If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
School of Nursing, University of Michigan, Ann Arbor, MIDepartment of Obstetrics and Gynecology, Medical School, University of Michigan, Ann Arbor, MIWomen’s Studies Department, University of Michigan, Ann Arbor, MI
We sought to describe occurrence, recovery, and consequences of musculoskeletal (MSK) injuries in women at risk for childbirth-related pelvic floor injury at first vaginal birth.
Evaluating Maternal Recovery from Labor and Delivery is a longitudinal cohort design study of women recruited early postbirth and followed over time. We report here on 68 women who had birth-related risk factors for levator ani (LA) muscle injury, including long second stage, anal tears, and/or older maternal age, and who were evaluated by MSK magnetic resonance imaging at both 7 weeks and 8 months’ postpartum. We categorized magnitude of injury by extent of bone marrow edema, pubic bone fracture, LA muscle edema, and LA muscle tear. We also measured the force of LA muscle contraction, urethral pressure, pelvic organ prolapse, and incontinence.
In this higher-risk sample, 66% (39/59) had pubic bone marrow edema, 29% (17/59) had subcortical fracture, 90% (53/59) had LA muscle edema, and 41% (28/68) had low-grade or greater LA tear 7 weeks’ postpartum. The magnitude of LA muscle tear did not substantially change by 8 months’ postpartum (P = .86), but LA muscle edema and bone injuries showed total or near total resolution (P < .05). The magnitude of unresolved MSK injuries correlated with magnitude of reduced LA muscle force and posterior vaginal wall descent (P < .05) but not with urethral pressure, volume of demonstrable stress incontinence, or self-report of incontinence severity (P > .05).
Pubic bone edema and subcortical fracture and LA muscle injury are common when studied in women with certain risk factors. The bony abnormalities resolve, but levator tear does not, and is associated with levator weakness and posterior-vaginal wall descent.
Childbirth is arguably one of the most dramatic musculoskeletal (MSK) events the human body undergoes. Passage of the newborn through the pelvis and its muscles requires an exceptional degree of soft-tissue stretch.
Childbirth exerts remarkable stresses on maternal pelvic bones from the pressures of the fetal head and the forces of abdominal muscles used during maternal pushing that originate from the pelvic bones. Such stretch and stress may produce injury in some women.
In the last decade, new imaging techniques have brought important new insights into understanding the mechanisms of soft-tissue and bony injury.
Special sequences in MSK magnetic resonance imaging (MRI) offer advantages over other imaging techniques for studying deep bony and soft-tissue changes. Fluid-sensitive sequences have the best combined specificity and sensitivity for revealing areas of injury and edema. Hence, they are the recommended diagnostic imaging test for stress injuries.
Soon after beginning a study of levator ani (LA) muscle injury following vaginal birth, it became evident we should add these standard MSK-MRI fluid-sensitive sequences to our existing protocol of anatomical MRI sequences to better characterize the full scope of possible injuries and pattern of recovery.
The purpose of this study is to report on the occurrence and severity of bony and LA muscle injuries observed and how magnitude of tissue trauma relates to clinical consequences in the first 8 months’ postpartum. Fluid-sensitive sequences are necessary for 3 of 4 indicators of bone and muscle injury evaluated in our study. The sequences show: (1) increased signal that indicates edema (extracellular fluid) in bone; (2) matched linear signal changes in bone that indicate a fracture; or (3) increased signal that indicates edema in muscle. The fourth indicator of injury, visual discontinuity of muscle seen with muscle tear, does not require fluid-sensitive sequences. However, use of fluid-sensitive sequences makes detection of tears and their magnitude much easier. The precision of these measures offered opportunity to assess more precisely the relationship between injury magnitude and relative consequences seen clinically in the first 8 months’ postpartum.
