of the antiin the department of
Heparin therapy was controlled on the kaolin-cephalin time (K.C.T.). The patients injected their heparin subcutaneously into the anterior abdominal wall, the doses varying between 6000 and 15,000 units 8-hourly, and blood-samples for the K.C.T. were taken 3-4 hours after the last injection. The K.C.T. was prolonged for 60-100 seconds (normal control 30-35 seconds). The K.C.T. was preferred to the thrombin-time, after trial, because of its better reproducibility. At aK.C.T. of 60-100 seconds the corresponding thrombin-times were prolonged 5-fold or more above the control time of 11-14 seconds and the end-points were very poor. Warfarin therapy was controlled by the one-stage prothrombintime using human brain thromboplastin (Manchester Comparative Reagent), the normal control times being 12-13 seconds and the therapeutic range 30-50 seconds. This therapeutic range corresponds to a prothrombin concentration of 15-5%, as determined from a dilution curve prepared from normal plasma diluted with barium-sulphate-absorbed normal plasma. This range has been in use for some years at Guy’s for all patients on oral anticoagulant therapy. "
"
The level of anticoagulation was, in greater than was implied in the paper.
Department of Hæmatology, Renal
Unit, Guy’s Hospital, London SE1 9RT.
general, somewhat P. BARKHAN P. J. BLACK T. THOMAS. C. B. BROWN
J. S. CAMERON C. S. OGG.
CALCIUM BALANCE IN PREGNANCY
SIR,-Dr Hytten and his colleagues (Nov. 9, p. 1152) make several criticisms of our paper (Oct. 19, p. 926) on this subject. The references used to support their contention that metabolic balance studies are of little value 1,2 do not take into account the use of faecal markers and do not fulfil the requirements set down by later workers,3,4 especially regarding previous dietary histories and the importance Our studies of subjects continuing on the same diet, &c. fulfilled these criteria and, as the agreement between results in the two balance periods was good, then stabilisation of the subject has almost certainly occurred. We do not recommend a calcium intake of 2 g. per day, but suggest that our results on the 7 subjects we studied indicate that this amount may be necessary for calcium balance. Further studies are obviously required. Radiological studies are notoriously unreliable in establishing loss of bone calcium since the loss has to be considerable before any change can be detected. Finally, the biochemical results in plasma and urine in 1. 2. 3. 4.
Duncan, D. Forbes, G. Hargreaves, Hartley, T. 1974, 52,
L. Nutr. Abstr. Rev. 1958, 28, 695. Nutr. Rev. 1973, 31, 297. T., Rose, G. A. Clin. Sci. 1965, 28, 537. F., Dawson, J. B., Hodgkinson, A. Clinica chim. Acta, 321.
37
compared with those obtained
from a females of the same age. We do group of non-pregnant is an that abnormal state nor do we not suggest pregnancy this was an abnormal state these that for particular imply subjects. We were aware that these results were normal for pregnant women. G. G. DUGGIN Renal Unit and Biochemistry NANCY E. DALE Department, R. C. LYNEHAM Royal Prince Alfred Hospital, R. A. EVANS Missenden Road, D. J. TILLER. Camperdown 2050, New South Wales. our
subjects
were
PHENOBARBITONE, LIVER TUMOURS, AND THOROTRAST
SIR,-Dr Schneiderman’s interpretation (Nov. 2, p. 1085) of the data on cancer in epilepsy patients1 suffers because of the necessary brevity of statistical details in medical journals, and because of some generalisations from American experience to Danish data. In the main, the Danish data showed cancer morbidity of about the expected values for patients treated less than ten years. For those treated over longer periods most cancers were fewer, while cancer of the liver showed some " excess, although not statistically significant. As a usually accepted " explanation Dr Schneiderman mentions that
institutionalisation, besides reducing exposure to oncogens, will also cause occasional under-reporting of deaths ascribed to one cause in a population known to have another, sometimes fatal, disease. This explanation does not apply to efficient research hospitals. The neuropsychiatric hospital, Filadelfia, served as the main centre for epilepsy treatment and research in Denmark from the late 1930s to the late 1950s, when treatment became less centralised. As indicated by the dates for the introduction of new drugs, the medical staff was alert to development, and research activities produced a series of publications. 