Tobacco Documents Online

on all examined subjects). Table II preterits a mo~t ret-med aml.vm. • m~ at ~ to ~ ~ to ~t~te hit ~fi~ ~ f~ th~ f~or (a mm~ si~ ~Ung a fall). ~ t~m~ of t~ ~vid~ c~n~ ~ ~ c~rc~ by sub~t~ f~ it ~e val~ for t~ control ~p to allow ~ for ~y drift L~t mi~t have ~u~cd indc~ndcmt of our inte~cntion and for rcgress~n towardg the n~n. ~i~lly thi~ corr~ed e,ti~te of ~ ~ect of inte~enti~ ~s exp~ as a ~rc~m~ of ~c initial ~an ~he for t~ intc~'~tion group at a whole. S~t~tical ~i~nce call,ted by a r te~t on the diffcr~e b~n the ~ for intervention and c~trol ~amp~. ~e ri~ factor le~h at ~ry gene~l~ ~ ~m ~tc~ bet~cn the intervcntion and control g~upL The ~st dis~ri~ 10 S 140 130 5-9 E 5.5 80 o ...... o Cmtrol group ~ Intervention group Plasma cholesterol b~ 2~ . . . 0 I 2 3 4 F Yec~r ~IG 2--T~stimated mean risk levels for men in random samples. w~ for body w~tght m the-high-risk racn (~ l). ~ich c~mbm~ ]l~ to o~I1 ~. ~'~O ~ ~ m~t~le ~tm ~+ ~i~ to mdivid~ls" val~s f~ c~ ~ed ~ ~y, ~t~ b~ p~ssu~, ph~ c~l~t~l c~t~, ~d ~y ~s ~d~. T~ wcigh~ ~c~ts) appli~ to ~ch factor were ~o~ derived fr~ t~ ~per~cc of t~ Em~an co,tries in the Scv~ ~mrics" Study (G Farchi and A N~enotti, unpublish~ ~ta, 1978). At entry ~se ove~II sco~s ~ alm~ i~t~d in hi~-risk men bnt averaged 5". hi~er in inte~¢nt~ ~n 1SIGH-RISK MEN High-risk men showed the largest changes in risk factor levels. They reported a decline in mean daily cigarette consumption of 29% at the end of the first year, and this effect was sustained. Thc controls changcd little, so that the corrected c~tlnmte for the effect of intervention at the final examination was again -29".. (table II); by then 12% of those who were smokcrs at entry had stopped completely. Blood pressure showed a large regression to thc mean in both groups, but throughout the trial the systolic pressure averaged about 3 mm Hg lower in the intervention group. Systolic prcssurcs of 160 mm Hg or more were recorded in 38", of the high-risk men in the intervention group initially and in 16% at final examination. An almost identical fall occurred in the control group. Advice on cholesteroI-towering diets was more concentrated in the early months of the trial. After one year the plasma cholesterol concentration in men in the intervention group fell by an average of 8~.~, and 75~. claimed to have changed their eating habits. Judging by the two-year results in the control group, about half this fall represented the effects of intervention. Little dietary advice was given in the second year, and the gain was lost entirely. After renewed efforts in the next two years the estimated effect of intervention was a fall of 6.9°,t~ at four years (table II). During the final stages of the trial the nurses were fully occupied with examinations, and once again most of the ground was lost. Individual advice on weight loss was given to those high-risk men who were I5% or more overweight. By the two-year examination the high-rlsk group as a whole had lost t.4 kg more than their controls. The whole of this advantage was later lost. The estimates of overall changes in risk suggested that the benefit at the end of year ] of about 20% had all been lost by a year later. But during the final three and a half years the net effect of intervention in these high-risk men averaged - 12%, or - 11% over the whole trial. REMAIND~ OF STUDY POPULATION Results for r~ndom samples of both intervention and control grow~ are give~x i~ fig 2 and table II. Reported cigarette consumption was substantially ~du~l by int*rvention throughout the trial: at the final examinations the corrected estimate of change was - 19% for the whole intervention g~ap, or -16% when hig.b,-risk men w,re ex- cluded. About 9~ ~ of smoker~ lad by then been persuaded to stop (7% excluding high-risk men). Special measures to reduce blood pre~ure TN]LE tl--Pereemage me,at ¢~ar~ in riA faaor .Car i~iduab in interoentian group, corr*c2ed for ~r~ol~adi~*~ clmr~es in ce.,~trol F~u~ High-ri=k inert Random sample T109782380

Ietter~, bookkts, and po~ers. Conversations whh ~ ~ su~d that t~e ~ a ~c ~, ~ at ~h of ~c ~ ~ " " a~ ~¢'~, of ~ose ~' ~amined (e~l~mg high-ri~ men) r~ ~at they ~d c~ged their ~ting ~bi~. O~ecfive ~ of c~nge ~, ~r, n~ligib]e. In t~ m~d~ 3~a~ of t~ tr~l ~e ~ ~s ~ to advise thc~ men individual~, the c~t~ c$t~te at four y~r$ x~$ of a fall averages 4% (ruble ]I). ~i$ ~t di~ in t~ final st~ ~ ~e trY, w~ the int~ntion effo~ ~d to be relaxed. No eff~ts on b~y ~i~t ~ ~t at any ~, ~ o~i~t ~ ~d ~i~ postal advice in the firs~ year. Overall there wcrc no clear differences bctw~n ~e intervention and control ~u~ in total ri~ estimates. Over ~e whoie per~d of the trial, the net eff~t of ~nte~ention on ove~ll risk averaged u~er 4~. Resulm o~ physical activhy have not been presented, there *~s no adequate m~surc of ~sponsc, though some indivi~a]s ~e~ised energetically and persistemly. Annual quemionnaires su~ested a persis~ modest advantage to ~e ime~enfion group. Thus at gout years, for example, vigorous exercise was repor~ by 35% of the intervention g~up compared with 22% of co~trols. This estimate of response could ~ll be exaggerated. EFFECT OF SCREENING IN CONTROL MEN" We did not know what effects screening might have on behaviour or use of medical care in the control group, and the assessment of changes in incidence will therefore be based on the 90°., of control men who had no contact with the trial until the final examination. A comparison of risk-factor levels at the final examination berc,'ecn this 90% and the remaining 10% of men who had been examined before showed almost identical results for smoking~ cholesterol, and weight, but a significant difference in blood pressure. The average was 4.1!2.0 mm Hg lower in the men who had been screened before. Since few had been referred for treatment, this presumably resulted from habituation to medical examinations. Discussion The control of coronary heart disease necessarily depends on prevention, since treatment so often comes too late. Mass medication is potentially dangerous~s and it would be better if risk factors could be controlled by changing habits. To be effective in middle age this would require that the risk factors did cause the disease, that the progress of the disease could be altered, ~nd thin those advised c~uld c1~age tl~ir habk~. This report deals with the last of these issues. The answers to the first .