Lawrence Leeman, MD, MPH

Family physicians providing maternity care often face a scenario in which an otherwise low-risk, term patient is incidentally noted to have a low amniotic fluid index (AFI). Common reasons for obtaining an AFI in a woman with a low-risk pregnancy include evaluation of decreased fetal movement, spontaneous variable decelerations during monitoring to evaluate for labor, or an ultrasound evaluation for fundal height measurements discordant with gestational age. How should “isolated” oligohydramnios—an AFI <5 cm—be interpreted, and should immediate induction be recommended for such patients?

Oligohydramnios occurs in about 1% to 5% of pregnancies at term.^1,2 Because adverse outcomes occur in high-risk pregnancies complicated by low amniotic fluid volume, oligohydramnios commonly prompts labor induction.^1,3,4 At one university center, oligohydramnios is now the leading indication for labor induction.⁵ Many centers may even induce labor when the AFI is between 5 cm and 8 cm, the so-called borderline AFI.³

Labor induction increases the use of cesarean delivery, particularly for the primiparous woman with an unripe cervix.⁶ Recent studies questioning the safety of labor induction in women who have had a cesarean may increase the number of elective repeat cesarean procedures when delivery is believed indicated for oligohydramnios.⁷ (See Underlying causes of oligohydramnios.)

Oligohydramnios is difficult to assess

True oligohydramnios can be difficult to confirm due to the questionable accuracy of amniotic fluid measurement by ultrasound. There is controversy, for example, about whether (and how) to include pockets of amniotic fluid containing umbilical cord.¹⁵ The AFI was introduced in 1987² to replace the 2 cm “pocket technique” of fluid assessment, and studies continue to question to what extent the AFI reflects actual amniotic fluid volume.

AFI measurements may vary with the amount of pressure applied to the abdomen and with fetal position or movement.¹⁶

Serial measurements taken by the same ultrasound operator have been shown to differ from the true volume by 1 cm, or 10.8%; serial measurements taken by multiple operators have differed by as much as 2 cm, or 15.4%.^17,18

O’Reilly-Green compared the diagnosis of oligohydramnios in 449 post-term patients with actual amniotic fluid volume measured at rupture of membranes.¹⁹ They found a positive predictive value of 50% for oligohydramnios at an AFI of 5 cm as the lower limit of normal. A study of 144 third trimester patients using the dye-dilution technique found that, to achieve 95% confidence for ruling out oligohydramnios, a cutoff AFI of 30 cm would need to be used, a value consistent with polyhydramnios.²⁰

What is the association between oligohydramnios and poor fetal outcomes?

A number of studies over the past 15 years have shown an association between oligohydramnios and poor fetal outcomes. These were predominantly retrospective studies, which failed to control for the presence of factors known to be associated with oligohydramnios such as intrauterine growth restriction (IUGR) and urogenital malformations.

No studies have directly addressed whether labor induction improves outcomes. A meta-analysis of 18 studies examining outcomes of pregnancies with AFI <5 cm found an increased risk of cesarean delivery for fetal distress and low Apgar scores at 5 minutes. Most of these studies, however, had high-risk patients including IUGR (level of evidence [LOE]: 2).²¹

A recent study of high-risk patients failed to detect a difference in the incidence of nonreactive nonstress tests, meconium-stained amniotic fluid cesarean delivery for fetal distress, low Apgar scores, or infants with a cord pH of <7.10 when oligohydramnios (AFI <5.0 cm) was present (LOE: 1).² The patients with oligohydramnios were all induced, while many of the other high-risk patients were expectantly managed. The study therefore provides no guidance on the safety of expectant management for patients with oligohydramnios. To eliminate the potential effect of induction versus expectant management the same authors performed a case-control study of 79 high-risk women with AFI <5 cm matched to 79 women with the same high-risk pregnancy complication who had an AFI >5 cm at the time of induction (LOE: 2).²³ They failed to detect any significant differences in neonatal outcomes between the groups.

