Saturday 3 October

1992

No 8823

ORIGINAL ARTICLES

Randomised controlled trial of cost-effectiveness of lithotripsy and open cholecystectomy as treatments for gallbladder stones

Inpatient extracorporeal shockwave lithotripsy for gallbladder stones has not previously been compared with open cholecystectomy in terms

treatment of

of cost-effectiveness. In

a

randomised controlled

trial, 163 patients, stratified by gallstone bulk (over 4 cm3

not), were randomised to lithotripsy or cholecystectomy (38 large-bulk and 27 small-bulk cholecystectomy; 37 large-bulk and 61 small-bulk lithotripsy) and followed up for 1 year. Both treatments gave significant health gains in terms of a reduction in episodes of biliary pain, improved perceived health status, and symptom or

relief, but few differences between found. There

treatments were

evidence that biliary-pain episodes were less severe after cholecystectomy. Cholecystectomy patients also had greater improvements in mean health gain for three related symptoms: vomiting, feeling sick, and fatty-food upset. However, there were no differences between groups in perceived health status. Among lithotripsy patients, health gain was not related to stone clearance. Lithotripsy was more expensive than cholecystectomy, principally because of the costs of the inpatient stay and adjuvant bile-salt therapy. Conventional lithotripsy appears at least as costeffective as cholecystectomy for patients with smallbulk stones but less cost-effective for those with large-bulk stones. To some extent treatment choice can be guided by patient preference. was some

adjuvant litholytic therapy for gallstones, 91 %

were free of 18 months. Other researchers have also reported varying degrees of clearance of gallstones with lithotripsy.4 However, the health gains after lithotripsy have not been assessed or compared with those achievable by open cholecystectomy. Accordingly, we have undertaken a randomised controlled trial to compare the effect on health status of elective cholecystectomy and ESWL for symptomatic patients with gallbladder stones, and to evaluate the relative cost-effectiveness of these procedures.

stones

after 12

to

Patients and methods Recruitment Between April, 1988, and July, 1990, all symptomatic patients with gallbladder stones seen by nine general surgeons at the Royal Hallamshire Sheffield, for whom elective Hospital, cholecystectomy was indicated as the sole major operative procedure were asked to enter the trial. Later in the study, to accelerate recruitment, consultant surgeons at two other major local hospitals were asked to enter patients. Patients were excluded by surgeons if they were pregnant, did not understand English, had had liver damage, were jaundiced, or had had acute pancreatitis. The general surgeons explained the nature of the trial to all potentially suitable patients and obtained their consent to be assessed for trial entry. At the baseline assessment, patients were interviewed, had biliary system examination by ultrasound, and had a plain film radiograph of the right upper abdominal quadrant to detect stone calcification. Measurements were made of stone number, size, volume, and calcification, and the diameter of the common hepatic duct. Platelet count, clotting screen, and liver-function tests were done, and gallbladder volume and ability to contract were also assessed. After assessments, patients with abnormal-liver function tests, dilated

Lancet 1992; 340: 801-07.

Introduction

Extracorporeal shockwave lithotripsy (ESWL) has been advocated for treatment of renal stones’ and also used to fragment gallstones.2 Sackmann et aP reported that among the first 175 selected patients treated by lithotripsy and

ADDRESSES Medical Care Research Unit, Department of Public Health Medicine, Sheffield University Medical School, Beech Hill Road, Sheffield S10 2BX, UK (J P. Nicholl, MSc, J E. Brazier, MSc, P C Milner, MRCP, C Westlake, BSc, B Kohler, BSc, Prof B. T. Williams, FRCP); and Departments of Radiology (B Ross, MD, E Frost, DMU) and Surgery (Prof A. G Johnson, MChir), Royal Hallamshire Hospital, Sheffield. Correspondence to Mr J P Nicholl.

