Many third molar teeth are probably removed unnecessarily in Sandwell

Many climate change scientists do not agree that global warmingis happening

Editor—The apocalyptic tone that Smithadopted in relation to the environmentbears little relation to reality.1 In his editorialSmith asserts, “virtually all scientists agreethat global warming is happening.” Globalwarming is now joining the list of “whateveryone knows.”

Whether most scientists outside clima-tology believe that global warming ishappening is less relevant than whether theclimatologists do. A letter signed by over 50leading members of the American Meteoro-logical Society warned about the policiespromoted by environmental pressuregroups. “The policy initiatives derive fromhighly uncertain scientific theories. They arebased on the unsupported assumption thatcatastrophic global warming follows fromthe burning of fossil fuel and requiresimmediate action. We do not agree.”2 Thosewho have signed the letter represent theoverwhelming majority of climate changescientists in the United States, of whomthere are about 60. McMichael and Haines

quote the 1995 report of the Intergovern-mental Panel on Climate Change (IPCC),which is widely believed to “prove” thatclimate change induced by humans hasoccurred.3 The original draft document didnot say this. What happened was that thepolicymakers’ summary (which became the“take home message” for politicians) alteredthe conclusions of the scientists. This led DrFrederick Seitz, former head of the UnitedStates National Academy of Sciences, towrite, “In more than sixty years as a memberof the American scientific community . . . Ihave never witnessed a more disturbing cor-ruption of the peer-review process than theevents that led to this IPCC report.”4

Policymaking should be guided byproved fact, not speculation. Most membersof the Intergovernmental Panel on ClimateChange believe that current climate modelsdo not accurately portray the atmosphere-ocean system. Measurements made bymeans of satellites show no global warmingbut a cooling of 0.13°C between 1979 and1994.5 Furthermore, since the theory ofglobal warming assumes maximum warm-ing at the poles, why have average tempera-tures in the Arctic dropped by 0.88°C overthe past 50 years?5

Gregory Gardner Locum general practitioner64 Chelworth Road, Birmingham B38 0AE

1 Smith R. Climate change: decision time in Kyoto. BMJ1997;315:1326. (22 November.)

2 Michaels P. Conspiracy, consensus or correlation? Whatscientists think about the ‘popular vision’ of global warm-ing. World Climate Review 1993;1:11.

3 McMichael AJ, Haines A. Global climate change: thepotential effects on health. BMJ 1997;315:805-9.

4 Seitz F. Major deception on global warning. Wall StreetJournal 1996 June 12;section A:16(col 3).

5 Balling RC. Global warming: messy models, decent dataand pointless policy. In: Bailey R, ed. The true state of theplanet. New York: Free Press, 1995:83-107.

BMJ is second most popularjournal among surgeonsEditor—The Editor’s choice of 27 Novem-ber expressed concern about the readershipof the BMJ among surgeons. A recent surveyof the candidates who passed the intercol-legiate specialty examination in general sur-gery in 1997, undertaken by the Intercol-legiate Board in General Surgery, identifiedthe BMJ as the second most regularly readjournal. The journal that was most com-monly read was the British Journal of Surgery.

Candidates may have been reading theBMJ in anticipation of passing the examin-ation and being in a position to apply for jobadvertisements for consultant surgeons. Thesurvey does, however, indicate that the BMJhas a considerable readership amongspecialist registrars in general surgery. Thisreadership would enjoy increased numbersof general surgical articles in addition to thebreadth of material already presented.D A D Macleod ChairmanIntercollegiate Board in General Surgery,Edinburgh EH8 9DR

Preventing RhD haemolyticdisease of the newborn

RhD negative women who haveintrauterine death may need anti-Dimmunoglobulin

Editor—Mayne et al reported evidence ofbenefit from routine antenatal administra-tion of anti-D immunoglobulin to primi-gravid women who are RhD negative.1 It hasbeen proposed that this policy be adoptednationwide. I would like to suggest anothermodification to the current guidelines forRhD negative women who experience anintrauterine death.

Isoimmunisation is a maternal immuneresponse to the passage of fetal red cells intothe maternal circulation. This occurs sponta-neously in the antenatal period. Manyantenatal transplacental haemorrhages passunrecognised clinically, but large transpla-cental haemorrhage may cause fetal death.2

Even large fetal bleeds are generally painfree, and the woman usually presents withabsence of fetal movement alone, thediagnosis not becoming apparent until theresult of a Kleihauer test (for fetal cells in thematernal circulation) is known.

Currently, there is no specific advice forthe care of RhD negative women with anintrauterine death,3 and many if not allmaternity units seem to manage them iden-tically to RhD negative women in normallabour; anti-D immunoglobulin is givenafter delivery (within 72 hours for optimaleffect). Women with an intrauterine deathoften present 24 hours or more after fetalmovement has stopped, and the fetus maynot be delivered for 48 hours, particularly asthe mother often returns home beforeinduction of labour. Given this, the deliverymay not take place until more than 72 hoursafter any lethal transplacental haemorrhage.It seems logical, therefore, that all non-sensitised RhD negative women presenting

Advice to authorsWe receive more letters than we can publish: wecan currently accept only about one third. Weprefer short letters that relate to articlespublished within the past four weeks. We alsopublish some “out of the blue” letters, whichusually relate to matters of public policy.

When deciding which letters to publish wefavour originality, assertions supported by dataor by citation, and a clear prose style. Lettersshould have fewer than 400 words (please give aword count) and no more than five references(including one to the BMJ article to which theyrelate); references should be in the Vancouverstyle. We welcome pictures.

Letters, whether typed or sent by email,should give each author’s current appointmentand full address. Letters sent by email shouldgive a telephone and fax number when possible.We encourage you to declare any conflict ofinterest. Please send a stamped addressedenvelope if you would like to know whether yourletter has been accepted or rejected.

We may post some letters submitted to us onthe world wide web before we decide onpublication in the paper version. We will assumethat correspondents consent to this unless theyspecifically say no.

Letters will be edited and may be shortened.

Letters

1164 BMJ VOLUME 316 11 APRIL 1998 www.bmj.com

with an intrauterine death should have aKleihauer test shortly after diagnosis is con-firmed with scanning and that this isfollowed immediately by an injection ofanti-D immunoglobulin. The result of thetest should be obtained reasonably quickly,and further doses of anti-D immunoglobu-lin can then be given if the result indicatesthat the first dose is insufficient.

Transplacental haemorrhage associatedwith intrauterine death is a rare cause ofsensitisation to the RhD antigen. This altera-tion to current practice would have only asmall impact on the number of women whobecome sensitised to the RhD antigen. How-ever, it is particularly important to actpromptly in women with an intrauterinedeath because any transplacental haemor-rhage causing fetal death would be large andthe chance of sensitisation is greater. And, ofcourse, the woman has already suffered oneobstetric catastrophe.Robert Fox Consultant obstetricianTaunton and Somerset Hospital, Taunton TA1 5DA

1 Mayne S, Parker JH, Harden TA, Dodds SD, Beale JA. Rateof RhD sensitisation before and after implementation of acommunity based antenatal prophylaxis programme. BMJ1997;315:1588. (13 December.)

2 Weiner CP. Fetal death. In: James DK, Steer PJ, Weiner CP,Gonik B, eds. High risk pregnancy: management options. Lon-don: Saunders, 1994.

3 National Blood Transfusion Service ImmunoglobulinWorking Party. Recommendations for the use of anti-Dimmunoglobulin. Prescribers’ J 1991;30:137-45.