Materials and Methods
The parent study Evaluating Maternal Recovery from Labor and Delivery (EMRLD) is an institutional review board–approved (University of Michigan Institutional Review Board HUM00051193) longitudinal cohort study following up primiparous women with recent history of childbirth. In this article, we report on those with higher-risk factors for LA injury. The first published reports from this work included: (1) details on EMRLD’s sampling strategies and a Strengthening the Reporting of Observational Studies in Epidemiology diagram; (2) specifics of using MSK-MRI methods; (3) ensuing anatomical detail of pelvic floor structures at rest, during dynamic activity, and by LA muscle subdivision and line of action; and (4) predominant demographic or obstetric variables associated with LA tear when evaluated early postpartum.
EMRLD data collection occurred from June 13, 2005, through March 14, 2012, collecting data at approximately 7 weeks after a first vaginal birth and again at about 8 months after first vaginal birth. In this article, we report the 7 weeks to 8 months’ postpartum longitudinal findings.
The enriched sampling relied on inclusion criteria of heuristically determined risk factors for LA tear (eg, prolonged second stage, anal sphincter tear, higher maternal age, forceps delivery) suggestive in 2005, the time of the study’s start.
Women were excluded from EMRLD if aged <18 years, spoke a primary health care language other than English, delivered at <36 weeks’ gestation, birthed >1 infant, or if the infant was admitted to neonatal intensive care.
Of the 90 women originally recruited into EMRLD, 22 women did not have a second MRI at 8 months’ postpartum. Our analysis was based on the 68 women with MRI data at both 7 weeks and 8 months’ postpartum.
MRIs were completed on a 3-T Philips Achieva (Philips Medical System, Eindhoven, The Netherlands) with an 8-channel cardiac coil positioned over the pelvis. The lower pelvis was imaged in the coronal, axial, and sagittal planes with proton density-weighted (PD) sequences; repeat time (TR) = 2107 milliseconds; echo time (TE) = 30 milliseconds; number sequence averages = 2; slice thickness = 4 mm, gap = 1 mm; and field of view (FOV) = 20 cm, matrix = 256 × 256. For better definition of the anterior pelvic floor anatomy, additional tailored imaging (slice thickness = 2 mm, gap = 0.2 mm; FOV = 18 cm, matrix = 256 × 256) included 3 planes of PD sequences and axial and coronal planes, either PD fat saturation (TR = 2355 milliseconds; TE = 30 milliseconds) or short tau inversion recovery (STIR) (TR = 5987 milliseconds; TE = 60 milliseconds; number sequence averages = 2) sequences. A single sacral PD fat saturation sequence was obtained in the axial plane (slice thickness = 4 mm, gap = 1 mm; FOV = 20 cm, matrix = 256 × 256).
The complete MRI protocol has been discussed in previously published reports.
MRIs were reviewed by 2 board-certified, fellowship-trained MSK radiologists who were blinded to details of an individual woman’s birth data and risk category. They were aware if the woman was having her initial 7 weeks or 8 months’ postpartum visit since the postpartum uterus was obvious in the pelvic MRI study.
We used standard MSK-MRI radiology grading categories to evaluate edema and fractures in bone and muscle injuries so our data could be compared across time and studies in the radiology literature.
When there was a difference in scoring between the 2 radiologists, the scans were reviewed together and graded by consensus, consistent with standard procedures in radiology. The radiologists measured and scored 4 sites of likely MSK injury.
Pubic bone marrow evaluation for edema
Bone marrow edema was assessed by grading signal intensity (none, mild, moderate, or intense) within the bone marrow of each pubic bone (right, left) as compared with the ischial tuberosity and other bones in the FOV.
The spectrum of bone stress injuries has been correlated with clinical findings in athletes and military recruits and is graded 0-4, based on the degree of bone marrow edema according to the MSK-MRI scale with STIR or fat-suppressed sequences.
Grade 0 is no abnormal signal. Grades 1-3 are mild, moderate, and intense bone marrow edema. Grade 4 is a true stress fracture with a line of increased signal (STIR or fat-suppressed sequences) with matching linear decreased signal on T1.
Pubic bone evaluation for fracture
Evidence of pubic bone fractures in the cortical or trabecular bone was assessed as matching lines of increased and decreased signal on the fluid-sensitive and standard T1 sequences visible in 2 imaging planes.
Fractures were recorded as none, subcortical, or cortical fractures for both pubic bones.
