Objective To study the prognosis for pelvic girdle pain, and to explore the association between presence of emotional distress during pregnancy and pelvic girdle syndrome 6 months after delivery.
Design Longitudinal population study.
Setting Norway, for the period 1999–2008.
Population A follow-up of 41 421 women in the Norwegian Mother and Child Cohort who reported pelvic girdle pain at 30 weeks of gestation.
Methods Data were obtained by self-administered questionnaires in weeks 17 and 30 of gestation, and 6 months after delivery.
Main outcome measure Pelvic girdle syndrome 6 months after delivery, defined as pain in the anterior pelvis and on both sides in the posterior pelvis.
Results Six months after delivery, 78.0% of the women had recovered, 18.5% reported persistent pain in one or two pelvic locations, 3.0% reported pelvic girdle syndrome and 0.5% reported severe pelvic girdle syndrome. The recovery rates decreased with increasing levels of pain severity in pregnancy. Being emotionally distressed at two time points during pregnancy was associated with the presence of pelvic girdle syndrome (adjusted OR 1.5, 95% CI 1.2–1.9) and severe pelvic girdle syndrome (adjusted OR 1.9, 95% CI 1.1–3.1), after adjustment for pain severity in pregnancy, other medical conditions, body mass index, age at menarche, previous low back pain, and smoking during pregnancy.
Conclusions In this follow-up of women with pelvic girdle pain in pregnancy, the recovery rates after delivery were high. Our findings suggest that the presence of emotional distress during pregnancy is independently associated with the persistence of pelvic girdle pain after delivery.
Pelvic girdle pain affects one in five pregnant women, gives rise to functional disability, and is a major cause of sick leave during pregnancy.1,2 Hence, it should be regarded as a major women’s health issue. The reported recovery rates after delivery are high.3 However, in some women, pelvic girdle pain turns into a state of persistent pain and disability,4 and may have substantial individual and socio-economic consequences. The causes and pathogenesis of pregnancy-related pelvic girdle pain remain poorly understood. The diagnosis is based on the woman’s subjective experience, and not from objective findings, and is thus categorized as a subjective health complaint. Nevertheless, the onset of pelvic girdle pain often occurs in the first half of pregnancy,5 and typically resolves shortly after delivery,3 indicating that reproductive factors probably play a significant role.6,7
Although the causes of pelvic girdle pain may differ from the causes of other pain conditions, the factors affecting the recovery process may well have similarities. High pain intensity and the number of pain locations have been indentified as important factors in the transition of acute and subacute pain into chronic pain and disability.8,9 These factors have also been reported to influence the prognosis in pelvic girdle pain.3,10–12 Except for pain severity during pregnancy, knowledge about prognostic factors for pelvic girdle pain is scarce, and most studies have methodological limitations, such as small study samples and/or limitations in design.
Psychological factors are associated with a poor prognosis in other types of musculoskeletal pain.13,14 The direction of the association between psychological factors and chronic pain conditions, however, is debated, leading to calls for longitudinal studies.14,15 It is plausible that psychological factors play a role in the recovery process of pelvic girdle pain, and emotional distress has previously been identified as a risk factor for pelvic girdle pain in pregnancy.7 However, the association of psychological factors with the persistence of pelvic girdle pain after delivery has scarcely been studied, and evidence is inconclusive.10,16–18
In the present study, our aim was to assess the prognosis according to pain severity in 41 421 Norwegian women who reported pelvic girdle pain at 30 weeks of gestation. We also explored the association between the presence of emotional distress during pregnancy and pelvic girdle syndrome (PGS, defined as pain in the anterior pelvis and in the posterior pelvis bilaterally) 6 months after delivery.
