1583
CLINICAL PRACTICE Intermediate-term results of heart-lung transplantation for cystic fibrosis
Between September, 1984, and March, 1991, 79 patients underwent heart-lung transplantation for end-stage cystic fibrosis at the Harefield Hospital. Short-term outcome has already been reported, and we now present intermediate-term results. The overall actuarial patient survival was 69% at 1 year, 52% at 2 years, and 49% at 3 years. 17 patients had diabetes mellitus with a survival of 62% to 1 year and 51% to 2 years. 23 patients had one or more other possible high-risk factors, and survival of these patients was 64% at 1 year and 57% at 2 years, compared with 71% and 49%, respectively, in the low-risk group (n=56). Pseudomonas aeruginosa infection was the most common respiratory infection encountered postoperatively. 92% of patients had at least one episode of acute rejection during the first 3 postoperative months. Lung function was greatly improved after transplantation, the mean forced expiratory volume in 1 s and forced vital capacity increasing from 22% and 35% predicted, respectively, preoperatively to 68% and 70% predicted, respectively, by the sixth postoperative month. This improvement was maintained at 1, 2, and 3 years after transplantation. Lymphoproliferative disorders (4 patients) were successfully treated. Obliterative bronchiolitis developed in 17 patients and the cumulative probability of getting this complication at 1, 2, and 3 years postoperatively was 17%, 23%, and 48%, respectively. Overall, 7 patients were retransplanted. There was no coronary artery disease in the 37 patients who underwent coronary angiography at 1 year, 14 at 2 years, and 9 at 3 years after surgery. 58 patients donated their hearts for subsequent "domino" heart transplantation. Our 5½-year experience with heart-lung transplantation is encouraging but the shortage of donor organs and the complication of obliterative bronchiolitis are the two main obstacles to be overcome.
Nowadays, many are surviving into adult life,! there 2000 CF adults over the age of 15 years in the about being UK.3 The estimated survival to 20 years is broadly similar in the UK (50%) Denmark (52%),5 the USA (43%),6 and Canada (55%).6 The gene responsible for CF has been identified,7 and since the disease seems to be limited to the cell membrane of the affected patient it should not affect the lungs of a transplanted patient. Against this background of a disease with a high mortality in adolescent and early adult life, the possibility of lung transplantation has become a of life.
promising therapeutic option. Initially there was much concern about operating on patients with CF because of the multisystem nature of the disease, together with chronic infection in the upper and lower respiratory tract. However, in October, 1985, the first successful transplants for CF were done at Harefield and Papworth Hospitals in the UK.8.9 The early results of selected series have been reported. Here we report our 5 2 ’-year experience with 79 CF patients who have undergone heart-lung transplantation, so that we can clarify the intermediate-term outcome of the procedure for CF and identify the high-risk factors. Patients and methods CF patients Between September, 1984, and March, 1991, 79 patients (39 male, 40 female; mean age 23 years, range 8-43) with CF underwent heart-lung transplantation in our unit. 3 patients underwent combined heart-lung and liver transplantation. The indications for surgery were severe respiratory failure with severely impaired quality of life. All patients were keen to have a transplant. At the time of acceptance for surgery, the mean forced expiratory volume in 1 s (FEV 1) was 22% predicted and the mean forced vital capacity (FVC) was 35 % predicted (n 78). When breathing air, mean Pa02 was 6.57 kPa (range 3-6-9-6) and mean PaC02 was 6-79 kPa (range 3-6-10-8, n=50). Preoperatively, sputum samples yielded Pseudomonas aeruginosa (70 patients), P cepacia (3), P maltophilia (1), Staphylococcus aureus (17), methicillin-resistant Staph aureus (4), Haemophilus influenzae (4), Streptococcus pneumoniae (2), and Aspergillusfumigatus (1); no pathogens were isolated from 1 patient. Many patients had more than one organism in their sputum. Contraindications to heart-lung transplantation were active aspergillus or mycobacterial infection, non-compliance with treatment, prednisolone therapy (> 10 mg/day), or other end-stage organ failure (unless this was also treatable at the time of transplantation-for example, by heart-lung and liver transplantation in patients with coexisting severe liver disease). =
Introduction
Cystic fibrosis (CF) is a multisystem disease, the main features of which are chronic bronchopulmonary infection, malabsorption, and high sweat-sodium concentration. There is a high incidence of diabetes mellitus, and most adults are colonised by Pseudomonas aeruginosa in the upper and lower respiratory tract.1,2 When the disorder was first described in 1938,80% of patients died within the first year
ADDRESSES: Harefield Hospital, Harefield, Middlesex, UK (B. P. Madden, MD, R. Radley-Smith, FRCP, A. Khaghani, FRCS, Prof M. Y. Yacoub, Kt), Royal Brompton Hospital, Sydney Street, London, UK (B. P Madden, M. E. Hodson, FRCP, V. Tsang, FRCS, M Y. Yacoub), and National Heart and Lung Institute, Dovehouse Street, London, UK (B. P. Madden, M E Hodson, V. Tsang, A. Khaghani, M. Y. Yacoub). Correspondence to Sir Magdi Yacoub, Harefield Hospital, Uxbridge, Middlesex UB9 6JH, UK.
1584
17 patients had diabetes mellitus and 23 had one or more other possible high-risk factors; these included pre-operative ventilation in 12 patients (6 conventional ventilation, 5 nasal ventilation, 1 ventilation with conventional extracorporeal membrane oxygenation), previous thoracic surgery in 10 (5 abrasion pleurodesis [1 bilateral], 1 double-lung transplantation, 3 thoracotomy [bilateral],1 pleurectomy), chemical pleurodesis in 2 (1bilateral), and combined heart-lung-liver transplantation for liver failure in 3. Some patients had more than one high-risk factor. 9 patients had a feeding gastrostomy tube inserted before surgery to improve nutrition.
Donor selection and organ procurement
’
The general criteria for donor selection and organ procurement have been described previously.lO Briefly, a portable cardiopulmonary bypass machine was taken by the retrieval team to the donor hospital. The donor was cooled on bypass to a core temperature of 8-10°C, the heart was preserved with cold blood potassium cardioplegia solution, and the organs were transported in cold donor blood and packed in ice.
