Uterine rupture and dehiscence: Ten-year review and case-control study.

Background. Previous cesarean section, oxytocin administration, and fetal macrosomia increase the risk of uterine rupture or dehiscence (URD).

Methods. All 25,718 deliveries at Riverside Regional Medical Center from January 1990 to June 2000 were reviewed to describe complications and identify risk factors for URD.

Results. Eleven uterine ruptures and 10 dehiscences occurred during this period (0.08%). One maternal death (5%) and three neonatal deaths (14%) occurred. Other complications included intrapartum nonreassuring fetal status (67%), 5-minute Apgar score <7 (52%), maternal blood transfusion (24%), neonatal hypoxic injury (14%), hysterectomy (14%), and endometritis (10%). Uterine rupture/dehiscence was independently associated with fetal weight [greater than or equal to]4,000 g, nonreassuring fetal status, use of oxytocin, and previous cesarean delivery; internal fetal monitoring reduced the risk of URD.

Conclusions. To reduce the risk of URD, a delivery plan should include assessment of cesarean history and fetal macrosomia, judicious use of oxytocin, and intrapartum monitoring for nonreassuring fetal status.

**********

UTERINE RUPTURE is an uncommon obstetric complication with potentially devastating outcome for both mother and baby. Medical terminology used to describe uterine injury during delivery is imprecise, with the overlapping terms "window," "dehiscence," and "rupture" often used to describe various clinical manifestations. Rupture and dehiscence describe complete separation of the uterine wall (endometrium, myometrium, and serosa). (1) Rupture is often traumatic and may occur in an intact uterus or involve the majority of a uterine scar from previous cesarean delivery. (2) Dehiscence is a separation that involves only a portion of the uterine scar. (2)

"Windows," believed to arise from lack of complete healing of the original scar, (3) are a partial rather than a complete separation of uterine wall layers. Operative reports often describe windows as membranes so thin they can be seen through.

Uterine rupture complicates 0.05% of all pregnancies, (3) a reported incidence with no appreciable change since 1930. (4) The etiology of uterine rupture has been affected by the changing trends of obstetric practice. Today, the most common cause of rupture is separation of a previous cesarean scar. (2,5) Whereas scar dehiscence has an incidence of 0.6%, the risk of rupture increases minimally to 0.8% after previous lower segment cesarean section and greatly (>5%) after classical cesarean section. (3) Other predisposing factors include previous uterine trauma, congenital anomaly, abnormal placentation, and inappropriate oxytocin administration. (2)

The aim of this study was to review all cases of URD seen at Riverside Regional Medical Center (RRMC) during a 10-year period. Comparison of these case-patients and a control group allows both the identification of risk factors and development of methods for prevention of URD.

METHODS

Using ICD9-CM codes, we identified 37 women as having URD at RRMC from January 1990 to June 2000. We retrospectively reviewed prenatal, labor, delivery, and postpartum records and interviewed the patients' personal physicians. Sixteen patients were excluded because of inappropriate coding of the diagnosis of URD (Table 1). Among the remaining 21 patients with URD, one maternal (and fetal) death occurred before delivery, allowing 20 patients with URD and complete perinatal records to be compared with control patients in a case-control study. Four control patients were randomly selected from all deliveries that occurred during the same month and year of each case-patient. The control group was selected from all deliveries regardless of mode of delivery.

Characteristics of the antepartum, intrapartum, and postpartum course were described and analyzed, crude odds ratios (OR) were estimated, and 95% confidence intervals (CI) were calculated with exact methods. Multiple risk factors were confirmed, and adjusted OR estimates were obtained with multiple unconditional logistic regression analysis. All exposures, known risk factors, and potential confounding factors for LTRD were entered into logistic regression models if the respective P value from univariate analysis was less than .20. Final models were determined with likelihood ratio test and Hosmer-Lemeshow goodness-of-fit test using EpiInfo 2000 and Statistical Analysis System (SAS) version 8. (6)

Since controls were selected randomly from RRMC deliveries listed in the month and year of case-patient deliveries, potential bias from unintentional time-based matching of case-patients and controls was evaluated. The analysis was repeated as a matched case-control study, using month and year of delivery as the matching factor. Conditional multiple logistic regression was used to provide effect measure estimates and CIs with EpiInfo 2000.

RESULTS

Of the 25,718 women whose infants were delivered at RRMC from 1990 to 2000, 21 women (0.08% of these deliveries) had URD (Table 2). Although cesarean deliveries decreased and vaginal birth after cesarean increased during this period, a time trend in URD incidence was not apparent. Among women with URD, mean age at delivery was 28 years (range, 15 to 38 years), mean gravidity was 4 (range, 1 to 8), and mean parity was 2 (range, 0 to 5).

