Lavender, Tina PhD, RM; Alfirevic, Zarko MD, FRCOG; Walkinshaw, Stephen MD, FRCOG
The partogram (or partograph) is considered a valuable tool in the improvement of maternity care by allowing midwives and obstetricians to record intrapartum details pictorially; belief that its use was applicable in developed and developing settings led to its introduction worldwide.1
A number of common partogram designs follow the work of Philpott and Castle,2 and most incorporate an action line. An action line allows unambiguous diagnosis of prolonged labor enabling the timing of intervention to be based on the rate of cervical dilatation. It is conventionally placed a number of hours to the right of another line, the alert line,3 which was developed from the rate of cervical dilatation of the slowest 10% of primigravidae in Zimbabwe (then Rhodesia) in the early 1970s,4 to aid the timely transfer of women from peripheral to tertiary units. More than 20 years after its introduction, and using a partogram adapted from that formulated by Philpott and Castle,2,3 the World Health Organization (WHO)5 conducted a prospective study of 35,484 women in Southeast Asia. They concluded that the partogram was a necessary tool in the management of labor and subsequently recommended its universal application. Like Philpott and Castle,2 who were the first to provide specific guidelines on the timing of intervention for prolonged labor, the WHO partogram contained an action line 4 hours to the right of the alert line.
In contrast to the WHO recommendations, the Dublin group6,7 proposed an active management package, which relied on early identification of prolonged labor with early correction by oxytocin. Interventions were triggered if cervical progress deviated more than an hour from the alert or normal progress line. Some randomized studies8 demonstrated a reduction in cesarean delivery rate using an active management package. However, despite inclusion of all the components of the National Maternity Hospital protocol for active management of labor, the largest randomized study of 1,934 nulliparous women9 failed to provide evidence that such a protocol reduces the cesarean delivery rate. Other studies of various forms of early intervention have shown reductions in duration of labor but not in cesarean delivery outcome.10 Few studies have considered maternal satisfaction as an important outcome.
Recently, practitioners in developed countries have begun to question the benefits of the partogram and its individual components.11,12 The timing of intervention is a particularly contentious issue because it is the trigger for interventions such as amniotomy, intravenous hydration, analgesia, oxytocic infusion, and operative delivery. The timing of such interventions has not been subjected to rigorous evaluation. Cartmill and Thornton13 hypothesized that the way a partogram is presented may affect an obstetrician’s perception of the labor progress and thus influence decision-making. This hypothesis has received some support from others,14 who suggested that the slope and position of the action line may have an effect on cesarean delivery, intervention, and maternal satisfaction. Evidence, however, in the form of randomized controlled trials is limited.
Two trials that have added to the debate15,16 had conflicting findings. Pattinson et al,15 in his study of 694 laboring nulliparous women in South Africa, found that women who received early intervention (2 hours after no progress), had fewer cesarean deliveries (16%) than those with expectant management (23.4%) (intervention after 4 hours without progress). In our earlier randomized controlled trial16 (n = 928) comparing partograms with action lines either 2, 3, or 4 hours to the right of the alert line, the cesarean delivery rate was lowest when using the 4-hour line (8.3%) compared with the 2-hour arm rate (11.1%), but this difference did not reach statistical significance. However, women in the 2-hour arm were most satisfied with their labor experience. The sample size was too small to show any significant differences in other important perinatal outcomes. Meta-analysis of these two trials showed no difference in cesarean delivery rate between the use of 2- or 4-hour action lines (relative risk [RR] 0.93, 95% confidence interval [CI] 0.48–1.78).17
Given the potential importance of action line positioning on clinical and emotional outcomes and the inconclusiveness of the available evidence, we carried out a large randomized trial to assess the effect of using a 2- or 4-hour partogram on cesarean delivery and maternal satisfaction.
MATERIALS AND METHODS
This trial was conducted in a regional, inner city, teaching hospital in the Northwest of England between August 1998 and March 2005, with 6,000 births per annum. Intrapartum care is provided in two units within the same building.
The midwife-led unit only admits women with uncomplicated pregnancies who present in spontaneous labor. The unit prides itself on a “home-like” environment, where furnishings are inviting and emergency equipment is concealed. The approach is one that actively promotes normal birth with minimal invasive intervention. If a woman requires epidural analgesia or a labor complication develops, she is transferred to the delivery suite (the other birthing unit).