Studies of the “borderline” AFI (between 5 cm and 8 cm) may also demonstrate an association with adverse neonatal outcomes if researchers include fetuses with IUGR or malformations. In one retrospective case review of 214 women with AFI of 5 cm to 10 cm, the only statistically significant finding was an association with IUGR.³ The authors recommended antepartum surveillance twice a week for mothers with borderline AFI, but they did not comment on induction (LOE: 2). Correspondence regarding this study argued that this recommendation was not supported by the evidence and would lead to unnecessary antenatal testing.²⁴

Studies of isolated oligohydramnios

Investigators have conducted studies (Table 1) excluding fetuses with intrauterine growth restriction or anomalies to try to determine if isolated oligohydramnios is associated with poorer outcomes.^25-30

Rainford’s study of outcomes in exclusively term, low-risk patients failed to show significant outcome differences in Apgar scores, NICU admissions, or rates of cesarean delivery for non-reassuring fetal heart rate monitoring (LOE: 2).²⁹ This study was limited due to its retrospective design. The authors comment that the relatively good outcomes in the oligohydramnios group may be due to the widespread practice of inducing such patients.

In a case-control study by Conway, 183 low-risk, term parturients with oligohydramnios were matched to 183 women of similar gestational age and parity who presented in spontaneous labor. The patients with isolated oligohydramnios were induced and showed an increased cesarean delivery rate. The increased rate of cesarean delivery was not due to nonreassuring fetal surveillance and was attributed to the induction process (LOE: 2).²⁵

An analysis of woman diagnosed with isolated oligohydramnios (AFI <5) at any gestational age in the multicenter prospective RADIUS trial demonstrated similar perinatal outcomes and fetal growth compared with pregnancies with a normal amniotic fluid (LOE: 2).³⁰

The only randomized clinical trial of labor induction vs expectant management for term isolated oligohydramnios showed similar outcomes in each group. But this study was small (n=61) and has only been published as an abstract.³¹

TABLE 1
Isolated oligohydramnios and perinatal outcomes

Effect of maternal hydration

Maternal hydration status and plasma osmolality have an affect on amniotic fluid volume (Table 2). Maternal hydration with oral water or intravenous hypotonic solutions has been shown to increase amniotic fluid volume.^8,11-13 Oral hydration with hypotonic fluid has been demonstrated to increase fetal urine production in one observational study.³² Another observational study demonstrated increased amniotic fluid volume and uteroplacental perfusion without alteration of fetal urine production suggesting the possibility that transmembranous fluid shifts from the placenta to the amniotic cavity may be involved.¹²

Two small, randomized controlled trials (RCTs) demonstrated an increase in amniotic fluid volume in women with oligohydramnios after oral hydration.^11,13 Doi demonstrated significant increases in AFI in women with oligohydramnios beyond 35 weeks when given oral hydration with free water (increase of 3.8 cm ± 1.9; P<. 001) or hypotonic intravenous solution (increase of 2.8 cm ± 1.9; P<.001) (LOE: 3).¹¹ Interestingly, this study did not demonstrate an increase in amniotic fluid volume with intravenous hydration with isotonic fluid.

Kirkpatrick demonstrated a 30% increase in amniotic fluid compared with controls in women of unspecified gestational age with oligohydramnios given 2 liters of oral water 2 to 5 hours before repeat amniotic fluid index (LOE: 3).¹³

A randomized trial in women with normal amniotic fluid demonstrated a 16% increase in amniotic fluid index 4 to 6 hours after hydration with 2 liters of oral water, compared with an 8% decrease after fluid restriction during the same period.⁸

A recent study of daily oral hydration in women with amniotic fluid volume <10% percentile showed increased amniotic fluid volume at 1 week, suggesting long-term benefit, although the study lacked an appropriate control group (LOE: 3).³³

There are no studies of clinical outcomes such as fetal heart rate decelerations during labor, or neonatal outcomes. A Cochrane systematic review concluded that maternal hydration appears to increase amniotic fluid and may be beneficial in management of oligohydramnios; however, it recommended controlled trials to assess clinical outcome benefits (LOE: 3).³⁴

TABLE 2
Effect of hydration on amniotic fluid index

Management recommendations

The AFI has low specificity and positive predictive value for oligohydramnios, and there is scant evidence that isolated term oligohydramnios causes adverse fetal outcomes. We recommend that an AFI under 5 cm should prompt additional antenatal testing rather than immediate induction in low-risk term pregnancies (SOR: B).

