Intradermal sterile water injections for the relief of low back pain in labour—A systematic review of the literature

Introduction 

A systematic literature review was conducted of the published research into the use of intradermal sterile water injections in order to determine the efficacy of this treatment for low back pain afflicted labours.

Melzack and Schaffelberg1 found that up to 30% of women in labour experience severe lower back pain, which is both continuous and frustrating. This type of pain has been described as quantitatively dissimilar to the pain derived from uterine contractions and is continuous throughout labour, affording no periods of respite. It intensifies to excruciating levels during the contraction on which it is superimposed, and then subsides to leave the low back pain as a baseline. Low back pain has been associated more commonly with labours in which the baby adopts a malrotated occipito-posterior position.2

In an Australian study, Paech3 found that in terms of maternal satisfaction unquestionably the gold standard for analgesia during labour is the epidural block; however an anaesthetic service is not available to all women, especially those in developing countries or in some rural and remote areas.4 Parenteral narcotic analgesia reduces pain but exhibits unwanted side effects such as neonatal respiratory depression and loss of maternal control.

The discovery of a non-pharmacological analgesic technique that is simple, safe and effective without serious side effects would prove to be beneficial for both mother and baby; an example being the use of intradermal water injections. This technique appears to produce the stimulation of skin nociceptors by injecting sterile water under the dermis to relieve referred pain. Research has previously shown this method to be effective in the treatment of renal colic, whiplash injuries and treating low back pain in labour.5, 6, 7, 8, 9, 10, 11, 12, 13 The aim of this literature review is to answer the question: “Do intradermal water injections relieve low back pain in labour?” by systematically reviewing six randomised controlled trials, comparing their respective methodologies and critically discussing their conclusions.

Methods 

A search of the published literature on this topic was undertaken. An Advanced Ovid Search included keyword search terms such as: injections, labour/labor, pain, water, subcutaneous, and yielded several results. The Cochrane Collaboration and CINAHL databases were searched and Scopus was accessed through the University of Newcastle online library. A total of 14 articles were retrieved for subject review and background context, ranging chronologically from 2 to 18 years old.

These articles provided a diverse international perspective on this topic, including research from Canada, Sweden, Denmark, Thailand and Iran.

Two articles were open studies with neither double blinding nor placebo control.7, 14 Three articles were by Canadian Dr Larry Reynolds, outlining the first practical guide for the use of intracutaneous sterile water for the relief of back pain in labour and providing colour photos of the desired injection sites. Reynolds2 cemented his position as one of the leading experts on this topic by publishing a didactic commentary on the Swedish randomised controlled trial of labour pain treated with cutaneous injections of sterile water by Martensson and Wallin.10

Simkin, an American physiotherapist, has been published twice on this subject in the course of her research into non-pharmacological pain relief methods for labour.15, 16 The six articles selected for systematic review are all randomised controlled trials from 1990 to 2006.8, 9, 10, 11, 12, 13 Publishing sources are reputable and well respected, including: Pain, American Journal of Obstetrics and Gynecology, The Journal of Family Practice, British Journal of Obstetrics and Gynaecology, Australian and New Zealand Journal of Obstetrics and Gynaecology and the Journal of the Medical Association of Thailand.

A number of methodological differences were identified between these studies. Written and verbal consent for treatment was gained by all except Martensson and Wallin,10 and all studies except Labrecque et al.11 were conducted in accordance with the Declaration of Helsinki and passed through local ethics committees. Participants in all studies were recruited in active labour once the experience of low back pain was verbalised. Studies other than Trolle et al.9 required that participants had received no analgesia in the 3h prior to study inclusion, and studies other than Ader et al.8 did not permit the use of other analgesia during the assessment period. Notably, the Canadian study11 did not specify any restrictions on analgesia use either prior to or during the course of the study.

Martensson and Wallin10 allowed administration of the sterile water injections to occur whilst the woman was inhaling nitrous oxide to mask the pain associated with administration, and to ensure the placebo effect (normal saline is not as painful to administer as sterile water).

All studies described the procedure with a similar technique. Using an insulin syringe with a small bore needle, varying amounts (0.1–0.5ml) of sterile water or normal saline (placebo) were injected under the skin at four sites over the sacrum. This area has been referred to as Michaelis’ Rhomboid. One injection was administered over each posterior superior iliac spine and then another two injections were administered 3cm below and 1cm medial to each of the first two injections.17 Studies except Trolle et al.9 and Bahasadri et al.12 administered the injections during the peak of a contraction to mask any administration pain and Bahasadri et al.12 used only one site, the most painful point over the sacrum, for administration. All studies except the Danish one9 had the injections administered by an independent midwife who was not caring for that woman during labour.