Operative technique The surgical methods for heart-lung transplantation used at Harefield are described elsewhere.1O Matching criteria were based on blood group, size of thoracic cage, and cytomegalovirus (CMV) status. 58 of the 79 CF patients donated their hearts to patients who required cardiac transplantation with a primed heart (with a well-developed right ventricle) for moderate elevation in pulmonary vascular resistance. This "domino" procedures involves removal of the CF recipient’s heart intact for subsequent transplantation. The heart-lung block is inserted into the CF patient with separate superior and inferior vena caval anastamoses instead of a conventional right atrial anastamosis.
Immunosuppression Cyclosporin 6-10 mg/kg and oral azathioprine 2 mg/kg were given preoperatively. 1 g methylprednisolone was given intravenously during surgery after release of the aortic cross clamp. Routine immunosuppression postoperatively was achieved with cyclosporin, azathioprine, and rabbit anti-thymocyte globulin (RATG). Cyclosporin was given intravenously in the immediate postoperative period and thereafter orally with the aim to attain levels of 500 ng/ml (whole blood monoclonal antibody assay) during the first month postoperatively and thereafter between 250 and 350 ng/ml. Renal function was carefully monitored; if serum creatinine increased above 200 unol/1, cyclosporin was stopped and patients were given prednisolone until renal function improved and cyclosporin could be recommenced. Azathioprine was administered at a dose of 2 mg/kg per day provided the white blood cell count remained above 4 x 109/1. RATG (100 mg intravenously) was given on alternative days for the first 10 postoperative days with the aim to maintain a T-cell count of less than 200/d.
Postoperative care Episodes of acute rejection were diagnosed according to clinical and radiological indices with respiratory function tests and analysis of transbronchial lung biopsy specimens obtained at fibreoptic bronchoscopy. Patients were treated with intravenous methylprednisolone 1 g/day on 3 consecutive days, occasionally supplemented with RATG. Fibreoptic bronchoscopy was done routinely at the end of the first week and subsequently only if clinically indicated (ie, reduction in lung function or in exercise capacity, abnormality on chest radiograph, pyrexia of unknown origin, or unexplained cough). At bronchoscopy, the anastomosis was inspected and bronchoalveolar lavage specimens were taken for culture and sensitivity, opportunistic pathogen screen, and immunocytochemistry. Transbronchial lung biopsy specimens were sent for histopathological examination and culture. Episodes of infection were diagnosed by means of clinical, serological, and radiological findings together with studies of bronchoalveolar lavage and
Fig 1-Cause of death after heart-lung transplantation for cystic fibrosis. Redo= retransplantation, OB = obliterative bronchiolitis. *After retransplantation elsewhere.
transbronchial lung biopsy specimens obtained at bronchoscopy. Patients were treated appropriately with antimicrobial drugs. Patients received duel antipseudomonas antibiotics (selected on the basis of most recent preoperative sputum cultures and adjusted in the light of postoperative sputum cultures) for the first 10 postoperative days. They also received intravenous flucloxacillin (unless allergic to penicillin) while central lines and chest drains were in situ. All patients with positive sputum cultures were treated with antibiotics in the first and second postoperative months and thereafter only if they had clinical respiratory infections. Patients were prescribed life-long colomycin inhalational therapy via a face mask at a dose of 1 MU twice a day in the first 3 postoperative months and thereafter 1 MU daily. Acyclovir and nystatin were also prescribed for the first 3 postoperative months, as was life-long cotrimoxazole 960 mg twice daily every third day, or, in patients intolerant to this antibiotic, pentamidine isethionate 600 mg by nebuliser every 2 weeks. Patients were fed as soon as bowel sounds returned, and pancreatic enzyme supplementation was given with food, snacks, and oral cyclosporin. 20 patients required nasogastric feeding with elemental diet after surgery, 12 required total parenteral nutrition, and a feeding gastrostomy tube was inserted in 5.
Results Survival Actuarial survival for the whole group was 69% to 1 year, 52% to 2 years, and 49% to 3 years. There were no deaths after 28 months and the longest surviving patient is now at 66 months after his operation. There were 31 deaths, which were due to infection (11), multiorgan failure (8), haemorrhage (6), cerebrovascular accident (2), obliterative bronchiolitis (2), anastomotic dehiscence (1), and hyperacute rejection (1). Mortality after transplantation is shown in fig 1. The actuarial survival of patients in the low-risk group (n 56) was 71 % to 1 year and 49% to 2 years (fig 2). When considered separately, the patients with diabetes mellitus had a survival of 62% to 1 year and 51% to 2 years. The actuarial survival at 1 and 2 years for patients at high risk was 64% and 57%, respectively (fig 2). The difference between both groups was significant to 3 months (p < 0’05) but thereafter there were no statistically significant differences between the groups. We believe that this finding relates to bleeding, infection, and multiorgan failure, which were more common in the high-risk group during the early postoperative period. 8 of the 12 ventilated patients, 1 of =
the
3 patients who had received heart-lung-liver transplantation, 5 of the 10 patients who had had previous
1585
Fig 2-Survival of high-risk vs low-risk patients in heart-lung transplantation for cystic fibrosis. High-risk group= preoperative ventilation, thoracic surgery, heartlung-liver transplantation, plasmodesis. thoracic surgery, and 1 of the 2 patients who had had previous chemical pleurodesis survived.
Rejection Episodes of acute rejection were commonest during the first 3 postoperative months, with 92 % of patients having at least 1 acute rejection episode during this time. Thereafter, acute rejection was much less frequent so that between 7 and 12 months, 13 and 24 months, and 25 and 36 months 74%, 84%, and 82% of patients, respectively, did not have any acute rejection episodes (fig 3). There were no episodes of acute rejection after 36 months. Infections
Respiratory infections that required treatment are shown Additionally, there were 2 cases of infectious mononucleosis, 1 of herpes simplex encephalitis, 1 of Clostridium difficile enteritis, 1 of systemic toxoplasmosis, and 1 of painful viral vulval condolomata. in the table.
Lung function Lung function improved quickly during the first 3 postoperative months with a substantial improvement over pretransplant values (fig 4). At 3, 6, 12, 24, and 36 months after surgery the mean FEV1 was 60, 68, 67, 70, and 60% predicted, respectively, with corresponding values for FVC of 61, 70, 71, 70, and 66% predicted, respectively. Most patients were well enough to return to their normal activities (eg, school, higher education, and work) between 6 and 12 months after transplantation.
100l
No of
episodes .0 1
1 0 i!!a3 !;34 4 Fl 5
Fig 4-Respiratory function after heart-lung transplantation for cystic fibrosis. Bars=SEM.
