Acupuncture for Pelvic and Back Pain in Pregnancy: a Systematic Review Carolyn C

Reviews Obstetrics www.AJOG.org

OBSTETRICS Acupuncture for pelvic and back pain in pregnancy: a systematic review Carolyn C. Ee, MBBS, BAppSci (Chinese Medicine/Human Biology); Eric Manheimer, MS; Marie V. Pirotta, MBBS, PhD; Adrian R. White, MA, MD, BM BCh

sleep in late pregnancy, and both acupunc- The objective of our study was to review the effectiveness of needle acupuncture in ture and physiotherapy may improve treating the common and disabling problem of pelvic and back pain in pregnancy. Two pain.13 Several case reports and -1 retro small trials on mixed pelvic/back pain and 1 large high-quality trial on pelvic pain met spective case series have suggested that the inclusion criteria. Acupuncture, as an adjunct to standard treatment, was superior to acupuncture may relieve pelvic and back standard treatment alone and physiotherapy in relieving mixed pelvic/back pain. Women pain in pregnancy.14-17 with well-defined pelvic pain had greater relief of pain with a combination of acupuncture Complementary and alternative ther- and standard treatment, compared to standard treatment alone or stabilizing exercises and apies are growing in popularity and are standard treatment. We used a narrative synthesis due to significant clinical heterogeneity used by more than a third of the US pop- between trials. Few and minor adverse events were reported. We conclude that limited ulation.18 They continue to be used- dur evidence supports acupuncture use in treating pregnancy-related pelvic and back pain. ing pregnancy,19 and 60% of women Additional high-quality trials are needed to test the existing promising evidence for this with lower back pain in pregnancy report relatively safe and popular complementary therapy. that they would accept complementary therapies for treatment of their20 pain. Key words: acupuncture, back pain, pelvic pain, pregnancy Acupuncture is used by more than 2 million people in the US 18annually. It involves stimulation of anatomical loca- elvic and back pain are amonguse ofthe different definitions, and some ex- tions on the skin (acupoints) by various P most common “minor” complica- perts advocate differentiating pelvic measures, most commonly by penetra- 1 3 tions in pregnancy.Estimates of preva- from back pain in pregnancy. tion of the skin by metallic needles (nee- lence of pelvic and back pain in preg- The exact etiology remains 4unclear 2 dle acupuncture). Acupuncture analge- nancy range from 24-90%.This and is thought to be related to the inter- sia involves complex neurohumoral difference is most probably due to the action between physiological changes in mechanisms involving endogenous opi- pregnancy and risk factors such as phys- ates and monoamines,21 with evidence From the Department of General Practice, ical work and previous back or pelvic of sustained depression of dorsal horn 2,5 University of Melbourne, Melbourne, pain. The pain can result in significantneurons in the spinal 22 cord.Adverse Australia (Drs Ee and Pirotta); Center for morbidity. Twenty-five percent of events are reported to be 23minimal, and Integrative Medicine, University of women with pelvic pain in pregnancy life-threatening events such as pneumo- Maryland School of Medicine, Baltimore, will seek medical help for their pain, 8% thorax are considered rare in the hands MD (Mr Manheimer); Peninsula School of are severely disabled, and 7% continue to of a trained practitioner.24 Medicine and Dentistry, University of have pain beyond the pregnancy.6 The Our aim in this review was to determine Plymouth, Plymouth, UK (Dr White). majority of women with back pain in whether acupuncture is more effective Received May 11, 2007; revised Oct. 3, 2007; pregnancy report disturbed sleep from than “standard treatment,” no additional accepted Nov. 5, 2007. their pain.7 Disability often involves- simtreatment, placebo acupuncture, “sham” Reprints: Carolyn Ee, 200 Berkeley St, 8 Carlton, Victoria 3053, Australia. ple activities of daily andliving can - reacupuncture, or other treatments in the 9 [email