4 Bowman JM. Rh immune disease: diagnosis, managementand prevention. In: Droegemueller W, Sciarra JJ, eds.Obstetrics and Gynaecology. Vol 3. Philadelphia: Lippincott,1991.

Guidelines are not followed in many cases

Editor—Routine antenatal prophylaxisagainst rhesus alloimmunisation from silentfetomaternal haemorrhage in first pregnan-cies reduces maternal sensitisation fourfoldprovided that it is combined with appropri-ate application of the guidelines for usinganti-D immunoglobulin to cover events dur-ing pregnancy and after delivery.1 The factthat Mayne et al achieved almost 100%cover within two years reflects considerableorganisation and cooperation by allinvolved. The authors’ paper was optimisticand forward looking, and its subtext was thatincreasing uptake of routine prophylaxisincreases awareness of the need to useanti-D immunoglobulin in antepartumhaemorrhage. It is this aspect of educationand motivation that is particularly impor-tant in the prevention of rhesus disease.

An audit of 922 rhesus negative womenin Mersey showed that potentially avoidablealloimmunisation arose from failure toobserve the guidelines in 39% of cases.2

Bleeding after the first trimester was treatedwith anti-D immunoglobulin in 141 of 171cases, but in 25 women who bled after 20weeks’ gestation only 250 IU was given,rather than the recommended 500 IU. Also,abdominal trauma was not realised to be anindication for anti-D immunoglobulin in 19of 24 cases.

The need for continuing education of allhealth professionals who care for pregnantwomen was shown by this and by misinter-pretation of Kleihauer tests, a negative result

frequently being wrongly interpreted tomean that prophylaxis was not necessary.Positive results were also misinterpreted.Among 14 women with a fetomaternalhaemorrhage greater than 4 ml the addi-tional dose of anti-D immunoglobulin indi-cated was not given in five. Misunderstandingof the rationale behind postnatal prophylaxisled to this being omitted in five women whohad been given anti-D immunoglobulin latein pregnancy to cover antepartum haemor-rhage because it was thought that this alsocovered labour and delivery.

The current guidelines can be madeeffective by audit of their use, and the intro-duction of routine prophylaxis will reducerhesus disease. Since a shortage of anti-Dimmunoglobulin makes widespread appli-cation of routine antenatal prophylaxisunlikely until the 21st century, the properuse of updated guidelines with frequentreminders and education of those enteringobstetric care is still the way forward. Thisshould be extended to include those in acci-dent and emergency departments and allother healthcare professionals required totreat pregnant women.Iain R McFadyen Senior research fellowDepartment of Obstetrics and Gynaecology,Women’s Hospital, Liverpool L8 7SS

Vanessa J Martlew Consultant haematologistHelen L Howard Clinical assistant in transfusionmedicineNational Blood Service, Mersey and North WalesCentre, Liverpool L7 8TW

Cyril A Clarke Emeritus professor of medicineDepartment of Biological Sciences, University ofLiverpool, PO Box 147, Liverpool L69 3BX

1 Mayne S, Parker JH, Harden TA, Dodds SD, Beale JA. Rateof RhD sensitisation before and after implementation of acommunity based antenatal prophylaxis programme. BMJ1997;315:1588. (13 December.)

2 Howard HL, Martlew VJ, McFadyen IR, Clarke CA.Preventing rhesus(D) haemolytic disease of the newbornby giving anti-D immunoglobulin: are the guidelines beingadequately followed ? Br J Obstet Gynaecol 1997;104:34-41.

Women should be counselled aboutsource of anti-D immunoglobulin

Editor—Van Dijk’s editorial on preventingRhD haemolytic disease of the newborn anda consensus conference on anti-D prophy-laxis last year clarify the need for continuedreliance on donors’ blood to prevent rhesusisoimmunisation.1 2 I am concerned aboutthe lack of consumer involvement in andawareness of the programme and theproposed extension to the antenatal periodfor women without obvious sensitisingstimuli.

My concern led to an ad hoc survey ofwomen who had received anti-D immu-noglobulin in the past and of rhesusnegative antenatal women who were sched-uled to receive anti-D immunoglobulin iftheir neonates were detected to be rhesuspositive. None of these 12 women knew thatanti-D immunoglobulin was a blood prod-uct. They had presumed this to be eithersynthetic or derived from some othernatural source. I wonder whether all health-care givers involved in prescribing andgiving the immunoglobulin realise that it isprepared from paid donors’ blood.

It is imperative that counselling of preg-nant women about rhesus isoimmunisationshould include telling them about thesource of anti-D immunoglobulin and thesmall possibility of acquiring bloodborneinfections after such injection. This infor-mation is vital, because introduction ofmonoclonal anti-D immunoglobulin is notimminent.Arabinda Saha Consultant obstetrician andgynaecologistNorth East Lincolnshire NHS Trust, GrimsbyDN33 2BA

1 Van Dijk B. Preventing RhD haemolytic disease of thenewborn. BMJ 1997;315:1480-1. (6 December.)

2 Consensus statement from the Royal College of Physiciansof Edinburgh/Royal College of Obstetricians and Gynae-cologists consensus conference on anti-D prophylaxis7 and 8 April 1997. Transfus Med 1997;7:1434.

New recommendations must beexplained to GPs and midwives

Editor—The revised guidelines on givinganti-D immunoglobulin for RhD prophylaxiswere outlined by van Dijk in his editorial andcame from a consensus conference held inEdinburgh in April 1997.1 It was agreed thatcompliance with existing guidelines about theuse of anti-D immunoglobulin must beimproved to reduce the number of isoimmu-nised women further. Failure is certainlyassociated with a lack of information inprimary care. Unfortunately, the previousguidelines in 1991 were given no publicityand as a result the information never reachedgeneral practitioners and midwives. How doour expert advisers intend informing generalpractitioners and midwives this time abouttheir recommendations to give anti-D immu-noglobulin in the last trimester of pregnancy?

We urgently need a fresh approach bythe blood transfusion experts to educate allthose who care for pregnant women,especially now that there is additional adviceabout antenatal prophylaxis. Women willneed to understand why some of them(those who are RhD negative) will requireantenatal anti-D immunoglobulin andothers (those who are RhD positive) will not.

We spoke at the annual scientific meetingof the British Blood Transfusion Society inWarwick in September 1997 about thecurrent lack of information and the need toinvolve women in their own care. Weexpressed the view that a multidisciplinaryworking party, including representatives fromRoyal College of General Practitioners andRoyal College of Midwives, should be set upurgently to recommend how to explain thenew recommendations. Unfortunately, noth-ing has happened. We support the advice togive anti-D immunoglobulin during the thirdtrimester of pregnancy, but if there is no con-sultation and collaboration with primary careand midwifery services the potential benefitsof the new recommendations could be lostonce again.Christopher Everett General practitionerAlton Health Centre, Alton, Hampshire GU34 2QX

Julie Wray Research fellow-midwifeRoyal College of Obstetricians and GynaecologistsAudit Unit, St Mary’s Hospital, Manchester M13 0JH

1 Van Dijk B. Preventing RhD haemolytic disease of thenewborn. BMJ 1997;315:1480-1. (6 December.)

Letters

1165BMJ VOLUME 316 11 APRIL 1998 www.bmj.com

Medline and Embasecomplement each other inliterature searchesEditor—Greenhalgh briefly mentions someimportant differences between the Medlineand Embase databases.1 To compare theusefulness of the databases in the druginformation setting we carried out aprospective study of all requests for druginformation that required a literature searchduring July and August 1996. For eachinquiry both databases were searched backto 1992 with current CD ROMs. Standardsearch strategies were used, based onindexed subject headings and textwords.Relevant articles were selected on the basisof the information in the abstract title orsubject headings, or both. When an articleseemed to be unique to a database this wasverified by cross checking in the other data-base, using an author or journal namesearch.

