alcohol consumption may influence onset of the menopause

Self monitoring of glucose by people with diabetes

Urine testing provides only historicalinformation

Editor—Marilyn Gallichan raised someimportant issues concerning self monitor-ing of blood glucose concentrations butgreatly underestimated its benefits in termsof health, safety, and quality of life to patientstaking insulin.1 The hypoglycaemic potentialof insulin is a constant worry to all patientswith insulin dependent diabetes, buthypoglycaemia can be confirmed only ifblood is tested. Urine testing will not help.

The comparison between the merits oftesting blood glucose with those of testingurine glucose missed the point that the twotests provide quite different information andshould not be regarded as old versus new oraccurate versus inaccurate. Urine tests arerelatively cheap and easy to use, but theirinterpretation for patients with insulindependent diabetes requiring tight controlis difficult because glucose can often befound in the urine at the same time as thepatient has hypoglycaemia.

While urine testing can be helpful,patients need to be taught that it provides

historical information. Thus data from urinetests on one day can be used to predict thenext day’s treatment and dietary and activityschedules but should never be used to deter-mine current action.

Blood testing is of benefit not only whenunexpected variations occur but also forthose patients who are prepared to performtests in order to “see” and thus learn abouthow their metabolism deals with individualmeals and activities. Exercise has profoundeffects, particularly on those patients whoare taking insulin. Blood testing before andafter sport is essential if hypoglycaemia is tobe avoided.

Gallichan quotes a patient who said “Ido not worry about my day to day controlprovided the [haemoglobin A1] readingsstay good"; this is a flawed philosophy. Manypatients have only annual measurements ofhaemoglobin A1. Is it acceptable for patientsto tolerate poor control (which may havebeen present for a year) when they couldhave been alerted much earlier if blood test-ing had been started?

Gallichan also says that “only a relativelysmall proportion of [patients with insulindependent diabetes] have the need or desireto make frequent adjustments to theirinsulin dosage.” This population numbersabout 170 000, so a small proportion maynumber many thousands. It would be ashame to see a decrease in self monitoringof blood glucose as a consequence ofmisconceptions.Nigel J Crossland*4 Furness Close, Holmes Chapel, CheshireCW4 7LG*Nigel Crossland is a patient with over 32 years’experience of insulin dependent diabetes.

1 Gallichan M. Self monitoring of glucose by people withdiabetes: evidence based practice. BMJ 1997;314:964-7.(29 March.)

Self monitoring improves quality of lifeand prognosis of people with diabetes

Editor—Marilyn Gallichan discusses selfmonitoring of blood glucose concentrationsby people with diabetes.1 Through discus-sions with people with insulin dependentdiabetes we have arrived at four recommen-dations for monitoring blood glucose.x Measure your blood glucose if knowing itwill help you to solve a specific problem—forexample, when suspecting hypoglycaemia,when ill, or before driving.x Many patients find it useful to measureblood glucose once or twice daily, especially

during holidays, during and after exercise,during and after alcohol intake, etc. If receiv-ing intensive treatment, you will find thatmeasuring blood glucose before every mealto adjust the insulin dose gives freedom ineating and excellent glycaemic control.x When your regimen is being changedbecause the haemoglobin A1c concentrationis high or because of changes in yourlifestyle, frequent blood glucose monitoringis advisable.x In women with diabetes, frequent moni-toring before and during pregnancy isessential to achieve normoglycaemia.

It is essential that blood glucose moni-toring is done because the person withdiabetes wants to know the results and notbecause the doctor has asked him or herto do it. Glucose monitoring will alertpatients when the treatment is not effective,and our guidelines often also apply topatients with non-insulin dependentdiabetes who are treated with insulin. Oneof us has lived with insulin dependentdiabetes for almost 45 years and enjoys a lifefree from complications.

Gallichan informs us that, in 1995,£42.6m was spent on home monitoring inBritain. If a price of 25p a test is assumed,this sum paid for 170 million glucose strips.In Britain at least 600 000 people havediabetes, which means that 283 tests couldbe performed per person with diabetes eachyear, or one test every one to two days. If thefifth of patients with diabetes who are insulindependent are the only ones who shouldtest their blood glucose this would meanthat around 1400 tests a year could beperformed, or three to four tests daily. Thesenumbers of tests are not too far from therecommendations given by Gallichan.

Home blood glucose monitoring is acost effective, simple, and relatively cheaptool for improving the quality of life and theprognosis of all people with insulin depend-ent diabetes and many with non-insulindependent diabetes. All controlled trialsshowing the beneficial effects of nearnormalisation of glycated haemoglobinconcentrations have used self monitoring asan important tool.2-4 It is for the sceptics toprove that tight glycaemic control main-tained over years can be achieved by othermeans.Maria de Alva President elect, International DiabetesFederation,Mol Del Valle Ote 400, Garza Garcia, Nuevo Leon,Mexico

Jak Jervell President, International Diabetes FederationMedical Department B, Rikshospitalet, Oslo,Norway

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

184 BMJ VOLUME 315 19 JULY 1997


2 Wang PH, Lau J, Chalmers TC. Meta-analysis of effects ofintensive blood glucose control on late complications oftype 1 diabetes. Lancet 1993;329:1306-9.

3 DCCT Research Group. The effect of intensive treatmentof diabetes on the long-term complications of insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977-86.

4 Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S,Motoyoshi S, et al. Intensive insulin therapy prevents theprogression of diabetic microvascular complications inJapanese patients with non-insulin-dependent diabetesmellitus: a randomised prospective 6-year study. DiabetesRes Clin Pract 1995;28:103-17.

Self monitoring is vital for people withimpaired awareness of hypoglycaemia

Editor—In her article questioning the valueof self monitoring of blood glucose fordiabetic management, Marilyn Gallichanstates that “newly diagnosed patients of allages are usually taught to measure theirblood glucose concentrations.”1 This pre-sumably describes her local clinic policy andis unlikely to represent national practice.In our clinic, patients are generally taughtroutine blood glucose testing only if theyhave insulin treated diabetes or are likely torequire treatment with insulin in theforeseeable future. Most patients withnon-insulin dependent diabetes monitorurine glucose tests, which suffice whenglycaemic control is stable. However, themain premise of Gallichan’s article, which isbased on a few, limited, clinical studies, isthat self monitoring of blood glucose is notsuperior to urine testing because it does notimprove glycaemic control and incursunnecessary cost. This undervalues one ofthe principal tenets of blood glucosetesting—namely, the early detection, andconfirmation, of hypoglycaemia.

Hypoglycaemia is the side effect of insu-lin treatment that patients fear most and isthe main factor limiting the maintenance ofgood glycaemic control. The diabetes con-trol and complications trial showed aninverse correlation between glycated hae-moglobin concentration and the frequencyof severe hypoglycaemia.2 Strict glycaemiccontrol and long duration of insulindependent diabetes are both associated withthe development of impaired awareness ofhypoglycaemia, causing a sixfold increase inthe frequency of severe hypoglycaemia.3 It isvital for people with impaired awareness ofhypoglycaemia to monitor blood glucosefrequently to detect asymptomatic bio-chemical hypoglycaemia and so take appro-priate action.

The risk of nocturnal hypoglycaemia,which is often asymptomatic, is greatlyincreased when blood glucose at bedtime is< 6.0 mmol/l.4 All patients treated withinsulin should therefore be encouraged toestimate their blood glucose before going tosleep. Furthermore, a knowledge of the pre-vailing blood glucose concentration isimportant in everyday situations wherehypoglycaemia could be dangerous, such asdriving, and during some employment andsporting activities.

