DOI: 10.1111/j.1471-0528.2008.01781.x
General obstetrics
www.blackwellpublishing.com/bjog
Maternal temperature during labour FD Schouten,a H Wolf,a BJ Smit,b DJ Bekedam,c R de Vos,d I Wahlenc a Department of Obstetrics and Gynaecology, Academic Medical Centre, Amsterdam, the Netherlands b Division of Neonatology, Department of Paediatrics, Erasmus MC-Sophia Children’s Hospital, University Medical Centre, Rotterdam, the Netherlands c Department of Obstetrics and Gynaecology, Onze Lieve Vrouwe Gasthuis, Amsterdam, the Netherlands d Department of Clinical Epidemiology and Biostatistics, Academic Medical Centre, Amsterdam, the Netherlands Correspondence: Dr FD Schouten, Academic Medical Centre, H4-253, PO Box 22660, 1100 DD Amsterdam, the Netherlands. Email
[email protected]
Accepted 10 April 2008. Published OnlineEarly 2 June 2008.
Objective The aim of this study was to describe the variation of
Main outcome measures Rectal measured temperature in degrees
normal maternal temperature during labour.
Celsius.
Design A prospective cohort study.
Results The mean temperature during labour in the complete study population increased from 37.1C at the beginning of labour to 37.4C after 22 hours. Temperature in the abnormal labour group was equal to the normal labour group during the first 3 hours of labour (P > 0.05) but increased thereafter.
Setting Two hospitals in Amsterdam, the Netherlands. Population All women with a live singleton pregnancy and
a gestational age of 36 weeks or more admitted to the delivery ward from June 2000 to January 2002. Methods Maternal temperature was measured rectally every 2–3
hours from admission until the beginning of second stage, and 1hour postpartum. Normal labour (n = 843) was defined as gestational age ‡37 weeks, spontaneous onset of labour, rupture of membranes 6 hours before delivery. Ten percent of the women with epidural were treated with antibiotics, while in those without epidural, 2% were treated. The median number of temperature measurements per patient before delivery was 3 (quartiles 2–4, range 1–12), and the average duration of labour was 10.5 hours (95% CI 3–23 hours). The mean temperature during labour in the Table 1. Description of the study group Total number eligible Total number included Maternal age (years) Non-European descent Primipara Group allocation Normal labour Abnormal labour In hospital 6 hours Duration of labour (hours) Ruptured membranes 18 hours Intrauterine infection Antibiotics during labour Diagnosis of IUI according to study protocol Diagnosis of IUI not according to protocol Urinary tract infection prelabour Prophylaxis for GBS Prophylaxis for endocarditis Gestational age at delivery (weeks) Birthweight (g) IUI, intrauterine infection. Data are presented as n (%) or mean (SD).
3358 3052 (90.9) 30.6 (5.8) 1111 (36.4) 1631 (53.1) 843 (27.6) 2209 (72.4) 1196 (39.2) 10.8 (6.0) 529 (17.4) 84 (2.7) 89 (2.9) 68 (2.2) 11 (0.4) 2 (0.1) 7 (0.2) 1 (0.0) 40.0 (1.4) 3435 (534)
complete study population increased from 37.1C (2SD 0.7C, 95% CI 36.5–37.7C) at the beginning of labour to 37.4C (2SD 1.2C, 95% CI 36.4–38.2C) after 22 hours (Figure 1A). The regression line for the upper 2SD limit was: temperature = 37.741 + 0.030 · time (hours). In the normal labour group, the mean temperature was 37.1C (95% CI 36.4–37.7C). Temperature in the abnormal labour group was equal to the normal labour group during the first 3 hours of labour (P > 0.05) but increased thereafter (Figure 1B). The regression coefficient for the 2SD limit was significantly higher in the abnormal labour group (0.030C/hour) than in the normal labour group (0.008C/hour). For temperature measurements during the first 3 hours of labour, a circadian pattern could be detected: mean temperature 37.1C, amplitude 0.1C and acrophase at 21.3 hours (Figure 2). For all measurements before delivery, the mean temperature was 37.1C, amplitude 0.1C and acrophase at 21.2 hours, which were similar in both the normal and abnormal labour groups. For measurements higher than 1SD, the amplitude (0.01C) was not significantly different from zero and a circadian pattern could not be recognised. For further reference, the complete study group 2SD limit was used. In the normal labour group, only 2% of women had a temperature during labour higher than the upper 2SD limit, while this occurred in 8% of the abnormal labour group. About 91% of women with the diagnosis of intrauterine infection had a temperature above the 2SD limit. Multivariate analysis demonstrated that epidural analgesia, in hospital ‡6 hours, meconium-stained liquor and rupture of membranes ‡18 hours were significantly associated with a temperature higher than the 2SD limit (OR depicted in Figure 3). The remaining classification parameters were ejected from the model.
Table 2. Temperature (mean and SD) during the last 3-hour period before delivery and after delivery Duration of labour (hours)
,6** 6–11** 12–17** 18 Total
Last 3-hour period*
Postpartum
Mean
SD
Mean
SD
37.01 37.09 37.18 37.33 37.13
0.39 0.45 0.47 0.52 0.47
37.13 37.26 37.29 37.38 37.26
0.40 0.46 0.50 0.52 0.47
ANOVA, analysis of variance. *Difference between duration of labour groups statistically significant (ANOVA). **Difference between temperature before and after delivery statistically significant (ANOVA).
ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology
1133
Schouten et al.
