328 antibody titre concentrations in the workforce were no different from those found in the general population. To study the ecology of Legionella i...

341KB Sizes 0 Downloads 6 Views


antibody titre concentrations in the workforce were no different from those found in the general population. To study the ecology of Legionella in power stations the CEGB sampled water from 17 power stations every month through the summer of 1983. Four of these stations were sampled more intensively in 1984. A total of 335 samples were taken from various parts of the cooling water system; 205 of these were taken from the condenser outlet (the warmest part of the system) and the cooling tower ponds. Legionella was isolated from eleven of the power stations sampled, the organism being present in low concentrations in 35% of the 205 samples of the circulating water. Six of the fourteen make-up sources of water tested were positive. In all but 3 of the 205 samples the numbers of Legionella were very low. 49 of the positive samples contained fewer Legionella than found in the make-up water with the highest number, and most were at the limits of detection. The 3 samples with high numbers were obtained on isolated occasions from two stations and the numbers were still too low for the water to be regarded as possible sources of infection. After the outbreak of legionnaires’ disease in Stafford in April/May, 1985, suggestions were made in the press and on television that one or more of the CEGB’s cooling towers could have been a source of the infection. The CEGB initiated a sampling programme of power stations countrywide, including some power stations which operate only rarely. Thirty-three power stations, all operating on inland rivers and using cooling towers were studied. Legionella was found in one river sample and in nine of the thirtythree stations sampled, 20% of the samples were positive; none of the samples from the other twenty-four stations was positive. On all occasions the concentrations were low and rarely above background levels, which confirm the findings in the 1983/84 programme. The results of these surveys, which will be published in full elsewhere, generally showed only an intermittent presence of Legionella in low numbers in power station waters. The numbers of organisms found cannot be considered a health risk and there is no requirement to modify current power station operational procedures. Further work is in hand to follow the seasonal patterns of bacterial growth and to identify possible sites of multiplication of Legionella within a power station and to ensure that the ecology of the organism in the cooling water system is fully understood. Additionally, a national surveillance programme will be established to monitor continuously the situation countrywide. Medical Branch, Central Electricity

Generating Board,

London EC4P 4EB


DOES ISOTRETINOIN CAUSE LIMB REDUCTION DEFECTS? p 1276) describes a child with limb reduction defects whose mother took isotretinoin (’Accutane’) during the pregnancy. Several important aspects of this case are missing from McBride’s letter. McBride examined the child in the capacityof a professional witness in litigation proceedings. When he saw the patient, corrective surgery had already been done on both hands, so his assessment of the affected hands was largely based on photographs. One of us (E. J. L.) examined the child before surgery. Both hands had complete syndactyly of all four fingers with absent middle and distal phalanges. McBride incorrectly stated that there were no fingers on the nght. The syndactylous left fingertips had a ring-like circumferential constriction with four terminal bulbous tips grouped in a grape-like cluster distal to the constriction. The appearance was very suggestive of a circumferential constriction and amputation from an amniotic band. McBride fails to mention that four expert witnesses concluded that the child’s limb reductions were probably caused by amniotic bands. This conclusion was based on the following points: (1) one of the child’s physicians reported that, during the newborn period, he had removed a band of tissue that was attached to one of the malformed limbs; (2) several of the malformed limbs had circumferential constrictions without complete amputation; and (3) the child’s pattern of limb reductions did not resemble that

SiR,-Dr McBride (June 1,

produced when retinoic acid was given to laboratory animals at the comparable time in gestation. The child’s limb reductions were all distal with normal proximal bones; in animals, retinoic acid usually causes more proximal limb deficiences when given before or during the initial appearance of the upper limb bud. McBride speculated that isotretinoin may have caused limb reductions by damaging the apical ectodermal ridge of the developing limb bud. This is inconsistent with a large volume of experimental data showing that1 retinoic acid does not affect the cells of the apical ectodermal ridge. McBride incorrectly stated that the child’s mother had taken isotretinoin for 14 days. She reported that she had taken it for only 7

days. Isotretinoin is a human teratogen. It may cause limb reductions in human fetuses, but we do not think that isotretinoin caused the malformations in this case. The evidence that this child’s limb reductions resulted from amniotic bands was very convincing, yet none of this information was provided in McBride’s letter. McBride’s report is misleading: he fails to provide the reader with the’information necessary to judge whether a relationship was possibly causal. The following advice from the thalidomide era is still valuable, "Let us be on the alert by all means and report our observations, but we must be critical, because many pregnant women and their husbands are unnecessarily upset by uncritical accusations of as being the cause of all congenual


