specifically related to factors such as conflicts between career and personal life and lack of female role models than to sex differences in doctors' responses to emotionally charged events like failure of treatment and talking to distressed relatives.' A study of veterinary surgeons2 has given rather different results in that women were consistently more likely than men to show both short and long term emotional reactions of a broadly depressive kind to both failure of treatment and carrying out euthanasia, which is a unique and common component of veterinary practice. Women have formed a noticeably increasing proportion of veterinary surgeons in recent years and this may be associated with the shift of emphasis towards the treatment of animals that fill an emotional rather than economic role. A direct comparison of the effects of the two professions might be of considerable interest. The less formal and hierarchical organisation of the veterinary profession may reduce some of the career conflicts found to be important among house officers while throwing other sources of stress into relief. DAVID ABRAHAMSON
Goodmayes Hospital,lEssex IG3 8XJ BRUCE FOGLE Portman Veterinary Clinic,,London W 1H lDP 1 Firth-Cozens J. Sources of stress in women junior house officers. Br MedJ3 1990;301:89-91. (14 JulNy.) 2 Abrahamson D, Fogle B. Pet loss: a survev of the attitudes and feelings of practising veterinarians. Anthrozoos 1990;J11: 143-50.
Loop diathermy excision SIR,-The paper by Dr D M Luesley and colleagues raises a number of questions.' They carried out diathermy loop excision of the cervical transformation zone in 616 patients with abnormal cervical smears and concluded that it was an effective treatment with low morbidity in this group of patients. Unfortunately the cytological terminology used by these workers does not correspond to that recommended by the British Society for Clinical Cytology,' which is now used by most British laboratories. This idiosyncratic terminology, which is regrettably not explained in the text, makes it impossible to interpret the cytological indications for treatment except by guesswork. More importantly, 45% of the patients who had a loop excision, which is essentially a small cone biopsy, were found to have no or only minor abnormalities of uncertain clinical3 and pathological importance4 (5% histologically normal, 22% koilocytosis only, 18% cervical intraepithelial neoplasia grade I). In addition to haemorrhage and cervical stenosis surgical manipulation of the cervix has been associated with cervical endometriosis,6 which has been found in 43% of women after cone biopsy in this hospital and often presents with postcoital and intermenstrual bleeding. This must be regarded as an important complication in this group of patients, 70% of whom were aged 30 years or less. Though diathermy loop excision may be a good treatment for cervical intraepithelial neoplasia, uncritical application of this technique in this instance has regrettably resulted in overtreatment in 45% of cases, which must be considered uneconomic if not positively harmful to the patient. I would ask for a more critical approach regarding the indications for treatment of cervical intraepithelial neoplasia. To those of us who are interested in cervical pathology the important question is not how but which cervical squamous lesions should be treated. S M ISMAIL
University of Wales College of Medicine, Cardiff CF4 4XN
BMJ
VOLUME
301
11
AUGUST
1990
I Luesley DM, Cullimore J, Redman CWE, et al. Loop diathermy excision of the cervical transformation zone in patients with abnormal cervical smears. Br Med J 1990;300:1690-3. (30 June.) 2 Evans DMD, Hudson EA, Brown CL, et al. Terminology in gynaecological cytopathology: report of the working party of the British Society for Clinical Cytology. J Clin Pathol 1986;39:933-44. 3 Robertson JH, Woodend BE, Crozier EH, Hutchinson J. Risk of cervical cancer associated with mild dvskaryosis. Br Med J 1988;297: 18-2 1. 4 Ismail SM, Colclough AB, Dinnen JS, et al. Observer variation in histopathological diagnosis and grading of cervical intraepithelial neoplasia. BrMedj 1989;298:707- 10. 5 Robertson AJ, Anderson JM, Beck JS, et al. Observer variability in histopathological reporting of cervical biopsy specimens. J Clin Pathol 1989;42:231-8. 6 Williams GA. Endometriosis of the cervix uteri-a common disease. Am J Obstet Gynecol 1960;80:734-41.