~nd second depend on showing dumges i~ the k~cideace of disease and will form the subject of a subsequent report from the WHO European Co|labo~at~vc Trial. Screening ~nd health education require t~e same rigorou~ evaluation as a new treaunent, but this has to be organised at community level, which makes it difficult to obtain a suitable control group. The North Karelia project in Finland" and the Th~e Communities study in California' each u~ed a sLugle community as control, but this gives no estimate of bility of results, and it is impo~ible to assess statistical significance* We therefore inck~cd many o~mmunitie~ in the study and then randomised these, rather u~n individuals, to intervention or control status. This produced two group~, and the conu'ols remained free from contamination by the intervention programme proceeding elsewhere. T~ne meab-~ selected for evak~ion were b~ed on the choice today. /v~re debatable was the level of remurcc~ appropriate to implementing Lhe programme. It could be argued that a marc hazard to public health m~it~ a ma~or inve~.mcnt in its ~xol, bu~: we had to make do lm'gely w~Ja what wa~ available. A bigger eff~'t might have achieved bigger r~uhs. Periodic examinatio~ of .~all random ~.'-aples, cxmpled ~ Mk"D[CAL JOL'I~*gAL 15 xu, dto~ 19~t with good respoose rates, pr~-/ded an excellent m-ml memas of ~itor~g p~s. ~itc ~pl~g c~ oval ~e~ ~ clear. ~ results s~w the n~ity ~ a c~t~ group to i~ntify ~d allow for rcgr~sion to the mean (b~ pressure and ch~esteml), habi~uati~ ,o ~minauon procures (bl~d ~), ~d ~x~ ch~s out thc t~ thc wholc stay ~pulation x~s cx~'d to con- flicting prcssurcs f~m advcrti~ ~d health cdu~tion). It p~cd rclativcly casy to disscminatc information ~d men's rcs~scs to q~fi~ing~f~ ~mplc, by thc end of first year 75' ,, of high-risk men and 30".. of the remainder claimed ro ~ve ch~d thor caring habits in r~ to advlce~but their claims were nor paralleled by ~rcspunding ~anges in plasma ~olesterol concentrations and ~dy weight. This was particularly evident in the men who were m~ at high- risk and who, during the e~ly stages of the trial, received letters and ~n~al p~p~anda. ~opaganda seemed ~cctive a me~s of spreading informatk~, but a substantial c~ngc habits seemed to require some personal The screening ex~inations proved to be a two-edged weapon. In gener~ ~ey stimulated interest in the aims of the project, and enhanced the receptiveness to advice of the high-risk men re, led for personal consultation. The remaining men, however, tended to regard a satisfactory report as tantamount to some sort of ~arantee of health, even though their letter warned that heart attacks could still happen. Only after four years, when many had re~ntly been recalled for personal dietary advice, was there any evidence of an effect on dietary habits in the inte~ention group as a whole. When the interv~mion effort was later relaxed this benefit was soon lost. The only strong and maintained effect of mass advice was on the number of cigarettes reported as smoked, which was persistently reduced in the intervention group. If this reduction was real this could be an important benefit, but the claims are unvalidated. Other centres in the WHO ~llaborative Trial are examining the use of plasma thiocyanate levels as a means of validation. The high-risk men responded better to advice. ~heir reported fall in cigarette coasumption was larger, and the percentage of smokers appeared to be reduced. Average blood pressure levels were lowered by a few millimetres. The measures of the effc~ of dietary c~ge on pl~ma cholesterol levels reflected the amour of recent personal advice~ being greatest in the first year (about 3% allowing for estimated regression to the mean) and the fourth year (6-9%). Even these modest responses were q~ckly lost when re~lar personal ~ntact c~ld not be main- tained. The same was true of weight loss. Perhaps dietary advice needs to be ~ded in ~e sa~ way as l~-term d~g ~eatm~t, ~i~ sustained ~ns~tations and en~u~gement over ~e T~e ~ial has sh~a that ~na~ risk factors ~n be ch~d ia the w~kiag popularly, Wen p~pcr org~i~i~ ~d some sapplemen~i~ ~ ~isting res~s. But cha~¢s were not large, ~d we ~ave not shown ~at they can be su~ined. The pa~icipativ~ oc~tlon~ physicim~s concluded that on p~sent evidence ~ would not ~mmend the gene~ introduction of ~s ~ of sc~ni~ ~d ~eal~ education s~vi~. At the same time ~ reprised ~eidenml ben~ts h ~e form of cl~er ~n~ wi~ em~oyees, ~d ~'~1 individual were d~ted ~ m~ n~ng ~tm~. If a r~acti~ ~ risk fa~o~ ~ to be a~cd by a ~s~ate f~l ~ risk then the ove~l ~ x~k of the hi~-fi~ men was r~u~d by an estimated a~ of 11% ~d of the inte~emi~ ~up as a whole by ~ ave~e of 4%. ~ ~ ~ b~ ~t a r~ ofcv~ IIOL ~ ~t ~ ~ ~th w~ ~ cqui~[e~t f~ m~ny other w~t • rare ~e~ive ~ ~u~mi~d cffor~ m~t achieve. We thank the medical staff, rnanagcrm.'nts, track- tmions, and em- ployees, all of whom co-operated Wlcndidly; Prof~so~ V H T Jam~ TI09782381

