Abstract
We conducted an observational retrospective cohort study to evaluate the risk factors and the maternal morbidity associated with unintended extensions of the hysterotomy during caesarean delivery. We evaluated 2707 women who underwent low-transverse caesarean deliveries in 2011 at an academic, tertiary-care hospital. Hysterotomy extensions were identified through operative reports. Of the 2707 caesarean deliveries, 392 (14.5%) had an unintended hysterotomy extension. On the multivariable regression modelling, neonatal weight (OR 1.42; 95%CI 1.17–1.73), the arrest of labour [first-stage arrest (2.42; 1.73–3.38); second-stage arrest (5.54; 3.88–7.90)] and a non-reassuring foetal status (1.65; 1.20–2.25) were significantly associated with hysterotomy extensions. Hysterotomy extensions were significantly associated with an increased morbidity including an estimated blood loss >1200 millilitres (2.06; 1.41–3.02), a decline in postoperative haemoglobin ≥3.7 g/dL (2.07; 1.35–3.17), an evaluation for lower urinary tract injury (5.58; 3.17–9.81), and a longer operative time (8.11; 6.33–9.88). Based on these results, we conclude that unintended hysterotomy extensions significantly increase the maternal morbidity of caesarean deliveries.
Keywords: Caesarean section, caesarean delivery, hysterotomy, extension, morbidity, retrospective
Introduction
Over 1.2 million women undergo caesarean delivery (CD) in the United States each year (Hamilton et al. 2015). In 2014, CD was the method of delivery for 32.2% of all births (Hamilton et al. 2015). The maternal morbidity associated with CD is significantly greater than vaginal delivery due to higher rates of blood loss, transfusion and unplanned hysterectomy (Curtin et al. 2015). Unintended extensions of the uterine incision frequently occur at the time of CD with an estimated incidence of 4–8% (Rodriguez et al. 1994; de la Torre et al. 2006; Cromi et al. 2008). These extensions can involve the broad ligament, cervix and vagina and thus, may contribute to the increased morbidity associated with CD by increasing the surgical complexity and blood loss (Sung et al. 2007).
Unintended hysterotomy extensions have often been included as secondary outcomes in studies assessing other risk factors for CD morbidity such as a prolonged second stage of labour (Sung et al. 2007; Asicioglu et al. 2014a). In an evaluation of CDs performed following labour, the hysterotomy extensions were significantly more common after a prolonged second stage of labour (Sung et al. 2007). Variations in surgical technique including blunt versus sharp creation of the hysterotomy have also been assessed (Rodriguez et al. 1994; Magann et al. 2002; Cromi et al. 2008; Sekhavat et al. 2010; Xu et al. 2013; Asicioglu et al. 2014b; Xodo et al. 2016). While there is a trend towards a reduced incidence of hysterotomy extensions after blunt creation, this has not been statistically significant in meta-analysis (RR 0.57; 95%CI 0.28–1.17) (Xu et al. 2013).
As the relationship between unintended hysterotomy extensions and CD morbidity has been largely limited to secondary outcome evaluations, the clinical sequelae and maternal morbidity associated with hysterotomy extensions have been incompletely evaluated. The purposes of this study were to assess the frequency of and identify risk factors for unintended hysterotomy extensions and to evaluate the impact of unintended extensions on maternal outcomes. We hypothesised that unintended hysterotomy extensions would be associated with clinically significant maternal morbidity including intraoperative haemorrhage, urinary tract injury and longer operative times.
Materials and methods
We conducted a retrospective cohort study of all CDs performed from 1 January 2011 to the 31 December 2011 at a large, tertiary-care academic medical centre. International Classification of Diseases, Ninth Revision (ICD-9) codes for CD were used to identify the sample. The data were extracted from the medical records and validated through a double entry process (Shree et al. 2016). Any discrepancies were resolved by a third review of the medical record. The University of Pittsburgh Institutional Review Board approved this study as a minimal risk and waived the need for informed consent.
The demographic variables included race, parity, insurance status, tobacco use and comorbidities. Gestational hypertension (HTN), included gestational HTN, pre-eclampsia, superimposed pre-eclampsia, or eclampsia and diabetes mellitus (DM) was dichotomised into pre-gestational or gestational. Chorioamnionitis was diagnosed by the physician based on clinical and laboratory criteria necessary for the diagnosis (Higgins et al. 2016). Private practice CDs were performed by a private practice physician with either a resident physician or a physician’s assistant. The resident teaching service CDs were performed by a fourth-year resident and a first or second-year resident under the supervision of a teaching service physician. While we did not specifically assess the impact of the physician training level on the hysterotomy extension, the ‘resident teaching service’ variable serves as a proxy for the training level.
