COMPARISON OF PATIENTS WITH ACUTE RENAL FAILURE, 1960-69 AND 1980-89
Symptoms in gastro-oesophageal reflux SIR,-Dr Klauser and colleagues (Jan 27, p 205) report the diagnostic validity of symptoms for gastro-oesophageal reflux disease (GORD). Their conclusion that acid regurgitation is insensitive but highly specific can easily be misinterpreted. In the study population acid regurgitation proved the dominant symptom in 4 of 166 (2%) patients with GORD and in 3 of 138 (2%) of those without this disease. Therefore Klauser et al do not demonstrate that acid regurgitation is discriminatory. The likelihood ratios1 are about 1. The observed high specificity of this symptom (98%) is only caused by chance: the prevalence of not having acid regurgitation in the total study population is 98% (297/304). Iota for specificity2 is almost zero. Generally, observed sensitivities and specificities should either be interpreted together, or corrected for
(Mann Whitney and chi-squared tests) ns=p>0 05.
ARF (creatinine over 600 lunol/1, treated at Leeds General Infirmary between 1956 and 1989, 336 were judged to have had ARF as a consequence of sepsis. Throughout the study period this group had the worst prognosis.’ Because of disparities in numbers between the two decades, only patients with medical or general surgical conditions, and only those with ARF solely or largely attributable to sepsis were compared (table). All patients received dialysis (9 peritoneal dialysis). APACHE scores were obtained from the worst recordings on the day of referral to the renal unit, but arterial oxygen and inspired oxygen values were excluded. The time between onset of ARF and referral has not changed between 1960-69 and 1980-89.4 This study of two diagnostically comparable groups of patients treated 20-30 years apart shows that more recent patients are older, have more complicated underlying conditions, are more likely to have multisystem organ failure, and are more severely ill. Why is mortality in ARF persistently high? Because the patients and their illness are worse, such that the outcome is likely to be poor despite severe
and/or requiring dialysis)
chance. Department of Epidemiology/Health Care Research, University of Limburg, 6200 MD Maastricht, Netherlands 1. Sackett DL, Haynes RB, Tugwell P. Clinical
epidemiology: a basic science for clinical
medicine. Boston/Toronto: Little Brown, 1985. 2. Knipschild P. Looking for gall bladder disease in the 1578-81.
**This letter follows.ED. L.
*1960-69 vs 1980-89
patient’s iris. Br Med J 1988; 297:
Dr Klauser, whose
SiR,—Professor Knipschild points out that the specificity of acid regurgitation as a dominant symptom of gastro-oesophageal reflux disease is high because regurgitation is uncommon in the sample overall. It is true that a low test sensitivity favours a high specificity. The low sensitivity of acid regurgitation as a dominant symptom in sample was caused by the definitions used: a clearly dominant symptom could be identified in only 124 of 304 patients. Additionally, dominant symptoms were mutually exclusive by definition. For this reason positive and negative predictive values were shown in table m. In a substantial number of patients (19 of 32) with heartburn as a dominant symptom regurgitation was also present. Therefore we felt it justified to consider severe acid regurgitation as a symptom suggestive of gastro-oesophageal disease. Nevertheless, Knipschild is right to state that the likelihood ratio in a positive test (acid regurgitation as the dominant symptom) is only slightly above 1 in our sampler other symptoms such as epigastric and retrosternal pain or retrosternal burning would result in likelihood ratios below 1 in such a definition of a "dominant the whole
symptom". Department of Gastroenterology, Medizinische Klinik Innenstadt, University of Munich, D8000 Munich 2, West Germany 1.
A. G. KLAUSER
Hope RA, Longmore JM, Moss PAH, Warrens AN. Oxford handbook medicine, 2nd ed. Oxford: Oxford University Press, 1989: 11.
Regional Renal Unit, General Infirmary at Leeds, Leeds LS1 3EX, UK
RB, Cameron JS, Ogg CS, Bewick M. Why the persistently high mortality in renal failure? Lancet 1972; ii: 75-79. Wheeler DC, Feehally J, Walls J. High risk acute renal failure. Q J Med 1986; 61:
977-84. 3. Beaman M, Tumey JH, Rodger RSC, McGonigle RSJ, Adu D, Michael J. Changing pattern of acute renal failure. Q J Med 1987; 62: 15-23. 4. Tumey JH, Marshall DH, Brownjohn AM, Ellis CM, Parsons FM. The evolution of acute renal failure, 1956-1988. Q J Med 1990; 74: 83-104. 5. Tumey JH, Ellis CM, Parsons FM. Obstetric acute renal failure: 1956-1987. Br J Obstet Gynaecol 1989; 96: 679-87. 6. Guly UM, Tumey JH. Post traumatic acute renal failure, 1956-1988. Clin Nephrol (in
press). WA, Draper EA, Wagner DP, Zimmerman JE. APACHE
disease classification system. Crit Care Med 1985; 13: 818-29.