Though we acknowledge the lack of high-quality studies with patient-oriented outcomes to support observation and maternal hydration, we have developed a management strategy that does not require immediate induction of labor in women with uncomplicated term pregnancies.

The following recommendations apply to women having oligohydramnios as defined by amniotic fluid volume of less than 5 cm and gestational age between 37 and 41 weeks.

Initial assessment

1. Assess for premature rupture of membranes with a thorough history and a sterile speculum exam
2. Reassess dating as oligohydramnios in post-dates pregnancy (>41 weeks) is an indication for induction (SOR: C)³⁵
3. Perform a nonstress test to assess fetal wellbeing
4. Assess for IUGR with an ultrasound for estimated fetal weight and for the ratio of head circumference (HC) to abdominal circumference (AC). A comparison with prior ultrasounds can aid in assessing interval growth. An estimated fetal weight below the 10%, an elevated HC/AC ratio, or poor interval growth would suggest IUGR
5. Arrange for an ultrasound anatomic survey for fetal anomalies, if not done previously
6. Determine if preeclampsia, chronic hypertension, diabetes, or other maternal conditions associated with uteroplacental insufficiency are present.

Action steps

With any positive findings in the initial evaluation, proceed to labor induction, as the patient does not have isolated, term oligohydramnios (SOR: C). If the initial assessment is unremarkable and the AFI is less than 5, consider hydration with oral water and repeating the AFI 2 to 6 hours later (SOR: B).

Persistent oligohydramnios at term, particularly with a ripe cervix, may lead you to consider labor induction. Continued expectant management of isolated term oligohydramnios with twice weekly fetal surveillance may also be a reasonable option due to the paucity of evidence that oligohydramnios is associated with an adverse outcome in this scenario (SOR: C). Normal results with umbilical artery Doppler flow studies have been used to decrease the need for induction in high-risk pregnancies with oligohydramnios, and this technique may eventually have a role in isolated term oligohydramnios.³⁶

It is essential that patients receive counseling and give informed consent regarding the risks and benefits of observation or induction for isolated term oligohydramnios. The ease of induction based on parity and cervical ripeness should be considered.

A primiparous woman with an unfavorable cervix who strongly desires a spontaneous, vaginal birth could be told that, although there may be a small risk for her baby, no study has demonstrated any increased long-term morbidity or mortality associated with low fluid in her situation and that labor induction may double her chance of cesarean delivery.^37,38 In such a situation, an acceptable approach for mother and clinician may be rehydration followed by a repeat AFI and close follow-up with testing for fetal well-being according to the algorithm (Figure). In a practical sense, rehydration with 2 liters of oral water for oligohydramnios may be done whether or not immediate induction is chosen, as this is a safe measure that has been shown to significantly increase AFI. Alternatively, the preferred approach for a multiparous woman with a ripe cervix by Bishop score may be labor induction.

As adverse fetal outcomes have not been demonstrated in women with isolated term oligohydramnios, there is no rationale for routinely inducing labor based on an isolated finding of a so called “borderline” amniotic fluid index in the 5-to-8 range. In this situation it is appropriate to perform the initial assessment described above and may be reasonable to repeat the amniotic fluid index in 3 to 4 days to determine if true oligohydramnios has developed.

FIGURE
Assessment of the pregnant woman with oligohydramnios at term

Acknowledgments

We appreciate the assistance of George Gilson MD, Lauren Plante MD, and William Rayburn MD in manuscript review.

Corresponding author
Lawrence Leeman, MD, MPH, University of New Mexico Depts of Family and Community Medicine, Obstetrics and Gynecology, 2400 Tucker NE, 3rd floor, Albuquerque, NM 87131. E-mail: [email protected].