2-7 From my experience during short-term appointments in 1930-32, I can testify to the high standard of Filadelfia’s case-records, maintained during the tenure of the chief physician, Stubbe Teglbjxrg, from 1928 to 1959. That his proverbial interest in the fate of his patients certainly included the risk of liver cancer is shown by the fact that he referred the use of ’Thorotrast’ (thorium dioxide) for angiography to neurosurgical units elsewhere following a warning in 1939which I based on a discussion in the Pathological Society for Great Britain and Ireland in November, 1938. Earlier application of this compound in a number of cases naturally turned attention at Filadelfia to liver cancer. Furthermore, since patients at Filadelfia developing cancer would be referred for treatment to other hospitals, there is no reason to suppose that reporting of cancer would be less efficient for them than average. With regard to the fraction known from death certificates, up to the introduction of international certificates a cancer present at death would in Denmark invariably be classified as the primary cause of death. As appears from table i,l consequently, cancers other than those with significant deviations from the expected values showed no reduction in frequency for patients treated less than ten years. Schneiderman agrees that the deficit in numbers for 1. 2. 3. 4. 5. 6.
Clemmesen, J., Frederiksen, V. F., Plum, C. M. Lancet, 1974, i, 705. Clemmesen, C. Acta psych. neurol. 1932, suppl. 3. Teglbjærg, H. P. S. ibid. 1936, suppl. 9. Faurbye, A. Blodets reaktion ved epilepsi. Copenhagen, 1942. Hendriksen, V. Spasmofili og epilepsi. Copenhagen, 1923. Munch-Petersen, C. J. Spinalvæskens sukkerindhold. Copenhagen, 1929.
Yde, A. Nyrefunktionen ved epilepsi. Copenhagen, 1938. 8. Clemmesen, J. Ugeskr. læg. 1939, 664; Selbie, F. Lancet, 1936, ii, 7.
847.
cervical uterine
compared with expected values is reduced sexual activity among epilepsy explainable by He at Filadelfia. adds that in a population with patients a low cervical-cancer rate one usually finds high breastcancer rates, suggesting that the low rates found at Filadelfia may be due to under-reporting, and thus disregarding long-established observations. In Denmark urban/rural comparison shows parallel occurrence of cervical and mammary carcinomas,9 and in Copenhagen the social trend for cervical carcinoma has no reverse counterpart for breast cancer, the latter showing no unambiguous social trend.10° Furthermore, age-standardised morbidity rates for Denmark, 1943-67, known to Schneiderman in proof, show parallel rising secular trends for the two sites in all urban and rural categories.ll No final explanation of this parallelism is yet available, but it may be noted that the age-distribution of cervical carcinoma in Denmark differs from that of many other countries. It is true that the total for other sites fail to reach the expected values for patients treated for more than ten years, with the exception of carcinoma of the liver. However, once having accepted cancers of different sites as of different causation, and thus forming a non-homogenous group, it seems somewhat illogical that Schneiderman attempts an estimate of liver cancer in proportion to their total (exclusive of brain tumours). Employing in this way the methods also used by Clemmesen et al., Schneiderman underlines that one more case of liver cancer would mean statistical significance-thus arriving at what may be termed hypothetical significancean evaluation we would hesitate to use in view of its consequences (for example, in estimating results in cancer cancer
therapy). It should be strongly emphasised that Clemmesen et al., among their cases of liver carcinoma, included a male patient, treated for less than ten years for epilepsy, and known to have received a dose of thorotrast equivalent to 440 rad in 1946-a dose that would have justified exclusion from the material because of its oncogenic effect. Since, nevertheless, Schneiderman seems to attach significance to the small excess of liver cancer, which, even including all cases, does not reach the level of statistical significance, we feel we should give further details of this and other cases for which the accurate dose of thorotrast is not known, together with cases not having received the compound: Males 28364: Born Oct. 13, 1903. Epilepsy diagnosed, 1936. Thorotrast 440 rad. 1946. Died Nov. 23, 1964. Cholangiocarcinoma. Cirrhosis hepatis. 11002: Born Oct. 10, 1915. Epilepsy diagnosed about 1926. Thorotrast 1938. Died April 7, 1965. Hepatocarcinoma. 17123: Born Sept. 14, 1930. Epilepsy diagnosed 1942. Thorotrast Nov. 30, 1942. Died July 17, 1967. Cholangiocarcinoma. 7783: Born Oct. 15, 1905. Epilepsy diagnosed 1947. Died April 18, 1967. Adenocarcinoma hepatis, probably cholangio-
carcinoma. Female 23198: Born March 3, 1899. Epilepsy diagnosed 1946. Died July 7, 1964. Necropsy: primary liver cancer (no histology).
