Women's health

In reply: We feel that Hayes and colleagues repeat the editorial comments of Dixon1 that accompanied our article.2 Our investigation, being retrospective, had biases and limitations normally associated with that type of study. In the third and fourth paragraphs of the discussion section of our article, we examined further the sources of bias and their impact. Hayes and colleagues state that the omission of tumour grade and receptor status from our analysis is a flaw. We agree that including these variables would have made our study more complete, even if we doubt that this would have altered the conclusions. So many of these data were missing that statistical analysis would have been unreliable. As was stated in the methods section, the analysis was adjusted for the covariate HRT use before diagnosis. Tamoxifen use was very similar among HRT users and non-users (58% and 60%, respectively), and any bias introduced by this difference would be insignificant. The scope of our article, as suggested by its title, was limited to the issue of cancer recurrence and mortality. We did not specifically investigate the impact of HRT on cardiovascular and thromboembolic events. Nonetheless, if car-diovascular disease had increased mortality, this would have been apparent as an increase in all-cause mortality. We agree with Hayes and colleagues that it is vital that the media interpret studies correctly for the general public. We are disappointed that Hayes and colleagues have misinterpreted our conclusion. Our article does not advocate HRT use by women with breast cancer. We explicitly stated that "These results need to be confirmed in a randomised trial before HRT can be advocated for all women who have had breast cancer." Furthermore, ". . . the observed association between HRT use and [reduced] risks of breast cancer recurrence and death cannot be inferred to be causal". We believe that we are justified in our conclusion that HRT use in women diagnosed with breast cancer is not associated with an increased risk of recurrence of breast cancer or a shortened life span. It is our clinical practice to use HRT only when alternative therapy fails.

Eva M Durna · Barry G Wren · Gillian Z Heller · Leo R Leader · Peter Sjoblom · John A Eden

In reply: Walters and Graham question the role of low-molecular-weight heparin (LMWH) as a replacement for unfractionated heparin during pregnancy, and cite the lack of randomised comparisons to support their view that the standard initial treatment for pulmonary embolism during pregnancy remains intravenous unfractionated heparin. We do not deny the lack of clinical trials of LMWH in pregnancy; we were simply referring to the increased use of LMWH.1,2 However, the lack of evidence of effectiveness does not equate with evidence of lack of effectiveness of LMWH in pregnancy. Clinical trials are needed to determine optimal anticoagulant strategies during pregnancy, particularly in patients with prosthetic heart valves. While awaiting the results of these trials, we believe that the major pharmaco-kinetic and safety advantages of LMWH over unfractionated heparin, coupled with an extensive body of evidence demonstrating their efficacy and safety in non-pregnant patients, should not be ignored in our pursuit of optimal anticoagulation therapies. Williamson and Street question the validity of asymptomatic deep vein thrombosis as a surrogate for symptomatic venous thromboembolism in patients undergoing major orthopaedic surgery. Further, they cite a 0.1%–0.2% incidence of pulmonary embolism in patients undergoing hip replacement surgery without prophylaxis3 to support the conclusion that any effect of thromboprophylaxis on reducing symptomatic events is likely to be irrelevant for individual orthopaedic surgeons. We believe that their argument is seriously flawed. Firstly, the "meta-analysis" they cite3 had major methodological limitations, as elegantly highlighted by "Sherlock Holmes" in his critical appraisal of systematic reviews of surgical thromboprophylaxis.4 Secondly, rigorously conducted randomised trials and meta-analyses of randomised trials have demonstrated the efficacy of antithrombotic therapy for the prevention of both symptomatic and fatal venous thromboembolism in high-risk surgical patients, including lower-limb orthopaedic surgery.5,6 Thirdly, the clear correlation between reduction in asymptomatic and symptomatic venous thromboembolism in patients undergoing elective joint replacement surgery7 suggests that asymptomatic thrombosis detected by screening venography is a valid surrogate for symptomatic events. Fourthly, we agree that an individual orthopaedic surgeon performing 50 joint replacements per year may remain unaware of a small reduction in fatal pulmonary emboli in his or her own practice (eg, a 0.1% absolute risk reduction would be equivalent to preventing one death in 20 years of practice). Yet, on a population basis, even a 0.1% absolute reduction (which is likely to be an underestimate — the PEP study showed a 0.3% absolute reduction in fatal pulmonary embolism with aspirin5) equates to 50 preventable deaths per year in Australia alone8 and many thousands worldwide. There is now overwhelming evidence of the efficacy of thromboprophylaxis for preventing venous thromboembolism, including symptomatic and fatal pulmonary embolism, in high-risk surgical patients. With the rapid ageing of the Australian population and the expected increase in joint replacement surgery in coming years,9 the failure to use effective thromboprophylaxis in orthopaedic patients will likely result in a growing burden of preventable morbidity and mortality from venous thromboembolism.

John W Eikelboom · Graeme J Hankey

Objective: To determine the chance of at least one live birth from one round of in-vitro fertilisation (IVF) treatment and the effect of the woman's age on that likelihood.Design: Retrospective analysis of outcomes from IVF treatment that did not involve donated gametes, but which included embryos cryopreserved in the retrieval cycle.Setting and patients: All IVF patients (median age, 36 years; range, 22–48 years) who attended a private IVF clinic in Sydney for an egg retrieval between 1 January 1998 and 31 December 1998, and had embryo placements (fresh and cryostored) performed up to 30 June 2001.Main outcome measure: Independently audited live births surviving the neonatal period.Results: 565 women had 648 egg retrievals during the period. The age of peak utilisation of IVF was 39 years. For women aged 34 years or less, the chance of a live birth from one round of egg retrieval and IVF treatment was 52.4% (95% CI, 47%–59%). For women aged 35–44 years, there was a linear decline in the live birth rate, and no babies were born from retrievals at age 45 years and over. There was an age-dependent rise in the frequency of miscarriages, from 10.5% (95% CI, 5%–18%) for women under 35 years, to 16.1% (95% CI, 9%–25%) for those 35–39 years, and 42.9% [95% CI, 24%–63%] for those over 40 years (P < 0.001). A third of the first births resulted from embryo transfers performed after a period of cryostorage.Conclusion: As fertility with IVF falls from the age of 34 years, and the age of peak IVF utilisation is 39 years, many Australian women are seeking IVF at an age when the likelihood of a live birth is reduced.

Robert P S Jansen MD, FRACP, FRANZCOG, CREI

To the Editor: It was estimated in 1989 that cervical screening in Australia was preventing only 46% of squamous malignancies, against a theoretical capacity of 90%.1 This suboptimal achievement after almost 25 years of cervical screening led to a major reorganisation of the program. The 1989 analysis has been repeated, using the most recent year (1998) for which incidence rates have been published (Box). The more recent figures suggest that cervical screening in Australia is now preventing 70% of squamous carcinoma of the cervix. This remarkable improvement can probably be attributed to the improved participation by women in regular screening, improved standards within laboratories, and better follow-up of cytological abnormalities. While there is still scope for improvement, there is clear evidence of a substantially better gain from cervical screening. Continued efforts to increase the participation rate among women aged 60–69 years is appropriate given that the prevented proportion appears to be lower in this age range. Percentage of squamous carcinoma of the cervix prevented, by age group, Australia 1998 Age group Number of women (estimated number with a cervix*) in Australia2 Expected rate (per 100 000 women) of squamous carcinoma without screening† Expected number of squamous carcinomas‡ Estimated number of squamous carcinomas observed in 1998§ Percentage prevented 20–24 665 691 (665 025) 5 33.3 9.6 71.1% 25–29 733 145 (732 412) 15 109.9 34.8 68.3% 30–34 706 925 (687 838) 25 172.0 62.9 63.4% 35–39 748 913 (728 692) 45 327.9 74.7 77.2% 40–44 702 629 (608 477) 45 273.8 76.2 72.2% 45–49 649 539 (562 501) 45 253.1 81.4 67.8% 50–54 570 287 (410 607) 45 184.8 48.1 74.0% 55–59 431 183 (310 452) 45 139.7 40.7 70.9% 60–64 370 123 (251 314) 45 113.1 40.7 64.0% 65–69 348 707 (236 772) 45 106.6 44.4 58.3% Total 5 927 142 (5 194 090) 1714 514 70.0% * Calculated by multiplying the number of women in Australia by the estimated age-specific hysterectomy fractions.3 † Methodology as for the study by the International Agency for Research on Cancer,4 using incidence in Norway at a time when the rates would have been little affected by screening. ‡ Calculated by multiplying the estimated number of women in Australia with a cervix by the expected rate of squamous carcinoma of the cervix in the absence of screening.4 § Calculated by multiplying the number of cases of cervical cancer observed in 1998 by 0.74, which was the proportion of all cervical cancers that were squamous.5

Heather S Mitchell

Clinical questionA physiotherapist at Monash Medical Centre wanted to know the most effective physiotherapy treatments for lymphoedema following treatment for breast cancer. Search questionPeople undergoing physiotherapy for lymphoedema formed the patient group of interest. Clinical outcomes were reduction or prevention of lymphoedema. A randomised controlled trial comparing physiotherapy treatment to no treatment, placebo, or other conservative treatment would be the most appropriate study design. SearchThe search term "lymph(o)edema" was combined with the general treatment search terms "physical therapy", "physiotherapy", "conservative treatment", "non-invasive treatment" and the more specific treatment terms "complex physical therapy", "compression sleeve/bandaging/garment", "manual lymphatic drainage", "mechanical compression" to identify all English-language controlled trials, meta-analyses and systematic reviews published between January 1970 and August 2000. We electronically searched the Cochrane Library, Best Evidence, MEDLINE, Cumulative Index of Nursing and Allied Health Library (CINAHL), Current Contents and the Physiotherapy Evidence Database (PEDro). We also searched the websites of 12 relevant health organisations, including the National Guidelines Clearinghouse, National Breast Cancer Centre, and Agency for Healthcare Research and Quality. Summary of findingsOur search identified two systematic reviews,1,2 and five randomised controlled trials3-7 and one non-randomised controlled trial8 published subsequently to, or not included in, the systematic reviews. There are few well-designed randomised controlled trials that examine the effectiveness of physical therapies for lymphoedema. The two identified systematic reviews1,2 included uncontrolled studies with methodological problems, and their conclusions cannot be confidently endorsed. Furthermore, as no single validated and widely acceptable outcome measure is available, comparison of results across trials or pooling of results in a meta-analysis is not appropriate. Some conclusions can be drawn by examining the controlled studies included in the systematic reviews, and the subsequently published randomised3,4 and non-randomised8 controlled trials. While there is some evidence that compression garments reduce limb swelling, the addition of other modalities, including multilayer bandaging,4 manual lymphatic drainage,3,8 pneumatic pumps or electrical stimulation,1 does not provide additional benefit. However, one randomised controlled trial found that compression sleeves provided only short term benefit (over four weeks),6 while the remaining randomised controlled trials reported that compression was no more effective than manual lymphatic drainage7 or no active treatment5 to reduce lymphoedema. The randomised controlled trials of Dini et al5 and Hornsby6 illustrate the importance of appropriate control groups. Both reported that adherence to skin-care guidelines and other self-care protocols reduces lymphoedema as effectively as the addition of active treatment. Although the International Society of Lymphology9 recommends complex physical therapy (skin hygiene, lymph drainage, bandaging, and exercises in treatment and maintenance phases) as the initial treatment for lymphoedema, we found no rigorous evidence to support this. OutcomeSubsequently published evidence-based clinical practice guidelines10 support our findings, namely that compression garments are of some but limited value, although they are as effective as other, more resource-intensive methods. Other physical therapies, such as complex physical therapy, pneumatic pumps, rigorous massage and compression bandaging, require further rigorous evaluation before recommendations about their use can be made. These recommendations were incorporated into a lymphoedema service for women in south-east Melbourne.

