The effectiveness and safety of complementary health approaches to managing postpartum pain: A systematic review and meta-analysis.

BACKGROUND: Women experience pain from a number of causes during the postpartum period. Although pharmacological pain relief has shown to be effective, the efficacy of non-pharmacological methods of pain relief will be of interest to breastfeeding women. The aim of this systematic review was to examine the efficacy and safety of complementary approaches to manage postpartum pain.
METHODS: A search of English language databases from their inception to 2020 was undertaken for randomised controlled trials and included primiparous and multiparous women who experienced postpartum pain up to two weeks post birth. The primary outcome was pain. The risk of bias was assessed using the Cochrane risk of bias tool.
RESULTS: Thirty trials were included in the review, 25 trials (2,413 women) were included in the meta-analysis. Two trials of massage found a reduction in pain following caesarean birth within the first 24 h post birth (MD -2.64, 95-2.82 to -2.46, 184 women, I2 0%), and at seven days postpartum (MD -1.91, 95%CI -2.42 to -1.40, 2 trials, 120 women I2 37%). Two trials conducted with women receiving an episiotomy found reduction in perineal pain from herbal ointments within 24 h (MD -1.33, 95% CI -.96 to -0.70, 221 women) and at 14 days postpartum (MD -0.74, 95% CI -1.02 to -0.47, 4 trials). Few trials reported on safety, few trials were at an overall low risk of bias, and overall the quality of evidence was very low.
CONCLUSION: Further high quality trials are needed to determine the safety and effectiveness of herbal ointment and massage during the early postpartum period.

Introduction

Pain relating to pregnancy and childbirth can have a significant impact on women during the postpartum period. Women may experience perineal pain, breast pain, low back pain, pain from uterine involution and incisional pain post caesarean Section 1.1 Experience of postpartum pain varies depending upon individual circumstance, mode of birth, previous conditions, level of psychosocial support and available options for pain management. Postpartum pain can be significant with potential detrimental impacts on maternal movement, mobility, sleep and mental health which may interrupt a woman's transition to motherhood.

Much of the research on complementary health approaches for childbearing women has focused to date on labour pain and techniques to reduce pain and minimise the need for pharmacological options. The ‘working with pain’ model has been the dominant theory when it comes to labour and birth but there is less focus on pain following the birth. The postpartum period is a time when the woman's body is recovering from birth physically and adapting physiologically to a non-pregnant state. This is happening at the same time that the woman is caring for a newborn baby, experiencing interruptions to sleep, changes in hormones and learning how to establish breastfeeding. As a result, there are physiological and psychosocial changes impacting on how pain may be perceived and responded to. Pain associated with the birth, such as perineal trauma, back pain and breast pain tend to be most intense in the first couple of weeks unless there is some unresolved pathology underlying the ongoing pain.

Whilst pharmacological pain relief may be effective, consideration needs to be given to use in women who may be breastfeeding. It is therefore essential that effective and safe pain management options and alternatives to mainstream medical treatments are available to women during the postpartum period.

Complementary health approaches are practised to maintain health and wellbeing and are used with conventional medicine. They include a wide range of practices including biological based therapies, traditional herbal medicines, and non-biologically-based therapies. These practices are regularly used by women during pregnancy, and in Australia, data indicate these modalities are widely used to manage pain in labour. Evidence of effectiveness from these therapies during labour show a small benefit from acupuncture and acupressure, relaxation, yoga and music, and massage, warm packs and thermal massage. The aim of this systematic review is to examine the effectiveness and safety of complementary health approaches to manage perineal pain, low back pain and postoperative caesarean pain.

Methods

Study selection and eligibility

We included primiparous and multiparous women who experienced perineal pain, low back pain and postoperative caesarean pain. Breast pain, nipple pain and chronic back pain were excluded as there is extensive literature available on breast/nipple pain and treatments and chronic back pain may have other causes than pregnancy hormones and the birth process. Studies were eligible if women had given birth by normal vaginal birth, instrumental birth or caesarean section and were breast or formula feeding. The care setting was unrestricted.

We included parallel randomised, cross over or quasi randomised controlled trials. Non randomised studies were excluded. Due to limited resources, studies not published in English were also excluded.

Types of interventions

The interventions evaluated in this review included non-pharmacological approaches used alone or as an adjunct to conventional treatment. Complementary health approaches included the following categories; mind and body interventions, natural product-based therapies, manipulative and body-based practices, energy therapies and alternative medical systems. Operational definitions were inclusive, utilising the Weiland et al., definitions for complementary and alternative therapies. Cryotherapy for perineal pain was excluded as it was considered a recommended routine intervention. Cognitive psychological interventions were excluded from the review. Interventions were included that commenced during labour or prior to the onset where the aim was to reduce pain postnatally.

Outcomes

The primary outcome was a reduction in self-reported post-partum pain. Secondary maternal outcomes included: maternal satisfaction, change in wound appearance, maternal quality of life, adverse events for the mother, use of pharmacological pain relief and duration of use, and full or partial breastfeeding.

Secondary outcomes were focussed on the infant and included: adverse effects (lethargy, reduced feeding, jaundice), delayed maternal/infant bonding, and maternal infant hospital admission.