SIR, -I am writing concerning the paper by Dr D M Luesley on loop diathermy excision in patients with abnormal cervical smears. ' I regard the figure of 27% overtreatment of cervical intraepithelial neoplasia (5% of patients with no evidence of preinvasive disease and 22% showing only koilocytosis) as unacceptable. Imagine the woman who presents to the outpatient department with an abnormal smear. By the loop diathermy modality of treatment she has a 15% risk of experiencing mild to moderate discomfort during the procedure, a 4-7% risk of secondary haemorrhage, a 5-6% risk of having vaginal discharge for more than six weeks, and a 1-3% risk of severe cervical stenosis. Admittedly she may have none of the above complications, but she could experience one or all of them and then has a 27% chance of being told that she had no disease in the first place. The potential risk of cervical incompetence and future spontaneous abortion must be borne in mind. The psychological aspects have not even been considered. I know that if I were a woman with, say, vaginal discharge six weeks after treatment for what many people see as cancer I would be worried sick. I believe that it is only humane to perform punch biopsies on women who present with abnormal smears before subjecting them to more radical treatment with potential risks. CHARLES HILLIER
University College Hospital, London WC 1 E 6BT 1 Luesley DM, Cullimore J, Redman CWE, et al. Loop diathermy excision of the cervical transformation zone in patients with abnormal cervical smears. Br Med J 1990;300:1690-3.
(30 June.)
AUTHOR'S REPLY,-Dr S M Ismail raises the possibility that 45% of our patients were overtreated; this was discussed in our paper. The issue of whether cervical intraepithelial neoplasia grade I or warty atypia should be treated, however, must be based on sound prospective data and not on personal opinion. We must also assume that most of her patients are treated on the basis of directed cervical biopsy, a technique that we have shown to be inaccurate in comparison with loop excision and, furthermore, tends to overgrade lesions. We are satisfied that our treatment has a low morbidity and think that it is deliberately misleading to compare transformation zone loop excision with knife conisation. Most workers in this specialty are aware of the differences in indication, technique, and morbidity between the two techniques. Though 43% of patients who have cone biopsies and a further representative biopsy may have endometriosis, this cannot be extrapolated to the total population of patients having cone biopsies (unless they are all given second biopsies or the sample was truly random). Thus such a suggestion lacks credibility, and this is compounded by the use of percentages without confidence intervals, which leaves the reader sceptical about sample size. The economics of the debate on treatment will
not be settled by personal prejudice, and one must be wary of any statement on health economics that does not count the cost of the alternative-that is, not treating patients with abnormal cervical smears. This cost must include not only the actual expense entailed in repeated cytological examinations (and eventual treatment in a proportion) but also the psychological cost to the woman, who is aware of her persisting abnormality. Dr Ismail also criticises our cytological terminology. At the start of our programme trial the guidelines of the British Society for Clinical Cytology were certainly not adhered to universally, and indeed many laboratories still have not adopted them. We are, however, quite reassured that most laboratories and clinicians can interpret the cytology gradings without recourse to guesswork. A point on which we would wholeheartedly agree is the need rationally to select patients for treatment. As clinicians managing patients we are aware of the need to develop our selective expertise and the appropriate clinical research protocols to enable this objective to be achieved. Mr Charles Hillier regards a 27% overtreatment rate as unacceptable and suggests that prior directed biopsy is a more humane approach. We also believe that there is scope for improved selectivity but are aware of the fairly poor accuracy ofdirected biopsy. He should be aware that this procedure is not uncommonly uncomfortable and, in relation to transformation zone excision, is diagnostically inferior. It would be premature to dismiss koilocytotic atypia as normal given the strong associations between infection with human papillomavirus and intraepithelial neoplasia, and though I would wish to reserve judgment on the preneoplastic importance of infection with papillomavirus I would not consider such infection as normal. Furthermore, the technique that Mr Hillier recommends (punch biopsy) has been shown both by us and by others to overall koilocytotic atypia as mild or moderate dysplasia, and we would suggest that the high incidence of koilocytotic atypia in our series reflects to some extent the superior diagnostic technique. A final consideration is the question of related anxiety. Mr Hillier would be "worried sick" as a result of associating a discharge with a diagnosis of cancer (although all our patients are counselled that they do not have cancer). I can assume only that a persistent cytological abnormality would be more acceptable to him as this is the alternative. I do not believe that Mr Hillier would enjoy majority support least of all from women with abnormal smears, in whom, unlike Mr Hillier, we have considered the psychological impact of such a situation and believe that by shortening the length of time that they have abnormal cytology we are considerably reducing their anxiety. Finally, the risk of cervical incompetence is theoretical and, in a procedure that not only conserves the internal cervical os but removes only the lower canal (like laser), is virtually
