SEARCH

SEARCH BY CITATION

Keywords:

  • Abnormal uterine bleeding;
  • caesarean section;
  • caesarean section scar;
  • niche;
  • postmenstrual spotting;
  • ultrasonography

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Objective

To study the prevalence of niches in the caesarean scar in a random population, and the relationship with postmenstrual spotting and urinary incontinence.

Design

A prospective cohort study.

Setting

A teaching hospital in the Netherlands.

Population

Non-pregnant women delivered by caesarean section.

Methods

Transvaginal ultrasound (TVU) and gel instillation sonohysterography (GIS) were performed 6–12 weeks after caesarean section. Women were followed by questionnaire and menstruation score chart at 6–12 weeks, 6 months, and 12 months after caesarean section.

Main outcome measures

Prevalence of a niche 6–12 weeks after caesarean section, using TVU and GIS. Secondary outcomes: relation to postmenstrual spotting and urinary incontinence 6 and 12 months after caesarean section; and niche characteristics, evaluated by TVU and GIS.

Results

Two hundred and sixty-three women were included. Niche prevalence was 49.6% on evaluation with TVU and 64.5% with GIS. Women with a niche measured by GIS reported more postmenstrual spotting than women without a niche (OR 5.48, 95% CI 1.14–26.48). Women with residual myometrium at the site of the uterine scar measuring <50% of the adjacent myometrial thickness had postmenstrual spotting more often than women with a residual myometrial thickness of >50% of the adjacent myometrial thickness (OR 6.13, 95% CI 1.74–21.63). Urinary incontinence was not related to the presence of a niche.

Conclusions

A niche is present in 64.5% of women 6–12 weeks after caesarean section, when examined by GIS. Postmenstrual spotting is more prevalent in women with a niche and in women with a residual myometrial thickness of <50% of the adjacent myometrium.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Concern regarding the association between delivery by caesarean section and long-term maternal morbidity is growing, as the rate of caesarean section continues to increase. In the Netherlands the caesarean section rate increased from 7.4 to 15.8% during the period 1990–2008,[1] and in the USA it increased from 21.2 to 32.8% from 1990 to 2011.[2] The long-term complications of caesarean section in relation to future reproduction have been comprehensively examined.[3, 4] In the past decade several articles have described a defect that can be seen on ultrasound at the site of the caesarean section scar, known as a ‘niche’.[5, 6] A niche is defined as a triangular anechoic structure at the site of the scar or a gap in the myometrium of the anterior lower uterine segment at the site of a previous caesarean section.[7, 8] Niche prevalence depends on the method used for evaluation and the population investigated.[9-16] In non-pregnant women the scar is visible with transvaginal ultrasonography (TVU) and contrast sonohysterography using either saline (saline infusion sonohysterography, SIS) or gel (gel instillation sonohysterography, GIS).[5, 6] There have been several reports of an association between abnormal bleeding and a niche.[8, 10-12, 15] In particular, postmenstrual spotting seems to be a predominant symptom in women with a niche[10, 11]; however, most of these studies included symptomatic patients, resulting in a possible selection bias. Given the proximity of the niche to the bladder, it has been postulated that the local accumulation of fluid and scarring might disturb bladder function[11]; however, prospective studies assessing both bleeding symptoms and urinary symptoms in relation to the presence of a niche in a random population of women with a history of caesarean section are scarce. As far as we are aware, none of the reported studies asked women to participate at the time of caesarean section. The aim of our study was to determine the prevalence of niches in the caesarean section scar in a random population of women enrolled immediately after caesarean section, using both TVU and GIS 6–12 weeks after caesarean section, and to examine the relationship between the niche and postmenstrual spotting and urinary incontinence.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Study design and population