Renea V Johnston PhD · Jeremy N Anderson MD, FRANZCP · Barbara L Walker BAppSci(Phty), GradDipHealthAdmin

To the Editor: It has been known for a long time that washing soda (crystalline sodium carbonate) is effective for removing fluid from joint effusions. The crystals are simply wrapped in a tea towel, crushed with a rolling pin and wrapped around the joint overnight. In the morning the sodium carbonate is rock solid and the joint effusion has markedly improved. I was discussing with one my patients her problem of gross upper-limb lymphoedema after surgery and radiotherapy for breast cancer to the axilla. She went home and made an appropriate pack of sodium carbonate, which she wears overnight. She can now use her arm throughout the day. Obviously, the lymph fluid reaccumulates because her lymphatic system is well and truly obstructed. This patient spoke to several other patients in the Day Centre at the Monash Medical Centre, who tried this in addition to other exercises for lymphoedema, and they have found it to be extraordinarily effective. One woman had gross oedema of her hand, which rendered it useless: she simply immersed her hand in a solution of sodium carbonate and thus dialysed the fluid from her hand. Thereafter she could use her hand for 12 hours before significant amounts of fluid reaccumulated. While this is obviously not the perfect solution to lymphoedema in patients with breast cancer, anything that may help them is worth noting. Perhaps a suitable linen device could be made which would cover the whole arm at night. Evidence for the use of this simple dialysis treatment is currently anecdotal. It might be appropriate to design a clinical trial to determine whether this method has potential in treating this extremely troublesome form of lymphoedema.

Graeme N Brodie

Objectives: To review the choices of couples relinquishing frozen embryos and the outcomes of embryo donation at a major in-vitro fertilisation (IVF) clinic.Design and setting: Retrospective audit of 11.5 years of data (1991–2002) from the Monash University IVF clinic, Melbourne.Participants: Couples who make decisions regarding the fate of their frozen embryos, and recipient couples taking part in embryo adoption.Main outcome measures: Couples' choices with regard to the fate of their frozen embryos, and the outcome of donated embryo treatment cycles.Results: Of 1246 couples relinquishing frozen embryos, 1116 (89.5%) opted to discard rather than donate their embryos. Sixty-six per cent of donated embryos survived thawing. From donated-embryo transfer to 50 women in 92 cycles, a 17.4% pregnancy rate per transfer cycle was achieved, and 10 women delivered 11 healthy babies at term. At the time of our audit there were 98 couples on the waiting list to adopt embryos.Conclusions: It is worth considering how couples can be encouraged to donate rather than discard their surplus frozen embryos. An educational program on relevant legal, social and clinical issues may facilitate this.

Gabor T Kovacs MD, FRANZCOG · Sue A Breheny BSc(Hons) · Melinda J Dear BAppSc

The death of a friend solved a centuries-old, oft-fatal mystery My doctrine is produced in order to banish the terror from lying-in hospitals, to preserve the wife to the husband, and the mother to the child... — Ignaz Semmelweis, 1861 Today, a very large proportion of women giving birth receive antibiotics, potent and sometimes in combination, during their accouchement. Routine prophylaxis is widely accepted for caesarean sections, which account for 20%–25% of deliveries.1 Of course, any pregnant woman presenting with an obvious infection will automatically receive an antibiotic. Further, so will most pregnant women with membranes ruptured for any length of time, either before or after labour begins, and any woman in labour with a raised temperature. There is also a more relaxed approach to many former midwifery routines — for example, the abandoning of masks and gowns and the admission of several support people to the delivery scene — which could diminish both younger obstetricians' and midwives' appreciation of the potentially deadly risk of puerperal infection. However, until relatively recently in developed countries, and still in many developing countries, puerperal sepsis was and is a killer.2 Fatal feverImmediately postnatally, the placental site is a large open wound — easily invaded by ascending bacteria. For thousands of years, it was recognised that puerperal women were at risk of a fever that could be fatal. The Hippocratic writings contain references to childbed fever, as do some Hindu texts dating back to 1500 bc.3 Moreover, the potential for birth attendants to initiate such infections seems to have been comprehended by some of the ancient writers, including the Greek physician Soranus, and the Hindus, since advice on hygiene for birth attendants was offered.3,4 Triptych showing the Hôtel Dieu in Paris, about ad 1500. The comparatively well patients (on the right) were separated from the very ill (on the left). Note there were always two patients to a bed. Nevertheless, in ancient and medieval times, mortality from puerperal sepsis was apparently relatively low, as women generally gave birth at home. Peculiar to the puerperium?The 17th century saw the establishment of "lying-in" hospitals in many European cities. While these institutions were, in some ways, an advance — in particular, by relieving obstructed labour with forceps or intrauterine manipulation — the crowding of patients, frequent vaginal examinations and the use of contaminated instruments, dressings and bedlinen spread infection in an era when there was no knowledge of antisepsis. The first recorded epidemic of puerperal fever occurred at the Hôtel Dieu in Paris in 1646. Subsequently, maternity hospitals all over Europe and North America reported intermittent outbreaks, and even between epidemics the death rate from sepsis reached one woman in four or five of those giving birth.5 Numerous bizarre theories as to the cause of childbed fever were expounded — among them that it was due to a "miasma", or the labouring woman's disturbed state of mind, or mechanical pressure from the distended uterus. Certainly, childbed fever was universally regarded as a condition peculiar to women in labour.6 Sepsis suspectedContagion as the basis for childbed fever was first suspected by a number of British physicians in the late 18th and early 19th centuries.7 The name of Thomas Watson, Professor of Medicine at King's College Hospital, London, is not well known, but in 1842 he wrote: "Wherever puerperal fever is rife, or when a practitioner has attended any one instance of it, he should use most diligent ablution." Watson recommended handwashing with chlorine solution and changes of clothing for obstetric attendants — everything, he said, "to prevent the practitioner becoming a vehicle of contagion and death between one patient and another." Unfortunately, Watson's advice seems to have been largely ignored by obstetric practitioners of the time — the contagion theory is completely absent from contemporary obstetric texts.7 Oliver Wendell Holmes Across the Atlantic, in Boston, Dr Oliver Wendell Holmes — pathologist, physician and president of the Boston Society for Medical Improvement — developed an interest in the condition after two related cases were presented to his society. A physician and a medical student both died of septicaemia after performing an autopsy on a woman who died of puerperal fever. Holmes read the existing literature, and became convinced that the condition was highly contagious, and that doctors, nurses and midwives were the active agents of its spread. He began to speak and write on the subject, and in 1843 published his classic essay The Contagiousness of Puerperal Fever.8-10 The essay contains eight rules for the obstetrician, which included not only handwashing and changes of clothing, but also the avoidance of autopsies if obstetric cases were being managed. Holmes' conclusions were ridiculed by many of his prominent contemporaries. For example, Charles Meigs, a well-known obstetrician, was incensed at the suggestion he may himself be transmitting disease. "Doctors," he said, "are gentlemen, and gentlemen's hands are clean."10 Connection comprehendedMeanwhile, in Vienna, Dr Ignaz Semmelweis, a native of Hungary, was beginning a life-long obsession with finding the cause of, and preventing, puerperal fever. However, knowing no English, and far from North America, Semmelweis was unaware of the work of Holmes. Ignaz Semmelweis In 1844, Semmelweis was appointed assistant lecturer in the First Obstetric Division of the Vienna Lying-In Hospital, the division in which medical students received their training. He was appalled by the division's high mortality rate from puerperal fever — 16% of all women giving birth in the years 1841–1843. In contrast, in the Second Division, where midwives or midwifery students did the deliveries, the mortality rate from the fever was much lower, at about 2%. Semmelweis also noted that puerperal sepsis was rare in women who gave birth before arriving at the hospital.6,11 Over the next few years, Semmelweis studied and rejected numerous hypotheses. He did note that medical students and doctors from the First Division performed autopsies each morning on women who had died in the hospital the previous day and that midwives were not required to perform such autopsies. However, he did not immediately appreciate the connection between the two observations.6,11 In March 1847, Jakob Kolletschka — professor of forensic pathology, colleague and friend of Semmelweis — died of septicaemia after sustaining an accidental wound to the hand during an autopsy. On reading the report of Kolletschka's autopsy, Semmelweis was struck by the similarity of the pathological findings to those of women who had died of puerperal fever. He later wrote: "Suddenly a thought crossed my mind: childbed fever and the death of Professor Kolletschka were one and the same. His sepsis and childbed fever must originate from the same source . . . the fingers and hands of students and doctors, soiled by recent dissections, carry those death-dealing cadavers' poisons into the genital organs of women in childbirth . . .".6,11,12 Semmelweis began experimenting with various cleansing agents and, from May 1847, ordered that all doctors and students working in the First Division wash their hands in chlorinated lime solution before starting ward work, and later before each vaginal examination. The results were extraordinary — the mortality rate from puerperal fever in the division fell from 18% in May 1847 to less than 3% in June–November of the same year.11 Doctrine dismissedLike Holmes, Semmelweis found that his conclusions did not receive immediate acclaim from his colleagues and superiors. Indeed, he was treated with scepticism and ridicule by many in the Viennese and wider European medical establishments, including his own professor, Johann Klein. In 1849, Semmelweis' contract with the Lying-In Hospital was not renewed and he returned to Hungary, joining the University of Pest. He presented his findings to the Medical Society of Vienna in 1850. They were not well received, other opponents at that time including the famous pathologist Rudolph Virchow and the prominent obstetrician Friedrich Scanzoni. Semmelweis did not publish his observations until 1861; again, they were greeted dismissively. Embittered, Semmelweis wrote a series of "open letters" to his former professors, accusing them — rightly, as it turned out — of being "medical Neros" and "murderers".11,12 Sadly, his last years were affected by depression and mental disturbance. In July 1865, he was committed to a psychiatric institution in Vienna, and died there two weeks later — ironically, probably from septicaemia following a cut to a finger.3,6 Acceptance of antisepsisParadoxically, within a few years of his death, Semmelweis' doctrine began to be accepted by the wider medical community. In 1874, Billroth demonstrated streptococci in pus from wound infections, and in 1879 Louis Pasteur identified the haemolytic streptococcus in the blood of a woman with puerperal sepsis.3 Joseph Lister, learning of Pasteur's work and germ theory, began to apply antiseptic principles to the practice of surgery, with a dramatic fall in postoperative deaths from infection. As with Semmelweis', Lister's ideas were also greeted with scepticism and it took nearly 30 years for "Listerism" to be universally accepted by medical practitioners.13 By the end of the 19th century, the need for obstetric asepsis was well appreciated. An authoritative text of 1905 gives detailed instructions for the personal hygiene of physicians and nurses attending confinements and instructions on the performance of internal examinations. The importance of "inculcating in the student the principles of obstetrical cleanliness, mechanical and chemical" is emphasised. The need for meticulous antiseptic care during operative vaginal deliveries and manipulations — more frequent then than now — is reiterated.14 Australian medical practitioners were apparently quick to follow the lead of their overseas colleagues in the application of hygienic measures in obstetrics, and of self-regulation when puerperal fever occurred in their practices. Most stopped attending midwifery cases for a time after one or two deaths among their patients.15 Surprisingly though, despite the new understanding of the importance of antisepsis, puerperal sepsis still occurred frequently in developed countries in which figures were kept. It appears the principles of antisepsis were not universally applied. In England and Wales, in the period 1870 to 1890, the maternal death rate in hospital births was around 1 : 20, of which about 40% were due to infection. In the United States, in the 1890s, 20 000 women a year died in childbirth.14,16,17 In New South Wales, in 1894–1896, among confinements of married women, both at home and in hospital, the government statistician found a death rate of 1 : 148, and he commented on the negligence of medical men in filling the certificates required by law. Up to that point, causes of death had been supposedly accurately recorded for more than 40 years, but, in fact, the puerperal nature of fatal infections in women was frequently omitted.14,15 Globally, the most common and most feared infecting organism at the time was the Group A haemolytic streptococcus, whose virulence appears to have diminished in recent years, possibly due to improved socioeconomic conditions and the use of antibiotics. Normally found on the skin, in the nose and throat, and in the vagina, as well as in skin lesions, the streptococcus was introduced into the genital tract during examinations and deliveries. Lacerations, blood loss and exhaustion from prolonged labour increased the possibility of postpartum infection. Staphylococci, gonococci, coliforms and other bowel flora, as well as anaerobes, were less likely culprits, but have assumed greater importance in recent years, as have Group B streptococci.3,7,15 Debating deliveriesIn Australia overall, as elsewhere overseas, the maternal mortality rate (MMR) actually remained steady from 1900 until the late 1930s (5.95 per 1000 women delivered in 1903; 5.13 in 1933). Among developed nations, the United States had the highest MMR and the Netherlands and Scandinavia the lowest, although there were some individual hospitals with remarkably low rates, including the Rotunda in Dublin, Ireland, and Crown Street Women's Hospital in Sydney, Australia.15,17 Causes of the continuing fatal role of puerperal sepsis were widely debated. The medical profession tended to blame untrained midwives, and moved towards their training and registration, which was achieved by the 1930s. Some attributed the rates of sepsis to high levels of interference in labour and delivery, especially forceps deliveries. In Australia, the "lodge" system of practice — whereby families purchased medical services through lodge or friendly society membership — was held to blame. Busy general practitioners contracted under this system were allegedly likely to try to conduct confinements hurriedly.15-17 There were moves both to increase instruction in obstetrics for general practitioners and to encourage specialist obstetricians to do deliveries.15 Antibiotic arsenalAfter 1935, the situation improved rapidly in developed countries. Early that year "a startling therapeutic success" was announced by Domagk in Germany — the prevention of septicaemia in mice experimentally infected with streptococci after the administration of prontosil, a sulfonamide dye.18 In June 1936, Colebrook and Kenny, in a landmark paper, reported their success in treating established puerperal sepsis in women using prontosil — the death rate in apparently similar cases dropped from around 27% to 8%. Colebrook and Kenny wrote (cautiously): ". . . the very low death rate, taken together with the spectacular remission of fever and symptoms observed in so many of the cases, does suggest that the drug has exerted a beneficial effect".18 History was to prove them correct, and in 1939 Domagk was awarded the Nobel Prize in Medicine and Physiology for his work. Prontosil and other sulfonamides were followed by penicillin, to which streptococci causing puerperal sepsis still remain sensitive, and the arsenal of antibiotics used for all other forms of postpartum fever today.3,19 Today, in Australia, deaths from puerperal sepsis are extraordinarily rare (the MMR is currently about 0.1 per 1000 births).2 However, infection and fever are not rare, and the microbes causing them are omnipresent. In caring for pregnant women, especially the many who have some intervention in labour or delivery, we would be wise to reflect that it is only the use of increasingly complex antibiotic regimens which prevents a return to "the terror of the lying-in hospitals".