impossible. DAVID LUESLEY
Dudley Road Hospital, Birmingham B18 7QH
Perioperative deaths among children SIR,-In his editorial' on the newly published report on perioperative deaths among children Mr Malcolm H Gough said that the inquiry underlined the accepted need for continuing postgraduate education for all consultants who care for children, particularly for those in district and single specialty hospitals.2 He suggested that the best way to do this is by joining and attending the meetings held by specialty associations such as the Association 343
of Paediatric Anaesthetists of Great Britain and Northern Ireland. A high percentage of children under 3 are anaesthetised in adult hospitals, and there has been much discussion in recent years about the organisation of paediatric anaesthetic services in the district general hospital. Some specialisation has been encouraged: it has been suggested that one or two consultant anaesthetists could take on particular responsibility for paediatric anaesthesia and intensive care.34 Such practice would seem to be supported by the report of the National Confidential Enquiry into Perioperative Deaths. Unfortunately, membership of the Association of Paediatric Anaesthetists is restricted to consultants with at least five sessions a week in paediatric anaesthesia. This precludes most consultants in district general hospitals and some in teaching hospitals from becoming members. These consultants are therefore unable to meet other colleagues with similar exposure to paediatric anaesthesia and to exchange experiences. Neither can they benefit from hearing experts discuss paediatric problems.The Association of Anaesthetists has played a prominent part in the initiation and completion of the National Confidential Enquiry into Perioperative Deaths and the original confidential inquiry into perioperative deaths.' I appeal to the officials of the Association of Paediatric Anaesthetists to heed the recommendations of the report and change the criteria for membership so that more anaesthetists with an interest in paediatric anaesthesia are allowed entry. Perhaps our paediatric patients would benefit from this action. PAUL M SPARGO Shackleton Department of Anaesthetics, Southampton General Hospital, Southampton S09 4XY 1 Gough MH. Perioperative deaths among children. Br Med J 1990;300:1606. (23 June.) 2 National Confidential Enquiry into Perioperative Deaths. Report 1989. London: NCEPOD, 1990. 3 Hatch DJ. Anaesthesia for children. Anaesthesia 1984;39:405-6. 4 Bowhay AR, Morgan-Hughes JO. Paediatric anaesthesia in a district general hospital. Anaesthesia 1989;44:139-42. 5 Buck N, Devlin HB, Lunn JN. The report of the confidential enquiry into perioperative deaths. London: Nuffield Provincial Hospitals Trust/King's Fund, 1987.
The cyclotron saga SIR,-Drs R Hugh MacDougall and Sydney J Arnott made numerous erroneous and misleading statements about a study by the Radiation Therapy Oncology Group on locally advanced prostate cancer.' The study compared the efficacy of fast neutron radiotherapy with that of conventional radiotherapy for locally advanced adenocarcinomas of the prostate gland.2' Given the poor penetration of the low energy neutron beams available at the time the study was initiated, it was elected to use a combination of neutrons and photons (mixed beam) instead of neutrons alone in the experimental arm. Patients with stages C and DI tumours were eligible for the study and a total of 91 patients were randomised between the two arms 11 had stage DI cancer (positive pelvic nodes) and 80 had stage C. Patients were stratified according to histological grade, history of hormonal treatment, and state of the pelvic nodes. Based on X2 tests the two patient arms were balanced according to the following major prognostic variables which were prospectively collected: stage of disease (C v D1), grade (Mustofi scheme), affected seminal vesicles, serum acid phosphatase activities, history of hormonal therapy, diagnostic procedure (transurethral resection v needle biopsy), method of nodal evaluation (lymphangiography v laparotomy), age distribution of patients, Karnofsky performance score, presence of cardiac disease, presence of intercurrent disease, ethnic origin,
344
and presence of benign prostatic hypertrophy. The only variable of marginal significance was the presence of benign prostatic hypertrophy (p=006), which was more common in the mixed beam arm. Gleason scores were retrospectively evaluated for 73 patients, and this subgroup was balanced between the two arms. Hence the allegation that the patients treated in the photon arm were in some respect "worse" than those patients treated in the mixed beam arm is simply incorrect. Ten years' results were presented at the 1990 meeting of the American Radium Society and will be presented at the 15th international cancer congress. For all major end points the group receiving mixed beam treatment did better than the group receiving photons alone (local control, 63% mixed beam v 52% photons (p=005); survival, 42% v 27% (p=005); and disease specific survival, 56% v 42% (p=004)). Whether or not differences between arms in a clinical trial achieve significance depends on both the number of patients in the trial and the observed differences in end points. For the above variables all differences were significant based on a two sided MantelHaenszel log rank test.4 Hence, the allegation by Drs MacDougall and Arnott that the numbers of patients were too small to draw valid conclusions is also incorrect. Differences in factors related to patients make comparisons between studies carried out at other times and places extremely hazardous. This is themain reason why randomised trials are conducted. Morbidity related to treatment was discussed in detail in both the five year and the eight year reports.2" Both acute and late morbidity were evaluated according to the joint