For 11 of 32 questions no relevant infor-mation was retrieved by either database,leaving 21 questions for comparison. Thesame articles were retrieved by bothdatabases for only four questions. Embaseretrieved articles that were unique to thatdatabase for 14 questions, while Medlineretrieved unique articles for two questions.For one question both databases retrieved aunique article.

Embase was able to answer adequately20 of the 21 questions, whereas Medlineanswered only 17 adequately; both data-bases were required to answer all 21questions. Articles in Embase are assignedmore index terms than those in Medline,and consequently people using Embase maybe less likely to miss an important article butmust spend more time browsing throughirrelevant material.

The overlap of journals in the two data-bases is only about 34%,2 and other studieshave shown that using them both improvesthe coverage of the literature.3–5 Despite this,Embase seems to be underused. On theassumption that the words “Embase” and“Medline” would be used in many publica-tions to describe the search strategy, we did atextword search of Medline (1993 to August1997) and retrieved 66 and 1669 hitsrespectively.

Our study shows some important differ-ences between the databases. Firstly,although Medline adequately answeredmost of the questions, a considerablenumber could be answered only by usingEmbase. Secondly, Embase complementsMedline and vice versa. Finally, to retrieveimportant articles that are not indexed inMedline it is essential to use Embase insearches done for a comprehensive reviewor to find rare case reports. Embase is moreexpensive, more time consuming to use, andperhaps less accessible than Medline. Webelieve, however, that all practitionerssearching for information about drugs andtherapeutics should use Embase whenMedline has not retrieved sufficient infor-

mation or when more comprehensivecoverage of the literature is required.David Woods Lecturer in drug information andclinical pharmacyKate Trewheellar Pharmacy studentSchool of Pharmacy, University of Otago,PO Box 913, Dunedin, New [email protected]

1 Greenhalgh T. How to read a paper: the Medline database.BMJ 1997;315:180-3. (19 July.)

2 Smith B, Darzins P, Quinn M, Heller RF. Modern methodsof searching the medical literature. Med J Aust 1992;157:603-11.

3 Odaka T, Nakayama A, Akazawa K, Sakamoto M,Kinukawa N, Kamakura T, et al. The effect of a multiple lit-erature database search. A numerical evaluation in thedomain of Japanese life science. J Med Syst 1992;16(4):177-81.

4 Biarez O, Sarrut B, Doreau CG, Etienne J. Comparison andevaluation of nine bibliographic databases concerningadverse drug reactions. Drug Intell Clin Pharm 1991;25:1062-5.

5 Barillot MJ, Sarrut B, Doreau CG. Evaluation of drug inter-action citation in nine on-line bibliographic databases. AnnPharmacother 1997;31:45-9.

What investigations andprocedures do patients inhospices want

Providing ineffective treatment is futile

Editor—In their short report Meystre et alsay that “legal advice suggests that if patientsrequest resuscitation it should be provided,”1

referring to our study on the views of elderlypatients on cardiopulmonary resuscitation.2

We contacted the Medical Protection Societyand the Medical Defence Union by tele-phone during our study because a patient inthe study who had metastatic lung cancer“demanded” cardiopulmonary resuscitationif he had a cardiac arrest. We were advisedthat cardiopulmonary resuscitation shouldbe provided if a patient requested it sincefailure to do so could result in litigation.Subsequently, the Medical Protection Soci-ety and the Medical Defence Union said thatthe advice they had given was specific to thecase cited and could not be generalised, andthat they would not agree with the assertion“legally if a patient requests cardiopulmo-nary resuscitation it should be provided.”3

We agree with this as providing treatmentthat is ineffective is futile.

Cardiopulmonary resuscitation is a use-ful treatment for some patients. “Do notresuscitate” orders should be made inaccordance with the resuscitation policiesthat all hospitals should have in place.4 Itmay be necessary to include patients in thediscussion but it is not appropriate to do soif cardiopulmonary resuscitation is deemedmedically futile, unless the patient initiatesthe discussion.5 If a patient requests cardio-pulmonary resuscitation when it is inappro-priate, discussion must show compassionand reasoning, which requires good com-munication skills. Doctors must rememberthat they are under no legal obligation toprovide cardiopulmonary resuscitation ifpatients or relatives demand it. They do,however, have to provide an explanation fortheir medical decision.

Undergraduate and postgraduate edu-cation programmes often omit to teach

when and how to discuss cardiopulmonaryresuscitation with patients. We have intro-duced regular tutorials to our preregistra-tion house officers so they can acquire theseskills. Unfortunately, the decisions aboutthese procedures are made by more seniordoctors, who have often had no experienceor training in this field.Rosemary Morgan Consultant physician in geriatricmedicineDebra King Consultant physician in geriatric medicineDepartment of Medicine for the Elderly, WirralHospital, Upton, Wirral, Merseyside L49 5PE

1 Meystre CJN, Burley NMJ, Ahmedzai S. What investiga-tions and procedures do patients in hospices want? Inter-view based survey of patients and their nurses. BMJ1997;315:1202-3. (8 November.)

2 Morgan R, King D, Prajapati C, Rowe J. Views of elderlypatients and their relatives on cardiopulmonary resuscita-tion. BMJ 1994;308:1677-8.

3 Wall JA, Palmer RN. Resuscitation and patients’ views. BMJ1994;309:1442-3.

4 Doyal L, Wilsher D. Withholding cardiopulmonaryresuscitation: proposals for formal guidelines. BMJ1993;306:1593-6.

5 Stewart K. Discussing cardiopulmonary resuscitation withpatients and relatives. Postgrad Med J 1995;71:585-9.

Patients with untreatable cancer almostnever survive cardiopulmonaryresuscitation

Editor—Meystre et al say that if terminallyill hospice patients request cardiopulmo-nary resuscitation then there is a legalobligation to provide it.1 They cite a study byMorgan et al to support this view,2 and,although Morgan et al were quoted accu-rately, they themselves had misinterpretedadvice which they had been given by thedefence organisations.3 In subsequent corre-spondence senior officers of the defenceorganisations denied that this would be theiradvice, saying that “it is not correct generallyto assert that ‘legally if a patients requestscardiopulmonary resuscitation it should beprovided.’ ”3 They also chastised journal edi-tors for publishing incorrect advice withoutfirst checking the source. A more compre-hensive search of the literature might haveprevented this inaccurate claim beingrepeated.

Meystre et al’s report raises a widerquestion about the use of cardiopulmonaryresuscitation in patients with untreatablecancer. We would be interested to know ifthe patients interviewed for this survey weretold that those with advanced cancer almostnever survive resuscitation attempts,4 sincethis information would have been likely toinfluence their decisions. In our hospital weaudited casenotes of all 264 patients havingcardiopulmonary resuscitation over 19months. Twenty eight of these patientssurvived and were discharged (unpublisheddata). Twenty of our patients had cancer, andin six it was metastatic. Of the patients withcancer only one, who had non-metastaticdisease, survived until hospital discharge,and this patient died in a nursing home fourdays later. Our local experience is thereforein keeping with previous reports,4 and weregard advanced cancer as a condition forwhich resuscitation is futile.

