Between 1965 and 1991, cesarean deliveries in the United
States increased from 4.5% to 23.5% of all births.1 There
is general agreement among health care professionals and consumers that
current cesarean rates have been too high. The American College of Obstetricians
and Gynecologists (ACOG) has offered no guideline regarding the "appropriateness"
of any cesarean delivery rate. The only opinion on this subject issued
by ACOG has referred to vaginal birth after a cesarean delivery.2
Although authors have suggested that rates varying from 8 to 15% are
appropriate targets,3,4 there has been little evidence on
which to base the choices of target rates for defined populations that
would result in acceptable rates of maternal and neonatal morbidity.
Strategies such as performing a trial of labor after cesarean delivery,
managing labor actively, obtaining second opinions, providing labor
support personnel (e.g., doulas) or using narcotic analgesia (instead
of epidural anesthesia) have resulted in lower cesarean delivery rates.5-13
Further, using gross measures of perinatal outcome (e.g., Apgar scores,
acidemia at birth, and mortality), these investigators found that lowered
cesarean delivery rates did not increase the rates of adverse outcomes.
There is a serious need for reports that compare varying cesarean delivery
rates to maternal and infant morbidity. Cesarean deliveries are still
being performed in some settings for more than 25% of births while being
held to less than 10% in others. Replicated studies of safety and risk
will be required if the performance of cesarean deliveries is to be
based on sound evidence of benefit rather than on personal or institutional
practice styles.
From 1986 through 1989, the cesarean delivery rate in the Colorado Region*
of Kaiser Permanente decreased from 16.2 to 12.6%. This downward trend
in cesarean delivery provided an opportunity to study modes of delivery,
associated maternal and neonatal morbidity, and length of hospital stay
among patients in a group-model health maintenance organization.
Methods
In this retrospective study, we wished to learn whether a lowered cesarean
delivery rate was accompanied by increased maternal and neonatal morbidity
and length of hospital stay. The study population was composed of mothers
(and their liveborn neonates) who delivered in the Colorado Region of
Kaiser Permanente from July 1, 1986, through December 31, 1989. All
deliveries occurred at a large, urban, nonprofit hospital (Saint Joseph
Hospital, Denver, Colorado). There was a standard physician staffing
pattern during the study, in which at least two obstetricians were present
in the hospital 24 hours per day on a rotating basis.
Obstetric practice guidelines did not undergo formal change during this
time, nor is there evidence that the demographic composition of the
population changed. Generally, the population base included employed
women and wives and daughters of employed men.
Discharge data were provided by the hospital for July 1, 1986 through
December 31, 1989. Maternal and neonatal records of more than 8,000
births were linked manually. Because mothers' and neonates' discharge
records were maintained separately and because some mothers' names differed
from those of their offspring, it was often necessary to search neonatal
charts in order to find mothers' names. The proportion of mothers who
could not be matched to neonates ranged from 8.4 to 9.7% per year. Our
outcome data refer only to matched pairs of mothers and neonates. We
classified maternal and neonatal morbidity, maternal postpartum length
of stay, and neonatal length of hospital stay by mode of delivery. Two
single breech vaginal births in 1988 and 1989 were omitted from analysis;
no morbidity was associated with these births.
Maternal morbid conditions (Figure 1) were identified by ICD-9 codes.
We selected codes that were most likely to be associated with events
of delivery rather than with underlying health problems. We differentiated
vaginal births by whether they were spontaneous, forceps, or vacuum
deliveries. Neonatal morbidity was identified using selected ICD-9 diagnostic
codes. In Figure 2, three general categories are shown: 1) birth trauma,
2) intrauterine hypoxia and birth asphyxia, and 3) other conditions
of the fetus and newborn. Data were first analyzed by the chi-squared
method. Odds ratios and 95% confidence intervals were calculated. Years
with the highest (1986) and lowest (1989) rates of cesarean delivery
were compared, with 1989 as the reference year. Continuous data (lengths
of stay) were analyzed with t-tests, comparing 1986 with 1989 (SAS version
6.10, Statistical Analysis Systems, Cary, NC). We considered a P value
of < 0.05 significant.
 |
|
 |
|
Results
Mode of Delivery
In Table 1, births are reported by mode of delivery for each year of
the study. The rate of cesarean delivery was lower among our study participants
in 1989 (11.4%) than in 1986 (15.9%). This decline was significant (P
< .001). Rates of spontaneous and vacuum deliveries did not differ
in these two years. The rate of forceps delivery in 1989 was higher
than in 1986 (P = .047). We also noted that the rate of cesarean delivery
in our study group was lower than that of the entire Kaiser Permanente
population for each year (e.g., 15.9% vs 16.2% in 1986). We attribute
this finding to attrition from the population of mothers and neonates
whom we could not match.
 |
|
Maternal Morbidity
Maternal morbidity is displayed in Table 2. Some mothers had more than
one condition. The percentages reflect the total number of morbid conditions
divided by the number of deliveries by year and mode of delivery. The
differences in morbidity were not significant for any mode of delivery.