protected]. sult in significant absenteeism.Back management of pain and disability due to Carolyn Ee was funded by General Practice pain in pregnancy also increases the risk pregnancy-related pelvic and back pain. Education and Training Australia. Eric of postpartum back 5pain. We chose to include both pelvic and back Manheimer was funded by grant number R24 Provision of education, advice, and the pain in our review, as many investigators AT001293 from the National Center for do, because of the ongoing debate and un- Complementary and Alternative Medicine prescription of exercise by a physiothera- (NCCAM) of the US National Institutes of pist appear to be the standard recommen- certainty regarding etiology and treatment Health. Adrian White was supported by the dations for treatment.10 Evidence for theof this problem. DH-National Coordinating Centre for Research benefits of physical therapies and support Capacity Development (NCC RCD). belts is inconclusive.3,11,12 A Cochrane - reMATERIALS AND METHODS 0002-9378/$34.00 view found that water gymnastics helps re- We searched the following electronic da- © 2008 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2007.11.008 duce sick leave in pregnancy, a specially tabases from their inception until July shaped pillow improves back pain and 2006:

254 American Journal of Obstetrics & Gynecology MARCH 2008 www.AJOG.org Obstetrics Reviews

The Cochrane Central Register of Controlled Trials (CENTRAL), National TABLE 1 Library for Health Complementary and Modified Cochrane Back Review Group (BRG) criteria for methodological Alternative Medicine Specialist Library, quality assessment of randomized, controlled trials CINAHL, EMBASE, AMED, and Acu- 1. Was the method of randomization adequate? briefs. We searched MEDLINE from its 2. Was allocation concealment adequate? 3. Was an appropriate sample size calculation used to ensure adequate inception until November 2006. Due to power to detect significant differences due to treatment? funding limitations, we only searched for 4. Were participants blinded to intervention? trials written in or translated into 5. Were caregivers blinded to intervention? English. 6. Were cointerventions avoided or similar? We based our MEDLINE search strat- 7. Were cointerventions reported for each group separately? 8. Was the acupuncture treatment adequate? egy on the Cochrane highly sensitive 9. Was the withdrawal and dropout rate described and acceptable? 25 search strategyand combined this with 10. Did the analysis include an intention-to-treat analysis? specific intervention and disease identi- 11. Was the outcome assessor blinded to the intervention received? 12. Was the timing of outcome assessment in both groups similar? fiers. The key MeSH terms and keywords ...... used were “acupuncture,” “acupuncture Details of operationalization available from the authors upon request. therapy,” “electroacupuncture,” “Medi- Ee. Acupuncture in treating pelvic and back pain in pregnancy. Am J Obstet Gynecol 2008. cine, Chinese traditional,” “pregnancy complications,” “pregnancy,” “peripartum,” “prenatal care,” “pelvic pain,” have had a nonmusculoskeletal cause for dles. According to acupuncture theory, “back pain,” “low back pain,” “lumbar their pain (eg, malignancy, urinary tract activation of de qi may be one indication back pain,” “sacroiliac joint pain,” and infection, obstetric complication). Trials that acupuncture is exerting its beneficial “symphysis pubis pain.” using laser therapy alone without the use effects). Details of the control group in- We attempted to identify unpublished of needles were excluded from our re- tervention and cointerventions were also trials by contacting prominent acupunc- view. We included auricular acupunc- extracted. We attempted to contact the ture researchers in the US, UK, Europe, ture trials, but only if this was combined chief investigators for missing trial data. Australia, and Sweden and by searching with body acupuncture, and intended to Trial quality was assessed by 2 inde- reference lists of identified trials. We also perform a separate analysis for such