The delivery suite is a consultant led unit that admits women with complicated pregnancies and those who choose to give birth there, for example if they wish epidural analgesia. Consequently, a medical approach is more prevalent and the environment is more clinical. For low-risk women choosing to give birth in this setting, midwives remain the primary caregivers.
Written information was given to all potentially eligible women in the antenatal period. Women were eligible for the trial if they were primigravid, in spontaneous labor, at term (37 weeks or more of gestation) with a live, singleton, cephalic presentation. Women were excluded if they had significant medical disease, had pregnancies complicated by fetal malformations, or required high-dependency intrapartum care. The Local Research Ethics Committee approved the study, and women gave written informed consent before enrolment.
On admission in spontaneous labor, consenting women were randomly allocated to one of two trial arms—a partogram with 2-hour or 4-hour action line (Fig. 1), using consecutively numbered sealed opaque envelopes. The partogram was contained inside the randomization envelope to make the process efficient and to ensure that the women received their allocated management. The randomization sequence was generated using a table of random numbers, stratified according to intended place of birth (ie, Midwife-led unit or Delivery suite). Given the type of intervention, blinding of clinicians or women to the allocation was not possible.
Women who declined participation had labor managed with the aid of a 2-hour partogram, because this had been preferred by women in our first trial.16 Eligible women were randomly assigned to the trial, by the attending midwife, once established labor had been confirmed by vaginal examination. Labor was confirmed if 1) the cervix was effaced, 2) the cervix was at least 3 cm dilated, and 3) regular uterine contractions at least every 5 minutes lasting a minimum of 20 seconds were present.
The management of randomly assigned women was unaffected if labor followed the expected rate of progress. However, if cervical dilatation crossed the allocated action line, then a clinical assessment was made and delivery suite guidelines for the management of prolonged labor were followed. If the woman was laboring on the midwife-led unit, she would be transferred to delivery suite. Where augmentation was required, this involved oxytocin alone when membranes were ruptured or amniotomy followed by oxytocin in the presence of intact membranes. The oxytocin infusion rate commenced at 2 milliunits/min and was doubled every 30 minutes until effective regular uterine contractions were achieved, the maximum rate of oxytocin being 32 milliunits/min. Women with oxytocin infusion or with epidural analgesia had continuous external fetal monitoring. After initiation of augmentation, clinical care was at the discretion of the on-call clinical team. Cardiotocograph changes denoting fetal distress were confirmed by fetal scalp blood sampling following the Royal College of Obstetricians and Gynaecologists guidance on electronic fetal monitoring.18
Primary outcome measures were cesarean delivery rate and maternal satisfaction. Secondary outcomes were need for augmentation, duration of labor, use of analgesia, postpartum hemorrhage, number of vaginal examinations, admission to special care baby unit and Apgar score.
Demographic and intrapartum data were extracted from hospital case records. Data sheets were completed for all randomly assigned women before putting the information into a database. Satisfaction data were obtained by a postal questionnaire, which included a stamped addressed envelope. A previously validated short questionnaire,19 presented in the form of a rating scale followed by an open question, was administered to all women in the immediate postnatal period (2–10 days). One reminder letter (incorporating a copy of the questionnaire) was sent to those who initially failed to reply, to encourage a greater response rate.
Sample size calculations were based on data from the previous randomized trial in the same unit.16 We anticipated an overall cesarean delivery rate of 10% and prevalence of women who were unsatisfied with their child bearing experience (score less than 15) of 14%. A 3% difference in both these outcomes between two groups was believed to be clinically significant by the trial investigators, who, following discussion with obstetricians and midwives from the study hospital, agreed that this difference would prompt a change in practice. A sample size of 3,000 women would have 90% power to detect such differences in both primary outcomes (alpha 0.05).
An interim analysis was carried out after the recruitment of 1,500 women by the Data Monitoring Committee; advice was given to continue recruitment until the full sample size was reached. Before analysis, the data were double entered to maximize the accuracy of the information collected. Statistical analyses were carried out using SPSS 12.01 (SPSS Inc., Chicago, IL) according to the intention-to-treat principle.
The groups were compared using the Fisher exact test for the analysis of binary outcomes and unpaired t test for normally distributed (Shapiro-Wilk W test) continuous data. A two-tailed P value of less than .05 was considered indicative of statistical significance. Relative risks and 95% confidence intervals were also calculated.
