The suggestion that the "requirement for pro- This can be answered satisfactorily only by con- 3 Jick H, Derby LE, Garcia Rodriguez LA, Jick SS, Dean AD. found anaesthesia will decline" certainly does trolled clinical trials and not by comparison with Non-steroidal anti-inflammatory drugs and certain rare, serious adverse events: a cohort study. Pharmacotherapy not apply to laparoscopic surgery-indeed, for contemporary series of patients, as the authors 1993;13:212-7. laparoscopic herniorraphy the opposite is true. have attempted. 4 Huskisson EC, Wojtulewski JA, Berry H, Scott J, Hart FD, Sedoanalgesia is not, in any case, inherently better, W H TURNER Balme HW. Treatment of rheumatoid arthritis with feno- safer, or preferred by patients, nor would its Department ofUrology, profen: comparison with aspirin. BMJ 1974;i: 176-80. widespread application allow operator-sedationists University ofBeme, Inselspital, to free themselves from the shackles of their former 3010 Beme, anaesthetists. Switzerland Long term use ofsumatriptan The most disturbing vignette of the future is, J C SMITH however, plausible unlike the foregoing. This is of Department ofUrology, EDITOR,-M J Osborne and colleagues report on a a large team of people who pass the patient among Churchill Hospital, Oxford OX3 7LJ patient who developed a pattern of daily headaches each other. The "director" has a brief interview and used excessive daily doses of sumatriptan.' and passes the patient on. The tedious task of They suggest that the long term use of sumatriptan preoperative and postoperative care is given to the I Kerbl K, Clayman RV, McDougall EM, Kavoussi LR. Minimally invasive surgery: laparoscopic nephrectomy. BMJ could have led to a dependent state. The patient otherwise redundant anaesthetist. The operators 1993;307:1488-9. (4 December.) had a 50 year history of incapacitating migraine in both open and minimal access surgery, freed attacks refractory to other treatments, but his from the tedium of actually seeing, diagnosing, attacks were rapidly and effectively treated with and caring for patients, practise their skills in the sumatriptan. theatre. The patients are no longer cared for by one Renal failure after topical use People who suffer from migraine report changes practitioner who (in theory at least) looks after ofNSAIDs in the frequency and nature of attacks over time them as a whole person. The new process might be and also experience other types of headache. The efficient but would be bad for patients. authors note that this patient experienced mild Named nurses now look after each patient. EDTOR,-We wish to amplify the case reports of C A O'Callaghan and colleagues' by reporting on a headaches every morning and that these frequently Allocation oftasks has ended; personal care is here. progressed to migraine. The patient took suma- Medicine neither wants nor needs to step away patient who developed acute renal failure twice: once after taking ibuprofen orally and once after triptan daily in anticipation of these attacks, in from personal care, it needs to step towards it. clear contrast to the recommendation on the data- Patients need named doctors as well as named topical administration ofthe drug. A 76 year old man was admitted with acute sheet. This does not suggest dependence on the nurses. effects ofsumatriptan. A C SKINNER anuric renal failure five days after taking four Warrington WA4 4AZ tablets of ibuprofen for a muscle strain. In the 48 Data have been published on long term experi- hours before admission he had vomited repeatedly, ence with sumatriptan.' In three studies lasting up 1 Wickham JEA. Minimally invasive surgery: future develop- and at presentation he was clinically dehydrated. to one year the tolerability profile of sumatriptan ments. BMJ 1994;308:193-6. (15 January.) The platelet count on admission was 35 x 109/1, and (the incidence and nature of reported adverse a blood film showed microangiopathic haemolytic events) was similar to that reported in short term anaemia without eosinophilia. Renal biopsy studies. There was no evidence of an escalation of Laparoscopic nephrectomy unproved showed