Feb 2, 1980 - pharynx, round the mouth, andon the lower legs. There was no ... was an alcoholic with bleeding varices who evidently had had thrombo-.
BRITISH MEDICAL JOURNAL
293
2 FEBRUARY 1980
polyradiculoneuropathy of subacute onset, seems to be more in keeping with an immunologically induced demyelination. I thank Dr E C Hutchinson, North Staffordshire Royal Infirmary, for his permission to report this case.
'Russell RWR. Giant-cell arteritis. A review of 35 cases. Q J7 Med 1959; 28:47 1-89. 2 Warrell DA, Godfrey S, Olsen EGJ. Giant-cell arteritis with peripheral neuropathy. Lancet 1968;i:1010-3. 3 National Institute of Health. Diagnostic criterions for Guillain-Barr6 meningitis caused by Haemophilus influenzae. J3AMA 1978;240: 1709-11. 4 Hazleman B, Goldstone A, Voak D. Association of polymyalgia rheumatica and giant-cell arteritis with HLA-B8. Br MedJ 1977;ii:989-91. 5Hughes RAC. In: Rose FC, ed. Clinical neuroimmunology. Oxford: Blackwell Scientific Publications, 1979. (Accepted 2 October 1979)
Regional Department of Neurology, Derbyshire Royal Infirmary, Derby DE1 2QY R N CORSTON, MB, MRCP, senior registrar
With the widespread use of low-dose oestrogen contraceptives such interactions may well occur more often. Tetracyclines are used for a wide range of infections including pelvic inflammatory disease and non-specific urethritis. One of us has seen a second case of breakthrough bleeding associated with a course of tetracycline. We therefore think that doctors should be aware of the potential problem. We would advise that women taking low-dose oral contraceptives should take extra precautions against conceiving in any cycle during which antibiotics are given. Stockley I. Interactions with oral contraceptives.
Pharmaceut3J 1976 ;216:
140-3.
Reimers D, Nocke-Finck L, Breuer H. Rifampicin and the "pill" do not go well together. JAMA 1974;227 :608. 3 Dossetor EJ. Drug interactions with oral contraceptives. Br MedJ3 1975; iv :467-8. 4 Buck DJ, Breckenridge AM. Drug interactions with oral contraceptives. IPPF Med Bull 1978;12:1-2. Pulkkinnen MO, Willman K. Reduced maternal plasma and urinary estriol during ampicillin treatment. Am J Obstet Gynecol 1971;109: 893-6. 2
(Accepted 19 October 1979) Family Planning Association of Western Australia, 233 Adelaide Terrace, Perth, Western Australia JANET F BACON, MB, medical officer
Pregnancy attributable to interaction between tetracycline and oral contraceptives Various drugs, including antibiotics, may cause failure of contraception and breakthrough bleeding in patients on oral contraceptives.' Early reports implicated rifampicin,' which was later shown to increase the rate of hepatic metabolism of ethinyloestradiol and norethisterone.' There have been three cases of pregnancy in women taking ampicillin.3 We cannot find any report of either breakthrough bleeding or pregnancy attributable to tetracycline in a woman taking oral contraceptives. We report such a case.
Case history The patient, a 20-year-old student, had been taking oral contraceptives for four years. Her only complaint had been of an increase in left-sided headaches. For the last two years she had taken Microgynon 30 (ethinyloestradiol 30 ,ug, D-norgestrol 150 ,ug) and had had no breakthrough bleeding. On 16 October 1978 she had a normal withdrawal bleed. On 20 October she started the next course of Microgynon and also (for sinusitis) a five-day course of tetracycline 500 mg six hourly for three days and 250 mg six hourly for two days. She completed the course on the 24 October. She did not forget to take her oral contraceptive. She had no diarrhoea or vomiting but had only a very light two-day withdrawal bleed on 14 and 15 November 1978. She continued to take her oral contraceptive for a further two months until seen at the family planning clinic on 8 January 1979. Her uterus was then the size of at least a 12-week pregnancy. This would have corresponded with a previous menstrual period on about the 15 October, and indicated that she must have ovulated either when she was taking the tetracycline or in the week after. Termination confirmed the clinical assessment of the duration of pregnancy.
Comment The time course of this history in a woman who had been taking oral contraceptives for four years with no problems strongly suggests that the tetracycline contributed to the failure in contraception. In addition to the reports of pregnancies associated with rifampicin' and ampicillin,' pregnancies have been reported in two women taking either sulphamethoxypyradizine or chloramphenicol4 and also an increased incidence of breakthrough bleeding in women on phenoxymethylpenicillin, neomycin, or nitrofurantoin.1 The mechanism of the interaction is unknown but a decrease in urinary oestriol output has been noted in pregnant women taking ampicillin,5 phenoxymethylpenicillin, or neomycin.1 Therefore the enterohepatic circulation of contraceptive steroids may be interrupted by inhibition of the gut bacteria normally responsible for hydrolysis of steroid conjugates.4 This would result in lower-than-normal concentrations of circulating steroids. There seems no reason why tetracycline should not cause a similar change.
