A randomised controlled trial of non-pharmacological approaches

for relief of low back pain during labour


Michel Labrecque1 MD, MSc, Professor

Arie Nouwen2 PhD, Associate professor

Marc Bergeron3 MD

Jean-François Rancourt3 MD


1 Department of Family Medicine, Laval University, Sainte-Foy, Quebec, Canada

2 School of Psychology, Laval University, Sainte-Foy, Quebec, Canada

3 Hôtel-Dieu de Montmagny, Montmagny, Quebec, Canada




Correspondence : Michel Labrecque, 1296 rue Green, Sainte-Foy, Qc, Canada.

Phone : (418) 659-3099; Fax : (418) 659-3173;

E-mail: michel.labrecque@mfa.ulaval.ca



Headline: Trial of three approaches for relief of back pain during labour


Objective: To compare the effectiveness of three non-pharmacological approaches for relief of low back pain during obstetrical labour.

Design: Randomised controlled trial.

Setting: A rural hospital with approximately 350 deliveries per year in the Province of Quebec (Canada).

Population: Thirty-four women suffering of low back pain during labour.

Methods: Participants were randomly assigned to receive one of three treatments : 1) intracutaneous sterile water injections (ISW) (n=10), 2) transcutaneous electrical nerve stimulation (TENS) (n=12), 3) standard care, including back massage, whirlpool bath, and liberal mobilisation (MBM) (n=12).

Main Outcome Measures: Women self-evaluated both intensity and affective dimensions of pain using visual analogue scales. Women's sense of control and satisfaction were assessed with adapted versions of the Labour Agentry Scale and the Labor and Delivery Satisfaction Index .

Results: Women in the ISW group rated both the intensity and unpleasantness of pain over the experimental period significantly lower than women in either the MBM or the TENS groups, (p= 0.0001 and 0.0003, respectively). Similar results were observed for intensity (p = 0.0), and for unpleasantness (p = 0.03) of the pain ratings just prior to the request for an epidural or delivery. Mean pain intensity at 15 and 60 minutes after randomisation was significantly reduced in the ISW compared to the two other groups. There was no significant difference between the three groups in the level of control and satisfaction towards labour and delivery.

Conclusion: Intracutaneous sterile water injections are more effective than standard care (back massage, bath, and mobilisation) or transcutaneous electrical nerve stimulation for relieving low back pain during labour.


Approximately one-third of women suffers from severe low back pain during labour. 1,2 Epidural analgesia, when available, is often the preferred method for relieving the pain. Although very effective, epidural analgesia, particularly when given at a cervical dilatation of five cm or less, is associated with an increase in Cesarean-section delivery (Morton, Ramin,Thorp- 3-5). Narcotics are effective in relieving pain but are associated with side effects and have potentially deleterious consequences for the newborn. Non-pharmacological approaches that are based on the Gate Control Theory (1) have been suggested as treatment alternatives. One of these techniques consists of injecting sterile water intracutaneously in the lower back. Intracutaneous sterile water (ISW) injections has shown to be effective in four experimental studies (trolle, ader, lytzen Aarnes, 6-9) of which two are double-blinded, placebo-controlled, randomised trials (trolle, ader- 6,7). Transcutaneous electrical nerve stimulation (TENS) is also recommended. Although according to a meta-analysis including six randomised placebo-controlled trials (cochrane 10), TENS received very favourable evaluations by women who reported that they would used it during other pregnancies, it does not seem to decrease the intensity of the pain nor to preclude the use of other forms of analgesia. This finding was confirmed in a more recent trial. Van der ploeg -11

Although the literature seems to favour ISW over TENS, no study has compared these two approaches for relief of low back pain during labour. Therefore, we carried out a randomised controlled trial in a rural hospital to compare the effect on low back pain of ISW, TENS and the current non-pharmacological approach used in this environment namely low back massage, whirlpool baths, and liberal mobilisation. We also explored the effect of these approaches on request for other means of analgesia.


Participants and Procedure

Participants were recruited between September 1995 and January 1997 at the Hôtel-Dieu de Montmagny Hospital (Canada), a rural hospital. Among the 403 pregnant women admitted for delivery during the study period, 304 (75%) low-risk pregnant women in active first stage labour were informed about the study. Only those who subsequently complained of low back pain during the labour were eligible to participate in the study.

Upon signing an informed consent form, participants were randomly assigned to either one of three intervention groups: 1) intracutaneous sterile water injections (ISW); 2) transcutaneous electrical nerve stimulation (TENS); and 3) standard care including back massage, whirlpool bath, and liberal mobilisation (MBM). The randomisation scheme was developed using a table of random numbers and balanced in blocks of six or nine, and a set of sealed, sequentially numbered, opaque envelopes was used for study group assignment.