evidence of acute interstitial nephritis and the dose, irrespective of the number of attacks of in controlled clinical trials acute tubular necrosis. After four days of haemo- migraine treated. Two of these studies (one of dialysis his renal function recovered and the serum subcutaneous and one of oral treatment) were ED1TOR,-An italic paragraph accompanying the creatinine concentration fell to 128 ,umol/l. extended up to two years. There was no evidence of series of articles on minimally invasive surgery Three years later the patient applied topical an increased frequency of migraine over the two states that the "articles have been written to inform ibuprofen once to his shoulders because of muscle years (table). Additionally, no evidence of depen- non-specialists of developments in this rapidly aches after he had cut a hedge. Thirty six hours dence has been noted from spontaneous post- moving subject." Two of the authors of the article later he was admitted with acute anuric renal marketing reports. on laparoscopic nephrectomy,' Ralph V Clayman failure. The platelet count fell transiently to and Louis R Kavoussi, are known by urologists to Median number of attacks of migraine per patient in 72x 109/l, and a blood film showed microangio- months 1, 2, 23, and 24 of two year studies of use of be innovators, and without such people urology pathic haemolysis with no eosinophilia. Methyl- subcutaneous or oral sumatriptan would not advance. Certain points should, prednisolone (0 5 g) was given intravenously on however, be made. admission. After 13 days, during which haemo- Subcutaneous Oral The summary of the article says that "laparo- dialysis was required, the patient's renal function scopic nephrectomy for benign disease has become Months 1, 2 6 7 recovered, and nine months after discharge his Months 23, 24 5 7 widely accepted."' This is not so, and the authors' serum creatinine concentration was 571 figures show this: the fact that only 30 laparoscopic p.moll. procedures have been performed in St Louis and Unlike O'Callaghan and colleagues, we Sumatriptan is indicated only for intermittent "more than 100 worldwide" shows that almost observed only limited renal recovery after a single short term treatment of migraine. When it is used all nephrectomies for benign disease are open topical application of a non-steroidal anti- appropriately there is no evidence of dependence operations. inflammatory drug. Severe idiosyncratic renal developing during long term treatment. Suma- It is appropriate that the authors comment on syndromes are recognised after administration triptan should not be used daily as prophylaxis the learning curve for laparoscopic nephrectomy, of non-steroidal anti-inflammatory drugs, and against migraine. but to imply that three major complications in the mechanisms other than interference with vaso- D K LLOYD first 12 patients and then one in the second 12 dilatation mediated by prostaglandin have been W S G ARNOLD represents an improvement due to experience implicated.2 In this case sensitivity to oral Product Strategy Directorate, is incorrect. There is no significant difference be- Glaxo Research and Development, ibuprofen had been shown previously. Haemolysis Uxbridge, tween these complication rates (x2=1-2, P=0-27). and thrombocytopenia, which were prominent Middlesex UB1 1 1BT A rate ofmajor complications of 16% in the first 24 features after each to have patients in the hands of presumably outstandingly exposure ibuprofen, been reported after oral administration of non- 1 Osborne MJ, Austin RCT, Dawson KJ, Lange L. Is there a good operators suggests that for most urologists steroidal anti-inflammatory drugs.34 problem with long term use of sumatriptan in acute migraine? the complication rate for their first 24 patients will Abrupt renal impairment may occur after BMJ 1993;308:113. (8 January.) be even higher; this may make many urologists 2 Tansey MJB, Pilgrim AJ, Martin PM. Long-term experience modest exposure to oral or topical non-steroidal with sumatriptan in the treatment of migraine. Eur Neurol sceptical about using the technique. It also sug- anti-inflammatory