Labour was defined as the onset of contractions, cervical change, spontaneous or artificial rupture of membranes and/or the administration of an induction or augmentation agent. Non-laboured primary CDs were elective or performed for malpresentation, placenta previa, multifetal gestation, foetal growth restriction or anomalies, previous shoulder dystocia, fourth-degree laceration or ano-rectal disease or previous uterine surgery. Non-laboured repeat CDs were defined as elective repeat CDs. A first-stage arrest was defined after the onset of labour and with a cervical exam of fewer than 10 cm. The second-stage arrest was defined after the onset of labour and a complete cervical dilation. Due to a small sample size (n = 56), CDs performed following a failed trial of labour after a previous CD were included in the category corresponding to the CD indication. A non-reassuring foetal status included CDs for placental abruption, umbilical cord prolapse or a non-reassuring foetal heart tracing.
A hysterotomy extension was defined as any documentation of an extension by the surgeon in the operative note. The diagnosis of a hysterotomy extension was at the discretion of the operating physician. The estimated blood loss (EBL) in millilitres (ml) was extracted from the CD operative note. An intraoperative haemorrhage was defined as the 95th percentile for EBL (>1200 ml in our dataset). A change in the haemoglobin was the difference between the preoperative haemoglobin and the lowest postoperative haemoglobin value obtained during the hospitalisation. A clinically significant change in the postoperative haemoglobin was defined as ≥3.7 g/dL (95th percentile for this change). The need for a blood transfusion was defined as any transfusion of packed red blood cells intraoperatively or during the postoperative hospital course. The operative time was defined as the time in minutes from the skin incision to the skin closure. Because of the low frequency of cystotomy, an evaluation for lower urinary tract (LUT) injury was assessed as a composite outcome defined as the backfilling of the bladder, intraoperative cystoscopy or an actual cystotomy repair. Evaluation of the LUT was at the discretion of the surgeon.
The variables were summarised using a mean and standard deviation, a median and an interquartile range (IQR), or the frequencies and percentages as appropriate. The Chisquare, t-tests or Fisher’s exact test were used for the binary outcomes. Pearson’s correlation was used to examine the associations with the operative time. The predictors exhibiting p < .2 in the bivariate analysis were included in the initial multivariable models. To check for collinearity, the variance inflation factor (VIF) was calculated for each predictor. The variable ‘labour’ was multi-collinear and removed as a predictor from the multivariable models. To select the final models, stepwise backward elimination was used based on minimising the Akaike Information Criterion (Lawless and Singhal 1978). The prevalence of blood transfusion (n = 12) and cystotomy (n = 11) were low, not associated with the hysterotomy extension in the bivariate analysis, and were not evaluated in the multivariate analysis. The missing values (2.8%) were excluded from the final models. All of the analyses were conducted using STATA® 12 (College Station, TX) and the R Foundation for Statistical Computing 3.2 (Austria).
Results
Between 1 January 2011 and 31 December 2011, 2876 CDs were performed at our institution. We included only term pregnancies (≥37 weeks of gestation) and excluded the non-Pfannenstiel skin incisions (n = 34) and the hysterotomies that were not low transverse (n = 46). In our dataset, there were no hysterectomies performed following an unintended zhysterotomy extension, so all of the caesarean hysterectomies (n = 8) were excluded from the analysis. We also excluded the CDs in which any extra procedures were performed (i.e. drain placement, oophorectomy/cystectomy). Due to one or more of the previous diagnoses, 169 (5.9%) CDs were excluded.
Our final sample consisted of 2707 low transverse CDs performed via a Pfannenstiel skin incision and 392 (14.5%) had documentation of an unintended hysterotomy extension (Table 1). The mean age of the patients in our sample was 30 years. The majority of the patients were Caucasian, had private insurance, and had their CD performed by a private attending. The mean BMI of our cohort was 33.3 kg/m2. Approximately 17% had either chronic or gestational HTN, 11% had either pre-gestational or gestational DM and 4% of patients had a diagnosis of chorioamnionitis. The most common indication for a CD was an elective repeat CD. Over half of the CDs were performed after the onset of labour. Approximately 6% of the CDs had an EBL >1200 ml and 5% had a postoperative drop in haemoglobin of ≥3.7 g/dL. 2.4% required an intraoperative evaluation for LUT injury [a backfilling of the bladder (n = 77), an intraoperative cystoscopy (n = 24), or a cystotomy requiring repair (n = 11)], and less than 1% required a blood transfusion. The mean duration of a CD operative time was approximately 51 minutes.