Family physicians providing maternity care often face a scenario in which an otherwise low-risk, term patient is incidentally noted to have a low amniotic fluid index (AFI). Common reasons for obtaining an AFI in a woman with a low-risk pregnancy include evaluation of decreased fetal movement, spontaneous variable decelerations during monitoring to evaluate for labor, or an ultrasound evaluation for fundal height measurements discordant with gestational age. How should “isolated” oligohydramnios—an AFI <5 cm—be interpreted, and should immediate induction be recommended for such patients?

Oligohydramnios occurs in about 1% to 5% of pregnancies at term.^1,2 Because adverse outcomes occur in high-risk pregnancies complicated by low amniotic fluid volume, oligohydramnios commonly prompts labor induction.^1,3,4 At one university center, oligohydramnios is now the leading indication for labor induction.⁵ Many centers may even induce labor when the AFI is between 5 cm and 8 cm, the so-called borderline AFI.³

Labor induction increases the use of cesarean delivery, particularly for the primiparous woman with an unripe cervix.⁶ Recent studies questioning the safety of labor induction in women who have had a cesarean may increase the number of elective repeat cesarean procedures when delivery is believed indicated for oligohydramnios.⁷ (See Underlying causes of oligohydramnios.)

Oligohydramnios is difficult to assess

True oligohydramnios can be difficult to confirm due to the questionable accuracy of amniotic fluid measurement by ultrasound. There is controversy, for example, about whether (and how) to include pockets of amniotic fluid containing umbilical cord.¹⁵ The AFI was introduced in 1987² to replace the 2 cm “pocket technique” of fluid assessment, and studies continue to question to what extent the AFI reflects actual amniotic fluid volume.

AFI measurements may vary with the amount of pressure applied to the abdomen and with fetal position or movement.¹⁶

Serial measurements taken by the same ultrasound operator have been shown to differ from the true volume by 1 cm, or 10.8%; serial measurements taken by multiple operators have differed by as much as 2 cm, or 15.4%.^17,18

O’Reilly-Green compared the diagnosis of oligohydramnios in 449 post-term patients with actual amniotic fluid volume measured at rupture of membranes.¹⁹ They found a positive predictive value of 50% for oligohydramnios at an AFI of 5 cm as the lower limit of normal. A study of 144 third trimester patients using the dye-dilution technique found that, to achieve 95% confidence for ruling out oligohydramnios, a cutoff AFI of 30 cm would need to be used, a value consistent with polyhydramnios.²⁰

What is the association between oligohydramnios and poor fetal outcomes?

A number of studies over the past 15 years have shown an association between oligohydramnios and poor fetal outcomes. These were predominantly retrospective studies, which failed to control for the presence of factors known to be associated with oligohydramnios such as intrauterine growth restriction (IUGR) and urogenital malformations.

No studies have directly addressed whether labor induction improves outcomes. A meta-analysis of 18 studies examining outcomes of pregnancies with AFI <5 cm found an increased risk of cesarean delivery for fetal distress and low Apgar scores at 5 minutes. Most of these studies, however, had high-risk patients including IUGR (level of evidence [LOE]: 2).²¹

A recent study of high-risk patients failed to detect a difference in the incidence of nonreactive nonstress tests, meconium-stained amniotic fluid cesarean delivery for fetal distress, low Apgar scores, or infants with a cord pH of <7.10 when oligohydramnios (AFI <5.0 cm) was present (LOE: 1).² The patients with oligohydramnios were all induced, while many of the other high-risk patients were expectantly managed. The study therefore provides no guidance on the safety of expectant management for patients with oligohydramnios. To eliminate the potential effect of induction versus expectant management the same authors performed a case-control study of 79 high-risk women with AFI <5 cm matched to 79 women with the same high-risk pregnancy complication who had an AFI >5 cm at the time of induction (LOE: 2).²³ They failed to detect any significant differences in neonatal outcomes between the groups.