Extracorporeal membrane oxygenation
SIR,-A 1988 Lancet editorial1 cast doubt on the place of extracorporeal membrane oxygenation (ECMO) in the treatment of severe neonatal respiratory failure. The morbidity of the procedure
SiR,—This question, first posed in 1972,1 continues to worry nephrologists and doctors specialising in intensive care. Survival in acute renal failure (ARF) does not seem to have improved, and may have worsened, since the advent of haemodialysis prevented death from uraemia. Workers have suggested that the failure to improve outcome results from increasing numbers of poor-risk patients2 and a changing case-mix,3 both of which adversely affect prognosis. Investigation of ARF treated by dialysis at Leeds General Infirmary since 1956 has shown that numbers of elderly patients with complicated medical or surgical conditions have risen lately,44 with the virtual disappearance of good prognosis groups such as those in which ARF arises from obstetrics5 or trauma.6 However, patients now and those treated previously appear not to have been directly compared. To test the widely held view that patients with ARF are more seriously ill than formerly, APACHE II severity of disease scores’ have been done on two cohorts of patients dialysed in 1960-69 and 1980-89.
and the lack of controlled clinical trials was criticised. The number of centres providing ECMO in the USA had been growing, and in 1988 over 1000 cases were reported with survival rates of 84-90%.2,3 ECMO is now safer than formerly. The use of double-lumen cannulae can spare the carotid artery and help avoid complications from femoral vein ligation.4 Successful repair of carotid artery has been reported in cases where veno-arterial circulation was required.’ The development of heparin-bonded bypass circuits should decrease morbidity and mortality caused by systemic heparinisation, especially intracranial haemorrhage, and should also permit consideration of ECMO in premature infants. ECMO has also been used more frequently as a cardiac support and as a bridge to lung and heart/lung transplantation.5,6 ECMO is again being used to support adults with acute respiratory failure, and 50-67% survival rates have been achieved/,8The imminent increase in patients with adult acute respiratory distress syndrome (ARDS) and the apparent efficacy of
Why is mortality persistently high in
ECMO should lead us to reconsider some aspects of our conventional management of patients with severe ARDS.9 Experience has demonstrated the safety and feasibility of transporting even the most critically ill to centres specialising in ECMO. In the UK ECMO should be available in a few centres capable of meeting a need of between two and five cases per million population per year. Our cardiothoracic unit with the support of the university department of surgery has been staffed, equipped, and funded by charitable donations to provide an ECMO service. To date we have treated four children, all referred with a 100% predicted mortality, in whom ECMO was used as a last resort. Two children survived; the other two died of complications unrelated to ECMO. Regional Cardiothoracic Unit, Groby Road Hospital,
A. W. SOSNOWSKI T. R. GRAHAM R. K. FIRMIN
Leicester LE3 9QE, UK, and Department of Surgery, University of Leicester
1. Editorial. Persistent fetal circulation and extracorporeal membrane oxygenation. Lancet 1988, ii: 1289-91. 2. Anon. ELSO registry report. Ann Arbor, Michigan: University of Michigan, 1990. 3. Lillehei C, O’Rourke P, Vacanti J, Crone R. Role of ECMO in selected pediatric respiratory problems. J Thoracic Cardiovasc Surg 1989; 98: 968-71. 4. Zwischenberger J, Toomasian J, Drake K, Andrews A, Kolobow T, Bartlett R Total
respiratory support with single lumen cannula veno-venous ECMO: double lumen continuous flow vs single lumen tidal flow. Am Soc Artif Intern Organs 1985; 31: 610-15. 5. Weinhaus L, Canter C, Noetzel M, McAlister W, Spray T. ECMO for circulatory support after repair of congenital heart defects Am Thoracic Surg 1989; 48: 206-12. 6. Egan T, Duffin J, Glynn M, et al. Ten year experience with ECMO for severe respiratory failure. Chest 1988; 94: 681-87. 7. Hickling K. Extra corporeal CO2 removal in severe adult respiratory distress syndrome. Anæsth Intens Care 1989; 14: 46-53. 8. Gattinoni L, Pesenti A, Mascheroni D, et al. Low frequency positive-pressure ventilation with extra corporeal CO2 removal of severe acute respiratory failure.
JAMA 1986; 256: 257-62. 9. Editorial. ARDS time. Lancet
1989; i: 140.