tri- pendent reviewers according to 2 scales. searched Computer Retrieval of Infor- als. We postulated that the neurohu- The first scale used was a modified Jadad mation on Scientific Projects (CRISP) moral mechanisms involved in these scale assessing adequacy and reporting of and Current Controlled Trials (CCT) for other therapies may differ from those in- the randomization method, participant ongoing trials. volved in needle acupuncture; in addi- blinding, testing of participant blinding Two reviewers independently assessed tion, we noted that most experts would after treatment, and reporting of drop- study eligibility. Our inclusion criteria agree that acupuncture by definition in- outs and withdrawals. We regarded a were randomized controlled trials com- volves insertion of needles into the score of 2 points or less out of a total of 5 paring acupuncture therapy against a skin.26,27 points as indicating a poor quality trial. control group for pelvic and back pain in Two reviewers independently ex- The second scale used was modified pregnancy. We defined acupuncture as tracted data from eligible trials as defined from Cochrane Back Review Group cri- needle insertion into acupoints, whether above. We used a modified version of a teria29 (Table 1).We added criteria for the acupuncture was described as data extraction spreadsheet that was pre- adequate acupuncture treatment30 and “traditional” Chinese acupuncture, viously used.28 Where possible, we- exadequate sample size calculation. Both “Western”/segmental/tender point acu- tracted baseline, end-of-treatment, and reviewers are practicing acupuncturists. puncture, or other. Comparison inter- interval data. We regarded a score of 5 or less out of a ventions could be placebo/“sham” acu- We extracted participant data regard- total of 12 points as indicating a poor puncture, no additional treatment, ing diagnosis, age, gestation, and parity. quality trial. “standard treatment,” or any other treat- We also extracted details of acupuncture We intended to combine data, if suffi- ment. Our accepted outcome measures treatment, including type of acupunc- cient data were available, in a metaanaly- were pain, disability, overall improve- ture (Chinese/“Western”/or mixed), sis using Cochrane Review Manager ment, analgesic use, time off work, and acupoints used, frequency and duration software (RevMan software, version 4.2, adverse events. We included unpub- of treatment, number of sessions, type of Nordic Cochrane Centre, Copenhagen, lished trials. stimulation, and whether or not de qi Denmark), first performing chi-square We excluded trials that were quasi- was obtained (De qi, literally meaning testing to assess heterogeneity. We in- randomized. If the trial had a crossover “arrival of energy,” is a term used in acu- tended to use a random effects model if design, we intended to analyze only the puncture and refers to a sensation of significant statistical heterogeneity (P Ͻ data prior to the crossover. We excluded numbness or distension sometimes gen- .1) was found. Alternatively, we planned trials that enrolled women who may erated by stimulating acupuncture nee- to use a narrative synthesis if significant

MARCH 2008 American Journal of Obstetrics & Gynecology 255 Reviews Obstetrics www.AJOG.org

als. None of the trials reported using an FIGURE intention-to-treat analysis; therefore, we Summary of trial flow have reported the numbers allocated, numbers analyzed, and the percentage of women who dropped out after allocation. No standard deviations were reported in any of the trials; hence, we were unable to calculate confidence intervals. Where available, we have reported P values. We obtained unpublished end-of- treatment data for Elden et al’s trial from the chief investigator. We were unable to establish communication with researchers from the other trials. Wedenberg et 33 alreported acupuncture to be superior to individualized physiotherapy in their small trial. It is important to note here that the acupuncture group received auricular acupuncture plus body acupuncture “if needed.” It was unclear how many patients received