LA evaluation for edema
The LA was evaluated for the presence and location of increased signal, indicating edema consistent with stress or injury, as compared with other pelvic muscles, including the obturator internus. Grading classification categories were none, mild, moderate, or intense for each side.
LA evaluation for tear
The LA was evaluated for discontinuity of muscle observed as loss of visible muscle in an area where it is known to occur, indicating muscle tear. Grading classifications were: 0% to <20% (none to subtle), 20% to <50% (low grade), or ≥50% (high grade) for each side. Based on previous pelvic floor imaging experience, it was assumed that normal muscles should be symmetric and that each should have the same morphological configuration as muscles seen in nulliparous controls reported in other studies.
Strength of the LA was evaluated at both time points by measuring the vagina closure force at rest and during maximal pelvic muscle contraction (average of 3 attempts) using an instrumented speculum modified to not be influenced by changes in abdominal pressure.
A urethral pressure profile was obtained at 8 months’ postpartum only, due to its invasive nature. These measures were made by a nurse practitioner with >5 years of experience in clinical examinations. She was blinded to the MRI findings.
To standardize rehabilitation during the study period, at the examination, the same nurse practitioner instructed each woman in Knack technique and individually prescribed home pelvic muscle exercises per the graduated strength-training protocol.
Of the 68 women with MRIs at 7 weeks and 8 months’ postpartum, 59 had fluid-sensitive sequences, and the remaining 9 women had only nonfluid-sensitive sequences. Discovery of the importance of these sequences occurred after study initiation. The missing sequences were due to early enrollment before the MSK-MRI protocol was in use. Since LA tear is readily observable without fluid-sensitive sequences, the full 68 women were retained for that analysis, but the 9 were not included in analysis of pubic bone edema, fracture, and LA edema.
A composite score for the degree of injury for each individual was derived by collapsing “left,” “right” sides to yield ordinal-level data: for the LA tears, a composite score of “0” indicated no or subtle tear on both sides, “1” indicated a low-grade unilateral tear, “2” indicated a bilateral low-grade or unilateral high-grade tear, and a score of “3” indicated a bilateral high-grade tear. Similar composite scores were constructed for pubic bone marrow edema, pubic bone fracture, and LA edema (Table 1, Table 2, Table 3, Table 4).
Table 1Maternal demographics and birth variables among women with 7 wks and 8 mos’ postpartum magnetic resonance images
Mean (SD) or frequency
Range or %
Maternal age, y
Maternal age >31 y
High school graduate or less
College/technical school graduate
Infant weight, g
Infant head circumference, cm
Second stage, min
Second stage >150 min
Active pushing, min
Passive second stage, min
Miller. Childbirth muscle/bone injury and recovery. Am J Obstet Gynecol 2015.
Descriptive statistics were conducted only for those who completed both MRIs, comparing magnitude levels on 7 weeks and 8 months’ postpartum in the 4 MSK sites. We excluded subjects from analysis when there was missing data due to lack of follow-up or, on select analysis, lack of fluid-sensitive sequences required for a particular variable.
The Stuart-Maxwell test was used to test for homogeneity between the marginal distributions for each of the 4 sites at the 2 evaluation times. To calculate the statistic, an inconsequential decimal (0.5) was added to each empty cell in the table.
We used repeated measures modeling techniques to investigate the association among the 4 types of MSK injury and the clinical measures. Analyses were performed using SAS, version 9.3 (SAS Institute, Cary, NC) and R, version 2.14.1 (R Foundation for Statistical Computing, Vienna, Austria).
Table 1, Table 2 show the obstetric, demographic, and 8 months’ postpartum clinical characteristics of the 68 women analyzed with 7 weeks and 8 months’ postpartum MRI scans (including 9 who had scans without the fluid-sensitive sequences). When compared with the 22 women without a second MRI, the group with scans at both evaluation times (n = 68) was older, more predominately white, and more highly educated. The 2 groups did not differ in terms of obstetric characteristics or distribution of MSK injury severity at 7 weeks’ postpartum. The early MRI scans were obtained on average 47.3 ± 21.0 days' postpartum; the late MRI scans were obtained 242.1 ± 60.8 days postpartum.