The prevalence of maternal and fetal complications was similar when women with uterine rupture were compared with women who had uterine dehiscence (Table 3). One maternal death and two fetal deaths occurred among the deliveries complicated by uterine dehiscence. Two cesarean-hysterectomies were done because the uterine defects were large and unrepairable; one of these patients had already had repair of uterine rupture with a previous pregnancy. Another woman had hysterectomy to control postpartum hemorrhage due to placenta accreta. Pediatric charts were reviewed for those neonates with 5-minute Apgar score <7. One infant had complications of severe anoxic brain injury and died at 7 months of age; three other infants lived with permanent neurologic deficits from anoxic brain injury.

Crude ORs for multiple exposures and risk factors for URD were compared for 20 case-patients with complete medical records and 80 control patients (Table 4). Because URD is rare, ORs derived from categorical analysis should approximate risk ratios. (7)

Demographic and Antepartum Factors

Risk of URD increased greatly with history of previous cesarean delivery. After one or two cesaran deliveries, risk of URD increased tenfold (OR 9.9 [95% CI 2.6, 38]), and after 3 or more deliveries risk increased 30-fold (OR 30 [1.9,1600]) (Table 4). Uterine anomaly, forceps delivery, twin gestation, previous myomectomy, external version, and placenta previa were not associated with URD in this population.

Intrapartum Factors

Fifteen case-patients labored; 5 had elective cesarean delivery, and 9 had a "trial of labor" (ie, labor allowed when two or fewer previous cesarean deliveries are noted) (Table 3). Labor induction or augmentation was used in 14 of the 20 case-patients (13 with oxytocin, 1 with prostaglandin). Labor alone reduced the risk of URD (OR 0.3 [0.1, 1.5]), whereas oxytocin augmentation doubled the risk (OR 2.2 [0.7, 7.1]). Internal monitors were used among 9 casepatients, and their use reduced the risk of URD (OR 0.6 [0.2, 1.8]). Nonreasuring fetal status noted on monitoring equipment was associated with a sevenfold increased risk of URD (OR 7.4 [2.3, 25]). When noted, nonreasuring fetal status was the primary indication for cesarean delivery. Epidural anesthesia was used in 13 case-patients but was not found to be a significant risk factor for URD.

Postpartum Factors

Two case-patients received postoperative antibiotics for endometritis, and two required blood transfusions. Although endometritis was associated with URD, control patients did not have the other maternal and fetal postpartum factors, and OR estimates for these postpartum factors were therefore undefined (ie, the estimate was infinite). Fetal macrosomia (birth weight [greater than or equal to]4,000 g) and 5-minute Apgar score <7 were associated with increased risk of URD (Table 4).

Multivariable Logistic Regression Analyses

To determine which exposures or risk factors were independent predictors of URD in this study, demographic, antepartum, and intrapartum factors were studied with multiple unconditional logistic regression models. Except for birth weight, postpartum factors were not modeled because this information would not be available to the obstetrician developing a delivery plan; birth weight was used as a surrogate for estimated fetal weight, since the latter measure was not available in these patient records.

Logistic regression analysis confirmed five independent factors associated with URD (Table 5). Internal fetal monitoring greatly reduced the risk of URD, while history of previous cesarean delivery, presence of nonreassuring fetal status, large birth weight, and augmentation of labor with oxytocin increased the risk of URD. When these data were reanalyzed as a matched case-control study with month and year of delivery as the matching factor, OR estimates from conditional logistic regression were found to be similar to those obtained from unconditional logistic regression (Table 5).

DISCUSSION

The incidence of URD at RRMC was low during this period (0.08%), similar to the reported national average (0.05%). (3,8) Maternal and neonatal morbidity from URD was evidenced by the high proportion of cesarean deliveries, cesarean-hysterectomy, treatment of postoperative endometritis, need for postoperative blood transfusion, low 5-minute Apgar scores, and permanent neurologic deficits among some neonates. Although uterine dehiscence is thought to be a more benign process, both maternal and fetal morbidity were high among women with dehiscence. According to this case-control study, previous cesarean delivery, nonreassuring fetal status, large fetal weight, and oxytocin augmentation of labor greatly increase the risk of URD, while internal fetal monitoring greatly reduces risk of URD.