802

TABLE I-BASELINE CHARACTERISTICS OF PATIENTS AND THEIR GALLSTONES

I

IQR = interquartile *Died from

an

i

range

unrelated lung

cancer

hepatic duct, non-contracting gallbladder (less than 50% volume after a fatty meal), abnormal clotting screen or platelet count, an inaccessible gallbladder for ultrasound or lithotripsy, or who were incapable of being followed-up were excluded. Patients were divided into two groups according to whether their stone bulk was over 4 cm3 or not. Those with large-bulk stones were randomised with equal probability either to ESWL with bile salts or to cholecystectomy, and, initially, those with small-bulk stones were randomised, also with equal probability, to receive one of the three treatments: cholecystectomy, ESWL with bile salts, or ESWL alone. Because of the widely held belief that bile salts were an essential part of ESWL therapy, we planned an early comparison of the outcome oflithotropsy with and without bile salts after about 25 patients with small-bulk stones had been randomised into each arm. After this comparison, all patients with small-bulk stones were randomised with equal probability to cholecystectomy or to lithotripsy with bile salts. The protocol allowed for addition of bile salts to those not started with them if stones showed no change from initial fragmentation by 6 weeks or if it was thought advisible on clinical grounds at 6 months. The results of the bile-salt subtrial will be reported in detail elsewhere.

about 4 weeks before intervention, and then 2 weeks, 5 weeks, 3 months, 6 months, and 1 year after intervention. The measures of effectiveness were health-related quality of life, biliary pain, and gastrointestinal and other symptoms. Quality of life was measured by the Nottingham Health Profile (NHP),5 a 38-item self-administered questionnaire that profiles six dimensions of perceived health status (energy, pain, emotional reactions, sleep, physical mobility, and social isolation). The NHP has been shown to be useful in measuring outcome after cholecystectomy.6 The number of biliary-pain episodes between assessments was recorded and the severity of the most recent episode was measured by a modified McGill pain questionnaire? Biliary pain was defmed, with the help of a picture, as any pain occurring in a specified part of the upper abdomen, lower thorax, and/or mid-back that lasted for more than 30 min. Patients also completed pain diaries between assessments recording the severity on visual analaogue scales (VAS) of up to five episodes. The severity of fourteen gastrointestinal and other symptoms were also measured by VAS. Complications and other adverse sequelae in the inpatient

Treatment Patients randomised to cholecystectomy had a single, elective procedure under general anaesthesia by the referring surgeon and had an inpatient stay. All surgeons did routine per-operative cholangiography. The lithotripsy regimen consisted of hospital admission followed by up to four treatment sessions on consecutive days of up to 3000 shocks per session (Wolf Piezolith 2200/2300 lithotripter, Wolf, Knittington, Germany) depending on the power setting. Treatment was stopped when stones were broken into fragments of about 2 mm diameter or when the maximum number of shocks (12 000) was reached. Lithotripsy was done without general anaesthetic or analgesia. Patients allocated to receive bile salts started them 2 weeks before lithotripsy at a daily dose of 75 mg/kg chenodeoxycholic acid and 6-5 mg/kg ursodeoxycholic acid, both taken in the evening.

Fig 1-Mean biliary-pain episodes per week and Kaplan-Meier estimates of proportion of lithotripsy patients not stone free.

(a) Mean biliary-pain episodes in all cholecystectomy (0-0) or lithotripsy (0- —W) patients. 8-8 lithotripsy patients not stone =

Outcomes and follow-up Outcome

assessed by research staff acting of those treating the patients at baseline, which was

measures

independently

were

free.

(b) Mean biliary-pain episodes for lithotripsy patients whose gallstones cleared (8- -8) and whose gallstones never cleared (0-0)

803

TABLE II-NHP SUMMARY HEALTH-STATUS GAINS BY TREATMENT GROUP AND GALLSTONE BULK I

C = cholecystectomy, L=lithotnpsy p for difference in mean summary responses between patients havi ng cholecystectomy or lithotnpsy adjusted for baseli ne scores and gallstone bulk energy = 0 65. emotional reactions = 58, sleep = 60, social !so!at!on—0 18 physical mob!!ity0 09

treatment phase were reported by the doctor responsible for patient management, and those that occurred after inpatient treatment and resulted in medical consultation were detected by systematic questioning of the patient.