drugs, and patients with a 1993;33:310-5. gests that a period of formal training should be history of this should be warned to avoid these required for urologists of any grade who wish to drugs irrespective of the route of administration or use this technique to avoid subjecting excessive dose. numbers ofpatients to the same learning curve. Mesalazine toxicity The authors' final paragraph contains entirely A H N FERNANDO S THOMAS reasonable comments. The authors should, how- RM TEMPLE H A LEE EDITOR,-The case reported by A G Lim and K R Wessex Renal and Transplant Unit, ever, follow their own advice, with "careful critical St Mary's Hospital, Hine, in which a patient developed a reaction to comparison of each newly developed procedure Portsmouth P03 6AD mesalazine,' prompts me to describe a case. with its counterpart in open surgery." It is A 30 year old woman presented with a five week intuitively obvious that an uncomplicated laparo- 1 O'Callaghan CA, Andrews PA, Ogg CS. Renal disease and use of history of bilateral pleuritic chest pain. She had scopic nephrectomy will result in a shorter and less topical non-steroidal anti-inflammatory drugs. BMJ increasing shortness of breath (though no wheez- convalescence than an 1994;308:110-1. (8 January.) ing), a non-productive cough, and intermittent painful open operation. 2 Clive DM, Stoff JS. Renal syndromes associated with non- Whether the complications of the two forms of steroidal anti-inflammatory drugs. N Engl J Med fevers and had lost 7 kg in weight. She had a 14 year surgery favour laparoscopy remains to be seen. 1984;310:563-72. history of ulcerative colitis, which had remained

BMJ VOLUME 308 19 FEBRUARY 1994 533 quiescent for 18 months. Her drug treatment and mice. This feature is characteristic of human If patients with non-insulin dependent diabetes consisted of oral mesalazine 800 mg twice daily, carcinogens. In fact, tamoxifen is associated with are divided into insulin requiring and non-insulin which she had been taking for the past eight increases in cancers of the endometrium34 and requiring on the basis of a glucagon stimulated months. She had previously taken sulphasalazine possibly liver3 in treated patients. C peptide concentration <0-6 nmol/1, however, for several years but had stopped this after develop- In answer to the question "Is tamoxifen safe?" only the insulin requiring patients show an ing arthralgia. Jordan compares the risk of tamoxifen with that increased frequency of HLA-DR4 (54% (37/69) On examination there was decreased air entry at of use of oral contraceptives on the basis of compared with 32% (41/127) in patients whose both lung bases and no other abnormality. Investi- his conclusion that it is the oestrogenic proper- condition was controlled with oral antidiabetic gations showed a white cell count of 11 3x 109/l ties of tamoxifen that result in the assumed, agents).4 Patients with DR4 had lower C peptide with 16% eosinophils and a high plasma viscosity; but not proved, ability of tamoxifen to promote concentrations than patients with other HLA-DR she was negative for rheumatoid factor, weakly hepatic tumours in rats. This conclusion fails antigens.2 positive for antinuclear factor, and strongly posi- to take into account other aspects of the toxic- Since high insulin and C peptide concentrations tive for antibody to neutrophil cytoplasm (titre ology of tamoxifen, as noted above, which are are associated with an increased risk of cardio- 1/400). The pattern of staining of the antibody to different from the effects of oestrogens. The vascular disease we propose that these patients neutrophil cytoplasm was cytoplasmic, and the weakness of the comparison with oestrogens is represented a subgroup of patients with non- neutrophils were negative for antibodies to myelo- further illustrated by the fact that toremifene, insulin dependent diabetes protected against peroxidase, lactoferrin, and cathepsin G. Initial which is related to tamoxifen and has comparable cardiovascular disease. Therefore, the increase in urine testing showed microscopic haematuria, an oestrogenic properties in liver, is not hepato- frequency