lable at ScienceDirect

Manual Therapy 19 (2014) 190e196

Contents lists avai

Manual Therapy

journal homepage: www.elsevier .com/math

Original article

Clinical course of pelvic girdle pain postpartum e Impact of clinicalfindings in late pregnancy

Hilde Stendal Robinson a,*, Nina K. Vøllestad a, Marit B. Veierød b

aDepartment of Health Sciences, Institute of Health and Society, University of Oslo, P.O. Box 1089, Blindern, 0317 Oslo, NorwaybDepartment of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, P.O. Box 1122, Blindern, 0317 Oslo, Norway

a r t i c l e i n f o

Article history:Received 12 June 2013Received in revised form6 January 2014Accepted 13 January 2014

Keywords:Lumbopelvic painLow back painPostpartumClinical tests

* Corresponding author. Tel.: þ47 22 84 53 94; faxE-mail address: h.s.robinson@medisin.uio.no (H.S.

1356-689X/$ e see front matter � 2014 Elsevier Ltd.http://dx.doi.org/10.1016/j.math.2014.01.004

a b s t r a c t

The aims were to study: prevalence of pelvic girdle pain (PGP) one year postpartum; clinical course ofPGP, physical functioning (PF) and bodily pain (BP) (from SF-36, 0 (worst) to 100 (best)) from gestationweek (GW) 30 to one year postpartum; and whether findings at GW30 were associated with develop-ment of PF and BP from GW30 to one year postpartum.

215 pregnant women were followed from GW30 to one year postpartum. Clinical examination andquestionnaire were used at GW30, questionnaire only were used at 12 weeks and one year postpartum.The women were categorised by GW30 clinical variables: self-reported PGP, pain locations in the pelvisand response to two clinical tests. Linear mixed models for repeated measures were used to study PF andBP during follow-up, within the categories of clinical variables.

PGP prevalence remained unchanged from 12 weeks to one year postpartum (31e30%). PF and BPscores improved markedly from GW30 to 12 weeks postpartum, and marginally thereafter. Median PFscores were 70, 95 and 100 at GW30, 12 weeks and one year postpartum, respectively. Correspondingmedian BP scores were 52, 84 and 84. We found significant interactions between each clinical variableand time (P � 0.01) for PF and BP. The most afflicted women at GW30 experienced largest improvement.

Despite high PGP prevalence one year postpartum, most women recovered in terms of PF and BPscores. Unfavourable clinical course postpartum did not appear to depend on self-reported PGP, painlocations in the pelvis, or response to clinical tests at GW30.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Pelvic girdle pain (PGP) is common during pregnancy, with areported prevalence from 20% to above 50% depending on the casedefinition (Olsson and Nilsson-Wikmar, 2004; Gutke et al., 2006;Mogren, 2006; Robinson et al., 2006; Vleeming et al., 2008; Bjel-land et al., 2010; Robinson et al., 2010a). Pain is usually locatedbetween the posterior iliac crest and the gluteal fold, predomi-nantly around the sacroiliac joints and may also include pain in thesymphysis (Vleeming et al., 2008). PGP has been associated withreduced capacity for weight-bearing activities such as walking andstanding (Rost et al., 2006; Robinson et al., 2006, 2010c). Althoughthe severity of PGP, in terms of disability or pain is modest in mostwomen, a considerable fraction does report severe disability(Olsson and Nilsson-Wikmar, 2004; Gutke et al., 2006; Robinsonet al., 2010a; Mens et al., 2012b). Several studies have also reportedthat PGP prevalence declines markedly in the first months

: þ47 22 84 50 91.Robinson).

All rights reserved.

postpartum (Albert et al., 2001; Mogren, 2006; Gutke et al., 2008;Robinson et al., 2010b), but the clinical course of PGP in longerfollow-up has been the object of few studies. One study found that8.5% of the women with PGP in late pregnancy reported PGP twoyears postpartum (Albert et al., 2001).