Overall, there was no increase in maternal morbidity during the year
of the lowest cesarean rate.
Neonatal Morbidity
Neonatal morbidity is shown in Table 3. Neonatal morbidity decreased
in each year of the study with the exception of a slight rise among
neonates delivered by vacuum extraction in 1989, reflecting a greater
proportion of injuries to the scalp. For each mode of delivery, there
were significantly fewer problems in 1989 compared with 1986.
 |
|
Several specific conditions were collectively identified as "birth
trauma" (Figure 2). An analysis of these conditions occurring among
all neonates during the study years revealed their infrequent occurrence.
In 8,387 deliveries, there were only 9 instances of subdural and cerebral
hemorrhage (0.1%); 243 (2.9%) injuries to the scalp, including cephalohematoma;
5 clavicular fractures (0.06%); 57 fractures of large bones or skull
(0.7%); 4 cases of facial palsy (0.05%); and 6 cases of Erb's palsy
(.07%). It was of interest, however, that there were proportionately
more injuries to the scalp resulting from vacuum births than other modes
of delivery in all study years. Overall birth trauma is shown in Table
4, with significantly fewer occurrences amongst neonates delivered spontaneously
in 1986 compared with 1989. When the 243 cases of injury to the scalp
(ICD-9, 767.1) were removed from this analysis, there was no statistically
significant difference between the two years.
|
|
 |
|
Maternal Postpartum Length of Hospital Stay
Mean maternal postpartum lengths of stay by mode of delivery and year
are compared in Figure 3. Overall, there was a significant difference
when the years of low (1989) and high (1986) cesarean delivery rates
were compared for all modes of delivery. The mean maternal postpartum
lengths of stay for all deliveries in 1986 and 1989 were 2.3 days ±
1.3 and 1.6 ± 1.1, respectively (P< .0001). Also, compared
with 1986, lengths of stay were significantly shorter for mothers in
1989 for each mode of delivery.
 |
FIGURE 3. Mean maternal postpartum lengths
of stay by mode of delivery and year. Compared with 1986, the overall
mean length of stay was significantly shorter in 1989.
(P < .0001). |
Neonatal Length of Stay
As anticipated, the shortest lengths of stay were experienced by vaginally
delivered neonates (mean for all years = 2.0 ± 3.8 days). During
the years studied, only 5% of those with cesarean birth stayed for two
days or less, while 91.6% of those delivered spontaneously, 85.6% of
those delivered by vacuum extraction, and 84.2% of those delivered by
forceps stayed for two days or less. Trends over the years can be seen
in Figure 4. The mean neonatal lengths of stay for all deliveries were
2.7 ± 3.0 days in 1986 and 2.3 ± 5.1 days in 1989 (P =
0.0078). However, for each mode of delivery, analyzed separately, there
was no difference in neonatal length of stay in 1986 and 1989.
 |
| FIGURE 4: Mean neonatal lengths of stay
by mode of delivery and year. Compared with 1986, the overall mean
length of stay was significantly shorter in 1989 (P = .0078). |
Comment
In this retrospective study, we found that a significant decrease in
the cesarean rate was not accompanied by an increase in important delivery-related
maternal and neonatal problems as identified from hospital discharge
data. In fact, we observed a significant decrease in overall neonatal
morbidity in the year of the lowest cesarean rate compared with the
highest year. We saw an increase in "birth trauma" in the
spontaneously delivered group in 1989 compared with 1986 because of
the number of injuries to the scalp (ICD-9, 767.1). The main conditions
of this code, caput succedaneum and cephalohematoma, are entities which
usually resolve without clinical sequelae. When these "injuries"
were removed from the analysis of birth trauma, the difference was no
longer significant.
During the two comparison years, overall maternal postpartum length
of stay decreased. This finding is compatible with national trends in
postpartum care over the period studied. Length of stay also decreased
for each of the three modes of vaginal delivery in 1989 compared with
1986. The lower rate of cesarean delivery in 1989 also contributed to
the decrease in maternal length of stay. We also found a decrease in
neonatal length of stay for all modes of delivery combined (P = .0078).