auricular acupuncture alone, and if the acupuncture group received any cointerventions. Duration of treatment and time of follow-up varied between the 2 groups. The trial’s major flaw was that large numbers of women dropped out of Ee. Acupuncture in treating pelvic and back pain in pregnancy. Am J Obstet Gynecol 2008. the physiotherapy group. Kvorning et 34 alcompared an -acu puncture group with a control group clinical heterogeneity and too few trials cluded in our review at the time of writing that received no additional treatment. were found. If appropriate, we intended (personal communication, Shu-Ming Cointerventions were allowed in both to generate sensitivity analysis and a fun- Wang, 2006). We contacted 12 acupunc- groups. The dropout rate was signifi- nel plot. ture researchers, and from this communi- cant in both groups, and we judged the cation we identified 1 other ongoing trial, trial to be underpowered. Interest- RESULTS which was investigating acupuncture for ingly, the acupuncture needles were The Figure summarizes the trial flow.pelvic We pain in pregnancy. No data were withdrawn after de qi was obtained, found 10 papers from our MEDLINE available from this trial either at time of whereas in accepted clinical practice search and 421 papers from searching writing (personal communication, Helen they are usually retained for 20 min- other databases. We excluded the vast ma- Elden, 28 March 2007). We retrieved full utes when treating pain. We noted the jority of papers after a careful initial screen text articles for 5 RCTs. risk of contamination in this trial, with the because we found them to be duplicates, We excluded 2 trials from the initial 5 authors reporting “12 patients were incor- not randomized controlled trials (RCTs), that were retrieved. One was inade- rectly included” because “rumors of suc- or not pertaining to acupuncture for pel- quately randomized (according to days cessful treatment . . . had made further po- vic/back pain in pregnancy. We found 1 of the week),31 and the other comparedtential 2 participants unwilling to accept the ongoing trial of acupuncture for back pain different types of acupuncture (superfi- risk of being randomized to the control in pregnancy (CRISP-Computer Retrieval cial vs deep stimulation of acupuncture group.” Although the authors report as- on Scientific Projects. Acupuncture and points) without a nonacupuncture con- sessing pain using a VAS, no data on VAS Low Back Pain during Pregnancy. Avail- trol group.32 We included the final- were3 tri published. It was reported that more able at: http://crisp.cit.nih.gov/crisp/als, with a total of 448 women analyzed, women in the acupuncture group reported CRISP_LIB.getdoc?textkeyϭ 7012732& in our review.33-35 As the trials were- aclin decrease in their pain at end of treatment, p_grant_numϭ 5R21AT00161302& ically heterogeneous and few in number, compared with the control group. p_queryϭ&ticketϭ27900274&p_audit_ we could not combine them in a The largest and highest quality trial session_idϭ 192845617&p_keywordsϭ . metaanalysis. in our review was a multicenter trial Retrieved July 14, 2006). The investigators Table 2 summarizes the characteristicsconducted by Elden 35et Womenal. did not have any data that could be in- and main findings of the 3 included tri- with well-defined pregnancy-related

256 American Journal of Obstetrics & Gynecology MARCH 2008 www.AJOG.org Obstetrics Reviews

TABLE 2 Study characteristics and main ﬁndings of randomized controlled trials of acupuncture for pelvic and back pain in pregnancy Participants