Previous studies on the clinical course of PGP in the first weekspostpartum have used slightly different criteria, but were basedmostly on PGP prevalence (Albert et al., 2001; Mogren, 2006; Gutkeet al., 2008). It has been reported that pain locations and responsesto clinical tests are associated with PGP prevalence, disability andpain intensity postpartum (Albert et al., 2001; Gutke et al., 2008;Robinson et al., 2010b). Albert et al. (2001) found that womenwith combined pain in the symphysis and posterior parts of thepelvis during pregnancy recovered to a lesser extent two years afterdelivery than womenwith fewer pain locations. Gutke et al. (2008)found that women with combined low back pain and PGP inpregnancy had a less favourable course till three months post-partum. We previously reported a low level of disability and painintensity 12 weeks postpartum, despite a PGP prevalence of 31%(Robinson et al., 2010b). However, 25% of these women had higher

H.S. Robinson et al. / Manual Therapy 19 (2014) 190e196 191

disability scores compared to healthy women of the same age withminor ailments (Salen et al., 1994), and 25% reportedmoderate painintensity. Based on these results it is of interest to explore theclinical course of PGP postpartum in the same cohort with a longerfollow-up time, in order to examine both long-term PGP preva-lence, and the degree of affliction (disability) as determined byphysical functioning and pain. Moreover, physical functioning andpain in the study sample one year after delivery should becompared with normative data from the general population.

The aims of this study were: 1) to determine the prevalence ofself-reported PGP one year postpartum, 2) to examine the clinicalcourse from gestation week (GW) 30 to one year postpartum in

Fig. 1. Flow chart of t

terms of prevalence of PGP, physical functioning and bodily painand 3) to examine whether presence of self-reported PGP, painlocations in the pelvis or responses to clinical tests at GW30 areassociatedwith the development of physical functioning and bodilypain over time from GW30 to one year postpartum.

2. Materials and methods

This paper is based on a prospective cohort study of pregnantwomen who were followed up from early pregnancy to one yearpostpartum (Robinson et al., 2010b, 2010c). The Regional Com-mittee for Medical Research Ethics and the Norwegian Social

he study sample.

H.S. Robinson et al. / Manual Therapy 19 (2014) 190e196192

Science Data Services gave formal approval of the study (referencenumber: S-05284, approved on 20 December 2005). All partici-pants gave written informed consent.

2.1. Procedures

Most women in Norway attend maternity care units (MCUs) forhealth services during pregnancy. Midwives and staff at four MCUslocated in, and nearby the capital Oslo handled the recruitmentprocess. All eligible Norwegian-speaking pregnant women(n¼385), were invited to participate in the study at their firstcontact with the MCUs, between January 2006 and June 2007.Women not expected to have a normal pregnancy (as determinedby the midwife) were excluded from participation. A total of 326women agreed to participate. Mean age at study enrolment was 31years (range 18e45 years) and 60% of participants were nulliparous.Data used in the present paper were collected by questionnairesand clinical examinations at GW30, and by questionnaires alone at12 weeks and one year postpartum. Out of the 326 womenrecruited to the cohort, 283 were examined in GW30 and 233 ofthese returned the questionnaire at one year postpartum. However,18 of these were excluded due to a new pregnancy and 215 womenconstituted the study sample (Fig. 1).

2.1.1. Outcome variablesPresence of PGP was assessed at GW30, 12 weeks and one year

postpartumbya simple question in the questionnaire: “Do youhavepain in your pelvic girdle?” (yes, no). The Short Form-36 (SF-36)wasalso filled in at GW30, 12 weeks and one year postpartum (Norwe-gian version 1.1) (Loge and Kaasa, 1998; Loge et al., 1998). Thisgeneric instrument covers central health aspects and canbe used forhealthy as well as patient groups (Ware, 2000). Disability and painare complaints associatedwith PGP (Vleeming et al., 2008), thus weapplied the subscales of physical functioning and bodily pain in thisstudy. These scales range from 0 (worst physical functioning/bodilypain) to 100 (best physical functioning/bodily pain).