Table 1.
Sample characteristicsa (n = 2707) and specific risk factors predictive of unintended hysterotomy extensionb.
| All (N = 2707) |
No extension (N = 2315) |
Extension (N = 392) |
p value | Unadjusted OR (95%CI) | Final adjusted ORa (95%CI) | |
|---|---|---|---|---|---|---|
| Patient characteristics | ||||||
| Age (years) | 29.8 (±5.7) | 29.9 (±5.7) | 28.9 (±5.7) | <.01 | 0.97 (0.95–0.99) | 0.99 (0.97–1.01) |
| Gestational age (weeks) | 38.7 (±2.2) | 38.6 (±2.3) | 39.5 (±1.6) | <.01 | 1.30 (1.21–1.39) | 1.06 (0.98–1.15) |
| Neonatal weight (grams) | 3290 (±672) | 3263 (±687) | 3446 (±555) | <.01 | 1.55 (1.30–1.85) | 1.42 (1.17–1.73) |
| Race | ||||||
| Caucasian | 2086 (77.7) | 1772 (77.2) | 314 (80.7) | .31 | 1.23 (0.91–1.67) | – |
| African-American | 453 (16.9) | 396 (17.3) | 57 (14.7) | ref | – | |
| Other | 144 (5.4) | 126 (5.5) | 18 (4.6) | 0.99 (0.56–1.75) | – | |
| BMI (kg/m2) | 33.3 (±7.1) | 33.3 (±7.2) | 33.3 (±6.8) | .76 | 1.00 (0.98–1.01) | – |
| Multiparous | 1405 (51.9) | 1287 (55.6) | 118 (30.1) | <.01 | 0.34 (0.27–0.43) | 0.90 (0.65–1.24) |
| Tobacco use | 219 (8.1) | 185 (8.0) | 34 (8.7) | .72 | 1.09 (0.75–1.60) | – |
| Insurance status | ||||||
| Private payer | 1551 (68.6) | 1323 (68.6) | 228 (68.5) | 1.00 | – | – |
| Medicaid | 710 (31.4) | 605 (31.4) | 105 (31.5) | 1.01 (0.78–1.29) | – | |
| Hypertensive disorders | ||||||
| Chronic HTN | 56 (2.1) | 47 (2.0) | 9 (2.3) | .85 | 1.12 (0.55–2.31) | – |
| Gestational HTN | 408 (15.1) | 352 (15.2) | 56 (14.3) | 0.93 (0.69–1.27) | – | |
| Diabetes mellitus | ||||||
| Pre-Gestational DM | 80 (3.0) | 67 (2.9) | 13 (3.3) | .89 | 1.15 (0.63–2.10) | – |
| Gestational DM | 213 (7.9) | 183 (7.9) | 30 (7.7) | 0.97 (0.65–1.45) | – | |
| Chorioamnionitis | 114 (4.2) | 79 (3.4) | 35 (8.9) | <.01 | 2.78 (1.84–4.20) | 1.01 (0.65–1.56) |
| Provider characteristics | ||||||
| Private practice | 1926 (73.0) | 1649 (73.2) | 277 (71.9) | .66 | – | – |
| Resident teaching service | 712 (27.0) | 604 (26.8) | 108 (28.1) | 1.064 (0.83–1.36) | – | |
| Procedure characteristics | ||||||
| Caesarean delivery indication | ||||||
| Non-laboured primary CD | 379 (14.0) | 364 (15.7) | 15 (3.8) | <.01 | 0.21 (0.13–0.36) | 0.34 (0.19–0.61) |
| Non-laboured repeat CD | 804 (29.7) | 759 (32.8) | 45 (11.5) | <.01 | 0.27 (0.19–0.37) | 0.47 (0.31–0.71) |
| First-stage arrest | 433 (16.0) | 326 (14.1) | 107 (27.3) | <.01 | 2.29 (1.78–2.95) | 2.42 (1.73–3.38) |
| Second-stage arrest | 307 (11.3) | 177 (7.6) | 130 (33.2) | <.01 | 5.99 (4.62–7.78) | 5.54 (3.88–7.90) |
| Non-reassuring foetal status | 498 (18.4) | 407 (17.6) | 91 (23.2) | .01 | 1.42 (1.10–1.83) | 1.65 (1.20–2.25) |
| Labour prior to CD | 1470 (54.3) | 1142 (49.3) | 328 (83.7) | <.01 | – | – |
| Intraoperative and postoperative factors | ||||||
| EBL >1200 ccc | 151 (5.6) | 110 (4.8) | 41 (10.5) | <.01 | na | na |
| Change in Hgb ≥3.7 g/Dlc | 123 (4.6) | 90 (3.9) | 33 (8.5) | <.01 | na | na |
| Blood transfusion | 12 (0.4) | 11 (0.5) | 1 (0.3) | 1.00 | na | na |
| LUT injury evaluationd | 64 (2.4) | 36 (1.6) | 28 (7.1) | <.01 | na | na |
| Operative time (minutes)e | 50.8 (±16.3) | 49.6 (±15.4) | 57.5 (±19.4) | <.01 | na | na |