Patients entering hospice care areusually not going to benefit from attempts atcurative surgery, which is, correctly, not

Letters


offered as a treatment option. The sameprinciple applies to cardiopulmonary resus-citation. Doctors have neither a moral nor alegal obligation to offer treatment that isineffective,5 even if it is requested by patients.Kevin Stewart Consultant physicianMike Bacon Specialist registrarDepartment of Medicine and Elderly Care, RoyalHampshire County Hospital, WinchesterSO22 5DG

Lesley Bowker Senior registrarRadcliffe Infirmary, Oxford OX2 6HE


2 Morgan R, King D, Prajapati C, Rowe J. Views of elderlypatients and their relatives on cardiopulmonary resuscita-tion. BMJ 1994;308:1677-8.

3 Wall JA, Palmer RN. Resuscitation and patients’ views. BMJ1994;309:1442-3.

4 O’Keeffe S, Ebell MH. Prediction of failure to survivefollowing in-hospital cardiopulmonary resuscitation: com-parison of two predictive instruments. Resuscitation1994;28:21-5.

5 Doyal L, Wilsher D. Withholding and withdrawing life sus-taining treatment from elderly people: towards formalguidelines. BMJ 1994;308:1689-92.

Doctors are under no legal obligation toprovide ineffective treatment

Editor—Meystre et al report that patientswho are terminally ill are prepared to acceptinvasive procedures more readily than theirnurses are.1 The article points out that caremust be taken to ensure that patients are notdenied the opportunity of therapeutic inter-ventions because of staff attitudes, but I thinkseveral of Meystre et al’s conclusions aremisleading.

I believe that the difference in attitudesexists not because nurses wish to deny theirpatients benefit from treatment but becausenurses are more aware of the true—andlimited—benefit of some of the interventionsproposed. To ask a patient with advancedcancer “If your heart stopped unexpectedlywould you want to be resuscitated?” withoutexplaining the likely outcome of thatresuscitation does not provide useful infor-mation. It has been shown that patientsacquire their knowledge of cardiopulmo-nary resuscitation predominantly from tele-vision and that most patients think that it is“a mostly successful procedure.”2 3 Studieshave shown, however, that cardiopulmonaryresuscitation of patients with advancedmalignancy almost invariably fails tore-establish cardiopulmonary function. Atbest it succeeds only to result in further car-diopulmonary arrest with no interveninghospital discharge.4 It is therefore inappro-priate in most of the hospice population.

Cardiopulmonary resuscitation is themost extreme example of the interventionsproposed. Other questions in the surveywere biased by the preface that medical staffthought that a particular intervention wouldimprove the patients’ care. The article doesnot say whether the patients were given anyindication of the nature or degree of thatimprovement. The nursing staff wouldcertainly have had that knowledge whenanswering the same questions.

In response to Meystre et al’s concludingparagraph I would emphasise that equal

care must be taken not to subject patients tointerventions of negligible or limited benefiton the basis of poorly informed choicesmade by them.Valerie O’ Donnell Senior registrar in palliativemedicineAintree Palliative Care Team, Fazakerley Hospital,Liverpool L9 7AL


2 Liddle J, Gilleard C, Neil A. The views of elderly patientsand their relatives on cardiopulmonary resuscitation. J RColl Physicians 1994;28:228-9.

3 Mead JE, Turnbull C. CPR in the elderly: patients’ andrelatives’ views. J Med Ethics 1995;21:39-44.

4 Faber-Langendon K. Resuscitation of patients withmetastatic cancer. Is transient benefit still futile? Arch InternMed 1991;151:235-9.

Questions to patients and nurses shouldhave been identical

Editor—Meystre et al compared answers onthe acceptability of invasive procedures andtreatments given by patients with cancerwith answers given by their nurses.1 We needto be much more aware of the different per-spectives of clinicians and patients, particu-larly when we consider how, and by whom,difficult choices and decisions about treat-ment should be made.

A flaw in the study, however, makes it dif-ficult to draw useful conclusions from thefindings. The patients and nurses were notasked the same questions, and so theiranswers cannot be compared meaningfully.Patients were asked to rate how much theywould want procedures that staff thoughtwould help to improve their care. Nurseswere asked to rate how appropriate it wouldbe to carry out the procedures. These arequite different considerations, which reflectthe distanced position held by professionalsin the the healthcare system—a position thatrequires impartiality and a recognition of rea-sonable limits—and the immediacy of theindividual patient’s predicament—whichmeans partiality and a concern that someonewill go to any lengths on the patient’s behalf.

The views of nurses and patients on theacceptability of procedures might have beenless divergent if they had been asked thesame questions, and even less so if the ques-tions had been reversed. Nurses may want todo all (if it were possible) that could be doneto improve patient care, while patients mayacknowledge (with regret) the constraintsimposed by their illness and by the clinicalcontext.

Gilligan differentiated ethic of justicefrom the ethics of care.2 In accordance withthis, further research and clinical practiceshould take into account the differing ethicalstances of those with greater and those withlesser power in the clinical situation.Annie Mitchell Lecturer in clinical and communitypsychology.Department of Psychology, University of Exeter,Washington Singer Laboratories, Exeter EX4 4QG


2 Gilligan C. In a different voice:psychological theory and women’sdevelopment. Boston, MA: Harvard University Press, 1982.

Responses are not comparing like withlike

Editor—Meystre et al investigated theattitudes of terminally ill patients in ahospice towards investigations and proce-dures.1 I am concerned about the validity oftheir results.

The responses from the nurses may wellhave been influenced by their professionalknowledge. Most medical and nursingprofessionals would agree to any investiga-tion if it were “thought necessary for themedical management of their patient.” Itherefore think that in fact the nurses wereanswering the question “How appropriatedo you feel it would be to carry out theseinvestigations?”

Patients who were offered investigationsor treatment that were thought necessary forthe medical management of their diseasewould clearly be more likely to consent tothem. The responses are therefore not com-paring like with like.Charles Daniels Macmillan consultant in palliativecareCraig Gannon consultant in palliative careMacmillan Team, Hounlow and SpelthorneCommunity and Mental Health NHS Trust, WestMiddlesex Hospital, Isleworth, Middlesex TW7 6AF.


Authors’ reply

Editor—Morgan and King pointed out thatwe were wrong1 and there is not always alegal obligation to provide cardiopulmonaryresuscitation if a patient requests it, and weagree with Stewart et al that cardiopulmo-nary resuscitation in patients with advancedcancer is usually futile.

Patients were given brief descriptions ofeach option available, and for cardiopulmo-nary resuscitation this included “Resuscita-tion involves calling two or three doctorsurgently . . . laying a patient flat, using a bagand a mask applied closely to the face, . . .pressing down hard rhythmically on thechest. . . . A drip is put up and drugs injected. . . the breathing mask would be replaced bya tube down into the lungs.. . . Attempts atresuscitation are not very successful but theoutcome depends on how ill the person isbefore the heart stops and how quicklyresuscitation can be started.” We believe thisis even handed and not overoptimistic.

The comments on the futility of cardio-pulmonary resuscitation are based on theassumption that all hospice patients withcancer are irreversibly near death. Manyhospices take patients earlier in the courseof their disease and offer more invasivemedical interventions, such as endoscopy,paracentesis, or spinal analgesia. Some ofthese patients may be put at risk of suddencardiac arrest, and it seems reasonable tooffer cardiopulmonary resuscitation tothose whose life expectancy otherwise couldbe several months.