We conducted a prospective observational cohort study. The trial was registered in the Dutch trial register (trial number NTR-2887). The study was performed in St Antonius Hospital, Nieuwegein: a teaching hospital in the Netherlands. The protocol was approved by the local medical ethics committee (VCMO NL18722.100.07 R-07.14A/SCAR.). Participants were recruited between November 2007 and September 2010. Specially trained midwives asked all women over 18 years of age who underwent a caesarean section at St Antonius Hospital during the inclusion period to participate, within 3 days of the caesarean section. Exclusion criteria were twin pregnancies, known uterine anomalies, or the suspicion of a uterine infection, defined as a positive culture or fever and abdominal pain with discharge. After providing written informed consent, all participants were evaluated by TVU followed by GIS 6–12 weeks after the caesarean section. A GIS was not performed when there were suspected placental remnants, intrauterine haematomas, intrauterine fluid accumulation, or a possible pregnancy (i.e. a visit during the midluteal phase in women not using adequate contraception). Women were followed by questionnaires and menstruation score charts at 6–12 weeks, 6 months, and 12 months after the caesarean section.

Transvaginal ultrasound

All TVU and GIS procedures were performed at St Antonius Hospital by three experienced examiners (L.V., S.V., and M.S.). TVU was performed using a 7.5–MHz transducer (Philips Sonicare HD 11.XE, Philips Medical Systems, Eindhoven, the Netherlands). Women were examined after emptying their bladder. The uterus, uterine scar, and niche, if present, were examined in a standardised way. The position, length, and width of the uterus and double thickness of the endometrium were registered in the midsagittal plane. A niche was defined as an anechoic space (with or without fluid) at least 2 mm deep at the presumed site of the caesarean section scar. The uterus was screened for the presence of caesarean section scar(s) using parallel sagittal planes going from left to right, until the plane with the largest niche depth was defined. The uterus was also scanned in the transverse plane. When a niche was identified, it was measured in the sagittal plane in which the niche had the greatest depth; the depths of the niche and of the residual myometrium at the site of the scar and of the adjacent normal myometrium were measured in real time (see Figure 1). In cases when more than one caesarean section scar was present the largest niche was measured.

image

Figure 1. Schematic presentation of niche measurement. The following measurements were performed: (1) niche depth (in the sagittal plane); (2) residual myometrium (RM), from the serosal surface of the uterus (without the white lining of the serosa) to the apex of the niche, perpendicular to the endometrium (in sagittal plane); (3) adjacent myometrial thickness (AMT). Adapted from Bij de Vaate et al.[11]

Download figure to PowerPoint

Gel instillation sonography

A speculum was inserted and a catheter (SIS Rudigoz Catheter, CCD International, Paris, France) was placed inside the uterus. Gel (Endosgel, Farco-Pharma GmbH, Köln, Germany) was flushed inside the uterine cavity and cervix while retracting the catheter. A maximum volume of 10 cm3 was instilled, or less if uterine cramps occurred or reflux was observed from the cervix. The speculum was removed, the vaginal probe was reinserted, and the presence of a niche was recorded in the sagittal plane where the niche had the greatest depth. Niche characteristics were measured and recorded as described above (Figure 1). The shape was classified according to the previous classifications published by Bij de Vaate et al.[11] The presence of uterine polyps, fibroids, or other intrauterine abnormalities was noted.

Data collection

The characteristics of participants, including contraceptive use, breastfeeding, bleeding characteristics before pregnancy, and detailed obstetric history (dates of previous caesarean section, indication, and characteristics of the last pregnancy and caesarean section) were noted immediately after informed consent was given. Current contraceptive use and bleeding pattern were recorded at 6–12 weeks after the caesarean section. Women completed and returned a questionnaire at 6 and 12 months after the caesarean section. The questionnaire included questions on current medication or contraceptive use, breastfeeding, bleeding pattern (including duration of menstruation, postmenstrual spotting, vaginal discharge, and urinary incontinence). They were also asked to complete a validated menstrual score chart: the Pictorial Blood Assessment Chart (PBAC).[17] All data were recorded in a web-based database by two research nurses. The results of the ultrasound scans were not recorded in the case notes, and women and their doctors were not informed of the ultrasound findings. All women who underwent TVU were included in the analyses. Only women who underwent both TVU and GIS were included in the comparison of TVU and GIS measurement outcomes.