scoring scheme of the Radiation Therapy Oncology Group and European Organisation for Research on the Treatment of Cancer and were found to be equivalent in the two arms. In all there were six reactions scored as severe or greater in the mixed beam arm and five in the photon arm. The only fatal complication occurred in the photon arm-a patient underwent a diverting colostomy, developed septicaemia, and died. To paraphrase Drs MacDougall and Arnott, the "morbidity watchdog did not bark because there was nothing to arouse it." This dog barked at Edinburgh because low energy neutrons alone were used unadvisedly to treat deep seated pelvic tumours.' The study concluded that the mixed beam form of treatment offered improved local control and survival for locally advanced prostate cancer at no increased morbidity. It was discontinued when new high energy cyclotrons capable of isocentric treatment became available in the United States and was replaced with a new study comparing neutrons alone and conventional photon irradiation. The planners of the protocol thought that the most important issue was not whether one form of neutron treatment was better than another (as would have been the case if the mixed beam arm had been the control arm of a new trial) but to ask the question again with larger patient numbers. The new trial has thus far accrued about 180 patients. If it confirms the results of the earlier study we predict that neutron radiotherapy will move to the private medical sector in the United States. This is the only way that sufficient numbers of machines can be built to accommodate the resulting patient load. GEORGE E LARAMORE Department of Radiation Oncology, University of Washington Medical Center, Seattle, WA 98195, United States I MacDougall RH, Arnott SJ. The cyclotron saga. Br Med J 1990;300:1721. (30 June.) 2 Laramore GE, Krall JM, Thomas FJ, Griffin TW, Maor MH, Hendrickson FR. Fast neutron radiotherapy for locally advanced prostate cancer: results of an RTOG randomized study. IntJ7 Radiat Oncol Biol Phys 1985;11:1621-7. 3 Russell KJ, LaramoreGE, KrallJM,etal. Eightyears' experience
with neutron radiotherapy in the treatment of stages C and D prostate cancer: updated results of the RTOG 7704 randomized clinical trial. Prostate 1987;11:183-93. 4 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease.J7NCI 1974;22:719-48. 5 Duncan W, Williams JR, Kerr GR, et al. An analysis of the radiation related morbidity observed in a randomized trial of neutron therapy for bladder cancer. Int J Radiat Oncol Biol Phts 1986;12:2085-92.
Precautions taken by orthopaedic surgeons to avoid infection with HIV and hepatitis B SIR,-The Editor's Choice and the article by Mr M J LeF Porteous in the same issue' criticise the NHS occupational health departments for allegedly not offering hepatitis B vaccine to surgeons. Few would disagree that this group should be offered vaccine and post-immunisation antibody testing. It is not reasonable to place all of the responsibility for failure to offer vaccine on occupational health staff. Although occupational physicians have the responsibility of advising their districts, the authority has to decide whether to take the advice and meet the resource implications. Regrettably, some health authorities have chosen to ignore the advice on financial grounds, and the responsibility for this rests squarely on them. M S GATLEY North Manchester General Hospital, Manchester M8 6RB
1 Porteous MJLeF. Operating practices of and precautions taken by orthopaedic surgeons to avoid infection with HIV and hepatitis B during surgery. Br MedJ 1990;301:167-9. (21 July.)
Management of hypovolaemic shock SIR,-Dr Peter J F Baskett's letter seems to be seriously misleading when he states that a venous cut down requires training and practice.' Of course, training and practice are desirable, but it is important for doctors of all disciplines to know that when a patient is dying of hypovolaemic shock it is perfectly appropriate for them to attempt a cut down on to a vein if venous access fails even if they have no experience. I have found this a feasible and lifesaving procedure. In the absence of contraindications the most suitable site for novices (such as myself) is the long saphenous vein as it courses anterior to the medial malleolus. As many surgeons are aware, cut downs at this site have saved many lives. It is worth quoting from one of our most well known surgical anatomists on this issue: "From the practical point of view the position of the long saphenous vein immediately in front of the medial malleolus is perhaps the most important anatomical relationship. No matter how collapsed and obese, or young and tiny the patient, the vein can be relied upon to be available at this site when urgently required for transfusion purposes."2 No one would suggest that someone tries to insert a central venous line without training and practice as so much damage could be done to vital structures, but at the ankle there is not much to lose and everything to gain if the patient is already dying. All that the doctor needs is a scalpel and artery forceps. He or she does not need to know how to tie surgical knots or use sutures. Such knowledge is helpful but in an emergency the cannula can simply be held in place while the doctor waits for skilled help to arrive-in those circumstances when postponing action even for a few minutes would be likely to be fatal. Of course it is necessary to bear in mind that the leg is not the ideal site when there is bleeding from pelvic or
BMJ
VOLUME
301
11 AUGUST 1990
Goodmayes Hospital,lEssex IG3 8XJ BRUCE FOGLE Portman Veterinary Clinic,,London W 1H lDP 1 Firth-Cozens J. Sources of stress in women junior house officers. Br MedJ3 1990;301:89-91. (14 JulNy.) 2 Abrahamson D, Fogle B. Pet loss: a survev of the attitudes and feelings of practising veterinarians. Anthrozoos 1990;J11: 143-50.