Airways complications Tracheal stenosis developed in 3 patients. 1 responded to repeated dilatation alone, and 1 to repeated dilatation and insertion of a tracheal stent. The remaining patient had a repeat heart-lung transplantation and subsequently died of a CMV infection. 3 patients had a clinically significant growth of granulation tissue in the left main bronchus. All were treated with repeated dilatation and cryotherapy and 1 also required the insertion of an endobronchial stent. Bronchiectasis developed in 1 of these patients, who awaits retransplantation, and 2 died (1 from bleeding and 1 from overwhelming infection). Coronary angiography 37 patients underwent coronary angiography 1 year after heart-lung transplantations, 14 at 2 years, and 9 at 3 years. There was no evidence of coronary artery disease in any of these patients. Obliterative bronchiolitis Late deterioration in lung function due to obliterative bronchiolitis after lung transplantation is the most serious late complication and is a clinical diagnosis of progressive airflow obstruction often in the presence of infection. The patient usually presents with cough, reduced exercise tolerance, and deterioration in lung function. Chest radiographs typically are clear or may show hyperexpanded lung fields. Obliterative bronchiolitis developed in 17 of our patients. The cumulative probability (70% confidence interval) of having this complication at 1, 2, and 3 years after surgery was 17% (12-23), 23% (16-29), and 48% (38-59), respectively. 7 patients are now being treated medically with
augmented immunosuppression (5
are
stable,
RESPIRATORY INFECTIONS REQUIRING TREATMENT
Fig 3-Number of episodes of acute rejection transplantation for cystic fibrosis.
in
heart-lung Data
are no
of mfections
1
is
1586
improving, 1 is deteriorating); 3 patients died after retransplantation for obliterative bronchiolitis; 2 died awaiting retransplantation; and 3 are awaiting retransplantation. 2 patients have been successfully retransplanted. Lymphoproliferative disorders 4 patients had lymphoproliferative disorders (3 B-cell and 1 T-cell), all of which occurred during the first postoperative year. The B-cell disorders
were confined to the thorax, whereas the patient in whom a T-cell disorder developed presented with a maculopapular rash, oral ulceration, and generalised lymphadenopathy with hepatosplenomegaly. Patients with a B-cell disorder (associated with Epstein-Barr virus infection) were successfully treated with a reduction in immunosuppression and high-dose acyclovir. The patient with the T-cell disorder was treated by reduction in immunosuppression alone and this patient remains well 13 months after successful treatment. Only 1 of the 3 patients with a B-cell disorder remains well 27 months after developing this complication; 1 died 5 months after retransplantation for obliterative bronchiolitis; and the other died 4 weeks after diagnosis from overwhelming sepsis.
Retransplantation 7 patients were retransplanted a median of 15 months after the first operation (range 5-30). Of the 5 patients retransplanted for obliterative bronchiolitis 2 are well at 37 and 9 months after surgery, and 3 died, 1 from bleeding, 2 from multiorgan failure (1had a retransplantation [single lung] at another centre). 1 patient had recurrent infection due to tracheal stenosis and after retransplantation was well until the fifth postoperative month when CMV pneumonitis developed and she died. The seventh patient was retransplanted because of graft failure and he died 5 months later due to peritonitis. 4 patients are currently awaiting retransplantation (3 obliterative bronchiolitis, 1 bronchiectasis). Discussion These medium-term results confirm the short-term
findings that heart-lung transplantation for CF leads to a pronounced improvement in lung function and good rehabilitation after surgery. There are still difficulties related to bleeding from adhesions in the pleural cavity, infection, and multiorgan failure, especially in patients with increased risk factors. The late deterioration in lung function due to obliterative bronchiolitis remains a serious difficulty but this is being addressed by aggressive detection and treatment of rejection, newer immunosuppressive agents, and revascularisation of the bronchial anastomosis. Malabsorption, diabetes mellitus, infection of the upper airways, and salt loss have been successfully managed, and survival for patients with CF after heart-lung transplantation is as good as that for non-CF patients who have undergone heart-lung transplantation at our unit. 10 We believe that long-term aerosol colomycin therapy reduces the incidence of pseudomonas infections in the transplanted
lung. Several surgical techniques are available for transplantation in patients with lung disease:lO single-lung, double-lung, heart-lung, and bilateral-single-lung is Single-lung transplantation. transplantation for with CF who have bilateral severe patients inappropriate pulmonary infection, and after double-lung transplantation without direct revascularisation of the bronchus many
patients have large airway complications.ll The heart-lung transplantation technique remains the treatment of choice. However, direct revascularisation of the bronchial arteries with the internal mammary artery, a procedure that has been developed at our centre, improves large-airway healing; therefore, double-lung transplantation with this technique is now an option for this group of patients. We remain concerned about the use of the bilateral-single-lung technique without revascularisation. There are many advantages in transplanting patients with CF. The patients are young, they and their families are well motivated, they are used to attending hospitals and having intravenous drugs and physiotherapy, and they take an active interest in the management of their disease. Furthermore, donor organ supply is improved by the use of CF hearts in the domino procedure. These hearts often have right ventricular hypertrophy and are especially suitable for transplantation into patients with moderate reversible elevation of pulmonary vascular resistance. Results of the domino transplants have been as good as for other heart
transplants.8 We have also shown that it is possible to carry out combined heart-lung and liver transplantation on patients with portal hypertension and that it is possible to successfully operate on patients who are on ventilators preoperatively. The use of the Brompton Pneupac ventilator is especially encouraging in this latter group.12 New
techniques to control bleeding (eg, use of aprotinin13 and the Argon coagulator) make it possible to operate successfully on many patients who have had previous thoracic surgery. Thus, we would not exclude patients who have had previous pleurodesis or who are on ventilators at the time of surgery as is the practice in some centres.l4 Shortage of suitable donor organs is a serious issue, and continued efforts must be made to ensure that the subject of organ donation is raised in a caring and compassionate way with relatives of potential donors who fulfil brain-death criteria. The cost of transplantation in CF is especially high because large doses of cyclosporin are needed for malabsorption but it is less expensive than caring for a chronically sick patient with CF who requires frequent hospital admissions and intravenous antibiotics. Furthermore, the quality of life is much better for the successfully transplanted patient. Initially there was concern about possible recurrence of CF in the transplanted lung but our experience indicates that there is no such recurrence. Wood et al 15 have shown that the potential difference across the respiratory mucosa (which is more negative in CF patients than in normal healthy controls) is in the normal range 2 years after transplantation. The timing of surgery is vital: it must not be too early because about 30% of patients die within the first postoperative year; nor must it be too late, because suitable donor organs may not be found in time or the patient may be too ill to be successfully transplanted. It is essential that only those patients who have had the maximum medical therapy should be considered for transplant surgery. Our results so far are encouraging but we regard the shortage of donor organs and obliterative bronchiolitis as the two main obstacles to be overcome. We hope that further modifications in the operative technique and after-care will improve the results. We thank Miss M. Rehahn for statistical advice, the Cystic Fibrosis Research Trust for fmancial support for B. M. and V. T., Miss Sally Hockley for typing the script, and Sir J. C. Batten, KCVO, for much help and encouragement during the development of the transplant programme.