The primary outcome was the prevalence of a niche as measured by TVU and GIS 6–12 weeks after caesarean section. Secondary outcome measures included details of the bleeding pattern, including postmenstrual spotting (defined as more than 2 days of brownish discharge after the end of the menstrual period), number of days of menstrual bleeding, urinary incontinence at 6 and 12 months after caesarean section, and the relationship of these with a niche, and niche characteristics (i.e. depth, residual myometrium, and the ratio of the residual myometrium divided by the total thickness of the adjacent anterior myometrium), identified by TVU and GIS.

Statistical analyses

Data were analysed using the spss 20 (SPSS Inc., Chicago, IL, USA). Differences in baseline characteristics were compared using the chi-square test, Fisher's exact test, Student's t–test, or the Mann–Whitney U–test, depending on the type and distribution of the variables. Logistic regression analysis was planned to analyse the relationship between the presence of a niche and postmenstrual spotting. Potential confounding factors were predefined, and included age, breastfeeding, body mass index (BMI), BMI above 25 kg/m2, oral contraceptive use, and use of the levonorgestrel intrauterine system (LNG-IUS). These potential confounding factors were tested using univariate analysis. All tests were two-sided. A two-tailed P < 0.05 was considered to be statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Between November 2007 and September 2010, 350 women gave informed consent to participate in the study within 3 days of their caesarean section at St Antonius Hospital. Of these, 276 were willing to receive a TVU at 6–12 weeks, and 263 women could be included. TVU was used to examine 263 women, and 197 underwent both TVU and GIS (see Figure 2). Two women underwent a second caesarean section during the study period, but only the results of the first caesarean section and the related TVU and GIS were included in the study. TVU and GIS were performed at a mean of 7.6 ± 1.7 (SD) weeks after caesarean section. In total, 72% (191/263) of all women who underwent a TVU completed the questionnaire at 6 months, and 69% (172/249) of the women who received a questionnaire at 12 months after the caesarean section completed it. Of all the women who completed the questionnaire at 12 months, 45 were pregnant again. Of the women who underwent both a TVU and GIS the response rate was 73% at 12 months. Baseline characteristics and sonography results at 6–12 weeks after caesarean section are shown in Table 1. The majority of the women (71%) analysed had had only one previous caesarean section. Sixty-four (24%) had had two and eleven (4%) had had three previous caesarean sections. One caesarean section was carried out by a J–shaped or ‘hockey stick’ incision. All other women had a transverse lower segment incision. The uterus was closed in two layers in four women and in one layer in all others. Most caesarean sections (55%) were performed because of an emergency. During the sonographic evaluation at 6–12 weeks after caesarean section, 55% of the women were breastfeeding and 77% were amenorrhoeic. At 6–12 weeks, the women reported using the following methods of contraception: no contraception, 36%; oral contraceptives, 10%; progesterone only pill, 3%; sterilisation, 4%; condoms, 42%; and other, 5%. The difference in contraceptive use was not statistically significant between patients with or without a niche.

Table 1. Characteristics of the study participants
 All women (n = 263)Niche (TVU) (n = 129)No niche (TVU) (n = 131) P a Niche (GIS) (n = 127)No niche (GIS) (n = 70) P b
  1. a

    P value between niche and no niche during TVU.

  2. b

    P value between niche and no niche during GIS.

  3. c

    No information on breastfeeding for 13 patients at 6–12 weeks after caesarean section.