Study design, study population, and follow-up
During the years 1999–2008, all pregnant women scheduled to give birth at 50 hospitals in Norway were targeted for recruitment into the Norwegian Mother and Child Cohort Study (http://www.fhi.no/morogbarn),19 conducted by the Norwegian Institute of Public Health. The women were recruited at the routine ultrasound examination at 17–18 weeks of gestation. This examination is part of the public antenatal care programme, and is offered to all pregnant women. The study had no exclusion criteria, and 38.5% of all eligible women who gave birth in Norway agreed to participate. The current study is based on the quality-assured data files released for research in 2010. The Norwegian Mother and Child Cohort Study is described in detail elsewhere.19
Data were obtained through three self-administered questionnaires, sent and returned by mail. The first questionnaire was completed in the second trimester (mean 17.2 weeks of gestation, standard deviation, SD, 2.2 weeks) and included questions about sociodemographic factors, general health, and reproductive history. In the third trimester (mean 30.5 weeks of gestation, SD 1.4 weeks) and at 6 months after delivery (mean 28.0 weeks, SD 3.0 weeks), the women completed the second and third questionnaires, respectively, which included questions about maternal health during pregnancy and after delivery.
The response rate was 91.6% at 30 weeks of gestation and 90.5% at 6 months after delivery (Figure 1). A total of 73 418 participants completed the three questionnaires used in our study. We used a wide definition of pelvic girdle pain, and included the 58% (42 289/73 418) who reported pain in at least one pelvic location at 30 weeks of gestation in our study sample. We excluded the 868 women who had not responded to the questions about emotional distress, leaving 41 421 women in our study sample. The excluded 868 women did not differ with respect to pelvic girdle syndrome or level of pain severity from the women in our study sample.
Pelvic girdle pain was classified on the basis of answers to the following questions at 30 weeks of gestation and at 6 months after delivery: do you have pain in the pelvic girdle; if you have pain in the pelvic girdle, where is the pain located? One or more locations could be specified: in the frontal part of the pelvis; on one side of the rear part of the pelvis; on both sides of the rear part of the pelvis. Pain intensity was scored as mild or severe at each location. Pelvic girdle syndrome (PGS) was defined as combined pain in the anterior pelvis and on both sides in the posterior pelvis.3 PGS was subdivided according to reported pain intensity: the presence of severe pain in all three locations was designated as severe PGS.7 Pelvic girdle pain at 30 weeks of gestation was classified into three mutually exclusive subgroups to grade the level of severity during pregnancy: pain in one or two pelvic locations; pain in three pelvic locations (PGS); and severe pain in all three pelvic locations (severe PGS). Information about the onset of pelvic girdle pain was obtained from the first questionnaire at 17 weeks of gestation. Functional disability at 30 weeks of gestation and at 6 months after delivery was addressed by questions about the use of crutches and waking up at night because of pelvic girdle pain.
Emotional distress in weeks 17 and 30 of gestation was measured by a short version of the Hopkins Symptom Checklist 25 (SCL-25), namely the Symptom Checklist 5 (SCL-5).20 SCL-25 is a widely used self-administered instrument, designated to measure symptoms of anxiety and depression in population surveys.21,22 It has been consistently shown in several populations that SCL-25 is acceptable as a screening instrument for depression, as defined by the 10th revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10).23,24 SCL-5 is highly correlated (r = 0.92) with SCL-25,25 and consists of the following questions. Have you been bothered by any of the following during the last 2 weeks: (1) feeling fearful; (2) nervousness or shakiness inside; (3) feeling hopeless about the future; (4) feeling blue; or (5) worrying too much about things. The response categories ranged from 1 (not bothered) through to 4 (very bothered). A sum score was calculated and divided by the number of items. In order to focus on emotionally distressed women the SCL-5 scores were dichotomized, and we defined a mean score ≥2.0 as the presence of emotional distress.20 In our data analyses, the presence of emotional distress in pregnancy was coded: no emotional distress (at either time point); emotional distress at one time point during pregnancy (at either 17 or 30 weeks of gestation); and emotional distress at two time points in pregnancy (at both 17 and 30 weeks of gestation).