Blood glucose monitoring is essential inchildren with diabetes, in whom daily carbo-hydrate intake and physical exertion often

vary considerably and the development ofhypoglycaemia may be unpredictable. It isonly by frequent monitoring of bloodglucose that parents can determine anappropriate daily dose of insulin and avoidlow blood glucose concentrations.

We agree that diabetic patients shouldnot be taught self monitoring of bloodglucose indiscriminately, but its prophylacticvalue in avoiding severe hypoglycaemia isfundamental to the management of insulintreated diabetes.Mark W J Strachan Clinical research fellowKay Malloch Specialist diabetes sisterBrian M Frier Consultant physicianDepartment of Diabetes, Royal Infirmary ofEdinburgh, Edinburgh EH3 9YW


2 Diabetes Control and Complications Trial ResearchGroup. The effect of intensive treatment of diabetes on thedevelopment and progression of long-term complicationsin insulin-dependent diabetes mellitus. N Engl J Med1993;329:977-86.

3 Gold AK, MacLeod KM, Frier BM. Frequency of severehypoglycemia in patients with type 1 diabetes withimpaired awareness of hypoglycemia. Diabetes Care1994;17:697-703.

4 Pramming S, Thorsteinsson B, Bendtson I, Ronn B, BinderC. Nocturnal hypoglycaemia in patients receiving conven-tional treatment with insulin. BMJ 1985;291:376-9.

Patients with non-insulin dependentdiabetes should monitor urine ratherthan blood glucose

Editor—Marilyn Gallichan’s article aboutabuses of blood glucose self monitoring wasoverdue.1 We face the same problem in Ger-many. Polls and Boehringer Mannheim’ssales figures indicate that in one year (1994)about 305 million blood glucose strips wereprescribed, costing about DM 450m(£169m); this was paid by the sick funds.About half of these strips were prescribed topatients with non-insulin dependentdiabetes (table), and this proportion isincreasing.

We have followed up 97 patients withnon-insulin dependent diabetes, who had tobe changed from tablet to insulin treatmentand who had participated in a teaching pro-gramme emphasising self monitoring ofblood glucose.2 After two years their haemo-globin A1c concentration was around 9%,which is equivalent to an average blood glu-cose concentration at the renal threshold ofabout 8-12 mmol/l. Eighty seven of thepatients continued monitoring their bloodglucose once or twice daily during the weekbefore follow up examination, but only 29had ever adjusted their insulin dose them-selves. Thus self monitoring of bloodglucose is neither mandatory for insulintreatment in non-insulin dependentdiabetes patients nor helpful.3

We agree with Gallichan that selfmonitoring of glycosuria is more appropri-

ate in patients with non-insulin dependentdiabetes and much less costly. We dispute,however, that strips for testing urine glucosethat do not indicate urine glucose concen-tration quantitatively are “unsuitable.” Byaiming at blood glucose concentrationsbelow the renal threshold of glucose, manypatients with non-insulin dependentdiabetes can achieve fasting blood glucoseconcentrations of < 15 mmol/l or freedomfrom hyperglycaemic symptoms.4 Strips thatindicate that the renal threshold of glucosehas been passed by showing traces (0.1%) ofglycosuria (Glucose P/N strips (Machereyand Nagel) or Clinistix (Bayer Diagnostics))would thus be suitable for this large group ofpatients. Such strips can be extremely cheap(about 0.03 DM (1p) each5 ).Ernst Chantelau Senior registrarStefan Nowicki Medical studentDiabetesambulanz MNR-Klinik, Postfach 10 10 07,D-40001 Düsseldorf, Germany


2 Jörgens V, Grüsser M, Kronsbein P. Mit Insulin geht’s mirwieder besser. Für Typ-2 Diabetiker, die Insulin spritzen.Mainz: Kirchheim Verlag, 1993.

3 Janz H, Muller UA, Pohle B, Rauchfuss J. Wie sinnvoll istdie Blutglukose-Selbstkontrolle bei alteren insulinbehan-delten Patienten mit Typ-2 Diabetes? Diabetes undStoffwechsel 1995;4:162.

4 Turner R, Cull C, Holman R. United Kingdom prospectivediabetes study 17: a 9-year update of a randomized,controlled trial on the effect of improved metaboliccontrol on complications in non-insulin-dependentdiabetes mellitus. Ann Intern Med 1996;124(1 pt 2):136-45.

5 Chantelau E. Diabetes treatment in developing countries.Lancet 1993;342:620.

Author’s reply

Editor—In general I agree with the pointsmade by the authors, and would differ onlyin my emphasis. I acknowledge the variouswell known benefits of blood glucose moni-toring, but the main purpose of my paperwas to highlight the problems that can arisewhen it is assumed that these benefits applyto the entire diabetic population. Theinappropriate use of blood testing wastesresources and can cause psychologicalharm. To achieve the maximum benefit fromself monitoring of glucose, better educationof people with diabetes and their profes-sional carers is needed.

Ernst Chantelau and Stefan Nowickipoint out that Clinistix (Bayer Diagnostics)may sometimes be an adequate monitoringmethod if the aim is simply to detect thepresence or absence of glycosuria. InBritain, however, both Diastix (Bayer Diag-nostics) and Diabur-Test 5000 (BoehringerMannheim), at 4.5p per strip, are slightlycheaper than Clinistix, at 5.5p per strip. It istherefore cheaper to perform a urine test,which offers a wider range of results.Marilyn Gallichan Diabetes specialist nurseEast Cornwall Hospital, Bodmin, CornwallPL31 2EN

Estimated use of self monitoring of blood glucose in Germany, 1994

Type of diabetes Total No of patients% Of patients using

self monitoring No of tests/week

Insulin dependent 200 000 85 16

Non-insulin dependent, treated with insulin 620 000 53 7

Non-insulin dependent, treated without insulin 2 950 000 19 1-2

Letters

185BMJ VOLUME 315 19 JULY 1997

Multiple myeloma

Surgery is often more effective thananalgesia for mechanical pain

Editor—Charles R J Singer’s recent discus-sion of multiple myeloma and related condi-tions in the ABC of clinical haematologyincludes a radiograph of a vertebral crushfracture due to myeloma with a caption say-ing: “Bone pain from mechanical effects ...often necessitates long term treatment withstrong analgesia.”1 It is well recognised bysurgeons regularly treating such patientsthat the most effective solution for mechani-cal pain may be a mechanical one—that is,surgery. My colleagues and I have shownthat this is a good option in many patientswith vertebral metastases.2 We appeal to allphysicians to refer any patient with vertebralmetastases to a spinal surgeon, at least for anopinion.S J Krikler Consultant orthopaedic surgeonCoventry and Warwickshire Hospital, CoventryCV1 4FH

1 Singer CRJ. Multiple myeloma and related conditions. BMJ1997;314:960-3. (29 March.)

2 Krikler SJ, Marks DS, Thompson AG, Merriam WF,Spooner D. Surgical management of vertebral neoplasia:who, when, how and why. Eur J Spin Surg 1994;3:342-6.

Surgical stabilisation often provides goodpain relief

Editor—There are two aspects to the man-agement of multiple myeloma: treatment ofthe disorder and palliative management ofsecondary complications such as pathologi-cal fracture or spinal instability. This secondtopic has not been fully covered in thearticle by Charles R J Singer.1

It is not possible to advise on treatmentpurely from looking at radiographs sincethere might be other factors which influencetreatment. However, the patient with multi-ple fractures of the humerus shown in thefirst radiograph might have received betterand more rapid pain control from internalstabilisation of the entire humerus, providedthat he or she was otherwise fit. The patientwith severe back pain, whose radiograph ofthe spine was shown second in the article,probably would have obtained more rapidand better pain relief from spinal stabilisa-tion. It was noted in the caption that thebone pain necessitated “long term treatmentwith strong analgesia despite response tochemotherapy.”1 Surgical stabilisation ofsuch lesions often gives the patient totalpain relief and may obviate the need foranalgesics.