Loop diathermy excision SIR,-The paper by Dr D M Luesley and colleagues raises a number of questions.' They carried out diathermy loop excision of the cervical transformation zone in 616 patients with abnormal cervical smears and concluded that it was an effective treatment with low morbidity in this group of patients. Unfortunately the cytological terminology used by these workers does not correspond to that recommended by the British Society for Clinical Cytology,' which is now used by most British laboratories. This idiosyncratic terminology, which is regrettably not explained in the text, makes it impossible to interpret the cytological indications for treatment except by guesswork. More importantly, 45% of the patients who had a loop excision, which is essentially a small cone biopsy, were found to have no or only minor abnormalities of uncertain clinical3 and pathological importance4 (5% histologically normal, 22% koilocytosis only, 18% cervical intraepithelial neoplasia grade I). In addition to haemorrhage and cervical stenosis surgical manipulation of the cervix has been associated with cervical endometriosis,6 which has been found in 43% of women after cone biopsy in this hospital and often presents with postcoital and intermenstrual bleeding. This must be regarded as an important complication in this group of patients, 70% of whom were aged 30 years or less. Though diathermy loop excision may be a good treatment for cervical intraepithelial neoplasia, uncritical application of this technique in this instance has regrettably resulted in overtreatment in 45% of cases, which must be considered uneconomic if not positively harmful to the patient. I would ask for a more critical approach regarding the indications for treatment of cervical intraepithelial neoplasia. To those of us who are interested in cervical pathology the important question is not how but which cervical squamous lesions should be treated. S M ISMAIL
University of Wales College of Medicine, Cardiff CF4 4XN
BMJ
VOLUME
301
11
AUGUST
1990
I Luesley DM, Cullimore J, Redman CWE, et al. Loop diathermy excision of the cervical transformation zone in patients with abnormal cervical smears. Br Med J 1990;300:1690-3. (30 June.) 2 Evans DMD, Hudson EA, Brown CL, et al. Terminology in gynaecological cytopathology: report of the working party of the British Society for Clinical Cytology. J Clin Pathol 1986;39:933-44. 3 Robertson JH, Woodend BE, Crozier EH, Hutchinson J. Risk of cervical cancer associated with mild dvskaryosis. Br Med J 1988;297: 18-2 1. 4 Ismail SM, Colclough AB, Dinnen JS, et al. Observer variation in histopathological diagnosis and grading of cervical intraepithelial neoplasia. BrMedj 1989;298:707- 10. 5 Robertson AJ, Anderson JM, Beck JS, et al. Observer variability in histopathological reporting of cervical biopsy specimens. J Clin Pathol 1989;42:231-8. 6 Williams GA. Endometriosis of the cervix uteri-a common disease. Am J Obstet Gynecol 1960;80:734-41.
SIR, -I am writing concerning the paper by Dr D M Luesley on loop diathermy excision in patients with abnormal cervical smears. ' I regard the figure of 27% overtreatment of cervical intraepithelial neoplasia (5% of patients with no evidence of preinvasive disease and 22% showing only koilocytosis) as unacceptable. Imagine the woman who presents to the outpatient department with an abnormal smear. By the loop diathermy modality of treatment she has a 15% risk of experiencing mild to moderate discomfort during the procedure, a 4-7% risk of secondary haemorrhage, a 5-6% risk of having vaginal discharge for more than six weeks, and a 1-3% risk of severe cervical stenosis. Admittedly she may have none of the above complications, but she could experience one or all of them and then has a 27% chance of being told that she had no disease in the first place. The potential risk of cervical incompetence and future spontaneous abortion must be borne in mind. The psychological aspects have not even been considered. I know that if I were a woman with, say, vaginal discharge six weeks after treatment for what many people see as cancer I would be worried sick. I believe that it is only humane to perform punch biopsies on women who present with abnormal smears before subjecting them to more radical treatment with potential risks. CHARLES HILLIER
University College Hospital, London WC 1 E 6BT 1 Luesley DM, Cullimore J, Redman CWE, et al. Loop diathermy excision of the cervical transformation zone in patients with abnormal cervical smears. Br Med J 1990;300:1690-3.