Embryology-Teratology Unit, Massachusetts General Hospital, Boston, 02114 Massachusetts, USA Medical

College of Georgia,


University of Michigan Ann Arbor




1 Kochnar DM. Cellular basis of congenital limb deformity induced in mice A. Birth Defects 1977; 13(1). 111-54. 2. McBride WG Drugs and congenital abnormalities. Lancet 1962, ii: 1332

by vitamin


SIR,—I wish to report a local rather than a systemic’ complication of self-poisoning with theophylline. A 31-year-old female who believed she was beginning to have one of her occasional bronchial asthma attacks as she was preparing for bed took a slow-release theophylline tablet (’Phyllocontin 350’) while recumbent, without any accompanying fluid. She slept well that night. Since she had been entirely well before this episode, she had not been taking any other drugs. When she awoke the next morning she noticed a retrosternal burning pain, which became excruciating when she tried to swallow breakfast juice. After about 8 h she crushed one aspirin tablet in water, but because of the pain was able to swallow only about half the mixture. She was unable to take anything (including water and patent antacids) orally that day because of pain and vomiting. Her pain was felt behind the sternum in the right breast and in the back on the right side. She slept poorly. The next morning she consulted her family physician, and a barium meal revealed no abnormality. The pain continued to be exacerbated by swallowing but was partly controlled by a great deal of anaesthetic gel. Within a few days she noticed that she was losing weight and becoming dry. When I saw her 9 days after the pain began, she was looking dehydrated and ill. At upper gastrointestinal endoscopy with a flexible fibreoptic endoscope the next dayI found, at 34 cm from the incisor teeth, a 2 cm long oesophageal ulcer which spared only the right lateral side of the oesophagus. The ulcer was deep, nodular at the edges, and had some shaggy areas in its base. Distal to the ulcer the oesophagus was completely normal, as were the stomach and duodenum. Unusually, the patient experienced extreme pain when biopsy and cytology (brush) samples were being taken. The pain subsided slowly over the next week, during which the patient took large amounts of liquid antacids. Histology showed acute nonspecific ulceration with granulation tissue. Re-endoscopy 20 days later, when the patient no longer had symptoms, showed normal



329 The unusually sited oesophageal ulcer was presumed to be due to theophylline since there is no evidence of any other cause. Drugs known to cause oesophageal erosion and ulceration2include emepronium bromide, slow-release potassium preparations, tetracycline, doxycycline, and clindamycin. I know of only one other report of oesophageal ulceration due to slow-release theophylline.3 The patient did not notice sticking of tablets in the oesophagus. Another unusual feature was the extreme pain experienced during biopsy and collection of the cytology specimen. In normal supine individuals, pills and capsules may remain in the oesophagus for 45 min and, if the oesophagus is abnormal, for 90 min.4 This case-report highlights some simple precautions which can be taken. Patients should be advised to take all oral medication while erect or sitting up and with enough fluid (about 200 ml). Adequate fluid with medication is particularly important for patients who for some reason have to remain recumbent. I would also repeat Evans’ advice4 that tablets should be taken either before or during meals and not afterwards, as often advised. Division of General

Surgery, Department of Surgery, University of British Columbia,


Vancouver, Canada 1. Editorial.


Self-poisoning with theophylline. Lancet 1985; i: 146-47. Collins FJ, Mathews HR, Baker SE, et al. Drug induced oesophageal injury. Br Med J

1979; i: 1673-76. 3. Enzenauer RW, Bass 4.

JW, McDonnell JT. Esophageal ulceration associated with oral theophylline. N Engl J Med 1984; 310: 261. Evans KT, Roberts GM. Where do all the tablets go? Lancet 1976; ii: 1237-39.


SIR,-In commenting on efforts to improve ampoule labels to reduce errors in drug administration the secretary of the Medical Defence Union has expressed concern about cases where a doctor or nurse requiring a solution of sodium chloride to dilute a drug picks out an ampoule of similar design which contains potassium chloride, with fatal results. A similar tragedy in this areal has focused our minds on possible ways of prevention. To our knowledge potassium chloride is never meant to be injected directly in the concentration supplied in the ampoule or vial, so one possible solution would be to supply potassium chloride only diluted ready for intravenous infusion in a plastic bag or bottle similar to the way lignocaine is supplied for control of arrhythmias. Alternatively, it could be supplied in powder form in a squeezable plastic container with an injection spike to puncture the container of intravenous fluid so that the potassium chloride could readily be dissolved and transferred directly into the container of intravenous fluid before infusion. This method is already used to prepare dilute solutions of both lignocaine (lidocaine) and suxamethonium (succinylcholine). Either method would greatly reduce the hazard to patients by removing the necessity of ever aspirating concentrated potassium chloride into a syringe, a method which has led to many fatal accidents. Once potassium chloride is available packaged in either of these two methods, it will require either an industry-wide agreement or government edict to remove concentrated potassium chloride in the present containers from ward and theatre shelves. An alternative would be a national (and, later, international) standard. Pettis Memorial Veterans Hospital, Loma Linda, California 92357, USA and Department of Anesthesiology, Loma Linda University School of Medicine 1 San Bernardino Sun

