Pregnancy Research on Osteopathic Manipulation Optimizing Treatment Effects: The PROMOTE Study A Randomized Controlled Trial

Abstract

Objective

To evaluate the efficacy of Osteopathic Manipulative Treatment (OMT) to reduce low back pain and improve functioning during the third trimester in pregnancy and improve selected outcomes of labor and delivery.

Study Design

PROMOTE was a randomized, placebo-controlled trial of 400 women in their third trimester. Women were randomized to usual care only (UCO), usual care plus OMT (OMT), or usual care plus placebo ultrasound treatment (PUT). The study included seven treatments over nine weeks. The OMT protocol included specific techniques administered by board-certified OMT specialists. Outcomes were assessed using self-report measures for pain and back-related functioning, and medical records for delivery outcomes.

Results

There were 136 women in the OMT group, 131 in PUT, and 133 in UCO. Characteristics at baseline were similar across groups. Findings indicate significant treatment effects for pain and back related functioning (P<.001 for both), with outcomes for the OMT group similar to that of the PUT, but both groups were significantly improved compared to UCO. For secondary outcome of meconium- stained amniotic fluid there were no differences between the groups.

Conclusion

OMT was effective for mitigating pain and functional deterioration compared to the UCO group; however OMT did not differ significantly from PUT. This may be attributed to PUT being a more active treatment than intended. There was no higher likelihood of conversion to high risk status based on treatment group. Therefore, OMT is a safe, effective adjunctive modality to improve pain and functioning during their third trimester.

INTRODUCTION

Low back pain (LBP) is encountered in an estimated 70% of pregnant patients ^1,2 but challenges persist in identifying safe, effective treatment options. For example, most pain medications are not recommended during pregnancy, leaving few options for pain control. Osteopathic manipulative treatment (OMT) is a body-based treatment that offers a conservative, non-invasive option for relieving pregnancy-related LBP while increasing back-related function. Osteopathic manipulative treatment is defined as “the therapeutic application of manually guided forces by an osteopathic physician to improve physiologic function and/or support homeostasis that has been altered by somatic dysfunction"³. Pregnancy brings dramatic musculoskeletal changes that alter normal biomechanics, accompanied by ligamentous strain, increased muscle tension, and decreased range of motion, causing pain. OMT treats somatic dysfunctions with a variety of techniques, thereby increasing range of motion, improving tissue texture, and decreasing pain. ^4–8.

Studies of OMT for LBP in non-pregnant populations have provided sufficient evidence of safety and efficacy for the establishment of practice guidelines recommending OMT for LBP⁹. Those studies reported greater pain relief, improved physical functioning, and reduced pain medication use with OMT compared to subjects not receiving manipulation ^10–12. Because of the unique biomechanical changes during pregnancy, it was not known if OMT would produce the same positive outcomes in pregnant women with LBP, although a pilot study found a slowing of decline in functional status¹³.

The Pregnancy Research in Osteopathic Manipulation Optimizing Treatment Effects (PROMOTE) study was designed to evaluate the efficacy of a nine-week regimen of OMT to reduce LBP during the third trimester. The primary objective was to evaluate the influence of OMT on self-reported pain and back-related functioning. A secondary objective was to corroborate an earlier study that found a relationship between receiving OMT during the third trimester and a decrease in meconium-stained amniotic fluid. We chose to examine this relationship based on a supposition that reducing maternal stress caused by pain may reduce meconium-staining, a potential objective measure of fetal stress ¹⁴.

MATERIALS AND METHODS

PROMOTE was a randomized, controlled clinical trial. Study visits were conducted from 2007–2011, at three Obstetrics and Gynecology clinics in Tarrant County, Texas. The Institutional Review Boards at the University of North Texas Health Science Center and John Peter Smith Health Network approved the study. Clinic personnel referred interested women to the research coordinators for screening. Eligible women aged 18 to 35, who had reached the 30^th gestational week and were willing to participate were consented and enrolled (Figure 1). Women with high-risk conditions, as determined by their OB provider, were excluded. High-risk conditions included but were not limited to pre-eclampsia/eclampsia, vaginal bleeding, oligohydramnios, gestational diabetes, and hypertension. Women were withdrawn from the study if their OB provider identified a high-risk condition during their participation. Women were also unable to continue in PROMOTE if they reported using other body-based therapies including additional OMT, massage, physical therapy, or chiropractic therapy during the study. OB providers included physicians (MD or DO), Certified Nurse Midwives, and Women’s Health Nurse Practitioners, and all providers were blinded to study group assignment. Providers knew their patient was participating in the study, but both subject and provider were asked to not discuss the study or treatments. Study subjects were compensated for their time and travel.