Statistical analysis The outcome considered for determining sample sizes was health gain measured by the NHP. There was no information available on the variability of NHP scores in patients with gallstone disease and it was judged that a difference between the two main treatment groups equivalent to half of one standard deviation of an NHP health-gain score would be clinically relevant. For a 90% chance of detecting such a difference, about 80 patients would be needed if the scores were normally distributed and no covariate adjustments were required. We assumed that appropriate tests might have the same power with 100 patients in each group8 (the actual number achieved fell just short of this). Results were analysed by intention to treat, so that patients in the lithotripsy arm of the trial who were referred for cholecystectomy after complications (in accordance with the protocol) were retained in the lithotripsy arm for analysis.

For NHP and VAS symptom scores, the summary measure9 of health gain after treatment calculated for every patient was the difference between the baseline score and a weighted average of responses at the five post-treatment assessments, with weights proportional to the time patients were presumed to be in each health state. This is equivalent to the area above the response curve measured with respect to the baseline score. A large value in these summary scores represents a large health gain. For biliary pain, outcome was summarised by estimating the number of episodes avoided (that is the number that might have been expected in the 12 months after treatment based on the number reported in the 3 months before baseline assessment, less the number that actually occurred). A weighted estimate of the number of episodes avoided was also calculated, with weights proportional to the severity of the most recent episode at each assessment. Another measure of pain experience was the sum of VAS scores for severity of up to five episodes of biliary pain recorded in the pain diaries completed

between every

assessment.

The

pain diaries before baseline

pain = 0

81,

patients did

assessments

not, of course, have and so there were no

baseline VAS pain measures. Mean and median summary responses were calculated (and 95 % confidence intervals [CI] by bootstrap methods which work by resampling from observed data and thus make no distributional assumptions), and stratification by gallstone bulk has been taken into account in all analyses. The distributions of responses of

patients having lithotripsy or cholecystectomy were compared by (two-sided) rerandomisation (or "permutation") analysis of covariance tests, taking into account baseline scores, which were sensitive to differences in mean response." In patients having lithotripsy, there was no evidence with respect to any outcome of additional health gains from having bile salts; therefore, the 24 lithotripsy patients with small-bulk gallstones who were randomised not to receive adjuvant bile-salt therapy have been pooled with those who did receive bile salts. Costs Total estimated National Health Service (NHS) cost per lithotripsy patient included the costs of a suitability assessment, lithotripsy, ultrasound, hospital stay (nursing, medical and "hotel" services), bile salts, and of any complications requiring NHS services. Cost per lithotripsy session was calculated as the sum of the annual equivalent capital cost and revenue costs divided by the expected total gallstone lithotripsy workload estimated from observations of lithotripsy sessions. Capital costs of the machine were provided by the manufacturer and annuitised assuming a 7-year life. NHS costs of cholecystectomy patients included costs of the cholecystectomy procedure, hospital stay, investigations before and after the procedure, outpatient follow-up, and complications requiring NHS services. Cholecystectomy theatre costs were estimated from observations of time in theatre, staff present, use of consumables, and other services. Patient time costs were recorded for both treatment regimens. Mean costs per patient in the two treatment groups were compared by a rerandomisation test procedure. Cost differences were also subjected to a sensitivity analysis of the main cost estimates.