~d Mrs P Hopgood for =holc~crol ettirtmtitms; oor c~ ~ A G~, ~ M~ K O~. ~ ~u~ ~ ~ ~ ~ ~lpful thro~t. ~N~-K ~ re~ m~ ~ prov~cd by Glaxo Ltd. The fo]~wing com~nie~ a~ ~pati~al ph~ician~ ~ipat~ : British Airwa}~ (Dr~ D M B~t~, I M ~n, C C O ~fl); British Steel ~rmion (Drs J A E Ricb~ds, G B D~m, A Si~ir, C F R~, the ~t¢ C R T~mas, J B ~¢~t~); ~db~- ~hw~s Ltd (Dr C ~¢); ~I M~ Indu~h~! Hcahh S~'~e (the late Dr G E Ff~h); G~t Ke~ Nettlefold Ltd (Drs P L ~I~ar: ] A R~', L E Tyler); Guin~ Ltd (Dr B M ~am~); Ilford L~d (Drs V O S~ew~, D ~ul]) ; Ko~k Ltd (D~ M Falconer, G Hughes, K W Harbord, T Kelly); May and Baker Ltd (Dr J Cmhbcrt); Phili~ Ind~tricz (Dr D J Terry); Royal Ordnance Factor" (~he late Dr Tyrer, Dr C Edward~); Shell Chemical~ UK L~ (Dr C M S ~pp~); Tare and Lyle Refinerie~ Ltd (Dr G L ~c~d). 751 Referem=es = l J~nt Worl~g Party o~ the Ro~I Co'ego ~ Ph~mc~ ~ ~ ~ ~ R Call P~uoa~ ~ 1~76;10~!3-75. ~ ~ G. A ~d t~ of h~ di~ ~cnt~n in md~t~'. Try, ~ Ot~p Aicd 1970~:1~-1 !. in t~ multff~t~ml ~tion o[ coma~ ~rt dts~. Intff &pidtmwl 1074~:219-24. t Waik~ SH, Dun~n DB. Estimati~ of the probability of an event as a fun~ ~ s~c~ in,pendent vafia~s. Biomnrtka 1967#4:167-70. t ~tt~ ~ Pri~i~l Instate. A c~rati~ tml in the ~m~" p~*cnt~n of i~h~mic heart di~asc using clofib~tc. Br Heart ~ 1978;40:10~-118. t ~s~ P. N~h Ka~lia ~t, a pr~ for ~uni~ c~tml of ~rdiovascular diseases. Publications of the Univcrsi~ of Kuopio. ~mmunity Health 1974; Series A:I. 7 F~qu~r JW, ~c~by N, W~ PD, a al. Community ~cation for ~rdiovas~lar health. Lancet 19~7 ;i : 1192-5. Vitamin D supplements in pregnant Asian women: effects on calcium status and fetal growth GBROOKE, IRFBROWN, CDMBONE, NDCARTER, HJWCLEEVE, MAXWELL, V P ROBINSON, S M WINDER Summary and conclusions In a double-blind trial of vitamin D supplements in pregnant Asian worsen calciferol (ergocalciferol, 1000 IU/day) was administered to 59 women and placebo to 67 controls during the last trimester. The two groups had similar distributions of maternal age, height, parity, number of vegetarians, countries of origin, and sex and gestation of the infants. At entry to the trial maternal serum 25-hydroxy vitamin D (25-OHD) concentrations were low in both treatment and control groups and significantly lower in vegetarians than non-vegetarians. Mothers in the treat- ment group gained weight faster in the last trimester than those in the control group~ and at term they and their infants all had adequate plasma 25-OHD con- centrations..~others a~ infants in the control group, however, had low plasma concentrations of 2fi-OH~D St George's Hosp|tal. London SWI? ORE Department of Child Hem|th 0 G BROOKE, ~n, ~ar.l% senior l~turer N D CA~TER, ~nD, ~ni~ lectu~r Dew.meat ~ ~e~ Pathology 1 R F BROWN, P~ ~a~Ar~, l~tu~r H ~ W CLEEVE, ~p~, senior l~t~r De~me~ of O~tet~ a~ Gyaa~gy V P RO~NSO~, u~ ~n~r ~tmr and calcium and raised plasma alkaline phosphatase (bone isoenzyme) activity. Five of these infants developed symptomatic hypocalcaemia. Almost twice as many infants in the control group were small for gestatlonal age (29% v 15%), but there were no significant differences between the two groups of infants in anthropometric measurements. Infants in the control group, however, had larger fontanelles, suggesting impaired ossification of the skull. Because of the benefits to mothers and infants in the treatment group and the absence of side effects, vitamin D supplements should be given to all pregnant Asian women in the United Kingdom. Introduction Despite improved living standards there is continuing clinical and biochemical evidence of vitamin D deficiency in As/an immigrants, both among children and adults? This problem is not confmed to Scotland and the north of England, where winter sunlight hours are short, and Turton er al recently showed that vitamin D deficiency occurred in pregnant Asian women in the south Londo~ di~ict of Tooting, which contains a relatively affluent Asian community.: Osteomaiacia is a wcll-recogniscd complication of pregnancy in Asians living in the United Kingdom." Asian women appear to be particularly at risk of vitamin D deficiency during pregnancy, since low co~ccntraticaas of 25-h.xxtVOgT vitamin D (25-OHD) are fuund at this time.: *-' Possible nemaatal con- sequences of this deficiency include hypocalcacmia,: craniot abes ,' and frank z4~ckz,~s.' Since the ~ to the fetus of subcl/nieal maternal vitamin D deficiency axe not ck~arly defined and since birth size of Indian Asians in Britain is le~s than that of north Europeans and Negroes," ~° we have undertaken a trial, using cakifel~l (crgocakZderol) supplements, to investigate the effects of the vitamin tm maternal and infant oJ~_-ittm honx~- costasis and fetal growth. T109782382

14 were seen with the leanmorgeat rel IUD with eke greater daily rdease ra~e, w~ ~nar~ occu~ ~ a ~ with an iml~ix~d release r~e well bdow the amount normally measured for this model,js The daily amotmt ofprogcstagen released can be reduced by using a potent c~,apound like levonorgcstrcl, and thus the lifelime of r~ach a medicated device can be prolonged by • several years. This could make the need for medical interven- tion more infrequent and the increase in the number ofdays of intermenstmal spotting usually occurring immediately after insertion ofa r/ew IUD can therefore be avoided. The levonorgestrd IUDs of the present study have been calculated m have a life,an of~t least seven years.)