Caroline M De Costa FRANZCOG, FRCOG

Christmas Eve, a couple of years ago. I was on call for the birth suite until 8 am the next, Christmas, morning but was hoping to be able to stay at home with my family. At 6 pm, I did a festive round with the registrar on duty. Good — only three women in the suite, and two delivered, both delighted with themselves for getting it over before Christmas Day. In the corner room, one woman in early labour — Grace, aged 34. Elderly for a first baby, especially for an Aboriginal woman, the registrar observes. I say hello to Grace, but don't examine her — that's why the midwives and junior staff are here. Surprisingly, no partner or family is with her. Then I realise that I have seen Grace about our town. She is one of the "park people". Virtually homeless, living mostly outdoors, drifting back and forth between town and some of the more remote communities of the region, the park people are frequently subjected to the ire of some of the town's better-heeled residents. Recently, these residents have demanded more stringent "move-on" laws, to keep the park people out of the sight of the tourists and restaurant patrons along the town's seashore. So far, the State's Anti-Discrimination Commissioner has successfully opposed such laws, but for the park people — rather like those people back in Bethlehem whose birth experience we are celebrating tonight — it seems there is no room at the inn. In her time, like other park people, Grace has had many visits to the hospital's emergency department. At every admission, the same comments appeared: "poor historian"; "C2H5OH"; "lacerations"; "bruises". She'd been sutured many times, with the new and old stab wounds noted. A large scar on her throat and another on her left breast were recorded. Unfortunately, no-one noted the scar on her lower abdomen. Longitudinal and midline. A laparotomy scar. Grace hasn't attended any formal antenatal clinics, but thankfully, during one of her visits to casualty a while back, someone did do an ultrasound scan, so we know she is labouring close to term. Routine antenatal blood tests are being done now. She is in established labour, progressing, and all appears well. I leave it to the registrar to check the results, wish everyone a Merry Christmas, and go home. Two hours later, I am rung by an agitated registrar and, on the strength of what I am told, go back to the birth suite. In conversation with one of the midwives on duty, an elderly Aboriginal woman, visiting another patient, remarked of Grace: "Long time since that girl had a baby!". "Oh no," replied the midwife, "this is her first". "No," the woman was firm. "Had the baby when she was 13. A caesarean. At . . . " — and she named a former mission station some hundreds of kilometres away. The midwife hastened to question Grace. Did she ever have a baby before? It was difficult for her to answer; she is indeed a poor historian. For a start, she has no teeth. Those that weren't knocked out in fights have rotted away. Also, chronic middle ear disease since childhood has made her rather deaf. But she does know that, yes, she did have a baby. A girl. Nobody in the hospital had ever asked her before; she didn't know it was important. Did she have a caesarean? Grace is unsure. It was a long time ago. Did the baby come out through that scar on her tummy? Yes, maybe. It was a long time ago. Where is her daughter now? Grace does not know. Obstetric dilemma: is the abdominal scar longitudinal because Grace had a classical caesar, in which case, because of the risk of uterine rupture, caesarean section should be repeated forthwith? Or — and more likely — was the longitudinal incision merely the route to a standard lower-segment operation, allowing Grace the possibility of a successful vaginal birth this time? For the moment, all seems well. Grace is contracting regularly, has accepted pethidine, and is making progress in labour. She now lies in a clean hospital bed, surprised to be — for perhaps the first time in her life — the centre of concerned attention. We try ringing the hospital near the former mission for more information. It's 9.30 pm on Christmas Eve! We're told: "You want records from more than 20 years ago? You must be joking! Ring back next week." An hour later, a further complication arises. Grace's blood has shown unusual antibodies and it will take some hours to find and crossmatch blood if we need it. We decide to ask for the crossmatch and hope that she delivers vaginally soon and that she won't need surgery or blood. Regularly, anxiously, we watch Grace's vital signs and the fetal monitor. Another two hours later, we have blood, Grace's cervix is 8 cm dilated and the fetal heartbeat has been fine. And then, she begins to bleed. Torrentially. Everyone swings into action. After all, this is what we do best. Acute care. Three wise men appear — anaesthetist, paediatrician and theatre porter — bringing not frankincense, myrrh and a manger but ropivacaine, oxygen and a trolley. In five minutes, Grace is on the operating table; another five, and a spinal block is in place. Soon, a rapid repeat caesarean section is under way. The old scar — in fact, a classical — has ruptured and is bleeding profusely, but it's repairable and the baby is alive. On the stroke of midnight, a baby boy arrives. He is small and scrawny, covered in meconium. But when he gives a feeble cry, Grace smiles and reaches out one arm for him (a blood transfusion is running into the other) and she names him, appropriately, Joseph Christopher. Joseph spends that night and the next in the special care unit. He starts to breastfeed. Grace is eating three meals a day, including turkey and plum pudding. But it's Christmas time, the hospital is short of staff and many beds are closed, so even at this inn Grace cannot stay too long. On Day 5 post-op, Grace and Joseph are discharged "home". As Grace has no home, a place is found for her in a hostel, with domiciliary visits planned. On the first visit, the domiciliary midwife finds things are OK; the next day, Grace and her baby have gone. A few days later — in fact, on New Year's Day — Grace presents to the emergency department again, this time with Joseph. He isn't feeding well and is bringing up feeds; and, he has a fever. But we can deal with all that. It's another acute problem, not one of those complicated social issues that, in hospital practice, just have to be put into the "too-hard" basket. The paediatric registrar arrives, Joseph is admitted and a drip is put up. And so the cycle of disadvantage starts all over again — unto a new generation.