Database search

AD and EH searched the following databases, CINAHL, Pubmed, Nursing & Allied Health, Johanna Briggs Institute, Embase, Cochrane Library and Ovid Medline from the inception of the databases to 27 July 2020. The grey literature was searched including conference proceedings, Google Scholar and reference lists of primary and reviewed articles. Each database search was performed utilising the terms postpartum, post-partum, postnatal, post-natal and pain and ‘yoga’ OR ‘hypnosis’ OR ‘hypnotherapy’ OR‘ meditation’ OR ‘aromatherapy’ OR ‘tai chi’ OR ‘rescue remedy’ OR ‘hydrotherapy’ OR ‘nutrition therapy’ OR ‘diet therapies’ OR ‘alexander technique’ OR chiropractic’ OR ‘osteopathic manipulation’ OR ‘massage’ OR ‘reflexology’ OR ‘healing touch’ OR ‘movement therapies’ OR ‘herbs’ OR ‘probiotics’ OR ‘vitamins’ OR ‘minerals’ OR ‘acupuncture’ OR ‘acupressure’ OR ‘breathing exercises’ OR ‘moxibustion’ OR ‘qi gong’ OR ‘magnetic therapy’ OR ‘therapeutic touch’ OR ‘reiki therapy’ OR “ultrasonic therapy’ OR ‘ayurvedic medicine’ OR ‘Chinese/Japanese/Tibetan traditional medicine’ OR ‘homeopathy’ OR ‘naturopathy’.

Retrieved articles were reviewed for further sources. No date restrictions were applied. Three authors retrieved the studies and assessed the study eligibility (EH, AD, CS). Data was extracted and assessed by two review authors (EH, AD). The following characteristics were retrieved for each study; country, setting, study design, inclusion criteria, description of the intervention and control groups, and outcomes.

Quality assessment was assessed independently by two authors (CS, AD) using the Cochrane risk of bias tool. Any disagreements were resolved by a third author (HD). The quality of the evidence was assessed using the Grade of Recommendation, Assessment, Development and Evaluation (GRADE) criteria. GRADE was applied to outcomes in the meta-analysis that included more than one included trial, and the evidence downgraded following an assessment of study limitations.

Data synthesis

Data was presented as risk ratios (RR) or a mean difference (MD) with 95% confidence intervals (CI) for the primary and secondary outcomes. We used the standardised mean difference (SMD) if outcomes were the same but used different methods. Data were pooled using Review Manager version 5.4, and effect sizes with a 95% confidence interval were reported. Statistical heterogeneity was assessed for each meta-analysis using the I2 and Chi square statistics. Heterogeneity was rated at substantial if the I2 statistic was greater than 30%. Where substantial heterogeneity was assessed a random effects model was applied. If we could not pool data a narrative reporting of the results was undertaken. For cross over trials we planned to analyse data from the first phase only before groups crossed into the second phase of the study.

Results

A total of 236 studies were identified through database searching (Fig. 1). A further nine studies were identified through other sources such as the reference review on retrieved studies. After duplicates were removed, a total of 128 studies underwent an abstract and title screen, which resulted in 80 articles being excluded due to failure to meet established inclusion criteria, and a further 20 studies were excluded following a full article assessment for eligibility. Of these, eighteen studies were excluded, this included ten studies that were not RCTs,,15, 16, 17, 18, 19, 20, 21, 22, 23 seven studies examined interventions outside the scope of the review,24, 25, 26, 27, 28, 29, 30 no outcomes were reported in the two week postpartum period in one study.

Fig. 1

Study flow diagram.

Study number and design

Thirty studies were included, and the characteristics are shown in Table 1. Seventeen studies were from Iran,32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48 and two each from Brazil,, England,, India,, and one each from Turkey, Croatia, Hong Kong, Taiwan, Denmark, Egypt, and the United States.

Table 1.

Characteristics of included studies.