Baseline characteristics
Age (years), mean (±SD)32.46 (±4.15)32.66 (±4.16)32.26 (±4.15)0.4432.51 (±4.11)32.73 (±3.85)0.72
Parity, mean (±SD)1.54 (±0.70)1.55 (±0.73)1.54 (±0.68)0.961.57 (±0.78)1.50 (±0.68)0.52
BMI (kg/m2), mean (±SD)25.70 (±5.58)26.15 (±5.48)25.28 (±5.66)0.2825.98 (±5.83)24.98 (±5.19)0.29
Number of caesarean sections, mean (±SD)1.33 (±0.56)1.35 (±0.57)1.31 (±0.54)0.551.35 (±0.59)1.26 (±0.50)0.26
Visit at 6–12 weeks after caesarean section
Weeks after caesarean section, mean (±SD)7.58 (±1.74)7.40 (±1.40)7.74 (±2.01)0.127.66 (±1.77)7.72 (±1.92)0.82
Breastfeeding, number (percentage)c142 (55%)67 (52%)75 (±59%)0.3864 (53%)41 (61%)0.29
Ultrasonographic results at 6–12 weeks after caesarean section
Double endometrial thickness TVU (mm), mean (±SD)5.49 (±3.80)5.61 (±3.85)5.36 (±5.77)0.615.36 (±3.57)4.86 (±3.33)0.34
Uterus length TVU (cm), mean (±SD)7.18 (±1.43)7.29 (±1.52)7.07 (±1.32)0.237.09 (±1.38)7.11 (±1.47)0.92
Uterus width TVU (cm), mean (±SD)4.23 (±0.94)4.26 (±0.81)4.19 (±1.07)0.594.27 (±0.84)4.17 (±1.20)0.49
Thickness of myometrium at site of caesarean section scar (mm), mean (±SD)9.59 (±4.18)8.51 (±3.90)10.73 (±4.18)<0.0018.59 (±3.81)10.20 (±3.59)0.005
image

Figure 2. Flow diagram of study participants.

Download figure to PowerPoint

There were no differences regarding the rate of primary or emergency caesarean section, age, or parity between women who were included in our analyses and those who refused to participate, those who did not undergo a GIS, or those who did not complete the questionnaire at 12 months.

Niche prevalence and characteristics

Niche prevalence was 49.6% when evaluated by TVU and 64.5% when evaluated by GIS. In three women it was not possible to identify the scar in the uterus with TVU. With GIS all scars were visible. In women with one caesarean section, 62% who underwent GIS had a niche, compared with 68.2% of women with two caesarean sections and 77.8% of women with three caesarean sections. Sonographic characteristics of the uterus are shown in Tables 1 and 2. The length and width of the uterus were no different between women who had or did not have a niche. The niche was deeper when examined by GIS than by TVU (2.32 ± 3.35 and 3.03 ± 3.1 mm, respectively; P < 0.001). The thickness of the residual myometrium at the site of the uterine scar was approximately 2 mm less in women with a niche compared with those without a niche, as measured by TVU and GIS (P < 0.001 and P = 0.005, respectively; see Table 1). The mean ratio between the thickness of the residual myometrium at the site of the niche and the thickness of the adjacent myometrium was 0.74 ± 0.58 in all women who underwent GIS. The ratio was <0.5 in 22% of these women. Most niches detected by GIS had a semicircular shape (55%); 24% had a triangular shape, 10% had a droplet shape, 6% were inclusion cysts, and 7% had another shape.

Table 2. Comparison of transvaginal ultrasound and gel instillation sonohysterography results (n = 197)a
 TVUGISP (95% CI)
  1. a

    Of the 263 women in total, only the 197 women who underwent both TVU and GIS were included in this table.

  2. b

    Niche depth was recorded as 0 when no niche was observed.