Studies of the “borderline” AFI (between 5 cm and 8 cm) may also demonstrate an association with adverse neonatal outcomes if researchers include fetuses with IUGR or malformations. In one retrospective case review of 214 women with AFI of 5 cm to 10 cm, the only statistically significant finding was an association with IUGR.³ The authors recommended antepartum surveillance twice a week for mothers with borderline AFI, but they did not comment on induction (LOE: 2). Correspondence regarding this study argued that this recommendation was not supported by the evidence and would lead to unnecessary antenatal testing.²⁴

Studies of isolated oligohydramnios

Investigators have conducted studies (Table 1) excluding fetuses with intrauterine growth restriction or anomalies to try to determine if isolated oligohydramnios is associated with poorer outcomes.^25-30

Rainford’s study of outcomes in exclusively term, low-risk patients failed to show significant outcome differences in Apgar scores, NICU admissions, or rates of cesarean delivery for non-reassuring fetal heart rate monitoring (LOE: 2).²⁹ This study was limited due to its retrospective design. The authors comment that the relatively good outcomes in the oligohydramnios group may be due to the widespread practice of inducing such patients.

In a case-control study by Conway, 183 low-risk, term parturients with oligohydramnios were matched to 183 women of similar gestational age and parity who presented in spontaneous labor. The patients with isolated oligohydramnios were induced and showed an increased cesarean delivery rate. The increased rate of cesarean delivery was not due to nonreassuring fetal surveillance and was attributed to the induction process (LOE: 2).²⁵

An analysis of woman diagnosed with isolated oligohydramnios (AFI <5) at any gestational age in the multicenter prospective RADIUS trial demonstrated similar perinatal outcomes and fetal growth compared with pregnancies with a normal amniotic fluid (LOE: 2).³⁰

The only randomized clinical trial of labor induction vs expectant management for term isolated oligohydramnios showed similar outcomes in each group. But this study was small (n=61) and has only been published as an abstract.³¹

TABLE 1
Isolated oligohydramnios and perinatal outcomes

Effect of maternal hydration

Maternal hydration status and plasma osmolality have an affect on amniotic fluid volume (Table 2). Maternal hydration with oral water or intravenous hypotonic solutions has been shown to increase amniotic fluid volume.^8,11-13 Oral hydration with hypotonic fluid has been demonstrated to increase fetal urine production in one observational study.³² Another observational study demonstrated increased amniotic fluid volume and uteroplacental perfusion without alteration of fetal urine production suggesting the possibility that transmembranous fluid shifts from the placenta to the amniotic cavity may be involved.¹²

Two small, randomized controlled trials (RCTs) demonstrated an increase in amniotic fluid volume in women with oligohydramnios after oral hydration.^11,13 Doi demonstrated significant increases in AFI in women with oligohydramnios beyond 35 weeks when given oral hydration with free water (increase of 3.8 cm ± 1.9; P<. 001) or hypotonic intravenous solution (increase of 2.8 cm ± 1.9; P<.001) (LOE: 3).¹¹ Interestingly, this study did not demonstrate an increase in amniotic fluid volume with intravenous hydration with isotonic fluid.

Kirkpatrick demonstrated a 30% increase in amniotic fluid compared with controls in women of unspecified gestational age with oligohydramnios given 2 liters of oral water 2 to 5 hours before repeat amniotic fluid index (LOE: 3).¹³

A randomized trial in women with normal amniotic fluid demonstrated a 16% increase in amniotic fluid index 4 to 6 hours after hydration with 2 liters of oral water, compared with an 8% decrease after fluid restriction during the same period.⁸

A recent study of daily oral hydration in women with amniotic fluid volume <10% percentile showed increased amniotic fluid volume at 1 week, suggesting long-term benefit, although the study lacked an appropriate control group (LOE: 3).³³

There are no studies of clinical outcomes such as fetal heart rate decelerations during labor, or neonatal outcomes. A Cochrane systematic review concluded that maternal hydration appears to increase amniotic fluid and may be beneficial in management of oligohydramnios; however, it recommended controlled trials to assess clinical outcome benefits (LOE: 3).³⁴

TABLE 2
Effect of hydration on amniotic fluid index

Management recommendations

The AFI has low specificity and positive predictive value for oligohydramnios, and there is scant evidence that isolated term oligohydramnios causes adverse fetal outcomes. We recommend that an AFI under 5 cm should prompt additional antenatal testing rather than immediate induction in low-risk term pregnancies (SOR: B).