A comparison of outcomes related to place of labor admission (ie, midwife-led unit or delivery suite) was the only planned subgroup analysis. The possible differences in the size and direction of intervention effects were compared using χ2.
The study took place between August 1998 and March 2005 in a unit with 42,748 births during this period. Of these total births, 19,012 were to primigravid women. Out of 9,990 eligible women at term, 2,024 declined participation. Of the 1,324 who provided a reason, 861 said they preferred a 2-hour action line, 240 said they were scared of “being left too long,” 167 said they wanted the midwife to decide the best management, and 56 did not want to take part in research. Four thousand nine hundred sixty-six women were never approached to participate, mainly because informed discussion about participation was deemed inappropriate by the midwife. This left a total of 3,000 randomly assigned women. However, 25 randomly assigned women could not be traced due to the inaccurate recording of demographic details; data were therefore collected on a total of 2,975 women (Fig. 2). Questionnaires were completed by 1,929 (65%) women.
The demographic details and cervical state at randomization is given in Table 1, showing, as expected, groups were comparable with respect to maternal age, gestational age, cervical dilatation, cervical effacement, and status of membranes.
There were no differences in the overall cesarean delivery rate between the 2-hour and 4-hour arms (136/1,490 compared with 135/1,485; RR 1, 95% CI 0.80–1.26; P=1) or the number of women who were dissatisfied with labor experience (72/962 compared with 81/967; RR 0.89, 95% CI 0.66–1.21; P=.5) (Table 2).
The intrapartum details (Table 3) show that when compared with the 4-hour arm, more women in the 2-hour arm crossed the partogram action line (854/1,490 compared with 673/1,485; RR 1.27, 95% CI 1.18–1.37; P<.001) and therefore received more interventions to augment labor (772/1490 compared with 624/1485; RR 1.23, 95% CI 1.14–1.33; P<.001). Women in the 2-hour arm had shorter labors than those in the 4-hour arm (mean difference 26.9 minutes, t=−2.66, df 2,973, P=.008). No further differences were found in secondary outcomes between the groups and none of the limits of the confidence intervals were of clinical importance.
When the 2- and 4-hour groups were compared by intended place of birth (Table 2) no differences were detected in either cesarean delivery rate (68/674 compared with 52/666; RR 1.29, 95% CI 0.92–1.82) or number of women dissatisfied with labor experience (31/674 compared with 38/666; RR 0.81, 95% CI 0.51–1.27). The size and the direction of effects were consistent in both settings with the exception of cesarean delivery and epidural rates; women on the midwife-led unit assigned to the 2-hour arm had disproportionally more epidurals (16.5% compared with 12.9%) and cesarean deliveries (10.1% compared with 7.8%) than those on the midwife-led unit in the 4-hour arm (Table 2).
If admitted to the midwife-led unit, women in the 2-hour arm were more likely to be transferred for consultant-led care than those in the 4-hour arm (366/674 [54.3%] compared with 285/666 [42.8%]; RR 1.26, 95% CI 1.13–1.42; P<.001).
Although the partogram is in widespread use, little research has been undertaken in the form of randomized control trials to assess its efficacy. Given that prolonged labor is the major contributor to cesarean delivery,20 knowing when and how to intervene is a vital component in modern obstetrics. This is the first study large enough to assess the effect of different positioning of the partogram action line, and subsequent timing of intervention, on the rate of cesarean delivery and maternal satisfaction.
This study demonstrates that earlier intervention, determined in this case by a visual prompt, simply increases obstetric interventions but neither reduces the likelihood of cesarean delivery nor improves women’s satisfaction with their birth experience. This is consistent with other studies of early intervention, whether based on early amniotomy,21 early augmentation with oxytocin, or based on nonvisual prompts.22 The seeming inability of earlier intervention to reduce cesarean delivery rates may not be an issue of timing but of the methods used to correct poor labor progress. It may be that in addition to an examination of the definitions of dystocia and timing of intervention, the effectiveness of interventions themselves needs to be reevaluated. There are data that suggest higher doses of oxytocin may reduce cesarean delivery rates,23 although the safety of this approach is unclear.