BMI: body-mass index; HTN: hypertension; DM: diabetes mellitus; CD: caesarean delivery; Hgb: haemoglobin; EBL: estimated blood loss; LUT: lower urinary tract; OR: odds ratio; CI: confidence interval; na: not applicable.
n (%) or mean (±standard deviation), bolded values are statistically significant.
Final multivariable model with the lowest Akaike Information Criterion (AIC), model controlled for variables listed in table, bolded values are statistically significant value.
Represents 95th percentile.
Defined as backfilling the bladder or cystoscopy or cystotomy repair.
Defined as time from skin incision to skin closure in minutes.
The patient, provider and procedure characteristics for the patients with and without a hysterotomy extension were compared (Table 1). Hysterotomy extensions occurred in 22.3% of the patients who laboured prior to their CD compared to only 5.2% of patients who did not labour prior to their CD (p < .01). The patients with a hysterotomy extension were more likely to be younger (28.9 vs. 29.9 years; p < .01), at a greater gestational age (39.5 vs. 38.6 weeks; p < .01), have an infant with a greater neonatal weight (3446 vs. 3263 grams; p < .01) and be diagnosed with chorioamnionitis (8.9% vs. 3.4%; p < .01). The patients with a hysterotomy extension were also significantly more likely to have their CD performed for a first-stage arrest (27.3% vs. 14.1%; p < .01), a second-stage arrest (33.2% vs. 7.6%; p < .01) or a non-reassuring foetal status (23.2% vs. 17.6%; p =.01) and were significantly more likely to be in labour prior to their CD (83.7% vs. 49.3%; p < .01). The patients who did not have an unintended hysterotomy extension were more likely to be multiparous (55.6% vs. 30.1%; p < .01) and have a non-laboured CD (primary [15.7 vs. 3.8%; p < .01], repeat [32.8% vs. 11.5%; p < .01]). Finally, patients with an unintended hysterotomy extension were significantly more likely to have an EBL of >1200 cc (10.5% vs. 4.8%; p < .01), a drop in postoperative haemoglobin of 3.7 g/dL (8.5% vs. 3.9%; p < .01), an evaluation for an LUT injury (7.1% vs. 1.6%, p < .01) and a longer CD operative time (57.5 vs. 49.6 minutes; p < .01).
Multivariable logistic regression models were developed to identify risk factors predictive of hysterotomy extension at the time of a CD (Table 1). A greater neonatal weight (OR 1.42; 95%CI 1.17–1.73) and the CDs performed for a first-stage arrest (2.42; 1.73–3.38), a second-stage arrest (5.54; 3.88–7.90) and a non-reassuring foetal status (1.65; 1.20–2.25) were significantly more likely to have a hysterotomy extension. Conversely, the non-laboured CDs [primary (0.34; 0.19–0.61) and repeat (0.47; 0.31–0.71)] were significantly less likely to have a hysterotomy extension.
Table 2 describes the adjusted multivariable models for four perioperative complications hypothesised to be associated with an unintended hysterotomy extension. In the multivariable analyses, a hysterotomy extension was significantly associated with an EBL >1200 ml (OR 2.06; 95%CI 1.41–3.02), a decline in the postoperative haemoglobin of ≥3.7 g/dL (2.07; 1.35–3.17), an evaluation for LUT injury (5.58; 3.17–9.81) and a longer operative time (8.11 minutes; 6.33–9.88).