of HLA-DR4 is due to an admixture of isotope lung scan did not show any abnormality, carcinogenic.2 patients with genes conferring susceptibility to and an electrocardiogram was normal. She was It is important to ascertain whether breast insulin dependent diabetes rather than a general unable to perform lung function tests because of cancer can be prevented without the potential feature of patients with non-insulin dependent coughing. A chest x ray film showed bilateral lung risks of tamoxifen being imposed. Considerable diabetes. In the search for genes for non-insulin infiltrates, especially at the apexes, and bilateral evidence supports the view that breast cancer can dependent diabetes correct definition of the pheno- small pleural effusions. be substantially prevented by reduction of dietary type is critical, since the disease is likely to be In view of the picture of pulmonary eosinophilia, consumption of fat,5 which entails no risk and heterogenous. haematuria, and strongly positive titre of anti- affords other benefits. If chemical intervention is LEIF GROOP bodies to neutrophil cytoplasm a thoracoscopic deemed necessary the antioestrogen toremifene Department ofEndocrinology, University ofLund, lung biopsy specimen was taken from the right could be used. Malmo General Hospital, upper lobe. It showed no evidence of Wegener's GARY M WILLIAMS S-214 01 Malmo, Sweden granulomatosis or pneumonia due to bronchiolitis American Health Foundation, GIAN FRANCO BOTTrAZZO obliterans: appearances were those of a chronic Valhalla, Department ofImmunology, NY 10595, USA London Hospital Medical College, eosinophilic pneumonia. Mesalazine was stopped, El and the patient refused steroid treatment. Her London 2AD condition improved over the next few weeks, and 1 Jordan VC. How safe is tamoxifen? BMJ 1993;307:1371-2. (27 November.) 1 Tuomilehto-Wolf E, Tuomilehto J, Hitman GA, Nissinen A, the abnormalities in the chest x ray film and blood 2 Hard GC, Iatropoulos MJ, Jordan K, Radi L, Kaltenberg 0, Stengard J, Pekkanen J, et al. Genetic susceptibility to non- eosinophilia resolved. The titre of antibody to Imondi AR, et al. Major difference in the hepatocarcino- insulin-dependent diabetes mellitus and glucose intolerance neutrophil cytoplasm was still 1/400 three weeks genicity and DNA adduct forming ability between toremifene are located in HLA region. BMJ 1993;307:155-9. (17 July.) after she stopped taking mesalazine but had fallen and tamoxifen in female crl:CD(BR) rats. Cancer Res 2 Groop L, Groop P-H, Koskimies S. Relationship between B-cell 1993;53:4534-41. function and HLA antigens in patients with type 2 (non- to 1/25 a few months later. 3 Fomander T, Rutqvist LE, Cedermark B, Glas V, Mattsson A, insulin-dependent) diabetes mellitus. Diabetologia 1986;29: This patient's illness was probably caused by Silfversward C, et al. Adjuvant tamoxifen in early breast 757-60. mesalazine. Pulmonary side effects from this drug cancer: occurrence of new primary cancers. Lancet 1989;i: 1 17- 3 Rich SS, French LR, Sprafka JM, Clements JP, Goetz FC. HLA- 20. associated susceptibility to type 2 (non-insulin-dependent) are rare and may occur soon after the drug is 4 De Muylder X, Neven P. Tamoxifen and potential adverse diabetes mellitus: the Wadena City health study. Diabetologia started, as in the case described by Lim and Hine, effects. CancerJournal 1993;6:1 11-5. 1993;36:234-8. or after many months of exposure,2 as in this case. 5 Cohen LA, Rose DP, Wynder EL. A rationale for dietary 4 Groop L, Miettinen A, Groop P-H, Meri S, Koskimies S, One report mentions a strongly positive titre of intervention in postmenopausal breast cancer patients: an Bottazzo GF. Organ-specific autoimmunity and HLA-DR update. Nutrtion and Cancer 1993;19:1-10. antigens as markers for 3-cell destruction in patients with type antinuclear anbitody, which fell after mesalazine II diabetes. Diabetes 1988;37:99-103. was stopped.3 Ulcerative colitis may cause a posi- tive result in a test for antibodies to neutrophil cytoplasm, but the staining is usually perinuclear.4 Genetic susceptibility to non- The high titre of antibody to neutrophil