Though we acknowledge the lack of high-quality studies with patient-oriented outcomes to support observation and maternal hydration, we have developed a management strategy that does not require immediate induction of labor in women with uncomplicated term pregnancies.

The following recommendations apply to women having oligohydramnios as defined by amniotic fluid volume of less than 5 cm and gestational age between 37 and 41 weeks.

Initial assessment

1. Assess for premature rupture of membranes with a thorough history and a sterile speculum exam
2. Reassess dating as oligohydramnios in post-dates pregnancy (>41 weeks) is an indication for induction (SOR: C)³⁵
3. Perform a nonstress test to assess fetal wellbeing
4. Assess for IUGR with an ultrasound for estimated fetal weight and for the ratio of head circumference (HC) to abdominal circumference (AC). A comparison with prior ultrasounds can aid in assessing interval growth. An estimated fetal weight below the 10%, an elevated HC/AC ratio, or poor interval growth would suggest IUGR
5. Arrange for an ultrasound anatomic survey for fetal anomalies, if not done previously
6. Determine if preeclampsia, chronic hypertension, diabetes, or other maternal conditions associated with uteroplacental insufficiency are present.

Action steps

With any positive findings in the initial evaluation, proceed to labor induction, as the patient does not have isolated, term oligohydramnios (SOR: C). If the initial assessment is unremarkable and the AFI is less than 5, consider hydration with oral water and repeating the AFI 2 to 6 hours later (SOR: B).

Persistent oligohydramnios at term, particularly with a ripe cervix, may lead you to consider labor induction. Continued expectant management of isolated term oligohydramnios with twice weekly fetal surveillance may also be a reasonable option due to the paucity of evidence that oligohydramnios is associated with an adverse outcome in this scenario (SOR: C). Normal results with umbilical artery Doppler flow studies have been used to decrease the need for induction in high-risk pregnancies with oligohydramnios, and this technique may eventually have a role in isolated term oligohydramnios.³⁶

It is essential that patients receive counseling and give informed consent regarding the risks and benefits of observation or induction for isolated term oligohydramnios. The ease of induction based on parity and cervical ripeness should be considered.

A primiparous woman with an unfavorable cervix who strongly desires a spontaneous, vaginal birth could be told that, although there may be a small risk for her baby, no study has demonstrated any increased long-term morbidity or mortality associated with low fluid in her situation and that labor induction may double her chance of cesarean delivery.^37,38 In such a situation, an acceptable approach for mother and clinician may be rehydration followed by a repeat AFI and close follow-up with testing for fetal well-being according to the algorithm (Figure). In a practical sense, rehydration with 2 liters of oral water for oligohydramnios may be done whether or not immediate induction is chosen, as this is a safe measure that has been shown to significantly increase AFI. Alternatively, the preferred approach for a multiparous woman with a ripe cervix by Bishop score may be labor induction.

As adverse fetal outcomes have not been demonstrated in women with isolated term oligohydramnios, there is no rationale for routinely inducing labor based on an isolated finding of a so called “borderline” amniotic fluid index in the 5-to-8 range. In this situation it is appropriate to perform the initial assessment described above and may be reasonable to repeat the amniotic fluid index in 3 to 4 days to determine if true oligohydramnios has developed.

FIGURE
Assessment of the pregnant woman with oligohydramnios at term

Acknowledgments

We appreciate the assistance of George Gilson MD, Lauren Plante MD, and William Rayburn MD in manuscript review.

Corresponding author
Lawrence Leeman, MD, MPH, University of New Mexico Depts of Family and Community Medicine, Obstetrics and Gynecology, 2400 Tucker NE, 3rd floor, Albuquerque, NM 87131. E-mail: [email protected].

Family physicians providing maternity care often face a scenario in which an otherwise low-risk, term patient is incidentally noted to have a low amniotic fluid index (AFI). Common reasons for obtaining an AFI in a woman with a low-risk pregnancy include evaluation of decreased fetal movement, spontaneous variable decelerations during monitoring to evaluate for labor, or an ultrasound evaluation for fundal height measurements discordant with gestational age. How should “isolated” oligohydramnios—an AFI <5 cm—be interpreted, and should immediate induction be recommended for such patients?