We did not ask patients and nursesexactly the same question because we werenot measuring whether nurses knew what

Letters


their patients would say (as we did regardingperformance status) but how far nursing wasinfluenced by patient autonomy. Therefore,we asked to what extent, if “we thought itwould improve your care,” individualpatients would give consent. A nurse whoknew the patient well was asked “Howappropriate do you think it would be to dothe following tests if the need arose duringthe management of your patient?” The lackof agreement on eight of 14 issues (thenurses always scored below the patients)may show, as Daniels says, that “the nursesmay well have been biased by theirprofessional knowledge.” It would be a pity ifthe autonomy of patients, freed from the“paternalism” of an earlier generation ofdoctors, was limited by the “maternalistic”bias of well meaning nurses, makingdemedicalised care the only option whenpatients themselves would opt for interven-tion. We would welcome larger, preferablymulticentre, studies to confirm or refutethese findings.Chantal Meystre Consultant in palliative medicineMary Ann Evans Hospice, George Eliot Hospital,Nuneaton CV10 7DJ

Neil M J Burley Senior house officer psychiatrySt Michael’s Hospital, Warwick, CV34 5QW

Sam Ahmedzai Professor of palliative medicine,Department of Surgical and Anaesthetic Sciences,University of Sheffield, Royal Hallamshire Hospital,S10 2JF

1 Ahmedzai S, Mallett K. Brooks D, Johnson IS. What dohospices do? (abstract). Palliat Med 1997;11:69-70

Culyer reforms will create newopportunities for researchEditor—The role of director of researchand development in a teaching hospital is achallenging one, particularly during theintroduction of the Culyer reforms. It doesnot include discouraging clinical staff whoare industrious in their service work, enjoyteaching, and are active in research fromcontinuing to develop evidence that canlead to improved clinical practice.1 Truststhat wanted to obtain money from the levyfund that was established to supportresearch and development had to satisfy theNHS Executive that they would, firstly,implement a corporate research strategy;secondly, prospectively review the quality ofproposed studies; thirdly, rigorously managesupport costs for research; and, fourthly,encourage but not subsidise commercialresearch. These obligations are the result ofa major evidence based reform of NHSfinance and translate directly into themessages from the author’s anonymousdirector of research and development thatwere quoted in the article.

Under the new support arrangementsfor research, clinicians are subject to ahigher level of research management, andtheir projects, properly peer reviewed andcosted, must be placed in a coherentstrategic framework. These changes aredeemed to be inimical to clinical research,but is this true? By rising to this challenge we

were able to put forward a competitive bidfor levy funds. Its success will allow us tostrengthen the research support that isavailable to our clinical researchers, therebyenhancing their ability to compete for grantfunding and further improving the highquality of their research output.

Although consultants in this trust thinkthat applying for grants is futile and believethat funding bodies favour basic science, theevidence contradicts this. Many sponsorsfund research into health services andassessment of health technology across therange of clinical specialties. Successful appli-cants show that authoritative proposals canbe put forward in any area of health care.Satisfactory scientific peer review should bea precondition for ethical approval.

The Culyer reforms will clarify howNHS resources support research and devel-opment, improve value for money, andwiden participation in clinical research.Some clinicians, whose pilot studies arenever followed up and whose substantivestudies are too small to inform clinical prac-tice or unnecessarily duplicate earlier work,will feel discouraged by the new environ-ment in clinical research. They will probablybe unwilling to change their established pat-tern of research and will therefore not comewithin the new support framework. I believe,however, that most will welcome the oppor-tunities created by a system that forcesresearchers to articulate their research plansexplicitly, encourages multicentred andmultidisciplinary studies, and demands thatfellow professionals respond to rigorousevidence.Clark Lewis Director of research and developmentKing’s College Hospital, London SE5 9RS

1 Conflict [personal view]. BMJ 1998;316:80-1. (3 January.)

Clinician scientists face difficultfutureEditor—Bitter criticism has been levelled atthe Calman changes,1 but nothing has beensaid of the 100 or so people in Britain whohold clinician scientist fellowships.2 Onlyfour funding bodies award these fellowships:the Medical Research Council, WellcomeTrust, British Heart Foundation, and CancerResearch Campaign. Together they providearound £30 million of support for theseindividuals. The success rate for applicantsin these competitive fellowships is consist-ently only 25%. These awards, universallyregarded as successful and cost effective,provide for three to seven years of personalfunding. Holders of the awards ultimatelyprogress to senior fellowships or full facultypositions.

The Calman changes have introduced arigid system requiring these individuals tofulfil inflexible criteria; the changes take noaccount of the fact that they ultimately willbe engaged principally in research, provid-ing only a minimal, and often superspecialty,clinical service. The absurdity of thissituation is that clinician scientists could

spend six years (after obtaining the MRCP)in research, followed by a further four to fiveyears in full time clinical training. Worse still,they spend part of their week on the wardsand part on the bench, achieving neitherintended goal satisfactorily.

As Calman deserts a sinking academicship for greener pastures, those he leavesbehind may begin to wonder what, if any,future there is in pursuing an academiccareer in the United Kingdom. The fundingbodies are trying to be as flexible as possible,while others apply pressure on the commit-tees that award specialist certificates. But thisis not a time for quiet diplomacy. Two of thefour funding bodies are already concernedabout the reduction of inquiries about thesefellowships (personal communication), andsome award holders are thinking aboutwithdrawing from awards already obtainedin order to return to the NHS to fulfil clinicalrequirements—a damning indictmentagainst the Calman changes.

So what is to be done? Clinicianscientists should automatically be awardedwith a training number in the specialty oftheir choice,3 and trusts should be requiredto develop fast track training programmessufficient for accreditation. Postgraduatedeans, who have their own interpretation ofthe rules as applied to academics, seem to beparty to unfair regional discrepancies; someseem to be more accommodating than oth-ers. They not only have the power to helpthese individuals but a duty and responsibil-ity to do so. If clinician scientists cannot begiven latitude in the new system then no oneis safe.Pankaj Sharma BHF clinician scientistClinical Pharmacology Unit, University ofCambridge, Addenbrooke’s Hospital, CambridgeCB2 2QQ

1 Barber P. The colleges, Calman, and the new deal. Lancet1997;350:974.

2 Chaudhry B, Winyard P, Cale C. Academia: the view frombelow. BMJ 1997;315:560-1.

3 College of Vice Chancellors and Principals. Clinicalacademic careers. Report of an independent task force. London:CVCP, 1997.

Early discharge after stroke

Use of Barthel index was not appropriate

Editor—There are three reasons why Ruddet al may be at fault when they conclude thatearly discharge of patients with stroke hasno clinical importance.1

Firstly, the authors used the Barthelindex as the main outcome measure. Thisscale is insensitive to change at higherscores.2 More than 1 in 10 of the patientshad a maximum Barthel score at randomi-sation, and improvement therefore cannotbe shown in these patients with this index.The study was powered to detect a 3.5 pointdifference in the Barthel index. To be able todetect such a difference, the Barthel scoreshould have been <16.5 at the start of thestudy. Only about two thirds of the patients,however, seem to have had this low score, soa difference of this magnitude is unlikely tohave occurred. Moreover, the most probable

Letters


effects of early discharge are on the psycho-social level3 and thus are not measured bythe index.

Secondly, outcomes were assessed afterone year. Most functional improvement afterstroke takes place during the first fewmonths.4 Different treatment strategies mayinfluence the speed of this, and such effectsmay be clinically important even if they dis-appear within a year.

Thirdly, nearly two weeks elapsed beforehalf the patients had been randomised, andsome were not randomised until consider-ably later. Thus the authors had noopportunity to start the intervention—thatis, prepare for early discharge—during thefirst important weeks.5

The study provides important thoughinsufficient information for clear conclu-sions to be drawn on the role of earlydischarge. Such conclusions should prefer-ably be based on several studies withcommon, appropriate outcome measures.5

Torgeir Bruun Wyller Research fellowDepartment of Geriatric Medicine, UllevaalHospital, N-0407 Oslo

1 Rudd AG, Wolfe CDA, Tilling K, Beech R. Randomisedcontrolled trial to evaluate early discharge scheme forpatients with stroke. BMJ 1997;315:1039-44. (25 October.)