804

TABLE III-IMPROVEMENTS IN BILIARY PAIN IN 12 MONTHS POST-TREATMENT BY TREATMENT GROUP AND GALLSTONE BULK

C = cholecystectomy, L = lithotnpsy p for difference

in mean

summary response between patients having cholecystectomy or lithotripsy adjusted for baseline frequency and pain episodes = 42, VAS pain daryscoreO 03. m 12 months post treatment, or monthly pain diary scores post treatment, in the latter case, a small value

episodes=0 61, McGill weighted *Estimated pain episodes avoided

Results Baseline

comparisons 179 patients were randomised into the trial, of whom 16 were excluded from analysis for the reasons shown in table I. Thus, 163 patients were included in analyses when data were available. Of 978 possible assessments (163 patients x 6 assessments), only 13 assessments in 10 patients (3 lithotripsy, 7 cholecystectomy) were missed. Randomisation produced groups with similar personal and stone characteristics; however, compared with the other groups the 27 patients with small-bulk stones who had cholecystectomy were older and had biliary colic less often as their main presenting symptom (table I). Cholecystectomy patients had a longer median waiting time from randomisation to intervention because they were treated like other NHS patients. Crossovers and gallstone clearance 9 patients randomised to lithotripsy had cholecystectomy within the 12 month follow-up period, principally because of recurrent episodes of biliary pain and no evidence of gallstone clearance. There were no crossovers from cholecystectomy to lithotripsy. Within 12 months of treatment, 20 (22%) of the 89 patients randomised to lithotripsy who did not crossover and have a cholecystectomy were clear of gallstones at two successive ultrasound scans. Of these 20 patients, 9 (45%) had large-bulk gallstones, all of whom had had bile-salt therapy, and 11 (55%) had small-bulk gallstones, 8 (73%) of whom had had bile salts at some time. During follow-up, there was a steady increase in the number of lithotripsy patients becoming stone free, and the Kaplan-Meier estimate of the proportion stone-free at 12 months post-treatment was 24-3% (fig la).

gallstone bulk unweighted implies

a

pain

better outcome

months had become similar to that in the general population with the same age distribution.5 However, in no health dimension was there a significant difference between the cholecystectomy and lithotripsy treatment groups in mean health gain over the whole 12 months after adjustment for baseline scores (table II). There was evidence of substantial improvements in all health dimensions for the large-bulk and small-bulk stone lithotripsy groups and the large-bulkstone cholecystectomy group, but the small-bulk-stone cholecystectomy group had only about half the health gains of the other three groups in each of the dimensions of the NHP apart from social isolation.

Biliary pain Both

lithotripsy and cholecystectomy patients significant reductions in episodes of biliary pain

had after treatment, but there was no evidence of any difference between lithotripsy and cholecystectomy patients in the number of episodes avoided (table III). Adjusting the episodes of biliary pain for severity with the McGill rating a

r

Health status

significant gains in health status after in both lithotripsy and cholecystectomy patients for the dimensions of pain, energy, emotional reactions, and sleep as measured by the NHP (table 11). Almost the whole of these gains were achieved within 2 weeks in the lithotripsy group and by 5 weeks in the cholecystectomy group. With the exception of sleep, the health status of all patients after 12 There

treatment

were

Fig 2-Distribution of estimated number of episodes of biliary pain avoided in 12 months post treatment. (a) small-bulk stones and lithotripsy, (b) small-bulk stones and cholecystectomy, (c) large-bulk stones and lithotripsy, (d) large-bulk stones and cholecystectomy G patients with no episodes in 3 months before baseline, 8 patients with at least one episode recorded at baseline Negative numbers represent more episodes of biliary pain post =

=

treatment than

expected.

805

TABLE IV-SUMMARY HEALTH-GAIN SCORES FOR SYMPTOMS BY TREATMENT GROUP

C = cholecystectomy, L = lithotripsy p for difference

in mean summary response between patients having cholecystectomy or lithotripsy adjusted for baseline score and gallstone bulk: vomiting = 02, feeling sick=0 02, fatty-food upset =0 05, heartburn = 0 70, belching = 0 32, stomach swelling = 0-15, !ndigest!on-0 30 *Results are shown only for the six symptoms that affected more than half the patients at baseline and for vomiting The other seven symptoms measured were loss of appetite, umntentional weight loss, constipation, diarrhoea, jaundice, fever, and shivering.