Numbers allocated, Age median (y), (numbers trimester, First author, analyzed), primipara/ Description of year [dropout rate] multipara (%) intervention Quality Main findings Wedenberg, 2000 Acupuncture 30 (28) [7%] 28.6, 2nd and Chinese acupuncture— Modified Jadad 3/5; Pain: Statistically significant reduction of 3rd, 30/70 auricular points plus Cochrane Back pain in both groups, with greater relief of body acupuncture Review Group (BRG) pain reported by the acupuncture group points if needed, for 4 3/12; overall quality (decrease in mean evening VAS from 7.4/10 weeks. Unclear if rated poor baseline to 1.7/10 end-of-treatment) cointerventions allowed compared with control group (6.6/10 in acupuncture baseline to 4.5/10 end-of-treatment). P Ͻ group. .01 in both groups. Control 30 (18) [40%] Physiotherapy plus Function: Nonstatistically significant physical therapies reduction of Disability Rating Index in (pelvic belt, warmth, acupuncture group, and increase in control massage, soft tissue group. mobilization- Other: 27/28 women in acupuncture group individualized) for 6-8 reported “good or excellent help from weeks. treatment” compared with 14/18 in control group. Statistical significance not provided...... Kvorning, 2004 Acupuncture 50 (37) [26%] 30, 3rd, unclear Chinese acupuncture Modified Jadad 3/5; Pain: More women in acupuncture group plus tender points, plus Cochrane BRG 6/ reported decrease in pain (60% acupuncture cointerventions as 12; overall quality group vs 14% control group, P Ͻ .01). desired. Duration of rated high Function: 43% of women in acupuncture treatment and end-of- group reported decrease in pain during treatment follow-up physical activity compared with 9% in point unclear. control group (P Ͻ .001). Control 50 (35) [30%] No additional Analgesic use: 14% in control group used treatment, but analgesics; no women in acupuncture group cointerventions needed analgesics (P Ͻ .05). (physiotherapy, analgesics, etc) allowed in both groups...... Elden, 2005 Acupuncture 125 (107) [14%] 30.3, 2nd, 27/73 Mixed Western and Modified Jadad 3/5; Pain (at 1 week): Statistically significant Chinese acupuncture Cochrane BRG 7/ difference in median evening VAS (0-100) at plus standard 12; overall quality 1 week between intervention and control treatment (see Control) rated high groups, with lowest VAS in acupuncture for 6 weeks. group (Acupuncture group: baseline 65, 1- Control 130 (108) [17%] Standard treatment week 31; Stabilizing exercises group: consisting of advice, baseline 60, 1-week 45; Control group: education, exercises, baseline 63, 1-week 58; P Ͻ .001 for pelvic belt for 6 weeks. acupuncture-control comparison; P Ͻ .05 for Stabilizing 131 (106) [19%] Stabilizing exercises stabilizing exercises-control comparison). exercises plus standard Pain (at end-of-treatment): Median evening treatment for 6 weeks. VAS at end-of-treatment was lowest in acupuncture group (35) compared with control group (59) and stabilizing exercises group (50). Statistical significance not provided and unable to be calculated from median values given. Other: Fewer women in acupuncture group complained of pain when turning in bed at end-of-treatment compared with those in the control and stabilizing exercises group (66% acupuncture group vs 88% control group vs 71% stabilizing exercises group; P Ͻ .001 for acupuncture-control comparison). Ee. Acupuncture in treating pelvic and back pain in pregnancy. Am J Obstet Gynecol 2008.

MARCH 2008 American Journal of Obstetrics & Gynecology 257 Reviews Obstetrics www.AJOG.org pelvic pain in their second trimester increase in lumbar lordosis, muscular in- other aspects.44 This would control for were randomized into 3 groups. All sufficiency)5,37,38 causing pelvic and/orpatient expectations, but it is probably groups received standard treatment, back pain appear similar. At any rate, the impossible to blind needle acupuncture with 1 group receiving acupuncture as etiology and treatment of the syndromes practitioners. an adjunct and another receiving sta- remains a matter of debate, and we de- Within these limitations, our review bilizing exercises as an adjunct. The re- cided to include both syndromes in our method was robust and its main findings searchers reported median instead of review. are consistent with Young and Jewell’s mean values. The combination of acu- A limitation of our review was the few 2002 Cochrane review13 which included puncture and standard treatment was trials found in our search. However, the Wedenberg’s trial. Our results are also found to be the superior treatment in search was comprehensive—we used the consistent with the hypothesis that acu- this