2 Wellwood I, Dennis MS, Warlow CP. A comparison of theBarthel index and the OPCS disability instrument used tomeasure outcome after acute stroke. Age Ageing1995;24:54-7.

3 Seale C, Davies P. Outcome measurement in strokerehabilitation research. Int Disabil Studies 1987;9:155-60.

4 Jørgensen HS, Nakayama H, Raaschou HO, Vive-Larsen J,Støier M, Olsen TS. Outcome and time course of recoveryin stroke. Part II. Time course of recovery. TheCopenhagen stroke study. Arch Phys Med Rehabil1995;76:406-12.

5 Rodgers H, Soutter J, Kaiser W, Pearson P, Dobson R, Skil-beck C, et al. Early supported hospital discharge followingacute stroke: pilot study results. Clin Rehabil1997;11:280-7.

Authors’ reply

Editor—We agree that the Barthel indexhas considerable limitations as an outcomemeasure for stroke. We chose it because it isthe most widely used outcome measure incomparable trials and because high comple-tion rates are attainable, with results beingvalid whether it is completed by the patients,carers, or researchers. We were seeking todetect a group difference of 3.5 points onthe Barthel index rather than this differencein each individual. The intervention wasaimed primarily at physical rather than psy-chosocial rehabilitation, and we thereforeexpected that this would be reflected in theoutcomes. Nevertheless, we measured a widerange of outcomes, including those coveringpsychosocial function, none of whichshowed significant differences between thetwo groups.

We chose one year as the key outcomepoint, arguing that sustained rather thantransient effects were more important toachieve. Outcomes were measured at two,four, and six months; preliminary analysis ofthese showed no significant differencesbetween the groups, even in the early phasesof recovery.

Within a trial setting it would not havebeen feasible to prepare patients for earlydischarge when half of them would subse-

quently end up receiving conventional care.We accept that greater benefits could poten-tially have been achieved, particularly inlength of stay, if the patients had not had towait to reach the randomisation pointsbefore discharge was planned.

As this is the first full trial report on earlydischarge after stroke we must wait forreports of other trials in which similaroutcome measures are used. These couldthen be combined to answer other key ques-tions, such as the role of early discharge fordifferent subgroups of patients.Anthony Rudd Consultant physicianElderly Care Unit, Guy’s and St Thomas’s NHSTrust, London SE1 7EH

Charles Wolfe Senior lecturerDepartment of Public Health, UMDS, St Thomas’sHospital, London SE1 7EH

Kate Tilling LecturerRoger Beech Senior research fellowCentre for Health Planning and Management,Keele University, Keele, Staffordshire ST5 5BG

Basing resource allocationformulas on standardisedmortality ratios would bewrongEditor—Sheldon proposes that resourceallocation formulas should revert to themethods proposed by the resource alloca-tion working party over 20 years ago.1 Hethinks that instead of messing around withcomplicated statistics we should use thestandardised mortality ratio (although hedoes not specify which variant2). It is anappealing simplification, but it won’t do forfour reasons.

Firstly, because allocations are having tobe made to much smaller populations, weneed accurate and robust formulas andguidance on where discretion is appropriateand what the likely sources of bias are; roughapproximations are worse than useless.

Secondly, up to date measures arerequired. Although standardised mortalityratios can be calculated annually, most of thedeaths concerned—except for a smallnumber of accidental and avoidabledeaths3—are a culmination of vulnerabilityover a long period. Annual movementsprobably reflect more the transient patternof migration of elderly people than changesin underlying social conditions. Whateverthe arguments in favour of “making do” withstandardised mortality ratios as an alterna-tive to proxies for material deprivationbased on the census, the ratios seem unlikelyto be more than a poor substitute for directmeasures of poverty (such as uptake of ben-efits) that are now becoming available.4

Thirdly, the presumption that the stand-ardised mortality ratio is the best all roundproxy is based on the view that it reflects thebundle of factors that lead to use of healthservices. But if there is one thing that hasbeen learnt over the past 15 years aboutwhether inequalities in health betweensocial classes have increased, it is thatarguments phrased in terms of overall death

rates are of little value, because differentcomponents of the standardised mortalityratio have different aetiologies.5

Fourthly, resorting simply to the stand-ardised mortality ratio means excluding thepossibility that material deprivation influ-ences use of health services other thanthrough the impact it makes on morbidityand mortality.

Twenty years ago most deaths wererelated to poverty so that an overall indexmore or less reflected the distribution ofpoverty in a population.5 Nowadays thestandardised mortality ratio reflects a mix ofoften contradictory tendencies in terms ofliving standards and diseases related tolifestyle. Knowing an area’s standardisedmortality ratio tells us much less than it usedto. Statements about the best way toapproach resource allocation should bebased on evidence, not just on appeals tosimplicity.Roy A Carr-Hill Senior research fellowCentre for Health Economics, York University, YorkYO1 5DD

James Q Jamison DirectorHealth and Health Care Research Unit, Queen’sUniversity of Belfast, Belfast BT12 6B

1 Sheldon TA. Formula fever: allocating resources in theNHS. BMJ 1997;315:964. (18 October.)

2 Carstairs V, Morris R. Deprivation and health in Scotland.Aberdeen: Aberdeen University Press, 1991.

3 Charlton JRH, Hartley RM, Silver R, Holland WW.Geographical variation in mortality from conditions ame-nable to intervention in England and Wales. Lancet1983;i:691-6.

4 Jamison J, O’Reilly D, Steveson M, Reid J, Merriman B,Carr-Hill RA. A model to predict acute hospital utilisationuptake of social security benefits and low birth weight[abstract]. J Epidemiol Community Health 1997;51:600.

5 Illsely R, Baker D. Inequalities in health: adapting thetheory to fit the facts. Bath: University of Bath, 1997. (Socialpolicy papers No 26.)

Are part time doctors betterdoctors?

Doctors need flexible training andflexible jobs

Editor—Gibson raised some serious pointsfor further discussion about part time work-ing in career grades.1 Kumar contributesnothing to an informed debate on this topic,and his argument runs counter to the viewsof respected national bodies such as theRoyal College of Physicians.2 He states: “Parttimers are no good for any medical specialtywhere there may be a need for quickaction. . . .” He says that he works 50 hours ofthe 168 in a week. This means that, for theremaining 118 hours, someone from histeam has to be available, or there is no oneto undertake the responsibilities. A teamprovides continuity of service to patients,and part timers can be highly effectivemembers of a team.

Overall, some 8% of specialist registrarsare training on a flexible basis; in Oxford,Kumar’s region, this proportion is 15.3%(Department of Health, annual medicalworkforce census). It is therefore unfortu-nate that he should seek a confrontation atthe very time when flexible and full timetraining are so much more closely inte-

Letters


grated and flexible training leads success-fully to consultant posts. The same con-straints that lead doctors to ask for flexibletraining are often still present when theyseek a career grade post, and these doctorswill seek part time consultant posts.