Niche depth of all women, mean (±SD)b2.32 (±3.35)3.03(±3.1)<0.001 (0.38–1.02)
Niche depth of women with niche on GIS, mean (±SD)b3.43 (±3.67)4.77 (±2.64)<0.001 (0.92–1.79)
Residual myometrial thickness of all women, mean (±SD)9.58 (±4.39)9.11(±3.8)0.06 (–0.24 to 0.96)
Residual myometrial thickness of women with niche on GIS, mean (±SD)9.30 (±4.67)8.59(±3.75)0.03 (0.09–1.38)

Bleeding pattern and urological symptoms

The questionnaire at 6 months was completed by 191 women. Information on the menstrual cycle was available for only 59 women. Given the large proportion of women who were still breastfeeding or who did not yet have a regular menstrual cycle we decided not to analyse the outcomes at 6 months. The questionnaire at 12 months was completed by 172 women. The menstrual pattern could not be analysed in 45 of these women because of a subsequent pregnancy. Ten of these women completed the information on the basis of their menstruation pattern before pregnancy. Information on the menstrual cycle and contraceptive use was available for 137 women, of whom 17 were amenorrhoeic because of LNG-IUD or another hormonal contraceptive. Postmenstrual spotting was reported by 13 out of 45 (28.9%) women with a niche detected by GIS, compared with two women out of 29 (6.9%) without a niche detected by GIS (OR 5.48, 95% CI 1.14–26.48; Table 3). Including only primiparae, again postmenstrual spotting was reported more frequently in women with a niche detected by both TVU or GIS compared with those without a niche (31 versus 4% for TVU and 32 versus 0% for GIS; OR 9.7, 95% CI 1.1–85.6). Women with a ratio of residual myometrium of less than half of the adjacent myometrium (ratio <0.5) measured by TVU or GIS reported postmenstrual spotting more often than women with a ratio of >0.5 (OR 7.2, 95% CI 1.74–21.62, and OR 6.1, 95% CI 1.94–26.70, respectively; Table 4).

Table 3. Uterine bleeding pattern and urological symptoms at 12 months after caesarean section
 Niche (TVU)No niche (TVU) P a Niche (GIS)No niche (GIS) P b OR (95% CI)
  1. Data are presented as number (%), unless indicated with mean (±SD).

  2. Three women for whom the presence of a niche could not be determined were excluded.

  3. a

    P value between niche and no niche during TVU.

  4. b

    P value between niche and no niche during GIS.

  5. c

    Number of women who completed the questionnaire for a specific parameter and who underwent TVU/GIS, respectively.

No. of women with postmenstrual spotting (n = 91/74)c13 (26)5 (12)0.1013 (28.9)2 (6.9)0.025.48 (1.14–26.48)
Days of blood loss during menstruation, mean (±SD) (n = 91/83)c5.9 (2.57)6.1 (2.84)0.755.5 (±2.1)5.5 (±1.4)0.98 
No. of women with vaginal discharge (n = 137/98)c11 (14)11 (18)0.847 (11)8 (23)0.15 
No. of women taking oral contraceptives (n = 137/109)c19 (25)13 (21)0.6119 (28)10 (24)0.82 
No. of women with LNG-IUD (n = 137/109)c9 (12)8 (13)0.8210 (15)2 (5)0.21 
No. of women with amenorrhoea (n = 137/109)c12 (16)5 (8)0.189 (13)3 (7)0.53 
No. of women with incontinence (n = 115/90)c20 (31)17 (34)0.7119 (33)10 (29)0.82 
No. of women with stress incontinence (n = 115/90)c15 (23)11 (22)0.8914 (25)6 (17)0.45 
No. of women with urge incontinence (n = 115/90)c4 (6)3 (6)0.972 (4)2 (6)0.63 
No. of pregnant women (n = 260/197)c17 (13)28 (21)0.0821 (16.5)15 (21)0.40 
No. lost to follow–up/non-responders (n = 263/197)42 (33)49 (37)0.4145 (35)18 (26)0.16 
Total (n = 260/197)c129 (49.6)131 (50.4) 127 (64.5)70 (35.5)  
Table 4. Niche characteristics in women with or without postmenstrual spotting (n = 91/74)a
 Postmenstrual spotting (TVU) (n = 18)No postmenstrual spotting (TVU) (n = 73)P (OR, 95% CI)bPost-menstrual spotting (GIS) (n = 15)No postmenstrual spotting (GIS) (n = 59)P (OR, 95%CI)c
  1. AM, adjacent myometrium; RM, residual myometrium.