Table 2.
Maternal morbidity associated with unintended hysterotomy extensiona.
| Adjusted OR (95%CI)
|
||||
|---|---|---|---|---|
| Change in Hgb ≥3.7 g/dLb | EBL >1200ccb | LUT injury evaluationc | Operative timed | |
| Primary predictor | ||||
| Hysterotomy extension | 2.07 (1.35–3.17) | 2.06 (1.41–3.02) | 5.58 (3.17–9.81) | 8.11 (6.33–9.88) |
| Patient characteristics | ||||
| Age (years) | 1.05 (0.99–1.09) | |||
| Gestational age (weeks) | – | – | – | – |
| Neonatal weight (grams) | – | – | – | – |
| Race | ||||
| Caucasian | – | – | – | – |
| African–American | – | – | – | – |
| Other | – | – | – | – |
| BMI (kg/m2) | 0.96 (0.93–0.99) | 1.03 (1.01–1.06) | – | 0.40 (0.31–0.48) |
| Multiparous | – | – | 3.09 (1.65–5.77) | 3.24 (1.96–4.53) |
| Tobacco Use | – | – | – | – |
| Medicaid | – | – | – | – |
| Hypertensive Disorders | ||||
| Chronic HTN | 0.68 (0.09–5.05) | – | – | – |
| Gestational HTN | 3.65 (2.44–5.48) | – | – | – |
| Diabetes mellitus | ||||
| Pre-gestational DM | – | – | – | – |
| Gestational DM | – | – | – | – |
| Chorioamnionitis | – | – | – | – |
| Provider characteristics | ||||
| Resident teaching service | – | – | 1.62 (0.94–2.80) | 3.42 (2.06–4.78) |
| Procedure characteristics | ||||
| CD indication | ||||
| Non-laboured primary CD | – | – | – | – |
| Non-laboured repeat CD | 0.38 (0.21–0.67) | 0.46 (0.29–0.73) | – | – |
| First-stage arrest | – | – | – | – |
| Second-stage arrest | – | – | 2.22 (1.09–4.54) | 2.67 (0.62–4.73) |
| Non-reassuring fetal status | – | – | – | −4.43 [−6.04–(−2.83)] |
OR: odds ratio; CI: confidence interval; Hgb: haemoglobin; EBL: estimated blood loss; LUT: lower urinary tract; BMI: body-mass index; HTN: hypertension; DM: diabetes mellitus; CD: caesarean delivery.
Final multivariable models with the lowest Akaike Information Criterion (AIC), models controlled for variables listed in table.
Bolded values are statistically significant. Odds ratios used for Hgb, EBL and LUT injury evaluation; mean difference used for operative time.
Value represents 95th percentile.
Defined as backfilling the bladder or cystoscopy or cystotomy repair.
Defined as time from skin incision to skin closure in minutes.
Discussion
Unintended hysterotomy extensions are frequent complications of low transverse CDs, particularly after the onset of labour. Unintended hysterotomy extensions subsequently increase maternal morbidity by increasing the blood loss, the frequency of an LUT evaluation and the operative time. Our findings indicate that the rate of a hysterotomy extension in the general obstetric population is approximately 15%, which is higher than the previously reported estimates (Rodriguez et al. 1994; de la Torre et al. 2006; Cromi et al. 2008).
The CDs performed after the onset of labour had frequency of extensions four times greater than the patients who did not labour prior to their CD. The most significant independent risk factor for the hysterotomy extension was a second-stage labour arrest. This is consistent with the prior literature, demonstrating that a second-stage arrest is the most consistent risk factor for an increased maternal morbidity at the time of a CD, including hysterotomy extensions (de la Torre et al. 2006; Sung et al. 2007; Asicioglu et al. 2014a). Rodriguez et al. (1994) (n = 296) found an increase in the prevalence of unintended hysterotomy extensions with the increasing stages of labour [1.4% without labour, 15.5% after the first stage and 35.0% after the second stage (p < .01)]. Due to a lower uterine segment thinning that occurs with labour progression combined with the force necessary to disengage the foetal head from the pelvis, an increase in the incidence of a hysterotomy extension after labour is not unexpected.