1587
REFERENCES 1. Penketh ARL, Wise A, Mearns MB, Hodson ME, Batten JC. Cystic fibrosis in adolescents and adults. Thorax 1987; 42: 526-32. 2. Hodson ME. Managing adults with cystic fibrosis. Br Med J 1989; 298: 471-72. 3. Report of the Royal College of Physicians. Recommendations for care of CF patients in the UK. London: Royal College of Physicians of London, November, 1990. 4. British Paediatric Association Working Party on Cystic Fibrosis. Cystic fibrosis in the United Kingdom 1977-1985: an improving picture. Br Med J 1988; 297: 1599-602. 5. Niellsen OH, Schiotz PO. Evaluation of centralised treatment 19451981. Acta Paed Scand 1982; (suppl 301): 107-19. 6. Corey M, McLaughlin FJ, Williams M, et al. A comparison of survival growth and pulmonary function in patients with cystic fibrosis in Boston and Toronto. J Clin Epidermiol 1988; 41: 583-91. 7. Rommers JM, Ianuzzi MC, Kerrem B, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science 1989; 245:
1059-65. 8. Yacoub MH, Banner NR, Khaghani A, et al. Heart-lung transplantation
for cystic fibrosis and subsequent domino heart transplantation. J Heart Transplant 1990; 9: 459-67. 9. Scott J, Higenbottam T, Hutter J, et al. Heart-lung transplantation for cystic fibrosis. Lancet 1988; ii: 192-94. 10. Yacoub MH, Banner NR. Recent developments in lung and heart-lung transplantation. In: Morris PJ, Tilney NL, eds. Transplantation reviews. Philadelphia: Saunders, 1990: 1-29. 11. Patterson GA. Double lung transplantation. Pediatr Pulmonol 1989; 4: 56-57. 12. Hodson ME, Madden BP, Steven MH, Tsang VT, Yacoub MH. Non-invasive mechanical ventilation for cystic fibrosis patients: a bridge to transplantation. Eur Respir J 1991; 4: 524-27. 13. Alajmo FR, Giancarlo C, Pema AM, et al. High dose aprotinin haemostatic effects in open heart operations. Ann Thorac Surg 1989; 48: 536-39. 14. Smyth RL, Higenbottam T, Scott J, Wallwork J. The current state of lung transplantation for cystic fibrosis. Thorax 1991; 46: 213-16. 15. Wood A, Higenbottam T, Jackson M. Airway mucosal bioelectric potential differences in cystic fibrosis after lung transplantation. Am Rev Respir Dis 1989; 140: 1645-49.
PUBLIC HEALTH Health effects of white-water canoeing
There is little quantitative information on the relation between water quality and disease attack rates after recreational activities in fresh water. We conducted a prospective cohort study to measure the health effects of white-water and slalom canoeing in two channels with different degrees of microbial contamination. Site A, fed by a lowland river, showed high enterovirus concentrations (arithmetic mean 198 pfu per 10 litre and moderate faecal coliform concentrations (geometric mean 285/dl); at site B, from an upland impoundment, all samples were free of enteroviruses and the geometric mean faecal coliform concentration was 22/dl. Between 5 and 7 days after exposure canoeists using site A had significantly higher incidences of gastrointestinal and upper respiratory symptoms than canoeists using site B or non-exposed controls (spectators). Like seawater bathers, fresh-water canoeists can be made ill by sewage contamination. The hazard of fresh water may be best measured by counting of viruses rather than bacteria. Introduction is made of inland waters by canoeists, water skiers, sailboarders, and casual bathers; and, as with seawater bathing, the effects of sewage contamination give cause for concern.1,2 Opinions differ on whether the same microbiological safety standards should be applied to marine and fresh recreational waters.2,3 Further, there is currently a debate about possible changes to the microbiological standards in the European Bathing Water Directive.4 Although the LTK maintains that this directive relates exclusively to coastal bathing waters, other member Much
use
countries apply it to inland freshwater sites and its standards are used as a guide by recreational groups. Notwithstanding earlier research5-7 there are insufficient data to determine the microbiological dose-response relation between water quality and disease attack rates for the various freshwater activities. We have used methods developed in epidemiological studies at marine bathing sites8 to measure the health effects of canoeing in two slalom channels with different water quality.
Methods At site A, an artificial concrete channel diverts water around a weir in the river’s main channel and provides facilities for white-water and slalom canoeing. The water comes from a large lowland river receiving effluent from several sewage treatment plants upstream. Site B receives water from a pristine upland impoundment. Water samples were taken at each site during the event. Bacteriological analysis included counts of faecal coliforms,9 faecal streptococci,9 and total staphylococci.10 Samples were also assessed for enteroviruses" by measurement of plaque forming units (pfu) with Buffalo green monkey kidney cells and the suspended cell method; poliovirus 2 was used as a control. Subjects were recruited by advertising in the canoeing press and by approaching canoeists and spectators before the events. Data on age, sex, occupation, place of residence, water activities including canoeing, consumption of various foods, and illness were obtained on a structured questionnaire by trained interviewers, initially on arrival. Participants were interviewed a second time on leaving the site to determine water exposure, and 5-7 days later a third interview was conducted by telephone to record subsequent illness. A fourth questionnaire was sent by post so that subjects could record illness experienced between the second interview and 28
days. ADDRESSES: Centre for Research into Environment and Health, University of Wales, Lampeter (L. Fewtrell, MSc, F. Jones, B Tech, D. Kay, PhD, M. D. Wyer, PhD); Acer Environmental, Daresbury, Cheshire (A. F. Godfree, BSc); and PHLS Communicable Disease Surveillance Centre, Welsh Unit, Cardiff (B. L. Salmon, MFPHM). Correspondence to Mrs L. Fewtrell, Centre for Research into Environment and Health, University of Wales, Lampeter, Dyfed SA48 7ED, UK.