2.1.2. Other variablesAge, education, employment and parity were recorded at study

enrolment (mean GW15). Pain locations in the pelvis were derivedfrom pain drawings. The active straight leg raise (ASLR) test andposterior pelvic pain provocation (P4) test were performed at clin-ical examinations at GW30. Active lifting of one extended leg at atime was performed and scored by the participant on a 6-pointLikert scale (0 ¼ not difficult to lift to 5 ¼ impossible to lift theleg). The scores for the two legs were added, thus the ASLR scoreranges from0 to 10. A sum score over 0was defined as a positive test(Mens et al., 2001, 2002). High sensitivity and specificity have beenreported for theASLR test inpregnancywith this cut-off value (Menset al., 2012a). The P4 testwasperformedwith thewoman supine andthe hip and knee flexed to 90�. Once in this position, the examinerapplied pressure, pushing the knee into the pelvis through thelongitudinal axis of the femur (Ostgaard et al.,1994). The participantreported whether this provoked a familiar pain in the posteriorpelvis (yes, no). Both sides were tested and the response to the P4test could thus be negative, unilateral positive or bilateral positive.The P4 test is reported to be reliable andhavehigh validity for PGP inpregnancy (Ostgaard et al., 1994) Both the ASLR test and the P4 testare widely used in clinical practice to examine potential PGP pa-tients as well as in research (Ostgaard et al., 1994; Mens et al., 2001,2002; Robinson et al., 2007; Gutke et al., 2009; Mens et al., 2010,2012a). Both tests are reported to distinguish between PGP andLBP (Ostgaard et al., 1994; Gutke et al., 2009; Mens et al., 2012a).

2.1.3. Statistical analysesMissing in SF36 was handled according to the standard proce-

dure for SF36 (Loge et al., 1998). A total of 17 participants hadmissing items on SF36: 4 at GW30, 7 at 12 weeks postpartum and 7oneyear postpartum(11of theseparticipantsmissedonlyone item).

Descriptive data are presented as frequencies, percentages,means with standard deviations and medians with first (Q1) andthird (Q3) quartiles. Expected age-adjusted mean scores of physicalfunctioning and bodily pain for the study sample were estimated asdescribed by Fayers and Machin (2007, page 436) using the Nor-wegian population norms for women in the age groups 19e29, 30e39 and 40e49 years (Loge and Kaasa, 1998) (corresponding to theage groups 18e29, 30e39 and 40e45 years in our study sample).

Study women were categorised according to clinical variables atGW30: 1) presence of PGP (yes, no), 2) pain locations in thepelvis (nopain, symphysis pain only, posterior pain only, combined symphysisand posterior pain), 3) P4 response (negative, unilateral positive,bilateral positive) and 4) ASLR score (negative ¼0, positive >0).

A linear mixed model for repeated measures (unstructuredcovariance matrix) was used to study the evolution of physicalfunctioning and bodily pain scores, within the categories of clinicalvariables over the three time points considered. Physical func-tioning and bodily pain scores were loge transformed in these an-alyses, while trends over time are illustrated by box plots on theoriginal scale.

All analyses were performed using SPSS (version 19) and a 5%level of significance was used.

3. Results

The study sample was similar to the total cohort and the womenbeing examined at GW30 in terms of age, education, marital status,employment and parity (Table 1), and in terms of the prevalence ofself-reported PGP and median physical functioning and bodily painscores at GW30 (Table 2). As compared to the study sample, drop-outs tended to have similar age, similar length of education andsimilar parity, while some discrepancies were found for maritalstatus, smoking and employment (Table 1).

Prevalence of self-reported PGP in the study sample declinedfrom 63% at GW30 to 31% and 30% 12 weeks and one year post-partum, respectively (Table 2). Sixty-nine percent of the womenwho reported PGP at 12weeks postpartum also reported PGP at oneyear postpartum, and 6 of thewomen (3%) who reported PGP at oneyear postpartum had not reported PGP either in pregnancy or 12weeks postpartum. These 6 women had comparable scores onphysical function with the rest of the study sample at 12 weeks(median 92 and 95, respectively) and one year postpartum (98 and100, respectively), but they reported less bodily pain comparedwith the rest of the study sample (92 and 84, respectively). Com-bined symphysis pain and bilateral posterior painwere reported by53, 12 and 5 women at GW30, 12 weeks and one year postpartum,respectively.