Importantly, a hysterotomy extension was found to be an independent risk factor for an intraoperative haemorrhage and a significant drop in postoperative haemoglobin. In the analysis by de la Torre et al. (2006), the intraoperative blood loss was also greater in the CDs complicated by the extensions (p < .01), but they did not find a significant change between the preand postoperative haematocrit levels (p = .376) or the blood transfusion (0.6% vs. 0.8%; NS). Despite our low prevalence of blood transfusion, the association between hysterotomy extensions, a greater EBL and a greater postoperative haemoglobin decline has important clinical implications. Obstetric haemorrhage is the fourth most common cause of a pregnancy-related mortality in the United States and accounts for approximately 11% of maternal deaths (CDC 2016). Thus, we highlight the importance of the optimisation of haemoglobin during pregnancy to avoid the morbidity associated with an increased blood loss surrounding delivery. Additionally, utilising the cell salvage mechanisms in the setting of an unintended extension can be considered given the association with an increased blood loss.
Hysterotomy extensions were found to increase the risk of intraoperative evaluation for LUT injury and to have longer CD operative times. The prevalence of LUT injury in our cohort was low (0.4%), but consistent with the rates found in previous evaluations of LUT injury during CD (Oliphant et al. 2014). In contrast, an intraoperative evaluation for an LUT injury was common during the CDs complicated by a hysterotomy extension and highlights the increased surgical complexity of these cases. The procedures used to evaluate for an LUT injury such as backfilling the bladder and cystoscopy may also require an additional training to the providers and staff and an increase in resource utilisation (Oliphant et al. 2014). Relatedly, the hysterotomy extensions had the most significant impact on the operative time in the multivariable analysis. An increase in the CD duration is understandable, as the identification and repair of the extension, achieving haemostasis and evaluating the patients for an LUT injury takes time.
Upon comparing our results to those of studies over the last 10 years, the rate of hysterotomy extensions during routine CD has potentially doubled from approximately 6.6% to 14.5%, which is of significant clinical importance (de la Torre et al. 2006). Recent re-evaluations of the labour curve with a focus on extending the second stage of labour to increase vaginal delivery rates may have increased the frequency of CDs performed after a prolonged second stage (Kilpatrick and Laros 1989; Laughon et al. 2014; Gimovsky and Berghella 2016). Although we were unable to quantify the duration of labour in our cohort, 11% of patients had a CD performed after a second-stage arrest. Given the associated morbidities in this group of women, prospective studies are needed to identify the risk factors predictive of arrest to better stratify the patients who may ultimately require CD.
Our study must be interpreted in light of certain limitations. This study was performed in a university tertiary-care hospital and may not be applicable to the community settings. Our study was retrospective and based on the available medical records. The presence of a hysterotomy extension was based on the surgeon documentation and could not otherwise be confirmed, which may have underestimated the true prevalence of the extensions in our cohort. Likewise, we were unable to determine the size or location of the extension from the CD operative notes. Due to limitations associated with the electronic medical record templates which were used for the documentation, we were unable to determine how the hysterotomy was created (a blunt versus a sharp dissection). However, in a recent systematic review and meta-analysis, a blunt versus a sharp expansion of the hysterotomy was not significantly associated with the development of a hysterotomy extension (Xu et al. 2013). We were also unable to determine the length of the first stage and the second stage of labour. Lastly, the interventions to decrease the trauma to the uterus during the delivery of the impacted foetal head have been evaluated (Jeve et al. 2016; Seal et al. 2016; Waterfall et al. 2016). The reverse breech extraction through the hysterotomy (the ‘pull technique’) has been compared to assistance from a vaginal hand (the ‘push’ technique) in the CDs performed for a second-stage arrest in four randomised trials. The reverse breech extraction was significantly associated with a decrease in the risk of hysterotomy extension across all four trials (Waterfall et al. 2016). Unfortunately, we were unable to assess if either of these methods were used to assist the delivery of the foetal head, as this information was not uniformly recorded in the medical record. Further evaluation of methods such as these to aid the obstetrician during difficult caesarean delivery should and will likely be an area of continued research.
In conclusion, unintended hysterotomy extensions are common and are associated with an increase in maternal morbidity. A second-stage arrest is a strong independent risk factor for a hysterotomy extension and the impact of a prolonged second stage on the CD morbidity should be factored into the clinical decision-making. Once the decision for a CD is made after a second-stage arrest, surgeons should be prepared for a potentially longer procedure with an increased possibility of an unintended hysterotomy extension. Upon identifying a hysterotomy extension intraoperatively, the preparedness for a possible haemorrhage is warranted. It is also important to notify and prepare the operating room staff for the potential need to evaluate for an LUT injury. As one out of every three births result in a CD, the efforts to mitigate maternal morbidity associated with CD are of significant clinical importance.