The percentages in the 4-hour group for the three related adverse perinatal outcomes (Table 2) are slightly higher. These small differences are neither clinically nor statistically significant and have almost certainly arisen by chance. This study was not adequately powered to detect differences of less than 1% in rare perinatal outcomes; such a study would require tens of thousands of participants.
It is of particular note that the use of the 2-hour action line in a midwife-led setting increased transfer rates out of this setting without demonstrable benefit to women or neonates. Subgroup analysis according to the intended place of birth has also raised a possibility of a significant interaction between the type of partogram, birth setting, and cesarean delivery rates. The observed differences in the direction of the effect may be a chance finding, as is often the case in subgroup analyses.24 Alternatively, a partogram that results in more transfers during labor (54% compared with 43% in our study) may lead to delayed implementation of interventions to correct slow progress (Syntocinon, Alliance Pharma plc, Chippenham, United Kingdom) or lower the threshold for cesarean delivery after a transfer of care between two birth settings. We did not collect data on the timing of various interventions and were therefore unable to explore these hypotheses further.
Based on our previous work,16 we predicted that 14% of women would be dissatisfied. This proved not to be the case, because only 7% were disappointed. It is possible that our previous study16 included more women to whom a perception of a shorter labor was important, and therefore allocation to 3- or 4-hour action lines did not meet their expectations and left them dissatisfied. Where reasons for declining this trial were recorded, concern over a longer interval to intervention was the commonest reason given, and thus these women may be underrepresented in the study.
In our uncomplicated primigravid population, 51.3% of women were diagnosed as being in “prolonged” labor (ranging from 57.3% in the 2-hour arm to 45.3% in the 4-hour arm), confirming others’ suggestions25 that the current definition of normal progressive cervical dilatation of 1 cm/h and an alert line that reflects that are unrealistic. This brings into question current definitions of poor progress. Many of these have been arrived at by consensus groups26 or are based on old data from different ethnic groups.2,27 Further observational research is required to assess whether the slope of labor progression for white women today is different from that defined by Philpott and Castle2 for African women in the early 1970s. It may be, as previously suggested,28 that the individual progress of each woman’s labor makes standardized guidance inappropriate. Work carried out in our unit in multiparous women suggests both that labors are individual, rather than linear in progress, and that definitions of dystocia need to be individualized.28 Ultimately, there is a need for a trial that compares the package of intrapartum care and obstetric interventions guided by the WHO 4-hour partogram with a package of care using partograms without lines. In the absence of fetal or maternal compromise, this latter partogram would rely on the observational skills of the clinicians alongside women’s individual wishes, to determine the need for intervention.
Only one half of eligible women were approached, and this raises concerns about generalizability. Recruitment to labor ward studies is fraught with ethical issues,29 and in most cases the reason given for failure to seek consent was the view of the attending midwife that the women in labor were not able to give truly informed consent. Women were excluded by midwives if perceived to be too distressed or had not had prior information about the trial. The role of the attending midwife as a gatekeeper was essential, although this consenting process raises important ethical dilemmas including denying women the choice of participating in research and the ability of laboring women to give truly informed consent for other clinical interventions.
Although clinical care other than the position of the partogram action line was not prescribed, it is possible that other aspects of care may have reduced any effect of either arm of the trial. The unit has well-established guidance for the management of labor, which was not altered during the trial. We feel it is unlikely that there was a significant confounder that had inadvertently influenced the results. The fact that similar results occurred in both birth settings argues against other factors affecting the differences in outcomes between the two types of care.
In summary this trial shows that intervention in slow labor based on an action line 2 hours to the right of the “normal” progress line is more common, but does not result in improved outcomes compared with intervention based on a 4-hour action line. Where clinicians believe that labor should be managed using graphically described cervical progress with predefined intervention times, then at present the 4-hour WHO partogram seems preferable in settings similar to that described in this trial. The trial highlights a number of issues with current definitions and beliefs of labor progress and management of slow labor that urgently need to be addressed in large clinical trials.
1. Lennox CE, Kwast BE. The partograph in community obstetrics. Trop Doct 1995;25:56–63.
2. Philpott RH, Castle WM. Cervicographs in the management of labour in primigravidae II: The action line and treatment of abnormal labour. J Obstet Gynaecol Br Commonw 1972;79:599–602.
3. Philpott RH, Castle WM. Cervicographs in the management of labour in primigravidae I: the alert line for detecting abnormal labour. J Obstet Gynaecol Br Commonw 1972;79:592–8.