(A)38.5 Mean ± 1 SE temperature (°C)
37.4
Temperature (°C)
38.0
37.5
37.0
36.5
37.3
37.2
37.1
37.0
36.9
36.0 0
4
8
12
16
20
0
3
6
Duration of labour (hours)
12
15
18
21
24
Figure 2. Mean temperature with standard error of the mean during the first 3 hours of labour specified for time of day. The continuous line represents the predicted values by the cosinor function: mean 37.1C, amplitude 0.13C and acrophase at 21 hours.
(B)38.5 38.0
Temperature (°C)
9
Time of day
37.1C, SD 0.5C), except in women who had duration of labour ‡18 hours, where it was similar (Table 2 and Figure 5).
37.5
Discussion 37.0
This prospective cohort study observed that normal temperature was 37.1C (2SD 0.7) at the beginning of labour and that temperature increased slowly during labour to 37.4C (2SD 1.2) after 22 hours. At the beginning of labour, temperature was comparable between women with normal or abnormal labour. Women with abnormal labour had an increase in
36.5
36.0 0
4
8
12
16
20
Duration of labour (hours) Figure 1. (A) Mean temperature with 2SD during labour in complete study group. Time is counted from the beginning of active labour. A linear regression line is drawn for the mean, +2SD and –2SD. (B) Separate plots for normal labour (asterisk) and abnormal labour (dot). The regression coefficient for the 2SD line was significantly higher in the abnormal labour (0.030) than in the normal labour group (0.008).
Figure 4 demonstrates the association with temperature in the last 3-hour period before delivery in the complete study group for the parameters that were used for group classification or were thought to be related to temperature. Epidural analgesia was associated with an elevation of 0.4C. Induction or augmentation of labour, meconium-stained amniotic fluid, rupture of membranes more than 18 hours before delivery, duration of hospital admission for more than 6 hours before delivery and more than five vaginal examinations were associated with a small, but statistically significant increase of maternal temperature of 0.05–0.2C. Temperature after delivery (mean 37.2C, SD 0.5C) was higher than during the 3-hour period before delivery (mean
1134
Epidural analgesia
In hospital >6 hours
Meconium-stained liquor
Ruptured membranes >18 hours
0
1
2
3
4
5
6
Odds ratio Figure 3. Odds ratios and their 95% CI for a temperature higher than 2SD, calculated by multivariate regression analysis. Variables at first step were in hospital ‡6 hours, duration of labour ‡8 hours, meconiumstained liquor, epidural analgesia, induction or augmentation of labour, rupture of membranes ‡18 hours, greater than or equal to five vaginal examinations and operative delivery.
ª 2008 The Authors Journal compilation ª RCOG 2008 BJOG An International Journal of Obstetrics and Gynaecology
Variation of maternal temperature during labour
Vaginal examination ≥5 (18%) In hospital ≥6 hours (39%) Rupture of membranes ≥18 hours (18%) Operative delivery (24%) Augmentation of labour (29%) Meconium (26%) Induction of labour (15%)
–0.4
–0.2
0.0
0.2
0.4
Temperature (°C) Figure 4. Point estimate of the mean temperature difference (C) with 95% CI between women with and without the indicated parameters during the last 3 hours of labour.
temperature during labour, while temperature remained stable in the normal labour group (Figure 1). A reference line at 2SDs of the mean was calculated for classification of elevated temperature. The diagnosis of intrauterine infection (by definition), the application of epidural analgesia, duration of admission in hospital ‡6 hours, meconium-stained liquor and rupture of membranes ‡18 hours were significantly associated with a temperature higher than the 2SD limit. It is well known that temperature follows a circadian pattern with higher values during daytime than during night-time.23 Acker et al. observed a circadian pattern of temperature at onset of labour, although at a slightly lower level, probably because they measured temperature orally.24 In our study, admission temperature and mean temperature followed a circadian pattern. However, this pattern was not detected for temperature higher than 1SD above the mean. Correction for the time of day is therefore not necessary for elevated temperature. Unlike Lieberman, we did not observe an association of elevated temperature with operative delivery. This might be due to a lower percentage of women with epidural analgesia in our study.15 The percentage of women with normal labour according to study definitions in our study was only 28%. This was due to selection in the Dutch obstetric care system, where approximately 35% of women deliver at home and the indication for hospital delivery is generally based on obstetric risk factors.25 The association of epidural analgesia with elevated temperature is extensively described in literature.9–12,15,26–30 Most studies did not observe an association of increased temperature after epidural analgesia with intrauterine or neonatal infection, but a number of studies describe an increase of antibiotic prescription, obstetric interventions or neonatal
sepsis work-up. In our study, the diagnosis of intrauterine infection was made five times more frequently compared with that in the women without epidural analgesia (10 versus 2%), based on strict diagnostic criteria. This elevated incidence of infection could be explained in part by the association of epidural analgesia and prolonged duration of labour, which was caused by selection because epidural analgesia was usually indicated by prolonged labour.31–33 Increase of temperature in the remaining women with epidural analgesia could be due to its interference with heat dissipation.34 Temperature measurements in most studies were obtained from the mouth, ear canal, axilla, uterus or rectum. In general, intestinal or pulmonary artery temperature measurement is regarded as the gold standard for measurement of body core temperature. As these methods are not feasible in routine clinical practice, other methods have to be used. Comparative studies demonstrate that rectal or intrauterine measurement is most reliable, sublingual measurement is acceptable and tympanic measurement is insufficient.6,8,35 Banerjee et al. compared intrauterine temperature with sublingual, tympanic and skin temperature. Sublingual temperature was 0.8C lower than intrauterine temperature measurement, with an interclass correlation coefficient of 0.8. Lower sublingual temperature during labour could be mediated by forced respiration. Bartholomew et al.,
39.5 39.0 38.5
Temperature (°C)
Epidural analgesia (8%)
38.0 37.5 37.0 36.5 36.0 35.5