Author	Study Design	Country setting	Participants	Intervention	Control	Outcomes
Essential Oils
Behmanesh 37	RCT double blind	IranSetting hospital	90 primiparous women post episiotomy or perineal tear grade 2.	Lavender essential oil 2% based olive oil sitz bath. 10 drops of essential oil distilled in 5 L of warm water. Used twice daily for 10 days. Pain relief available to all.	Group 2: olive oil and Group 3 distilled water sitz bath. 5 L of warm water.Pain relief available to all.	Pain VAS, perineal redness, oedema, echymosis, discharge assessed using REEDA at 2 h, 5 and 10 days
Olapour 43	RCT double blind	Iran:Setting: delivery suite	60 pregnant women, admitted for planned caesarean. Postpartum with VAS score > 3.	Aromatherapy inhalation occurred at 4, 8, and 12 h. Three drops of aromatherapy blend containing lavender essence 10% applied to cotton in cast containers and the woman was asked to inhale it for five minutes from a distance of 10 cm. Routine medication available.	Using the same procedure as in the intervention group, three drops of placebo oil applied.	Pain severity measured using the VAS. Satisfaction at 4, 8, and 12 h.
Vaziri48	RCT double blind	Iran	56 postpartum primiparous women post episiotomy with pain ≥to 4 on the VAS	1% lavender essential oil utilised. 5 drops were placed on a cotton ball an inhaled about 20 cm from their nose for 10–15 min. Intervention applied 3 times in 24 h post birth. Analgesia provided.	Sesame oil used as a placebo.Routine care including analgesia	Pain using VAS assessed at the end of the intervention.
Dale52	RCT double blind	United KingdomSetting postnatal ward	386 primiparous and multiparous women, vaginal birth.	1. Extract of lavender oil: 6 drops as bath additive.2. Synthetic laboratory produced oil: 6 drops synthetic laboratory produced oil as bath additive. Daily for 10 days.	6 drops of inert compound. Placebo had odour similar odour to the intervention groups.	Pain using VAS, mood, analgesia use, bruising, infection number of days to heal. Collected until day 10 postpartum.
Sheikhan46	RCT single blind	IranSetting: postnatal ward	60 primiparous women with episiotomy	Lavender oil essence with a concentration of 0.96%; linalyl acetate 20% was extracted and made soluble. 30 min sitz bath (+25 ml lavender oil essence per 5 L of water) twice a day for 5 days.	Routine sitz bath care 30 min (+10 ml betadine per 4 L of water) twice a day for 5 days.	Pain using VAS. Wound healing using REEDA scale. Assessed 4, 12 h and 5 days. Analgesic use
Vakilian47	RCT single blind	IranSetting: unclear	120 primiparous women with episiotomy.	1.5% lavender essential oil was prepared in olive oil as a carrier A sitz bath using 5–7 drops of essential lavender oil in 4 SSL of water twice a day for 10 days.	Routine care. Povidone-iodine to sitz bath	Pain using VAS, wound healing using measurements of size on day 10.
Marzouk60	double blind RCT	EgyptSetting labour ward	69 primiparous women with episiotomy	Lavender flowers essence applied to jojoba oil as a carrier. Women were instructed to add 7 drops of the oil to 4 L of warm tap water. The incision was cleaned by directing the nozzle of a bottle towards the incision site twice daily for 7 days.	Placebo essential oil	Pain using VAS, wound healing using REEDA analgesic intake. Assessed on the 7th day postpartum
Acupressure
Afravi33	RCTdouble blind	IranSetting:	62 multiparous women post vaginal birth, complaining of medium to severe postpartumpain	Acupressure to point Hugu LI4 over 20 min. Pressure applied for 1 min, pressure released and repeated.	A sham acupressure point applied to the hand	Pain intensity assessed using the VAS at 2, 4, 6, 8 h post-partum.Use of analgesics.
Akbarzade34	RCT single blind	Iran, Setting: hospital	150 postpartum women, 18–40 years, 4–8 post birth	Dry cupping therapy or acupressure at Shenshu BL23 for 15–20 min up to three times every other day for 3 weeks. The pressure continued for 20 min.	Routine care	Pain using short form Magill questionnaire at 24 h post-partum and 2 weeks postpartum
Kwan57	RCTdouble blind	Hong KongSetting: labour ward	256 women both primiparous or multiparous with singleton women with episiotomy or first or second degree of perineal laceration.	Vaccaris segatalis garcke seeds were taped to acupuncture points on both ears (shenmen point, the point of external genitalia and the anus). Women were instructed to press 30 s onto each of the seeds every 4 h while awake.The intervention commenced during perineal suturing.	Tape applied to the same acupuncture points but with no seeds attached.All women had access to pain relief.	Pain using VAS, analgesia intake, maternal views.Assessments at 2 h post-birth and 12, 18, 24, 26 and 48 h post-birth.
Acupuncture
Jaic 56	RCTsingle blind	CroatiaSetting: hospital	60 primiparous multiparous women post episiotomy, ≥18 years of age, ≥ 36 weeks gestation.	Ear acupuncture consisted of three acupuncture points internal genital area, external genital area and Shen Men point. Needles inserted 6-8 h post-birth by certified practitioner. Needles retained for 3 days. Oral analgesics supplied on patients request.	Oral Analgesics (NSAID) supplied on patients request.	Pain using VASneed for analgesics.3 days postpartum
Kim 61	RCT single blind	USASetting: hospital	70 women at least 6 h post vaginal birth, pain score ≥4.	Battle-field Acupuncture (BFA) and standard analgesia. Qualified physician placed semi-permanent acupuncture needles into 5 bilateral acupuncture points Shenmen, Point Zero, Omega 2, Thalamus, Cingulate Gyrus.	Standard analgesia alone	Pain using the NRS, use of analgesia up to 10 days postpartum
Kindberg59	RCT single blind	Denmark	207 primiparous women a spontaneous or instrumental birth using ventouse after 36 weeks gestation.	Ear acupuncture during postpartum perineal repair. Two 15 mm needles placed on the ear plus points (Shenmen, Genital, Chengfu Bl36 and Baihui GV20 point. Manual stimulation of the needles to evoke needle sensation (deQi).	Local anaesthetics: Lidocaine 10 mg/ml applied directly into the wound.	Pain using VAS at 24–48 h and 14 days postpartum.Analgesia intake. Maternal views on the intervention.Wound healing used the REEDA scale
Music
Reza 44	RCT double blind	IranSetting: delivery suite	100 womenscheduled for elective caesarean section under generalanaesthesia,	Routineanaesthesia.Spanish guitar music therapy played through earphones to women during caesarean up until the end of wound dressing	White music	Post-operative pain. Assessed by VAS immediately in at 0.5, 1, 2, 4 & 6 h postop.
Simavli 55	RCT single blind	TurkeySetting: hospital clinic	161 primiparous women aged 18– 35 years, 37–41 weeks gestation	Music: pre-recorded CD listen to 30 min/day for two weeksParticipants choose types of music. Music commenced >2 cm cervical dilatation until the end of the third stage.	Routine prenatal care.	Pain (VAS), satisfaction (VAS) recorded at 1, 4, 8, 16, 24 h postpartum
Massage