Reliable knowledge about prognostic factors for pelvic girdle pain is scarce. Hence, when building the multivariate statistical models, we assessed factors that are plausible confounders of the association of emotional distress with persistence of pelvic girdle pain, as well as factors previously identified as risk factors for developing pelvic girdle pain in pregnancy.6,7
Maternal age upon inclusion at 17 weeks of gestation was coded as follows: <25, 25–34, and ≥35 years. The number of deliveries was coded as no previous delivery and one or more previous deliveries. Age at menarche was coded: ≤10, 11, 12, ≥13 years, and missing. Educational level was coded: ≤12, 13–16, ≥17 years, and missing. Smoking during pregnancy was coded as non-smoker, occasional smoker, daily smoker, and missing. Body mass index (BMI) at 17 weeks of gestation was calculated as weight/height2 (kilograms per square metre), and coded: <25, 25–29, ≥30, and missing. A history of low back pain was defined as the presence of any low back pain before being pregnant for the first time (yes/no). A medical co-morbidity index (ranging from a minimum of 0 to a maximum of 15) was constructed on the basis of questions assessing the presence of diseases including high blood pressure, asthma/allergy, migraine, fibromyalgia, dermatological disease, anaemia, diabetes, cancer, hyper/hypothyroidism, and organ-specific diseases (cardiovascular disease, disease in the gastrointestinal tract, disease in the kidney and urinary tract, gynaecological disease, rheumatic disease, and neurological disease). With the intention to grade the level of morbidity, the co-morbidity index was coded: no disease; 1 disease; 2–3 diseases; ≥4 diseases.
The occurrence of pelvic girdle pain 6 months after delivery according to subgroups is presented as proportions (%). The associations between presence of emotional distress during pregnancy and PGS and severe PGS after delivery were estimated as crude and adjusted odds ratios (ORs), with 95% confidence intervals (95% CIs), using logistic regression analyses. We studied the association in all 41 421 women, and we also made separate analyses for the 32 680 women who experienced the onset of pelvic girdle pain after 17 weeks of gestation. We used univariate and multivariate binary logistic regression analyses to assess the relationship of potential confounding factors with PGS. Only study factors that caused 10% change in the coefficient of emotional distress,26 and/or were significantly associated with the outcomes, were included in the final adjusted models. Potential interactions were tested by including interaction terms in the multivariate binary logistic regression analyses. A 5% significance level was chosen for the analyses. The statistical software package pasw 17.0 was used for the statistical analyses (SPSS Inc., Chicago, IL, USA).
Upon inclusion at 17 weeks of gestation, the mean maternal age was 29.7 years (SD 4.5 years, range 14–46 years). Forty-one percent of the women were first-time mothers, and their mean BMI was 25.4 (SD 4.3). One in five women (21.1%) reported the onset of pelvic girdle pain at 17 weeks of gestation. Emotional distress (SCL-5 ≥ 2.00) at both 17 and 30 weeks of gestation was present in 3.4% of the women, whereas emotional distress at one time point only was present in 8.5%. In women without pelvic girdle pain at 30 weeks of gestation, the corresponding presence of emotional distress was significantly lower, 1.6 and 5.4% (P < 0.001).
Six months after delivery, 22.0% (9099/41 241) of the women reported pain in the pelvic girdle. PGS was present in 3.0% (1252/41 421) and severe PGS was present in 0.5% (196/41 421) of the women (Figure 2). The recovery rate was 82.5% in women reporting pain in one or two pelvic locations at 30 weeks of gestation; it decreased to 68.5% in women reporting PGS and to 45.8% in women reporting severe PGS (Table 1). Nevertheless, 95.0% of women with severe PGS reported an improvement after delivery. The absolute risk of using crutches 6 months after delivery was 5.8% (73/1252) and 29.1% (57/196), respectively, in women with PGS and severe PGS. Likewise, 16.1% (202/1252) of women with PGS and 45.9% (90/196) of women with severe PGS frequently woke up at night 6 months after delivery because of pelvic girdle pain.