+ No case of pelvic infection was noted daring the present study. A predisposing factor for ascending i~fecrions is an in- creased number of d~ys of bleeding, and the fewer days of bleeding associated with the levonorgestrel IUDs thus pro- tect against pelvic inflammatory disease. Although the differences were not statistically significant, the rates of pregnancy ~nd expulsion were numerically lower with the levonorgestrel IUDs than with the control IUD. Amenorrhoea as a reason for removal of the device was found only with the levonorgestrel tUDg experimental studies;3'1( suggest that it was not associated with a deterioration in the hormonat function of the subjects. Levonorgestrel-rdeasing IUDs significantly reduce menstrual blood loss: the total number of days of bleeding, except during the first 2-3 months, is less than with the nova-T device and continues to fail with time. There seems to be no problem of infection with levonorgestrel IUD% and the pregnancy rates are very low and do not seem to be affected by the skill of the inserter. initial spotting was not the reason for many removals and was neither a health problem nor an inconvenience. In view of. these findings the use of the potent progestagen, levonorgestrel, for intrauterine release seems to be a most promising new approach to achieving a long-lasting intrauterine contraceptive device equally suitable for both developing and developed countries. This work was undertaken as part of the contraceptive development programme sl~raored and coordinated by the International Committee for Contr~ceptlon Re.arch of the Pop~laHon Cooncil, Inc., l~¢wYork. We thank the staffofthe ¢linles for their eoopera lion and the Intemationa! Devdopm ent Resea~ca C~nrte of C.annda, the U.S. Agency for Internalion:l Development (gram A|Dtpha,.~ l ] 16)~ the Fe~d Fo~andalJoa~ the Rockefeller Foundation, and the (3ooJ. Hecht Fund for fmanchlsupl~ort. Thceontcnt ofth~s work does not necessarily reflect the policy ofany of the Funding murves. gc~uots lot rcpdms should h¢ ~dr~.*ed tO C. G. H. M. J. SHIPLEY BRIOHY J. TBO~S Iummar3, Inn longitudinal study of civil servant% 1422 men were classified according to their average daily alcohol lotto:Over 10 years of follow-up, the mortality rate was lower in men reporting moderate alcohol intake tha~n in either non-drinkers or heavier drinkers (>34 g alcohol per day). Cardiovascular mort alley was greater in non- drinkers and non-cardiovascular mortality was greater in the heavier drinkers. The heavier drinkers had higher mean blood pressures and contained- a greater proportion of smokers.~ multivariate analysis showed this U-shaped relationship between reported alcohol consumption and subsequent mortality to be largely independent of differences in ~blood pressure, plasma cholesterol, and grade of cmploym~ Introduction THE question of a safe limit o falcohot consumption has not been resolved. Alcoholics have a higher mortality rate than the average,~ but this tells us little of the risks of more moderate drinking. There have been few studies on non- institutionalised popuJations of the relation between alcohol consumption and mortality. There are reports that moderate drinkers have a lower mortality rate than either heavy drinkers or abstainers~'~ and that moderate drinking may DR HLI._SSON AND OTHER'S: REF'EXENCE~--conI~nu~d 9. Nilnon CG. A d-notges.d-rdeaging IUD. M.D. t hesit, Univer~iv/ofHdsi~, 1977. 10. Nilsson CG, Luuk~in~ T, Arko H. Endomeu~l mowholo~ in women using ! I. ~il~ ~. ~m~ndve quancks~i~ of m~ns~al bl~ ~s with a ~n~gcstreb ~d~fing IUD and * ~ev~T~p~r d~. ~nua~r ~a 19771 ! ~: 12. ~ ~ Lu.~ T. lm~o~m~t of t 6norg~t~el.rcl~ng IUD. ~ntra~t~a I977; t3: TI09782383

We have invesdg~ed the rdation between estimated akehoi comumtxkm and mort-Airy over the subsequent ]0 yearn of follow-up in a longltud|n#l study ofa sarn~ td'mcn ha active employment, the Whltchall Srody ofclvil ~crva~ts? The analysis takes into account smokJatg and o~er risk Methods In 1967-69) a screening examination was carried out on 18 403 male civil ae-rvant~ al~ed 40-64, wocking in several departments in and around London" The ~xamlnati~ included a questionnaire, determination of Mood pre~ure, and mcgstlrcmc~t Of chol~ co~emWa~i~a. A I -i~10 sample of~he f'~rst I~ 000 men screened Were asked to return by mail a completed dietary record that listed everything eaten or d.r~nk ~ a 3-day period (Sm~hy, Monday, Tuesday). The response rate to this request was 89% and 1422 completed dietary records were received. The food items consumed were coded according to type and amount. Nutrlcnt intake was calculat~ by means ofcomputcriscd analysis based ou food tables made available by Misa Jean Marr and UlXlated by us. The mc~ure of alcohol Latake comes from this 3~day record of dietary intake. For over 99% of the men, ~ccords in the N.H.S. Cer~tfa| Registry were idemL6¢d and flagged. A copy ofthc death certificate ha~ beta provided for each man who died, in the LLK., during the 10 years from examination. Dcatlx certificates wcrc coded by the O~cc of Population Ccususcs and Surveys. Age-standardlsation was by the direct method using the total population zs standard. Rdativc risks (strictly, relative betting odds) and t~ts of significance w-re calculated by fitting multiple iogisdc regressions to the data using the statistical package G.L.I.M.