Fig. 1.

Flow of women through the PROMOTE trial.

After consent, subjects were randomized to one of three groups: osteopathic manipulative treatment (OMT), placebo ultrasound treatment (PUT), or usual obstetrical care only (UCO). A computer-generated randomization program was used to allocate women in blocks of 15, by clinic location, to each of the three treatment groups. At the first study visit, the research coordinator collected baseline data. Then the OMT specialist opened the randomization envelope to reveal the subject’s group assignment and provided the assigned treatment. PROMOTE included seven study visits, scheduled to correspond with ongoing routine prenatal care at weeks 30, 32, 34, 36, 37, 38, and 39. Study visits occurred within 24 hours after routine prenatal appointments. The research coordinator scheduled each study appointment, confirmed medical clearance from the OB provider, and collected data prior to treatment.

Both the OMT and PUT protocols addressed the same specific body regions chosen for their relationship to common musculoskeletal changes in pregnancy, including bilateral cervical, thoracic and lumbar paraspinal musculature, thoracolumbar junction, sacroiliac joint, hip and anterior pelvis. Both OMT and PUT treatments were applied over clothing. OMT consisted of approximately 20 minutes of treatment provided by a physician board-eligible or certified by the American Osteopathic Board of Neuromusculoskeletal Medicine and trained in the PROMOTE protocol (Table 1). Each OMT technique was performed on every subject for 1–2 minutes until an adequate tissue response was felt.

Table 1.

Summary of OMT techniques applied in the PROMOTE study

OMT Technique	Description
Seated Thoracic Articulation	Patient’s arms rest on physician’s chest. Physician reaches around patient to contact the articular pillars, then the rib heads Thorax is extended and sidebent/rotated as necessary for articulation
Cervical Soft Tissue	Apply kneading, traction, inhibition and/or stretching to cervical muscles
OA decompression	Contact near the occipital condyles with sustained anterior, lateral and cephalad tension
Thoracic Inlet Myofascial Release	Contacting the clavicles, upper thoracic spine and ribs, the fascia is assessed in rotation, side-bending, and forward/backward bending motions Tissues are positioned either directly or indirectly and held until released
Lateral Recumbent Scapulo-thoracic Soft Tissue	With positioning to isolate the scapula, kneading, traction, and/or stretching is rhythmically applied to the musculature medial to the scapula Positioning is altered to contact the posterior axillary fold, with kneading, inhibition, and stretching rhythmically applied
Lateral Recumbent Lumbar Soft Tissue	Kneading, lateral, rotational, and longitudinal stretching is applied in a rhythmic motion to the lumbar paraspinal musculature
Abdominal Diaphragm Myofascial Release	Using a lateral contact over the lower ribs, the motion of the diaphragm is assessed in rotation, side-bending, and forward/backward bending motions Tissues are taken either directly or indirectly and held until released
Pelvic Diaphragm Myofascial Release	Using anterior contact just superior to the pubic symphysis and posterior contact at the level of the coccyx and ischial tuberosities, the motion of the pelvic diaphragm is assessed in rotation, side-bending, and forward/backward bending motions Tissues are taken in either a direct or indirect position until released
Sacro-Iliac Articulation	Hip and knee flexed, with the leg supported Compression is applied through the knee to engage the femur into the acetabulum The leg is internally rotated and circumducted, then externally rotated and circumducted several times, while maintaining compression
Pubic Symphysis Decompression	Patient supine, physician at side of patient Hips and knees are flexed, with feet together on the table. Muscle energy is applied, first with the knees pulled apart against resistance, then being pulled together against resistance
Frog-Leg Sacral Release	Hips and knees flexed, with feet together on table Physician’s contact is cupping the sacrum with fingers at the base Sacrum is held at a balance point as the legs are externally rotated Inferior traction is applied while maintaining balance and legs are straightened with the heels sliding down the table
Compression of the Fourth Ventricle (CV4)	Contact medial to the occipital-mastoid suture with thenar eminences Encourage the occipital motion in extension phase, while resisting flexion until a still point is reached Allow the cranial rhythmic impulse to return to normal before disengaging