TABLE V-MEAN COST PER PATIENT

*Estimated ward cost per

BY TREATMENT AFTER 12 MONTHS

day was E128

before each

did not substantially alter the findings. However, the VAS pain diary scores showed that, overall, there was a significant difference in pain experience post-treatment in favour of cholecystectomy (p=0-03). Improvements in both cholecystectomy and lithotripsy patients were usually achieved within the first 3 months after treatment and showed the same temporal pattern (fig la). In lithotripsy patients, the pattern of improvement in biliary pain bore little resemblance to the temporal pattern of stone clearance; furthermore, the pattern of improvement was similar for lithotripsy patients whose stones cleared and those whose stones did not clear (fig 1b). There was evidence of a substantial reduction in the mean number of episodes of biliary pain per week during the 12 months after treatment in each gallstone-bulk stratum in both treatment groups apart from the small-bulk-stone cholecystectomy group. Patients in the latter group reported more biliary-pain episodes in the 3 months after treatment than in the 3 months before treatment, and only in the last 6 months of follow-up was there a reduction in number of pain episodes. Although some patients in all groups, including both cholecystectomy groups, continued to report biliary pain after treatment, most patients in each of the four treatment and gallstone-bulk groups avoided some pain episodes, and a few patients in every group improved

for the last

(f)

episode

assessment

considerably (fig 2).

Symptoms Within 2 weeks of treatment there was symptoms experienced in both the

a

reduction in all and

lithotripsy

cholecystectomy groups (table IV). This early benefit was particularly striking for vomiting, nausea, fatty-food upset, indigestion, heartburn, and belching. There was little further improvement after the first 2 weeks, but improvements were still present at 12 months. No consistent pattern of differences in symptoms was observed between cholecystectomy and lithotripsy groups and any differences were usually small. However, there was some evidence that cholecystectomy produced a greater reduction than lithotripsy in three related symptoms: vomiting, nausea, and fatty-food upset.

Complications and other adverse sequelae During the first week of treatment, 16 cholecystectomy patients had a complication, usually minor (eg, postoperative pyrexia and bruising), but also including pneumonia (2 patients), hemicolectomy for a diverticular mass (1), cardiac dysrhythmia (1), and bile leak (1), whereas only 2 lithotripsy patients had a minor complication (skin bruising). During the 12 month follow-up, 3 patients (5%) in the cholecystectomy group had major complications (bile leak, hemicolectomy, and subphrenic abscess) as did 4 patients (4%) in the lithotripsy group (all acute cholecystitis). Minor complications were reported by more cholecystectomy patients (32%; mostly postoperative pyrexia and scar pain) than lithotripsy patients (13%; mostly diarrhoea). By contrast, more lithotripsy (34%) than cholecystectomy (18%) patients had other adverse sequelae, principally because more lithotripsy patients reported "biliary colic" and other non-specific abdominal pain necessitating medical care.

806

improvement in symptoms in lithotripsy patients whose gallbladders had not cleared of fragments and the parallel improvements in cholecystectomy patients raise questions about how symptoms are produced and how benefits are brought about. It seems unlikely that benefits of treatment were entirely due to placebo effects since improvements