cytoplasm insulin dependent diabetes Breast feeding and diabetes with a cytoplasmic staining pattern, which fell mellitus when mesalazine was stopped, raises the possi- EDITOR,-Eva Tuomilehto-Wolf and coworkers bility ofa link with the drug. have presented the interesting hypothesis that EDITOR,-K G M M Alberti briefly mentions that D HONEYBOURNE Department ofThoracic Medicine, genetic susceptibility to non-insulin dependent "breast feeding has been shown to protect against Dudley Road Hospital, diabetes mellitus is located in the HLA region. the development of insulin dependent diabetes Birmingham B18 7QH Their conclusions are based on findings in 157 mellitus."' Two papers have strongly suggested elderly Finnish men (age 70-89), in whom HIA that infants born into families with a family history 1 rim AG, Hine KR. Fever vasculitic rash, arthritis, pericarditis, haplotypes associated with insulin dependent of diabetes have a lesser chance of developing and pericardial effusion after mesalazine. BM3' 1994;308:113. (8 January.) diabetes could explain 98% of non-insulin depend- insulin dependent diabetes if they are breast fed up 2 Reinoso MA, Schroeder KW, Pisani RJ. Lung disease associated ent diabetes and 79% ofimpaired glucose tolerance. to the age of 9-12 months.23 This breast feeding with orally administered mesalazine for ulcerative colitis. Chest This association depended strongly on the pre- must be exclusive of top up formula milk feeds as 1992;lOl:1469-71. sence of HLA-DR4, which was observed in 57% these initiate the autoimmune process that may 3 Welte T, Hamm H, Fabel H. Mesalazine alveolitis. Lancet 1991;338:1273. (56/98) of patients with non-insulin dependent result in insulin dependent diabetes in later life. 4 Cambridge G, Rampton DS, Stevens TRJ, McCarthy DA, diabetes but only 13% (3/23) of controls (P= This message seems not to be getting across to Kamm M, Leaker B. Anti-neutrophil antibodies in inflamma- 0 003). The frequency of HLA-DR4 in the diabetic people in the community. Doctors, health tory bowel disease: prevalence and diagnostic role. Gat patients with non-insulin dependent diabetes is in visitors, and midwives must try to make breast 1992;33:668-74. the same range as that previously reported from feeding the norm in diabetic families. Finland2 and the United States3 (table). The Alberti mentions a possible link between con- frequency in the control population is considerably sumption of bovine serum albumin and the Safety oftamoxifen lower than that previously reported, but this is development of insulin dependent diabetes. most probably due to the small number of control Karialainen et al found raised antibodies to bovine EDrrOR,-V Craig Jordan reports some of the subjects studied. serum albumin in most newly diagnosed insulin preclinical toxicological findings on tamoxifen in dependent diabetic patients.4 More importantly, his editorial.' Though he states that tamoxifen Prevalence of HLA-DR4 in patients with non-insulin Dahl-Jorgensen et al showed a close correlation "promotes hepatic tumours in rats," he fails to dependent diabetes. Figures are numbers (percentages) of between the amounts of cows' milk consumed per note that tamoxifen is a strong liver carcinogen by patients head of the population in various countries and the itself in rats, producing tumours within six months incidence of insulin dependent diabetes.5 This and a high incidence by one year at doses that yield Non-insulin leaves little doubt that consumption of cows' milk to dependent blood concentrations comparable those in Study Controls diabetes P value is a trigger for diabetes mellitus. Bovine serum treated women.2 In addition to causing the albumin is 97% denatured by ultraheat treatment formation of DNA adducts in the liver of rats, Groop etaP 86/322(28) 51/121(51) 0-02 of milk. We are assessing data to see whether which Jordan mentions, tamoxifen causes the Rich et aP 62/221(27) 36/86 (42) 0-02 ultraheat treated milk is less diabetogenic than formation of DNA adducts in the liver of hamsters Tuomilehto-WolfetaP 3/23 (13) 56/98 (57) 0 03 pasteurised milk.

534 BMJ VOLUME 308 19 FEBRUARY 1994