Although these results support the view that cesarean rates can be lowered
without harm, we recognize limitations of this study. First, although
we saw a significant decrease in rates of cesarean birth over the study
years, we are unable to substantiate the reasons for the decline. We
speculate that the decrease was due to fewer cesarean deliveries for
dystocia and to fewer repeated cesarean deliveries. However, hospital
discharge records, the source of our study data, do not provide data
regarding the demographic characteristics of mothers such as obstetric
history or chronic medical conditions which would clearly influence
the selection of a mode of delivery. Second, we cannot provide evidence
that clinical practices (e.g., management of labor, use of oxytocin)
remained exactly the same during the study period.
As noted earlier, several studies have demonstrated that some strategies
lower the cesarean delivery rate. Programs using trials of labor after
cesarean birth, active management of labor, second opinions before performing
cesarean delivery, labor support personnel, or using narcotic analgesia
rather than epidural anesthesia have all lowered cesarean birth rates.5-13
Traditionally troubling perinatal outcomes such as low Apgar scores,
acidemia at birth, admissions to NICU, and perinatal mortality have
not increased when cesarean rates were lowered. These papers reported
no increase in other adverse outcomes, including maternal endometritis,
maternal transfusion, need for ventilatory support, hyperbilirubinemia,
neurologic abnormalities, hypoglycemia, and seizures. In contrast to
the aforementioned papers, the decrease in our cesarean rate occurred
without any implementation of formal program or guidelines. This phenomenon
has been previously reported by Sandmire and DeMott,14 who
also observed no change in traditional measures of perinatal morbidity.
Although "target rates" of 8 to 15% were achieved routinely
in the past, they occur infrequently today. To assure that good perinatal
outcomes are maintained when cesarean rates are lowered, precise definitions
of maternal and neonatal morbidity must be used. Organizations should
develop both perinatal and neonatal data bases to better track the effect
of changes in the rates. Only in this way can appropriate cesarean birth
rates for defined populations be established. Some hospitals serve higher-risk
mothers than others and hospital cesarean rates may vary with the number
of mothers with high risk factors. In 1986, our cesarean rate was already
near the target range. Our data indicate that the further lowering to
the 1989 rate was not accompanied by a clinically significant increase
in maternal or neonatal morbidity.
*Now part of the Rocky Mountain Division.
References:
1. Rates of cesarean deliveryUnited States, 1991. MMWR 1993;42:285-9.
2. American College of Obstetricians and Gynecologists. Vaginal delivery
after a previous cesarean birth. Committee Opinion 143. Washington,
DC: American College of Obstetricians and Gynecologists October
1994.
3. Quilligan E. Cesarean section: modern perspective. In: Queenan JT,
ed. Management of high-risk pregnancy. 2nd ed. Oradell, NJ: Medical
Economics Books; 1985:594-600.
4. United States. Public Health Service. Healthy people 2000: national
health promotion and disease prevention objectives: full report, with
commentary. Washington, DC: USGPO, 1990. (DHHS publication; no. (PHS)
91-50212.
5. Porreco RP. High cesarean section rate: a new perspective. Obstet
Gynecol 1985;65:307-11.
6. Myers SA, Gleicher N. A successful program to lower cesarean-section
rates. N Engl J Med 1988;319:1511-6.
7. Sanchez-Ramos L, Kaunitz AM, Peterson HB, Martinez-Schnell B, Thompson
RJ. Reducing cesarean sections at a teaching hospital. Am J Obstet
Gynecol 1990;163:1081-7; discussion 1087-8.
8. Socol ML, Garcia PM, Peaceman AM, Dooley SL. Reducing cesarean births
at a primarily private university hospital. Am J Obstet Gynecol 1993;168(6
Pt 1):1748-54; discussion 1754-8.
9. Akoury HA, Brodie G, Caddick R, McLaughin VD, Pugh PA. Active management
of labor and operative delivery in nulliparous women. Am J Obstet
Gynecol 1988;158:255-8.
10. Boylan P, Frankowski R, Rountree R, Selwyn B, Parrish K. Effect
of active management of labor on the incidence of cesarean section for
dystocia in nulliparas. Am J Perinatol 1991;8:373-9.
11. Lopez-Zeno JA, Peaceman AM, Asashek JA, Socol ML. A controlled trial
of a program for the active management of labor. N Engl J Med
1992;326:450-4.
12. Kennell J, Klaus M, McGrath S, Robertson S, Hinkley C. Continuous
emotional support during labor in a US hospital: a randomized controlled
trial. JAMA 1991;265:2197-201.
13. Thorp JA, Hu DH, Albin RM, McNitt J, Meyer BA, Cohen GR, et al.
The effect of intrapartum epidural analgesia on nulliparous labor: A
randomized, controlled prospective trial. Am J Obstet Gynecol
1993;169:851-8.
14. Sandmire HF, DeMott RK, The Green Bay cesarean section study. III.
Falling cesarean birth rates without a formal curtailment program. Am
J Obstet Gynecol 1994;170:1790-9; discussion 1799-802.