Since the latest symposium on the toxicity of thorotrast 12 failed to determine the oncogenic dose of this compound, and with expected values for liver cancer of 1-1 for males and 0-7 for females treated with anticonvulsants for more than ten years, we trust that Dr Schneiderman will consider our statistics conservative, stating as we did that we found no evidence that phenobarbitone is 9. 10. 11. 12.
Clemmesen, J. J. natn. Cancer Inst. 1951, 12, 1. Clemmesen, J. Acta path. microbiol. scand. 1965, suppl. 174; ibid. 1969, suppl. 209. Clemmesen, J. ibid. 1974, suppl. 247. Proceedings of III International Meeting on Toxicity of Thorotrast, 1973. Danish Atomic Energy Commission Riso Rep. 294. Roskilde, Denmark.
The non-invasive diagnosis of the Zollinger-Ellison syndrome hinges on the secretin test: a marked rise in a serum-gastrin, which is almost all big " gastrin, is very suggestive of a tumour. This test does require further evaluation before its routine use can be accepted. The therapeutic effect of metiamide in this case was dramatic, suggesting a major therapeutic role for metiamide in patients with refractory peptic ulceration, and possibly in the Zollinger-Ellison syndrome. We are at a loss to explain the striking changes in the fasting serum-gastrin during therapy, but they were greater than any of the day-to-day variations we have seen in this patient. Another two similar patients have been treated with
not contribute to cerebral symptoms and it is serious problem in acute falciparum malarias treated adequately in the endemic area.
tion does not a
Ramathibodi
Hospital, Bangkok 4, and Phrabudhabat Hospital, Saraburi, Thailand.
"
metiamide over a shorter period. The initial response assessed symptomatically and endoscopically was as good as in the above patient. Department of Surgery, University of Newcastle upon Tyne, Newcastle upon Tyne NE1 4LP. Sunderland Royal Infirmary, Sunderland, Co. Durham.
Department of Physiology, University of Newcastle upon Tyne.
M. H. THOMPSON C. W. VENABLES. I. T. J. D. D. J. E. R. E. L.
MILLER.
REED SANDERS GRUND BLAIR.