The following baseline characteristics were collected before randomisation: obstetrical and socio-demographic characteristics, intensity and affective dimensions of back pain using 10-cm Visual analogue scales (VAS) (Price et al.- 12), and self-efficacy (Bandura, 1977 -13) to tolerate labour-related back pain for 15, 60, 90, 120 and 180 min, using a 11-point numerical graphic rating scale.


Standard care (MBM group) consisted of an initial 15-min back massage with a moisturising cream either by the nurse or their partner, followed by either continued massages, whirlpool baths, or liberal mobilisation (walking or frequent change of position) according to the woman’s choice. For women assigned to the TENS group, the attending nurse or physician installed two pairs of disposable electrodes on the skin of the low back region connected to a 3Mâ portable TENS unit. Units were set in normal mode. Initial current intensity of each two channels was adjusted according to tolerance starting with a rate of 80-125 pulses per second and a pulse width of 60-100 msec. Women were instructed to proceed with subsequent adjustment according to their needs. Women in the ISW group received four intradermal injections of 0.1cc of sterile water over the lumbosacral area according to the technique described by Reynolds (Reynolds -<14>). The injections were administered during a contraction either by the attending nurse or physician and were repeated upon request. In both TENS and ISW groups, women were allowed to use any manoeuvre of the standard care approach 15 minutes after receiving their assigned interventions. All participants were told that if at any moment they considered that their pain was not sufficiently relieved, they could request other form of analgesia including epidural.

Outcome measures

The main outcome measures were the intensity and affective components of pain at 15, 60, 90, 120 and 180 minutes after the initial intervention, assessed using a 10-cm Visual analogue scale (VAS). Time and cervical dilatation at the moment of request and administration of epidural analgesia or narcotics were recorded. Data on other obstetrical outcomes were extracted from the medical records by a research nurse. Participants completed a self-administered questionnaire within a few days after delivery, addressing their feelings of control and satisfaction with the delivery as assessed with the French-Canadian brief version of the Labour Agentry Scale (LAS) <15> and an adapted version of the Labor and Delivery Satisfaction Index (LADSI) <16>. Higher score on these scales indicates higher levels of perceived control or satisfaction.

Data analysis

Differences between the three groups in terms of socio-demographic and obstetrical variables, self-efficacy to tolerate pain, and back pain prior to randomisation were examined by bilateral Fisher’s exact test and ANOVA for categorical and continuous variable, respectively.

Difference between the groups on the intensity and unpleasantness of back pain were evaluated by analysis of variance (ANOVA) using two models, one with the last rating of back pain attained just before the request of epidural or delivery, and one on the average pain rating during labour until the request of epidural or delivery. Baseline intensity (or unpleasantness) of back pain and baseline characteristics not equally distributed between the groups were used as co-variable in the models. Total mean scores of the Labour Agentry Scale and the Labor and Delivery Satisfaction Index between the groups were compared simultaneously using a multivariate analysis of variance (MANOVA). All post-hoc analyses were done using Tukey-Kramer tests.


Among the 304 informed women, 45 (15%) complained of low back pain and 35 were recruited. One women randomised to the ISW group was excluded for further analysis because she did not receive the injections and most data were missing due to a precipitate delivery. Table 1 shows the characteristics of the 34 participants according to study groups. Although none of the difference between the groups were statistically significant (all p values > 0.05), mean age was higher in ISW group, education level was lower in the TENS group and perceived self-efficacy to tolerate pain was higher in the MBM group.

Comparison of average pain ratings over the experimental period indicated significant differences between the three groups for intensity and unpleasantness (Table 2). Subsequent post-hoc analyses showed that the women in the ISW group rated both the intensity and unpleasantness significantly lower than women in either the BBM or the TENS groups. Similarly, comparison of the pain ratings just prior to the request for an epidural or delivery showed significant difference for intensity and for unpleasantness (Table 2). Subsequent post-hoc analyses showed that just before leaving the experiment women in the ISW group rated both the intensity and unpleasantness of pain significantly lower than women in the MBM and TENS groups. All pain rating differences between the groups remained statistically significant when baseline pain (intensity or unpleasantness), age, education level and mean self-efficacy were included in the ANOVA models.

As illustrated in Figure 1, reduction of pain intensity in the ISW group was rapidly achieved and persisted over the first hour of intervention period. After the first hour, the number of subjects with pain assessment was to small for reliable estimation of effect.