Table 1. Presence of subgroups of pelvic girdle pain 6 months postpartum, according to subgroups at 30 weeks of gestation (n = 41 421)
Pelvic girdle pain at 30 weeks of gestation
Pelvic girdle pain at 6 months postpartum
No pain n (%)
Pain in 1 or 2 locations n (%)
Pain in 3 locations (PGS) n (%)
Severe pain in 3 locations (severe PGS) n (%)
Pain in 1 or 2 locations (n = 30 903)
25 505 (82.5)
Pain in 3 locations (n = 8823)
Severe pain in 3 locations (n = 1695)
The presence of emotional distress at one time point during pregnancy was associated with PGS after delivery (crude OR 1.7, 95% CI 1.5–2.0). Likewise, emotional distress at two time points during pregnancy was associated with PGS after delivery (crude OR 2.5, 95% CI 2.0–3.1). The positive association remained after adjustment for potential confounding factors, although the magnitude of the associations was attenuated (adjusted OR 1.3, 95% CI 1.1–1.5, and adjusted OR 1.5, 95% CI 1.2–1.9, respectively; Table 2). Emotional distress at one or both time points during pregnancy was also significantly associated with severe PGS after delivery (crude OR 2.6, 95% CI 1.8–3.9, and crude OR 3.5, 95% CI 2.1–5.9, respectively). After adjustment, the corresponding odds ratios were attenuated, but the positive associations remained (adjusted OR 2.0, 95% CI 1.4–2.9, and adjusted OR 1.9, 95% CI 1.1–3.1, respectively; Table 2). Pain severity at 30 weeks of gestation and the co-morbidity index were the main confounding factors of the associations, but no interactions between these factors and emotional distress regarding the risk of PGS or severe PGS were identified. The associations between emotional distress during pregnancy and PGS were also present in analyses restricted to women with onset of pelvic girdle pain after 17 weeks of gestation (Table 3).
Table 2. Prognostic factors for pelvic girdle syndrome and severe pelvic girdle syndrome at 6 months postpartum (n = 41 421)
Pelvic girdle syndrome
Severe pelvic girdle syndrome
Adjusted OR (95% CI)
Adjusted OR (95% CI)
The associations are estimated as adjusted odds ratios with 95% confidence intervals. All study factors are mutually adjusted in the logistic regression model.
*P < 0.05; **P < 0.01; ***P < 0.001.
Yes, at one time point
Yes, at two time points
Level of severity in pregnancy
Pain in 1 or 2 locations
Pain in 3 locations
Severe pain in 3 locations
BMI at inclusion (kg/m2)
Age at menarche (years)
Previous low back pain
Smoking during pregnancy
Table 3. The associations of emotional distress with pelvic girdle syndrome and severe pelvic girdle syndrome at 6 months postpartum in 32 680 women with onset of pelvic girdle pain after 17 weeks of gestation
Pelvic girdle syndrome
Severe pelvic girdle syndrome
Adjusted OR (95% CI)
Adjusted OR (95% CI)
The associations are estimated as adjusted odds ratios with 95% confidence intervals, adjusted for pain severity in pregnancy, co-morbidity index, body mass index, age at menarche, previous low back pain, and smoking during pregnancy.
*P < 0.05; **P < 0.01; ***P < 0.001.
Yes, at one time point
Yes, at two time points
The level of pain severity at 30 weeks of gestation was the strongest prognostic factor (Table 2). The adjusted ORs for reporting PGS after delivery were 4.2 (95% CI 3.7–4.8) and 16.3 (95% CI 14.0–18.9) in women who reported pain in three locations and severe pain in three locations, respectively, at 30 weeks of gestation, as compared with the reference group. Other factors associated with PGS were a high co-morbidity index, early menarche, BMI ≥ 30, previous low back pain, and occasional smoking during pregnancy. All of these factors, with the exception of smoking, were also associated with severe PGS (Table 2). Accordingly, a high level of pain severity at 30 weeks of gestation was the strongest predictor for severe PGS after delivery, but early menarche and a high co-morbidity index were also strongly associated with severe PGS. Maternal age, number of deliveries, and educational level were not important confounding factors of the associations of emotional distress with PGS, nor were they associated with the outcomes, and were not included in the final models.
In this follow-up study of 41 421 pregnant women reporting pelvic girdle pain in pregnancy, 78.0% had recovered 6 months after delivery. Nevertheless, 3.5% of the women still reported significant complaints. The recovery rates decreased with increasing levels of pain severity at pregnancy week 30. Women who experienced emotional distress during pregnancy were at an increased risk of reporting PGS after delivery, even after adjustment for pain severity.