Costs The costs (at 1989-90 prices) of the 12 months of and follow-up have been estimated separately for patients with large-bulk and small-bulk gallstones in the two arms of the trial and for patients with small-bulk gallstones who had lithotripsy without bile salts (table v). For patients with large-bulk stones, lithotripsy with bile salts cost 778 more than cholecystectomy. However, in the small-bulkstone group, lithotripsy with bile salts cost only C230 more than cholecystectomy. The cost difference was lower in the small-bulk-stone group because of a longer than expected average inpatient stay in the cholecystectomy patients, which was mainly the result of 1 patient being resident for 44 days after an associated hemicolectomy. After adjusting for stone bulk, the overall mean cost per patient of cholecystectomy was significantly less (p < 0-001) than that of lithotripsy with bile salts. The principal cost elements of ward stay, treatment procedure, and, for lithotripsy patients, bile salts accounted for over 80% of the total costs for every group. A one-way sensitivity analysis showed that changing the values of variables between their likely minimum and maximum reduced the cost per patient of lithotripsy below that of cholecystectomy only when there were no bile-salt costs or no ward stay. Results of the analysis were insensitive to large changes in assumptions about the length of life or workload of the lithotripter, its cost, or the discount rate used to annuitise the capital cost of the machine. In terms of patient-time costs, lithotripsy patients visited hospital more often than cholecystectomy patients, but spent less time waiting for treatment and avoided 4 extra days in hospital. treatment

Discussion The expectation at the time of planning the trial that after 12 months all the stones in most lithotripsy patients would have cleared was not realised. Although over 50% clearance has been found for patients selected by wide criteria for lithotripsy,12 the low clearance rate we found (24%) was probably due in part to the heterogeneity of the stone characteristics of our lithotripsy patients,13 who were not restricted to those meeting the "Munich criteria".3 Despite the low rate of gallstone clearance in our lithotripsy patients, we observed substantial improvements in many self-reported symptoms, including pain, and in most dimensions of health. This improvement cannot be the result of stone clearance because it occurred in the majority of patients, most of whose stones did not clear during 12 months’ follow-up, and because nearly all the improvement was achieved within 2 weeks of lithotripsy when most patients still had stone fragments in their gallbladders. Furthermore, the facts that biliary-pain experience posttreatment in lithotripsy patients did not depend on whether or not stones cleared and that the health status of our lithotripsy patients, most of whose stones did not clear, reverted to levels similar to those found in the general population, indicate that stone clearance is unlikely to result in substantial health gains over and above those achieved by lithotripsy alone, at least in the 12 months after treatment that we have studied. The time profile for health improvement and symptom relief after cholecystectomy was similar to that after

lithotripsy, although improvements patients rather

sometimes than 2 week were

not

in

cholecystectomy

apparent until the 5 week

assessment.

The

immediate

sustained for 12 months. Patients were stratified by gallstone bulk at randomisation because it was felt that there would be differences in the effectiveness of lithotripsy in patients with large or small bulk stones. However, in practice, this difference was not found. On the other hand, although the large-bulk stone patients who had cholecystectomy had substantial health gains, these were not matched in the small-bulk stone patients who had cholecystectomy. It seems that patients with large-bulk stones tend to improve more if treated with cholecystectomy and those with small-bulk stones tend to do better with lithotripsy and bile salts. However, the differences are not so great as to make choice of treatment clear. Patients’ treatment preference or access to treatment may continue to influence treatment choice. Mean cost per patient of lithotripsy was higher than that of open cholecystectomy; however, this difference might disappear if lithotripsy were given without bile salts or on an outpatient basis, and provided these changes in treatment regimen did not alter other costs (eg, those associated with complication rates, readmissions, numbers of ultrasound scans). Outpatient lithotripsy would, of course, increase the number of outpatient visits and the time cost to a patient of extra visits would depend on the distance the patients travelled to the lithotripsy facility. Similarly, although lengths of stay for open (small incision) and laparascopic operations14 are often substantially less than reported here, there has been no formal assessment of the relative costs of the different procedures, and costs other than those associated with length of stay, such as those in theatre, may be greater than reported here for conventional open were

z

’