~° This allowed the significancc and size of the alcohol-mortality relationship to be assessed after taking into account the ctTects of age, cigarette smoking, blood pressure, plasma choIcsv'rol c-oncen- trafion~ and grade of employment. It also aliowed us to test for a linear or quadratic U-shaped relationship ofalcuhoi with morxality. Results The nutrient analysis of the dietary record produced figures for alcoholconsumption in g/day~ here cazcgoris~l as 0~ >0-9~ ~9-34~ >34 g/day. The General Household Survey~ referred to "onc drink" (8-10 g alcohol) as being halfa p'mt of beer (2~t mt), one mcasureo~'spirits(23-25 ml), on~ 2 oz gl~ss of port or sherry (56 ml), or one 4 oz glass of wine (1|2 ml). Our highest alcohol category therefore is roughly equivalent.to four or more &inks consumccl per day. Table ~ shows the per ~nt of men who died in 10 y~rs of f~Row-u!b according to age and ~h¢ amount of alcohol tb.cy reported having consumed at entry into the study. Among men aged 50-59 and 60-64, the albcau~e mortality ha those who consumed no alcohoI on the days ofdictary recording is ~er than in those whose reported alcohol intake corr~- spond~ m 0" 1-34 g/day. The "heavy'* ~rir~s C~.34 g/day) haw a Mghe.r mortally than Llte mod¢~ dfird~ers. ~he overaR rdativc risk~ adiusdng for age, is shown arbitrarily ~ the men ha ~e 0-1-9 g/day ca ~t~ory m have a r~lativc risk of 1-0. To ~est f~r sig:aificance of the U-shaped relatioashi~ a qna&-afic carve was fi~ed m the data. Th~s was significant (p=0-021), The axal variation ia ~clativc risks, as estimamd by X~ was mahaly ac~oumed for by Lhc wdafion du~ m the This U-shaped relat~ of alcohol m total atonality 4O-49 5O-59 6O-64 0 0-1-9 2.9 5.2 12-~ 4.7 po~ 072) 20-0 15-9 1 "$ 9-1-34 >~4 4.9 5-0 (!~4) (~) 7-3 !1-4 (177) $-7 21-7 1"! I-~ *Fitting a quadratic (U-sha~) curve |o ~bcs¢ tel=five fi=~ ~e~ Z; =5.29 U-shaped relationship is largely the result of a higher cardio- vascular mortality in the "non-drinkers" and a higher non- cardiovascular mortality in the "heavy" drinkers. Of the 63 cardiovascular deaths, 49 were due to coronary heart disease. Restricting the analysis to those who died from coronary heart disease did not change the pattern of a higher mortality among the "non-drinkers" and little difference in mortality between the other drinking groups. The mortality from non-cardiovascular causes is shown for cancer and other causes in table II. For both these categories there is the suggestion of a U-shaped relationship (not sig- n~cant) with the highest mortality in the men reporting the heaviest drinking. The 7 deaths due to "other" causes in the heaviest drinking category were from pneumonia (2), chronic bronchitis (1), disease of the larynx (1), hiatus hernia (1), and suicide (2). Only 1 of the total of 21 "other" deaths was due to cirrhosis and 2 to motor vehicle accidents. The high mortality among "abstainers" could be the result of people who were already sick at the time of examination, having given up alcohol and subsequently having a higher than average mortality rate. If this were so the higher mor- tality in the abstainers would be most marked in the first years of follow-up and become less marked the longer the follow-up period. HowcveG when we examined the mortality rates after excluding deaths in the first 2 y~ars of the I0 years follow-up, ~e U-shaped relation between alcohol intake and mortality was scea a~ strongly as shown above. The possibi~ty must be coo-sidereal that the observed rela- tionship b~tweca alcohol and mortality may be the result of 10' 8. %6" Z- ALcohol, (giddy} T100782384

582 Othe~ TABLE |H-- BLOOD PRE.SSURI~, SMOK LI¢O, AND PIM~MA CHOLI~'~EROL LEVEI~ I~ ALCOHOL C~NS~MPTION: ~E.AD~USTI~D ME.~aM$ (AND STAI~ RD ERROP~J *rnrno~1= m~dl+ 38-7. ° 0 136"9 (0-96) 85.0 (o-e5) 42- ! 16-6 (0-56) 5-19 (o-o63) ]37-8 134-2 139-7 0"~) (l.It) (I-~) 85-3 ~,~ 87.3 (o-~ (o.~) (o-9~) 36-5 37.5 52.8 (0-62) (0-75) (0-98) 5-27 5-22 5-26 (o-~e) (o.~4) (o-oe4) other factors. Table III shows mean systolic and diastolic blood pressures to be highest in the heaviest drinkers. Both the per cent of men who were smokers and the mean number of cigarettes consumed per day per smoker show a U-shaped relationship to alcohol consumption, higher in the non- drinkers and higher in the heavy drinkers. There was no asso- ciation between alcohol consumption and plasma cholesterol concentration. L~able IX.' shows the all-cause mortality (~ge-standardlsed) according to alcohol consumption and smoking. The moderate drinkers (0-1-34 g/day) have a lower mortality rate than the non-drinkers, regardless of smoking status. Only among the ex-smokers do the moderate drinkers not also have a lower mortality tha~ the heavy drinker~ This analysis was done separately for eardiovasoalar and non-cardiovascular mortality~,.~..mong smokers the previous pattern was seen--higher ¢ardiovasoalar mortality in non- drinkers, higher non-cardiovascular mortality in heavy drinkers. It w~ less clearly seen in the non-smokers, bat there were Few dea~ To take into account not }use smoking but also the number of'clgarettes smoked, a multivariate analysis was done which 0"1-9 9-1-34 >M 3"6 3-4 7-I ~-4 6-2 4.2 7-) 7-9 13-8 ~=0-~5) and ~ overa~ ~ ~tween ~lcohol c~tegofi~ ~c~ 75-{.e., the qua~adc rehfiomMp =~nts f~ m~t ~e diffe~c~ ~lween ~legoriet. adjusted for age, smoking habit (non-smoker, ex-smoker, and 1-9, 10-19, or >20 cigarettes/day), systolic blood pressure, plasma cholesterol, and grade of.employment. The results are summarised in table V as relative risks, the group reporting 0.1-9 g alcohol/day being assigned a relative risk of 1- 0. The magnitude oftbe relative risks shown in table I are reduced slightly (p = 0- 065) but the overall pattern is the same. Thus, to a major extent, the association between alcohol and mortality is independent of" the factors included in this analysis. Discussion In this study, alcohol intake was assessed by analysing the report of 3 days' dietary intake. This dietary method has been studied for other nutrients and has been shown to result in substantial errors in classifying indlvlduals,t2 It is unlikely that random errors could account for the lCmttem ofmortality observed. A serious bias would arise h'heavy drinkers failed to report their use of alcohol and were wrongly classified as non.drinkers. This could east doubt on the high mortality of the non-drlnkers, although not of the heavy drinkers. A spuriously high mortality among non-drinkers could al~o occur if men who were already sick and at high risk of dying gave up the use ofalcohol. When we confined our analysis to the last 8 years of Follow-up the association did not weaken. We have no information on past drinking histories. However, Room and Day examined this poss~ility in a study ofsamples 6ftbe U.S. popu.lation.