Caroline M De Costa FRANZCOG, FRCOG

When asked about features of women's health services that would best meet their needs, specific groups of Aboriginal and Torres Strait Islander women, despite their diversity, have given very similar responses.1-3 They want women's healthcare that takes a holistic rather than a narrow "single-disease" or biomedical approach; services that are accessible, flexible and supportive; and providers they can trust, who are respectful and who can communicate well. For many Aboriginal and Torres Strait Islander women, having access to a female provider is critical to their acceptance of women's healthcare services. The higher cervical cancer incidence and mortality for Aboriginal and Torres Strait Islander women compared with other women, and the available evidence about screening effectiveness, provide a strong imperative for healthcare providers and funders to listen carefully and respond to what women say they want.4 The article by Coory and colleagues in this issue of the Journal (page 544) quantifies and compares women's participation in cervical screening by analysing data from the Queensland Health Pap Smear Registry.5 Participation for women living in rural and remote Aboriginal and Torres Strait Islander communities in Queensland was generally lower than for women living in other areas. Proportions of women in these communities who had had a Pap smear over a two-year period ranged from 19% to 63%. These results suggest women's needs for women's health services are being better met in some communities than others. In interpreting their analysis, Coory et al used residence in a community where most people were Aboriginal and/or Torres Strait Islander as a proxy for Indigenous status. We believe this is a resourceful and reasonably valid way around Indigenous status not being identified on the Pap smear register. However, one limitation is that we can learn nothing about Aboriginal and Torres Strait Islander women living in other localities (ie, the majority of Aboriginal and Torres Strait Islander women in both Queensland and Australia more generally). It is important that the needs of these women are not neglected because of the lack of quantitative data with which to measure them. We commend the researchers for acknowledging the sensitivities of identifying data from individual Aboriginal and Torres Strait Islander communities in their research. However, rather than only obtaining permission to do so from a government department, we believe consulting directly with members of the communities concerned at an early stage of the project may have been beneficial. Although such a practice is uncommon in this type of research, and may be challenging and more time-consuming, it may also create or strengthen trust, links and understanding, which could be useful when implementing and evaluating subsequent interventions. Coory et al suggest that the higher cervical screening participation rates in some communities are an indication of what is achievable, and express support for a strategy of strengthening primary health care. We agree with these conclusions, but disagree that an intervention study where communities are randomised would be an ideal next step. Although randomised-community intervention trials have been implemented in other settings,6 for Aboriginal and Torres Strait Islander communities the barriers to delivery of women's health services are likely to be highly location-specific and the means to overcome them not amenable to random allocation. We believe any available resources would be better spent on (i) exploring in more detail the factors contributing to high and low levels of participation, and (ii) responding actively to identified issues in communities with lower levels of participation. Barriers to Aboriginal and Torres Strait Islander women accessing women's cancer screening services, and ways of responding to them, have been reviewed — most recently in the context of considering how to support the roles of general practitioners.1,4 We would like to highlight the need to also support the roles of Aboriginal Health Workers (AHWs). Because of their key role in providing primary health care for Aboriginal and Torres Strait Islander people, the need for improved clarity, recognition and support of AHW roles has been identified as a national priority.7 We have worked with many female AHWs who have had personal experience of the impact of cervical cancer on Aboriginal and Torres Strait Islander women and their communities, and are keen to be involved in women's health education and promotion activities. Some AHWs also want to provide women's clinical care, including taking Pap smears. Some of the specific areas needing attention are the provision of better training for AHWs in women's health, and issues of accreditation, legal cover and quality assurance for those wanting to take Pap smears. Finally, we urge caution about evaluating cervical screening programs solely on the basis of participation rates. Recent commentaries have begun to question a primary aim for screening programs of maximising participation, arguing that this may lead to the positive effects of screening being overstated, and the limitations and possible negative effects of screening and its sequelae being ignored or downplayed.8,9 These commentators acknowledge that providing more balanced information about screening may have a negative impact on participation rates, but stress the importance of individuals being informed about screening and being able to choose for themselves whether or not to participate.8 Qualitative research conducted with women in one rural Aboriginal community with high rates of participation in cervical screening found that many of the women had little understanding of cervical screening or its implications.10 For programs successful in terms of participation, questions may remain about the extent to which women are making an informed choice about screening. In many localities, providers' attempts to consistently give adequate information to Aboriginal and Torres Strait Islander women can be constrained by many factors, including lack of time, and language and cultural differences. These barriers, combined with a high level of concern about cervical cancer and evaluation criteria based mainly on participation rates, may lead to an emphasis on persuading women to have a Pap smear rather than on providing information and an opportunity for informed choice. We strongly advocate that evaluators of cervical screening programs take into account not only participation rates, but also Aboriginal and Torres Strait Islander women's views about available health services and their understanding of screening-related issues.

Jennifer M Hunt MB BS MPH FAFPHM Public Health · Lynore K Geia BN, RM, MPH

Objective: To investigate the extent of participation in cervical cancer screening among women who live in discrete rural and remote Indigenous communities in Queensland.Design: Descriptive analysis of data from the Queensland Health Pap Smear Registry for the period March 1999 to February 2001.Subjects: Women aged 20–69 years who had given their address of usual residence as one of 13 discrete rural and remote Indigenous communities in Queensland.Main outcome measures: Proportion of women who participated in cervical screening over a two-year period ("biennial participation percentage") and variation in participation across the 13 communities.Results: Overall, the biennial participation percentage in the Indigenous communities was 41.1%. This was 30% lower (risk ratio, 0.70; 95% CI, 0.67–0.72) than that for the rest of Queensland. There was statistically significant variation among communities, with biennial participation percentage ranging from 19.9% to 63.5%.Conclusions: The variation in participation across the communities suggests that the problem of low participation among Indigenous women is not intractable. Achieving participation rates similar to the highest rates found in our study would be of major benefit to Indigenous women.

Michael D Coory MB BS, PhD · Jennifer M Muller MEnvCommHealth, GradDipHealthProm · Nathan A M Dunn BSc(Hons) · Patricia S Fagan MB BS, FAFPHM

It’s time to adopt a cost-effective approach Hypertensive disorders in pregnancy, and particularly pre-eclampsia, remain major causes of maternal and perinatal mortality,1,2 accounting for 15% of maternal deaths and 4% of perinatal deaths. Therefore, a key aim of modern antenatal care is the timely detection and management of pre-eclampsia.1,2 A traditional belief is that this is best achieved by regular, and increasingly frequent, antenatal visits, allowing for both blood pressure measurement and dipstick urinalysis to detect new-onset proteinuria. This strategy underpins the schedule of antenatal care that is still most commonly followed in Australia; namely, monthly visits until 28 weeks of pregnancy, fortnightly visits until 36 weeks and weekly visits thereafter.3 However, it has been apparent for some time that the frequency of visits could be safely reduced without adversely affecting outcomes,4 a notion now confirmed by randomised controlled trials both in the developed and developing world.5 Similarly, it has long been recognised that dipstick urinalysis performs poorly in the detection of proteinuria,1 requiring confirmation by either a formal 24-hour urine collection or a spot urine protein/creatinine ratio.2 However, the accuracy of a dipstick reading is significantly improved if it is read with an automated device rather than visually,6 offering the possibility that routine automated testing for proteinuria may have a place in the detection of pre-eclampsia. In this issue of the Journal, the study by Murray and her colleagues (page 477) explores this possibility.7 The authors prospectively evaluated automated dipstick urinalysis in the diagnosis of pre-eclampsia in almost 1000 unselected women. In a quarter of the women who developed pre-eclampsia proteinuria arose before hypertension. From this, the authors concluded that if the initial screening urinalysis is negative then routine urinalysis thereafter is unnecessary in women with no high-risk factors for pre-eclampsia. These findings and conclusions should encourage providers of antenatal care to reflect on their own practice and to consider whether routine urinalysis is justified, thereby facilitating the provision of the most cost-effective care. The report by Murray et al should also stimulate us to reflect on the cost-effectiveness of the other tests routinely undertaken during antenatal care. It is of concern that there is considerable variation in routine antenatal testing in our hospitals, and that practice is often at odds with available evidence.8 These inconsistencies are not only indicative of inequalities in care, but also suggest wastage of precious and limited resources. Standardisation of antenatal care across Australia, through the development of clinical practice guidelines, might reasonably be expected to reduce this wastage. In the United Kingdom, the National Institute of Clinical Excellence has commissioned the development of such guidelines with 43 recognised stakeholders and a projected completion date by September 2003 (www.nice.org.uk). In Australia, through a project funded by the Victorian Department of Human Services, the three largest public hospital providers of maternity services in Victoria have already developed consensus guidelines on antenatal care. These encompass the delivery of antenatal care, including guidelines for most of the routine tests undertaken in pregnancy.8 These guidelines provide an evidence-based foundation for the rational delivery of antenatal care in these three hospitals. However, they offer far more. The guidelines could be used as a catalyst for the development of national guidelines for antenatal care. An important component of any such development, and one missing from the Three Centres Consensus Guidelines, must be a thorough cost-effectiveness analysis of the various tests and interventions recommended. A cost-effectiveness analysis is important because much of the evidence for the various antenatal testing is imported from overseas and may not be readily applicable to Australia. For example, a recent cost appraisal of screening methods for Down's syndrome in the United Kingdom costed a first-trimester ultrasound examination at about £4 ($12),9 a fraction of the Medicare cost in Australia ($60–$70). In addition, the prevalence of the various infections, such as syphilis and HIV, varies in different regions of Australia, and consequently the currently recommended strategies for screening may need to be modified on a regional basis.10 Such analyses are critical components of the further development of evidence guidelines, but, frustratingly, there has been little support at a national level for the funding necessary for their development and implementation. This is despite an estimate that between $75 million and $100 million is spent annually on antenatal screening in Australia.11

Euan M Wallace · Jeremy J N Oats

Objectives: To determine whether routine urinalysis in the antenatal period facilitates diagnosis of pre-eclampsia. Can routine urinalysis during pregnancy be discontinued in women with normal results of dipstick urinalysis and microscopy at the first antenatal visit?Design: Prospective observational study.Setting: A metropolitan public hospital and a private hospital in Sydney (NSW).Participants: One thousand women were enrolled at their first antenatal visit (March to November 1999), and 913 completed the study.Outcome measures: The primary outcome was a diagnosis of de novo hypertension (gestational hypertension, pre-eclampsia, or pre-eclampsia superimposed on chronic hypertension).Results: Thirty-five women had dipstick proteinuria at their first antenatal visit. In 25 (71%) of these women, further dipstick proteinuria was detected during pregnancy, and two (6%) were diagnosed with pre-eclampsia. Of the 867 without dipstick proteinuria at the first visit, 338 (39%) had dipstick proteinuria (> 1+) at some time during pregnancy. There were no statistically significant differences in the proportion of women with and without dipstick proteinuria at their first visit who developed hypertension during pregnancy. Only six women developed proteinuria before the onset of hypertension. Women who had an abnormal result of a midstream urine test at their first visit, compared with women with a normal result, were more likely to have a urinary tract infection diagnosed during pregnancy; however, the numbers were small.Conclusion: In the absence of hypertension, routine urinalysis during pregnancy is a poor predictor of pre-eclampsia. Therefore, after an initial screening urinalysis, routine urinalysis could be eliminated from antenatal care without adverse outcomes for women.