IMPACT STATEMENT.
What is already known on this subject? Maternal morbidity associated with caesarean delivery (CD) is significantly greater than that in vaginal delivery. Unintended extensions of the hysterotomy occur in approximately 4–8% of CDs and are more common after a prolonged second stage of labour. The morbidity associated with hysterotomy extensions has been incompletely evaluated.
What do the results of this study add? We demonstrate a rate of hysterotomy extension in a general obstetric population of approximately 15%, which is higher than previously reported estimates, and represents a potential doubling of the rate of the unintended hysterotomy extensions in recent years. The most significant risk factor for a hysterotomy extension was a second-stage labour arrest with a fourfold increase in the frequency of extensions. A hysterotomy extension is a significant independent risk factor for an intraoperative haemorrhage, a drop in postoperative haemoglobin, an intraoperative evaluation for lower urinary tract injury, and longer CD operative times.
What are the implications of these findings for clinical practice and/or further research? A second-stage arrest is a strong independent risk factor for a hysterotomy extension. Recent re-evaluations of the labour curve that extend the second stage of labour will likely increase the frequency of CDs performed after a prolonged second stage. In these scenarios, obstetricians should be prepared for an unintended hysterotomy extension and for the possibility of a longer procedure with the increased risks of blood loss and the need for evaluation of the lower urinary tract.
Acknowledgments
Funding
This work was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award Number [KL2TR000146] (Dr Krans). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Disclosure statement
The authors report no declarations of interest.
References
- Asicioglu O, Gungorduk K, Asicioglu BB, Yildirim G, Gungorduk OC, Ark C. Unintended extension of the lower segment uterine incision at cesarean delivery: a randomized comparison of sharp versus blunt techniques. American Journal of Perinatology. 2014b;31:837–844. doi: 10.1055/s-0033-1361934. [DOI] [PubMed] [Google Scholar]
- Asicioglu O, Gungorduk K, Yildirim G, Asicioglu BB, Gungorduk OC, Ark C, et al. Second-stage vs first-stage caesarean delivery: comparison of maternal and perinatal outcomes. Journal of Obstetrics and Gynaecology. 2014a;34:598–604. doi: 10.3109/01443615.2014.920790. [DOI] [PubMed] [Google Scholar]
- CDC. Centers for Control Disease and Prevention: Causes of pregnancy-related death in the United States. 2011–2012 [Internet]; [cited 2016 Oct5]. Available from: http://www.cdc.gov/reproductivehealth/maternalinfanthealth/pmss.html.
- Cromi A, Ghezzi F, Di Naro E, Siesto G, Loverro G, Bolis P. Blunt expansion of the low transverse uterine incision at cesarean delivery: a randomized comparison of 2 techniques. American Journal of Obstetrics and Gynecology. 2008;199:292 e1–296. doi: 10.1016/j.ajog.2008.07.013. [DOI] [PubMed] [Google Scholar]
- Curtin S, Gregory K, Korst L, Uddin S. National vital statistics reports. Hyattsville, MD: National Center for Health Statistics; 2015. Maternal morbidity of vaginal and cesarean deliveries, according to previous cesarean history: New data from the birth certificate, 2013; p. 64. [PubMed] [Google Scholar]
- de la Torre L, González-Quintero VH, Mayor-Lynn K, Smarkusky L, Hoffman MC, Saab A, et al. Significance of accidental extensions in the lower uterine segment during cesarean delivery. American Journal of Obstetrics and Gynecology. 2006;194:e4–e6. doi: 10.1016/j.ajog.2006.01.026. [DOI] [PubMed] [Google Scholar]
- Gimovsky AC, Berghella V. Randomized controlled trial of prolonged second stage: extending the time limit vs usual guidelines. American Journal of Obstetrics and Gynecology. 2016;214:S30–S36. doi: 10.1016/j.ajog.2015.12.042. [DOI] [PubMed] [Google Scholar]
- Hamilton B, Martin J, Osterman M, Curtin S, Mathews TJ. National vital statistics reports. Hyattsville, MD: National Center for Health Statistics; 2015. Births: Final data for 2014; p. 64. [PubMed] [Google Scholar]