CLINICAL PRACTICE Intermediate-term results of heart-lung transplantation for cystic fibrosis
Between September, 1984, and March, 1991, 79 patients underwent heart-lung transplantation for end-stage cystic fibrosis at the Harefield Hospital. Short-term outcome has already been reported, and we now present intermediate-term results. The overall actuarial patient survival was 69% at 1 year, 52% at 2 years, and 49% at 3 years. 17 patients had diabetes mellitus with a survival of 62% to 1 year and 51% to 2 years. 23 patients had one or more other possible high-risk factors, and survival of these patients was 64% at 1 year and 57% at 2 years, compared with 71% and 49%, respectively, in the low-risk group (n=56). Pseudomonas aeruginosa infection was the most common respiratory infection encountered postoperatively. 92% of patients had at least one episode of acute rejection during the first 3 postoperative months. Lung function was greatly improved after transplantation, the mean forced expiratory volume in 1 s and forced vital capacity increasing from 22% and 35% predicted, respectively, preoperatively to 68% and 70% predicted, respectively, by the sixth postoperative month. This improvement was maintained at 1, 2, and 3 years after transplantation. Lymphoproliferative disorders (4 patients) were successfully treated. Obliterative bronchiolitis developed in 17 patients and the cumulative probability of getting this complication at 1, 2, and 3 years postoperatively was 17%, 23%, and 48%, respectively. Overall, 7 patients were retransplanted. There was no coronary artery disease in the 37 patients who underwent coronary angiography at 1 year, 14 at 2 years, and 9 at 3 years after surgery. 58 patients donated their hearts for subsequent "domino" heart transplantation. Our 5½-year experience with heart-lung transplantation is encouraging but the shortage of donor organs and the complication of obliterative bronchiolitis are the two main obstacles to be overcome.
Nowadays, many are surviving into adult life,! there 2000 CF adults over the age of 15 years in the about being UK.3 The estimated survival to 20 years is broadly similar in the UK (50%) Denmark (52%),5 the USA (43%),6 and Canada (55%).6 The gene responsible for CF has been identified,7 and since the disease seems to be limited to the cell membrane of the affected patient it should not affect the lungs of a transplanted patient. Against this background of a disease with a high mortality in adolescent and early adult life, the possibility of lung transplantation has become a of life.
promising therapeutic option. Initially there was much concern about operating on patients with CF because of the multisystem nature of the disease, together with chronic infection in the upper and lower respiratory tract. However, in October, 1985, the first successful transplants for CF were done at Harefield and Papworth Hospitals in the UK.8.9 The early results of selected series have been reported. Here we report our 5 2 ’-year experience with 79 CF patients who have undergone heart-lung transplantation, so that we can clarify the intermediate-term outcome of the procedure for CF and identify the high-risk factors. Patients and methods CF patients Between September, 1984, and March, 1991, 79 patients (39 male, 40 female; mean age 23 years, range 8-43) with CF underwent heart-lung transplantation in our unit. 3 patients underwent combined heart-lung and liver transplantation. The indications for surgery were severe respiratory failure with severely impaired quality of life. All patients were keen to have a transplant. At the time of acceptance for surgery, the mean forced expiratory volume in 1 s (FEV 1) was 22% predicted and the mean forced vital capacity (FVC) was 35 % predicted (n 78). When breathing air, mean Pa02 was 6.57 kPa (range 3-6-9-6) and mean PaC02 was 6-79 kPa (range 3-6-10-8, n=50). Preoperatively, sputum samples yielded Pseudomonas aeruginosa (70 patients), P cepacia (3), P maltophilia (1), Staphylococcus aureus (17), methicillin-resistant Staph aureus (4), Haemophilus influenzae (4), Streptococcus pneumoniae (2), and Aspergillusfumigatus (1); no pathogens were isolated from 1 patient. Many patients had more than one organism in their sputum. Contraindications to heart-lung transplantation were active aspergillus or mycobacterial infection, non-compliance with treatment, prednisolone therapy (> 10 mg/day), or other end-stage organ failure (unless this was also treatable at the time of transplantation-for example, by heart-lung and liver transplantation in patients with coexisting severe liver disease). =
Introduction
Cystic fibrosis (CF) is a multisystem disease, the main features of which are chronic bronchopulmonary infection, malabsorption, and high sweat-sodium concentration. There is a high incidence of diabetes mellitus, and most adults are colonised by Pseudomonas aeruginosa in the upper and lower respiratory tract.1,2 When the disorder was first described in 1938,80% of patients died within the first year
ADDRESSES: Harefield Hospital, Harefield, Middlesex, UK (B. P. Madden, MD, R. Radley-Smith, FRCP, A. Khaghani, FRCS, Prof M. Y. Yacoub, Kt), Royal Brompton Hospital, Sydney Street, London, UK (B. P Madden, M. E. Hodson, FRCP, V. Tsang, FRCS, M Y. Yacoub), and National Heart and Lung Institute, Dovehouse Street, London, UK (B. P. Madden, M E Hodson, V. Tsang, A. Khaghani, M. Y. Yacoub). Correspondence to Sir Magdi Yacoub, Harefield Hospital, Uxbridge, Middlesex UB9 6JH, UK.
1584
17 patients had diabetes mellitus and 23 had one or more other possible high-risk factors; these included pre-operative ventilation in 12 patients (6 conventional ventilation, 5 nasal ventilation, 1 ventilation with conventional extracorporeal membrane oxygenation), previous thoracic surgery in 10 (5 abrasion pleurodesis [1 bilateral], 1 double-lung transplantation, 3 thoracotomy [bilateral],1 pleurectomy), chemical pleurodesis in 2 (1bilateral), and combined heart-lung-liver transplantation for liver failure in 3. Some patients had more than one high-risk factor. 9 patients had a feeding gastrostomy tube inserted before surgery to improve nutrition.
Donor selection and organ procurement
’
The general criteria for donor selection and organ procurement have been described previously.lO Briefly, a portable cardiopulmonary bypass machine was taken by the retrieval team to the donor hospital. The donor was cooled on bypass to a core temperature of 8-10°C, the heart was preserved with cold blood potassium cardioplegia solution, and the organs were transported in cold donor blood and packed in ice.
Operative technique The surgical methods for heart-lung transplantation used at Harefield are described elsewhere.1O Matching criteria were based on blood group, size of thoracic cage, and cytomegalovirus (CMV) status. 58 of the 79 CF patients donated their hearts to patients who required cardiac transplantation with a primed heart (with a well-developed right ventricle) for moderate elevation in pulmonary vascular resistance. This "domino" procedures involves removal of the CF recipient’s heart intact for subsequent transplantation. The heart-lung block is inserted into the CF patient with separate superior and inferior vena caval anastamoses instead of a conventional right atrial anastamosis.