In accordance with previous studies, the recovery rates of pelvic girdle pain were high, and the risk of non-recovery increased with pain severity in pregnancy.3,11,17 A previous study has reported a substantially poorer prognosis in women with PGS compared with other subgroups.3 A poorer prognosis has also been found in women with combined low back and pelvic girdle pain, than in women with either low back pain or pelvic girdle pain.11
An association between depressive symptoms and pelvic girdle pain after delivery has previously been reported in a cross-sectional study.16 For a better understanding of the nexus of cause and effect between psychological factors and chronic pain, longitudinal studies are needed.14,15 Only three previous longitudinal studies have assessed the association between emotional distress and persistence of pelvic girdle pain, finding none.10,17,18 The lack of a significant association in these studies may be explained by methodological limitations, such as low statistical power or insufficient control for confounding. To our knowledge, this longitudinal study is the first to confirm an association between the presence of emotional distress during pregnancy and persistent pelvic girdle pain in a large population that enabled control for several confounding factors.
Although we used a longitudinal design and measured emotional distress at two time points during pregnancy, we cannot rule out bidirectional causality. It is likely that the presence of pain during pregnancy induced increased levels of emotional distress. Nevertheless, the association between emotional distress during pregnancy and PGS after delivery was also present in women who did not report pain at the first measurement point of emotional distress at 17 weeks of gestation. In the analyses, we adjusted for several potential confounding factors, including the level of pain severity in pregnancy and co-morbidity, but the effect of emotional distress on PGS remained, suggesting a true relationship. It is also possible that PGS at 6 months after delivery has been over-reported by women with high levels of emotional distress. Therefore, it may be argued that the analyses should be adjusted for emotional distress 6 months after delivery. We performed additional analyses, adjusting for emotional distress after delivery, but the results remained the same (data not shown). However, adjusting for emotional distress at the end point may lead to over-adjustment because persistence of pain may have affected the level of emotional distress after delivery.
The location and intensity of pelvic girdle pain were based on self-reports and not on clinical examination. Self-reports of pelvic girdle pain have been reported to correlate well with the number of positive clinical tests.27 Even though attempts have been made,1 there is no international gold standard for clinical signs and symptoms that should result in a diagnosis of pelvic girdle pain. To ensure that the outcome measure had high specificity for pelvic girdle pain, we used pain in the anterior pelvis and on both sides in the posterior pelvis as our outcome. This subgroup has been associated with greater functional disability, more frequent sick leave, and significantly worse prognosis than pain in one or two pelvic locations.2,3,28 Furthermore, it has previously been shown that the number of pain sites has a strong predictor utility.9 This strengthens our classification of severity level during pregnancy, which was based on pain intensity, pain location, and the number of pain sites.
In the Norwegian Mother and Child Cohort Study, the women’s answers to SCL-5 were the only available measure of mental health. Generally, questionnaire studies enable the evaluation of larger study populations than studies using clinical interviews for data collection. A clinical interview, however, remains the gold standard for assessing emotional distress. It is important to note that SCL-5 cannot be used for diagnosing depression, anxiety, or other psychiatric disorders. The advantage of using SCL is that it is designed to measure symptoms of anxiety and depression in population surveys.20–22 It has been validated in several populations and is documented as an acceptable screening instrument for depression, as defined by the ICD-10.23,24 Furthermore, the presence of emotional distress at two time points during pregnancy is likely to have a higher specificity for emotional distress than the presence of emotional distress at one time point only.
The participation rate in the Norwegian Mother and Child Cohort Study was rather low, which is a common problem in large epidemiological studies.29 A low participation rate may result in skewed selection of participants and affect the prevalence estimates of the disease under study. However, the exposure–disease associations are more robust to selection bias than the prevalence estimates.29,30 A recent study on the potential biases of skewed selection in the Norwegian Mother and Child Cohort found the prevalence estimates, but not the exposure–disease associations, to be influenced by the selection.30 Compared with the general pregnant population in Norway, first-time mothers and women with a high socio-economic status were over-represented, whereas young mothers were under-represented. In a previous study, we found that the number of previous deliveries, a low educational level, and young maternal age were independent risk factors for pelvic girdle pain in pregnancy.7 Hence, the women included in our study sample may not be representative for the general population because the prevalence of pelvic girdle pain may have been underestimated. The women lost to follow-up after delivery had higher levels of emotional distress and reported more pelvic girdle pain at 30 weeks of gestation than the women making up our study sample. Consequently, our estimated recovery rates may have been overestimated, whereas the estimated association of emotional distress with PGS may have been underestimated as a result of selective attrition, although the direction of the association is not likely to be biased.30 Random errors associated with imprecise reporting may also have deflated the estimated association. On the other hand, this is probably counteracted because women with high levels of emotional distress during pregnancy may be more prone to report pain symptoms.