O’Donoghue et al reviewed 269 womenwith breast cancer who had developed bonemetastases.2 They identified 82 episodes ofstructurally significant bone destruction in47 women. The radiographs shown in thereview by Singer are of lesions which showstructurally significant bone destruction, andthe same principles apply to the manage-ment of such lesions in patients withmyeloma as those in patients with skeletalmetastases from breast cancer.

O’Donoghue et al concluded that aclinical review by an orthopaedic surgeon

would have been appropriate in 89% of epi-sodes but was sought in only 46%.2 Surgerywould have been feasible for 65% of theepisodes of structurally significant bonedestruction but was carried out in only 31%.Bracing would have been appropriate in40% of cases but was provided in only 18%.They also found that there was a signifi-cantly longer symptom free survival inpatients whose lesions were treated with sur-gery plus radiotherapy compared with thosewhose lesions were treated with radio-therapy alone, as well as a significantly lowerincidence of symptom recurrence and agreater survival.

Much has been written about the ortho-paedic management of pathological frac-tures and spinal instability, including chap-ters in the Oxford Textbook of Oncology3 andthe Oxford Textbook of Palliative Medicine,4 yetpatients are still being inadequately treated.Why has oncological education failed somiserably with respect to the optimum man-agement of patients with structurally signifi-cant bone destruction secondary to malig-nant disease?Charles S B Galasko Professor of orthopaedic surgeryUniversity of Manchester, Department ofOrthopaedic Surgery, Clinical Sciences Building,Hope Hospital, Salford M6 8HD

1 Singer CRJ. Multiple myeloma and related conditions. BMJ1997;314:960-3. (29 March.)

2 O’Donoghue DS, Howell A, Bundred NJ, Walls J.Orthopaedic management of structurally significant bonedestruction in breast cancer bone metastases. J Bone JointSurg 1997;79B(suppl l):98.

3 Peckham M, Pinedo HM, Veronesi U, eds. Oxford textbook ofoncology. Oxford: Oxford University Press, 1995:2286-95.

4 Doyle D, Hanks GWC, MacDonald N. Oxford textbook ofpalliative medicine. Oxford: Oxford University Press,1993:274-82.

Catheters smaller than 24French gauge can be used forchest drainsEditor—It is a pity that, in their editorial onreducing morbidity from chest drains,Jonathan Hyde and colleagues have per-petuated another medical myth.1 Their con-tention that a 24 French gauge catheter isadequate for air or low viscosity effusionsseems to ignore much of the evidence thatpleural collections, including empyemas,can be successfully drained with muchsmaller catheters.2 3 It is my experience thatlow viscosity fluid drains adequately through8 French catheters. Although I am not awareof any study showing that smaller cathetersare associated with reduced morbidity, suchcatheters are more comfortable for patientsand better tolerated. Perhaps the authors donot refer to the use of smaller cathetersbecause many of the reports of their useappear in the radiological literature. In thiscontext the authors also failed to acknowl-edge the value of ultrasonography when thesite is selected for placement of the drain.Although not a prerequisite for the insertionof drainage catheters, ultrasonography canidentify loculation within cavities, differenti-ate pleural thickening from fluid, detectunderlying pleural masses, and localise the

hemidiaphragm. This information will aidthe safe insertion of chest drains, which wasthe aim of the editorial.P M Taylor Consultant radiologistCentral Manchester Healthcare NHS Trust,Department of Clinical Radiology, ManchesterRoyal Infirmary, Manchester M13 9WL

1 Hyde J, Sykes T, Graham T. Reducing morbidity from chestdrains. BMJ 1997;314:914-5. (29 March.)

2 Reinhold C, Illescas FF, Atri M, Bret PM. Treatment ofpleural effusions and pneumothorax with catheters placedpercutaneously under imaging guidance. Am J Roentgenol1989;152:1189.

3 Lee KS, Im J-G, Kim YH, Hwang SH, Bae WK, Lee BM.Treatment of thoracic multi-loculated empyemas withintracavitary urokinase: a prospective study. Radiology1991;179:771-5.

Screening for prostate cancer

Randomised trials of treatment in earlydisease are important

Editor—The debate about the desirabilityof screening for prostate cancer as outlinedin the editorial by Steven H Woolf hascentred on the ability to detect, and thencure, localised disease.1 Early detection iscertainly possible, albeit at a cost. Whetherearly radical treatment can cure the cancerremains to be seen.

Cure is not the only objective of cancercare, however. Prolongation of life andimprovement in the quality of life in theabsence of cure are also valuable. In the caseof prostate cancer, improving the quality oflife translates into delaying the appearanceof metastases and local symptoms andavoiding iatrogenic morbidity. Radicalsurgery or radiotherapy may delay meta-static spread and retard localised disease butwe do not yet know. Since the recentpublication of a randomised trial by theMedical Research Council in the BritishJournal of Urology,2 however, the state-ment that there is no evidence that earlydetection improves outcome needs to beexpanded on.

Early androgen deprivation improvesoutcomes for patients with locally advanced(T2-4) or metastatic disease. The MedicalResearch Council trial found improvementsin all cause mortality, disease specificmortality, rate of local and metastaticprogression, and catastrophic complications(spinal cord compression, ureteric obstruc-tion, and pathological fracture). Benefitswere measurably greater for patients withlocally advanced disease who did not havedemonstrable metastases than for those withestablished metastases. Similar results hadbeen suggested but not established conclu-sively by earlier work.3 4

The possibility now arises that screeningmight be justified by the early detection ofnon-curable disease. Hypothetically, the rea-son for wanting to detect localised prostatecancer might be to watch it until it becomesadvanced (when an intervention of provedeffectiveness is available), rather than toattempt a cure by radical prostatectomy orradiotherapy (neither of which has goodevidence to support it). The problem of psy-chological morbidity resulting from early

Letters


diagnosis followed by inaction would remain.There would also be problems with iatrogenicmorbidity caused by androgen deprivation ifthis were instituted very early and maintainedfor many years. Both of these things wouldoffset the benefits of slowing disease progres-sion and lengthening life. Caution will also benecessary in interpreting the results ofrandomised trials of screening currentlyunder way in the United States and Europe;these include aggressive, mainly surgical,treatment for localised disease. A potentialalternative explanation for a positive resultmight be the earlier institution of adjuvant orpalliative hormonal treatment. The case forrandomised trials of treatment in earlydisease remains strong, despite oppositionfrom many urologists.Rowan H Harwood Consultant physicianUniversity Hospital, Nottingham NG7 2UH

1 Woolf SH. Should we screen for prostate cancer? BMJ1997;314:989-90. (5 April.)

2 MRC Prostate Cancer Working Party Investigators Group.Immediate versus deferred treatment for advancedprostatic cancer: initial results of the Medical ResearchCouncil trial. Br J Urol 1997;79:235-46.

3 Fellows GJ, Clark PB, Beynon LL, Boreham J, Keen C,Parkinson MC, et al. Treatment of advanced localisedprostate cancer by orchiectomy, radiotherapy or combinedtreatment. A Medical Research Council Study. Br J Urol1992;70:304-9.

4 Byar DP. The Veterans Administration Co-operativeUrological Research Group studies of cancer of the pros-tate. Cancer 1973;32:1126-30.