Physical functioning and bodily pain scores improved markedlybetween GW30 and 12 weeks postpartum, and remained high atone year postpartum. The observed means of these scores one yearpostpartumwere slightly higher (95 for physical functioning and 83for bodily pain) than those reported for women aged 19e49 yearsin the general Norwegian population (Table 2), and slightly higherthan the expected age-adjusted mean scores for the study sample,based on the norms for the general Norwegian population (92 and77, respectively).

The linear mixed model analyses of physical functioning andbodily pain scores showed significant interactions between theclinical variable and time for all four considered clinical variables(p � 0.01). The womenwith the lowest scores (highest affliction) at

Table 1Selected characteristics of the study sample, the total cohort, the women examined at GW30 and missing women (drop-outs) at one year postpartum.

Study sample n¼215 Total cohort n¼326 Women examined atGW30 n¼283

Drop-outs from GW 30 till1 year postpartum n¼47

Mean (SD) Mean (SD) Mean (SD) Mean (SD)

Age (years) 31 (4) 31 (4) 31 (4) 31 (4)Education (years) 16 (3) 16 (2) 16 (3) 16 (2)

n (%) n (%) n (%) n (%)Married/cohabitant (yes) 211 (98) 316 (97) 275 (97) 44 (94)Smoking (yes) 6 (3) 15 (5) 12 (4) 4 (9)Employed, yes 203 (94) 393 (93) 261 (92) 38 (81)Parity 0 121 (56) 196 (60) 167 (59) 28 (60)

1 75 (35) 103 (32) 92 (33) 15 (32)�2 20 (9) 27 (8) 24 (8) 4 (9)

GW: gestation week, SD: standard deviation.

H.S. Robinson et al. / Manual Therapy 19 (2014) 190e196 193

GW30 showed the largest changes across the time periods. Thedifference in scores across the categories of clinical variables nar-rowed 12 weeks postpartum, with small changes observed be-tween 12 weeks and one year postpartum (Figs. 2 and 3).

Some differences between categories of clinical findings wereseen at GW30. Women that reported PGP or some pain locations inthe pelvis had lower median physical functioning and bodily painscores (were more afflicted) compared to those without pain(Figs. 2 and 3). Similar differences in physical functioning andbodily pain scores were seen in women with and without positiveresponse to the P4 and ASLR tests.

At 12 weeks postpartum, the median scores for physical func-tioning and bodily pain were quite similar in all categories ofclinical variables, except for bodily pain scores in the variable painlocations in the pelvis (Figs. 2 and 3). The women with combinedsymphysis and posterior pain had a median bodily pain score of 72(Q1 ¼ 62, Q3 ¼ 84) whereas for those with no pain locations in thepelvis the corresponding value was 100, i.e. no pain (Q1 ¼ 84,Q3 ¼ 100).

At one year postpartum, the median (Q1, Q3) scores across thecategories of clinical variables for physical functioning were verysimilar (Fig. 2). Median bodily pain scores were lowest (i.e. highestaffliction) for women reporting PGP, pain locations in the pelvis orhaving positive responses to the P4 and ASLR tests (Fig. 3).

4. Discussion

We found large improvements in physical functioning andbodily pain across the three time points considered in our analysis(GW30, 12 weeks and one year postpartum) in this sample ofwomen. In general, physical functioning and bodily pain scoreswere high one year postpartum, irrespective of the clinical variablesconsidered, i.e., pain locations in the pelvis, or responses to clinical

Table 2The clinical course of pelvic girdle pain (PGP) prevalence, and physical functioning (PF) andin the study sample. Data at gestation week 30 for the entire cohort and the general Norw

Study sample (n¼215) Total cohort (

Gestationweek 30

12 weekspostpartum

One yearpostpartum

Gestation we

PGP, yes n (%) 136 (63) 66 (31) 64 (30) 193 (62)PF score Mean (SD) 67 (20) 92 (13) 95 (9) 100 67 (21)

Median (Q1,Q3) 70 (55, 85) 95 (90, 100) (95,100) 70 (55, 80)BP score Mean (SD) 57 (23) 78 (22) 83 (20) 57 (23)

Median (Q1,Q3) 52 (41, 74) 84 (62, 100) 84 (72, 100) 52 (41, 74)

PGP: pelvic girdle pain, PF score: physical functioning score, BP score: bodily pain scorePF and BP, 0e100 scales; 0 ¼ worst physical functioning/bodily pain, 100 ¼ best physica

a Loge and Kaasa (1998).

tests at GW30. Importantly, the significant interaction effects be-tween the clinical variables and time indicated that the largestimprovements postpartum were seen among the women who re-ported the highest affliction in late pregnancy.