Figure 1 shows one woman’s MRI findings and types of injuries observed and reported by category.
Pubic bone marrow edema
Of the 59 women with complete MSK-MRI studies, 39 of 59 (66%) showed increased signal, indicating edema in the pubic bone marrow at 7 weeks’ postpartum. At 8 months’ postpartum, only 8 of 59 (14%) still showed increased signal (Table 3). Of the 8 women with continued marrow edema, 5 did not show any improvement, while the other 3 who all initially had the highest category of signal intensity, improved by 1 category. For 7 weeks to 8 months’ postpartum, the distribution of women’s signal intensity scores across the 4 magnitude category levels (intense, moderate, mild, no signal intensity) was statistically different (P < .0001), with a strong shift toward lower severity categories at 8 months’ postpartum, indicating resolving marrow edema (Figure 2, A).
Pubic bone fracture
Fractures in the bone marrow of the pubic bone were observed in 17 of 59 women (29%) on the 7 weeks’ postpartum scans. By 8 months’ postpartum, 2 still showed evidence of healing fractures (Table 3). Both were low grade. Distributions of women’s signal intensity scores across the 4 magnitude category levels of signal intensity compared 7 weeks to 8 months’ postpartum were significantly different (P = .01), indicating overall fracture healing (Figure 2, B).
LA edema was initially observed in 53 of 59 women (90%) and was still evident in 4 women (7%) on follow-up (Table 3). Of the 4 without complete resolution of LA edema, only 1 did not show any improvement (level II magnitude at both evaluation times), while the other 3 had incomplete resolution. Similar to pubic bone marrow edema, severity for LA edema, compared 7 weeks to 8 months’ postpartum, was significantly different (P < .0001), indicating progress toward full resolution by 8 months’ postpartum. However, LA edema was faster to resolve than bone marrow edema, as only 4 women continued to have some increased signal in the LA at 8 months’ postpartum (Figure 2, C). There was no abnormal signal in the other pelvic muscles (eg, obturator internus) for any of the women, either 7 weeks or 8 months’ postpartum.
LA discontinuity was seen at 7 weeks’ postpartum in 28 of 68 women (41%). The magnitude of discontinuity for these 28 women was distributed as follows: 8 having unilateral low-grade tears, 13 having bilateral low-grade or unilateral high-grade tears, and 7 having bilateral high-grade tears. When all 68 women were evaluated at 8 months’ postpartum, no new tears were identified, and of the 28 who showed tear at 7 weeks’ postpartum, all but 3 showed the same magnitude level at 8 months’ postpartum (Table 3). Of the 3 who showed a difference, 2 dropped 1 category from unilateral low-grade tear to none-to-subtle tear and 1 dropped from bilateral high-grade tear to the category labeled as bilateral low-grade or unilateral high-grade tear (Figure 2, D). Statistically, there was no difference in the magnitude of the tears 7 weeks to 8 months’ postpartum (P = .86), indicating that tears did not resolve.
Clinical measures at 8 months
None of the clinical measures (prolapse, urethral pressure, incontinence, LA strength) obtained were significantly associated with higher magnitude of MRI-documented pubic bone marrow edema, pubic bone fracture, or LA edema injury. LA contraction force and the degree of descent of the posterior vaginal wall were significantly associated with higher magnitude of LA tear (P = .008 and P = .005). None of the fecal or urinary incontinence measures were associated with magnitude of LA tear scores (Table 4).
This study of women with risk factors for LA injury from first childbirth demonstrates that 91% of this selected sample show some form of MSK injury involving the pubic bone or LA when studied with MSK-MRI 7 weeks’ postpartum. Most of the observed injuries resolve by 8 months’ postpartum. LA muscle tears identified 7 weeks after birth in 41% of this selective sample did not resolve. Overall, 9% were high-grade lesions involving >50% of the muscle. In all of those with an LA tear, 89% had the same magnitude of levator injury 8 months’ postpartum, and for those with a change, it was by only 1 severity level.