The Visual Analogue Scale has been shown to have high validity and reliability in pain quantification.18 It provides a graphic representation of pain from 0 (no pain) to 10 (unbearable pain) and was the measuring tool used by all studies except the Iranian one, which stated that “this method is difficult to explain to patients with varying degrees of education especially when they are in pain” (p. 104). Alternatively, this study used the Faces Rating Scale for its ease of understanding and administration, citing favourable comparison to the Visual Analogue Scale in terms of pain assessment properties.19

A systematic review of these six randomised controlled trials is presented graphically in Table 1, providing an extension of research initiated by Simkin and O’Hara in 2002.15

Table 1. Systematic review of 6 RCT’s using SWI for relief of labour low back pain

	Author	Entry criteria	Groups/interventions	Validity
	Ader et al.,8 Sweden, N=45	1st stage labour. Severe low back pain. No analgesia last 3/24. >37/40. Consent. Ethics	SG (n=24); 0.1ml SW i/c. CG (n=21); 0.1ml NS s/c. 4× injection sites. Similar age, parity, gestation, dilatation, membrane status	− R not described; + double blind; + placebo controlled; + well matched
	Trolle et al.,9 Denmark, N=272	Severe low back pain. Active labour. Consent. Ethics	SG (n=141); 0.1ml SW i/d. CG (n=131); 0.1ml NS i/d. 4× injection sites. Similar age, parity, dilatation, gestation, fetal size, membrane status	+ Largest RCT; + R described; + double blind; + placebo controlled; + well matched
	Labrecque et al.,11 Canada, N=35	1st stage labour. Low risk. Low back pain. >36/40. Consent	SG1 (n=11); 0.1ml SW repeated. SG2 (n=12); TENS. CG (n=12); usual care, massage, bath. 4× injection sites. Similar age, education, parity, dilatation	− Smallest RCT; + R described; − not blinded; + well matched
	Martensson and Wallin,10 Sweden, N=99	1st stage labour. Severe low back pain. No analgesia last 3/24. >37/40. Ethics	SG1 (n=33); 0.1ml SW i/d. SG2 (n=33); 0.5ml SW s/c. CG (n=33); 0.1ml NS s/c. 4× injection sites. Similarities not tested	+ R described; + double blind; + placebo controlled
	Bahasadri et al.,12 Iran, N=100	1st stage labour. Low back pain. >37/40. No prior analgesia. Ethics. Consent	SG (n=50); 0.5ml SW s/c. CG (n=50); 0.5ml NS s/c. 1× injection only to sacrum. Similar age, weight, parity, gestation, dilatation	+ R described; + double blind; + placebo controlled; + well matched
	Wiruchpongsanon,13 Thailand, N=50	1st stage labour. No analgesia last 3/24. >37/40. Consent. Ethics	SG (n=25); 0.1ml SW i/c. CG (n=25); 0.1ml NS i/c. 4× injection sites. Similar age, parity, gestation, dilatation, membrane status	− R not described; + double blind; + well matched