Abbaspoor 32	RCTsingle blind	IranSetting:hospital ward	80 multiparous women post elective C/S; 18–35 years; 37–42 weeks gestation; previous C/S.	Foot and hand massage included petrissage, kneading, and friction applied to the patients' hands and feet. Massage initiated 1.5–2 h after spinal anaesthesia medication. Hand massage applied to each hand for 5 min. Foot massage followed.	Nurse talked to women for 20 min.Analgesics available for both groups	Pain: using NRS, 90 min after the intervention, use of medication
Saatsaz 45	RCT single blind	IranSetting postnatal ward	156 primiparous women undergoing caesarean section.	Group 1: Massage therapy included petrissage, kneading and friction to the target area with the anterior surface of the last phalanx or the palm. Massage applied to hands and feet for 5 min.Group 2: Foot massage only. Neither groups received oral analgesia during the 90 min duration of the assessment.	Routine care	Pain measured using VAS 90 min post intervention and 4 h after administration of the last dose on analgesia.
Sharma 53	RCT single blind	IndiaSetting: postnatal ward	60 postnatal mothers who had undergone planned or emergency C/S in the previous 24 h with moderate to severe pain.	Foot and hand massage was given for 20 min (5 min in both upper and lower extremities) after 4 h of the analgesic. Massage procedure: (Effleurage, Friction, Petrissage) applies squeezes and strokes. Each hand and foot massaged for five minutes. Repeated over a 3 day period.	Routine care.Analgesics available to both groups	Pain using NRS at day 1, 2 and 3 postnatal.
Lee58	RCT single blind	TaiwanSetting: post-partum bed rest centre	60 primiparous or multiparous women with LBP symptoms defined by a score of ≥1 on VAS scale post vaginal birth	Massage therapist administered 20 minute back massage involving effleurage, kneading, acupressure and friction. The session conducted once every evening for 5 consecutive days (postpartum days 9–13)	Routine care	Lower back pain assessed by the PVAS at 5 and 14 days postpartum
Herbal ointment
Aradmehr 35	RCT double blind	IranSetting hospital	114 primiparous women, aged 18–35 years, post episiotomy	Chamomile cream 1.3% combined with cold cream was provided to participants 2 h after episiotomy recovery applied twice a day for 10 days.	Cold-cream containing no chamomile	Pain intensity using VAS and use of medication 2 h and day 1, 7 and 14.
Asgharikhatooni36	RCT double blind	IranSetting hospital	108 nulliparous mothers, aged 18–35 years, post episiotomy	Equisetum Arvanse (horse tail) herbal ointment applied topically twice a day at 12 h intervals for 10 days.	Placebo ointment. Each participant was provided with 20 acetaminophen 500 mg pills	Pain intensity measured using the VAS, wound healing, side effects and medication used
Hajhashemi39	RCT double blind	IranSetting hospital;	140 nulliparous women with an episiotomy	Group 1 Hypericum perforatum infused into vaseline (5% weight ratio).Group 2. Achillea millefollium with anti inflammatory effects, extracts packed with sterile vaseline base (5% weight ratio). Both groups administered using 30 g tubes, 1 cm of ointment applied to perineum. Treatment twice a day for 10 days.	Group 3 vaseline ointment with no extract. Group 4 Control group - no intervention	Pain using VAS, redness, oedema, ecchymosis, wound dehiscence and wound secretion. Outcomes assessed on days 7, 10 and 14
Kaviani40	RCT double blind	IranSetting hospital	90 primiparous women following episiotomy.	Group1 Olive leaf extract. 1 g of ointment applied onto the suture area every 8 h for 10 days post birth.	Placebo ointment	Pain using VAS, wound healing using REEDA scale 4 h, 3, 7, 10, and 14 days post birth
Mohammadi41	RCTdouble blind	IranSetting delivery ward	114 multiparous postpartum women vaginal birth with episiotomy	Cinnamon extract ointment applied twice daily at 12 h intervals (+2 h for 10 days). Participants were provided 10 400mg mefenamic acid capsules	Placebo ointment applied in the same manner as the intervention group	Pain measured using the VAS. Wound healing using the REEDA, analgesic use Data recorded 1 h and 8 h after perineal repair 10–11 days postpartum.
Moudi42	RCT single blind	IranSetting postnatal ward	147 primiparous women following an episiotomy	15 gm of mastic oleoresin provided to women to administer twice a day over 3 days. Women lay the MO on hot coal. Standing with legs apart and knees slightly bent to a distance of 30 cm from a brazier to smoke the episiotomy incision Mefenamic and cephalexin pills administered during the first three and 7 days postpartum	Routine care	Perineal pain using the VAS and wound healing using REEDA scale. Assesses postpartum day 3,7 and 10.
Laser
Alvarenga49	RCT double blind	BrazilSetting: postnatal ward	54 primiparous postpartum women, 37–42 weeks, following episiotomy. Invited 6–10 h post birth	Low-level laser therapy applied to specific points of the episiotomy. Irradiation time was 90 s. Three sessions with intervals of 24 h before discharge (1st, 2nd and 3rd sessions carried out at 6–10 h postpartum and 40–48 h after the first session)	Same procedure but with no irradiation. Routine care	Perineal pain NRS, wound healing, maternal views on pain relief. Assessment at 7–10 days post birth.
De Santos50	double blind RCT	BrazilSetting: postnatal rest unit	114 primiparous women with mediolateral episiotomy with perineal pain ‡3 on a numeric scale of 1–10	Three sessions of irradiation with laser therapy. Laser performed by touching the tip of the device to the incision for 10 s at three points of the episiotomy (central, upper and lower portions).Group 1 Red. Group 2 Infrared.	No active laser therapy unit applied.All groups 24 h post-birth before intervention commenced	Perineal pain assessed before and immediately after and 30 min after application of intervention.
Reflexology
Varghese54	RCT single blind	India	60 post-caesarean section women 1st–3rd post-operative Parity not specified.	Feet warmed and soaked for 10 min. Therapist applied thumb and finger walking to the reflexology zones of the feet relating to the brain, eye, adrenal gland, respiratory organ, spine, circulating systems and lower abdomen. 15 minute session applied for 5 consecutive days.	Routine post-partum care.	Pain using the VAS and sleep quality using the Pittsburgh sleep quality index. Assessed at day 5.
Herbal tablet
Golezar38	RCT double blind	IranSetting:	82 primiparous women, with episiotomy	Bromelain 3 tablets daily for 6 days. Derivative from pineapple.	Placebo	Pain (VAS), days 3, 7, and 14 post-partum, wound healing assessed using REEDA scale.
Electro-magnetic therapy
Grant51	RCT double blind	United Kingdom Setting	414 women with operative vaginal birth; or perineal trauma within 24 h of birth.	Group 1: Pulsed electromagnetic energy intervention at frequency of 27MHz, pulse repeat rate of 100 pulses, pulse width of 65us, pulse ratio normal. Three treatments of 10 min duration. Therapy commenced within 12-24 h post-birth. Maximum of three treatments in a 36 h period.Group 2: Ultrasound therapy intervention also administered.	Placebo therapy	Pain, wound healing, analgesia use. VAS 0-10 scale. Midwife assessed wound healing. Assessments pre and post treatment and at day 10.

