Editor’s choice


Improving the prediction of premature delivery

Around 50% of women presenting with symptoms of preterm labour do not progress to preterm delivery. Inability to predict which women progress to delivery causes unnecessary admission and use of antenatal steroids. Multiple courses of the latter may have adverse effects on fetal development (Murphy et al. Lancet 2008;372:2143–51; Wapner et al. N Eng J Med 2007;357:1190–8; Crowther et al. Cochrane Database Syst Rev 2007 [3]). Two years ago we published evidence confirming the difficulty of predicting preterm birth in women with threatened preterm labour using history and clinical examination alone (Mahony et al. BJOG 2010;117:963–7). The most common approaches to improving the prediction of delivery include the use of fetal fibronectin or transvaginal ultrasound estimation of cervical length. However, it is likely that inflammation is an important factor in the mechanism of spontaneous preterm labour and that this is associated with increased production of cytokines. In the study on page 866, Tsiartas et al. examine the ability of 27 maternal serum proteins alone and in combination with cervical length to predict delivery within 7 days in 142 women with singleton pregnancies who had threatened preterm labour. They assessed the performance of multiple models of serum proteins alone, cervical length alone and combinations of serum proteins and cervical length. The best cutoff values for each protein and the cervical length were determined by receiver operated curve analysis, including the area under the curve. In a combined multivariable model of maternal serum proteins and cervical length, high levels of interleukin-10, RANTES (CCL5—chemokine [C-C 166 motif] ligand 5) and short cervical length (≤18 mm) made statistically significant contributions to the prediction of spontaneous preterm delivery within 7 days. The receiver operated curve for this combination of tests showed an area under the curve of 0.88, sensitivity of 74%, specificity of 87% and positive and negative predictive values of 76 and 86%, respectively. In contrast, cervical length alone had corresponding values of 0.77, 93, 62, 57 and 94%. Of course, this model has been derived post hoc from the data so before it can be concluded that this collection of tests may be useful to clinicians it needs to be applied to a new cohort of high-risk women to confirm its predictive ability. However, this study suggests that a combination of maternal serum inflammatory proteins and cervical length may be useful in predicting preterm delivery.

Recurrence of the hypertensive disorders of pregnancy

Hypertensive disorders of pregnancy may have a significant psychological impact on women and their partners, and many parents avoid or delay a future pregnancy because of their experiences. Post-traumatic stress disorder has the crucial diagnostic criterion ‘Persistent avoidance of stimuli associated with the trauma’ (American Psychiatric Association, 2000. Diagnostic and statistical manual of mental disorders, revised 4th edn) and has been reported in 11% of women 15 months after a pregnancy complicated by pre-eclampsia (Stramrood et al. Reprod Sci 2011;18:645–53). The post-traumatic stress disorder symptom of avoidance has been noted in 8% of such women 12 weeks after delivery (Hoedjes et al. J Psychosom Obstet Gynaecol 2011;32:126–34). It is becoming increasingly clear that the hypertensive disorders of pregnancy are heterogeneous, with differing times of onset, pathophysiology and maternal and fetal effects. Knowledge of the risk of recurrence of each type of hypertensive disorder should be helpful in counselling women about the risks of future pregnancies, in formulating plans of care and improving the outcome of a future pregnancy. In a retrospective cohort study van Oostwaard et al. (page 840) report a study of birth data over a 3-year period from six hospitals in the Netherlands, identifying women with a hypertensive disorder in the index pregnancy who required delivery between 34 and 37 weeks. They were able to contact 351 (of 425) women who matched the criteria, of whom around half had a subsequent pregnancy. Hypertensive disorders recurred in half of the women, of whom about 10% delivered again before 37 weeks. The most novel aspect of this study was the examination of the recurrence of individual hypertensive syndromes, categorised as; small for gestational age, gestational hypertension, pre-eclampsia and HELLP (haemolysis, elevated liver enzymes and low platelet count) syndrome. It appeared that each syndrome was most likely to be the recurrent syndrome in the next pregnancy. For women who delivered a small-for-gestational-age baby in the index pregnancy the risk of the same happening in the next pregnancy was 31%, a complication that was more common than other syndromes (for example, gestational hypertension complicated 28%, pre-eclampsia 9% and HELLP 4%). The recurrence risk of gestational hypertension was 44%, also more common than other syndromes. However, for index cases of pre-eclampsia the risk of recurrence was 19%, whereas 30% had gestational hypertension, and for those women who had HELLP syndrome in the index pregnancy the risk of HELLP recurrence was 10% and the risk of pre-eclampsia was 21%.

Preventing cerebral palsy

It is well established that in developed countries only around 15% of cases of cerebral palsy are clearly attributable to events during labour, but that these events are often avoidable. Even within the setting of a clinical trial, errors in the management of fetal monitoring occur. For example, in the Dublin trial of continuous versus intermittent fetal monitoring (MacDonald et al. Am J Obstet Gynecol 1985;152:524–39) approximately one-third of the neonatal deaths or intrapartum stillbirths were associated with a failure to follow either the monitoring policy or the protocols for the management of abnormal monitoring findings. The denominator in this study, the number of women who had normal babies when protocols were significantly violated, is unknown. However anyone who has spent enough time on a labour ward or in the preparations for a CNST (UK Clinical Negligence Scheme for [Hospital] Trusts) assessment visit knows that deviations from fetal monitoring protocols are common. Much time is spent in training staff to follow monitoring protocols, but could some of this time be better spent in training staff to understand other important factors that lead to adverse neonatal outcomes?