	Author	Outcome	Implications	Satisfaction
	Ader et al.,8 Sweden, N=45	<LBP per VAS at 10/60*** 45 and 90/60*. <LBP per midwife (blinded)***. No change: Narcotics EDB (<4%). Labour length. Instrumental. LSCS (<8%)	Injecting SW i/d over the sacrum decreases low back pain in labour for at least 90min. No difference in use of analgesia	Study group expressed a greater desire to repeat water blocks in future compared to control group
	Trolle et al.,9 Denmark, N=272	<LBP per VAS at 1/24*** and 2/24***. No change: Narcotics. Labour length. Instrumental. Blood loss. Apgar scores. <LSCS (4% vs. 11%)**	Injecting SW i/d over the sacrum decreases low back pain in labour for at least 2h and may decrease LSCS rate. Increased maternal report of pain relief (89% vs. 45%)***	Maternal satisfaction is high in this setting which doesn’t provide an epidural service. Increased maternal desire to use again in future labours**
	Labrecque et al.,11 Canada, N=35, SG1 vs. CG	<LBP per VAS at 10/60 and 1/24***. No change: Narcotics. EDB (<33%). LSCS (<8%)	Injecting SW over the sacrum decreases low back pain in labour for at least 1h. Transient pain with administration may impact on maternal desire to repeat	Increased maternal feelings of control, but no change in maternal satisfaction with labour. Decreased maternal desire to use again in future labours
	Martensson and Wallin,10 Sweden, N=99, SG1 and 2 vs. CG	<LBP per VAS at 10/60*** and 45/60**. <LBP per midwife (blinded)***. No change: LSCS (3%). Pain SG1vs. 2	Similar decrease in low back pain between s/c 0.5ml vs. i/d 0.1ml administration. Both s/c and i/d methods cause transient injection site pain	Increased maternal desire to repeat in future labours SG1 and SG2 vs. CG***
	Bahasadri et al.,12 Iran, N=100	<LBP per Faces Rating Scale at 10 and 45/60**. No change: LSCS (4%). Instrument	Injecting at only one site over the sacrum decreases low back pain in labour for at least 45min	Maternal satisfaction not assessed
	Wiruchpongsanon,13 Thailand, N=50	<LBP per VAS at 30/60** 1/24* and 2/24**. <LSCS (0% vs. 12%)	Injecting SW over the sacrum decreases low back pain in labour for up to 2h and may decrease LSCS rate	Maternal satisfaction not assessed

R, randomised; RCT, randomised controlled trial; N, number; >, increased; <, decreased; SW, sterile water; NS, normal saline; i/c, intracutaneous; s/c, subcutaneous; i/d, intradermal; SG, study group; CG, control group; EDB, epidural block; TENS, transcutaneous electrical nerve stimulation; VAS, visual analogue scale for rating pain, Faces Rating Scale, for rating pain; LBP, low back pain; LSCS, lower segment caesarean section; +, strengths; −, limitations. Statistical significance criteria; *p<0.05; **p<0.01; ***p<0.001. Method: 4× injections: 1 over each posterior superior iliac spine, plus 1 injection 3cm below and 1cm medial to each of the first two injections.

Results 

Comparison of results from these six studies elicited many statistically significant findings. Whilst participant characteristics were well controlled in areas of age, gestation, weight, cervical effacement and dilatation, intact/ruptured membranes and even education, Ader et al.8 also compared the length of, and duration between, contractions; and Labrecque et al.11 asked women whether back pain was worse than abdominal pain, and expressed this as a percentage.

Most studies found that the pain associated with administration of the injections of sterile water lasted approximately 2min, during which time a burning sensation was noted. Minimal pain was reported from administration of the placebo normal saline; in fact Martensson and Wallin10 showed a VAS pain score of treatment administration to be >7cm from that of placebo administration (p<0.001) which is statistically significant and may explain why the women in the Canadian11 study were significantly more reluctant to repeat this treatment in future labours, despite the obvious reductions in pain intensity they received from it. Conversely, Trolle et al.9 found that a significantly greater number of women would utilise this treatment again in future labours (p<0.0005).

Perhaps the most notable results derived from these studies are the statistically significant findings of reduction in low back pain during labour at intervals of:

•10min (p<0.001)8, 10, 11, 12

•30min (p<0.01)13

•45min (p<0.01),10, 12 (p<0.05)8

•1h (p<0.001),9, 11 (p<0.05)13

•90min (p<0.05)8

•2h (p<0.001),9 (p<0.01)13

The midwives’ blinded assessment of the treatment’s effectiveness revealed a significant difference in favour of the sterile water group over the placebo in two of the randomised controlled trials.8, 10 All studies except Bahasadri et al.12 reported losing women to birthing throughout their studies, but not in numbers great enough to compromise the validity of their findings.

Interestingly, the largest randomised controlled trial conducted in Denmark by Trolle et al.9 of 272 women found that 89% of women experienced an analgesic effect from the sterile water injections when compared with the placebo group at 45% (p<0.0005). She also found that the caesarean birth rate was statistically significantly higher in the placebo group than the sterile water treatment group (p<0.05), with 11% vs. 4%. She believed that this finding may be attributable to a “higher incidence of cephalopelvic disproportion and malposition of the occiput.” (p. 1279). No other study made this finding but Wiruchpongsanon13 did find that there were no caesarean births in his treatment group but 12% in the placebo group; however, it was not of sufficient statistical significance (p=0.235). Labrecque et al.11 concurs with no caesarean births in the sterile water group vs. 33% in the TENS group and 8% in the usual care group (p=0.10).