In the present study, women experiencing emotional distress during pregnancy were at an increased risk of reporting PGS 6 months after delivery. This finding may be explained by cognitive, behavioural, and neurophysiological factors.31 Emotionally distressed women may have an increased awareness of pain symptoms,32 and their interpretation of pain may differ from that of women without emotional distress. Patients with depression are more likely to catastrophize about pain.33 Thus, pain in emotionally distressed women may promote adaptation of dysfunctional movement patterns, physical inactivity, and hinder participation in social activities. Such behaviour may further increase the risk of chronic pain. Depression has also been suggested to promote central sensitisation and to influence endogenous pain inhibition,31 mechanisms likely to interfere with the recovery process of pelvic girdle pain. Furthermore, dysfunctions of the hypothalamic–pituitary–adrenal axis may represent a link between emotional distress and the persistence of pain.34
Although emotional distress during pregnancy affected the recovery process of PGS in our study, the level of pain severity in pregnancy was the most important prognostic factor. High pain intensity and a high number of pain sites during pregnancy and the postpartum period may cause central sensitisation,35 and increase the risk of persistent pain.36 In contrast, an increased risk of persistence of pain after delivery may result from individual differences in pain sensitivity. Most pain conditions are associated with an increased sensitivity to pain, with typically little correspondence between pathological findings and reported pain.37,38 Hence, normal changes in the pelvic girdle may be perceived as painful only in women with increased pain sensitivity. An influence of central pain mechanisms in the persistence of pelvic girdle pain is supported by the associations with a high co-morbidity index and previous low back pain. The increased risk of non-recovery in women with high co-morbidity is supported by a previous study.17 We also found associations of high BMI and early menarche with the presence of PGS after delivery. Obese women and women with early menarche have higher morbidity and mortality,39,40 and the increased risk of disease may be linked to hormonal and metabolic factors. Such factors are also likely to be associated with pelvic girdle pain. Therefore, women with persistent pelvic girdle pain may represent a group more susceptible to disease than the general female population.
The recovery rates in women with pelvic girdle pain during pregnancy were high at 6 months after delivery. Nevertheless, more than 3% still reported significant complaints. Presence of emotional distress during pregnancy was associated with non-recovery, even after adjustment for previous pain severity. An implication from our results is that if we could clinically impact emotional distress during pregnancy, we may be able to prevent the persistence of pelvic girdle pain after delivery. Hence, practitioners should identify and then pay careful attention to women with severe pelvic girdle pain during pregnancy, particularly if they are emotionally distressed. However, further research is required to confirm our results.
Disclosure of interests
Contribution to authorship
All authors planned the study and performed the analysis. E.K.B. and M.E.G. wrote the majority of the article. E.K.B., B.S., B.E., and M.E.G. discussed the design, edited the paper, and agreed on the final version. All authors had full access to all of the data in the study, and can take responsibility for the integrity of the data and the accuracy of the data analysis.
Details of ethical approval
The Norwegian Mother and Child Cohort Study was approved by all Regional Committees for Medical Research Ethics in Norway and by the Norwegian Data Inspectorate. All participants signed an informed consent form.
The Norwegian Mother and Child Cohort Study is supported by the Norwegian Ministry of Health and The Ministry of Education and Research, NIH/NIEHS (contract no. N01-ES-75558), NIH/NINDS (grant no. 1 UO1 NS 047537-01 and grant no. 2UO1 NS047537-06A1), and the Norwegian Research Council/FUGE (grant no. 151918/S10). The present study was supported by the Norwegian Research Council.
The authors thank Eli Heiberg and Anne Eskild for assistance in initiating this project, and for valuable advice and support. We are grateful to all of the participating families in Norway who take part in this continuing cohort study.