INTRAVASCULAR COAGULATION AND MALARIA
conflicting reports 14 of the incidence of increased intravascular coagulation in malaria. Most reports of increased intravascular coagulation came from selected groups of hospital inpatients in non-endemic We have found evidence contrary to the majority areas. opinion. We studied 24 cases of acute falciparum malaria first seen and admitted in a hospital in the endemic area. All were treated promptly with quinine. 6 were complicated by clinical jaundice, 2 had cerebral symptoms, and 1 developed fatal pulmonary oedema. The others recovered. SIR,-There
are
Parasite counts and platelet counts were made on admission. Fibrinogen levels were determined by radial immunodiffusion. Serum fibrinogen degradation products !F.D.P.) were determined according to Das,,’ and F.D.P. and fibrinogen values on the third day after admission were available in 16 cases. In 8 cases there was a slight increase in serum-F.D.P. (above 10 tLg. per ml.) and hypofibrinogenmmia (below 200 mg. per 100 ml.), singly or together. In these 8 cases, all except 1 had heavy parasitaEmia above 100,000 per c.mm. and clinical complications. Those which did not have increased intravascular coagulation were 2 patients with cerebral malaria, 1 with jaundice, and 1 with no complications and a parasite count above 100,000 per c.mm. Thrombocytopenia was found in 23 cases. Thrombocytopenia in malaria could be caused by other mechanismsand it should not be taken as evidence of intravascular coagulation. Highest serum-F.D.p. levels in these patients are still much lower than the value found in classical disseminated intravascular coagulation. In our patients a mildly increased intravascular coagulation could be detected in a third and it was associated closely with heavy parasitoemia.
Our conclusion is that increased intravascular
cbagula-
Reid, H. A., Nkrumah, R. K. Lancet, 1972, i, 218. Jaroonvesama, N. ibid. p. 221. 3. Neva, F. A., Sheagren, J. N., Shulman, N. R., Canfield, C. J. Ann. intern Med. 1970, 73, 295. 4. Punyagupta, S., Srichaikul, R., Nitiyanant, P., Petchclai, B. Am. J. trop. Med. Hyg. 1974, 23, 551. 5. Das, P. C. J. clin. Path. 1970, 23, 299.
1. 2.
BENCHA PETCHCLAI SOPAPORN PRASONGSOM WISITH BENJAPONGS.
ANTICOAGULANT CONTROL, IMMUNOSUPPRESSION, AND GLOMERULONEPHRITIS
SIR,-In a paper (Nov. 16, p. 1166) on anticoagulant immunosuppressant treatment of rapidly progressive glomerulonephritis, the data on the degree of anticoagula-
and
tion and its control
were
incorrect.
coagulant regimen was carried hoematology, as follows:
out
Control of the antiin the department of
Heparin therapy was controlled on the kaolin-cephalin time (K.C.T.). The patients injected their heparin subcutaneously into the anterior abdominal wall, the doses varying between 6000 and 15,000 units 8-hourly, and blood-samples for the K.C.T. were taken 3-4 hours after the last injection. The K.C.T. was prolonged for 60-100 seconds (normal control 30-35 seconds). The K.C.T. was preferred to the thrombin-time, after trial, because of its better reproducibility. At aK.C.T. of 60-100 seconds the corresponding thrombin-times were prolonged 5-fold or more above the control time of 11-14 seconds and the end-points were very poor. Warfarin therapy was controlled by the one-stage prothrombintime using human brain thromboplastin (Manchester Comparative Reagent), the normal control times being 12-13 seconds and the therapeutic range 30-50 seconds. This therapeutic range corresponds to a prothrombin concentration of 15-5%, as determined from a dilution curve prepared from normal plasma diluted with barium-sulphate-absorbed normal plasma. This range has been in use for some years at Guy’s for all patients on oral anticoagulant therapy. "
"
The level of anticoagulation was, in greater than was implied in the paper.
Department of Hæmatology, Renal
Unit, Guy’s Hospital, London SE1 9RT.
general, somewhat P. BARKHAN P. J. BLACK T. THOMAS. C. B. BROWN
J. S. CAMERON C. S. OGG.
CALCIUM BALANCE IN PREGNANCY
SIR,-Dr Hytten and his colleagues (Nov. 9, p. 1152) make several criticisms of our paper (Oct. 19, p. 926) on this subject. The references used to support their contention that metabolic balance studies are of little value 1,2 do not take into account the use of faecal markers and do not fulfil the requirements set down by later workers,3,4 especially regarding previous dietary histories and the importance Our studies of subjects continuing on the same diet, &c. fulfilled these criteria and, as the agreement between results in the two balance periods was good, then stabilisation of the subject has almost certainly occurred. We do not recommend a calcium intake of 2 g. per day, but suggest that our results on the 7 subjects we studied indicate that this amount may be necessary for calcium balance. Further studies are obviously required. Radiological studies are notoriously unreliable in establishing loss of bone calcium since the loss has to be considerable before any change can be detected. Finally, the biochemical results in plasma and urine in 1. 2. 3. 4.