The approach of Labrecque et al.11 was unique but fruitful in rating the intensity and unpleasantness of pain, finding that this was significantly lower in the sterile water group when compared to the usual care or TENS groups. Similarly, the comparison of the last pain rating collected prior to birth or epidural block showed significant differences for intensity and unpleasantness. A decrease in pain intensity was rapidly achieved in the sterile water group and maintained throughout the first hour of intervention; however, after this first hour there remained too few women for a reliable estimation of effect. Notably all pain rating differences between these groups were of statistical significance. A criticism of the Iranian study’s12 graphical representation of results from the Faces Rating Scale would lie in the omission to portray the scale from 0 to 5 as it is when presented to the women for assessment. The graphs in this article are shown from 1 to 5.

Discussion 

Critical review and synthesis regarding quality of the studies reviewed and their results have afforded recognition of perceived strengths and limitations by the reviewer. Without exception, all conclusions were evidence-based, eliciting statistically significant results which are of contemporary clinical significance to midwives and birthing women.

All studies provide a universal agreement that the origin of uterine contraction pain which is experienced in the low back area is commonly attributable to the baby’s occipito-posterior position, which causes stretching of the posterior part of the lower uterine segment to be more predominant. Bonica20 states that pain travels along a pathway of impulses from the posterior cervix and uterus which is then transmitted to spinal cord nerves at the thoraco-lumbar level T11-L1. Cutaneous branches from these segments supply the nervous innervations over the low back or lumbo-sacral area at L3-S2. This appears to explain how pain which originates in the cervix can be localised to the back and experienced as referred pain.

Ader et al.8 did not attempt to define the mechanism of action more exactly. Consequently neither did Wiruchpongsanon,13 whose whole discussion is obviously derivative of the earlier Swedish study8 with no credit to her work cited in his references. Wiruchpongsanon13 also failed to relate his study back to the original question and determine the efficacy for use particularly in Thai women.

Interruption of the pathway of pain is the desired action of sterile water injections with Trolle et al.9 first suggesting the area of Michaelis’ Rhomboid as the recommended site for injections because, subjectively, this is the area where the pain is felt most acutely by labouring women. When the sterile water is injected under the skin it raises a small bleb or “papule” which causes local irritation and a strong sensory stimulation of the surrounding skin nociceptors for about 30s. A hyperaemic zone is observed around the bleb for some hours afterwards demonstrating a prolonged irritation of the cutis. The analgesia induced by this stimulation may be caused by gate control at the spinal level21 and seems to be an important factor for the treatment to be effective, but this area provokes conjecture in the literature. Bahasadri et al.12 concurs that the gate control theory may provide an explanation for the mode of action of sterile water injections but because the inhibition of pain is not restricted to one specific segment, he suggests that there must be a more non-specific modulation of pain. Ader et al.8 speculates that the analgesic effect is similar to the stimulation of endogenous opioids seen with the use of acupuncture or TENS in labour and may act as “a long lasting segmental acupuncture” (p. 574). The Danish perspective from Trolle et al.9 tells us that hyperstimulation of a skin area can affect perception of visceral pain and it appears that “the mechanism of referred pain can be reversed to produce referred analgesia” (p. 1280).

Counterirritation was described by Melzack22 as the phenomenon of one painful stimulus reducing the pain caused by another noxious stimulus and may explain the pain-reducing effect of both sterile water injections and acupuncture. The Swedish10 and later Iranian studies12 discussed the concept of diffuse noxious inhibitory control (DNIC) which is a physiological mechanism produced to explain the effects of counterirritation.23 DNIC is the “inhibition of multireceptive neurons in the dorsal horn of the spinal cord, when a noxious stimulus is applied to a region of the body remote from the neurone’s excitatory receptive field” (p. 105).12 This ultimately means that pain is reduced in areas remote from those where stimuli are present, thus supporting the use of sterile water injections during labour. Administration pain associated with the sterile water injections proved to be problematic for the researchers. Despite providing significant reductions in pain levels, many women stated they were reluctant to repeat this treatment in future labours due to the transient sharp stinging sensation.11 Several researchers tried to modify the administration technique. Martensson and Wallin10 argue that according to the concept of DNIC,23 it can be assumed that an intense stimulation, such as that obtained from intracutaneous sterile water injections, providing both osmotic stimulation from the salt-free water and distension of the firm cutaneous layers, would be more effective than subcutaneous injections which merely induce osmotic stimulation. However, they found that the two active treatments did not differ in effectiveness. Therefore, the supposed beneficial effect of subcutaneous administration being less painful than intracutaneous administration was not shown in this study. The12 Iranian study cited the findings of Martensson et al.24 where she did find that the subcutaneous route of administration proved to be less painful and more tolerable than intracutaneous. This prompted him to trial a single subcutaneous injection over the most painful point of the sacrum in order to reduce the pain associated with administration. A limitation of this study was that it did not compare the pain reducing potency of one vs. four injections, or the pain associated with subcutaneous vs. intracutaneous administration. Attention to these questions would have significantly enhanced his study, by providing a comparison of the subjective experience of administration pain.