The PUT provided tactile and manual stimulation over the same regions as OMT. The ultrasound wand was applied with a circular, steady contact for approximately two minutes to each of the specified areas resulting in treatment duration similar to OMT. Although the ultrasound machine provided credible cues such as a ticking timer or digital display, the machines were set to not emit any ultrasound waves.

The UCO group completed study questionnaires, but received no study interventions, nor additional time or interaction with the treating physician.

There were two primary outcomes for PROMOTE: pain and back-related functioning. LBP was measured with the Quadruple Visual Analog Scale (QVAS), a self-reported reliable measure of four aspects of pain (now, average, worst, and best), used to calculate Characteristic Pain Intensity (CPI). ¹⁵ The CPI is a composite score reported on a scale of 0 to 100, with higher values indicating higher pain. Back-related functioning was measured with the Roland-Morris Low Back Pain and Disability Questionnaire (RMDQ), a validated 24-item self-reported questionnaire where higher scores indicate greater functional disability, and sensitive to change over time ^16–18. Labor and delivery records provided information for secondary outcomes. Following delivery, the research coordinator collected data from the subject’s medical records. All data were collected using paper forms, and transferred to an electronic dataset using double data entry. Subjects completed the Short Form 12 version 2 Health Survey (SF-12v2) at the beginning of the study to estimate general health at baseline. Collected demographic information included race, ethnicity, age, weight, height, and parity.

Sample size calculations for primary outcomes indicated a sample size of 71 per treatment group would be needed to detect a 30% change in pain outcomes; however this was less restrictive than the sample size required for the planned secondary outcome of meconium-staining. Thus calculations based on the secondary outcome were used for final sample size estimates. A sample size of 110 subjects per treatment group was estimated for 80% power at a 5% significance level (P<0.05) to detect a 62% reduction in the incidence of meconium-staining. The incidence and percent reduction estimate was based on a retrospective case-control study finding of a lower incidence of meconium-staining in women who received OMT compared to a control group ¹⁴. To allow for study attrition, the target recruitment goal was 400 subjects.

Data management and analyses were performed using the statistical software package IBM SPSS (version 19.0; IBM Corp., Armonk, NY, USA). Demographic characteristics were described by frequencies and percentages for categorical data, and by means and standard deviations for continuous data. Between-group comparisons were computed using chi-square for categorical data and analysis of variance for continuous data. Primary outcomes were analyzed using a linear mixed model suitable for repeated measures with fixed effects for treatment group and treatment visit, including an interaction term for treatment group and visit. For each of these variables, change from baseline was calculated and used as the data point for that visit. To address differences in baseline pain and disability, which might influence rates of change, models were adjusted for baseline values. Post-hoc analyses with Bonferroni adjustment were used to elucidate pairwise differences between treatment groups when main effects were detected. The linear mixed model analysis allowed for the use of all available data without censoring or imputation. Analysis was first performed using an intention-to-treat approach. An additional per-protocol analysis was performed on subjects who completed all seven treatment visits. Finally, logistic regression was used for analysis of the secondary outcome of meconium-staining.

RESULTS

In PROMOTE, a total of 400 subjects were enrolled and randomized. As shown in Figure 1, there were 136 (34%) women assigned to OMT, 131 (33%) to PUT, and 133 (33%) to UCO. Of these 400 women, 99 (25%) completed all seven treatment visits and were considered to have completed the protocol.

Baseline characteristics were similar across groups with the exception of BMI (Table 2). The OMT group had a lower mean pre-pregnancy BMI than the other two groups. Table 3 shows that baseline pain and functioning were similar among the three groups, except on pain subscales at baseline, the OMT group reported higher levels of pain now than the PUT group and higher levels of pain at best than the UCO group.

Table 2.