The observed mean scores for physical functioning and bodilypain postpartumwere slightly higher than mean scores for womenof the same age in the general Norwegian population (Loge andKaasa, 1998). One explanation for the higher scores (i.e. less painand better physical function) in our study sample than in thegeneral population might be the different age distribution withinthe age groups. For instance, compared to the general populationage group of 40e49 years, the corresponding age group in oursample was in the young range (40e45 years). Moreover, it hasbeen shown that physical functioning and pain reported by patientstend to be more favourable than those reported by the generalpopulation (Fayers and Machin, 2007). The same response shiftmay also exist among the afflicted women in our sample. However,the differences between the observed and the estimated means (3for physical functioning and 6 for bodily pain) were small when thevariability in responses was taken into account.

We have previously reported that 62% of the women in thissame cohort reported PGP at GW30 (Robinson, 2010), and impor-tantly, only 31% reported PGP at 12 weeks postpartum (Robinsonet al., 2010b). These results are in concordance with previousstudies, showing that the PGP prevalence declines most markedlyin the first months postpartum (Albert et al., 2001; Noren et al.,2002; Rost et al., 2006). Only 5 women reported combined sym-physis pain and bilateral posterior pelvic pain, which is in agree-ment with the results reported in a large Norwegian populationbased study (Bjelland et al., 2013). Having pain in all these threelocations has previously been termed pelvic girdle syndrome andused to identify those with a severe condition (Albert et al., 2001;Robinson et al., 2006). Hence, our data indicate that 2e3% of the

bodily pain (BP) scores at gestationweek 30, and 12weeks and one year postpartumegian population data (Loge and Kaasa, 1998) for women are given for comparison.

n¼326) General Norwegian population (women)

ek 30 19e29 yearsa 30e39 yearsa 40e49 yearsa Expected age-adjustedmean scores

94 (11) 92 (13) 89 (17) 92

80 (23) 77 (25) 74 (26) 77

from Short form 36.l functioning/bodily pain.

Fig. 2. Box-plots of physical functioning as reported in the SF-36 at gestation week 30, 12 weeks and one year postpartum. The study sample (n¼215) was categorised according topresence of pelvic girdle pain, pain locations in the pelvis, posterior pelvic pain provocation (P4) test and active straight leg raise (ASLR) test at gestation week 30. Median, quartilesand range are presented. The length of the box is the distance between the first (Q1) and third (Q3) quartile, i.e., the interquartile range (IQR). Circles and asterisks represent outliers(>1.5 IQR and >3 IQR above the third quartile, respectively).

H.S. Robinson et al. / Manual Therapy 19 (2014) 190e196194

women have severe PGP one year postpartum. Since the proportionof women with PGP is stable from 12 weeks till one year post-partum, a transition to a more chronic PGP status seems to occuraround or before 12 weeks postpartum. This is in accordance withstudies on low back pain that defined 3 months as the “limit” forthe condition to be defined as chronic (Frank et al., 1996; Waddell,2004). However, the proportion of women with PGP seems sur-prisingly high one year postpartum. The lack of decline in preva-lence of PGP might be due to altered demands when the womenshift from being in early postpartum period to a life combiningwork, leisure time and family.

A different pattern is seen when using data on physical func-tioning and bodily pain. Both scores improved markedly and werecomparable with expected age-adjusted mean scores within thefirst 12 weeks postpartum. The scores remained high until one yearpostpartum and only small changes were observed. Relatively largevariations in physical functioning and bodily pain scores werefound in categories of clinical variables, but 75% of the women hadphysical functioning scores above 95 (Q1 ¼ 95) and 75% had bodilypain scores above 72 (Q1 ¼ 72). Furthermore, clinical variables inlate pregnancy do not seem to be associatedwith the clinical courseof PGP postpartum. There were few women in each of the

categories of clinical variables scoring lower than population normsof physical functioning and bodily pain one year postpartum, butsome of these women scored markedly lower.