Trolle et al.’s9 Danish study identified that the analgesic effect of sterile water injections was not connected with any impairment of labour progress. Although the incidence of labour dystocia and oxytocin augmentation was high in both groups this was not an unexpected finding because only women with low back pain entered the study and low back pain commonly occurs with malrotation of the vertex, which often requires augmentation. Coincidentally it was found that although the rate of caesarean birth in the placebo group was comparable to the usual rate of caesarean birth in that hospital, the rate of caesarean birth in the sterile water group was statistically significantly lower. This appears to suggest that the sterile water injections may have an effect on the relaxation of pelvic and cervical tension, which is required for correcting a malrotated presentation.

The validity of placebo could be argued in these randomised controlled trials, with the administration of the placebo normal saline being less painful to administer, but not completely painless, theoretically inducing some degree of analgesia. In order to make these studies double blind some kind of injection into the low back area had to be used in the control group to ensure the placebo effect was not lost. Since normal saline injections are almost painless it could be argued that women knew they had received placebo treatment. However, Bahasadri et al.12 stated that women were not told how the different kinds of injections would be experienced during administration; thus, they should presumably be unable to judge whether they had received active treatment or placebo. Martensson and Wallin10 found that whilst placebo produced “neither osmotic nor distension pain,” it was shown to have an analgesic potency. Further, they state that a considerable risk is implied with overestimating the potential beneficial effect of active treatment as it must be compared with placebo. “If the placebo injections possess minor analgesic power in themselves in addition to that of placebo then the differences between a true placebo treatment and active treatment is even greater than shown in this study.” (p. 636).

Conclusion 

The studies reviewed consistently demonstrated that sterile water injections induce a statistically significant, dramatic analgesic effect on the low back pain experienced by women during labour; lasting from 10min8 and up to 2h13 post-administration. Sterile water injections have proved to be a justifiable alternative to the use of narcotics for birthing women and their midwives who are concerned about unwanted side effects for both mother and baby. Their effect has been described as powerful, rapid and effective; with the potential to decrease or delay the use of epidural anaesthesia. The analgesia produced from one injection of sterile water compared favourably to that obtained from four injections, further research is required to determine whether the four injection technique is warranted, particularly in terms of resultant patient satisfaction and willingness to repeat the treatment in future.

Placebo treatment also had an analgesic effect which was statistically significantly more pronounced in the sterile water group than in the normal saline group; which indicates that it is both the act of injecting the sterile water and the use of the sterile water itself that induces analgesia. A summary of the proven benefits is outlined in Table 2.

Table 2. Summary of proven benefits of sterile water injections during labour

	Sterile water injections:
	Provide rapid and effective low back pain relief during labour
	Have no apparent side effects
	Offer simplicity of use and a high level of success
	Are non-pharmacological
	Decrease the need for epidural anaesthesia
	Delay the use of epidural anaesthesia
	May be used whilst waiting for an anaeasthetist
	May decrease the caesarean birth rate
	May be used in first and second stage labour
	Have applications for use in rural/remote areas and developing countries where alternative treatments are not available
	Have a role to play with their analgesic effect on pelvic floor tone, cervical tension and fetal rotation

Although we do not clearly understand the exact mode of action of sterile water injections, the proof of relief found consistently throughout different studies makes their use for analgesia in labour a justifiable option. If this technique is to succeed as an innovation in Australian birthing services then it must be championed by midwives for women; with its simplicity, cost-effectiveness and high level of success demonstrating clear therapeutic merit. The opportunity to relieve or reduce the amount of pain and suffering experienced by women during the birthing process has long been the goal of midwives, and sterile water injections clearly provide a tool to safely and effectively achieve this end.