Mefloguine prophylaxis SiR,-Neurological reactions, ranging from dizziness and inability to concentrate to psychotic reactions," after mefloquine treatment and prophylaxis have raised the question of the risks of mefloquine prophylaxis. This is of special concern to doctors who have to recommend prophylaxis for those travelling to regions with multiply resistant Plasmodium falciparum. We report our experience with mefloquine prophylaxis in US soldiers training in Thailand for 6 weeks in 1988. We conducted a double-blind, placebo-controlled study comparing the effect on diarrhoeal disease of prophylaxis with daily
100 mg or weekly mefloquine hydrochloride 250 mg. 119 soldiers were randomised to receive doxycycline and 134 to receive mefloquine. All were healthy men aged 18-40 years (mean age 24). At the end of the fourth and sixth weeks of medication, the soldiers were interviewed for side-effects and blood was obtained serum doxycycline or mefloquine determinations by highperformance liquid chromatography. The frequencies of side-effects were not significantly different between the two groups (table). Nausea and vomiting were reported more frequently by those taking doxycycline, and headache was reported more often with mefloquine; only 1 soldier thought that his symptom (nausea) was attributable to mefloquine. No soldier needed to discontinue mefloquine prophylaxis because of perceived
adverse side-effects. The mean serum mefloquine level was 450 ng/ml (95% confidence interval 417-484) after the fourth weekly dose and 437 REPORTED SIDE-EFFECTS OF MALARIA PROPHYLAXIS
after the sixth dose. Reported side-effects could be related to raised mefloquine levels. 15 soldiers taking doxycycline were potentially non-compliant with their prophylaxis regimen by week 6, as indicated by doxycycline levels or refusal to have blood drawn. Only 7 soldiers were potentially non-compliant with mefloquine (p=0’04); all reported for the side-effect interview but 4 refused to have their blood drawn and 3 had mefloquine levels less than 2 SD below the mean. 4 of the 7 reported diarrhoea; all had confirmed infection with enterotoxigenic Escherichia coli; 2 of the 7 reported no side-effects; and the remaining soldier reported diarrhoea and headache but no causative agent was identified. This is the largest group of US nationals reported to date in whom mefloquine malaria prophylaxis was studied in the tropics. That more soldiers were potentially non-compliant with the doxycycline regimen than the mefloquine regimen, and that no soldier failed to perform during the demanding training period because of what might have been a neurological reaction, indicate that neurological reactions to weekly mefloquine prophylaxis that impair individual performance must be unusual. We are confident that no serious neurological reactions occurred from the weekly mefloquine malaria prophylaxis, although our study was not designed to detect infrequent and subtle neurological events.
ng/ml (404-470) not
JAMES D. ARTHUR Armed Forces Institute of Medical Sciences, Bangkok 10400, Thailand
G. DENNIS SHANKS PETER ECHEVERRIA
B, Bricaire F, Michon C, Franssen G, Lebras J, Bernard J. Mefloquine and brain syndrome. Ann Intern Med 1989; 110: 577-78. 2. Stuiver PC, Ligthelm RJ, Goud ThJLM. Acute psychosis after mefloquine. Lancet 1989; ii. 282 3 Patchen LC, Campbell CC, Williams SB. Neurologic reactions after a therapeutic dose of mefloquine. N Engl J Med 1989; 321: 1415-16. 4. Bjorkman A. Acute psychosis following mefloquine prophylaxis. Lancet 1989; ii: 865 an acute
experience with Japanese encephalitis vaccine
SiR,—Formalin-inactivated Japanese encephalitis vaccine has been shown to be effective in Thai children and has been recommended in the United States for use in certain travellers visiting the Orient.1,2 However, efficacy data in Thailand were derived in a population heavily exposed to flaviviruses (dengue and Japanese encephalitis) which may not be representative of most travellers. To learn about the safety and immunogenicity of the vaccine in a more representative population, we evaluated it in US soldiers. Soldiers were immunised with two doses of monovalent vaccine (NakayamaNIH strain; Biken, Osaka, Japan). Between May, 1987, and August, 1989, 4034 Department of Defense personnel were evaluated. Two doses 1 week apart were given; a small number received a third dose about 6 months later. Soldiers were briefed on the investigational nature of the vaccine and expected side-effects. Questionnaires asking about arm soreness, redness at injection site, febrile episodes, headaches, and rashes were given to all vaccinees 1 to 4 weeks after vaccination and the forms were completed within a week of each dose. 3573 (89%) vaccinees submitted completed forms. No severe adverse reactions occurred. Arm soreness occurred in about one-fifth of vaccinees (table!); this was mild, lasting no more than a few hours, and was not associated with significant swelling, erythema, or limitation of function. Headaches were significantly (p<0-05) more common after the first than after the second dose (13-7% vs 9-8%) whereas redness was slightly more common (p < 0-05) after the second dose (4-5% vs 3-4%). Blacks reported a lower frequency of vaccineassociated febrile episodes (p < 005). Higher rates of soreness, fever, and headaches after either dose were recorded among those who gave a history of yellowfever immunisation in the preceding 10 years. To evaluate
immunogenicity, sera from 27 vaccinees were collected at least three times (at 0, 8, 12, 26, 26-6, and 30 weeks, vaccine being given at times 0, 1, and 26 weeks) and tested for neutralising antibody.3 Antibody prevalence was 0% at time 0, 85%