Characteristics of Subjects by Study Group

	OMT (n = 136)	PUT (n = 131)	UCO (n = 133)	P value
Age	23.99 ± 4.13	24.11 ± 4.10	24.70 ± 4.54	.351
Marital Status
Single	62 (45.6)	64 (48.9)	58 (43.6)
Married	65 (47.8)	61 (46.6)	70 (52.6)	.712
Divorced or Separated	9 (6.6)	6 (4.6)	5 (3.8)
Race/Ethnicity
White	32 (23.5)	36 (27.5)	31 (23.5)
Black	25 (18.4)	23 (17.6)	22 (16.7)	.972
Hispanic	76 (55.9)	69 (52.7)	77 (58.3)
Other	3 (2.2)	3 (2.3)	2 (1.5)
Education Level
Grade School	11 (8.2)	11 (8.4)	8 (6.1)
High School	75 (56.0)	69 (52.7)	67 (50.8)	.761
Some College and Above	48 (35.8)	51 (38.9)	57 (43.2)
SF-12v2 Health Survey	37.43 ± 7.22	38.96 ± 7.30	38.74 ± 7.08	.185
Baseline BMI*	25.51 ± 4.56	27.50 ± 6.44	27.54 ± 6.61	.028
Weight Gain in kg*	14.57 ± 6.48	15.14 ± 6.63	13.40 ± 7.27	.211
Nulliparous	41 (34.2)	37 (35.6)	32 (29.9)	.658

OMT, Osteopathic Manipulative Treatment; PUT, Placebo Ultrasound Treatment; UCO, Usual Care Only.

Data are mean ± standard deviation or n (%).

^*Data were missing for 30% of cases.

Table 3.

Baseline Measurements of Back Pain and Functioning

	OMT (n = 136)	PUT (n = 131)	UCO (n = 133)	P value
RMDQ	6.70 ± 4.97	5.90 ± 4.68	6.55 ± 5.09	.375
CPI	53.38 ± 20.33	48.53 ± 20.13	49.21 ± 19.82	.103
Pain Now	3.49 ± 2.54	2.75 ± 2.42*	2.93 ± 2.21	.030
Pain Average	5.07 ± 2.39	4.71 ± 2.34	4.78 ± 2.41	.430
Pain Best	2.49 ± 2.04	2.02 ± 1.80	1.65 ± 1.92*	.002
Pain Worst	7.46 ± 2.36	7.12 ± 2.32	7.08 ± 2.17	.334

CPI, Characteristic Pain Intensity; OMT, Osteopathic Manipulative Treatment; PUT, Placebo Ultrasound Treatment; RMDQ, Roland Morris Disability Questionnaire; UCO, Usual Care Only.

Data are mean ± standard deviation.

^*Different from OMT.

Analysis of primary outcomes used an intention-to-treat model examining changes in pain and back-related functioning for each group across the study (N=400). The intention-to-treat analysis included all randomized subjects regardless of the number of visits completed. These findings indicated significant treatment effects for both pain as assessed by CPI and back-related functioning assessed by RDMQ (P<.001 for both). Specifically, OMT was effective for mitigating the progression of pain and deterioration of back-related functioning compared to the UCO group.

For each of the outcomes, the OMT group either demonstrated improvement or remained stable across time (Figure 2). In contrast, the UCO group experienced worsening pain and functioning. Although there was an effect of treatment and OMT was different from UCO, it is important to note that OMT outcomes did not differ significantly from those of the PUT group (Table 4). The addition of baseline BMI, maternal age, and parity adjustments into the model yielded findings consistent with the non-adjusted analysis.

Fig. 2. Mean change in self-reported pain and back functioning for each group by study visit.

2A. For RDMQ, there was an effect for treatment group, time, and interaction (all P<.001).

2B. For CPI, there was an effect for treatment group, time, and interaction (P<.001, P=.048 and P<.001 respectively).

2C. For Pain Now, treatment group was significant (P<.001), but not time (P=.139). However there was an interaction for group and time (P<.002).

2D. For Pain Average, both treatment group and time were significant (P=.004 and P=.012 respectively), however the interaction was not (P=.185).

2E. For Pain Best, treatment group was significant (P<.001), but not time (P=.845). However there was an interaction for group and time (P=.001).