Uterine rupture remains an uncommon obstetric problem, with an incidence essentially unchanged for 70 years despite increases in cesarean delivery rates. (4) Uterine rupture may be considered the most significant adverse outcome of vaginal birth after cesarean because of the potentially disastrous results to mother, fetus, or both. Of note, vaginal birth after cesarean increased during this 10-year period, yet there was no increase in the annual incidence of URD. Although this study revealed oxytocin augmentation as an independent risk factor for URD, a trial of labor was associated with reduced risk of URD and was not found to be an independent predictor of URD. These findings imply that the current definition of a "trial of labor" (ie, labor allowed when two or fewer previous cesarean deliveries are noted) is appropriate. It is accepted that the risk of URD increases greatly with a history of three or more cesarean sections as confirmed in this study. (9)

Despite these findings, caution and sound clinical judgment must be used when developing a delivery plan. This study suggests that planning is most critical when the patient has had previous cesarean delivery and if fetal macrosomia is suspected. In the past, oxytocin stimulation of labor has been associated with uterine rupture, especially in women of higher parity. (3) Today, uterine rupture due to oxytocin stimulation is rare even among parous women, but uterine rupture rates are increased if the uterus is scarred. (3) Oxytocin augmentation is not considered an independent cause of URD, (10,11) and the findings of this study may be controversial. Differences in the effect of oxytocin on URD may arise between studies because the protocols for oxytocin use may differ between institutions.

Bias from at least five causes may influence the findings of this study. First, all retrospective studies may be limited, since needed clinical information may be missing at the time of study. This study did have missing exposure information for some risk categories, but these were not found to be important predictors of URD in the final logistic regression models. Second, previous studies suggest that manual extraction of the placenta and uterine surgery or trauma increase subsequent risk of URD. (2) Since manual extraction of the placenta may be associated with cesarean delivery and since dilation and curettage may be associated with abortion, assessment of risk from these previous exposures may be inadequate because precise exposure information is lacking. Third, controls were selected independent of delivery mode even though every case-patient had a cesarean delivery. Repeat analysis with a second control group having cesarean deliveries resulted in a logistic model with similar findings. Fourth, random s election of controls from lists recorded during the month and year of case-patient delivery may suggest that this was a matched case-control study. Matching was not intended because of the inherent bias introduced by this method of control selection. (12) Repeat analysis post hoc with conditional logistic regression confirmed the risk factor estiamtes provided by unconditional logistic regression, lending some reassurance that inadvertent bias from time-based matching had not occurred.

Finally, ICD9-CM coding errors appeared to have occurred because of the vague terminology regarding URD as well as human error. Although true ruptures are less likely to be overlooked because of severe adverse clinical consequences, benign dehiscence may have been overlooked in some patients at the time of discharge. The effect of this bias cannot be accurately estimated with these data because manual exploration of the uterus after successful vaginal birth after cesarean is needed to diagnose benign dehiscence, and this practice was not documented in these delivery records.

TABLE 1

Exclusion Criteria for Patients With Uterine Rupture or Dehiscence
(ICD9-CM), Riverside Regional Medical Center, 1990 to 2000

      Reason for Exclusion        No. Patients Excluded

Skin/fascial wound dehiscence            8 (50%)
No documentation of uterine              5 (31%)
 Injury
Iatrogenic extension of uterine          1 (6%)
 incision at time of cesarean
 section
Operative report describe window         2 (13%)

  Total                                 16 (100%)
TABLE 2

Uterine Rupture and Dehiscence, Riverside Regional Medical Center, 1990
to 2000

                                  Vaginal Birth    No. Cases
           Total      Cesarean    After Cesarean  of Rupture/
Year     Deliveries  Deliveries      Delivery     Dehiscence

1990        2,003      701 (35%)     13 (0.06%)        2
1991        1,997      732 (37%)     19 (0.9%)         2
1992        3,001      667 (22%)    126 (4.2%)         0
1993        2,922      629 (22%)    138 (4.7%)         3
1994        2,798      639 (23%)    108 (3.8%)         1
1995        2,602      508 (20%)    155 (6.0%)         3
1996        2,518      481 (19%)    165 (6.5%)         2
1997        2,308      452 (20%)    114 (4.9%)         4
1998        2,314      461 (20%)    132 (5.7%)         1
1999        2,215      487 (22%)    119 (5.3%)         1
2000        1,040      208 (20%)     59 (5.6%)         2