Description of interventions

There were four studies using massage,,,, seven trials evaluating aromatherapy,,,46, 47, 48,, six trials that applied a topical application of a herbal ointment,,,39, 40, 41, 42 and acupuncture and acupressure,,,,,, two studies each of music,, and laser therapy,, and one trial each respectively of electro-magnetic therapy, reflexology, and a nutritional supplement.

The intervention was administered during labour in one study , and one trial applied music during caesarean section , with five interventions administered post caesarean section.,,,, A further 23 trials were administered post vaginal birth 33, 34, 35, 36, 37, 38, 39, 40, 41, 42,46, 47, 48, 49, 50, 51, 52,55, 56, 57, 58,,.

Pain was assessed in all studies; the majority using a Visual Analogue Scale (VAS).

Women's views’ of pain relief or satisfaction with pain relief was assessed in six studies.,,,,, Wound healing was assessed in 11 trials37, 38, 39, 40, 41, 42,,,,,, and most frequently used the redness, oedema, ecchymosis, discharge (REEDA) scale. Use of medication was assessed in 14 trials,,,,,,,,,,,,59, 60, 61 and side effects assessed in two trials.,

Outcomes were assessed within 24 h post birth or application of the intervention in seven studies.32, 33, 34,,,, An additional six trials collected outcomes up to and including day seven postpartum,,55, 56, 57, and 16 trials assessed pain up to two weeks postpartum.35, 36, 37, 38, 39, 40, 41, 42,,,,,,,, It is worth noting that all studies predominantly focused on effectiveness of pharmacological pain relief with little focus on safety during breastfeeding.

Participant parity and mode of birth

Twenty-six studies reported details of the women's previous pregnancy and childbirth. Fourteen studies recruited primiparous women only,,,,,,45, 46, 47, 48, 49, 50,,, and five studies recruited multiparous and primiparous women.,,56, 57, 58 Three studies recruited multiparous women only.,, No details were reported in six trials.,,,,, The intervention was performed on women following a caesarean section in six trials,,43, 44, 45,, post vaginal birth in seven trials,,,,,,, and specifically post episiotomy in 16 trials.35, 36, 37, 38, 39, 40, 41, 42,46, 47, 48, 49, 50,,,

Risk of bias

The overall risk of bias for the included trials was variable across trials (Fig. 2). No single trial was at a low risk of bias although three trials were at a low risk of bias on six of the seven domains, but insufficient reporting resulted in an unclear risk of bias for this single domain for these three trials. Eleven trials were assessed at a low risk of bias on both randomisation domains.,,,,,,,,,, Most trials were not designed as a placebo controlled. Eleven trials were assessed at low risk on both blinding domains.,,,,,48, 49, 50, 51,, Performance bias assessed as low risk in 13 trials.,,,,,,,48, 49, 50, 51,, Twenty five trials had a low risk of bias for attrition with two trials assessed at high risk., For the majority of trials, we assessed the risk of selective reporting to be unclear due to insufficient reporting. The other sources of bias were rated at low risk in 21 trials, with one trial at high risk, and eight trials where the risk was unclear.,,,52, 53, 54,,

Fig. 2

Risk of bias assessment.

Effect of the interventions

Twenty-five trials and 2413 women were included in the meta-analysis reporting on pain outcomes. For this outcome there was significant heterogeneity and a random effects model was applied, different scales were used for some interventions and the standard mean difference was calculated. Data from five trials that was not in a form to be included in the meta-analysis and is reported narratively.

Essential oils

Primary outcome

Four trials,,, of essential oils (lavender or a blend) reported on pain (site specific to mode of birth) within 24 h of being applied, and found no difference between groups (MD 0.33, 95% CI -0.71 to 1.36, 423 women, I2 83%). There was also no difference in pain between groups at 7 days (MD 0.76, 95% CI -1.9 to 3.41, 3 trials,, 368 women), and at 8 to 14 days (MD 0.22, 95% CI -0.22 to 0.67, 2 trials,, 296 women, I2 16%).