Duncan, D. Forbes, G. Hargreaves, Hartley, T. 1974, 52,
L. Nutr. Abstr. Rev. 1958, 28, 695. Nutr. Rev. 1973, 31, 297. T., Rose, G. A. Clin. Sci. 1965, 28, 537. F., Dawson, J. B., Hodgkinson, A. Clinica chim. Acta, 321.
37
compared with those obtained
from a females of the same age. We do group of non-pregnant is an that abnormal state nor do we not suggest pregnancy this was an abnormal state these that for particular imply subjects. We were aware that these results were normal for pregnant women. G. G. DUGGIN Renal Unit and Biochemistry NANCY E. DALE Department, R. C. LYNEHAM Royal Prince Alfred Hospital, R. A. EVANS Missenden Road, D. J. TILLER. Camperdown 2050, New South Wales. our
subjects
were
PHENOBARBITONE, LIVER TUMOURS, AND THOROTRAST
SIR,-Dr Schneiderman’s interpretation (Nov. 2, p. 1085) of the data on cancer in epilepsy patients1 suffers because of the necessary brevity of statistical details in medical journals, and because of some generalisations from American experience to Danish data. In the main, the Danish data showed cancer morbidity of about the expected values for patients treated less than ten years. For those treated over longer periods most cancers were fewer, while cancer of the liver showed some " excess, although not statistically significant. As a usually accepted " explanation Dr Schneiderman mentions that
institutionalisation, besides reducing exposure to oncogens, will also cause occasional under-reporting of deaths ascribed to one cause in a population known to have another, sometimes fatal, disease. This explanation does not apply to efficient research hospitals. The neuropsychiatric hospital, Filadelfia, served as the main centre for epilepsy treatment and research in Denmark from the late 1930s to the late 1950s, when treatment became less centralised. As indicated by the dates for the introduction of new drugs, the medical staff was alert to development, and research activities produced a series of publications. 2-7 From my experience during short-term appointments in 1930-32, I can testify to the high standard of Filadelfia’s case-records, maintained during the tenure of the chief physician, Stubbe Teglbjxrg, from 1928 to 1959. That his proverbial interest in the fate of his patients certainly included the risk of liver cancer is shown by the fact that he referred the use of ’Thorotrast’ (thorium dioxide) for angiography to neurosurgical units elsewhere following a warning in 1939which I based on a discussion in the Pathological Society for Great Britain and Ireland in November, 1938. Earlier application of this compound in a number of cases naturally turned attention at Filadelfia to liver cancer. Furthermore, since patients at Filadelfia developing cancer would be referred for treatment to other hospitals, there is no reason to suppose that reporting of cancer would be less efficient for them than average. With regard to the fraction known from death certificates, up to the introduction of international certificates a cancer present at death would in Denmark invariably be classified as the primary cause of death. As appears from table i,l consequently, cancers other than those with significant deviations from the expected values showed no reduction in frequency for patients treated less than ten years. Schneiderman agrees that the deficit in numbers for 1. 2. 3. 4. 5. 6.
Clemmesen, J., Frederiksen, V. F., Plum, C. M. Lancet, 1974, i, 705. Clemmesen, C. Acta psych. neurol. 1932, suppl. 3. Teglbjærg, H. P. S. ibid. 1936, suppl. 9. Faurbye, A. Blodets reaktion ved epilepsi. Copenhagen, 1942. Hendriksen, V. Spasmofili og epilepsi. Copenhagen, 1923. Munch-Petersen, C. J. Spinalvæskens sukkerindhold. Copenhagen, 1929.