~ They similarly found a higher mor- tality in abstainers than in moderate drinkers. Removing from the analysis people who reported their health to be poor did not change this. Amoa~g tt~e *'healthy", moderate drini~ra e~periet~eed a lo~.r mortality than a~stalners.) Other studies rapport the f'mding of an apparem|y pro- tectlve effect of moderate alcohol consumption in cardio- vasoalar disease?n Dyer et 81., reviewing the relation of alcohol to cardiovascolar disease, point out that some studiea show a harmful effect of atcohoL4 In general the poskive smdi¢~ have shown the at-risk category to consist o~problem drinker~ of'wh0m there were likely to be t"ew in our group with the highest alcohol im~ke of >34 g/day. A protective effect of alcohol o~ ¢ardio,caseular di$c~-'e could be medimed by the eft .e.~t ofa|ooh~t ~ fairing HDL cholesterolI~ ]-~te~, Keys has sugge~ttml that ~ HDL Ievek predi*ixxse to higher ratea off" non-cardiova~oahr mort~ty.~ Ifconfirmed th~ ~ld be a mediator of the adverse effect~ ¢ff~ghet ~ of alcohol emmampdoaL It lm~ been suggested that the protecdv~ effect ofalenhol i, confirmd ~o wine.~ Our ara~,~is did not di~tlngu~ the form of~ drunk, k k po~s~l~ ~t it is not the ethanol itseff T109782385

sumcd more alcohol than men in the lower grades; inclmion of grade of empkffmem ha the multivariate analysis made lit-de d'~K'reu~ te the resuks. Hence, k is untigdy that the higher mortality among heavy drinkers ob, er~ed in this pop,orion was the reault of general social ch~ differences. Similarly, the association between smoking and drinking appeared to account for llttle of the higher mortality. The most likdy conclusion is that men who consume more than 34 g of alo~hol (4 or more 6riffles) pox day incur an Lnc~eascd mortality risk. The size of the study does not permit us to estimate mortality for subgroups of this >34 g alcohol categmy. There is historical precedent for this limit. Anstie in 1864 proposed that an amount of drink equivalent to about 35 g of alcohoI W..r day was the safe limit,ts The public health message is not a slmlSte one. Our data on alcohol suggest that moderation is associated with longevity, but that small increases above that are associated with a shortening of life. Before recommending that everyone take one or two drinks per day, we must bear in mind that a higher mean level of alcohol intake in a population is usually associated with a higher proportionofpeople suffering from alcohol-associated problems--not only mortality but also social and biological disability,t6 We thank the civil servants who gave their time to participate in this study; the Civll Service Medical Advisory DeFaixment (in particular, the late Sir Daniel Thompson, DrA. Blair Harlngton, and Dr Adrian Semmence) for their hdp; and Miss Lind:~ Colwell who processed the death certiiicales. M. G. M. is supported in part by a grant from the British Heart Fou ndadon, arid the ~tudy was supported by a g~am from the Tobacco Research Council. Requests for reprints should be addressed to M. G. M., Department of Medical $lali*tles and Epidemiology, London School of Hygiene and Trolfical Medicine, Keppel Streth London WCIE 7HT. I. Surglby P. Ak'otm|ixm am~ moiety. ~ Univ~tsltc~sforlagct, 1967. 2. Pearl ~ ~koh~l a~ I~gcvhy. New Y~ Alfred & Kno~, 1926. ~. R~m ~ Day ~. ~hol t~ ~ality. In: alcohol ~d health: ~¢w k~wled[e "" (~ t~l ~n m the U.~ ~te~ Wu~n~ D~ U.~ Govtr~ P~ti~ 0~, 1~4.79-~. M3 Prelimimry Communications VI SEROLOGY I1~ THE DETECTIO~ OF TYPHOID CARRIERS C~ZLES M. NOL~ Departmem ~f M~dio'~% U,f~ersity ~ PAUL C. WHITE, JR Ar~as ~ate D~runem of Hmlt~ Li~tk R~ Ar~ns~ JOHN ~ FEEL~ ED~H A. HAM~IE SUSAN L. BRO%~ KwEI-HAY Wo~ ~t~ for Dixmu Cont~ At~nm, G~ Summa~ A new haemagglutination assay for Vi antibodies was evaluated in searches for symptom-free carriers of Salmonella typhi associated with sporadic cases of typhoid fever. The assay differs from pre- vious ones in that a purified (instead o fcrude)Vi antigen from Girrobacter was used to sensitise the red blood cells. In ten sporadic outbre~s of typhoid stool c~lmre identified seven enteric carriers orS. typhi among the patients' families or other close contacts. All seven carriers had Vi antibodies in titres ranging from 1:40 to 1:256.0. Moreover, among thirty- seven stool-culture-negative contacts of patients, only one had Vi antibodies, in a titre ofl: 10. Thus, the new assay for Vi antibodies was as sensitive and as specific as faecal culture in detecting symptom-free typhoid carriers. It could become a convenient screening test. INTRODUCTION THE response ofpublic-heakh agencies to a sporadic case of typhoid fever includes a search for a symptom-free enteric carrier ofSalmonella ~yphi. In the United States, such carriers have been discovered among the patients' close contacts in 30% of cases,t Full advantage should be taken ofthese oppor- tunities to identify symptom-free typhoid carriers in order to attempt to eradicate the infection or at least to institute sur- veillance measures that will minimise the risk of further spread of the organism. Searches for typhoid carriers are often costly and tedious. Three negative facto!cultures are traditionally required toex- el~de each s~tspeet, ~nd many of" the~e c~ltures mast be ob- tained from individuals who are elderly, have a poor under- standing oFthe circumstances, and may consequently be pieious and uncooperative. A reliable screening test would greatly facilitate typlmid- carrier evaluations. We have reported that 71% ofsymptom- free chrot~ic typhoid carriers had ,emm-amibodies to a puff- fm:i highly p.olymerised Vi antigen-' and that these antibodies specifically correlated wkh faecal excretion of,g C~phi. The present study was designed m assess the use~he~ of this Vi serological test in the sea~h for undiagnoscd enteric carriers associated with sporadic cases oftyphold fever. MATERIA/.~ AND METHOD~ For a fottr-year period ~i~ in Jaa~, 1~6, ~h ~¢ ~ TI09782386