Secondary outcomes

Wound healing to the perineum assessed by the REEDA scale demonstrated no difference between groups when assessed at 24 h post intervention. There was significant heterogeneity and a random effects model was applied (MD -0.92, 95% CI -2.52 to 0.68, 120 women, two trials,, I2 94%). There was no effect at 7 days (MD 0.50, 95% CI -0.15 to 1.15, one trial 60 women), and at 8–14 days post intervention (MD -0.20, 95% CI -0.57 to 0.17, one trial 60 women).

There were fewer women in the intervention group reporting use of pharmacological medication compared with the control (RR 0.58, 95% CI 0.45 to 0.75, 2 trials, 322 women). One trial reported pain as the presence or absence of pain symptoms. There was no difference in the number of women reporting no pain at the end of the intervention (RR 0.68, 95%CI 0.41 to 1.12, 42 women).

Acupressure

Two small trials, of acupressure found no reduction in perineal pain within 24 h of the intervention being applied compared to the control (SMD -0.54, 95% CI -1.13 to 0.05, 135 women, I2 0%), and no difference in pain assessed at 8 to 14 days (MD -1.20, 95% CI -3.58 to 1.19, one trial, 75 women). One trial 60 reported self-reported pain as the presence or absence of symptoms. There was no difference in the number of women reporting no pain at the end of the intervention (RR 0.95, 95% CI 0.72 to 1.24, 256 women).

Acupuncture

One trial reported pain as the presence or absence of symptoms. There was no difference in the number of women reporting no perineal pain at the end of the intervention (RR 1.34, 95% CI 0.97 to 1.85, 207 women).

One trial did not report data that could be included in the meta-analysis, this study found that pain was significantly reduced in the acupuncture group on the third day postpartum (P = 0.22).

Secondary outcome

Use of pharmacological medication did not differ between groups (RR 1.23, 95% CI 0.79 to 1.93, 2 trials,, 267 women I2 0%). One study reported no adverse events in both groups.

One trial did not report data that could be used in the meta-analysis. This study found the mean time to a 50% reduction of pain was six days in the control group compared with 5 days in the acupressure group (p = 0.35). There was also no difference in the use of pharmacological medication

Music

Two trials, found no reduction in pain (perineal and at site of caesarean) within the first 24 h of administration compared to a control (MD -0.57, 95% -1.74 to 0.59, 241 women, I2 93%).

One trial found higher maternal satisfaction with pain relief in the music group compared with the control (RR 2.91, 95% CI 2.66 to 316, 141 women).

Massage

Two trials, found a reduction in pain following recovery from caesarean birth within 24 h of being administered (MD -2.64, 95% -2.82 to -2.46, 184 women, I2 0%). A benefit from massage with reducing pain was present at 7 days (MD -1.91, 95% CI -2.42 to -1.40, 2 trials, 120 women I2 37%), (Fig. 3).

Fig. 3

Effect of massage on pain using.

There was a reduction in the amount of pharmacological pain relief used by women in the massage group (RR 0.21, 95% CI 0.10 to 0.46, one trial, 80 women) One trial reported on breastfeeding and found a significant increase in the frequency of breastfeeding in the massage group (data not reported)

Herbal ointment

Two trials, found a reduction in perineal pain within 24 h of administration (MD -1.33, 95% CI -1.96 to -0.70, 221 women, I2 0%), whilst no difference at 7 days MD -0.50, 95% CI -1.90 to 0.89, 4 trials,,,, 334 women I2 90%. There was evidence of a reduction in perineal pain from the different herbal ointments group at 8 to 14 days (MD -0.74, 95% CI -1.02 to -0.47, 4 trials,,,, 372 women) (Fig. 4).

Fig. 4

Effect of herbal ointment on pain.

Wound healing assessed by the REEDA scale demonstrated improved perineal healing in the intervention group at 24 h post intervention (MD -1.40, 95% CI -1.92 to -0.88, 123 women, one trial, I2 94%) There was no effect at 7 days (MD -0.67, 95% CI -1.36 to 0.02), 4 trials, 334 women,,,, and at 14 days (MD -0.52, 95% CI -0.67 to -0.38, I2 97% 3 trials, 289 women).,, One trial reported on adverse events and found no difference between groups (RR 1.4, 95% CI 0.95 to 2.09, 108 women). These events included nausea, vomiting, diarrhoea, fever and skin reactions. One trial found no difference in the dose of medication used (MD 0.16, 95% CI -0.03 to 0.35, 98 women). The authors of the paper reported higher satisfaction with using the chamomile cream on the day 10 although data was not reported in the paper.

One trial reported data not in a form that could be included in the meta-analysis. This study found a significant difference between groups with an improvement in perineal pain level at 7th, 10th and 14th days postpartum, and an improvement in wound healing at 7th and 10th days postpartum.

Cupping

One small trial of cupping (included in the trial of acupressure) found reduced perineal pain in the first 24 h following administration (MD -9.90, 95% CI -12.73 to -7.07, 75 women). A reduction in pain was also found at eight to 14 days postpartum (MD -7.80, 95% CI -9.90 to -5.70, one trial, 75 women).

Laser

One small trial of laser therapy found no difference in perineal pain between groups following administration within the first 24 h postpartum (MD 1.10, 95% CI -0.39 to 2.59, 41 women) There was no difference at eight to 14 days (MD 0.10, 95% CI -1.37 to 1.57, one trial 41 women).