Immunosuppression Cyclosporin 6-10 mg/kg and oral azathioprine 2 mg/kg were given preoperatively. 1 g methylprednisolone was given intravenously during surgery after release of the aortic cross clamp. Routine immunosuppression postoperatively was achieved with cyclosporin, azathioprine, and rabbit anti-thymocyte globulin (RATG). Cyclosporin was given intravenously in the immediate postoperative period and thereafter orally with the aim to attain levels of 500 ng/ml (whole blood monoclonal antibody assay) during the first month postoperatively and thereafter between 250 and 350 ng/ml. Renal function was carefully monitored; if serum creatinine increased above 200 unol/1, cyclosporin was stopped and patients were given prednisolone until renal function improved and cyclosporin could be recommenced. Azathioprine was administered at a dose of 2 mg/kg per day provided the white blood cell count remained above 4 x 109/1. RATG (100 mg intravenously) was given on alternative days for the first 10 postoperative days with the aim to maintain a T-cell count of less than 200/d.
Postoperative care Episodes of acute rejection were diagnosed according to clinical and radiological indices with respiratory function tests and analysis of transbronchial lung biopsy specimens obtained at fibreoptic bronchoscopy. Patients were treated with intravenous methylprednisolone 1 g/day on 3 consecutive days, occasionally supplemented with RATG. Fibreoptic bronchoscopy was done routinely at the end of the first week and subsequently only if clinically indicated (ie, reduction in lung function or in exercise capacity, abnormality on chest radiograph, pyrexia of unknown origin, or unexplained cough). At bronchoscopy, the anastomosis was inspected and bronchoalveolar lavage specimens were taken for culture and sensitivity, opportunistic pathogen screen, and immunocytochemistry. Transbronchial lung biopsy specimens were sent for histopathological examination and culture. Episodes of infection were diagnosed by means of clinical, serological, and radiological findings together with studies of bronchoalveolar lavage and
Fig 1-Cause of death after heart-lung transplantation for cystic fibrosis. Redo= retransplantation, OB = obliterative bronchiolitis. *After retransplantation elsewhere.
transbronchial lung biopsy specimens obtained at bronchoscopy. Patients were treated appropriately with antimicrobial drugs. Patients received duel antipseudomonas antibiotics (selected on the basis of most recent preoperative sputum cultures and adjusted in the light of postoperative sputum cultures) for the first 10 postoperative days. They also received intravenous flucloxacillin (unless allergic to penicillin) while central lines and chest drains were in situ. All patients with positive sputum cultures were treated with antibiotics in the first and second postoperative months and thereafter only if they had clinical respiratory infections. Patients were prescribed life-long colomycin inhalational therapy via a face mask at a dose of 1 MU twice a day in the first 3 postoperative months and thereafter 1 MU daily. Acyclovir and nystatin were also prescribed for the first 3 postoperative months, as was life-long cotrimoxazole 960 mg twice daily every third day, or, in patients intolerant to this antibiotic, pentamidine isethionate 600 mg by nebuliser every 2 weeks. Patients were fed as soon as bowel sounds returned, and pancreatic enzyme supplementation was given with food, snacks, and oral cyclosporin. 20 patients required nasogastric feeding with elemental diet after surgery, 12 required total parenteral nutrition, and a feeding gastrostomy tube was inserted in 5.
Results Survival Actuarial survival for the whole group was 69% to 1 year, 52% to 2 years, and 49% to 3 years. There were no deaths after 28 months and the longest surviving patient is now at 66 months after his operation. There were 31 deaths, which were due to infection (11), multiorgan failure (8), haemorrhage (6), cerebrovascular accident (2), obliterative bronchiolitis (2), anastomotic dehiscence (1), and hyperacute rejection (1). Mortality after transplantation is shown in fig 1. The actuarial survival of patients in the low-risk group (n 56) was 71 % to 1 year and 49% to 2 years (fig 2). When considered separately, the patients with diabetes mellitus had a survival of 62% to 1 year and 51% to 2 years. The actuarial survival at 1 and 2 years for patients at high risk was 64% and 57%, respectively (fig 2). The difference between both groups was significant to 3 months (p < 0’05) but thereafter there were no statistically significant differences between the groups. We believe that this finding relates to bleeding, infection, and multiorgan failure, which were more common in the high-risk group during the early postoperative period. 8 of the 12 ventilated patients, 1 of =
the
3 patients who had received heart-lung-liver transplantation, 5 of the 10 patients who had had previous
1585
Fig 2-Survival of high-risk vs low-risk patients in heart-lung transplantation for cystic fibrosis. High-risk group= preoperative ventilation, thoracic surgery, heartlung-liver transplantation, plasmodesis. thoracic surgery, and 1 of the 2 patients who had had previous chemical pleurodesis survived.
Rejection Episodes of acute rejection were commonest during the first 3 postoperative months, with 92 % of patients having at least 1 acute rejection episode during this time. Thereafter, acute rejection was much less frequent so that between 7 and 12 months, 13 and 24 months, and 25 and 36 months 74%, 84%, and 82% of patients, respectively, did not have any acute rejection episodes (fig 3). There were no episodes of acute rejection after 36 months. Infections
Respiratory infections that required treatment are shown Additionally, there were 2 cases of infectious mononucleosis, 1 of herpes simplex encephalitis, 1 of Clostridium difficile enteritis, 1 of systemic toxoplasmosis, and 1 of painful viral vulval condolomata. in the table.
Lung function Lung function improved quickly during the first 3 postoperative months with a substantial improvement over pretransplant values (fig 4). At 3, 6, 12, 24, and 36 months after surgery the mean FEV1 was 60, 68, 67, 70, and 60% predicted, respectively, with corresponding values for FVC of 61, 70, 71, 70, and 66% predicted, respectively. Most patients were well enough to return to their normal activities (eg, school, higher education, and work) between 6 and 12 months after transplantation.
100l
No of
episodes .0 1
1 0 i!!a3 !;34 4 Fl 5
Fig 4-Respiratory function after heart-lung transplantation for cystic fibrosis. Bars=SEM.