4. Philpott RH. Graphic records in labour. Br Med J 1972;4:163–5.
5. World Health Organization partograph in management of labour. World Health Organization Maternal Health and Safe Motherhood Programme. Lancet 1994;343:1399–404.
6. O’Driscoll K, Foley M, MacDonald D. Active management of labor as an alternative to cesarean section for dystocia. Obstet Gynecol 1984;63:485–90.
7. Turner MJ, Webb JB, Gordon H. Active management of labour in primigravidae. J Obstet Gynaecol 1986;7:79–83.
8. Gerhardstein LP, Allswede MT, Sloan CT, Lorenz RP. Reduction in the rate of cesarean birth with active management of labor and intermediate-dose oxytocin. J Reprod Med 1995;40:4–8.
9. Frigoletto FD, Lieberman E, Lang JM, Cohen A, Barss V, Ringer S, Datta S. A clinical trial of active management of labor. N Engl J Med 1995;333:745–50.
10. Thornton JG, Lilford RJ. Active management of labour: current knowledge and research issues. BMJ 1994;309:366–9.
11. Lavender T, Malcolmson L. Is the partogram a help or a hindrance? An exploratory study of midwives’ views. Pract Midwife 1999;2:23–7.
12. Groeschel N, Glover P. The partograph. Used daily but rarely questioned. Aust J Midwifery 2001;14:22–7.
13. Cartmill RS, Thornton JG. Effect of presentation of partogram information on obstetric decision-making. Lancet 1992;339:1520–2.
14. Tay SK, Yong TT. Visual effect of partogram designs on the management and outcome of labour. Aust N Z J Obstet Gynaecol 1996;36:395–400.
15. Pattinson RC, Howarth GR, Mdluli W, Macdonald AP, Makin JD, Funk M. Aggressive or expectant management of labour: a randomised clinical trial. BJOG 2003;110:457–61.
16. Lavender T, Alfirevic Z, Walkinshaw S. Partogram action line study: a randomised trial. Br J Obstet Gynaecol 1998;105:976–80.
17. Cesarean section: National Collaborating Centre for Women’s and Children’s Health, commissioned by the National Institute for Clinical Excellence. London (UK): RCOG Press; 2004.
18. Royal College of Obstetricians and Gynaecologists. The use of electronic fetal monitoring: the use and interpretation of cardiotocography in intrapartum fetal surveillance. Evidence-based Clinical Guideline Number 8. London (UK): Royal College of Obstetricians and Gynaecologists; 2001.
19. Lavender T, Wallymahmed AH, Walkinshaw SA. Managing labor using partograms with different action lines: a prospective study of women’s views. Birth 1999;26:89–96.
20. Neilson JP, Lavender T, Quenby S, Wray S. Obstructed labour. Br Med Bull 2003;67:191–204.
21. Fraser WD, Turcot L, Krauss I, Brisson-Carrol G. Amniotomy for shortening spontaneous labour. Cochrane Database Syst Rev 2000;CD000015.
22. Fraser W, Vendittelli F, Krauss I, Breart G. Effects of early augmentation of labour with amniotomy and oxytocin in nulliparous women: a meta-analysis. Br J Obstet Gynaecol 1998;105:189–94.
23. Merrill DC, Zlatnik FJ. Randomized, double-masked comparison of oxytocin dosage in induction and augmentation of labor. Obstet Gynecol 1999;94:455–63.
24. Rothwell PM. Treating individuals 2. Subgroup analysis in randomised controlled trials: importance, indications, and interpretation. Lancet 2005;365:176–86.
25. Albers LL, Schiff M, Gorwoda JG. The length of active labor in normal pregnancies. Obstet Gynecol 1996;87:355–9.
26. National Consensus Conference on Aspects of Cesarean Section Planning Committee. Criteria used for the diagnosis of dystocia. Hamilton (Canada): McMaster University; 1985.
27. Friedman E. The graphic analysis of labor. Am J Obstet Gynecol 1954;68:1568–75.
28. Lavender T, Hart A, Walkinshaw S, Campbell E, Alfirevic Z. Progress of first stage of labour for multiparous women: an observational study. BJOG;112:1663–5.
29. Spencer SA, Dawson A. Implications of informed consent for obstetric research. Obstetrician Gynaecologist 2004;6:163–7.