Early screening is important despite lackof data from trials

Editor—In his editorial Steven H Woolfstates (a) that routine screening for prostatecancer is unwarranted because there is nogood evidence that treatment improves out-come, (b) that the complications of treat-ment may offset the potential benefits, and(c) that the decision to screen requires anassessment of the benefits and harms tosociety.1

However, the balance of evidence isshifting to indicate that early detection andtreatment of prostate cancer improvesoutcome. The National Cancer Instituterecently reported a 6.3% overall decline inmortality from prostate cancer from 1991 to1995.2 White men showed greater decline inmortality than did black men, who lessfrequently seek screening and treatment.These decreases are almost certainly due toscreening on the basis of prostate specificantigen and to effective treatment.

A recent population based analysis of59 876 men treated for localised disease(from 1983 to 1992) showed that 10 yearsurvival rates were higher in men with inter-mediate or high grade cancer who weretreated by prostatectomy than in those whowere treated with radiotherapy or managedconservatively.3 This strongly suggests thatprostatectomy improves survival.

Recent studies show that, althoughpatients actively treated for cancer detectedby screening for prostate specific antigenhad more sexual dysfunction than untreatedpatients, there were no differences in urinarydysfunction, general health, emotional well-being, or satisfaction with treatment.4 Theoverall quality of life of surgically treated

patients was comparable to that of thegeneral population.

The evidence based approach considersscreening from the societal perspectiverather than the perspective of the patientdestined to develop prostate cancer. Hard-line insistence on the proof of the benefits ofscreening unrealistically and callously setsthe bar too high for the interests of patientswith prostate cancer and the doctorsresponsible for their health care. Notscreening would result in unnecessarysuffering and a painful death for countlessmen before the evidence to convincehardliners that early detection is beneficial isin hand. Current randomised clinical trialsare fatally flawed, and they are unlikely toprovide definitive data.

Early detection and treatment remaingoals in the approach to most cancers eventhough the benefits have not been formallyestablished for many cancers. Given themagnitude of the problem, doctors shouldoffer screening to men who may benefitfrom the opportunity for cure. A balancedassessment of the evidence suggests that inappropriately selected men screening willallow curative treatment of presymptomaticcancers that otherwise would cause substan-tial morbidity and mortality.William J Catalona ChiefDivision of Urologic Surgery, WashingtonUniversity School of Medicine, St Louis, MO 63110,[email protected] of interest: My research is supported in part by agrant from Hybritech, a manufacturer of assays for prostatespecific antigen.

1 Woolf SH. Should we screen for prostate cancer? BMJ1997;314:989-90. (5 April.)

2 Hoeksema MJ, Law C. Cancer mortality rates fall: a turningpoint for the nation. J Natl Cancer Inst 1996;88:1706-7.

3 Lu-Yao GL, Yao S-L. Population-based study of long-termsurvival in patients with clinically localised prostate cancer.Lancet 1997;349:906-10.

4 Smith DS, Schneider MK, Catalona WJ. Short-termoutcomes with screen-detected prostate cancer [abstract]. JUrol 1997;157(suppl):426.

Trials of streptokinase instroke depend on early,accurate diagnosisEditor—Paul Dorman and Peter Sander-cock’s survey of access to computed tomo-graphy provides scant reassurance to thosewith experience of trials of streptokinase inacute stroke.1 Randomised controlled trials(involving over 1200 patients) have providedample evidence of the hazard of strepto-kinase in stroke. The availability of com-puted tomography is only a minor factor inthe safe conduct of further trials. Despitecomputed tomography being carried outbefore treatment, intracerebral haemor-rhage induced by streptokinase caused 50additional early deaths per 1000 patients inthree recent randomised controlled trials.While survivors benefited, one may questionwhether doctors’ duty to individual patientsis compatible with risking a repeat of thisexperience in larger numbers.

How useful is computed tomography inhyperacute stroke? It excludes intracerebral

haemorrhage but is not a diagnostic test forischaemic stroke. Many scans are normalwithin six hours. Unlike with electro-cardiography in chest pain, no widelyavailable substitute exists for clinical diagno-sis. Diagnosis is incorrect, however, in a fifthof patients seen in hospital emergencyrooms2 and even higher among patientsseen by paramedical staff or in primary care.Giving streptokinase to patients who havenot had a stroke can be avoided only if staffexperienced in diagnosing stroke are imme-diately available.

The thrombolysis in acute stroke poolingproject and others have identified the risk ofhaemorrhage induced by treatment as beinggreatest when subtle computed tomographicsigns of extensive cerebral infarction areevident at presentation.3 4 These signs arenot yet recognised widely, and substantialdisagreement on the interpretation of earlycomputed tomograms has occurred evenamong specifically trained neuroradiologistsin trials.4 Few British centres have staffwith experience of interpreting computedtomograms in hyperacute ischaemic stroke.

The wide availability of computedtomography and the reported willingness touse it are welcome but are inadequate forfurther trials. Many issues regarding strepto-kinase in stroke remain unresolved, but fur-ther trials cannot be justified unless theyavoid the known dangers. A prerequisite isacute stroke services whose infrastructureincludes the 24 hour availability not just ofcomputed tomography but of stroke physi-cians and radiologists with diagnostic skillsin hyperacute stroke.Keith W Muir RegistrarDepartment of Neurology, Institute of NeurologicalSciences, Southern General Hospital, GlasgowG51 4TF

1 Dorman P, Sandercock P. Access to computed tomographyin British accident and emergency departments. BMJ1997;314:440-1. (8 February.)

2 Libman RB, Wirkowski E, Alvir J, Rao TH. Conditions thatmimic stroke in the emergency department. Implicationsfor acute stroke trials. Arch Neurol 1995;52:1119-22.

3 Hachinski V. Thrombolysis in stroke: between the promiseand the peril. JAMA 1996;276:995-6.

4 Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, vonKummer R, et al. Intravenous thrombolysis with recom-binant tissue plasminogen activator for acute hemisphericstroke: the European cooperative acute stroke study(ECASS). JAMA 1995;274:1017-25.

Tuberculosis treatment isexpensive for patients indeveloping countriesEditor—I was disappointed to read of theattitude of the World Health Organisationand others to the use of directly observedtreatment short course (DOTS) for tubercu-losis.1 Such treatment is described as a cheapand effective public health breakthrough.The failure of the method is blamed ondoctors explicitly by Sir John Croftonand implicitly by the director of the WorldHealth Organisation.

Tuberculosis is a major health problemin Kenya. Of the 70 medical beds currentlyunder my care, over one fifth are occupiedby patients with this disease. Patients are

Letters


more likely to complete treatment foreconomic reasons than for any other reason.Jessica Westall quotes the cost of directlyobserved treatment short course at $11(£6.90) per patient for a six month supply ofmedicines (rough current cost in Kenya).1

This ignores other costs that patients mustmeet: that of inpatient care if required at $4daily, travel to a designated centre at $1 daily,and the cost of syringes and needles forstreptomycin at $10 monthly. For patientswho are subsistence farmers, or unskilledlabourers earning $2 daily, these amountsare not minimal.

The epidemic of tuberculosis in thedeveloping world cannot be blamed on“ignorance and complacency.” Tuberculosisaffects the least affluent people in society:those who have the least resources to seekand complete treatment. Directly observedtreatment short course in developing coun-tries already uses drugs and protocols thatare unlicensed and unacceptable in devel-oped countries. Few individuals would be soignorant that they would not accept the bestmedical care if it were available andaffordable.