Noreen Murray RM, BN · Caroline S E Homer RM, PhD · Gregory K Davis MD, FRACOG · Julie Curtis RN, RM · George Mangos MD, FRACP · Mark A Brown MD, FRACP

Objectives: To describe the epidemiology of gestational diabetes mellitus (GDM) in Victoria.Study design: Population study of all women having singleton births in Victoria in 1996.Methods: Probabilistic record linkage of routinely collected data and capture–recapture techniques to provide an estimate of the incidence of GDM.Main outcome measures: Risk factors for and the adverse outcomes associated with GDM compared with the non-diabetic population by univariate and multivariate analysis.Results: The estimated incidence of GDM was 3.6% (95% confidence interval [CI], 3.60%–3.64%). GDM is associated with women who are older, Aboriginal, non-Australian born, or who give birth in a larger hospital. The adverse outcomes associated with GDM pregnancies were hypertension/pre-eclampsia (adjusted odds ratio [OR], 1.6; 95% CI, 1.4–1.9), hyaline membrane disease (1.6; 1.2–2.2), neonatal jaundice (1.4; 1.2–1.7) and macrosomia (2.0; 1.8–2.3). Interventions during childbirth were also associated with GDM — for example, induction of labour (3.0; 2.7–3.4) and caesarean section (1.7; 1.6–1.9).Conclusion: Women with GDM had increased rates of hypertension, pre-eclampsia, induced labour, and interventional delivery. Their offspring had a higher risk of macrosomia, neonatal jaundice and hyaline membrane disease.

Christine A Stone GradDipEpidBiostat, MPH, MHSc(PHP) · Kylie A McLachlan MB BS, FRACP · Jane L Halliday PhD · Peter Wein MB BS, FRANZCOG, GradDipEpidBiostat · Christine Tippett MB BS, FRANZCOG, MRCOG

To the Editor: Thank you to McVeagh1 for highlighting some more of the amazing scientific evidence regarding the benefits of breastfeeding. It is extremely important to continue to emphasise the benefits to mother and child in order to strengthen the individual's resolve and the community's support for breastfeeding. However, my concern is whether the question "Is breastfeeding best practice?" should ever be posed in the first place. Do our natural physiological processes now need to be supported by an evidence base and scrutinised in terms of whether they conform to notions of "best practice"? And should the question about choice between breastfeeding and artificial feeding continue to be asked? McVeagh also posed the question, "Is there justification in the argument that women are being pushed too hard to breastfeed?". Can there ever be too much encouragement given to women to provide nutrition and nurturing hand-in-hand to their baby? We must keep emphasising that breastfeeding is not only about good nutrition, reduction in childhood obesity and other measurable health outcomes. Surely there must remain some areas in our lives that do not require evidence and scientific support. Breastfeeding is about loving and nurturing a baby. It is about human relationships. When one experiences a newborn latching on to feed, the clearly felt surge of hormones from breast to brain, the physical expression of these hormones as a palpable "let-down" reflex and the incredible sight of milk rushing from the breast on demand, we do not need science to tell us that this is one of life's most amazing and wonderful experiences, nor to confirm what breastfeeding mothers innately know to be best practice.

Sandra L Neate

In reply: I thank Neate for her comments and for challenging the need to ask "Is breastfeeding best practice?". I appreciate her sentiment that there is more to infant feeding than nutrition and health. However, while there are mothers who don't find breastfeeding pleasurable or easy and are deciding how long to persist; while there are mothers who opt not to exclusively breastfeed for six months or to wean before a year of age; while mothers' advisors prescribe solids or complementary feeds or weaning for myriad problems without evidence for effectiveness beyond a placebo effect; while some believe that the disadvantages of not breastfeeding only apply to infants in developing countries; while there are commercial interests promoting products that undermine exclusive breastfeeding; while the Australian government has not fully implemented the recommendations of the World Health Organization Code and subsequent resolutions;1 while health professionals are receiving "educational material" implying that a new additive makes commercial infant formula more like human milk; and while the scientifically minded among us just need to satisfy our curiosity, we need to know to what extent it matters if an infant is breastfed at all, breastfed exclusively, or breastfed for longer periods. Thank you for challenging the question. The weight of the evidence is such that the real question is not "Is breastfeeding best practice" but "By how much?".

Patricia McVeagh

Trial: Hurskainen R, Teperi J, Rissanen P, et al. Quality of life and cost-effectiveness of levonorgestrel-releasing intrauterine system versus hysterectomy for treatment of menorrhagia: a randomised trial. Lancet 2001; 357: 273-277. QuestionWhat is the effectiveness of the levonorgestrel-releasing intrauterine system (IUS) compared with hysterectomy for the management of women with menorrhagia? Trial details Design: Randomised controlled trial, double-blind. Setting: Five university hospitals in Finland. Patients: Premenopausal women aged 35–49 years, with menorrhagia and no major gynaecological abnormalities were recruited to the study. Exclusion criteria were submucous fibroids, endometrial polyps, ovarian tumours or masses greater than 5 cm, cervical disease, urinary and bowel symptoms or pain resulting from large fibroids, lack of indication for hysterectomy, history of cancer, menopause, severe depression, metrorrhagia as a main complaint, previous treatment failure with levonorgestrel-releasing intrauterine system (IUS), severe acne, and uterine malformation. Interventions: 119 patients were allocated to the IUS group and 117 were allocated to the hysterectomy (any method) group. Hysteroscopy was performed before randomisation only if clinically indicated. Main outcome measures: Quality-of-life measures (including health-related, measured on a scale of 0–1 with the EuroQuol [EQ-5D] instrument), anxiety measures, resource use, menstrual blood loss, haematological indices. Main results: In the IUS group, 24 women (20%) had had a hysterectomy and 81 (68%) continued to use the system at 12 months. Of the women assigned to the hysterectomy group, 107 underwent the operation (5 cancelled before surgery and 5 withdrew from the study group). Health-related quality of life improved significantly in both the IUS and hysterectomy groups (change in EQ-5D, 0.10 [95% CI, 0.06–0.14] in both groups), as did other indices of psychological wellbeing. There were no significant differences between the treatment groups except that the hysterectomy group had less pain than the IUS group at 12 months. Overall costs were about three times higher for the hysterectomy group than for the IUS group. Conclusion: The significant improvement in health-related quality of life highlights the importance of treating menorrhagia. During the first year the levonorgestrel-releasing IUS was a cost-effective alternative to hysterectomy for treating this disorder. CommentaryRationale for the trialInternational hysterectomy rates vary considerably; by the age of 60 years, a third of women in the United States will have undergone a hysterectomy and menorrhagia is one of the most common indications.1 Satisfaction rates with hysterectomy have been consistently reported as 95% or higher.2 Quality-of-life measures have also been reported to improve following hysterectomy.3 It is perceived as a permanent solution to the problem of menorrhagia. However, short-term complications following hysterectomy are not uncommon; recent studies report rates of moderately severe complications of 16% while in hospital.4 The IUS is an intrauterine system that can be sited during a clinic visit and its duration of action is five years. The IUS contains 52 mg of levonorgestrel which is released through a rate-limiting membrane (20 μg/24 h), resulting in endometrial shrinkage and a reduction in menstrual bleeding. As the IUS has been proposed as an alternative to hysterectomy, it was important to establish if improvements in quality of life are reported with the IUS. Trial methodsThe trial was well conducted and reported according to published guidelines.5 Patients were allocated to the two treatment groups without the prior knowledge of the investigators, although blinding of patients and investigators was not possible in the follow-up period. An instrument measuring disease-specific QOL may have made QOL differences between the two groups more apparent; however, no disease-specific QOL measure exists. The completeness of follow-up was 97% (228/236), and an intention-to-treat analysis was done. In the IUS group, a third of devices were removed and 20% of the women underwent hysterectomy during the first year of follow-up. Reasons for removal included spotting and other bleeding disorders. This removal rate is higher than the two smaller trials comparing IUS with endometrial resection, which report a 20% removal rate, possibly reflecting the severity of symptoms among women who seek hysterectomy instead of endometrial resection.6,7 The use of health-related quality of life (EuroQol) and other measures of psychological wellbeing was appropriate, and the measures had been validated in the population before the study. Not all women referred for menorrhagia were included. Most of those not participating either did not want to take part or did not meet the eligibility criteria for either of the two treatments. Therefore, it is likely that the study group was representative of women who were candidates for both treatments, and the results can be generalised to the population of women with menorrhagia. The planned five-year follow-up will be welcome. New informationTwo small trials had previously reported that the IUS was an alternative to hysterectomy.8 However, this large randomised trial has improved on those studies, as it included a wider range of outcomes, and shows that, after a year, no statistically significant differences were present between the treatment groups in terms of health status, health-related quality of life, and psychological wellbeing. Cost-effectiveness data favour the IUS. Implications for clinical practiceThis study has shown that, in women with menorrhagia, there were no statistically significant differences between IUS and hysterectomy for the outcomes of quality of life and psychological wellbeing. The quality-of-life data show that the IUS scored as well at 12 months as hysterectomy in seven of the eight areas assessed, with the only difference between the groups being pain scores, despite a 66% reduction in the IUS group. The IUS is by no means the final answer to the problem of menorrhagia — a third of women had the device removed by one year because of bleeding problems. Costs to both the healthcare system and the patient were reduced with the use of the IUS, although these cost differences may reduce with time. The costs of hysterectomy may be higher because of the 30% complication rate reported in this study — similar rates have been reported by others.4,9 This study provides convincing evidence that the IUS is an effective alternative for women with uncomplicated menorrhagia, and should be offered as a treatment choice.