Yde, A. Nyrefunktionen ved epilepsi. Copenhagen, 1938. 8. Clemmesen, J. Ugeskr. læg. 1939, 664; Selbie, F. Lancet, 1936, ii, 7.
847.
cervical uterine
compared with expected values is reduced sexual activity among epilepsy explainable by He at Filadelfia. adds that in a population with patients a low cervical-cancer rate one usually finds high breastcancer rates, suggesting that the low rates found at Filadelfia may be due to under-reporting, and thus disregarding long-established observations. In Denmark urban/rural comparison shows parallel occurrence of cervical and mammary carcinomas,9 and in Copenhagen the social trend for cervical carcinoma has no reverse counterpart for breast cancer, the latter showing no unambiguous social trend.10° Furthermore, age-standardised morbidity rates for Denmark, 1943-67, known to Schneiderman in proof, show parallel rising secular trends for the two sites in all urban and rural categories.ll No final explanation of this parallelism is yet available, but it may be noted that the age-distribution of cervical carcinoma in Denmark differs from that of many other countries. It is true that the total for other sites fail to reach the expected values for patients treated for more than ten years, with the exception of carcinoma of the liver. However, once having accepted cancers of different sites as of different causation, and thus forming a non-homogenous group, it seems somewhat illogical that Schneiderman attempts an estimate of liver cancer in proportion to their total (exclusive of brain tumours). Employing in this way the methods also used by Clemmesen et al., Schneiderman underlines that one more case of liver cancer would mean statistical significance-thus arriving at what may be termed hypothetical significancean evaluation we would hesitate to use in view of its consequences (for example, in estimating results in cancer cancer
therapy). It should be strongly emphasised that Clemmesen et al., among their cases of liver carcinoma, included a male patient, treated for less than ten years for epilepsy, and known to have received a dose of thorotrast equivalent to 440 rad in 1946-a dose that would have justified exclusion from the material because of its oncogenic effect. Since, nevertheless, Schneiderman seems to attach significance to the small excess of liver cancer, which, even including all cases, does not reach the level of statistical significance, we feel we should give further details of this and other cases for which the accurate dose of thorotrast is not known, together with cases not having received the compound: Males 28364: Born Oct. 13, 1903. Epilepsy diagnosed, 1936. Thorotrast 440 rad. 1946. Died Nov. 23, 1964. Cholangiocarcinoma. Cirrhosis hepatis. 11002: Born Oct. 10, 1915. Epilepsy diagnosed about 1926. Thorotrast 1938. Died April 7, 1965. Hepatocarcinoma. 17123: Born Sept. 14, 1930. Epilepsy diagnosed 1942. Thorotrast Nov. 30, 1942. Died July 17, 1967. Cholangiocarcinoma. 7783: Born Oct. 15, 1905. Epilepsy diagnosed 1947. Died April 18, 1967. Adenocarcinoma hepatis, probably cholangio-
carcinoma. Female 23198: Born March 3, 1899. Epilepsy diagnosed 1946. Died July 7, 1964. Necropsy: primary liver cancer (no histology).
Since the latest symposium on the toxicity of thorotrast 12 failed to determine the oncogenic dose of this compound, and with expected values for liver cancer of 1-1 for males and 0-7 for females treated with anticonvulsants for more than ten years, we trust that Dr Schneiderman will consider our statistics conservative, stating as we did that we found no evidence that phenobarbitone is 9. 10. 11. 12.
Clemmesen, J. J. natn. Cancer Inst. 1951, 12, 1. Clemmesen, J. Acta path. microbiol. scand. 1965, suppl. 174; ibid. 1969, suppl. 209. Clemmesen, J. ibid. 1974, suppl. 247. Proceedings of III International Meeting on Toxicity of Thorotrast, 1973. Danish Atomic Energy Commission Riso Rep. 294. Roskilde, Denmark.