The results indicated that most women did recover from PGP,even though 30% still reported having the condition one yearpostpartum. One possible explanation for this is that women didnot experience their pain as bothersome anymore, but insteadexperienced it more like discomfort. Yet, when asked specifically,they reported discomfort as pain. Hence, the dichotomous questionabout having PGP could be less sensitive than a continuous mea-sure as the SF-36. It might also be explained as an adaptation to thesituation or a change in reporting behaviour. Both explanationsimply that the affliction becomes less important to the person.According to Schwartz et al. (2005), adaptation to PGP can suggestan experienced alteration in health during and after pregnancy,which changed the woman’s internal standards and values, andthus their reporting behaviour. Furthermore, physical functioningand bodily pain may interact. For example, a reduction in physicalfunctioning scores may influence, and thus reduce bodily painscores or vice versa. Length of pregnancy and personal experienceof pregnancy and birth may also recalibrate a woman’s painthreshold, or idea of pain, and result in an adjustment to the

Fig. 3. Box-plots of bodily pain at gestation week 30, 12 weeks and one year postpartum. The study sample (n¼215) was categorised according to presence of pelvic girdle pain, painlocations in the pelvis, posterior pelvic pain provocation (P4) test and active straight leg raise (ASLR) test at gestation week 30. Median, quartiles and range are presented. The lengthof the box is the distance between the first (Q1) and third (Q3) quartile, i.e., the interquartile range (IQR). Circles and asterisks represent outliers (>1.5 IQR and >3 IQR above thethird quartile, respectively).

H.S. Robinson et al. / Manual Therapy 19 (2014) 190e196 195

situation. The results underpin the importance of evaluatingaffliction together with the prevalence of PGP.

The prevalence of PGP one year postpartum in this study wasclose to the point prevalence reported for LBP in general pop-ulations (Waddell, 2004), indicating that this could be the sameentity. However, previous studies have reported that the responseon the clinical tests used here, are different in patients with LBP andPGP (Ostgaard et al., 1994; Gutke et al., 2009; Mens et al., 2012a).Furthermore, we have previously found that response on the P4and ASLR tests contributed independently to disability in multi-variable analysis during pregnancy (Robinson et al., 2010a). Thus, itseems more likely that PGP and LBP are different entities inconnection to pregnancy. However, retrospective studies on LBPhave reported that several women with chronic LBP identifypregnancy as an initial appearance of the pain (Biering-Sorensen,1983; Svensson et al., 1990). Knowledge of the association be-tween PGP and LBP in women throughout life is lacking.

Six of thewomenwho reported PGP at one year postpartum (3%)did not report PGP at any of the previous follow-ups in the study,indicating that they developed PGP from 12 weeks till one yearpostpartum. Their scores on physical function were comparablewith the total cohort at 12 weeks and one year postpartum, butthey reported less bodily pain. Since no responses on clinical tests

are available at one year follow up, we cannot exclude LBP in thisgroup.

The prospective design and long follow-up is an importantstrength of this study. Furthermore, the use of pain drawings andclinical examinations at GW30 allowed us to examine the impact ofclinical findings in late pregnancy on the course until one yearpostpartum. As compared to the study sample, drop-outs weresimilar in most aspects, however, slightly fewer women among thedrop-outs were married/cohabitant, non-smokers and employed.These findings are consistent with non-compliance present in otherstudies (Galea and Tracy, 2007; Tough et al., 2009; Stacey et al.,2011). Self-reporting of physical functioning is a potential weak-ness. We do not knowwhether the reported physical functioning isreal, an adaptation to the situation, or a result of reporting behav-iour. This could have been explored further by use of tests forphysical function. Likewise, bodily pain is also self-reported andmight suffer from similar weakness.

5. Conclusion

Thirty percent of the women in this cohort reported PGP oneyear postpartum. Yet the validity of the single question for assess-ment of PGP one year postpartum may be questioned since most

H.S. Robinson et al. / Manual Therapy 19 (2014) 190e196196

women were fully recovered with regard to physical functioningand bodily pain. Difference between groups defined by PGP, painlocations in the pelvis and responses to clinical tests in late preg-nancy seem to have little influence on the clinical course of physicalfunctioning and bodily pain postpartum.