trial, at both the 1 and 6 weeks fol- Cochrane highly sensitive Medline puncture relieves pain and that it is a rel- 25 low-up visits. search strategy, contacted key -acuatively safe procedure. Few and minor No serious adverse events were re- puncture researchers, and searched trial adverse events were reported, which may ported across the 3 trials. There were databases. Its main limitation was the be reassuring for practitioners who feel less than 6 reports each of minor ad- English language restriction, hence ex- nervous about treating pregnant verse events, such as local pain or cluding most studies from China. We felt women. Though the sample sizes were bruising, sweating, nausea, weakness this would not impact greatly, as there is relatively small, they add to the findings and tiredness in Kvorning 34 ettrial. al’sa significant bias toward publishing pos- 33 39 from Ternov and colleagues’ case series Wedenberg et alreported that “mostitive findings in Chinaand studies tendof 167 women treated in pregnancy with women” complained of tiredness and 2 to suffer from poor design. It is interest- no serious adverse events.15 women reported subcutaneous auricu- ing to note that all 3 trials were con- Overall, our review finds limited, lar hematomata. We judged the acu- ducted in Sweden, although we see no though promising, evidence for the ef- puncture regime to be adequate in all 3 reason why these results should not ap- fectiveness of acupuncture in managing trials. ply in other countries. pelvic and back pain in pregnancy. In A third problem was that no trials in- particular, there seems to be good evi- cluded in our review had a placebo acu- dence that acupuncture, in addition to COMMENT puncture arm. The issues of placebo and standard treatment, is superior to stan- Limited evidence suggests that acupunc- nonspecific effects in acupuncture are dard treatment alone and stabilizing ture given in addition to standard treat- complex and have been extensively de- exercises for well-defined pregnancy-re- ment is more effective than standard bated in the literature.30,40 A detailed lated pelvic pain. Given that acupunc- treatment alone, physiotherapy, or stabi- discussion is beyond the scope of this pa- ture is a relatively safe procedure, these lizing exercises in relieving pelvic and per. The issues relevant to this review are back pain in pregnancy. Although the whether the expectations, beliefs, and findings should encourage primary trials found in our review were small in subsequent behaviors of patients and/or health care providers, obstetricians, and number and clinically heterogeneous, practitioners influenced outcomes, as midwives to consider referring women the trial by Elden35 iset well-conal - none of the trials were blinded. As such, to trained acupuncturists for manage- ducted and provides good evidence for our findings reflect the overall effective- ment of this common, painful, and dis- the effectiveness of acupuncture in preg- ness (specific and nonspecific effects) of abling condition. Stronger evidence is nancy-related pelvic pain. acupuncture for pregnancy-related pel- needed, and we look forward to the re- Difficulties in this review included vic and back pain, rather than its efficacy sults of the 2 ongoing trials identified. making the decision to include both pel- (specific effects). Additionally, we feel that a consensus as vic and back pain in pregnancy. There is It has been demonstrated that both pa- to the nature, etiology, and “standard disagreement in medical circles as to tient expectation and practitioner be- treatment” of pregnancy-related pelvic whether pelvic and back pain in preg- havior can result in greater placebo anal- and back pain is essential to prevent fur- nancy are separate clinical entities. Some gesia.41,42 In fact, some have postulatedther dilution of the evidence through experts believe that the 2 can and should that acupuncture may have a “placebo- heterogeneity. Systematic reviews of be distinguished clinically and that they enhancing effect.”43 Kvorning et al- acupuncturere frequently conclude with a respond to different