If the world’s developed countriescannot give financial support to low incomecountries for tuberculosis treatment frommotives of compassion and justice, perhapsfear of a worldwide epidemic might movegovernments and non-government organi-sations to act.Elizabeth Bevan Medical officer in chargePCEA Tumutumu Hospital, Private Bag, Karatina,Kenya

1 Westall J. Tuberculosis levelling off worldwide. BMJ1997;314:921. (29 March.)

Alcohol consumption mayinfluence onset of themenopauseEditor—Patrick McElduff and Annette JDobson have shown an apparent protectiveeffect of alcohol consumption on coronaryevents.1 We have reported two studies whichseem to show that alcohol consumption isassociated with a delay in menopausaldevelopment.2 3 In a cross sectional surveyof 2240 women aged between 45 and 49 wefound a significant association between alco-hol consumption and current menopausalstatus (defined as six months of amenor-

rhoea): women who consumed the leastalcohol were more likely to be postmeno-pausal.2 Controlling for possible confound-ing factors, such as social class or smokingstatus, did not materially alter the associ-ation. Two years later we followed up all thepremenopausal women, and we noted thatthose women who drank the least alcoholwere more likely to have become postmeno-pausal.3 Again, this association remainedafter adjustment for potential confoundingfactors. The table summarises the results ofthe univariate analyses.

Our data suggest that alcohol may have arole in menopausal development and thatthis association merits further research; it maybe a cofactor involved in the protective effectof alcohol on cardiovascular risk in women.David J Torgerson Research fellowCentre for Health Economics, University of York,York YO1 SDD

Marion K Campbell Senior statisticianRuth E Thomas Research assistantHealth Services Research Unit, University ofAberdeen, Aberdeen AB9 1FX

David M Reid Senior lecturer in rheumatologyDepartment of Medicine and Therapeutics,University of Aberdeen

1 McElduff P, Dobson AJ. How much alcohol and how often?Population based case-control study of alcohol consump-tion and risk of a major coronary event. BMJ1997;314:1159-64. (19 April.)

2 Torgerson DJ, Avenell A, Russell IT, Reid DM. Factorsassociated with the onset of menopause in women aged45-49. Maturitas 1994;19:83-92.

3 Torgerson DJ, Thomas RE, Campbell MK, Reid DM. Alco-hol consumption and age of maternal menopause areassociated with menopause onset. Maturitas 1997;26:21-5.

Chronic venous ulcer

Some references may be misleading

Editor—Niren Angle and John J Bergan areto be congratulated in their stated pursuit ofevidence based information in the treatmentof chronic venous ulcer.1 However, in what isnecessarily a brief summary of the topic, Imust take issue with their choice ofreference in the paragraph on excision andskin grafting.2 Their only reference illustrat-ing the success in covering venous ulcersusing split thickness skin grafting is aretrospective review of 26 patients treatedover three years. In only 10 of these patientswas the aetiology thought to be venousinsufficiency. Despite the small number ofpatients, healing rates are given in percent-ages and a rate of only 20% was achieved for

venous ulcers. This study neither fulfils thecriteria described in the methods section ofthe review nor provides evidence to supportthe use of skin grafting in the treatment ofchronic venous ulcer.

The next paragraph talks about excisionand free flap coverage. An unreferencedstatement indicates that a free tissue transferprovides its own venous system containing“hundreds of competent valves.” In a freetissue transfer where the venae comitantesof the flap are used for the microvenousanastomosis, the ability of small valves in avein of 2-3 mm diameter to withstand thepressure generated by a column of bloodfrom the right atrium through the abdomi-nal and pelvic veins down to the dysfunc-tional limb veins is questionable.

Finally, with regard to maintaining thehealing of chronic leg ulcers (whetherachieved by conservative or surgical means),even in a study reporting one of the besthealing rates in published reports, the recur-rence rate in 111 compliant patients withhealed ulcers using compression stockingswas 22% at one year.3 The rate of recurrenceof ulceration in non-compliant patients withhealed ulcers was 45%.

Even though no clinical review can becomprehensive by definition, I think that theabove points should be drawn to theattention of anyone reading this review.R P Cole Consultant plastic and reconstructivesurgeonOdstock Centre for Burns, Plastic andMaxillo-Facial Surgery, Salisbury District Hospital,Salisbury, Wiltshire SP2 8BJ

1 Angle N, Bergan JJ. Chronic venous ulcer. BMJ1997;314:1019-23. (5 April.)

2 Wood MK, Davies DM. Use of split-skin grafting in thetreatment of chronic leg ulcers. Ann R Coll Surg Engl1995;77:222-3.

3 Blair SD, Wright DDI, Backhouse CM, Riddle E, McCollumCN. Sustained compression and healing of chronic venousulcers. BMJ 1988;297:1159-61.

Hyperbaric oxygen treatment is a costeffective option

Editor—We were surprised at the dismissalof hyperbaric oxygen treatment by NirenAngle and John J Bergan in their clinicalreview article on chronic venous ulceration.1

They referred to a review article by Tibblesand Edelsberg,2 which quoted a study byHammarlund and Sundberg on the healingof chronic leg ulcers.3 This study found asignificant improvement when chronic legulcers were treated with hyperbaric oxygen.

The mechanism for ulceration isdescribed as the distal adherence, trapping,and activation of leucocytes which causestissue destruction. In an earlier editorialKindwall stated that hyperbaric oxygentreatment reduces white cell adhesion tocapillary walls after ischaemic insult.4 Kind-wall also described how hyperbaric oxygenstops lipid peroxidation—by sparing cellmembranes; reducing oedema in post-ischaemic muscle through the preservationof adenosine triphosphate; increasing redcell flexibility and improving white cellkilling of aerobic bacteria; and, probablymore importantly, stimulating capillary andcollagen formation. This explains the effec-

Association between alcohol consumption and onset of the menopause

Amount of alcohol consumed

% (No) of postmenopausal women

Cross sectional study2 Prospective study3

At least one drink a day 12% (12) 6% (3)

One drink most days but not every day 17% (16) 9% (5)

One drink 2 or 3 days a week 14% (67) 18% (48)

One drink 2 or 3 times a month 16% (49) 22% (35)

One drink about once a month 17% (25) 19% (17)

Only on special occasions (eg Christmas) 13% (41) 25% (23)

Never 27% (48) 25% (21)

Multivariate odds ratio* (95% CI) 0.86 (0.79 to 0.94) 0.91 ( 0.84 to 0.98)

*Adjusted for smoking, social class, parity, and age; reference group = those who never drank.

Letters


tiveness of hyperbaric oxygen as treatmentfor chronic venous ulceration.

We recognise that chronic leg ulcerationhas different causes and that treatmentneeds to be focused on the cause. However,whatever the cause, at the cellular level themechanism of tissue damage is the sameand is related to reduced oxygen delivery.

In the double blinded study by Hammar-lund and Sundberg, patients with chronicleg ulcers but without diabetes or largevessel arterial disease showed a 36%reduction in ulcer size at six weeks after acourse of hyperbaric oxygen compared witha 3% reduction in the control group treatedwith hyperbaric air.3

Treatment of leg ulcers is expensive. Itcosts an estimated £5000 per patient forthree months of conventional treatment.5

Augmentation of conventional treatmentwith hyperbaric oxygen reduces ulcerhealing time and should allow earlier,successful skin grafting of clean wounds. Acourse of 30 treatments of hyperbaricoxygen in the adjunctive management ofchronic venous ulcer in the United Stateswould cost about $9000.2 In British treat-ment units we estimate, using our own busi-ness plan, that these costs could be reducedby as much as half.