Acknowledgements

This study has been supported by the EXTRA funds from theNorwegian Foundation for Health and Rehabilitation, The Norwe-gian Fund for Post-Graduate Training in Physiotherapy and theUniversity of Oslo.

We thank Hans and Olaf Physiotherapy clinic and the MaternityCare Units (MCU) for kindly making it possible to collect the data inthis study. In particular we want to thank Professor Anne MaritMengshoel for valuable contribution in planning the study andElisabeth K Bjelland, PhD, RPT for valuable help with the clinicalexaminations of the participating women. Furthermore we want tothank Anne Karine Bergva, Sigrunn Anmarkrud, Grete Kristiansen,Hege Kaspersen, Astrid Stormoen, Eva Marie Flaathen, HeidiArnesen, Tove Mols and Wenche Sjøberg at the MCU for help withrecruiting the pregnant women. We also thank Trudy Perdrix-Thoma for language review.

References

Albert H, Godskesen M, Westergaard J. Prognosis in four syndromes of pregnancy-related pelvic pain. Acta Obstet Gynecol Scand 2001;80(6):505e10.

Biering-Sorensen F. A prospective study of low back pain in a general population. II.Location, character, aggravating and relieving factors. Scand J Rehabil Med1983;15(2):81e8.

Bjelland EK, Eskild A, Johansen R, Eberhard-Gran M. Pelvic girdle pain in pregnancy:the impact of parity. Am J Obstet Gynecol 2010;203(2):146.

Bjelland EK, Stuge B, Vangen S, Stray-Pedersen B, Eberhard-Gran M. Mode of de-livery and persistence of pelvic girdle syndrome 6 months postpartum. Am JObstet Gynecol 2013;208(4):298e307.

Fayers PM, David Machin. Quality of Life. The assessment, analysis and interpre-tation of patient-reported outcomes. Chichester, England: John Wiley & SonsLtd; 2007.

Frank JW, Brooker AS, DeMaio SE, Kerr MS, Maetzel A, Shannon HS, et al. Disabilityresulting from occupational low back pain. Part II: what do we know aboutsecondary prevention? A review of the scientific evidence on prevention afterdisability begins. Spine (Phila Pa 1976) 1996;21(24):2918e29.

Galea S, Tracy M. Participation rates in epidemiologic studies. Ann Epidemiol2007;17:643e53.

Gutke A, Ostgaard HC, Oberg B. Pelvic girdle pain and lumbar pain in pregnancy: acohort study of the consequences in terms of health and functioning. Spine2006;31(5):E149e55.

Gutke A, Ostgaard HC, Oberg B. Predicting persistent pregnancy-related low backpain. Spine 2008;33(12):E386e93.

Gutke A, Hansson ER, Zetherstrom G, Ostgaard HC. Posterior pelvic pain provoca-tion test is negative in patients with lumbar herniated discs. Eur Spine J 2009Jul;18(7):1008e12.

Loge JH, Kaasa S. Short form 36 (SF-36) health survey: normative data from thegeneral Norwegian population. Scand J Soc Med 1998;26(4):250e8.

Loge JH, Kaasa S, Hjermstad MJ, Kvien TK. Translation and performance of theNorwegian SF-36 Health Survey in patients with rheumatoid arthritis. I. Dataquality, scaling assumptions, reliability, and construct validity. J Clin Epidemiol1998;51(11):1069e76.

Mens JM, Vleeming A, Snijders CJ, Koes BW, Stam HJ. Reliability and validity of theactive straight leg raise test in posterior pelvic pain since pregnancy. Spine2001;26(10):1167e71.

Mens JM, Vleeming A, Snijders CJ, Koes BW, Stam HJ. Validity of the active straightleg raise test for measuring disease severity in patients with posterior pelvicpain after pregnancy. Spine 2002;27(2):196e200.

Mens JM, Pool-Goudzwaard A, Beekmans RE, Tijhuis MT. Relation between sub-jective and objective scores on the active straight leg raising test. Spine (Phila Pa1976) 2010;35(3):336e9.