In some specialties, including generalpractice, the shortage of doctors to deliverthe service is severe. We need to facilitatetraining and career grade opportunities foryoung doctors who might otherwise leave,temporarily or permanently, by providingflexible (part time) opportunities. As Allensays: “. . . a radical assessment is needed ofhow we are to make use of the talents of thebrightest and best of successive generationsof young people who enter medicine.Assumptions still predicated on a medicalworkforce made up of men working full timemainly in one specialty for 40 years arehopelessly misguided. . . .”3

Ilfra Goldberg Associate dean, postgraduate medicine,Thames RegionRi Hornung Dean of postgraduate general practice,South ThamesThames Postgraduate Medical and DentalEducation, University of London, LondonWC1N 3EJ

1 Gibson H. Are part time doctors better doctors? [Careerfocus]. BMJ 1997;315 (classified supplement):2-3.(11 October.)

2 Kumar V. Full time backlash? [Career focus]. BMJ1997;315 (classified supplement):3. (6 December.)

3 Allen I. Career preferences of doctors. BMJ 1996;313:2.

Service depends on teamwork

Editor—Kumar seems to think he can per-sonally provide continuity of care while he isonly working 50 hours a week.1 Whomanages his patients when they are illduring the remaining 118 hours? Noindividual doctor, full or part time, can beavailable for all the hours their patients mayneed them. A continuous service dependson a team of health professionals workingtogether and communicating effectively,whether or not their consultant is present.Perhaps the wisdom that part time doctorsacquire during their “non-working” hours isthe realisation that they are not indispensa-ble but part of a team, both at work and athome.Anna Lipp Senior registrar in anaestheticsNorfolk and Norwich Hospital, Norwich NR1 3SR


Fulfilled doctors are better doctors

Editor—The London initiative zoneeducational incentives programme (LIZEI)is using education as an incentive to attractgeneral practitioners to enter, stay in, orreturn to, general practice in inner cities.The essence of the programme’s approachis flexibility. Providing opportunities fordoctors to develop their skills and interestsenhances their enjoyment, and thus thequality, of their work.

Kumar obviously disagrees.1 The numberof hours a doctor spends in contact withpatients, rather than the quality of work donewithin those hours, seems to be the issue for

him. Continuity of care is important and can-not be maintained if a doctor is not availablefor a sufficient number of hours. It is simplynot true, however, that someone who hasgone through medical training while givinglow priority to other aspects of his or her per-sonal development is therefore a betterdoctor. As Gibson argued, the opposite isusually the case.2

Patients benefit from having doctorswho are fulfilled in their personal lives.Would Kumar suggest that children ofdoctors remain relatively unparented, orwould he prefer medicine to return to beingthe exclusive domain of men? He failed tomention who looks after his children whilehe is pursuing his many, varied, and timeconsuming activities.

One of us (CS-V) is a mature medicalstudent, and the other (LM) is completing ayear as a returner in general practice. Weboth have children and have both exploredother interests before settling on our careerpaths. We are both sure that we are doingwhat we are doing—combining work andfamily—because we want to and not becausewe are fulfilling the expectations of others.We noted with interest that, in the surveyreported recently in Career focus, 95% ofmedical students considered flexible train-ing a good idea.3

A scheme proposed at Guy’s and StThomas’s Hospitals offers students leavingschool and selected for admission tomedical school a place which would then bekept for them until they were 21 years old. Inthe intervening time, they would be encour-aged to pursue their own interests with aplace guaranteed at the end. In case theyrealised during that time that medicine wasnot for them, they could withdraw with norecrimination. This is an excellent proposal.

Doctors have the same needs as anyoneelse to be good parents and fulfilled people.If a lack of flexibility in working practicesand attitudes does not allow these needs tobe expressed, doctors, and consequentlypatients, will suffer.Cynthia Sahleen-Veasey Medical studentKing’s College Medical School, Weston EducationCentre, London SE5 7DF

Lesley Morrison General practitionerGoodinge Health Centre, London N7 9EW


2 Gibson H. Are part time doctors better doctors? [Careerfocus]. BMJ 1997;315 (classified supplement):2-3.(11 October.)

3 Sinha A, Cook A. What do medical students think of flex-ible training? BMJ 1997;315 (classified supplement):2-3.(6 December.)

Continuity of care is likely to suffer

Editor—Kumar says that his comments maynot be applicable universally.1 There is nodoubt in my mind that the strongest featureof British primary care has been its intentionand often its ability to provide continuity ofpatient care. This is now being undermined aslow morale, exhaustion, and frustration sapthe patience of many. Canaries taken by min-ers down the pit quickly detected poor airquality, and, similarly, the most sensitive andcreative people in today’s work environments

are often the first to start wilting. I know manypractices that are now surviving on the backsof the booming locum industry—a phenom-enon also repeated in the teaching and nurs-ing professions.

The increases in numbers of job sharersand “portfolio practitioners” may be goodfor the health of the particular people, butthey are going to make it much harder forthose left in the job to ensure continuity. Ifthe people of Britain think that it is difficultto see the same doctor more than once—they haven’t seen anything yet.Chris Manning General practitionerSt John’s Health Centre, Twickenham TW11 9HG


*** We received eight other letters, six of themfrom doctors working flexibly. They all agreethat “part time” merely describes the waythat all doctors work, to a greater or lesserdegree. The debate should consequently notbe about whether part time or full time isbetter, but how flexible training and workpatterns can be developed that accommo-date different doctors’ different needs in achanging society, and therefore offerpatients better care and greater continuity.

More patients should beexcluded from being tissuedonorsEditor—In their editorial, Allan and Tuftnote the United Kingdom’s criteria aimed atpreventing patients with Creutzfeldt-Jakobdisease from becoming corneal donors.1 Inthe United Kingdom, as elsewhere, discrep-ancies occur in practice between groupsexcluded as blood donors and as corneal(and other) graft donors. Among thesediscrepancies is the fact that relatives ofpatients with Creutzfeldt-Jakob disease areexcluded from donating blood but notnecessarily corneas. Asymptomatic relativesof the 10-15% of patients with hereditaryCreutzfeldt-Jakob disease (which oftenoccurs without a clear family history) mayshare the same dominant genetic defect andhave a theoretical potential for transmittingthe disease in the presymptomatic stage oftheir illness.

We recommend that blood relatives ofpatients with Creutzfeldt-Jakob disease or therelated hereditary conditions Gerstmann-Sträussler-Scheinker disease and fatal familialinsomnia should be added to the list of thoseexcluded as corneal donors. Ideally, it wouldbe advisable to ensure the absence ofCreutzfeldt-Jakob disease by histologicalexamination of a small piece of frontotempo-ral cortex from any cadaver used as a sourceof tissue grafts (a requirement soon to beenacted for dura mater grafts in the UnitedStates, according to the transcript of themeeting of the transmissible spongiformencephalopathy advisory committee on 6October 1997). However, the large number ofcorneal transplant operations that areperformed each year, coupled with the

Letters


exceptional rarity of recognised transmissionof Creutzfeldt-Jakob disease from contami-nated grafts, make this proposal a legitimatesubject for argument.Martin Zeidler Specialist registrar in neurologyWestern General Hospital, Edinburgh EH4 2XU

Paul Brown Senior research scientistLaboratory of Central Nervous System Studies,National Institute of Neurological Disorders andStroke, National Institutes of Health, Bethesda MD20892, USA

1 Allan B, Tuft S. Transmission of Creutzfeldt-Jakob diseasein corneal grafts. BMJ 1997;315:1553-4. (13 December.)

Acute excited states andsudden death

Death after restraint can be avoided

Editor—Farnham and Kennedy’s recenteditorial on sudden death in patients withacute psychiatric illness1 who are beingrestrained does not reflect the forensicliterature fairly. It omits evidence from keypublications2 3 and twice misattributes state-ments to a paper that specifically excludedsuch deaths from the reported study.4

The current consensus on “restraintasphyxiation” is reflected in a guidancestatement of the United States Departmentof Justice.5 Any person who is restrainedprone has trouble breathing when pressureis applied to his or her back or when hand-cuffed and “hog tied,”2 3 and obesity exacer-bates this. People’s natural reaction is tostruggle more violently, which may be metwith still more restraining force, resulting ingreater oxygen deficiency and, in somecases, death. Risk factors include any condi-tion, such as ischaemic heart disease orexcited delirium,2 that increases susceptibil-ity to cardiac arrhythmias, as well asintoxication with alcohol or drugs.