Cynthia M Farquhar MD, FRANZCOG, CREI

Joe Correy, an outstanding contributor to the field of obstetrics and gynaecology in Tasmania, died on 18 April 2002 of severe Parkinson's disease. Joe was born in Manilla, New South Wales, on 17 June 1925, and moved to Sydney in 1930. Although his family was not well off, he won scholarships that enabled him to study medicine at the University of Sydney, from which he graduated in 1947. After completing his residency at Sydney Hospital, he began training in obstetrics at the Royal Hobart Hospital. He married Lucy Denne in 1950 and went to Oxford, UK, in 1951. He obtained his Membership of the Royal College of Obstetricians and Gynaecologists in 1952 and returned home in the following year. With his great personality, energy and ability, Joe built up a very successful private practice in Hobart over the next 15 years. In 1966, he received a William Morton Lemon scholarship to study ovulation induction in Melbourne, and returned to establish the Gynaecological Endocrinology and Infertility Clinics in Hobart and Launceston later that year. In 1969, accepting an appointment as Director of Obstetrics and Gynaecology at the University of Tasmania's medical school, Joe began a remarkable transition from private practitioner to full-time academic, involving research work, training to run a university department, study and teaching. By 1977 he had been promoted to the position of Professor, which he held until his retirement in 1990. With boundless energy and enthusiasm, he continued to take on new challenges throughout his academic career. In 1972, under a Brown Craig Travelling Fellowship, he studied ultrasound in the United Kingdom and returned to set up and run the first ultrasound service for pregnant women in Tasmania. In 1973–1974, he trained to acquire proficiency in pelvic cancer treatment. In 1981, he organised the IVF Clinic in Hobart, which he ran successfully for many years. He also initiated the collection of statistics for all public and private obstetric procedures in Tasmania. Joe was an active member of many medical committees, including (at various times) secretary and president of the Tasmanian branch of the Australian Medical Association, president of the Royal Australian College of Obstetricians and Gynaecologists, an examiner with the RACOG and the Royal College of Obstetricians and Gynaecologists, and president of the Tasmanian Medical Council and the Australian Medical Council. Joe's enthusiasm for life encompassed car trials, great parties, trips with his travellers' group, bridge and lawn bowls. Integrity, honesty and wit were inherent in his character. He was made a Member of the Order of Australia in 1986 for services to medicine. He also became an Honorary Fellow of the Royal Australasian College of Surgeons in 1975 and received an Advance Australia Award in 1994. Joe is survived by his second wife, Pam, and his three children and their families.

Gerard Gartlan FRANZGOG FRCOG

Probably in the short term, but results of ongoing trials are needed to determine longer-term safety Oestrogens play an important role in the development of breast cancer. This is most evident in postmenopausal women: circulating levels of endogenous oestradiol are higher in those who develop breast cancer,1 while use of hormone replacement therapy (HRT) increases breast cancer risk.2 Recent results from the Women's Health Initiative randomised trial showed a 26% excess rate of breast cancer development in women who took combined continuous equine oestrogens and medroxyprogesterone acetate for a mean of 5.2 years compared with placebo.3 This finding is consistent with results of earlier epidemiological studies that suggest breast cancer incidence is increased more by combined preparations than by oestrogen alone.2 Further evidence that oestrogen is important in breast cancer development comes from a study of over 9300 postmenopausal women with early breast cancer.4 This found that anastrozole (an aromatase inhibitor that dramatically reduces oestrogen production) significantly reduced the rate of new contralateral breast cancers compared with tamoxifen (hazard ratio, 0.42; 95% CI, 0.22–0.79; P = 0.005).4 Despite the increased incidence of breast cancer in women who use HRT, most studies have shown either no effect on mortality or a decrease.5 The reason appears to be that breast cancers that develop in HRT users are smaller and clinically less advanced, with a lower rate of node positivity, better differentiation and more favourable histological type, than cancers that develop in women not using HRT.2 Menopausal symptoms are reported by two-thirds of postmenopausal women with breast cancer.6 Can HRT be safely used by these women, or does it have the same impact on breast cancer recurrence as it appears to have on breast cancer development? A number of publications have addressed this issue. One systematic review documented 11 studies involving 214 women who took HRT after a diagnosis of breast cancer, and found that the risk of breast cancer recurrence was lower in HRT users (relative risk [RR], 0.64 (95% CI, 0.36–1.15) than in control women who did not use HRT.5 Durna and colleagues report similar findings in this issue of the Journal (page 347).7 They found significantly lower rates of recurrence (RR, 0.62; 95% CI, 0.43–0.87) and death from breast cancer (RR, 0.40; 95% CI, 0.22–0.72) in women who used HRT compared with non-users.7 These are important data and are also consistent with those of a recently published United States case–control study of 174 women who chose to use HRT after breast cancer diagnosis and matched non-users.8 These three studies reported a consistent reduction in recurrence and death from breast cancer in breast-cancer survivors who used HRT to treat menopausal symptoms, but all had potential confounding factors.5,7,8 All studies to date have been observational and are thus subject to a variety of biases. In the Australian study, women who used HRT after treatment of breast cancer had smaller tumours and fewer involved nodes compared with non-users, and were more likely to have used HRT before diagnosis.7 Although the final model adjusted for a number of prognostic factors, these did not include tumour grade, concurrent use of tamoxifen or oestrogen-receptor status. Concurrent tamoxifen is a particular confounding factor, as it was prescribed for almost 60% of women who used HRT, and would have limited the effects of oestrogen on normal and malignant breast epithelium.9 Duration of HRT use after a diagnosis of breast cancer was short in all three studies — a median of only 1.75 years in the Australian study. Surprisingly, in both the Australian7 and American8 studies there appeared to be a lower rate of breast cancer recurrence in patients taking progestogen alone, vaginal oestrogen alone, or a combination of the two. It is difficult to believe that the small amounts of oestrogen absorbed from vaginal preparations could have a positive influence on breast cancer recurrence and survival. This suggests that other characteristics of women who use HRT, whether vaginal or oral, may influence outcome. Socioeconomic status is an independent predictor of breast cancer recurrence and survival: women with more education and higher socioeconomic class have a lower recurrence rate and better survival.10 Hot flushes are more commonly reported by educated women,6 and these women are more likely to take HRT. Thus, socioeconomic factors could conceivably be part of the reason for the better outcome of women with breast cancer who take HRT. What other possible reasons are there to explain why HRT use by women with breast cancer might improve survival? Most of the oestrogens used in HRT preparations are conjugated, a form that does not occur naturally in humans. They are termed "impeded oestrogens", as they interfere with the effect of more powerful, naturally occurring oestrogens, such as oestradiol, and their biological effect on breast cancer cells is unclear. In the pharmacological doses used, they are likely to have direct anti-oestrogenic effects and may also downregulate the oestrogen receptor. The progestogens used in combined HRT preparations may also have anti-oestrogenic effects and are weak aromatase inhibitors. This may be relevant in postmenopausal women, as much of the oestrogen present within their breast cancers is produced locally from androgens by aromatase.11 How should women with breast cancer who develop menopausal symptoms be treated? For vaginal dryness, water-based lubricating gels and vaginal moisturisers significantly improve symptoms. If these measures fail, then locally delivered oestrogens are effective.12 For systemic symptoms, such as hot flushes, evening primrose oil, soya and black cohosh are rarely effective, but low-dose megestrol acetate and the antidepressants venlafaxine and fluoxetine were shown to have benefits in randomised trials in breast-cancer survivors.12,13 More recently, isoflavones from red clover were shown to reduce hot flush symptoms in postmenopausal women,14 although there have been no studies in breast-cancer survivors. When these remedies fail, then HRT can be given in the knowledge that current data do not show any detriment in terms of recurrence or survival. Effective agents for osteoporosis in women with breast cancer include bisphosphonates, tamoxifen, raloxifene, diet and exercise.12 Ongoing randomised trials of HRT in breast-cancer survivors will determine whether longer-term HRT is safe. These trials will evaluate whether the increased incidence of breast cancer and reduced sensitivity of mammography in women using HRT2 are important issues in women with breast cancer. Even if these trials show HRT to be safe, the problem in future will be how to treat menopausal symptoms in women taking one of the new aromatase inhibitors, which are already replacing tamoxifen in postmenopausal women with hormone-responsive breast cancer.4 It makes no sense to give these women oestrogen. Ongoing studies are investigating the role of a variety of agents, including tibolone (a synthetic corticosteroid with oestrogenic, androgenic and progestational activity).

J Michael Dixon

Significant perinatal distress and depression affects 14% of women, producing short and long term consequences for the family. This suggests that measures for early detection are important, and non-identification of these women may exacerbate difficulties. Screening provides an opportunity to access large numbers of women and facilitate pathways to best-practice care. A valid, reliable, ...

Anne E Buist MD, FRANZCP · Jeannette Milgrom PhD, FAPS · Bryanne E W Barnett MD, FRANZCP · Sherryl Pope PhD · John T Condon MD, FRANZCP · David A Ellwood MA, DPhil, FRACOG, FRANZCOG · Phillip M Boyce MD, FRANZCP · Marie-Paule V Austin MD, FRANZCP · Barbara A Hayes DNSc, FRCNA