"I'I2TR~I~LOID F/~ROBLA~I'S ~ FAa~ILIA.L POLYPOSIS CO~ ~ract~ b7 muMpte ~t~ ~ ~ ~ ~c m~li~nL F~ ~ t~ G~s ~ va~- ant of F~ s~w ~ni~ ~ ~ in ~id~ to ~lonlc ~i~. S~nn~ ~* e( ~. ~a~* that t~trap~y but not familial ~ly~[s wi*bom ~tra~nlc l~iom. Su~- qu~ly ~ ~up ~ o~e v~on in the ~ of tc~rap~ skin fi~ ~ ~h ~mff~l ~ly~ z~ Ga~ncPs synd~me and su~cst~ t~at th~s w~ duc to genetic het~enehy,z ~e ~uH ~ke to ~ke a p~l~inary re~rt ~ our findings in thirt~n ~ly~ ~d~s, tw~ve of whom a~ on ~e ~r of St Mark*s ~i~. Fib~b~zs~s we~ ~hu~ by stand~ ~eth~s f~m skin samples, 3 mm squa~, no~ally taken from the in~ su~acc of tb~ fo~arm. S~c~I p~mary ~Ru~ wc~ ~t u~ wilh h~man or fetal c~If scr~m as m~ium supplemcnl~ but som~ times only one cuhurc p~uccd a viable line of fibmblasts. Ch~m~ome prc~radons wc~ usua{ly made at s~nd or third passage aRer 5-7 weeks' culture. The ~lypo~is patients ranged in age from 13 to 66 years; the cont~]s were ten nor- mal Peo#e s# ~tw~n I I and 56 yca~. Clinical information Dec footnote) and Parentage of tetraploidy in cuhurcd fib~ btast~ are given in the table. TETRAPLOID FIBROBLASTS IN FPC PATIENTS AND CONTROLS FPC TetrapIoidy i 6-5% 0-8 2 33.0" 3 65.0" 4 34.0" $ 1.0% 2.7", 1.2, 5.0 6 4-5" 7 7.5", 2.0 8 I-0", 1-4" 9 4.5 I0 3.5 11 3.0", 1.4 12 1.5", 1.5, 1-3 13 8.1", 0-4, 6-0, 5.1 Contro[~ Tetraploldy 1 3.0", 10-0°, l-0 2 10,5% ii.5", 2.0 3 I-0", 5.0" 4 i7-0" 5 0-0", 4-5", 1.0, 4.0 6 1-5", 5-0", 1.0 7 8-0°, 4.0", 8.5* 8 3.5% 2.0 9 4-5 10 4-0 "Primar cultures initiated in medium with human serum. Tetra- ploidy expressed as % of 200 dividing cells per chromosome harvest. In FPC patient 2 only 90 cells were available. All FPC patients had adcnomas with cancer (cases 1, 2, 4), epldermuld cysts (cases I0, 11, ? 2), ostenma (case 2), ?fibroma (case 5), and congenital retinal pig- mentation (ease 8). All except patients 1 and 13 have a family history of polyposis coil There was considerable variation in levels of-tetraploidy in independently established li~es from the same patient or con- trot. The higher perecm~g~cs for ~ach person usually correlate wi~h initiation o/the culture in medlu,m supplemcmcd with human serum. Sin~ one comml wlue was as high ~ 11.5~ we wouM ~mider this only FFC lines 2, 3, and ~ show e~ascd mtmptoldy. Shannon Danes has c~m~ that the prow ence of cpithdlal cells is n~mry to demonstrate incr~sM tu~s &PC 2 and 4) w~ic~ sho~ b~ tetraplo~y had no ¢pith~al out~wth in the pri~ ~lm~, hcn~ incased ~t~oidy n~ n~ ~ ~at~ to the p~ce ~ cultu~ epithe~al ~. ~r ~ ~ dam wi~ sampl~g at later ~a~ (~n or mo~), ind~ thin tctra#oidy d~ not in~a~ wi~ ~ge level ~ ~in fib~bl~.ts f~ aff~t~ or ~m~ls. ~e ~ tkat the ~1 line with the hlg~s~ tetra- pb~y (~ 3) ex~it~ th~ ~ ~r~ ~ ~gc t~ ~me~ a: di~,d~r~'/.~a¢~ 19"/~; ~ leAS. 1365 ~o. / Piti~ ~ 7 ~ 9 m ~ ~ ~m~ 12; ~a~s. Patinas FPC 2 a~ FPC 11 a~ a~ ~ ~d are t~ t~ ~ticn~ d~t to ~ving h~c G~uer sy~ ph~otype. Howcv~, the lords of tctrapl~y in ~cir ~ltur~ ~in fi~ m ~y ~t. ~ f~ ~m ~ ~i~ that the ~ vafiat~ ~ ~n ~mi~ tctraploidy ~vels in tither FPC or ~Mner's syndrome sampling or cuhufing effects. ~ ~k ~ su~rt~ by the ~r R~a~ U~ty ~ ~n, Joy D. A. D~L ~n NW) 2HE MARY B. PRLTCHARD St Mark's H~#tal, H.J.R. BUSS EY ~ndon ECt B.C. MORSON I'~"~ LOW TAR CIGARETTE SMOKING Sm,--ln the survey of over 18 000 male Civil Servants, Dr Higenbottam and his colleagues (Feb. 23, p. 409) conclude that the tar yield of cigarettes smoked influences phlegm pro- ductlon but not the degree of airflow obstruction. Closer exam- ination of the data, kindly provided by Prof. G. A. Rose (see table), demonstrates not only that dccrcas~Id is associ- ated with an increase in forced expiratory volume but also that this association is similar in magnitude to that between dc- crcased tar yield and decreased phlegm prevalence. MEAN FEV (ADJUSTED FOR AGE AND HBIGHT) AND ~o PHLEGM INCIDENCE (ADJUSTED FOR AGE) BY NUMBER OF CIGARETTES SMOKED AND TAR LEVEL Number of ~ FEV~ cigarettes .............. per day" n n 1-9 I 3"2810"015 102~ $[ /2~_~ I 15801 13"31 10"055 II 1 3.1410.047 23 [ _333_ I 197 I 3"171 0.035 I0=19 [ 18-23 II096 I 3.1610-016 178 , 24-27 I lO+l 3.1slo.o52 16. 28-32 I 231 I 3-07 I 0-034 39: , 33+ I 26913.0310.032 20+ tI~-23II0,~6 [ 3.05 1 0.0~5 ~08~ 2~27 [ g51~.0310.055 . ~32 ] 230[ 3.011~.032 n=number of mac obse~ S~:sttnderd er~r. A~em~ a~{~ mi~ints of 5, 15 and ~ ~tt~ ~r day ~ in reg~ssioa ana~sis. A~m~ appellate mid~nts of 2~ 2~, 30, and 3~ mg us~ in ~es~ioa analysis. ~ phlegm ; [ 29.2s I s.so ' [ 32.03 1 2.40 '112.0112-09 ' I ~.40 I ~-00 ,1 44.7Z I ~4-24 My Conc|usioas were st~ggestcd by ~¢ o~mion that, ~r ~th the f~ expiratory vot~me (FE~ and phlegm datg s~nses in the hi~est tar ~up ~r 1-9 d~rcttes a day sm~ ~ fimi~r ~ thee h ~e b~ tar ~up for 1~19 a day smokes ~d that t~ in the hi~¢st tar ~up ~r t~19 a ~y ~ ~re fimilar to thee h t~ lowest tar g~p ~r ~e~ of ~ ~ mo~ ~tt~ a ~y. ~is tug- ~st~ that ~ucing tar y~ld by a~ut haff(f~m 18-23 ~ had a ~milar ~t mm~n ~ a~ut half (f~ 1~19 m 1-9 a~y) ~h f~ FEV~ #~ ~dmlve ch~ m T109782387