2F. For Pain Worst, there were differences for treatment group, time, and interaction (P<.001, P=.004 and P<.001 respectively).

Table 4.

Primary Outcomes Estimated in a Linear Mixed Effects Model adjusted for Baseline

	OMT (n=136)	PUT (n=131)	UCO (n=133)	Difference Between OMT and PUT Groups		Difference Between OMT and UCO Groups
				Mean (95%CI)	P value	Mean (95%CI)	P value
RMDQ	.676	.469	2.926	0.21 (−0.73 to 1.14)	>.999	−2.25 (−3.18 to −1.32)	<.001
CPI	−3.341	−3.488	3.769	.15 (−3.07 to 3.36)	>.999	−7.11 (−10.30 to −3.93)	<.001
Pain Now	−.299	−.034	.707	−0.27 (−0.70 to 0.17)	.439	−1.01 (−1.44 to −0.57)	<.001
Pain Average	−.205	−.364	.175	0.16 (−0.24 to 0.56)	>.999	−0.38 (−0.78 to 0.02)	.065
Pain Best	−.202	−.154	.478	−0.05 (−0.38 to 0.28)	>.999	−0.68 (−1.00 to −0.36)	<.001
Pain Worst	−.482	−.641	.296	0.16 (−0.22 to 0.54)	.946	−0.78 (−1.15 to −0.4)	<.001

CI, confidence interval; CPI, Characteristic Pain Intensity; OMT, Osteopathic Manipulative Treatment; PUT, Placebo Ultrasound Treatment; RMDQ, Roland Morris Disability Questionnaire; UCO, Usual Care Only.

Values are estimates for mean change in pain and P values are pairwise comparisons using Bonferonni adjustment.

Although the study was designed to include seven visits, it was expected that some participants would not be able to make all visits, and 44% delivered by 39 weeks. Therefore, we investigated if adherence to the protocol influenced findings by performing per-protocol analysis for subjects who received all seven treatments as well as an analysis for those who received at least four of seven treatments, which had previously been shown to be a therapeutic dose¹⁴. Results for the 99 subjects who completed all seven visits as planned and the 357 who completed at least four visits were consistent with the intention-to-treat findings for significant treatment effect.

Analyses of secondary outcomes focused on objective delivery measures. For the 329 women with available delivery information, only 61 (18.5%) had meconium-staining documented. Logistic regression indicated meconium-staining was not influenced by treatment group, and the overall model was not significant (P=.611). Analysis of gestational age at delivery indicated no difference based on group, with mean gestational age of 39 weeks 2 days. Conversion to high-risk status occurred for 12.5% of subjects (OMT=11, PUT=19, UCO=20). In a logistic regression model using conversion to high-risk status as an indirect measure of safety, there was no higher likelihood of conversion to high risk based on treatment group (P=.141).

COMMENT

The primary objective of PROMOTE was to investigate if adding OMT to usual obstetric care mitigated the progression of LBP and improved back-specific functioning during the third trimester. Our results show that OMT has benefits compared to UCO, demonstrating clinically and statistically significant improvements in pain and back-related functioning scores. Overall, the UCO group experienced increasing pain and decreasing function during the course of the study while the OMT and PUT groups’ pain and functioning did not deteriorate. These findings were consistent even when potential confounders such as BMI, maternal age, and parity were considered. In addition, incidences of conversion to high-risk status and meconium- staining were not higher in the OMT group, suggesting there is no additional risk associated with OMT for pregnant women in the third trimester. Additionally, there were fewer conversions to high-risk status in the OMT group, which may be clinically important even though it was not statistically significant. A more nuanced examination of the plausible protective nature of OMT is warranted.

A secondary objective was to attempt to elucidate the mechanism by which OMT may cause benefit; therefore the study was designed with a placebo group that would control for some of the potentially therapeutic components of the OMT group. Our results failed to show differences between the OMT and PUT, which may indicate that the common treatment components of time, touch, intention and interaction may be at least partially responsible for the observed beneficial effects.