While the results above indicate that women in the ISW group , when compared to either BBM or TENS, rated pain intensity and unpleasantness ratings lower, this intervention may still be less attractive in terms of treatment satisfaction. The MANOVA failed to show significant difference in the level of control and satisfaction between the three groups as assessed by the LAS and the LADSI, respectively (Table 2). However, the proportion of women who would have liked to receive the same treatment for relief of low back pain at another delivery was significantly lower in the ISW group. In the latter group, the four women who would receive the same treatment tended in average to have more pain before the injections (84mm vs 69 mm) and to be better relieved at 15 minutes (17 mm vs 26 mm ) and 60 minutes (15 mm vs 42 mm) than the six who would not . The proportion of women who would recommend the same treatment to a friend was not statistically different between the groups. (Table 2)

No significant differences were found between the three groups regarding the number of Caesarean-section, the number of epidural requested and effectively received, the dilatation on the moment of request for epidural, and the time between randomisation and the time of request for an epidural (Table 2). Only one woman, in the MBM group, received narcotics during labour.



Our study was the first to compare three non-pharmacological approaches for the relief of low back pain during obstetrical labour. ISW rapidly and effectively reduced low back pain compared to either MBM or TENS which both proved to have no significant analgesic effect.

The magnitude of the reduction of pain with ISW in the present study (-55 mm at 15 min and -45 mm at 60 minutes) is comparable to previous reports by Trolle et coll. <6> (-53 mm at 60 min.) and Ader and coll.<7> ( -42 mm at 10 min. and -35 mm at 45 min). The absence of significant effect on low back pain observed with TENS is also in agreement with earlier studies demonstrating no effect of this approach on labour pain <10,11>. Women in MBM group which integrated liberal mobilisation as a component of the approach, did not benefit any pain reduction. A decrease of 83% of back pain during early labour had been described with standing position <17> but as we did not measure the time spent lying or standing, it is difficult to compare these results to ours.

Perceived control and overall satisfaction towards labour and delivery does not seem to be influenced by any of the interventions studied. However, the absence of difference observed between the groups may be due to a lack of statistical power. On the other hand, women in the ISW group appeared to be less satisfied of the treatment received to relief their back pain. Only four women out of 10 intended receiving ISW again at another delivery. ISW is associated with a sharp pain lasting 20-30 seconds during administration. Although the ISW injections were usually done during a contraction, for some women, this very short high intensity pain may be less acceptable then suffering low back pain. In earlier studies, 68% <6> to 81% <8> of participants said they would request ISW at another delivery. Cultural disparities and participant selection may explained these differences. Nevertheless, women must be informed of this only < 6-9 > but bothersome side-effect.

We did not observed any statistically significant difference between the groups in request for epidural, epidural received and other birth characteristics and outcomes. However, the study was primarily design to evaluate the effect of the approaches on pain and did not have the statistical power to rule out effect on other labour and delivery characteristics. Epidural analgesia is associated with a higher risk of caesarean—section delivery<3-5>, especially if provided at a cervical dilatation of five cm or less <3,4>. Analgesia with ISW could potentially help decreasing the use of epidural or retard its administration. Further studies with larger sample size are needed to verify this hypothesis.

According to previous reports, up to one-third of women suffer low back pain in labour. <1,2> We observed a lower incidence of this problem in the present study (15%) as no data on low back pain were collected in women who, at admission, refused further involvement with the study. Although external validity may be questionable, this underestimation does not influence internal validity of the study.

In conclusion, among the three non-pharmacological approaches studied for relief of low back pain during labour, only ISW proved to be effective. Back pain is reduced rapidly and relief last for at least one hour. Although it does not seem to modify the sense of control and satisfaction towards labour and delivery experience compared to standard care and TENS, acceptability may be a problem for some women due to short duration stinging pain induced by the injections.


This study was supported by a grant from the Fonds de recherche en santé du Québec.