  2. a

    Number of women who completed the postmenstrual spotting questionnaire and who underwent TVU/GIS, respectively. All patients who completed the questionnaire on postmenstrual spotting were included.

  3. b

    P value between patients with and without postmenstrual spotting who underwent TVU.

  4. c

    P value between patients with and without postmenstrual spotting who received a GIS.

  5. d

    Ratio RM/AM = residual myometrial thickness at the site of the caesarean scar/myometrial thickness of the adjacent myometrium, presented in numbers (percentage).

Niche depth (mm), mean (±SD)5.46 (±2.50)5.15 (±2.77)0.725.45 (±2.76)4.86 (±2.50)0.47
Residual thickness of myometrium at the site of the caesarean scar (mm), mean (±SD)8.92 (±4.76)9.31 (±4.28)0.738.19 (±5.06)9.22 (±4.11)0.40
No. of women with ratio RM/AM <50%d (TVU n = 69;GIS n = 73)8 (61%)10 (18%)0.03 (7.2,1.94–26.70)8 (57%)10 (18%)0.005 (6.13,1.74–21.63)
Mean ratio RM/AMd0.500.720.0080.530.700.03

Using univariate analyses none of the predefined confounding factors or baseline characteristics were related to postmenstrual spotting: all had P > 0.1 (LNG-IUD use, P = 0.11; oral contraceptive use, P = 0.24; breastfeeding, P = 1.0; age, P = 1.0; BMI, P = 1.0; BMI > 25 kg/m2, P = 0.40), and therefore the planned logistic regression was not performed. We did not find any significant differences between women with or without a niche concerning the existence of (combined) urinary incontinence, urge incontinence, or stress incontinence (see Table 3).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

Main findings

All caesarean section scars could be identified using GIS. Using this technique the prevalence of niches at the site of these caesarean section scars was high (64.5%) in a random population. Comparing the results of GIS with TVU, niche prevalence was higher, measured niche depth was greater, and residual myometrium was thinner when detected by GIS. Postmenstrual spotting 1 year after caesarean section was strongly related to the presence of a niche detected by both TVU and GIS. One out of three women with a niche detected by GIS reported postmenstrual spotting, compared with one out of ten women without a niche. Postmenstrual spotting was related to a residual : adjacent myometrium ratio of <50%, which might indicate that this is a relevant parameter for niche size. The prevalence of urinary incontinence was not related to the presence or absence of a niche.

Strengths and limitations of the study

To the best of our knowledge this is the first study that has recruited participants at a very early stage after their caesarean section (within 3 days), followed by early examination (within 3 months) by both TVS and GIS, and with the sequential completion of menstrual questionnaires up to 1 year after the caesarean section. Only two prospective studies have evaluated niches in an unselected population.[11, 18] Both these studies recruited patients at 6–12 months after caesarean section. The Bij de Vaate et al.[11] study was carried out by our study group, but in a different population and with the examinations performed by a different examiner. In the present study, early scanning (3 months after the caesarean section) meant that the vast majority of the women were still amenorrhoeic, so the influence of abnormal uterine bleeding on the willingness of women to participate (selection bias) is expected to be limited. An additional advantage of early sonographic evaluation is that, because of amenorrhoea in most patients, the potential confounding effect of variations in timing of the measurements during the menstrual cycle and related endometrial thickness is limited. In addition, women were consecutively asked to participate by research midwives, and the registration of the results was independent of clinical considerations. Women and their doctors were not informed whether or not a niche was observed to prevent possible bias in reporting their bleeding pattern. Although complete post-caesarean section scar healing may take up to 6 months,[6] we do not know the exact time needed to develop a niche. We defined a niche as an anechoic space (with or without fluid) with a depth of at least 2 mm at the presumed site of the caesarean section scar. This should not be confused with the caesarean section scar itself. As demonstrated in this study, a scar is almost always visible and is reflected as a hypoechoic indentation of the myometrium at the location of the caesarean section. We expect the uterine scar to continue to heal beyond 6–12 weeks, and this might differ in performance with time. However, we consider it very unlikely that niches (i.e. discontinuations and fluid-stained spaces in the myometrium) will heal after this period. However, changes in a niche over time are not unlikely. Sustained accumulation of fluid may, in theory, increase niche size over time, but no data are currently available to confirm this theory.