We speculate that torn muscle edges retract from their origin because of the significant resting tone,
Severity of LA muscle tear was associated with degree of reduced LA force developed during a maximal contraction on instrumented speculum testing and with degree of posterior wall descent. There was no association with anterior wall descent, which differs from case-control studies when women were classified by only 2 LA tear categories (major or none) and were older.
Figure 3 shows how various degrees of tear do or do not explain various clinical findings.
Edema within the LA and pubic bone marrow and pubic bone fracture tended to recover. These observations are similar to general MSK-MRI studies where resolution of muscle strains or stress fractures found in athletes (eg, marathoners) resolve over time.
The most likely explanation for persistence of LA muscle edema in 4 women (7%) at 8 months’ postpartum is that each had a high-grade LA muscle tear. In MSK-MRI studies, the magnitude of muscle and bone injury is proportional to the intensity of signal abnormalities and length of time to heal.
Urethral pressure and incontinence signs and symptoms, fecal or urinary, were not sensitive indicators of magnitude of injury in any of the MSK aspects studied in the first few months after birth at this sample size. This differs from data reported in 3 studies,
which showed that postpartum women with “major” LA tear, compared with those “without major” LA tear, experienced lower urethral pressures and more incontinence.
Study limitations include that 22 of the original 90 women evaluated did not return for follow-up visit. The retained women were, on average, older and more frequently white; however, injury severity at 7 weeks' postpartum did not differ between those retained and those who dropped out. However, the radiologists could not be completely blinded to study design and time points, which may have introduced bias toward maximally identifying injuries. It is logical to assume from other, better-understood MSK trauma (eg, hamstring tear and sacral fractures) that symptoms such as degree of pain (which we did not measure in EMRLD) correlate with injury severity, and time needed for healing will vary accordingly.
Establishing the most effective rehabilitation protocol and time when activities, eg, high-intensity exercise or sexual intercourse, should occur, will require further study.
This study’s findings suggest we cannot rely on clinical examination to intuit the full spectrum of MSK issues associated with higher-risk birth. In select women, MSK-MRI may be warranted clinically and in research, since MSK-MRI can provide the full spectrum of MSK injuries. Consideration should be given to including MSK-MRI for evaluation of high-risk symptomatic patients. The role of MSK-MRI in evaluation of postpartum women could be expanded in situations of unexplained or prolonged pain after delivery, nonrecoverable ability to contract the LA (Kegel), or nonresolving postpartum pelvic organ descent. The differential diagnosis will suggest patient care aspects, ie, reassurance of no structural lasting injury, expected timeline of resolution, or follow-up considerations (short- or long-term) for more serious findings.
The degree to which women with postpartum levator defects and other MSK injuries will eventually develop symptomatic urinary incontinence, fecal incontinence, or chronic pelvic organ prolapse is not known. The longitudinal EMRLD study is continuing with long-term follow-up and will help answer these questions over time.
The authors acknowledge the EMRLD study staff: Ruta Misiunas, Lee Park, Caroline Garcia, and Meg Tolbert, and all the women participating in EMRLD. Heather Van Doren, MFA, coordinating senior editor with Arbor Research Collaborative for Health, provided editorial assistance on this manuscript. We also gratefully acknowledge James Ashton-Miller as Core B Director of the University of Michigan Specialized Center of Research on Sex and Gender Factors Affecting Women’s Health.
Levator ani muscle stretch induced by simulated vaginal birth.
The Evaluating Maternal Recovery from Labor and Delivery study is supported by grant number P50 HD044406 002 from the Office for Research on Women’s Health Specialized Center of Research on Sex and Gender Factors Affecting Women’s Health , National Institutes of Health , and the Eunice Kennedy Shriver National Institute on Child Health and Human Development (NICHD), and by grant number R21 HD049818 from NICHD.
The authors report no conflict of interest.
The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or its institutes. The sponsor had no involvement in study design; collection, analysis, or interpretation of data; writing; or decision to submit the article for publication.
Cite this article as: Miller JM, Low LK, Zielinski R, et al. Evaluating maternal recovery from labor and delivery: bone and levator ani injuries. Am J Obstet Gynecol 2015;213:188.e1-11.