Oligohydramnios occurs in about 1% to 5% of pregnancies at term.^1,2 Because adverse outcomes occur in high-risk pregnancies complicated by low amniotic fluid volume, oligohydramnios commonly prompts labor induction.^1,3,4 At one university center, oligohydramnios is now the leading indication for labor induction.⁵ Many centers may even induce labor when the AFI is between 5 cm and 8 cm, the so-called borderline AFI.³

Labor induction increases the use of cesarean delivery, particularly for the primiparous woman with an unripe cervix.⁶ Recent studies questioning the safety of labor induction in women who have had a cesarean may increase the number of elective repeat cesarean procedures when delivery is believed indicated for oligohydramnios.⁷ (See Underlying causes of oligohydramnios.)

Oligohydramnios is difficult to assess

True oligohydramnios can be difficult to confirm due to the questionable accuracy of amniotic fluid measurement by ultrasound. There is controversy, for example, about whether (and how) to include pockets of amniotic fluid containing umbilical cord.¹⁵ The AFI was introduced in 1987² to replace the 2 cm “pocket technique” of fluid assessment, and studies continue to question to what extent the AFI reflects actual amniotic fluid volume.

AFI measurements may vary with the amount of pressure applied to the abdomen and with fetal position or movement.¹⁶

Serial measurements taken by the same ultrasound operator have been shown to differ from the true volume by 1 cm, or 10.8%; serial measurements taken by multiple operators have differed by as much as 2 cm, or 15.4%.^17,18

O’Reilly-Green compared the diagnosis of oligohydramnios in 449 post-term patients with actual amniotic fluid volume measured at rupture of membranes.¹⁹ They found a positive predictive value of 50% for oligohydramnios at an AFI of 5 cm as the lower limit of normal. A study of 144 third trimester patients using the dye-dilution technique found that, to achieve 95% confidence for ruling out oligohydramnios, a cutoff AFI of 30 cm would need to be used, a value consistent with polyhydramnios.²⁰

What is the association between oligohydramnios and poor fetal outcomes?

A number of studies over the past 15 years have shown an association between oligohydramnios and poor fetal outcomes. These were predominantly retrospective studies, which failed to control for the presence of factors known to be associated with oligohydramnios such as intrauterine growth restriction (IUGR) and urogenital malformations.

No studies have directly addressed whether labor induction improves outcomes. A meta-analysis of 18 studies examining outcomes of pregnancies with AFI <5 cm found an increased risk of cesarean delivery for fetal distress and low Apgar scores at 5 minutes. Most of these studies, however, had high-risk patients including IUGR (level of evidence [LOE]: 2).²¹

A recent study of high-risk patients failed to detect a difference in the incidence of nonreactive nonstress tests, meconium-stained amniotic fluid cesarean delivery for fetal distress, low Apgar scores, or infants with a cord pH of <7.10 when oligohydramnios (AFI <5.0 cm) was present (LOE: 1).² The patients with oligohydramnios were all induced, while many of the other high-risk patients were expectantly managed. The study therefore provides no guidance on the safety of expectant management for patients with oligohydramnios. To eliminate the potential effect of induction versus expectant management the same authors performed a case-control study of 79 high-risk women with AFI <5 cm matched to 79 women with the same high-risk pregnancy complication who had an AFI >5 cm at the time of induction (LOE: 2).²³ They failed to detect any significant differences in neonatal outcomes between the groups.