University of Western Australia, Nedlands, Western Australia 6009 GILLIAN M SHENFIELD, DM, MRCP, senior lecturer in clinical pharmacology
Diet and ulcerative colitis The apparent rarity of ulcerative colitis in developing countries together with the ability of dietary fibre to affect colonic function and bacterial metabolism have led to suspicions that a low intake of fibre may be a causal factor in this disease.' 2 People who develop Crohn's disease (a closely related condition) show significant differences in diet from matched controls, eating less raw fruit and vegetable fibre and more refined sugar.3 To investigate whether ulcerative colitis is preceded by an unusual diet we have compared the diet before illness, of patients with newly diagnosed colitis with that of closely matched, healthy subjects.
Subjects, methods, and results Thirty consecutive newly diagnosed patients (17 men, 13 women) were studied. The diagnosis of ulcerative colitis was based on the findings of barium radiology, sigmoidoscopy, and rectal biopsy. In 10 patients the disease extended proximal to the splenic flexure, in five it was limited to the rectum, and in the remaining 15 it was of intermediate extent. The median duration of symptoms before diagnosis was two months (range 1-18). All patients were interviewed within three months of diagnosis. A single dietitian used the dietary history method4 to question each patient about his or her habitual diet before illness. Using a specially designed questionnaire, she determined the frequency of consumption and size of helpings of each of a comprehensive range of foods and drinks.3 The data were analysed to determine the average daily intake of different food components with the use of a computer programme compiled from McCance and Widdowson's food tables.5 Dietary intake of 30 patients with ulcerative colitis before their illness and that of 30 matched, healthy controls. Results given in glday as means i SE of mean unless stated otherwise Patients 199 ±11 .. Total dietary fibre . 8-6 Cereal fibre* .. 101 +0 6 Total fruit and vegetable fibre 1-4±0-2 Raw fruit and vegetable fibre . .. 96±9 Refined sugar . 291 ±20 . . Total carbohydrate 78 ±4 .. . Protein . .109±7 Fat . . 10-2±0-6 Energy (MJ/day)
*Median value. Conversion: SI to traditional units-Energy: 4-2 MJz 1000 kcal.
Controls 18-3 ±1-2 7-2
10-5 ±0-8 1-4±0-2 97±9 269 ± 16 82 ±3 106±5
10-0±0-5
294
The control subjects had recently attended the fracture clinic and were otherwise healthy. They were matched with the patients for age (± 5 years), sex, social class, and marital status and were not significantly different from the patients in height or weight distribution. The controls were interviewed about their diet in the same manner as the patients. The statistical significance of differences was calculated by Student's t test or by Wilcoxon's rank sum test for paired data. The table shows the dietary intake of the patients and controls. There was no significant difference in any food component between the two groups. In particular, there was no difference in the intake of dietary fibre or refined sugar.
BRITISH MEDICAL JOURNAL
2 FEBRUARY 1980
Tests for platelet agglutinins and incomplete complement-fixing platelet antibodies both with and without the addition of cimetidine detected no abnormalities. After 18 months the patient remained well with no recurrence of purpura. His platelet count is now 193 x 109/1 (193 000/,41), similar to that found on presentation (see figure). 250-
*
200
_ CClimetidine \
Comment
x~~~~~~~~~~~~e °1o50i 0~~~~~~~~~
There is apparently no difference between the diet before illness of patients with ulcerative colitis and that of closely matched controls. This contrasts with our findings in patients with Crohn's disease3 and suggests that, despite their similarities, these two diseases do not have an identical aetiology. Our data rule out the possibility that individuals develop ulcerative colitis because their intake of fibre is lower than the British average. Nevertheless, the average diet in this country differs in many respects from that of communities in which colitis is apparently rare. Hence, it remains possible that the average British diet favours the development of ulcerative colitis in susceptible individuals. Trowell H. Ulcerative colitis and Crohn's disease. In: Burkitt DP, Trowell HC, eds. Refined carbohydrate foods and disease. London: Academic Press, 1975:135-40. 2 Painter NS. Diverticular disease of the colon. London: Heinemann, 1975. 3 Thornton JR, Emmett PM, Heaton KW. Diet and Crohn's disease: characteristics of the pre-illness diet. Br MedJ 1979;ii:762-4. 4 Burke BS. The dietary history as a tool in research. Jf Am Diet Assoc 1947 ;23 :1041-6. McCance and Widdowson. The composition of foods. Paul AA, Southgate DAT, eds. London: HMSO, 1978.