One trial did not report data that could be used in the meta-analysis, there was no difference in perineal pain at the end of the intervention

Wound healing of the perineum assessed by the REEDA scale demonstrated no difference between groups when assessed at 24 h post intervention (MD 0.33, 95% CI -0.5 to 1.16, one trial, 43 women), and at 8–14 days post intervention (MD 0.47, 95% CI -0.15 to 1.09, one trial, 41 women).

Herbal tablets

One trial of herbal tablets(containing Bromelain) found a reduction in perineal pain from the intervention at eight to 14 days (MD -0.35, 95% CI -0.40 to -0.30, 82 women).

There was an improvement in perineum wound healing using the REEDA scale at 8–14 days post intervention for the intervention group (MD -0.90, 95% CI -0.97 to -0.83, one trial 82 women). There was no difference in the use of pain relief between groups (RR 0.64, 95% CI 0.38 to 1.06), one trial, 82 women).

Electromagnetic therapy

One trial reported pain as the presence or absence of symptoms. There was no difference in the number of women reporting perineal pain at the end of the intervention (RR 0.70, 95% CI 0.48 to 1.01), 260 women).

There was no difference in pharmacological pain relief between the 2 groups (RR 1.26, 95% CI 0.79 to 2.01, one trial, 260 women).

Reflexology

There was no data on pain from one trial of reflexology that could be presented in the results, however it found a benefit from the intervention (p (X= 4.75, X = 7.65, t= -10.627, p < .001).

Quality of evidence

The quality of the evidence was downgraded for all outcomes that included multiple studies in the meta-analysis (Table 2). Most interventions and outcomes were rated as very low having been downgraded due to significant heterogeneity, high risk of bias on multiple domains and effect sizes with wide confidence intervals. Only one intervention (herbal ointment) assessing the outcome of pain within 14 days was assessed at moderate quality of evidence.

Table 2

Quality of Evidence for postpartum, wound healing and use of analgesics.

Quality of Evidence	Intervention	Outcome	References	Trials	Participants	MD 95% CI
VERY LOWa,b,c,d	Essential oils	Pain within 24 h	Behmanesh37Dale52Olapour43Vaziri48	4	423	0.33, 95–0.71 to 1.36
VERY LOWa,c,d,e	Acupressure	Pain within 24 h	Afravi33Akbarzade34	2	135	-2.61, 96% CI -4.13 to -1.09
VERY LOWa,b,c,e	Music	Pain within 24 h	Reza44Simavli55	2	241	-0.57, 95% CI -1.74 to 0.59
VERY LOWa,c,e	Massage	Pain within 24 h	Abbaspoor32Saatsaz45	2	184	-2.64, 95%CI -2.82 to -2.46
MODERATEc	Herbal ointment	Pain within 24 h	Aradmehr35Mohammadi41	2	221	-1.33, 95% CI -1.96 to -0.70
VERY LOWb,c,d,e	Essential oils	Pain with 7 days	Behmanesh37Dale52Sheikhan46	3	368	0.76, 95–1.90 to 3.41
VERY LOWa,c,e	Massage	Pain with 7 days	Lee58Sharma53	2	120	-1.91, 95% CI 2.42 to -1.40
VERY LOWb,c,e	Herbal ointment	Pain with 7 days	Asgharikhatooni36Kaviani40Marzouk60Moudi42	4	334	-0.50, 95% CI -1.90 to 0.89
LOWc,e	Essential oils	Pain with 14 days	Behmanesh37Dale52	2	296	0.22, 95% CI -0.22 to 0.67
VERY LOWb,c,e	Herbal ointment	Pain within 14 days	Kaviani 40Marzouk60Moudi42Aradmehr35	4	372	-0.74, 95–1.02 to -0.47
VERY LOWa,b,c,e	Essential oils	Wound healing within 24 h	Behmanesh37Sheikhan46	2	120	0.96, 9%CI -1.36 to -0.56
VERY LOWb,c,e	Herbal ointment	Wound healing within 7 days	Asgharikhatooni36Kaviani40Marzouk60Moudi42	4	334	-0.67, 95% CI -1.36 to 0.02
VERY LOWb,c,e	Herbal ointment	Wound healing with 14 days	Asgharikhatooni36Marzouk60Moudi42	3	289	-0.52, 95% CI -0.67 to -0.38
VERY LOWa,c,e	Acupuncture	Use of analgesics	Jaic56Kindberg59	2	267	RR 1.23, 95% CI 0.79 to 1.93
LOWc,e	Essential oils	Use if analgesics	Dale52Olapour43	2	322	RR 0.58, 95% CI 0.45 to 0.75

Downgraded one level due to serious imprecision: small sample size

Downgraded one level due to serious inconsistency: substantial heterogeneity present

Downgraded one level due to serious risk of bias: one study high or unclear risk of bias on six domains.

Downgraded one level for imprecision due to wide confidence intervals.

Downgraded one level due to serious risk of bias: one study did not blind participants to group allocation, or it was unclear.

Discussion

This review aimed to examine the effectiveness and safety of complementary health approaches identified little evidence that the modalities included in the review are beneficial in reducing postpartum pain. Overall a small number of studies were found for the complementary medicines and therapies included in the review. Twenty-five trials and 2413 women were included in this systematic review and meta-analysis. Ten complementary pain relief modalities were included in the review with studies conducted in 11 countries including high; middle and low income countries. Few trials reported on outcomes other than pain, use of pharmacological medication and wound healing. Few trials reported on adverse events and the overall safety of these therapies is uncertain One trial of acupuncture included this outcome in the study and no adverse events were reported. One trial evaluating a herbal ointment found no difference in adverse events between groups. Overall there was an absence of the inclusion of infant outcomes.