(30 June.)
AUTHOR'S REPLY,-Dr S M Ismail raises the possibility that 45% of our patients were overtreated; this was discussed in our paper. The issue of whether cervical intraepithelial neoplasia grade I or warty atypia should be treated, however, must be based on sound prospective data and not on personal opinion. We must also assume that most of her patients are treated on the basis of directed cervical biopsy, a technique that we have shown to be inaccurate in comparison with loop excision and, furthermore, tends to overgrade lesions. We are satisfied that our treatment has a low morbidity and think that it is deliberately misleading to compare transformation zone loop excision with knife conisation. Most workers in this specialty are aware of the differences in indication, technique, and morbidity between the two techniques. Though 43% of patients who have cone biopsies and a further representative biopsy may have endometriosis, this cannot be extrapolated to the total population of patients having cone biopsies (unless they are all given second biopsies or the sample was truly random). Thus such a suggestion lacks credibility, and this is compounded by the use of percentages without confidence intervals, which leaves the reader sceptical about sample size. The economics of the debate on treatment will
not be settled by personal prejudice, and one must be wary of any statement on health economics that does not count the cost of the alternative-that is, not treating patients with abnormal cervical smears. This cost must include not only the actual expense entailed in repeated cytological examinations (and eventual treatment in a proportion) but also the psychological cost to the woman, who is aware of her persisting abnormality. Dr Ismail also criticises our cytological terminology. At the start of our programme trial the guidelines of the British Society for Clinical Cytology were certainly not adhered to universally, and indeed many laboratories still have not adopted them. We are, however, quite reassured that most laboratories and clinicians can interpret the cytology gradings without recourse to guesswork. A point on which we would wholeheartedly agree is the need rationally to select patients for treatment. As clinicians managing patients we are aware of the need to develop our selective expertise and the appropriate clinical research protocols to enable this objective to be achieved. Mr Charles Hillier regards a 27% overtreatment rate as unacceptable and suggests that prior directed biopsy is a more humane approach. We also believe that there is scope for improved selectivity but are aware of the fairly poor accuracy ofdirected biopsy. He should be aware that this procedure is not uncommonly uncomfortable and, in relation to transformation zone excision, is diagnostically inferior. It would be premature to dismiss koilocytotic atypia as normal given the strong associations between infection with human papillomavirus and intraepithelial neoplasia, and though I would wish to reserve judgment on the preneoplastic importance of infection with papillomavirus I would not consider such infection as normal. Furthermore, the technique that Mr Hillier recommends (punch biopsy) has been shown both by us and by others to overall koilocytotic atypia as mild or moderate dysplasia, and we would suggest that the high incidence of koilocytotic atypia in our series reflects to some extent the superior diagnostic technique. A final consideration is the question of related anxiety. Mr Hillier would be "worried sick" as a result of associating a discharge with a diagnosis of cancer (although all our patients are counselled that they do not have cancer). I can assume only that a persistent cytological abnormality would be more acceptable to him as this is the alternative. I do not believe that Mr Hillier would enjoy majority support least of all from women with abnormal smears, in whom, unlike Mr Hillier, we have considered the psychological impact of such a situation and believe that by shortening the length of time that they have abnormal cytology we are considerably reducing their anxiety. Finally, the risk of cervical incompetence is theoretical and, in a procedure that not only conserves the internal cervical os but removes only the lower canal (like laser), is virtually
impossible. DAVID LUESLEY
Dudley Road Hospital, Birmingham B18 7QH
Perioperative deaths among children SIR,-In his editorial' on the newly published report on perioperative deaths among children Mr Malcolm H Gough said that the inquiry underlined the accepted need for continuing postgraduate education for all consultants who care for children, particularly for those in district and single specialty hospitals.2 He suggested that the best way to do this is by joining and attending the meetings held by specialty associations such as the Association 343