Airways complications Tracheal stenosis developed in 3 patients. 1 responded to repeated dilatation alone, and 1 to repeated dilatation and insertion of a tracheal stent. The remaining patient had a repeat heart-lung transplantation and subsequently died of a CMV infection. 3 patients had a clinically significant growth of granulation tissue in the left main bronchus. All were treated with repeated dilatation and cryotherapy and 1 also required the insertion of an endobronchial stent. Bronchiectasis developed in 1 of these patients, who awaits retransplantation, and 2 died (1 from bleeding and 1 from overwhelming infection). Coronary angiography 37 patients underwent coronary angiography 1 year after heart-lung transplantations, 14 at 2 years, and 9 at 3 years. There was no evidence of coronary artery disease in any of these patients. Obliterative bronchiolitis Late deterioration in lung function due to obliterative bronchiolitis after lung transplantation is the most serious late complication and is a clinical diagnosis of progressive airflow obstruction often in the presence of infection. The patient usually presents with cough, reduced exercise tolerance, and deterioration in lung function. Chest radiographs typically are clear or may show hyperexpanded lung fields. Obliterative bronchiolitis developed in 17 of our patients. The cumulative probability (70% confidence interval) of having this complication at 1, 2, and 3 years after surgery was 17% (12-23), 23% (16-29), and 48% (38-59), respectively. 7 patients are now being treated medically with
augmented immunosuppression (5
are
stable,
RESPIRATORY INFECTIONS REQUIRING TREATMENT
Fig 3-Number of episodes of acute rejection transplantation for cystic fibrosis.
in
heart-lung Data
are no
of mfections
1
is
1586
improving, 1 is deteriorating); 3 patients died after retransplantation for obliterative bronchiolitis; 2 died awaiting retransplantation; and 3 are awaiting retransplantation. 2 patients have been successfully retransplanted. Lymphoproliferative disorders 4 patients had lymphoproliferative disorders (3 B-cell and 1 T-cell), all of which occurred during the first postoperative year. The B-cell disorders
were confined to the thorax, whereas the patient in whom a T-cell disorder developed presented with a maculopapular rash, oral ulceration, and generalised lymphadenopathy with hepatosplenomegaly. Patients with a B-cell disorder (associated with Epstein-Barr virus infection) were successfully treated with a reduction in immunosuppression and high-dose acyclovir. The patient with the T-cell disorder was treated by reduction in immunosuppression alone and this patient remains well 13 months after successful treatment. Only 1 of the 3 patients with a B-cell disorder remains well 27 months after developing this complication; 1 died 5 months after retransplantation for obliterative bronchiolitis; and the other died 4 weeks after diagnosis from overwhelming sepsis.
Retransplantation 7 patients were retransplanted a median of 15 months after the first operation (range 5-30). Of the 5 patients retransplanted for obliterative bronchiolitis 2 are well at 37 and 9 months after surgery, and 3 died, 1 from bleeding, 2 from multiorgan failure (1had a retransplantation [single lung] at another centre). 1 patient had recurrent infection due to tracheal stenosis and after retransplantation was well until the fifth postoperative month when CMV pneumonitis developed and she died. The seventh patient was retransplanted because of graft failure and he died 5 months later due to peritonitis. 4 patients are currently awaiting retransplantation (3 obliterative bronchiolitis, 1 bronchiectasis). Discussion These medium-term results confirm the short-term
findings that heart-lung transplantation for CF leads to a pronounced improvement in lung function and good rehabilitation after surgery. There are still difficulties related to bleeding from adhesions in the pleural cavity, infection, and multiorgan failure, especially in patients with increased risk factors. The late deterioration in lung function due to obliterative bronchiolitis remains a serious difficulty but this is being addressed by aggressive detection and treatment of rejection, newer immunosuppressive agents, and revascularisation of the bronchial anastomosis. Malabsorption, diabetes mellitus, infection of the upper airways, and salt loss have been successfully managed, and survival for patients with CF after heart-lung transplantation is as good as that for non-CF patients who have undergone heart-lung transplantation at our unit. 10 We believe that long-term aerosol colomycin therapy reduces the incidence of pseudomonas infections in the transplanted
lung. Several surgical techniques are available for transplantation in patients with lung disease:lO single-lung, double-lung, heart-lung, and bilateral-single-lung is Single-lung transplantation. transplantation for with CF who have bilateral severe patients inappropriate pulmonary infection, and after double-lung transplantation without direct revascularisation of the bronchus many
patients have large airway complications.ll The heart-lung transplantation technique remains the treatment of choice. However, direct revascularisation of the bronchial arteries with the internal mammary artery, a procedure that has been developed at our centre, improves large-airway healing; therefore, double-lung transplantation with this technique is now an option for this group of patients. We remain concerned about the use of the bilateral-single-lung technique without revascularisation. There are many advantages in transplanting patients with CF. The patients are young, they and their families are well motivated, they are used to attending hospitals and having intravenous drugs and physiotherapy, and they take an active interest in the management of their disease. Furthermore, donor organ supply is improved by the use of CF hearts in the domino procedure. These hearts often have right ventricular hypertrophy and are especially suitable for transplantation into patients with moderate reversible elevation of pulmonary vascular resistance. Results of the domino transplants have been as good as for other heart
transplants.8 We have also shown that it is possible to carry out combined heart-lung and liver transplantation on patients with portal hypertension and that it is possible to successfully operate on patients who are on ventilators preoperatively. The use of the Brompton Pneupac ventilator is especially encouraging in this latter group.12 New
techniques to control bleeding (eg, use of aprotinin13 and the Argon coagulator) make it possible to operate successfully on many patients who have had previous thoracic surgery. Thus, we would not exclude patients who have had previous pleurodesis or who are on ventilators at the time of surgery as is the practice in some centres.l4 Shortage of suitable donor organs is a serious issue, and continued efforts must be made to ensure that the subject of organ donation is raised in a caring and compassionate way with relatives of potential donors who fulfil brain-death criteria. The cost of transplantation in CF is especially high because large doses of cyclosporin are needed for malabsorption but it is less expensive than caring for a chronically sick patient with CF who requires frequent hospital admissions and intravenous antibiotics. Furthermore, the quality of life is much better for the successfully transplanted patient. Initially there was concern about possible recurrence of CF in the transplanted lung but our experience indicates that there is no such recurrence. Wood et al 15 have shown that the potential difference across the respiratory mucosa (which is more negative in CF patients than in normal healthy controls) is in the normal range 2 years after transplantation. The timing of surgery is vital: it must not be too early because about 30% of patients die within the first postoperative year; nor must it be too late, because suitable donor organs may not be found in time or the patient may be too ill to be successfully transplanted. It is essential that only those patients who have had the maximum medical therapy should be considered for transplant surgery. Our results so far are encouraging but we regard the shortage of donor organs and obliterative bronchiolitis as the two main obstacles to be overcome. We hope that further modifications in the operative technique and after-care will improve the results. We thank Miss M. Rehahn for statistical advice, the Cystic Fibrosis Research Trust for fmancial support for B. M. and V. T., Miss Sally Hockley for typing the script, and Sir J. C. Batten, KCVO, for much help and encouragement during the development of the transplant programme.