A possible limitation of our study is that not all women who were delivered by caesarean section during the study period were asked to participate. Women were asked consecutively by trained midwives during their day and night shifts, but these midwives were not always present. Because the absence of the midwives was randomly distributed over the week, we consider this potential effect on selection bias to be minor. This is underlined by the lack of differences in baseline characteristics between women who were asked to participate and those who were not, with regards to elective or emergency caesarean section, age, and parity. In addition, we did not observe any differences in these parameters between included and excluded women, or those who were not willing to undergo a GIS or who did not complete the questionnaire.

Patients with a previous twin pregnancy or a known uterine anomaly were excluded in order to assemble a homogeneous group and to prevent undesired effects of unknown confounding factors; however, whether these factors would in fact introduce significant confounding effects is a topic for further discussion. We also decided not to perform a GIS in patients with a placental remnant, haematoma, or fluid in the uterine cavity detected by TVU. Consequently, a relatively small proportion of all patients who provided informed consent were analysed. Whether it is correct to exclude women with intrauterine fluid collection from examination by GIS should be further discussed. This was decided before the start of the study, and was included in the study protocol; however, it can be postulated that this group might have been particularly symptomatic and prone to postmenstrual spotting. We could not observe any relationship between niches and urinary incontinence, but we did not report other urinary symptoms, such as urgency. An additional shortcoming of the study is that we did not measure the niche width in the sagittal plane or the niche length in the transverse plane, as recently proposed by Naji et al.[5] (our study had already begun before that publication). Therefore, we could not analyse the relationship between these parameters and niche symptomatology. We also did not determine the reproducibility of our measurements.

Clinical implications of the findings and relation to published data

We were able to identify the caesarean section scar with GIS in 100% of the women. This is in line with another study examining women within 3 months of a caesarean section.[19] The observed niche prevalence rate of 49.6% using TVU is comparable with the reported prevalence (42%) in a prospective cohort study of an unselected population.[9] Different prevalence rates (28 and 61%) were reported in prospective studies in populations evaluated at 6 months and at 6–12 months after a caesarean section.[11, 13] Different definitions of a niche and different timing after the caesarean section may both affect the niche detection rate. Vikhareva Osser et al.[13] used the definition ‘any visible defect’, and Bij de Vaate et al.[11] used ‘any indentation of at least 1 mm’. We used a cut-off level of 2 mm; however, consensus on the exact cut-off levels is lacking. Early scanning may facilitate the recognition of the location of the caesarean section scar in the uterine wall as a result of incomplete scar healing, and this may increase the detection of small niches. In addition, the related thin endometrium resulting from breastfeeding in the majority of the women may also improve niche recognition and measurement.