treatments and have ported that 12 women had to be ex- recommendation for more trials. How- different risk factors.3 The counter argu- cluded because rumors were circulating ever, it is not simply a matter of num- ment has been that the prognosis for about the success of acupuncture treat- bers. Larger trials would be a start. There both is similar, previous studies have not ment; the expectancy of the acupuncture is also a pressing need for these to be been able to convincingly distinguish be- group in this trial may have overinflated based on rigorous methodology and a tween the 2, and that back pain probably the positive results. “Placebo needles” careful consideration of the issues inher- forms a subset of pregnancy-related pel- have been developed that do not involve ent in acupuncture research, such as its vic pain.36 In our view, the proposedpenetration of the skin or induce de qi, nonspecific effects, and reported accord- physiological mechanisms (joint laxity, but simulate true acupuncture in most ing to the STandards for Reporting In-

258 American Journal of Obstetrics & Gynecology MARCH 2008 www.AJOG.org Obstetrics Reviews terventions in Controlled Trials of Acu- 16. Cummings M. Acupuncture for low back pain after acupuncture: a randomized puncture (STRICTA) guidelines.45 f pain in pregnancy. Acupunct Med 2003;21: controlled single-blind study. Acta Obstet Gy- 42-6. necol Scand 2006;85:12-9. 17. Forrester M. Low back pain in pregnancy. 33. Wedenberg K, Moen B, Norling A. A pro- REFERENCES Acupunct Med 2003;21:36-41. spective randomized study comparing acu- 1. Zib M, Lim L, Walters WA. Symptoms during 18. Barnes PM, Powell-Griner E, McFann K, puncture with physiotherapy for low-back and normal pregnancy: a prospective controlled Nahin RL. Complementary and alternative med- pelvic pain in pregnancy. Acta Obstet Gynecol study. AustNZJObstet Gynaecol 1999;39: icine use among adults: United States, 2002. Scand 2000;79:331-5. 401-10. Adv Data 2004;343:1-19. 34. Kvorning N, Holmberg C, Grennert L, Aberg 2. Mogren IM, Pohjanen AI. Low back pain and 19. Pinn G, Pallett L. Herbal medicine in preg- A, Jonas A. Acupuncture relieves pelvic and pelvic pain during pregnancy: prevalence and nancy. Complement Ther Nurs Midwifery low-back pain in late pregnancy. Acta Obstet risk factors. Spine 2005;30:983-91. 2002;8:77-80. Gynecol Scand 2004;83:246-50. 3. Ostgaard HC, Zetherstrom G, Roos-Hans- 20. Wang S-M, DeZinno P, Fermo L, et al. 35. Elden H, Ladfors L, Olsen MF, Ostgaard son E, Svanberg B. Reduction of back and pos- Complementary and alternative medicine for H-C, Hagberg H. Effects of acupuncture and terior pelvic pain in pregnancy. Spine 1994;19: low-back pain in pregnancy: a cross-sectional stabilising exercises as adjunct to standard 894-900. survey. J Altern Complement Med 2005;11: treatment in pregnant women with pelvic girdle 4. Wu WH, Meijer OG, Uegaki K, et al. Pregnan- 459-64. pain: randomised single blind controlled trial. cy-related pelvic girdle pain (PPP), I: terminol- 21. Stux G, Pomeranz B. Scientific bases of ogy, clinical presentation, and prevalence. Eur BMJ 2005;330:761. acupuncture. New York: Springer-Verlag; 36. Bastiaanssen JM, de Bie RA, Bastiaenen Spine J 2004;13:575-89. 1991. 5. MacEvilly M, Buggy D. Back pain and preg- CHG, Essed GGM, van den Brandt PA. A his- 22. Sandkuhler J. Learning and memory in pain torical perspective on pregnancy-related low nancy: a review. Pain 1996;64:405-14. pathways. Pain 2000;88:113-8. 6. Wu WH, Meijer OG, Uegaki K, et al. Preg- back and/or pelvic girdle pain. Eur J Obstet Gy- 23. White A, Hayhoe S, Hart A, Ernst E. Adverse nancy-related pelvic girdle pain (PPP), I: termi- necol Reprod Biol 2005;120:3-14. events following acupuncture: prospective sur- nology, clinical presentation, and prevalence. 37. Vleeming A, de Vries HJ, Mens JMA, van vey of 32 000 consultations with doctors and Eur Spine J 2004;13:575-89. Wingerden J-P. Possible role of the long dorsal physiotherapists. BMJ 2001;323:485-6. 