1587
REFERENCES 1. Penketh ARL, Wise A, Mearns MB, Hodson ME, Batten JC. Cystic fibrosis in adolescents and adults. Thorax 1987; 42: 526-32. 2. Hodson ME. Managing adults with cystic fibrosis. Br Med J 1989; 298: 471-72. 3. Report of the Royal College of Physicians. Recommendations for care of CF patients in the UK. London: Royal College of Physicians of London, November, 1990. 4. British Paediatric Association Working Party on Cystic Fibrosis. Cystic fibrosis in the United Kingdom 1977-1985: an improving picture. Br Med J 1988; 297: 1599-602. 5. Niellsen OH, Schiotz PO. Evaluation of centralised treatment 19451981. Acta Paed Scand 1982; (suppl 301): 107-19. 6. Corey M, McLaughlin FJ, Williams M, et al. A comparison of survival growth and pulmonary function in patients with cystic fibrosis in Boston and Toronto. J Clin Epidermiol 1988; 41: 583-91. 7. Rommers JM, Ianuzzi MC, Kerrem B, et al. Identification of the cystic fibrosis gene: chromosome walking and jumping. Science 1989; 245:
1059-65. 8. Yacoub MH, Banner NR, Khaghani A, et al. Heart-lung transplantation
for cystic fibrosis and subsequent domino heart transplantation. J Heart Transplant 1990; 9: 459-67. 9. Scott J, Higenbottam T, Hutter J, et al. Heart-lung transplantation for cystic fibrosis. Lancet 1988; ii: 192-94. 10. Yacoub MH, Banner NR. Recent developments in lung and heart-lung transplantation. In: Morris PJ, Tilney NL, eds. Transplantation reviews. Philadelphia: Saunders, 1990: 1-29. 11. Patterson GA. Double lung transplantation. Pediatr Pulmonol 1989; 4: 56-57. 12. Hodson ME, Madden BP, Steven MH, Tsang VT, Yacoub MH. Non-invasive mechanical ventilation for cystic fibrosis patients: a bridge to transplantation. Eur Respir J 1991; 4: 524-27. 13. Alajmo FR, Giancarlo C, Pema AM, et al. High dose aprotinin haemostatic effects in open heart operations. Ann Thorac Surg 1989; 48: 536-39. 14. Smyth RL, Higenbottam T, Scott J, Wallwork J. The current state of lung transplantation for cystic fibrosis. Thorax 1991; 46: 213-16. 15. Wood A, Higenbottam T, Jackson M. Airway mucosal bioelectric potential differences in cystic fibrosis after lung transplantation. Am Rev Respir Dis 1989; 140: 1645-49.
PUBLIC HEALTH Health effects of white-water canoeing
There is little quantitative information on the relation between water quality and disease attack rates after recreational activities in fresh water. We conducted a prospective cohort study to measure the health effects of white-water and slalom canoeing in two channels with different degrees of microbial contamination. Site A, fed by a lowland river, showed high enterovirus concentrations (arithmetic mean 198 pfu per 10 litre and moderate faecal coliform concentrations (geometric mean 285/dl); at site B, from an upland impoundment, all samples were free of enteroviruses and the geometric mean faecal coliform concentration was 22/dl. Between 5 and 7 days after exposure canoeists using site A had significantly higher incidences of gastrointestinal and upper respiratory symptoms than canoeists using site B or non-exposed controls (spectators). Like seawater bathers, fresh-water canoeists can be made ill by sewage contamination. The hazard of fresh water may be best measured by counting of viruses rather than bacteria. Introduction is made of inland waters by canoeists, water skiers, sailboarders, and casual bathers; and, as with seawater bathing, the effects of sewage contamination give cause for concern.1,2 Opinions differ on whether the same microbiological safety standards should be applied to marine and fresh recreational waters.2,3 Further, there is currently a debate about possible changes to the microbiological standards in the European Bathing Water Directive.4 Although the LTK maintains that this directive relates exclusively to coastal bathing waters, other member Much
use
countries apply it to inland freshwater sites and its standards are used as a guide by recreational groups. Notwithstanding earlier research5-7 there are insufficient data to determine the microbiological dose-response relation between water quality and disease attack rates for the various freshwater activities. We have used methods developed in epidemiological studies at marine bathing sites8 to measure the health effects of canoeing in two slalom channels with different water quality.
Methods At site A, an artificial concrete channel diverts water around a weir in the river’s main channel and provides facilities for white-water and slalom canoeing. The water comes from a large lowland river receiving effluent from several sewage treatment plants upstream. Site B receives water from a pristine upland impoundment. Water samples were taken at each site during the event. Bacteriological analysis included counts of faecal coliforms,9 faecal streptococci,9 and total staphylococci.10 Samples were also assessed for enteroviruses" by measurement of plaque forming units (pfu) with Buffalo green monkey kidney cells and the suspended cell method; poliovirus 2 was used as a control. Subjects were recruited by advertising in the canoeing press and by approaching canoeists and spectators before the events. Data on age, sex, occupation, place of residence, water activities including canoeing, consumption of various foods, and illness were obtained on a structured questionnaire by trained interviewers, initially on arrival. Participants were interviewed a second time on leaving the site to determine water exposure, and 5-7 days later a third interview was conducted by telephone to record subsequent illness. A fourth questionnaire was sent by post so that subjects could record illness experienced between the second interview and 28
days. ADDRESSES: Centre for Research into Environment and Health, University of Wales, Lampeter (L. Fewtrell, MSc, F. Jones, B Tech, D. Kay, PhD, M. D. Wyer, PhD); Acer Environmental, Daresbury, Cheshire (A. F. Godfree, BSc); and PHLS Communicable Disease Surveillance Centre, Welsh Unit, Cardiff (B. L. Salmon, MFPHM). Correspondence to Mrs L. Fewtrell, Centre for Research into Environment and Health, University of Wales, Lampeter, Dyfed SA48 7ED, UK.