of Paediatric Anaesthetists of Great Britain and Northern Ireland. A high percentage of children under 3 are anaesthetised in adult hospitals, and there has been much discussion in recent years about the organisation of paediatric anaesthetic services in the district general hospital. Some specialisation has been encouraged: it has been suggested that one or two consultant anaesthetists could take on particular responsibility for paediatric anaesthesia and intensive care.34 Such practice would seem to be supported by the report of the National Confidential Enquiry into Perioperative Deaths. Unfortunately, membership of the Association of Paediatric Anaesthetists is restricted to consultants with at least five sessions a week in paediatric anaesthesia. This precludes most consultants in district general hospitals and some in teaching hospitals from becoming members. These consultants are therefore unable to meet other colleagues with similar exposure to paediatric anaesthesia and to exchange experiences. Neither can they benefit from hearing experts discuss paediatric problems.The Association of Anaesthetists has played a prominent part in the initiation and completion of the National Confidential Enquiry into Perioperative Deaths and the original confidential inquiry into perioperative deaths.' I appeal to the officials of the Association of Paediatric Anaesthetists to heed the recommendations of the report and change the criteria for membership so that more anaesthetists with an interest in paediatric anaesthesia are allowed entry. Perhaps our paediatric patients would benefit from this action. PAUL M SPARGO Shackleton Department of Anaesthetics, Southampton General Hospital, Southampton S09 4XY 1 Gough MH. Perioperative deaths among children. Br Med J 1990;300:1606. (23 June.) 2 National Confidential Enquiry into Perioperative Deaths. Report 1989. London: NCEPOD, 1990. 3 Hatch DJ. Anaesthesia for children. Anaesthesia 1984;39:405-6. 4 Bowhay AR, Morgan-Hughes JO. Paediatric anaesthesia in a district general hospital. Anaesthesia 1989;44:139-42. 5 Buck N, Devlin HB, Lunn JN. The report of the confidential enquiry into perioperative deaths. London: Nuffield Provincial Hospitals Trust/King's Fund, 1987.
The cyclotron saga SIR,-Drs R Hugh MacDougall and Sydney J Arnott made numerous erroneous and misleading statements about a study by the Radiation Therapy Oncology Group on locally advanced prostate cancer.' The study compared the efficacy of fast neutron radiotherapy with that of conventional radiotherapy for locally advanced adenocarcinomas of the prostate gland.2' Given the poor penetration of the low energy neutron beams available at the time the study was initiated, it was elected to use a combination of neutrons and photons (mixed beam) instead of neutrons alone in the experimental arm. Patients with stages C and DI tumours were eligible for the study and a total of 91 patients were randomised between the two arms 11 had stage DI cancer (positive pelvic nodes) and 80 had stage C. Patients were stratified according to histological grade, history of hormonal treatment, and state of the pelvic nodes. Based on X2 tests the two patient arms were balanced according to the following major prognostic variables which were prospectively collected: stage of disease (C v D1), grade (Mustofi scheme), affected seminal vesicles, serum acid phosphatase activities, history of hormonal therapy, diagnostic procedure (transurethral resection v needle biopsy), method of nodal evaluation (lymphangiography v laparotomy), age distribution of patients, Karnofsky performance score, presence of cardiac disease, presence of intercurrent disease, ethnic origin,
344
and presence of benign prostatic hypertrophy. The only variable of marginal significance was the presence of benign prostatic hypertrophy (p=006), which was more common in the mixed beam arm. Gleason scores were retrospectively evaluated for 73 patients, and this subgroup was balanced between the two arms. Hence the allegation that the patients treated in the photon arm were in some respect "worse" than those patients treated in the mixed beam arm is simply incorrect. Ten years' results were presented at the 1990 meeting of the American Radium Society and will be presented at the 15th international cancer congress. For all major end points the group receiving mixed beam treatment did better than the group receiving photons alone (local control, 63% mixed beam v 52% photons (p=005); survival, 42% v 27% (p=005); and disease specific survival, 56% v 42% (p=004)). Whether or not differences between arms in a clinical trial achieve significance depends on both the number of patients in the trial and the observed differences in end points. For the above variables all differences were significant based on a two sided MantelHaenszel log rank test.4 Hence, the allegation by Drs MacDougall and Arnott that the numbers of patients were too small to draw valid conclusions is also incorrect. Differences in factors related to patients make comparisons between studies carried out at other times and places extremely hazardous. This is themain reason why randomised trials are conducted. Morbidity related to treatment was discussed in detail in both the five year and the eight year reports.2" Both acute and late morbidity were evaluated according to the joint scoring scheme of the Radiation Therapy Oncology Group and European Organisation for Research on the Treatment of Cancer and were found to be equivalent in the two arms. In all there were six reactions scored as severe or greater in the mixed beam arm and five in the photon arm. The only fatal complication