June 14,







uroprotective agent


has been


used, hypersensitivity reactions have not been recorded, as far as we know. We report the

case of a severe hypersensitivity reaction to (’Uromitexan’). A 25-year-old woman with Hodgkin’s disease (stage IIa) had a rash on day 9 of the first cycle of combination chemotherapy (C-MOPP). mesna

Since procarbazine was suspected to be the most probable allergenic agent, it was replaced by methotrexate. 2 weeks later, early in the second cycle, the rash recurred with fever, nausea, and vomiting. After the rash had cleared challenge tests were done with the drugs of C-MOPP. Procarbazine (150 mg orally), cyclophosphamide (1110 mg orally), vincristine (2 mg intravenously), methotrexate (60 mg intravenously), and supporting drugs (dimenhydrinate, metoclopramide) yielded negative results. Then we applied uromitexan (10% mesna and 0-025% edetic acid). When we gave 50 mg mesna intravenously no reaction was seen. But after the administration of 150 mg 4 h later an extensive maculopapular rash appeared on the face and trunk together with conjunctivitis and oedematous periorbital swelling and chills, nausea, and pain in the extremities. 13 h later the patient had recovered. After 2 more days an oral challenge with 250 mg mesna was negative, but 2000 mg 4 h later again resulted in an extensive rash after 5 h. The results of intracutaneous skin tests with uromitexan and its components were: Substance Uromitexan (pure) Uromitexan 1:10 Uromitexan 1:100 Mesna 10% Mesna 1 % Mesna 0’ 1%

Result ++ -


+ + -


Substance EDTA1% EDTAO-1% Histamine 0 - 0107o

Result -


+ ++

(positive control) Buffer liquid


(negative control)

A control group of ten persons intracutaneously exposed to uromitexan 10007o showed a positive response in each person. This indicates an unspecific effect in the tests with pure uromitexan. Although the signs and symptoms were identical to those of IgE-mediated anaphylactic reactions, the negative cutaneous tests with the single dilutions (1:10, 1:100) of uromitexan and mesna do not favour an immunological mechanism. A pseudoallergic reaction similar to that caused by aspirin is supported by the dose dependence of the oral challenge test. The widespread use of mesna and the rarely reported side-effects seem to contradict this interpretation. But skin reactions during cytostatic




continued with



explored, cytostatic therapy being

different regimen.

Department of Dermatology, University Hospital, D-4300 Essen, West Germany



SiR.—Dr Zaman and colleagues (June 15, p 1357) report a prospective study of 590 cirrhotic patients enrolled between March, 1978, and April, 1983. By October, 1983, 229 patients had died and 42 of those deaths were attributable to hepatocellular carcinoma (HCC). Zaman et al compare the frequency of various attributes (age, sex, hepatitis B virus [HBV] markers, nationality, and duration and aetiology of cirrhosis) between the 42 patients with HCC and the remaining 548 cirrhotic patients without HCC, and concluded that "seropositivity for hepatitis B surface antigen and the duration and aetiology of cirrhosis were not related to the development of HCC". The cirrhotic patients were enrolled over 5 years and were at various stages of the disease on entry to the study. Thus the duration of the follow-up period is a potential confounding factor which must be adjusted for when comparing the duration and aetiology of cirrhosis between patients with HCC and those without. Table IV shows a relative risk (RR) of 2’ 53 for HBsAg positivity after adjusting for nationality, age, and sex. When only UK patients were considered, the RR for HBsAg positivity was 3-71 after adjusting for age and sex. Zaman et al conclude that HBsAg positivity is not an independent risk factor for HCC, presumably because these adjusted RRs were not significant. Several factors influence the p value of a statistical association; besides the strength of the association, there is the prevalence of the factor in the study population and the sample size. Table IV shows that among UK patients, the RR associated with being a male was 2 - 64 and the RR associated with HBsAg positivity was 3-71. It is not surprising to