  Total    25,718    5,965 (23%)  1,148 (4.4%)        21


         Incidence of
Year      Dehiscence

1990        0.10%
1991        0.10%
1992        0.00%
1993        0.10%
1994        0.04%
1995        0.12%
1996        0.08%
1997        0.17%
1998        0.04%
1999        0.05%
2000        0.19%

  Total     0.08%
TABLE 3

Maternal and Fetal Morbidity and Mortality After Uterine Rupture or
Dehiscence, Riverside Regional Medical Center, 1990 to 2000

                              Uterine     Uterine
                              Ruptures  Dehiscenses
Complication                  (N = 11)   (N = 10)    Total

Maternal

  Endometritis                 2 (18%)    0 (0%)       2
  Hysterectomy                 1 (9% )    2 (20%)      3
  Transfusion                  3 (27%)    2 (20%)      5
  Death                        0 (0%)     1 (10%)      1

Fetal

  Nonreassuring fetal status   9 (82%)    5 (50%)     14
  5-minute Apgar score <7      7 (64%)    4 (40%)     11
  Neurologic deficit           2 (18%)    1 (10%)      3
  Death                        1 (9%)     2 (20%)      3
TABLE 4

Prevalence of Known Risk Factors and Other Exposures for Uterine Rupture
or Dehiscence, Riverside Regional Medical Center, 1990 to 2000

Exposure or                          Case-Patients  Controls
Risk Factor                            (N = 20)     (N = 80)

Demographic

   Maternal age >20 years               1 (5%)      10 (13%)
   Gravidity [less than or equal
   to]2                                 9 (45%)     37 (46%)
   Parity [less than or equal to]1     11 (55%)     54 (68%)
   Gestation [greater than or equal
   to]41 weeks (*)                      5 (25%)      6 (8%)

Antepartum

   Previous uterine surgery             6 (30%)     32 (40%)
   History of uterine anomalies         1 (5%)       0 (0%)
   Placenta previa or accreta (+)       1 (10%)      1 (1%)
   Attempted external version           0 (0%)       0 (0%)
   Previous cesarean sections
      None                              7 (35%)     69 (86%)
      1 or 2                            9 (50%)      6 (13%)
      3 or 4                            4 (15%)      5 (1%)

Intrapartum

   Epidural anesthesia                 13 (65%)     47 (59%)
   Labor                               15 (75%)     72 (90%)
   Trial of labor (**)                  9 (69%)      9 (82%)
   Internal fetal monitoring (++)       9 (45%)     45 (58%)
   Oxytocin use                        13 (65%)     37 (46%)
   Misoprostol use                      1 (5%)       2 (13%)
   Nonreassuring fetal status          13 (65%)     16 (20%)
   Cesarean delivery                   20 (100%)    18 (23%)

Postpartum

   Birth weight [greater than or
   equal to]4,000 g                     5 (25%)      4 (10%)
   5-minute Apgar score <7              9 (45%)      7 (3%)
   Endometritis                         2 (10%)      2 (3%)

Exposure or                            Odds Ratio
Risk Factor                            (95% of CI)

Demographic

   Maternal age >20 years            0.4 (0.0, 2.9)
   Gravidity [less than or equal
   to]2                              1.0 (0.3, 2.8)
   Parity [less than or equal to]1   0.6 (0.2, 1.8)
   Gestation [greater than or equal
   to]41 weeks (*)                   3.7 (0.8, 6.6)

Antepartum

   Previous uterine surgery          0.6 (0.2, 2.0)
   History of uterine anomalies      Undefined
   Placenta previa or accreta (+)    4.2 (0.0, 330)
   Attempted external version        Undefined
   Previous cesarean sections
      None                           Referent
      1 or 2                         9.9 (2.6, 38)
      3 or 4                         30  (1.9, 1600)

Intrapartum

   Epidural anesthesia               1.3 (0.4, 4.3)
   Labor                             0.3 (0.1, 1.5)
   Trial of labor (**)               0.5 (0.0, 4.8)
   Internal fetal monitoring (++)    0.6 (0.2, 1.8)
   Oxytocin use                      2.2 (0.7, 7.1)
   Misoprostol use                   0.8 (0.0, 17)
   Nonreassuring fetal status        7.4 (2.3, 25)
   Cesarean delivery                 Undefined

Postpartum

   Birth weight [greater than or
   equal to]4,000 g                  3.0 (0.7, 12)
   5-minute Apgar score <7           31  (5.2, 310)
   Endometritis                      4.3 (0.3, 62)

Data were available for:

(*)73 Controls.

(+)10 Case-patients and 16 controls.

(**)Previous cesarean in 13 case-patients and 11 controls.