7. Wang SM, Dezinno P, Maranets I, Berman sacroiliac ligament in women with peripartum 24. Macpherson H, Scullion A, Thomas KJ, MR, Caldwell-Andrews AA, Kain ZN. Low back pelvic pain. Acta Obstet Gynecol Scand Walters S. Patient reports of adverse events as- pain during pregnancy: prevalence, risk factors 2002;81:430-6. sociated with acupuncture treatment: a pro- and outcomes. Obstet Gynecol 2004;104: 38. Mogren IM. Previous physical activity de- spective national survey. Qual Saf Health Care 65-70. creases the risk of low back pain and pelvic pain 2004;13:349-55. 8. Hansen A, Jensen DV, Wormslev M, et al. during pregnancy. Scand J Public Health 25. Robinson KA, Dickersin K. Development of Pregnancy associated pelvic pain. II: symptoms 2005;33:300-6. a highly sensitive search strategy for the retrieval and clinical findings. Ugeskr Laeger 2000;162: 39. Vickers A, Goyal N, Harland R, Rees R. Do of reports of controlled trials using PubMed. Int 4813-7. certain countries produce only positive results? J Epidemiol 2002;31:160-4. 9. Larsen EC, Wilken-Jensen C, Hansen A, et A systematic review of controlled trials. Control 26. Birch S. Systematic reviews of al. Pregnancy associated pelvic pain. I: preva- Clin Trials 1998;19:159-66. acupuncture—are there problems with these? lence and risk factors. Ugeskr Laeger 2000; 40. Birch S. Clinical research on acupuncture. Clinical Acupuncture & Oriental Medicine 162:4808-12. Part 2. Controlled clinical trials, an overview of 10. Stones RW, Vits K. Pelvic girdle pain in 2001;2:17-22. 27. White P, Lewith G, Berman B, Birch S. Re- their methods. J Altern Complement Med pregnancy. BMJ 2005;331:249-50. 2004;10:481-98. 11. Stuge B, Hilde G, Vollestad N. Physical views of acupuncture for chronic neck pain: pit- falls in conducting systematic reviews. Rheu- 41. Gracely RH, Dubner R, Deeter WR, Wols- therapy for pregnancy-related low back and kee PJ. Clinicians’ expectations influence pla- pelvic pain: a systematic review. Acta Obstet matology (Oxford) 2002;41:1224-31. cebo analgesia. Lancet 1985;1:43. Gynecol Scand 2003;82:983-90. 28. Manheimer E, White A, Berman B, Forys K, 42. Pariente J, White P, Frackowiak RSJ, 12. Depledge J, McNair PJ, Keal-Smith C, Wil- Ernst E. Meta-analysis: acupuncture for low Lewith G. Expectancy and belief modulate liams M. Management of symphysis pubis dys- back pain. Ann Intern Med 2005;142:651-63. the neuronal substrates of pain treated function during pregnancy using exercise and 29. van Tulder M, Furlan A, Bombardier C, by acupuncture. Neuroimage 2005;25: pelvic support belts. Phys Ther 2005;85: Bouter L. Updated method guidelines for sys- 1161-7. 1290-300. tematic reviews in the Cochrane Collaboration 13. Young G, Jewell D. Interventions for pre- Back Review Group. Spine 2003;28:1290-9. 43. Lewith GT, White PJ, Kaptchuk TJ. Devel- venting and treating pelvic and back pain in 30. White AR, Filshie J, Cummings TM. Clinical oping a research strategy for acupuncture. Clin pregnancy. Cochrane Database Syst Rev trials of acupuncture: consensus recommenda- J Pain 2006;22:532-8. 2002(1):CD001139 tions for optimal treatment, sham controls and 44. Streitberger K, Kleinhenz J. Introducing a 14. Thomas CT, Napolitano PG. Use of acu- blinding. Complement Ther Med 2001;9: placebo needle into acupuncture research. puncture for managing chronic pelvic pain in 237-45. Lancet 1998;352:364-5. pregnancy. A case report. J Reprod Med 31. Guerreiro da Silva JB, Nakamura MU, Cor- 45. MacPherson H, White A, Cummings M, 2000;45:944-6. deiro JA, Kulay LJ. Acupuncture for low back Jobst K, Rose K, Niemtzow R. Standards for 15. Ternov NK, Grennert L, Aberg A, Algotsson pain in pregnancy—a prospective, quasi-ran- reporting interventions in controlled trials of L, Akeson J. Acupuncture for lower back and domised, controlled study. Acupunct Med acupuncture: the STRICTA recommendations. pelvic pain in late pregnancy: a retrospective 2004;22:60-7. STandards for Reporting Interventions in Con- report on 167 consecutive cases. Pain Med 32. Lund I, Lundeberg T, Lonnberg L, Svens- trolled Trials of Acupuncture. Acupunct Med 2001;2:204-7. son E. Decrease of pregnant women’s pelvic 2002;20:22-5.

MARCH 2008 American Journal of Obstetrics & Gynecology 259