A strength of PROMOTE is that it is one of the largest randomized controlled trials ever conducted on the effectiveness of OMT, and addressed pregnant women as a unique population. Other strengths include the use of reliable and valid questionnaires for pain and function, objective delivery outcomes, and close supervision by the obstetrical provider to ensure safety. Analysis included both intention-to-treat and per-protocol analyses and utilized a linear mixed model, less sensitive to censoring and missing data, to detect treatment effects. An important limitation was a higher than expected attrition rate in all groups. Of the 400 enrolled subjects, 99 (24.75%) completed the seven-visit protocol. While 80 women (20%) became ineligible or declined to continue, additional attrition was related to missed visits and delivery prior to week 39. This results in a less than robust sample for detecting effects in low-frequency outcomes such as conversion to high-risk status and meconium-staining.

The specific OMT protocol designed for this study includes both strengths and limitations. It includes a standardized set of techniques that could be applied in a clinical or research setting, and each technique is described in contemporary osteopathic literature ^19–23. However, this protocol is a departure from OMT provided by a physician in a clinical setting that would be customized to fit the specific musculoskeletal diagnoses of each patient. This may have resulted in suboptimal treatment for subjects who may have had better response with individualized treatment.

Including and selecting a placebo in manual therapy research is challenging, especially where outcome may be affected by many factors such as touch, time, intention, interaction, and expectation of benefit.²⁴ In designing PROMOTE, we reviewed placebo research and chose sub-therapeutic ultrasound because it has been shown to be credible.²⁹ We expected PUT would be inert because the machines were not actually emitting any waves. However, since the study began in 2007, much research has been done on placebo potency, and we now suspect that components of PUT likely caused a therapeutic response, even though other studies have not shown the same degree of placebo effect.¹³ In this trial, all treatments were provided by the same OMT specialists and contained study visit elements of assessment and observation, therapeutic ritual, and a supportive patient-practitioner relationship, that have been found to additively contribute to a significant effect ²⁴. In addition, the repetitive pressure from the ultrasound wand, in combination with the friction resulting from the contact on skin or clothing (without gel) may have affected the selected areas similar to light-to moderate pressure massage, that has also been shown to decrease LBP in pregnancy ³⁰.

There have been few studies of the safety and efficacy of OMT during pregnancy for LBP, and previous research failed to detect reduction in pain or in the incidences of various high-risk outcomes¹³. PROMOTE has confirmed earlier findings of safety and the slowing of progression of back-related disability and also demonstrated significance for pain outcomes. With few options for safely treating LBP during pregnancy, these findings are clinically meaningful. A common treatment for LBP is acetaminophen (paracetamol); however, considering the recent controversy associating acetaminophen use during pregnancy with childhood asthma, adverse neurodevelopmental outcomes, and serious skin reactions ^25–28 , it emphasizes the need to provide safe, effective, non-invasive treatments for women seeking non-pharmacological options to reduce pain and increase back-related function. Based on these findings, OB providers should consider adding body-based treatments for LBP into the care of pregnant women. While OMT may be an important option, there are few board-certified OMT specialists readily available, as DOs comprise only 7.2% of United States physicians, and less than 1000 DOs are board-certified in OMT.³¹ Broad applicability may rely on training providers in the OMT protocol, creating synergistic clinical partnerships between OB providers and OMT-trained physicians, and considering other body-based therapies where OMT is not available.

Future studies need to carefully consider the placebo choice to select one that is truly inert, controls for the time and treatment interaction, and allows for isolation of the effect of different manual modalities. They will also need to choose between a standardized treatment protocol similar to PROMOTE, versus an individualized, more clinical approach. Earlier enrollment of subjects may be considered to evaluate the effect of earlier intervention on musculoskeletal and physiological changes that have already manifested by 30 weeks gestation. In addition, while there were not statistically significant differences in the incidence of conversion to high-risk status, there may be some clinically meaningful protective effects of OMT that would warrant a more nuanced examination in future studies.

The treatment goals for LBP in pregnancy differ from other conditions in that it may not be reasonable to expect to eliminate pain and disability, but instead to slow the progression related to advancing pregnancy. In PROMOTE, both OMT and PUT demonstrated significant mitigating effects on pain and functional disability compared to UCO. Therefore, the inclusion of body-based therapies such as OMT that provide touch, time and interaction may offer patients and providers a safe, effective adjunctive option to improve comfort and reduce the impact of pain commonly associated with third trimester pregnancy.