We believe that the use of hyperbaricoxygen treatment as an adjunct in the treat-ment of chronic venous ulceration is costefficient. It is denied patients because of alack of medical hyperbaric oxygen facilitiesand an ignorance of its value by medicalpractitioners.Andrew W McEwen Consultant in anaesthesiaMark B Smith Consultant in anaesthesiaDepartment of Anaesthesia, Royal LancasterInfirmary, Lancaster LA1 4RP


2 Tibbles PM, Edelsberg JS. Hyperbaric oxygen therapy:review article. N Engl J Med 1996;334:1642-6.

3 Hammarlund C, Sundberg T. Hyperbaric oxygen reducedsize of chronic leg ulcers: a randomized double blind study.Plast Reconstr Surg 1994;93:829-33.

4 Kindwall EP. Hyperbaric oxygen. BMJ 1993;307:515-6.5 Freak L, Simon D, Kinsella A, McCollum C, Walsh J, Lane

C. Leg ulcer care: an audit of cost-effectiveness. HealthTrends 1995;27(4):133-6.

Causes are often multifactorial and aholistic approach is required

Editor—Niren Angle and John J Bergan doa great service in underlining theimportance of early surgical intervention invenous ulcers, an approach we wholeheart-edly endorse.1 All too often referral tosurgeons takes place when long termconservative treatment has failed or whenulceration recurs, as happens in nearly allconservatively treated ulcers.2 The result isan intractable and often large ulcer embed-ded in a cuirass of chronic granulationtissue, which makes any form of treatmentdifficult.

We wish to qualify some aspects of themanagement of chronic venous ulcer out-lined by Angle and Bergan in relation toexperience and practice in the United King-dom, using the population of patients insouth east and central Scotland as anexample.

Venous disease is but one, albeit thelargest, component of a range of diseasesassociated with chronic leg ulcers. Arterialinsufficiency occurs in 20% of the popula-tion aged over 60.3 Rheumatoid disease,diabetes, and many other chronic problemsare common.2 4 5 Chronic leg ulcers, includ-ing the venous variety, therefore have multi-factorial causes in nearly all cases. Particu-larly important are disorders which interferewith venous function by their impact on thecalf pump mechanisms. A high proportionof patients with leg ulcer attending our clin-ics are obese or have gait disorders,arthropathies, neuropathies, foot deformi-ties, and are frail; furthermore, they havevarious physical, social, and psychologicalfactors contributing to chronic limbdependency or the inability to walk enoughto activate the calf pump mechanism.

The conclusions that we draw from thisexperience are that a holistic approach tothe treatment of patients with chronic legulcer is necessary; that every patient shouldbe referred at an early stage in the disorderfor diagnostic screening in a specialistcentre; that associated problems should becorrected as far as possible; and that venousinsufficiency should be corrected promptlywhen it is a predominant factor, as advo-cated by Angle and Bergan.

We believe, however, that the indicationsfor and benefits of certain surgical tech-niques—including perforator interruption—need to be evaluated in controlled trials.C V Ruckley ProfessorA W Bradbury Senior lecturerW Stuart Research fellowVascular Surgery Office, Department of Surgery,University of Edinburgh, Royal Infirmary ofEdinburgh, Edinburgh EH3 9YW


2 Callam MJ, Ruckley CV, Harper DR, Dale JJ. Chroniculceration of the leg: clinical history. BMJ 1987;290:1855-6.

3 Fowkes FGR, Callam MJ. Is arterial disease a risk factor forchronic leg ulceration? Phlebology 1994;9:87-90.

4 Nelzen O, Bergqvist D, Lindhagen A. Leg ulcer aeti-ology—a cross sectional population study. J Vasc Surg1991;14:557-64.

5 McRorie ER, Jobanputra P, Ruckley CV, Nuki G. Clinicalreview: leg ulceration in rheumatoid arthritis. Br JRheumatol 1994;33:1078-84.

Authors’ reply

Editor—We were pleased by the responseto our article on chronic venous ulcer.

R P Cole is quite correct that ourreference was an unfortunate choice. How-ever, at the University of Massachusetts,Worcester, Dunn et al have proved convinc-ingly that free flap transfer to a wellprepared bed which includes wide excisionof scar tissue, down to and including fascia,does indeed convert that area of the leg tonormal venous physiology while theremainder of the leg continues to havesevere venous dysfunction.1 Cole makes ourpoint that simple healing of venous ulcer isnot the ultimate goal. Prevention of recur-rence should be the objective of treatment ofsevere venous insufficiency.

While we appreciate the point made byAndrew W McEwen and Mark B Smithabout the use of hyperbaric oxygen, we are

sure that they would agree that this modal-ity of ulcer treatment has made virtually noimpact either in the United Kingdom orin the United States. Regardless of the find-ings of their interesting reference, it is clearthat doctors treating venous ulcerationhave found little use for hyperbaric oxy-genation unless they are involved with aninstitution that supports such an expensivetreatment.

Finally, we greatly appreciate the viewexpressed by the Edinburgh surgeons C VRuckley, A W Bradbury, and W Stuart. Wecontinue to look to them for leadership inthe diagnosis and treatment of severevenous insufficiency.2 Their plea for control-led trials, if heeded, will provide doctors car-ing for patients with venous insufficiencywith Level 1 evidence based information,which will greatly improve the care ofpatients with terribly disabling venousproblems.John J Bergan Professor of surgeryNorth Coast Surgeons Medical Group, 9850Genesee Avenue, Suite 560, La Jolla, CA 92037,USA

Niren Angle Resident in surgeryUniversity of California, San Diego, UCSD MedicalCenter, 200 West Arbor Drive, San Diego,CA 92103

1 Dunn RM, Fudem GM, Walton RL, Anderson FA Jr,Malhotra R. Free flap valvular transplantation forrefractory venous ulceration. J Vasc Surg 1994;19:525-31.

2 Bradbury AW, Ruckley CV. Foot volumetry can predictrecurrent ulceration after subfascial ligation of perforatorsand saphenous ligation. J Vasc Surg 1993;18:789-95.

Counselling should beprovided before parents aretold of presence ofultrasonographic “softmarkers” of fetal abnormalityEditor—Martin Whittle alluded to theanxiety caused to parents after the disclo-sure of a “soft marker”—that is, “a minor ...structural change which may indicate a riskof serious fetal anomaly but ... is probablyinconsequential”—following a routine ultra-sound scan.1 The anxiety caused can be bothconsiderable and long lasting, as shown bythe following case.

A couple were referred for amniocen-tesis during the wife’s second pregnancy onthe grounds of maternal age (35 years) andanxiety. Their 3 year old son played happilyduring the consultation. When his wife andson had left the room after the procedurethe husband confided that they had optedfor amniocentesis to avoid having another“brain damaged” child. On questioning itbecame apparent that an ultrasoundexamination before their son’s birth hadshown a choroid plexus cyst. Despite havinga healthy child, the husband remainedconvinced that this cyst would cause his sonto be disabled.

The psychological sequelae of falsepositive results of routine prenatal orneonatal screening tests are considerable;numerous studies have reported high levelsof parental anxiety and worry even after

Letters


subsequent normal test results.2 Worrieshave been reported up to 12 months afterfalse positive results of screening.3

In the case of ultrasound screening,these psychological costs may be exacer-bated by the visual imagery which is an inte-gral part of the procedure.4 It is not easy toconvey to couples the idea that an indicatorof abnormality that can be visualised isprobably inconsequential. Women whosefetuses are deemed to be at an increasedrisk of chromosomal abnormality on thebasis of an ultrasound examination mightinterpret this result quite differently to thosereceiving a report of a similar risk as a resultof maternal serum screening. We suggestthat specialist counselling methods need tobe developed, and we support the report ofthe Royal College of Obstetricians andGynaecologists highlighting the fact thatfurther research is needed into methods ofprescreening counselling.5

The case above not only highlights a fail-ure of counselling at the appropriate time butalso raises the question of how often thedisclosure of the presence of a soft markerdoes more harm than good. Such cases haveled us to stop reporting isolated choroidplexus cysts in the fetuses of younger womenas we believe that the anxiety caused to thesewomen far outweighs the potential gain.Gerald Mason Consultant in fetomaternal medicineClarendon Wing, Leeds General Infirmary, LeedsLS2 9NS

Catherine Baillie Research assistantDepartment of Psychology, University of Leeds,Leeds LS2 9JT

1 Whittle M. Ultrasonographic “soft markers” of fetalchromosomal defects. BMJ 1997;314:918. (29 March.)

2 Marteau TM, Cook R, Kidd J, Michie S, Johnston M, SlackJ, et al. The psychological effects of false-positive results inprenatal screening for fetal abnormality: a prospectivestudy. Prenat Diagn 1992;12:205-14.