- Higgins RD, Saade G, Polin RA, Grobman WA, Buhimschi IA, Watterberg K, et al. Evaluation and management of women and newborns with a maternal diagnosis of chorioamnionitis: summary of a workshop. Obstetrics and Gynecology. 2016;127:426–436. doi: 10.1097/AOG.0000000000001246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jeve YB, Navti OB, Konje JC. Comparison of techniques used to deliver a deeply impacted fetal head at full dilation: a systematic review and meta-analysis. BJOG : An International Journal of Obstetrics and Gynaecology. 2016;123:337–345. doi: 10.1111/1471-0528.13593. [DOI] [PubMed] [Google Scholar]
- Kilpatrick SJ, Laros RK., Jr Characteristics of normal labor. Obstetrics and Gynecology. 1989;74:85–87. [PubMed] [Google Scholar]
- Laughon SK, Berghella V, Reddy UM, Sundaram R, Lu Z, Hoffman MK. Neonatal and maternal outcomes with prolonged second stage of labor. Obstetrics and Gynecology. 2014;124:57–67. doi: 10.1097/AOG.0000000000000278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lawless JF, Singhal K. Efficient screening of nonnormal regression models. Biometrics. 1978;34:318–327. [Google Scholar]
- Magann EF, Chauhan SP, Bufkin L, Field K, Roberts WE, Martin JN., Jr Intra-operative haemorrhage by blunt versus sharp expansion of the uterine incision at caesarean delivery: a randomised clinical trial. BJOG. 2002;109:448–452. doi: 10.1111/j.1471-0528.2002.01296.x. [DOI] [PubMed] [Google Scholar]
- Oliphant SS, Bochenska K, Tolge ME, Catov JM, Zyczynski HM. Maternal lower urinary tract injury at the time of Cesarean delivery. International Urogynecology Journal. 2014;25:1709–1714. doi: 10.1007/s00192-014-2446-2. [DOI] [PubMed] [Google Scholar]
- Rodriguez AI, Porter KB, O’Brien WF. Blunt versus sharp expansion of the uterine incision in low-segment transverse cesarean section. American Journal of Obstetrics and Gynecology. 1994;171:1022–1025. doi: 10.1016/0002-9378(94)90026-4. [DOI] [PubMed] [Google Scholar]
- Seal SL, Dey A, Barman SC, Kamilya G, Mukherji J, Onwude JL. Randomized controlled trial of elevation of the fetal head with a fetal pillow during cesarean delivery at full cervical dilatation. International Journal of Gynaecology and Obstetrics. 2016;133:178–182. doi: 10.1016/j.ijgo.2015.09.019. [DOI] [PubMed] [Google Scholar]
- Sekhavat L, Dehghani Firouzabadi R, Mojiri P. Effect of expansion technique of uterine incision on maternal blood loss in cesarean section. Archives of Gynecology and Obstetrics. 2010;282:475–479. doi: 10.1007/s00404-009-1251-5. [DOI] [PubMed] [Google Scholar]
- Shree R, Park SY, Beigi RH, Dunn SL, Krans EE. Surgical site infection following cesarean delivery: patient, provider, and procedure-specific risk factors. American Journal of Perinatology. 2016;33:157–164. doi: 10.1055/s-0035-1563548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sung JF, Daniels KI, Brodzinsky L, El-Sayed YY, Caughey AB, Lyell DJ. Cesarean delivery outcomes after a prolonged second stage of labor. American Journal of Obstetrics and Gynecology. 2007;197:306 e1–305. doi: 10.1016/j.ajog.2007.07.005. [DOI] [PubMed] [Google Scholar]
- Waterfall H, Grivell RM, Dodd JM. Techniques for assisting difficult delivery at caesarean section. Cochrane Database of Systematic Reviews. 2016:CD004944. doi: 10.1002/14651858.CD004944.pub3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xodo S, Saccone G, Cromi A, Ozcan P, Spagnolo E, Berghella V. Cephalad-caudad versus transverse blunt expansion of the low transverse uterine incision during cesarean delivery. European Journal of Obstetrics, Gynecology, and Reproductive Biology. 2016;202:75–80. doi: 10.1016/j.ejogrb.2016.04.035. [DOI] [PubMed] [Google Scholar]
- Xu LL, Chau AM, Zuschmann A. Blunt vs. sharp uterine expansion at lower segment cesarean section delivery: a systematic review with metaanalysis. American Journal of Obstetrics and Gynecology. 2013;208:62 e1–68. doi: 10.1016/j.ajog.2012.10.886. [DOI] [PubMed] [Google Scholar]