1366 C;garctles ~ ~r ~ a~ T ~s t~ 1~ y~ ~n rcgr~s~ c~i~nts ~or c~[[es ~ tar y~c~ wcrc in ~nit~ and ~th ~ndc~ntly highly significant (p<O.001), their 95~ ~fi~nce limits ~Jn~ r~tivcly -O.Ol~m -O.~a~ ~.Ol13m ~]e my ~vi~ c~s~s do ~ n~arily ~mo~ s~rmc an cff~t of tar ~r ~ ~ FEV ~cls~ ~n~ ~me o~h~ ~ invoDcd i~stcad, they do not pro~de any a priori sup~rtlng Higcn~ttam and ~ll~gues' su~cstion that it is the gas~s p~ of ¢obac~ smoke rather than the tar that to study the relationship ~[w~n FEV and smoke yields in currem ciga~tte smokers, now thai many smogers smoke cigarettes wkh a subs~amiatly lower tar ~ield than smoked at ~he time ~he ~urvey was carried out. 25 Ctdar R~, Sunon SM2 5~ PREVALENCE OF MITRAL VALVE PROLAPSE IN AN AUSTRALIAN POPULATION StR,--Mrs Bonella and her co-workers (April 19, p. 880) describe an increase in crosslinked procollagen in floppy mitral valves removed at surgery and 39~ of single-chain procollagen in these valves with none detectable in normal valves. Mitral vane prolapse is common in the general population and these findings raise the question of the significance of this diagnosis. We have assessed 200 volunteers (100 men and 100 women) for mltral valve prolapse by M-mode echocardiography using rigid and accepted criteria for the diagnosis' with additional criteria to exclude false positives. The volunteers responded to an advertisement in the hospital magazine stating that we wished to establish normal values for our echocardiography laboratory. This yielded a group with a wider age spectrum (mean age +SD, 38.8+12.1 years; range 18-79) than that reported in previous studies. There were eight subjects with unequivocal late systolic mitral valve prolapse, 4 men and 4 women, giving a prevalence of 4%. In 5 of these 8, systolic clicks and/or late systolic murmurs were present on auscultation or on the phonocardiogram, but in the other 3 mitral-valve prolapse could not be detected clini- cally or by phonoeaediography. There were eltmtrocardio- graphic abnormalities in 2~T-wave changes in 1 and left axis deviation in the othea'. 1 had atypical chest l~in and I breath- }e~nes~ for which no reason could bc identified. None plain~ of palpitations. A voiuntocr group may not be r~presentative of the whole population. However, the prevalence we found was remark- ably similar to that reported in two of the thr~e major Ameri- can studies, one in young womena and the other in young men.) "rh¢ one other study4 reported a much higher frequency in young women and the possible reasons for this will be dis, c~ssed elsewhere. Ia a series of 294 routine necropsies Hi~i iBtr LAI~CET, JU~E 21,19g0 ai3 found ~ vaiv¢~ in 5'L (::kady mitral valvc laXla¢sc is common in the general r, otmlation and the frequency is approximatc~ the same in men a~d women. The complications of mitral val~c prolapse, the most com- mon being severe mitral regurgitation, ar~ uncommon, bow- ever. For example, only 11~. of ari single mitrabvalve replace- merits undertaken over the past two years in this cardiopulmonary unit were for Sloppy mitral valves, and a similar incidence bat bee~ roported from the U.S.~ and the U.K.~ The disparity between these numbers and a prevalence rate of 4'~ in the general population is striking. There are two immediate questions. How can indlvioMals with mitrabvalvc prolapse and a high risk of complications be defined? Bariow and Pococks suggest that of people with mitral-vatve prolapse those with a Marfan babitus may r~re- sent a group at risk of having sever~ mhral ~eg~rgha,ion. But there are as yet no data to support this. The other question is whether the myxomatous transformation found in patients with floppy mitral valves and scvero mhrat regurgitation is a non-specific reaction in a congenitally malformed valve or whether it is an expression of a specific mesoderma! abnor- mality with the mhralvalve being the main target organ. The interesting observation of accumulation of procollagcn in floppy mitral valves removed at surgery for severe mitral regnrghation reported by Bonclla ¢t aL support this possibility but clearly does not setde the matter. I'rmc¢ Hene.~ Hospha|, ANDREW J. HICKEY I.mle Hay, New South Wales 2036, Australia DAVID E. L. W1LCKEN A/'/LNION FOR LEG ULCERS Sm,--I was delighted to read the papers by Professor Page Faulk and colleagues (May 31) on the use of human amnlon in the treatment of chronlc ulceration of the legs. "l'wenty years ago, whilst a student nurse at Grey's Hospi- tal, Pietcrmaritzburg, South Africa, a G~y's trained physician, the late Dr R. R. McKenzie, used fresh amnion for the treat- ment of his patients with leg ulceration. It was one of my daily tasks to fetch fresh amnion suspended in saline from the ma- ternity ward. As my training progressed I was permitted to do the dressings. We were instructed to use the amnion as soon as it arrived on the ward. It was applied directly to the wound (which had been cleaned with normal saline), and covered by a non-adherent dressing. The limb was firmly bandaged with cr~pe bandages and elevated. This was standard practice for all Dr McKenzle's patients with chronic ulceration of the legs. We never expected the treatment to fail or the supply of amnion to dwindle. In 1967, as a staff nurse on a gynmcologieal wa~ at a I.on- don teaching h~pital, I suggested the same treatment for a patient with a chronic-varicose ulcer. My suggestion was greeted with scorn but an adventurous registrar allowed me one chance to prove the ~x:acy of amnioa, ~rovid~l tl~ con- sultant was not informed. Within two w~eks ~to my great relief) the wound granulated well and autografting was not thought nee.essary. A ddighted patient was di,schargcd cured of both her gynecological and varicose problems. However, I unable to live down the reputation of rattllng bones and hav- ing consone~ with African wkch-doctors. Recently, having returned to nursing after a g~p of some 5. Hill l~ Oa~ M.J. ~rti~ /~V. The ~t~ 6. ~y DA, Geraml S, HaBmIn G~ ~kar~ ~ Hall 1972; ~ 3~. T109782388

CLIPS