3 Tymstra T. False positive results in screening tests:experience of parents of children screened for congenitalhypothyroidism. Fam Pract 1986;3:92-6.

4 Marteau TM. Screening in practice: reducing thepsychological costs. BMJ 1990;301:26-8.

5 Royal College of Obstetricians and Gynaecologists. Reportof the RCOG working party on ultrasound screening for fetalabnormalities. London: RCOG, 1997.

Harm resulting from screeningis likely to be high whereprevalence of breast canceris lowEditor—Matti Hakama and colleaguesevaluated an organised programme ofscreening for breast cancer in Finland.1 Theyreport a 24% reduction in mortality frombreast cancer due to screening, which, asthey point out, is close to the protectiveeffect reported in the early randomised con-trolled trials. In their conclusions they implythat having a breast cancer screeningprogramme is worth while and a good useof health service resources.

Such a conclusion must, however, betreated with caution. In this study the authorscalculate that about 200 000 women werescreened to prevent 20 deaths from breastcancer. They also quote other benefits ofscreening but make no mention of the well

documented disadvantages.2 The most seri-ous potential harm is that attributable to afalse positive result, which can occur in about14% of those screened.3 Thus the 20 deathsprevented must be balanced against the anxi-ety, trauma, and potential operative complica-tions encountered by an estimated 28 000women with a false positive result. To this canbe added the cost to the health service of theinvestigations that these women will have had.

Hong Kong does not yet have an organ-ised mammography screening programme.Since the prevalence of breast cancer is rela-tively low compared with that in other coun-tries,4 the harm resulting from screening islikely to be high. This factor must also beconsidered when a public health decision ismade on whether to introduce a screeningprogramme.Peymane Adab Lecturer in public health medicineDepartment of Community Medicine, University ofHong Kong, Hong Kong

1 Hakama M, Pukkala E, Heikkila M, Kallio M. Effectivenessof the public health policy for breast cancer screening inFinland: population based cohort study. BMJ1997;314:864-7. (22 March.)

2 Wright CJ, Mueller CB. Screening mammography andpublic health policy: the need for perspective. Lancet1995;346:29-32.

3 Baines CJ, McFarlane DV, Miller AB. Sensitivity andspecificity of first screen mammography in 15 NBSScentres. J Can Assoc Radiol 1988;39:273-6.

4 Hong Kong Cancer Registry. Cancer incidence in HongKong 1992. Hong Kong: Hospital Authority, 1996.

Minimising factitioushyperkalaemia

Samples should be centrifuged aftercollection in general practices

Editor—The advice given by J D Johnstonand S W Hawthorne on how to minimisefactitious hyperkalaemia in blood samplesfrom patients in general practice is mislead-ing and potentially dangerous.1 We agreethat samples should never be refrigerated,but it is not safe to assume that plasmapotassium concentrations will remainunchanged when samples are left at roomtemperature. One of us (PWM) showed thatin hot weather potassium concentrationsmay be falsely low in the first few hours aftercollection if blood samples are not sepa-rated.2 Large differences in changes inpotassium concentrations during storagecan occur, depending on time and tempera-ture.2 3 It is almost impossible to be sure thatany potassium concentration measured in ablood sample from a general practice is cor-rect unless the sample is separated withinone hour. Though it is easy to disregardabnormal results—some of which will actu-ally be correct—the main problem ispseudonormokalaemia in patients who trulyhave abnormally high or low potassiumconcentrations. This is more than “anannoying trait”1 —it is potentially dangerous.Unless samples remain at a constant 23˚Cthe only satisfactory way to ensure correctresults in samples from patients in generalpractice is to centrifuge samples in gel sepa-rator tubes.2 This has been implemented inWakefield, where all practices served by the

laboratory at Pinderfields General Hospitalhave been equipped with a benchtop centri-fuge costing about £500 and staff havereceived appropriate training. We believethat the initial capital outlay is justified bythe improved quality of results and, there-fore, patient care.P W Masters Senior registrarDepartment of Clinical Chemistry, NottinghamCity Hospital, Nottingham NG5 1PB

A A A Ismail Director of pathologyDepartment of Clinical Biochemistry, PinderfieldsGeneral Hospital, Wakefield WF1 4DG

1 Johnston JD, Hawthorne SW. How to minimise factitioushyperkalaemia in blood samples from general practice.BMJ 1997;314:1200-1. (19 April.)

2 Masters PW, Lawson N, Marenah CB, Maile LJ. Highambient temperature: a spurious cause of hypokalaemia.BMJ 1996;312:1652-3.

3 Moore D, Walker P, Ismail A. The alteration of serumpotassium level during sample transit. Practitioner1989;233:395-7.

Centrifuging samples may help preventfalse readings

Editor—After reading J D Johnston and S WHawthorne’s letter on problems with fac-titious hyperkalaemia in blood samplesreceived from general practices,1 our practiceaudited serum potassium concentrations insamples analysed during the past calendaryear. We found that in 74/82 (90%) ofsamples potassium concentration was below5.0 mmol/l (normal range 3.5-5.0 mmol/l);and in 8/82 (10%) of samples potassium con-centration was raised, seven of these (87%)falling into the 5.0-5.5 mmol/l range.

There is a difference between our figuresand the figures quoted in Johnston andHawthorne’s letter, although this may not besignificant because of the different basenumbers. It remains of interest, however,given that our mean transport time to thelaboratory was about 34 hours. The reasonfor the improvement over the Londonresults is that, like many rural doctors, wecentrifuge our own samples on site. Thisnormally happens within 45 minutes of tak-ing the blood. If laboratories have legitimateconcerns about false hyperkalaemia read-ings they should consider helping practicesinvest in centrifuges.N J Gourlay Principal general practitionerC McArthur Associate general practitionerMuasdale Surgery, Argyll PA29 6XD

1 Johnston J, Hawthorne S. How to minimise factitioushyperkalaemia in blood samples from general practice.BMJ 1997;314:1200-1. (19 April.)

HIV positive doctors deservesupportEditor—We were disturbed by the reportsthat Dr Patrick Ngosa had been struck off themedical register for refusing to take an HIVtest and continuing to work after learningthat his former lover was HIV positive.1-3 Wewould like to raise two points of concern.Firstly, was Dr Ngosa treated in a reasonableway? He was portrayed as an irresponsibledoctor who put his own interests before thesafety of his patients. Is that really the case, ordid he react as many of us would have done in

Letters


the same circumstances? None of the articlesexamined the support and counsellingavailable to Dr Ngosa or other HIV positivedoctors. Are such services available? Do theyconsider issues such as job security, retrainingopportunities, and confidentiality?