cholecystectomy. In comparing the costs of lithotripsy and cholecystectomy it should also be borne in mind that if lithotripsy is given on a conventional basis with adjuvant bile-salts, then in attempting to achieve gallstone clearance, lithotripsy patients may continue to incur costs beyond 12 months, either because of continuing bile-salt therapy or because of possible stone recurrencels leading to further treatment cycles at some future time. On the other hand, if lithotripsy patients are managed conventionally with adjuvant bile salts, but these are continued only until symptoms resolve, then

our

results suggest that bile-salt costs may be

dramatically reduced. Thus, the relative costs of lithotripsy and cholecystectomy depend not only on length of stay after operation but also on duration of any bile-salt therapy. If lithotripsy is given on the current conventional outpatient basis with adjuvant bile salts to aid stone dissolution for large-bulk-stone patients, lithotripsy would be more costly than cholecystectomy and is no more effective. For small-bulk-stone patients, conventional lithotripsy is at least as effective as cholecystectomy but relative cost will depend on length of stay. We thank the consultant surgeons who participated in this study and allowed us access to their patients: Mr G. L. Cohen, Mr W. P. Morgan, Mr G. Jacobs, Mr A. J. Shorthouse, Mr W. Morris-Jones, Mr C. J. Stoddard, Mr A. Raftery, Mr C. H. Talbot, Mr W. E. G. Thomas, Prof R. G. Clark, Prof K. Rogers, Mr J. A. R. Smith, and Mr M. Simms; and CP Pharmaceuticals for supplying the bile salts. We also thank Dr Stephen Birth for help with the initial stages of the economic evaluation, members of the finance, information, and pharmacy departments, Royal Hallamshire

807

administrative support, and Mrs Maria Platts for assistance with data collection. The Department of Health and Trent Regional Health Authority funded ttus research, and provide core funding for the Medical Care Research Unit. We are grateful to both bodies for their continuing support.

REFERENCES 1.

Chang CR, Webb DR, Payne SR, Wickham JEA. Comparison of

of renal calculi by open surgery, percutaneous treatment nephrolithotomy, and extracorporeal shockwave lithotripsy. BMJ 1986; 292: 879-82. 2. Sauerbruch T, Delivs M, Paumgartner G, et al. Fragmentation of gallstones by extracorporeal shock waves. N Engl J Med 1986; 314: 818-22. 3. Sackmann

M, Delius M, Sauerbruch T, et al. Shock wave lithotripsy of stones. The first 175 patients. N Engl J Med 1988; 318:

gallbladder

393-97. 4. Sackmann M. Electrohydraulic lithotripsy of radiolucent gallbladder calculi. In: Paumgartner G, Sauerbruch T, Sackmann M, Burhenne HJ, eds. Lithotripsy and related techniques for gallstone treatment. St Louis: Mosby Year Book, 1991: 49-58. 5. Hunt SM, McEwen J, McKenna SP. Measuring health status. Beckenham, Kent: Croom Helm, 1986. 6. Bardsley MJ, Venables CW, Watson J, et al. Evidence for validity of a

20 years

or more

of

status measure in assessing short term outcomes of cholecystectomy. Quality in Health Care 1992; 1: 10-14. Melzack R. The McGill pain questionnaire: major properties and scoring methods. Pain 1975; 1: 277-99. Milner PC, Nicholl JP, Westlake L, et al. The evaluation of lithotripsy as

health

Hospital, for calculating the costs of treatment, Mrs Linda Hawkesworth for 7.

8.

for gallstones: a randomised controlled trial approach in England. J Lithotripsy Stone Disease 1989; 1: 122-32. 9. Mathews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial measurements in medical research. BMJ 1990; 300: 230-35. 10. Effron A. Bootstrap and jacknife procedures. Am Statistician 1983; 37: a treatment

36-48.