Research papers should have a short section on how the results should be communicated to the public By all accounts, many of the half million Australian women who regularly take combined oestrogen and progestin hormone replacement therapy (HRT) were alarmed by the news on Wednesday, 10 July 2002, reporting that a United States study had shown HRT to increase the risk of breast cancer by 26%, as well as causing more vascular disease. Subsequently, numerous media reports, based on press releases from organisations such as the US National Institutes of Health (NIH)1 and the Cancer Council of New South Wales,2 continued to highlight the apparently large increases in risks caused by combined HRT and called for restrictions on the use of this treatment. General practitioners and cancer help-lines were inundated by enquiries from frightened women and reports of mass withdrawals from therapy soon followed. The source of this concern was the early termination of the NIH-funded Women's Health Initiative (WHI) trial comparing combined HRT and placebo among healthy postmenopausal women. The study was stopped after five years by an independent Safety and Data Monitoring Committee when a predetermined safety boundary for the risk of invasive breast cancer was crossed at an interim analysis. The report of the trial, published in JAMA,3 suggested that women allocated to combined HRT experienced increased risks of invasive breast cancer, coronary heart disease, stroke and venous thromboembolism, and decreased risks of colorectal cancer and hip fracture (Box 1). It was argued that, when all these outcomes were summed in a "global index", the adverse effects outweighed the benefits. However, treatment effects on only two of the outcomes — fractures and venous thromboembolism — met conventional criteria for statistical significance when appropriate (and prespecified) account was taken of the multiplicity of statistical tests performed. Even without adjustment for the dozens of statistical tests, the 95% confidence intervals for each of the other outcomes reportedly affected by combined HRT (including the global index) were consistent with a broad range of possible effects, including little or no effect. For example, any effect on the relative risk of invasive breast cancer appeared to lie somewhere in the range from no effect to an increase of about a half to two-thirds, whereas any effect on the absolute risk of the same outcome appeared to lie somewhere in the range from no excess cases to about 17 extra cases per 10 000 women per year. This very large degree of uncertainty about the true size (and arguably the existence) of most of the treatment effects reported is not reflected in any of the press releases we have seen, including that from JAMA,4 all of which report apparently precise estimates of the excess risks. However, the controversy that followed publicity about the results of the trial did not reflect concerns about the strength of the evidence, but rather dissatisfaction with the way in which the study outcomes were described. While the original report published in JAMA provided estimates of both relative-risk and absolute-risk differences, press releases from most sources focused on the relative increases in risk — in particular, the 26% increase in invasive breast cancer. That this increase in relative risk reflected a difference in incidence of eight cases per 10 000 women per year was much less emphasised. It was suggested by the economics editor of the Sydney Morning Herald that the reported 26% increase was likely to have been misinterpreted by many women as meaning that combined HRT conferred a one-in-four chance of developing invasive breast cancer.5 Others argued that the use of relative risks in press releases was a deliberate effort to dramatise results that would appear much less newsworthy if described in absolute terms. The same commentators suggested that press releases should focus instead on absolute treatment effects, as these are of most direct relevance to the advice provided by doctors and the decisions made by women. Are these criticisms justified? Certainly, there is little doubt that the way in which risk data are presented influences treatment preferences.6-9 In a recent randomised trial in which general practitioners were asked whether they would prescribe a preventive treatment that had negligible side effects, 91% of those given information about relative risks alone said they would do so, compared with 63% of those given information about absolute risks.6 Other studies suggest that the way in which risk information is presented to consumers can generate even greater divergence in preferences.7 Should we therefore abandon the use of relative risks entirely in interpreting the results of clinical trials? Almost certainly not — although a strong case can be made for not allowing relative risks to dominate press releases without appropriate reference to absolute risks. Arguably, each has a place in communications to doctors and patients, and neither should be relied upon exclusively, as both have strengths and weaknesses. For example, while relative risks are usually generalisable to a variety of different patient subgroups (since the proportional effects of treatments are often broadly similar in most major patient subgroups), absolute risks are not (since absolute effects are determined in large part by background disease risks, which can vary substantially). Conversely, relative-risk estimates do not provide sufficient information for assessing the ratio of benefit to harm, as this can only be calculated from estimates of absolute treatment effects. Given the obvious complexity of identifying and delivering the most appropriate message to consumers (whether doctors or patients), medical journals might well consider taking a more substantive role in overseeing the broader dissemination of information about the results of major randomised trials. At a recent seminar ("The HRT debate: how should the new evidence affect policy?") conducted by the Australian Health Policy Institute at the University of Sydney, it was suggested that research papers should have a short section on how the results should be communicated to the public (Sally Crossing, Chair, Breast Cancer Action Group NSW, personal communication). Journals could assume more responsibility in two ways. Firstly, by ensuring compliance with a checklist of essential statistical components to be included in press releases issued by journals (Box 2); and secondly, by publishing a section within the main journal article that summarises the key messages for consumers, with reference to the same checklist. Such a checklist should include requirements for information about absolute as well as relative treatment effects, and for information about the full range of possible effects consistent with the observed result. If journals were to adopt this policy, it would be less likely that consumers would be misled, unintentionally or otherwise, by information released through the press. One can only speculate as to whether providing such information after the termination of the WHI would have altered the subsequent 30% fall in sales of the most commonly prescribed HRT preparations in Australia.10 1: Main results of the Women's Health Initiative trial of oestrogen plus progestin in healthy postmenopausal women3 Outcome Hazard ratio* Adjusted 95% CI† Unadjusted 95% CI Cardiovascular disease 1.22 1.00–1.49 1.09–1.36 Coronary heart disease 1.29 0.85–1.97 1.02–1.63 Stroke 1.41 0.86–2.31 1.07–1.85 Venous thromboembolism 2.11 1.26–3.55 1.58–2.82 Cancer 1.03 0.86–1.22 0.90–1.17 Invasive breast 1.26 0.83–1.92 1.00–1.59 Endometrial 0.83 0.29–2.32 0.47–1.47 Colorectal 0.63 0.32–1.24 0.43–0.92 Fractures 0.76 0.63–0.92 0.69–0.85 Hip 0.66 0.33–1.33 0.45–0.98 Vertebral 0.66 0.32–1.34 0.44–0.98 Deaths from other causes 0.92 0.62–1.35 0.74–1.14 Total deaths 0.98 0.70–1.37 0.82–1.18 Global index‡ 1.15 0.95–1.39 1.03–1.28 * Hazard ratios from Cox regression analyses of outcome among 8506 women randomly allocated to oestrogen plus progestin and 8102 women allocated to placebo. † Adjusted using group sequential methods to correct for multiple analyses over time. ‡ First event for each participant from among the following: coronary heart disease, stroke, pulmonary embolism, breast cancer, endometrial cancer, colorectal cancer, hip fracture, and death from other causes. 2: Essential statistical components for medical journal press releases describing the results of randomised clinical trials A. Provide estimates of absolute treatment effect in addition to estimates of relative treatment effect Estimates of relative treatment effect should not be provided without accompanying information about absolute treatment effect (or, at least, absolute disease rates). If the rates observed in the trial are substantively different from absolute disease rates in major patient subgroups, the limited generalisability of the observed absolute treatment effects should be acknowledged. For example, among perimenopausal women beginning hormone replacement therapy (HRT), whose average age is 10–15 years younger than those recruited to the Women's Health Initiative (WHI), any absolute increase in invasive breast cancer incidence is likely to be less than that observed in WHI, as breast cancer rates are strongly age related. B. Describe the full range of possible effects consistent with the observed result Avoid inappropriate focus on point estimates of either relative or absolute effect when confidence intervals indicate a broad range of potential effects. For example, the WHI result for invasive breast cancer risk was reported in press releases as a 26% increase in relative risk (and, occasionally, as an absolute excess of 8 cases per 10 000 women per year). However, the observed result is consistent with no increase in risk, as well as with an increase in relative risk of half to two-thirds and an increase in absolute risk of up to about 17 cases per 10 000 women per year (based on unadjusted 95% confidence intervals).

Anushka Patel MB BS, MS, FRACP · Robyn Norton PhD, MPH · Stephen MacMahon PhD, FACC, FAHA

Objective: To determine whether hormone replacement therapy (HRT) after treatment for breast cancer is associated with increased risk of recurrence and mortality.Design: Retrospective observational study.Participants and setting: Postmenopausal women diagnosed with breast cancer and treated by five Sydney doctors between 1964 and 1999.Outcome measures: Times from diagnosis to cancer recurrence or new breast cancer, to death from all causes and to death from primary tumour were compared between women who used HRT for menopausal symptoms after diagnosis and those who did not. Relative risks (RRs) were determined from Cox regression analyses, adjusted for patient and tumour characteristics.Results: 1122 women were followed up for 0–36 years (median, 6.08 years); 154 were lost to follow-up. 286 women used HRT for menopausal symptoms for up to 26 years (median, 1.75 years). Compared with non-users, HRT users had reduced risk of cancer recurrence (adjusted relative risk [RR], 0.62; 95% CI, 0.43–0.87), all-cause mortality (RR, 0.34; 95% CI, 0.19–0.59) and death from primary tumour (RR, 0.40; 95% CI, 0.22–0.72). Continuous combined HRT was associated with a reduced risk of death from primary tumour (RR, 0.32; 95% CI, 0.12–0.88) and all-cause mortality (RR, 0.27; 95% CI, 0.10–0.73).Conclusion: HRT use for menopausal symptoms by women treated for primary invasive breast cancer is not associated with an increased risk of breast cancer recurrence or shortened life expectancy.

Eva M Durna MBioeth · Leo R Leader MD, FRANZCOG · Peter Sjoblom PhD · John A Eden MD, FRANZCOG · Barry G Wren MD, FRANZCOG · Gillian Z Heller PhD

Clinical question"What treatments are available for hot flushes in women receiving tamoxifen?" A 62-year-old woman asked her radiation oncologist this question. She was taking tamoxifen as adjuvant treatment for node-positive breast cancer, but was experiencing persistent and frequent hot flushes. Search questionThe search question was refined to "What treatments can be added to tamoxifen to reduce the frequency or severity of hot flushes? What are the benefits and risks?" The ideal study to answer these questions is a randomised controlled trial that compares various treatments in women taking adjuvant tamoxifen for breast cancer and prospectively assesses changes in flushing. SearchWe used a comprehensive strategy to search electronic databases, including MEDLINE, the Cochrane Library and SUMSearch <http://sumsearch.uthscsa.edu/searchform45.htm>. The search terms "hot flashes" / "hot flushes" and "tamoxifen" were combined to identify the relevant trials. Summary of findingsSeven agents have been tested in randomised, placebo-controlled trials. Appropriate randomisation procedures included stratification for tamoxifen use where applicable. Sample sizes ranged from 85 to 194 women, and the duration of baseline and evaluation periods ranged from 4 to 7 days and 28 to 84 days, respectively. Concurrent tamoxifen was an eligibility requirement in two studies, but otherwise between 59% and 81% of women were taking tamoxifen. Each study used frequency of hot flushes, as well as "activity scores" (which incorporate frequency and severity of flush episodes), to evaluate the medications. These were assessed using daily patient diaries, with a similar format in each study. Megestrol acetate (40 mg/day),1 venlafaxine (37.5–150 mg/day),2 transdermal clonidine (at a dose eqivalent to 0.1 mg/day orally)3 and oral clonidine (0.1 mg/day)4 were all significantly more effective than placebo at reducing the frequency of flushes after four weeks (P < 0.05). They resulted in reductions in the median number of flushes by 73%, 30%–58%, 44%, and 34% from baseline levels, respectively (ie, 4.5–2.7 fewer flushes daily from baselines of 6.1–8.0). The activity scores showed greater percentage reductions. Oral clonidine was also effective at eight weeks, but long-term effectiveness was not examined in any study. The three other agents examined — soy phytoestrogens,5 vitamin E6 and a "herbal remedy" black cohosh (Cimicifuga sp.)7 — were found not to be useful. Hormone replacement therapy is an established treatment for postmenopausal flushing. However, no randomised trials assessing its value in patients receiving tamoxifen for breast cancer were identified. Its safety in women with a history of breast cancer is controversial and it cannot be routinely recommended.8 Adverse reactions greater than those with placebo1-4 were, for megestrol acetate — withdrawal menstrual bleeding (31%); for venlafaxine — dry mouth, anorexia, nausea, and constipation; for oral clonidine — difficulty sleeping (41%); and for transdermal clonidine — itchiness under the patch, drowsiness, dry mouth and constipation. Each side effect may be dose-dependent. Each medication also has a number of recognised contraindications and precautions. OutcomeEach of the medications assessed in randomised trials is a reasonable option for treatment. However, venlafaxine's sole indication on the Pharmaceutical Benefits Scheme is major depression, while transdermal clonidine is not available in Australia. The radiation oncologist discussed the available choices with his patient. She declined venlafaxine on the basis of cost, and both megestrol acetate and clonidine on the basis of potential side effects, but continued to take tamoxifen.