(++)77 Controls.
TABLE 5

Crude and Adjusted Odds Ratios for Independent Risk Factors for Uterine
Rupture or Dehiscence, Riverside Regional Medical Center, 1990 to 2000

                                   Adjusted Estimates

                                Crude       Unconditional
Risk                         Odds Ratio      Odds Ratio
Factor                        (95% CI)        (95% CI)

Internal fetal monitoring   0.6 (0.2,1.8)  0.08 (0.01,0.7)
Previous cesarean sections
  None                      Referent       Referent
  1 or 2                    9.9 (2.6,38)   28 (4.4,180)
  3 or 4                    30 (1.9,1600)  113 (4.3,2980)
Nonreassuring fetal status  7.4 (2.3,25)   33 (4.1,257)
Fetal weight [greater than
 or equal to]4,000 g        3.0 (0.7,12)   8.7 (1.3,58.7)
Use of oxytocin             2.2 (0.7,7.1)  7.6 (1.1,53.9)

                            Adjusted Estimates

                              Conditional
Risk                           Odds Ratio
Factor                          (95% CI)

Internal fetal monitoring   0.05 (0.003,0.9)
Previous cesarean sections
  None                      Referent
  1 or 2                    29 (2.6,326)
  3 or 4                    135 (3.1,592)
Nonreassuring fetal status  37 (2.1,652)
Fetal weight [greater than
 or equal to]4,000 g        6.9 (0.7,68.0)
Use of oxytocin             6.9 (0.5,93.8)

References

(1.) Depp R: Cesarean delivery. Obstetrics: Normal and Problem Pregnancies. Gabbe SG, Niebyl JR, Simpson JL (eds). New York, Churchill Livingstone Inc, 3rd Ed, 1986, pp 606-607

(2.) Cunningham FG, MacDonald PC, Cant NF, et al: Obstetrical hemorrhage. williams Obstetrics. Stamford Conn, Appleton and Lange, 20th ed, 1997, pp 772-773, 777-778

(3.) Lynch JC, Prady JP: Uterine rupture and scar dehiscence. a five-year survey. Anaesth Intensive Care 1996; 24:699,702

(4.) Eden RD, Parker RT, Gall SA: Rupture of the pregnant uterus: a 53-year review. Obstet Cynecol 1986; 68:671

(5.) Yussman MA, Haynes DM: Rupture of the gravid uterus. a 12-year study. Obstet Gunecol 1970; 36:116

(6.) Hosmer DW, Lemeshow S: Applied Logistic Regression. New York, John Wiley & Sons, 2000, pp 145-156

(7.) Clayton D, Hills M: Statistical Models in Epidemiology. Oxford, Oxford University Press, 1993, p 8

(8.) Farmer RM, Kirschbaum T, Potter D, et al: Uterine rupture during trial of labor after previous cesarean section, Am J Obstet Gynecol 1991; 165:996

(9.) ACOG Practice Bulletin: Clinical Management Guidelines for Obstetricians-Gynecologists. Compendium of Selected Publications, July 1999, No. 5, p 1110

(10.) Suner S, Jagminas L, Peipert F, et al: Fatal spontaneous rupture of a gravid uterus: case report and literature review of uterine rupture. J Emerg Med 1996; 14:181-185

(11.) Rosen MG, Dickinson JC, westhoff CL: Vaginal birth after cesarean: a meta-analysis of morbidity and mortality. Obstet Cynecol 1991; 77:465470

(12.) Breslow NE, Day NE: Statistical Methods in cancer Research. The Analysis of Case-Control Studies. Lyon, International Agency for Research on Cancer, Vol 1,1980, p 162

RELATED ARTICLE: KEY POINTS

* The incidence of uterine rupture in our study was 0.08%.

* Significant morbidity is associated with uterine rupture/dehiscence, including maternal or fetal death, transfusion, endometritis, hysterectomy, and fetal anoxic brain injury.

* The risk of uterine rupture/dehiscence is increased with a history of previous cesarean section, oxytocin use, and fetal macrosomia.

From the Department of Obstetrics and Gynecology and the Department of Family Practice, Riverside Regional Medical Center, Newport News, Va. (Riverside Regional Medical Center is associated with Virginia Commonwealth University, Richmond.) (Dr. Seryakov is currently with Multi-Care Associates, Minneapolis, Minn, and Dr. Mann is currently with Jersey Shore Medical Center, Neptune, NJ.)

Reprint requests to Sumac D. Diaz, MD, 316 Main St, Second Floor, Newport News, VA 23601.