(Accepted 20 November 1979)
University Department of Medicine, Bristol Royal Infirmary, Bristol BS2 8HW J R THORNTON, MB, MRCP, research fellow P M EMMETT, BSC, SRD, senior dietitian K W HEATON, MD, FRCP, reader in medicine
Cimetidine and thrombocytopenia Since the introduction of cimetidine there have been sporadic reports of associated blood dyscrasia. Uncomplicated thrombocytopenia has been recorded in only three patients,1 2 in whom factors other than cimetidine treatment may have been implicated. I report a case in which the relation of treatment with cimetidine to the development of thrombocytopenia is apparently clear cut and which may elucidate the mechanism by which thrombocytopenia is induced.
Case report A 43-year-old Tanzanian Asian man with sarcoidosis manifesting as arthritis and lymphadenopathy (treated only with indomethacin) was admitted with polyuria, polydipsia, and melaena. He had developed hepatosplenomegaly and had anaemia and hypercalcaemia (plasma calcium concentration 3-2 mmol/l; 12-8 mg/100 ml). His platelet count was 184 x 109/1 (184 000/,ul). A small, lesser-curve gastric ulcer was detected radiologically and endoscopically. Indomethacin was stopped, and he was given cimetidine 1 g daily and prednisolone 20 mg daily, with resolution of his arthritis, lymphadenopathy, and hypercalcaemia. Forty-nine days after beginning cimetidine, while taking prednisolone 10 mg daily, he was readmitted with a two-day history of sore throat and oral ulceration and bleeding. Examination disclosed haemorrhagic blisters on the tongue and buccal mucosa and petechiae in the pharynx, round the mouth, and on the lower legs. There was no lymphadenopathy, and his spleen was barely palpable. His platelet count was 6 x 109/l (6000/t1l), haemoglobin concentration 13-7 g/dl, and white cell count 6-6 x l09/l (6600//1; distribution 62 % polymorphs, 24 % lymphocytes, and 14 % monocytes). Cimetidine was stopped but prednisolone 10 mg daily continued. The figure shows the subsequent rise in the platelet count. Examination of the bone marrow five days after admission detected an increased proportion of megakaryocytes, consisting chiefly of mature forms.
. I*
I/
_
100-
50; March 1977
Nov Jan 17 18 19 20 21 22 23 24 25 April June 1978 March
Dec March Sep Feb 1979
Relationship between platelet count and cimetidine treatment.
Comment
The American Food and Drug Administration studied 26 cases of haematological abnormalities (usually granulocytopenia) in about 13 million patients treated with cimetidine. In many cases the drug was exonerated or the data were insufficient. The other cases were complicated by factors such as renal failure, septicaemia, and treatment with cytotoxic drugs. Complicating factors are also apparent in other reported cases of thrombocytopenia associated with cimetidine treatment. Two patients''2also had large-bowel cancer and a third2 was an alcoholic with bleeding varices who evidently had had thrombocytopenia before taking cimetidine; after portacaval shunting he developed respiratory failure and subsequently renal and hepatic failure. In the present case the appearance of severe thrombocytopenic purpura was temporally related to the administration of cimetidine and reversed when cimetidine was stopped without altering the dose of prednisolone. The association with sarcoidosis is particularly interesting because this is a recognised though rare cause of thrombocytopenia, which may prove fatal if not treated with-steroids or by splenectomy.4 Sarcoidosis alone is unlikely to have caused thrombocytopenia in the present case, since the platelet count was normal when the disease was most active and fell during remission. Cimetidine probably acted as a precipitant in a disease that predisposes the patient to autoimmune platelet destruction; the negative results of the antibody tests do not exclude this mechanism. A recent report of autoimmune haemolytic anaemia during cimetidine treatment5 is relevant because this may also be associated with sarcoidosis. Increased peripheral destruction of blood cells sensitised by cimetidine (including leucocytes in some reported cases of neutropenia5) rather than bone-marrow toxicity may explain some of the haematological disturbances, particularly in predisposed patients. In view of the widespread use of cimetidine, the fact that there have been few reported cases of blood dyscrasia (none of them fatal) in uncomplicated peptic ulcer is reassuring.
'McDaniel JL, Stein JJ. Thrombocytopenia with cimetidine therapy. 2 3
N EnglJ_ Med 1979;300:864. Idvall J. Cimetidine-associated thrombocytopenia. Lancet 1979;ii:159. Freston JW. Cimetidine and granulocytopenia. Ann Int Med 1979;90: 264-5.
4Dickerman JD, Holbrook PR, Zinkham WH. Etiology and therapy of thrombocytopenia associated with sarcoidosis.y Pediatr 1972;81 :758-64. 5 Rotoli B, Formisano S, Alfinito F. Autoimmune haemolytic anaemia associated with cimetidine. Lancet 1979;ii:583.
(Accepted) Department of Medicine, Westminster Hospital, London SWlP 2AP A J ISAACS, BM, MRCP, senior registrar