We used the same terminology as that suggested by Naji et al.,[5] and measured the depth and residual myometrial thickness in the sagittal plane as described. Niche prevalence using GIS was 64.5% in our study, which is comparable with the reported prevalence using GIS or SIS in random, unselected populations (56, 59.5, 59, and 78%).[10, 11, 14, 18] Niche prevalence was higher using GIS than with TVU only. These findings are in line with the reported findings of three studies comparing GIS or SIS with TVU.[10, 11, 18] These studies showed a higher niche prevalence with sonohysterography than with TVU without contrast, and thus GIS may be more accurate at detecting niches. Although the niches that were detected using GIS but were missed using TVU were smaller than those detected by both TVU and GIS, they can be clinically relevant. This is underlined by the strong relationship we found between postmenstrual spotting and niches on GIS.[11] An additional argument for using GIS rather than TVU is that the residual myometrium measured using GIS was significantly thinner. The residual myometrium is considered to be the main limiting factor for eventual hysteroscopic niche resections to treat related bleeding symptoms. Most publications report a required residual myometrium of 2–3 mm for hysteroscopic niche resection, given the risk of perforation and/or bladder injury.[20] In order to prevent undesired complications during hysteroscopic niche resections we propose that a GIS or SIS be used rather than TVU for the preoperative evaluation of the residual myometrium. Several previous studies indicated a relationship between large niches and postmenstrual spotting. Bij de Vaate et al.[11] reported the niche volume to be related to postmenstrual spotting. We were not able to calculate niche volume because of the lack of stored three-dimensional volumes. In order to assess the relationship between large niches and postmenstrual spotting, we decided not to take an absolute measurement of residual myometrium, but to relate it to the adjacent myometrium. We defined large niches as those with a residual myometrium with thickness of <50% of that of the adjacent myometrium. This parameter was significantly related to postmenstrual spotting. This parameter and cut-off level were also used by Ofilli-Yebovi et al.,[15] who reported a high prevalence of women with a ratio of less of than 50% in a population with gynaecological symptoms. A potential relationship between niche size and postmenstrual spotting is in line with the hypothesis that spotting is induced by the accumulation of blood inside the niche.[6, 7] A depth of more than half the myometrial thickness makes the anterior part of the niche possibly large enough to obstruct the direct outflow of menstrual blood. This, in combination with lower contractility as a result of fibrosis, may induce the accumulation of blood in a niche.

One out of three women with a niche observed by GIS reported postmenstrual spotting. This is in line with the previous reported prevalence of postmenstrual spotting in patients with a niche.[10, 11] Given the high prevalence of postmenstrual spotting after caesarean section, this should be part of routine counselling before elective caesarean section.

Future perspectives

Future research should focus on the relationship of niches to subsequent fertility, obstetric complications such as uterine rupture, and on the impact of a niche on a woman's well-being. There is a lack of information on the impact of niche-related bleeding disorders on women's quality of life, their sexual function, and their willingness to undergo treatment for related symptoms.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

In conclusion, the caesarean section scar was visible in all women at 6–12 weeks after caesarean section using GIS. The prevalence of niches detected by GIS is high after caesarean section (64.5%), and more niches are detected using GIS than using TVU, with a larger observed niche size and reduced residual myometrial thickness. The presence of a niche is related to postmenstrual spotting. Postmenstrual spotting is more frequent in patients with large niches (defined as a residual myometrium of thickness <50% of that of the adjacent myometrium) than in patients with smaller niches.

Disclosure of interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

None of the authors have any relevant financial, personal, political, or religious interest linked to the subject of this article.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

The study was conceived by L.V., J.H., H.B., and S.V., and L.V. and S.V. performed the ultrasounds and collected the data. Analysis of the data was performed by L.V. and J.H. The first draft was written by L.V. and J.H. S.V., H.B., and A.B. helped to supervise the article up to the final draft.

Details of ethics approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

The study received ethics approval from the united committee on research involving human subjects (VCMO), Nieuwegein, the Netherlands (27 September 2007; ref. no. NL18722.100.07 R-07.14A/SCAR).

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References

We thank K.W. Van den Berg-Swart and C. van Dam-Bourens, research nurses, for recording all the collected data in the database. We thank M. Stigter-de Vries MD for performing the ultrasounds.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgements
  13. References