Manly BFJ. Randomisation and Monte Carlo methods in biology. London: Chapman and Hall, 1991. 12. Darzi A, Monson JRT, O’Morain C, Tanner WA, Keane FBV. Extension of selection criteria for extracorporeal shock wave lithotripsy for gallstones. BMJ 1989; 299: 302-03. 13. Ross BR. Are bile salts necessary with gallstone lithotripsy? In: Paumgartner G, Sauerbruch T, Sackman M, Burhenne HJ, eds. Lithotripsy and related techniques for gallstone treatment. St Louis: Mosby Year Book, 1991: 103-09. 14. Grace PA, Quereshi A, Coleman J, et al. Reduced post-operative hospitalization after laparascopic cholecystectomy. Br J Surg 1991; 78: 11.

160-62. 15. Sackmann

M, Ippisch E, Miller H, et al. Gallstone recurrence after lithotripsy. IVth International Symposium on Biliary Stone Therapy; 1991 Oct 2-4; Rochester, Minnesota.

follow-up of living kidney donors

and long-term risks for living donors are of concern. We have studied kidney donors at the University of Minnesota 20 years or more (mean 23·7) after donation by comparing renal function, blood pressure, and proteinuria in donors with siblings. In 57 donors (mean age 61 [SE 1]), mean serum creatinine is 1·1 (0·01) mg/dl, blood urea nitrogen 17 (0·5) mg/dl, creatinine clearance 82 (2) ml/min, and blood pressure 134 (2)/80 (1) mm Hg. 32% of the donors are taking antihypertensive drugs and 23% have proteinuria. The 65 siblings (mean age 58 [1·3]) do not significantly differ from the donors in any of these variables: 1·1 (0·03) mg/dl, 17 (1·2) mg/dl, 89 (3·3) ml/min, and 130 (3)/80 (1·5) mm Hg, respectively. 44% of the siblings are taking antihypertensives and 22% have proteinuria. To assess perioperative mortality, we surveyed all members of the American Society of Transplant Surgeons about donor mortality at their institutions. We documented 17 perioperative deaths in the USA and Canada after living donation, and estimate mortality to be 0·03%. We conclude that perioperative mortality in the USA and Canada after living-donor nephrectomy is low. In long-term follow-up of our living donors, we found no evidence of progressive renal deterioration The

or

perioperative

other serious disorders. Lancet 1992; 340: 807-10.

Introduction

Despite increased need, the number of cadaver kidneys available in the USA remained stable between 1986 and

1990. In 1990, the supply grew, but not as much as the waiting list1 which, in December, 1991, had 19 337 patients.2 Recipients of living-related donor transplants have better long-term survival (patient and graft) than do recipients of cadaver transplants.1,3 Yet the annual number of living-related donor transplantations in the USA has not increased in the past decade.2 Many transplantation centres do not encourage living-related donation, and many rarely (or never) do such a transplantation. Perioperative morbidity and mortality is one issue.4-7 Concern about long-term risk to the donor after uninephrectomy was accentuated by reports that in animals renal ablation was associated with compensatory changes in the remaining nephrons which led to renal failure;8-lO an early marker is proteinuria. Renal function has been evaluated in donors in the first two decades after donation. Although there was no functional deterioration, some donors had proteinuria and hypertension. To study these concerns, we surveyed all members of the American Society of Transplant Surgeons (ASTS) about perioperative mortality, studied donors at our institution for morbidity 20 years or more after uninephrectomy, and compared renal function, hypertension, and proteinuria in donors with their siblings.

Subjects and methods Living donors and sibling controls Between Jan 1, 1963, and Dec 31,1970,135 patients underwent living-donor transplantation (62 parent, 55 sibling, 6 offspring, 12 other) at the University of Minnesota.’1 The donor had to have perfect health with minimum anaesthesia risks, no transmissible ADDRESSES. Departments of Surgery and Pediatrics (J S Najarian, MD, L. E. McHugh, RN, A. J. Matas, MD), and the Department of Pediatrics (B. M. Chavers, MD), University of Minnesota, Box 328, 420 Delaware Street SE, Minneapolis, Minnesota 55455, USA. Correspondence to Dr A J Matas