Recommendations given to police offic-ers on how to prevent such deaths includereleasing people from the prone position assoon as they are handcuffed, not sitting onthem, and never tying handcuffs to a leg orankle restraint (hog tying).5 A form of hogtying restraint known as the police leg lockwas prohibited in Denmark in 1994, after acampaign by the human rights organisationAmnesty International when an incidenthad left an 18 year old permanently braindamaged. There is evidence that restraintprocedures that compromise respiratorymovements are potentially dangerous, andthis needs to be taken into account in anydebate on their use in a healthcare setting.Derrick Pounder Professor of forensic medicineDepartment of Forensic Medicine, University ofDundee, Royal Infirmary, Dundee DD1 9ND

1 Farnham FR, Kennedy HG. Acute excited states andsudden death. BMJ 1997;315:1107-8. (1 November.)

2 O’Halloran RL, Lewman LV. Restraint asphyxiation inexcited delirium. Am J Forensic Med Pathol 1993;14:289-95.

3 Reay DT, Fligner CL, Stilwell AD, Arnold J. Positionalasphyxia during law enforcement transport. Am J ForensicMed Pathol 1992;13:90-7.

4 Bell MD, Rao VJ, Wetli CV, Rodriguez RN. Positionalasphyxiation in adults. Am J Forensic Med Pathol1992;13:101-7.

5 Department of Justice. Positional asphyxia—sudden death.Rockville, MD: National Institute of Justice, 1995. (NationalLaw Enforcement Technology Center Bulletin.)

Acute excited states are not caused byhigh blood concentrations of cocaine

Editor—Much more is known about acuteexcited states and sudden death thanFarnham and Kennedy’s editorial suggests.1

Such states are not, as they say, commonlyassociated with high blood concentrationsof cocaine—subjects do not have highconcentrations of cocaine. In 45 cases inMiami, concentrations were similar to thoseseen in asymptomatic recreational users.2 Inpeople whose deaths are related to cocaine,whether due to psychosis or not, blood con-centrations completely overlap those ofpeople whose death was from trauma withcocaine as an incidental finding.3

Neurochemical abnormalities have beenidentified. There are fewer D2 receptors inthe centres of the hypothalamus that opposeD1 mediated temperature increases.4 Thisexplains why those who die are invariablyhyperthermic. In the series from Miami,mean body temperature was 40.6°C at thetime of the first medical examination.2

Changes induced by cocaine in the numberand distribution of ê2 opiate receptorswithin the amygdala may explain thedistinctive psychotic symptoms and violentagitation.5 These changes can be shown atany neurochemistry reference laboratory,provided the brain is removed and slices fro-zen within 12 hours. The mere presence ofthese changes does not prove that policemisconduct or medical mismanagement didnot occur, but it does prove that the patienthad a disease that is usually fatal.

Death is often attributed to capsicum(pepper spray) poisoning rather than to theunderlying psychiatric state. In trying tospray agitated psychotic people, police offic-ers often miss their face and spray their backor shoulder and sometimes the faces ofother officers. We use saline swabs to recovercapsicum from skin and clothing of peoplewho are alive and methanol to recover itfrom those who have died. There are noassays for capsicum in biological matrices,but toxicity is effectively ruled out ifcapsicum cannot be recovered from thefacial area.

Prudent management also includes ameticulous, well photographed, neck dissec-tion to rule out neck compression, andmeasurement of drug concentrations in thebrain. These seem to provide an accuratepicture of the pattern of cocaine use, and theuse at the time of death, whereas bloodmeasurements are often unreliable.

Finally, protocols must be put in place toensure that all the appropriate measures areimmediately taken. These are high profilecases. Since “popular journalism favourscontroversy and blame rather than balanceand exploration,”1 it is probably a good ideato implement such protocols before thebacklog of cases gets too large.Steven B Karch Assistant medical examinerBoyd G Stephens Chief medical examinerCity and County of San Francisco, Hall of Justice,San Francisco, California CA94103, USA

1 Farnham FR, Kennedy HG. Acute excited states andsudden death. BMJ 1997;315:1107-8. (1 November.)

2 Wetli C, Mash D, Karch, S. Cocaine-associated agitateddelirium and the neuroleptic malignant syndrome. Am JEmerg Med 1996;14:425-8.

3 Karch S, Stephens B, Ho CH. Relating cocaine blood levelsto toxicity—an autopsy study of 99 cases. J Forensic Sci(in press.)

4 Staley J, Hearn L, Ruttenber A, Mash D. High affinitycocaine recognition sites on the dopamine transporter areelevated in fatal cocaine overdose victims. J Pharmacol ExpTher 1994;271:1678-85.

5 Staley J, Rothman R, Rice K, Partilla J, Mash D. K2 opioidreceptors in limbic areas of human brain are up regulatedby cocaine in fatal overdose victims. J Neurosci1997;17:8225-33.

Many third molar teeth areprobably removedunnecessarily in SandwellEditor—Jones et al report their interviewstudy to assess dentists’ agreement on thetreatment of asymptomatic impacted thirdmolar teeth.1 They state that a quarter ofthese teeth are removed without beingdiseased, and give as a reference for thisstatement a paper by colleagues and myself;in fact the figure was half, not a quarter.2

It is reassuring to see that Jones et alused as the gold standard guidelines thathad been outlined at a consensus develop-ment conference of the National Institutesof Health. Their findings are alarming butnot unexpected. The purchasers for Bir-mingham have prepared guidelines for thirdmolar surgery largely based on the consen-sus guidelines. This document, however, stillsays that impacted third molars can beremoved “on the decision of aconsultant”—in other words, without anyclinical basis that can be justified.

The increase in the number of thirdmolars being extracted by general dentalpractitioners is alarming. In Sandwell (adja-cent to Birmingham) during 1994-5 theratio of the actual to expected episodes ofhospital extraction of third molars was thesecond lowest in the West Midlands region.3

Even more worrying, however, is the factthat the number of claims per thousandpopulation aged 15-39 made by dentists inSandwell for the removal of third molars in1994-5 was the highest in the WestMidlands.3 If the dentists in Sandwell havethe same attitudes as those in the studygroup then I suspect that many third molarsare removed unnecessarily. The challenge isfor the Dental Practice Board to monitorand audit this, and for the profession tobring about pressure to stop such activity.Victor Lopes Senior house officer, surgical rotationCity Hospital, Birmingham B18 7QH

1 Jones CM, O’Brien K, Blinkhorn AS, Rood JP. Dentists’agreement on treatment of asymptomatic impacted thirdmolar teeth: interview study. BMJ 1997;315:1204.(8 November.)

2 Lopes V, Mumenya R, Feinmann C, Harris M. Third molarsurgery: an audit of the indications for surgery,post-operative complaints and patient satisfaction. Br JOral Maxillofac Surg 1995;33:33-5.

3 Taylor K, Elley K. Improving oral health in Sandwell. WestBromwich: Department of Public Health, Sandwell HealthAuthority, 1997.

Letters


Many third molar teeth are probably removed unnecessarily in Sandwell

Documents