Secondly, did the General MedicalCouncil respond in a rational or emotionalmanner? What exactly is the risk posed topatients by an HIV positive surgeon? Howdoes the risk compare with other risks unre-lated to HIV infection? Is it safer to be oper-ated on by a competent HIV positivesurgeon or by an incompetent HIV negativeone? The response of the General MedicalCouncil seems to have been a vain attemptto appease the hostile, tabloid press. Patients’safety could have been ensured withoutremoving Dr Ngosa’s name from themedical register.

We agree with Richard Smith that weneed to move from a culture that encour-ages doctors to hide distress and difficultiesto one in which we learn to share them andask for help. We wonder if this ruling is astep in the right direction.Roger Drew Project directorFamily AIDS Caring Trust, PO Box 970, Mutare,Zimbabwe

Geoff Foster Consultant physicianMutare Provincial Hospital, PO Box 30, Mutare

1 Smith R. All doctors are problem doctors. BMJ1997;314:841-2. (22 March.)

2 Dyer C. Doctor who refused HIV test is struck off register.BMJ 1997;314:847. (22 March.)

3 Craft N. Trust me—I’m a doctor. BMJ 1997;314:910. (22March.)

Support and treatment ofserious comorbidity alsoimprove survival in breastcancerEditor—In their analysis of the prognosis ofbreast cancer diagnosed between 1982 and1989 Diane Stockton and colleagues founda significant improvement in overall sur-vival, which disappeared when adjustmentswere made for stage.1 Because of the minorstage-specific improvement they concludedthat the better overall survival over time wasmainly owing to earlier detection, with bettertreatment having only a small impact. Theystate that similar results, showing noimprovement in prognosis with increaseduse of treatment modalities, were found inthe south east Netherlands.2 This, however, isa misinterpretation of the results of ourstudy, reporting the survival of patients withbreast cancer diagnosed between 1970 and1984 in south east Netherlands.

In that study we compared the relativesurvival rates of women with breast cancerdiagnosed in 1970-4, 1975-9, and 1980-4. Inmultivariate analysis we adjusted for age(five groups) and stage (four groups).Relative survival greatly improved overall,and also after adjustment for age and stage.In non-metastatic disease, five year relativesurvival (adjusted for age and stage)improved by 30% (95% confidence interval

20% to 40%) for diagnoses made in 1975-9versus 1970-4 or in 1980-4 versus 1975-9.This improvement was present in all agegroups (adjusted for stage) and in stages I, II,and III (adjusted for age); it was notsignificantly different between the age orstage groups. Only in stage IV disease wasthere no change in prognosis.

In our region an improvement in stagespecific survival from breast cancer has beenobserved since 1955,3 although before 1970we could differentiate between only threestage groups. During 1955-74, diagnosisevery five years later (1960-4 v 1955-9,1965-9 v 1960-4, or 1970-4 v 1965-9)improved five year relative survival (adjustedfor age) in localised disease by 28% (14% to42%), in regional disease by 7% (−1% to15%), and in distant disease by 22% (6% to36%).

We agree that a large part of theimprovement in overall survival in bothregions was owing to earlier detection oftumours and that only a small effect oftamoxifen or chemotherapy was likely.4

However, better supportive care (includingprevention of complications such as throm-bosis and infections) and better treatment ofserious comorbidity (present in about half ofthe patients) may also have improved theprognosis. Furthermore, a change in themalignancy of the tumours over time—thatis, a relatively large increase in lessaggressive tumours—cannot be ruled out.5

H W Nab Clinical epidemiologistJ W W Coebergh Clinical epidemiologistComprehensive Cancer Centre South (IKZ),Eindhoven, Netherlands

1 Stockton D, Davies T, Day N, McCann J. Retrospectivestudy of reasons for improved survival in patients withbreast cancer in East Anglia: earlier diagnosis or bettertreatment? BMJ 1997;314:472-5. (15 February.)

2 Nab HW, Hop WCJ, Crommelin MA, Kluck HM, CoeberghJWW. Improved prognosis in breast cancer since 1970 insouth-east Netherlands. Br J Cancer 1994;70:285-8.

3 Nab HW, Hop WCJ, Crommelin MA, Kluck HM, van derHeijden LH, Coebergh JWW. Improved long-termprognosis in breast cancer: survival rates since 1955 in aDutch cancer registry. BMJ 1994;309:83-6.

4 Nab HW, Voogd AC, Crommelin MA, Kluck HM, van derHeijden LH, Coebergh JWW. Breast cancer in south-eastNetherlands, 1960-1989: trends in incidence andmortality. Eur J Cancer 1993;29A:1557-60.

5 Joensuu H, Toikkanen S. Comparison of breast carcino-mas diagnosed in the 1980s with those diagnosed in the1940s to 1960s. BMJ 1991;303:155-8.

How widespread is prejudiceagainst sick doctors by selfhelp groups?Editor—The self help group movement isflourishing: there are thousands of groupsactive in the United Kingdom and theirnumbers are increasing. The suspicions ofdoctors towards self help groups have beendiscussed,1-4 and they are understandable,though perhaps misinformed. However, Ican find no published work on the attitudesof self help groups towards sick doctors whowish to join such groups.

Two years ago I was diagnosed as havinga major psychiatric disorder; I was ill for atime but made a good recovery. I had neverbeen seriously ill before, let alone been aninpatient in a psychiatric hospital. I became

motivated to share experiences with otherpeople with the same problem rather thanjust reading about them in textbooks. I joineda well known charity dedicated to people withmy condition and decided to set up a self helpgroup in my area. I contacted the head officeand was told that the charity did not allowdoctors to run self help groups, especiallythose practising psychiatry. I protested at thisdiscrimination and was eventually given itsblessing. However, the organisation’s “how toset up a self help group” package mysteri-ously vanished in the post a number of times.After I had had numerous, increasinglyfrustrating telephone calls with theorganisation it finally arrived two monthslater. The group is now flourishing, and all ofthe members think that I am an asset becauseof my psychiatric knowledge, particularlywhen questions about technical issues such asdrug treatment arise.

A colleague who was due to give birthhad a similar experience with a nationalcharity for pregnant women. She waspractising obstetrics at the time and was toldbluntly by a self help group leader that shewould not be welcome to join the group. Shewas left in floods of tears. I wonder howmany other doctors have experiencedsimilar problems. I am interested to knowhow widespread such practices are. If youhave had a similar problem please write tome giving details of the illness, the charity,and how you felt about their attitudestowards you. I will deal confidentially withinformation about your illness, although youcan of course write anonymously.N J Stafford Senior house officer in psychiatryShrodells Unit, Watford General Hospital, Watford,Hertfordshire WD1 8HB

1 Black ME. Self help groups and professionals—what is therelationship? BMJ 1988;296:1485-6.

2 Ratcliffe MA. Self help groups and professionals. BMJ1988;296:1797.

3 Atkinson JM. Self help groups and professionals. BMJ1988;296:1797.

4 Logan RA. Self help groups and professionals. BMJ1988;296:1797.

Man’s fractured sternum wasprobably due to snake’s weightwhen it fellEditor—B F Moran and W J Newmanbriefly reported on a man with a sternalfracture after a snake fell from the treeunderneath which he was sleeping.1 I wassurprised that they attributed the fracture tothe snakebite. Surely it was more likely to bedue to the direct blow that the man receivedwhen the snake landed on him. The snakewas described as large and so possiblyweighed several kilograms. Thus, dependingon the height of its fall, it could haveimparted enough kinetic energy to breakthe man’s sternum. A bite alone is unlikely tohave the necessary energy to break a bone.Brian W Davies Specialist registrar in paediatricsurgerySt James’s University Hospital, Leeds LS9 7TF

1 Moran NF, Newman WJ. Minerva. BMJ 1997;314:1-58.(5 April.)

Letters


alcohol consumption may influence onset of the menopause

Documents