Sean A Bydder MB ChB, FRANZCR · Nigel A Spry MB BS, FRANZCR

With our high standards of sanitation and healthcare, how important is breastfeeding? Any breastfeeding advocate can rattle off a list of the advantages of breastfeeding to infants, their mothers and society. The list of benefits ranges from better emotional attachment to a lower risk of childhood leukaemia or dental malocclusion. However, claims of benefit have been contentious, with many studies failing to demonstrate a clear advantage and some even showing increased risk of a negative health outcome with breastfeeding. The constituents of human milk suggest that it has evolved to promote infant health and human growth, particularly to protect children from infections and to support the rapid growth of the human brain. Despite advances in science and manufacturing, infant formula will always be a poor copy of human milk. It is improbable that a substitute will ever reproduce the full spectrum of human milk proteins, milk sugars (numbering over 100), live white cells and antibodies programmed by infections in the infant's environment, together with constant variations in milk content to meet the needs of the growing infant. But does this matter in a developed country such as Australia? Is there justification in the argument that women are being pushed too hard to breastfeed, or is there more to infant nutrition than the provision of clean water and biologically safe artificial feeds? Many studies reporting the effects of breastfeeding, both positive and negative, have major methodological flaws. Indeed, robust studies are difficult to carry out. Random allocation of study participants to either a "breastfeeding" or an "artificial feeding" group is unethical, especially if we accept the belief that "the epidemiologic evidence is now overwhelming that, even in developed countries, breastfeeding protects against gastrointestinal and (to a lesser extent) respiratory infection, and that the protective effect is enhanced with greater duration and exclusivity of breastfeeding".1 So evidence for the effects of breastfeeding must rely largely on observational cohort and case–control studies. Blinding to the intervention group is difficult, except at the analysis level. Moreover, mothers who elect to breastfeed differ from those who don't — the former group are generally better-educated women of higher socioeconomic status, who are more likely to have partners and less likely to smoke. All of these factors are likely to be independently associated with positive health outcomes, so failure to adequately adjust for these confounders generally favours positive breastfeeding outcomes. On the other hand, apparently poorer outcomes may result if mothers preferentially breastfeed more vulnerable infants. Difficulties with defining terms such as "breastfeeding" and "exclusively breastfed", and with defining endpoints such as "atopic disease", also make comparisons between studies difficult. So, have the roughly 2000 papers on breastfeeding and human milk listed in Medline since the year 2000 helped answer the question of whether or not it matters if a term infant born in a country like Australia is breastfed? I will confine my comments to the three most common health problems in Australian children: infection, obesity, and asthma. In Australia, there are almost 20 000 hospital admissions a year for acute gastroenteritis in children under five years old. Rotavirus accounts for about half of these episodes and is also one of the most important nosocomial infections in paediatrics. A Belarussian study2 (involving 17 000 healthy mother–infant pairs intending to breastfeed) showed that, in centres randomly assigned to deliver support for breastfeeding (as outlined by the Baby Friendly Hospital Initiative3), there were increases in exclusive and continued breastfeeding and reductions in episodes of gastrointestinal infection. In an Italian study4 of infants aged 1–18 months admitted to an infant ward, fewer breastfed infants contracted rotavirus infection (10.6% v 32.4%), and none of these became symptomatic. It has been shown that human milk lactadherin prevents symptoms in breastfed infants infected with rotavirus by binding to the virus and inhibiting its replication.5 Obesity is the commonest chronic health problem in Australian children. This has health implications both in and beyond childhood, as about half of obese children become obese adults. The effect of breastfeeding on later overweight/obesity remains contentious. A 2001 review of 18 studies6 concluded that most studies examining the effects of breastfeeding on later obesity had found an insignificant effect, although two of the studies actually found a positive association between breastfeeding and later body fatness. Since that review, four large studies7-10 (involving between 2000 and 14 000 children) have shown a lower prevalence of overweight in previously breastfed children, with several showing decreasing risk with longer duration of breastfeeding. The three studies of older children (aged 5–14 years)8-10 also found a negative association between breastfeeding and obesity (with adjusted odds ratios [ORs] ranging from 0.66 to 0.8). Similar results in these three studies, despite differences in method, suggest that the finding may be robust. The second most common chronic health problem in Australian children is asthma. A large Western Australian study showed a substantial reduction in risk of childhood asthma, as assessed at six years of age, if exclusive breastfeeding is continued for at least the first four months of life.11 This was included in a systematic review with meta-analysis of 12 prospective studies,12 which gave a summary OR of 0.70 (95% CI, 0.60–0.81) for asthma among children who had been breastfed, with a greater effect in children from families with a history of atopy (OR, 0.52; 95% CI, 0.35–0.79). The studies were mainly of younger children but included children up to 8.4 years. Another prospective study of more than 330 000 children followed to age 24 months supported these findings: breastfeeding of less than nine months' duration was associated with an increased risk of asthma or wheezing, and a dose–response effect was observed with breastfeeding duration.13 However, a recent report has suggested that atopic children with asthmatic mothers are more likely to develop asthma in later childhood if they have been exclusively breastfed.14 There may be interesting possibilities for intervening to further reduce atopy in at-risk families. Mothers of infants who develop atopic sensitisation have a higher intake of saturated fat,15 and their breast milk is much lower in some long-chain unsaturated fatty acids (which may have antiallergenic properties16). Perhaps improving maternal diets and the use of probiotics17 may reduce the prevalence of atopic disease in their children. I have selected three childhood conditions for which small changes in prevalence or disease severity would have a major impact on the health of the nation. The US Department of Agriculture, Food and Nutrition chose otitis media, gastroenteritis and necrotising enterocolitis, and estimated that the United States would save a minimum of US$3.6 billion dollars a year if breastfeeding rates increased by 10% at initiation and by 20% at the age of six months.18 While we should be careful not to convey to mothers that breastfeeding will make their child a trim, non-atopic, infection-free genius, the balance of the evidence is that breastfeeding is of benefit in many ways. Every healthcare professional in contact with young children and parents should be familiar with their responsibilities under the World Health Organization Code19 to encourage and promote breastfeeding — for the health of the nation.

Patricia McVeagh MB ChB FRACP

Improving health outcomes for women with this enigmatic disease Endometriosis remains an enigma despite having been extensively studied. Its aetiology remains unclear, although there is a positive correlation with retrograde menstruation, probably with a background genetic predisposition. Traditionally, the symptoms of endometriosis are said to be pelvic pain and infertility, but, while a causal link between these symptoms and the disease is sometimes clear, the link is not always established. Treatment aimed at eliminating endometriotic deposits is effective in controlling symptoms in a proportion of patients, but recurrence is common. The diagnostic dilemmaEndometriosis is considered difficult to diagnose because the extent of the disease (endometriotic deposits and scar tissue) does not correlate with symptoms. The only accepted diagnostic test is direct visualisation (usually laparoscopically) by an experienced surgeon.1 Arguably, as endometriosis can be diagnosed in up to 40% of menstruating women at laparoscopy,2 it may be detected in all women at some stage of their reproductive lives, only to regress spontaneously in some. In addition, some symptoms associated with endometriosis are common. For instance, dysmenorrhoea occurs in up to 60% of women with regular menses3 and infertility occurs in about 15% of the population (of whom 30% have endometriosis).4 Hence, common conditions (mild endometriosis) may be associated with common symptoms (eg, dysmenorrhoea) by chance, not causally. The enormous variation in symptomatic expression and lack of clear correlation with the extent of endometriosis suggest the need for epidemiological studies of each symptom and of women with either mild or severe disease. This would exclude the possibility that mild endometriosis may be a normal finding in women. Epidemiological studies of women with endometriosis associated with dysmenorrhoea may find a causal link only in women with progressive dysmenorrhoea. In addition, epidemiological studies confined to specific symptoms causally associated with endometriosis (eg, dyspareunia and endometriosis in the uterosacral ligaments, progressive dysmenorrhoea and American Fertility Society stage III and IV endometriosis5) may result in consistent findings of the role of genes in disease development. 6 Management issuesFrom observations that pregnancy resolves symptoms, particularly of dysmenorrhoea, "mimicking pregnancy" became a treatment option. Medications that resulted in amenorrhoea (such as GnRH analogues, danazol and high-dose progestogens) have been used. These work by antagonising the growth-promoting effect of oestrogens or reducing the oestrogenic stimulus for growth, as well as increasing apo-ptosis. Each of these medications has been shown in a recent Cochrane meta-analysis of randomised controlled trials to be equally effective in control of dysmenorrhoea, dyspareunia and pelvic pain, and in decreasing the bulk of endometriotic tissue.7 However, the follow-up period in these trials was usually only three to six months. Annual recurrence rates of endometriosis after treatment are generally accepted to be of the order of 10%.8 Hence, further trials with longer follow-up, and possibly comparisons with the oral contraceptive pill (which has a better side-effect profile than the other drugs), in three-monthly cycles, would be beneficial. That medical therapies are not effective for infertility associated with mild endometriosis has been well established in large-scale, randomised controlled trials.9 Another Cochrane review showed that laparoscopic ablation or excision of the endometriotic tissue is effective in treating endometriosis associated with pelvic pain. However, this conclusion was based on one randomised controlled trial10 and further trials are needed to confirm this finding. Similarly, further evidence to support the effectiveness of laparoscopic surgery for subfertility associated with endometriosis is needed to support the beneficial finding in one randomised controlled trial. This is the aim of a protocol developed by the Cochrane Library.11 While the above conclusions are based on Cochrane reviews of the best available evidence, doubt remains. All trials are subject to a chance of showing a real effect when there is a probability that there is no effect. Continued development of well conducted randomised trials will reduce the probability of an incorrect conclusion. In addition, expert opinion based on the current extensive body of literature, and taking into account the basic mechanisms of the disease process of endometriosis, will help distinguish the causal relationship between each symptom associated with endometriosis. In turn this would lead to improved clinical outcomes for patients with associated dysmenorrhoea, chronic pelvic pain, dyspareunia and infertility. It is argued that current evidence for effective treatment of each of the symptoms associated with endometriosis could be further improved by further randomised controlled trials with clearly defined patient groups (eg, chronic pelvic pain in association with mild endometriosis, in-vitro fertilisation for three to five years of infertility associated with mild endometriosis). In addition, if results of epidemiological studies show no probable causal link between endometriosis and dysmenorrhoea then therapeutic regimens targeted at both symptom relief and monitoring progression of the disease may be more efficacious than excision of endometriotic tissue alone. While many of our current treatment regimens are effective in some patients, continued improvement in long term health outcomes for women with endometriosis requires a mindset open to critical evaluation.

Kevin L Forbes MB BS, FRACOG