Mens JM, Huis In ’t Veld YH, Pool-Goudzwaard A. The Active Straight Leg Raise testin lumbopelvic pain during pregnancy. Man Ther 2012a;17(4):364e8.

Mens JM, Huis In ’t Veld YH, Pool-Goudzwaard A. Severity of signs and symptoms inlumbopelvic pain during pregnancy. Man Ther 2012b;17:175e9.

Mogren IM. BMI, pain and hyper-mobility are determinants of long-term outcomefor women with low back pain and pelvic pain during pregnancy. Eur Spine J2006;15(7):1093e102.

Noren L, Ostgaard S, Johansson G, Ostgaard HC. Lumbar back and posteriorpelvic pain during pregnancy: a 3-year follow-up. Eur Spine J 2002;11(3):267e71.

Olsson C, Nilsson-Wikmar L. Health-related quality of life and physical abilityamong pregnant women with and without back pain in late pregnancy. ActaObstet Gynecol Scand 2004;83(4):351e7.

Ostgaard HC, Zetherstrom G, Roos-Hansson E. The posterior pelvic pain provocationtest in pregnant women. Eur Spine J 1994;3(5):258e60.

Robinson HS, Eskild A, Heiberg E, Eberhard-Gran M. Pelvic girdle pain in pregnancy:the impact on function. Acta Obstet Gynecol Scand 2006;85(2):160e4.

Robinson HS, Brox JI, Robinson R, Bjelland E, Solem S, Telje T. The reliability ofselected motion- and pain provocation tests for the sacroiliac joint. Man Ther2007;12(1):72e9.

Robinson HS, Mengshoel AM, Bjelland EK, Vollestad NK. Pelvic girdle pain, clinicaltests and disability in late pregnancy. Man Ther 2010a;15(3):280e5.

Robinson HS, Mengshoel AM, Veierod MB, Vollestad N. Pelvic girdle pain: potentialrisk factors in pregnancy in relation to disability and pain intensity threemonths postpartum. Man Ther 2010b;15(6):522e8.

Robinson HS, Veierod MB, Mengshoel AM, Vollestad NK. Pelvic girdle pain - asso-ciations between risk factors in early pregnancy and disability or pain intensityin late pregnancy: a prospective cohort study. BMC Musculoskelet Disord2010c;11(1):91.

Robinson HS. Pelvic girdle pain and disability during and after pregnancy. TheUniversity of Oslo; 2010.

Rost CC, Jacqueline J, Kaiser A, Verhagen AP, Koes BW. Prognosis of women withpelvic pain during pregnancy: a long-term follow-up study. Acta ObstetGynecol Scand 2006;85(7):771e7.

Salen BA, Spangfortke L, Nordemar R. The disability rating index: an instrument forthe assessment of disability in clinical settings. J Clin Epidemiol 1994;47(12):1423e35.

Schwartz C, Sprangers M, Fayers P. Response shift: you know it’s there but how doyou capture it? Challenges for the next phase of research. In: Fayers P, Hays R,editors. Assessing quality of life in clinical trials, vol. 2. Oxford: Oxford Uni-versity Press; 2005. pp. 275e90.

Stacey T, Thompson JM, Mitchell EA, Ekeroma AJ, Zuccollo JM, McCowan LM. TheAuckland Stillbirth study, a case-control study exploring modifiable risk factorsfor third trimester stillbirth: methods and rationale. Aust N Z J Obstet Gynaecol2011;51:3e8.

Svensson HO, Andersson GB, Hagstad A, Jansson PO. The relationship of low-backpain to pregnancy and gynecologic factors. Spine 1990;15(5):371e5.

Tough SC, Siever JE, Benzies K, Leew S, Johnstom DW. Maternal well-being and itsassociation to risk of developmental problems in children at school entry. BMCPediatr 2010;10:19.

Vleeming A, Albert HB, Ostgaard HC, Sturesson B, Stuge B. European guidelines forthe diagnosis and treatment of pelvic girdle pain. Eur Spine J 2008;17(6):794e819.

Waddell G. The back pain revolution. Edinburgh: Churchill Livingstone; 2004.Ware JE. Using generic measures of functional health and well-being to increase

understanding of disease burden. Spine 2000;25(12):1467.