occurred in the photon arm-a patient underwent a diverting colostomy, developed septicaemia, and died. To paraphrase Drs MacDougall and Arnott, the "morbidity watchdog did not bark because there was nothing to arouse it." This dog barked at Edinburgh because low energy neutrons alone were used unadvisedly to treat deep seated pelvic tumours.' The study concluded that the mixed beam form of treatment offered improved local control and survival for locally advanced prostate cancer at no increased morbidity. It was discontinued when new high energy cyclotrons capable of isocentric treatment became available in the United States and was replaced with a new study comparing neutrons alone and conventional photon irradiation. The planners of the protocol thought that the most important issue was not whether one form of neutron treatment was better than another (as would have been the case if the mixed beam arm had been the control arm of a new trial) but to ask the question again with larger patient numbers. The new trial has thus far accrued about 180 patients. If it confirms the results of the earlier study we predict that neutron radiotherapy will move to the private medical sector in the United States. This is the only way that sufficient numbers of machines can be built to accommodate the resulting patient load. GEORGE E LARAMORE Department of Radiation Oncology, University of Washington Medical Center, Seattle, WA 98195, United States I MacDougall RH, Arnott SJ. The cyclotron saga. Br Med J 1990;300:1721. (30 June.) 2 Laramore GE, Krall JM, Thomas FJ, Griffin TW, Maor MH, Hendrickson FR. Fast neutron radiotherapy for locally advanced prostate cancer: results of an RTOG randomized study. IntJ7 Radiat Oncol Biol Phys 1985;11:1621-7. 3 Russell KJ, LaramoreGE, KrallJM,etal. Eightyears' experience
with neutron radiotherapy in the treatment of stages C and D prostate cancer: updated results of the RTOG 7704 randomized clinical trial. Prostate 1987;11:183-93. 4 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease.J7NCI 1974;22:719-48. 5 Duncan W, Williams JR, Kerr GR, et al. An analysis of the radiation related morbidity observed in a randomized trial of neutron therapy for bladder cancer. Int J Radiat Oncol Biol Phts 1986;12:2085-92.
Precautions taken by orthopaedic surgeons to avoid infection with HIV and hepatitis B SIR,-The Editor's Choice and the article by Mr M J LeF Porteous in the same issue' criticise the NHS occupational health departments for allegedly not offering hepatitis B vaccine to surgeons. Few would disagree that this group should be offered vaccine and post-immunisation antibody testing. It is not reasonable to place all of the responsibility for failure to offer vaccine on occupational health staff. Although occupational physicians have the responsibility of advising their districts, the authority has to decide whether to take the advice and meet the resource implications. Regrettably, some health authorities have chosen to ignore the advice on financial grounds, and the responsibility for this rests squarely on them. M S GATLEY North Manchester General Hospital, Manchester M8 6RB
1 Porteous MJLeF. Operating practices of and precautions taken by orthopaedic surgeons to avoid infection with HIV and hepatitis B during surgery. Br MedJ 1990;301:167-9. (21 July.)
Management of hypovolaemic shock SIR,-Dr Peter J F Baskett's letter seems to be seriously misleading when he states that a venous cut down requires training and practice.' Of course, training and practice are desirable, but it is important for doctors of all disciplines to know that when a patient is dying of hypovolaemic shock it is perfectly appropriate for them to attempt a cut down on to a vein if venous access fails even if they have no experience. I have found this a feasible and lifesaving procedure. In the absence of contraindications the most suitable site for novices (such as myself) is the long saphenous vein as it courses anterior to the medial malleolus. As many surgeons are aware, cut downs at this site have saved many lives. It is worth quoting from one of our most well known surgical anatomists on this issue: "From the practical point of view the position of the long saphenous vein immediately in front of the medial malleolus is perhaps the most important anatomical relationship. No matter how collapsed and obese, or young and tiny the patient, the vein can be relied upon to be available at this site when urgently required for transfusion purposes."2 No one would suggest that someone tries to insert a central venous line without training and practice as so much damage could be done to vital structures, but at the ankle there is not much to lose and everything to gain if the patient is already dying. All that the doctor needs is a scalpel and artery forceps. He or she does not need to know how to tie surgical knots or use sutures. Such knowledge is helpful but in an emergency the cannula can simply be held in place while the doctor waits for skilled help to arrive-in those circumstances when postponing action even for a few minutes would be likely to be fatal. Of course it is necessary to bear in mind that the leg is not the ideal site when there is bleeding from pelvic or
BMJ
VOLUME
301
11 AUGUST 1990