The overall risk of bias was low with only three trials where the majority of domains at a low risk of bias. The quality of evidence overall was low or very low for all modalities and the findings conflicting. The results from this review should be interpreted with caution due to the uncertainty arising from the quality of evidence due to risk of bias and the small number of study participants.

We found that massage may reduce pain in the short term and within the two-week period postpartum for women undergoing a caesarean birth. There was an associated decreased use of pharmacological pain relief and an increased reporting of breastfeeding. Trials of different herbal ointments found a reduction in perineal pain within 24 h of initial administration, no effect with the first week and a reduction in the second week of the intervention, a benefit in perineal wound healing within 24 h but no effect thereafter. These trials were all undertaken on women following an episiotomy. One trial of cupping was found to reduce perineal pain in the short term and at the end of the two weeks postpartum. One trial of herbal tablets found reduced perineal pain in the second week postpartum and an improvement in perineal wound healing. One trial of music was found to increase women's overall satisfaction during the period of the intervention.

Essential oil aromatherapy, utilised at 24 h, and 8–14 days identified no difference in pharmacological pain relief for women in the postpartum period. Acupressure showed no improvement in perineal pain two hours after use and at 8–14 days post intervention. One acupuncture study found no difference in the number of women reporting perineal pain post intervention. Whilst one music study found an increase in maternal satisfaction, neither of the two music studies included in the review established a reduction in pain compared to a control group within the first 24 h. Electromagnetic therapy found no difference in the number of women reporting perineal pain post intervention.

A high quality recent Cochrane systematic review examined the evidence for pharmacological and non-pharmacological interventions to reduce pain due to uterine cramps following childbirth. This review found evidence for pharmacological medication including non-steroidal anti-inflammatory agents and opioids with reducing pain associated with uterine cramps, with the primary outcome focussing on adequate pain relief. Four studies of herbal preparations were compared to non-steroidal anti-inflammatory medication, it was unclear if there were differences between groups for pain relief, the need for additional pain relief, maternal adverse events. One study of TENS was compared to no TENS found no difference between groups for the outcome adequate pain relief. A number of complementary health interventions identified in this review were not included in the Cochrane review. Some complementary health therapies were excluded from the review due to uncertainty if the study met the inclusion criteria, and some studies that were included in our review were ongoing at the time the Cochrane review was being prepared. Although this Cochrane review included 28 studies and 2749 women, the focus was on women who gave birth vaginally only, and therefore the findings from the Cochrane review can be generalised to women with singleton pregnancies and uncomplicated births. A narrative review of aromatherapy examined the effects of essential oils on postpartum women which included the outcome of postpartum pain, and delivered the intervention over an eight week period. This review by Tsai et al, included additional studies not reported in our review, this may be explained by the inclusion of women receiving an intervention in the eight week period of the post-partum period. An earlier review from 2019 included fewer studies of aromatherapy reporting on postpartum pain. This review concluded therapeutic effects from essential oils but highlighted the limitations arising from the variation in the duration of the postpartum period, the inclusion of women undergoing different modes of birth and the different treatment regimens used.

The studies in our review can be generalised to women who gave birth vaginally and who underwent a caesarean birth. We aimed to reduce the review bias by including two authors assessing the eligibility review, and undertaking both data extraction and the risk of bias assessment. There are some limitations to our study with non-English studies being excluded. As a result, it is likely that some relevant literature may have been excluded from the review. Furthermore, it is possible that some literature that relates to non-pharmacological modalities has not been included within the databases included in our search. We excluded pain interventions outside of the immediate two-week postpartum period and nipple/breast pain as well as chronic back pain. As a result, our findings are not generalised to these groups. Whilst grey literature was searched extensively it is possible that some studies have been missed. A sensitivity analysis and analysis of publication bias were not undertaken due to the small number of trials included in the analysis.

This review has highlighted some evidence of benefit from massage and herbal ointment and herbal medication with reducing pain and improving wound healing. There was significant heterogeneity identified within comparisons and this may be explained by different clinical features of the intervention including aspects of dosage or the composition of the intervention its self. There was some heterogeneity in the comparison to a control and this is acknowledged as a limitation of the review. Further high quality research is needed focussing on the therapies identified as showing positive and consistent treatment effects on several outcomes, and future trials should include larger samples of women and comprehensive inclusion of both maternal and neonatal outcomes.

In conclusion, massage, cupping, herbal tablets and herbal ointments may be somewhat effective in the management of postpartum pain. Essential oils, acupressure, acupuncture and electromagnetic therapy studies found no improvement in postpartum pain management. The safety of use of these modalities by postpartum women remains uncertain due to a lack of data, and the overall very low quality of evidence. These conclusions need to be considered in light of the identified review limitations. The area of complementary approaches in the management of postpartum pain requires further study.

Author contributions

Conceptualisation: CS, HD. Methodology and data acquisition: CS, EH, AD, HD. Formal analysis: CS, AD. Writing – original; draft: CS, EH, AD, HD, CT. Writing review and editing: CS. EH, AD, HD, CT.

Conflict of interest

The authors declare no conflict of interests.

Funding

None

Ethical statement

No ethical approval was required for this manuscript as this study did not involve human subjects or laboratory animals.

Data availability

Data associated with this paper will be provided upon request.