  1. Bonica JJ. Peripheral mechanisms and pathways of parturition pain. Br J Anesth 1979;51:53-9.
  • Melzack R, Schaffelberg D. Low-back pain during labor. Am J Obstet Gynecol 1987;156:901-5.
  • Thorp JA, Hu DH, Albin RM. et coll. The effect of intrapartum epidural analgesia on nulliparous labor: A randomized, controlled, prospective trial. Am J Obstet gynecol 1993;169:851-8.
  • Morton SC, Williams MS, Keeler EB, Gambone JC, Kahn KL. Effect of epidural analgesia for labor on the cesarean delivery rate. Obstet Gynecol 1994;83:1045-52.
  • Ramin SM, Gambling DR, Lucas MJ, Sharma SK, Sidawi JE, Leveno KJ. Randomized trial of epidural versus intravenous analgesia during labor. Obstet Gynecol 1995;86:783-9.
  • Trolle B, Moller M, Kronborg H, Thomsen S. The effect of sterile water blocks on low back labor pain. Am J Obstet Gynecol 1991;164:1277-81.
  • Ader L, Hansson B, Wallin G. Parturition pain treated by intracutaneous injections of sterile water. Pain 1990;41:133-8.
  • Lytzen T, Cederberg L, Möller-Nielsen J. Relief of low back pain in labor by using intracutaneous nerve stimulation (INS) with sterile water papules. Acta Obstet Gynecol Scand 1989;68:341-3.
  • Dahl V, Aarnes T. Sterile water papulae for analgesia during labor. Tidsskr Nor Laegeforen 1991;111(12):1484-7.
  • Howell CJ. Transcutaneous nerve stimulation (TENS) in labor. In: Keirse MJNC, Renfrew MJ, Neilson JP, Crowther C (eds.) Pregnancy and Childbirth Module. In: The Cochrane Database of Systematic Reviews. Published through the Cochrane Collaboration; Issue 2, Oxford: Update Software; 1995.
  • van der Ploeg JM, Vervest HA, Liem AL, Schagen van Leeuwen JH. Transcutaneous nerve stimulation (TENS) during the first stage of labor: a randomized clinical trial. Pain 1996;68:75-8.
  • Price DD, McGrath PA, Rafii A, Buckingham B. The validation of visual analogue scales as ratio scale measure for chronic and experimental pain. Pain 1983;17:45-56.
  • Bandura A. Self-efficacy: Toward a unifying theory of behavioral change. Psychol Rev 1977;84:191-215.
  • Reynolds JL. Intracutaneous sterile water for back pain in labor. Can Fam Physician 1994;40:1785-92.
  • Hodnett ED, Simmons-Tropea DA. The labour agentry scale: Psychometric properties of an instrument measuring control during childbirth. Res Nurs Health 1987;10:301-10.
  • Lomos J, Dore S, Enkin M, Mitchel A. The labour and delivery satisfaction index: The developpement and evaluation of a soft outcome measure. Birth 1987:14(3):125-9.
  • Melzack R. Labour pain as a model of acute pain. Pain 1993;53:117-20.

    Table 1 Baseline characteristics of the participants according to study groups

    Baseline Characteristics






    Mean age (yr)
    27 ± 6*
    27 ± 5
    29 ± 6
    Nulliparae (%)
    Mean education level (yr)
    13 ± 2
    11 ± 2
    13 ± 3
    Mean gestational age (wks)
    40 ± 1
    39 ± 3
    40 ± 1
    Mean cervical dilatation (cm)
    3 ± 2
    3 ± 2
    3 ± 1
    Mean intensity of back pain (mm)
    81 ± 14
    77 ± 14
    75 ± 15
    Mean unpleasantness due to back pain (mm)
    80 ± 15
    70 ± 23
    79 ± 15
    Back pain worst than abdominal pain (%)
    Mean self-efficacy to tolerate back pain over 180 min (%)
    54 ± 18
    36 ± 26
    39 ± 32
    Mean time to less then 50 % self-efficacy (min)
    98 ± 43
    51 ± 66
    70 ± 69
    Mean birth weight (g)
    3406 ± 425
    3563 ± 326
    3387 ± 422
    Baby’s head in occipito-posterior position (%)

    * ± standard deviation


    Table 2. Outcome variables according to study groups

    Outcomes Variables
    P value






    Primary Outcomes        
    Mean pain intensity during intervention period (mm)
    79 ± 6*
    66 ± 6
    32 ± 6
    Mean pain unpleasantness during intervention period (mm)
    73 ± 7
    78 ± 7
    30 ± 7
    Last pain intensity before epidural or delivery (mm)
    82 ± 9
    68 ± 9
    40 ± 9
    Last pain unpleasantness before epidural or birth (mm)
    78 ± 9
    68 ± 9
    41 ± 10
    Secondary Outcomes        
    Mean Labor Agentry Scale score
    2.6 ± 0.6
    3.4 ± 1.0
    3.5 ± 1.2
    Mean Labour and Delivery Satisfaction Index score
    5.1 ± 0.7
    5.3 ± 0.4
    5.4 ± 0.4
    Would receive the same treatment for back pain at another delivery (%)
    Would recommend the same treatment for back pain to a friend (%)
    Delivery by Cesarean-section (%)
    Epidural requested (%)
    Epidural received (%)
    Mean cervical dilatation when epidural requested (cm) §
    6 ± 2
    6 ± 2
    6 ± 2
    Mean time between randomization and request for epidural (min) §
    103 ± 94
    145 ± 80
    130± 68

    * ± standard deviation

    With MANOVA model including both the Labor Agentry Scale and the Labour and the Labour and Delivery Satisfaction Index

    § In 25 women who requested an epidural

    Fig. 1. Mean intensity of pain (± 95% confidence intervals) at 0, 15, 60 and 90 minutes after beginning intervention according to study groups