Criteria for ascribing cerebral palsy to intrapartum hypoxia include ‘Evidence of a metabolic acidosis in intrapartum fetal, umbilical arterial cord, or very early neonatal blood samples (pH <7.00 and base deficit >12 mmol/l)’ and ‘Early onset of severe or moderate neonatal encephalopathy in infants of >34 weeks of gestation’ (MacLennan BMJ 1999;319:1054–9). Whether the recommendation concerning pH is appropriate is the subject of the study by Yeh et al. from Oxford, UK (page 824) who in an observational cohort study collected 53 855 correctly paired samples of neonatal umbilical blood from singleton non-anomalous live births at or over 37 completed weeks of gestation and examined the neonatal outcome (seizures with evidence of encephalopathy) in relation to the pH values. The sampling rate was around 50% of all deliveries. This cohort is larger than any previous reported study.

The results are intriguing. Although a pH <7.00 was associated with 20–24% of adverse neurological outcomes (with a further 10–15% occurring with a pH <7.11), for babies that developed seizures within the first 24 hours at least half had a pH of >7.10 and nearly 40% had a pH of >7.2. The results may have been biased by the well-reported phenomenon of a greater chance of cord sampling in women for whom an adverse outcome was suspected at delivery: seizures occurred in 2.2/1000 neonates where samples had been taken but in only 1.3/1000 where they had not. Nevertheless, the high number of babies with evidence of encephalopathy with ‘normal’ values for cord pH is likely to be genuine and may be a manifestation of the ‘acidosis paradox’ (Hermansen Dev Med & Child Neurol 2003;45:353–6), in which neonates are born with encephalopathy secondary to catastrophic events in labour, without time for the fetus to develop acidaemia, or because of the effects of adverse intrauterine environmental factors such chorioamnionitis.

Neonatal seizures are clearly associated with an increased risk of cerebral palsy, but as was demonstrated in the follow-up data from the Dublin trial, cerebral palsy is not inevitable. Yeh et al. plan to report the results of longer-term follow up. In the meantime, the authors conclude that ‘intrapartum fetal surveillance that relies almost entirely on detection of acidaemia, and follows National Institute for Health and Clinical Excellence (NICE) Guidelines for intervention, will lead to both a failure to prevent most adverse neurological outcomes and a high obstetric intervention rate’.

The finish of fertility

A longitudinal study of bone density spanning over 30 years using the same methods at the start and finish is a rare phenomenon. On page 810, Svejme et al. report the results of a prospective study comparing the effects of early (<47 years) and normally timed menopause on bone density, fracture risk and mortality with a follow-up period of more than three decades. Nearly 400 women aged 48 from Malmö Sweden were recruited in 1977 to this prospective observational study and underwent single-photon absorptiometry (SPA) of the distal forearm. At age 77, all eligible participants were re-invited for bone mineral density measurements. By this time nearly a quarter of the women had died and a quarter had relocated or declined further participation, leaving around 200 women to attend for follow-up SPA measurements. These were performed using the same apparatus, the same anatomical site, and the same technician analysed all SPA data. Measurements of a standardised phantom throughout the study period did not reveal any change in accuracy of the densitometer. Osteoporosis was defined according to World Health Organization criteria. The authors also identified fractures that occurred in all the women by repeated searches of hospital databases, including those women who declined follow-up bone mineral density measurements. The collection of these data was made easier because all patients in the Malmö region attend the same trauma unit. At the final visit, 56% of women with early menopause had osteoporosis, compared with 30% of women with late menopause. During the 30 years of follow up one-third of the women had sustained at least one fragility fracture. In the early menopause group, the fracture incidence per 1000 person-years was 19.45 compared with 11.60 in the late menopause group, giving a risk ratio of 1.68 (95% CI 1.05–2.57).

Perhaps the most interesting data from this study concern mortality. The mortality rate was 52% in the early menopause group and 35% in the late menopause group, giving a relative mortality risk of 1.59 (95% CI 1.04–2.36). Although a higher fracture incidence in women with an early menopause could be associated with increased fracture-related mortality this effect is likely to be unimportant compared with differences in mortality risk caused by cardiovascular disease, which is known to be increased by early menopause (Rivera et al. Menopause 2009;16:15–23, Løkkegaard et al. Maturitas 2006;53:226–33, Parker et al. Obstet Gynecol 2009;113:1027–37). It still remains unclear whether the increased risk of mortality due to cardiovascular disease with early menopause is related to changes in lifestyle factors that influence cardiovascular health, or if early menopause leads to organ dysfunction via a hormonal mechanism, or whether premature menopause is the result of an accelerated aging process throughout the body determined by genetic or non-genetic causes that involves all organs, including the ovaries.

Finally, on page 795 we publish a report of the largest single-centre study to date of outpatient hysteroscopic sterilisation, involving 4000 women with a median follow up of 3–4 years. This study indicates that the technique is safe, feasible and effective in the short and medium term. I have little to add to the excellent accompanying commentary by T. Justin Clark on page 775, which summarises the current issues surrounding the common methods of sterilisation and I would urge readers to read this commentary alongside the paper.

BJOG on Twitter

I have recently taken on the role of the Editor in charge of the BJOG Twitter site (@BJOGTweets), launched last year. As a novice to the world of tweeting who cannot obtain any clear advice from others involved in medical publication on the role of Twitter in the dissemination of medical information, I would be very interested in the opinion of our Twitter-savvy readership about what you would like from your @BJOGTweets. I can be tweeted directly on @MikeMarshFRCOG.