Studies of the “borderline” AFI (between 5 cm and 8 cm) may also demonstrate an association with adverse neonatal outcomes if researchers include fetuses with IUGR or malformations. In one retrospective case review of 214 women with AFI of 5 cm to 10 cm, the only statistically significant finding was an association with IUGR.³ The authors recommended antepartum surveillance twice a week for mothers with borderline AFI, but they did not comment on induction (LOE: 2). Correspondence regarding this study argued that this recommendation was not supported by the evidence and would lead to unnecessary antenatal testing.²⁴

Studies of isolated oligohydramnios

Investigators have conducted studies (Table 1) excluding fetuses with intrauterine growth restriction or anomalies to try to determine if isolated oligohydramnios is associated with poorer outcomes.^25-30

Rainford’s study of outcomes in exclusively term, low-risk patients failed to show significant outcome differences in Apgar scores, NICU admissions, or rates of cesarean delivery for non-reassuring fetal heart rate monitoring (LOE: 2).²⁹ This study was limited due to its retrospective design. The authors comment that the relatively good outcomes in the oligohydramnios group may be due to the widespread practice of inducing such patients.

In a case-control study by Conway, 183 low-risk, term parturients with oligohydramnios were matched to 183 women of similar gestational age and parity who presented in spontaneous labor. The patients with isolated oligohydramnios were induced and showed an increased cesarean delivery rate. The increased rate of cesarean delivery was not due to nonreassuring fetal surveillance and was attributed to the induction process (LOE: 2).²⁵

An analysis of woman diagnosed with isolated oligohydramnios (AFI <5) at any gestational age in the multicenter prospective RADIUS trial demonstrated similar perinatal outcomes and fetal growth compared with pregnancies with a normal amniotic fluid (LOE: 2).³⁰

The only randomized clinical trial of labor induction vs expectant management for term isolated oligohydramnios showed similar outcomes in each group. But this study was small (n=61) and has only been published as an abstract.³¹

TABLE 1
Isolated oligohydramnios and perinatal outcomes

Effect of maternal hydration

Maternal hydration status and plasma osmolality have an affect on amniotic fluid volume (Table 2). Maternal hydration with oral water or intravenous hypotonic solutions has been shown to increase amniotic fluid volume.^8,11-13 Oral hydration with hypotonic fluid has been demonstrated to increase fetal urine production in one observational study.³² Another observational study demonstrated increased amniotic fluid volume and uteroplacental perfusion without alteration of fetal urine production suggesting the possibility that transmembranous fluid shifts from the placenta to the amniotic cavity may be involved.¹²

Two small, randomized controlled trials (RCTs) demonstrated an increase in amniotic fluid volume in women with oligohydramnios after oral hydration.^11,13 Doi demonstrated significant increases in AFI in women with oligohydramnios beyond 35 weeks when given oral hydration with free water (increase of 3.8 cm ± 1.9; P<. 001) or hypotonic intravenous solution (increase of 2.8 cm ± 1.9; P<.001) (LOE: 3).¹¹ Interestingly, this study did not demonstrate an increase in amniotic fluid volume with intravenous hydration with isotonic fluid.

Kirkpatrick demonstrated a 30% increase in amniotic fluid compared with controls in women of unspecified gestational age with oligohydramnios given 2 liters of oral water 2 to 5 hours before repeat amniotic fluid index (LOE: 3).¹³

A randomized trial in women with normal amniotic fluid demonstrated a 16% increase in amniotic fluid index 4 to 6 hours after hydration with 2 liters of oral water, compared with an 8% decrease after fluid restriction during the same period.⁸

A recent study of daily oral hydration in women with amniotic fluid volume <10% percentile showed increased amniotic fluid volume at 1 week, suggesting long-term benefit, although the study lacked an appropriate control group (LOE: 3).³³

There are no studies of clinical outcomes such as fetal heart rate decelerations during labor, or neonatal outcomes. A Cochrane systematic review concluded that maternal hydration appears to increase amniotic fluid and may be beneficial in management of oligohydramnios; however, it recommended controlled trials to assess clinical outcome benefits (LOE: 3).³⁴

TABLE 2
Effect of hydration on amniotic fluid index

Management recommendations

The AFI has low specificity and positive predictive value for oligohydramnios, and there is scant evidence that isolated term oligohydramnios causes adverse fetal outcomes. We recommend that an AFI under 5 cm should prompt additional antenatal testing rather than immediate induction in low-risk term pregnancies (SOR: B).

Though we acknowledge the lack of high-quality studies with patient-oriented outcomes to support observation and maternal hydration, we have developed a management strategy that does not require immediate induction of labor in women with uncomplicated term pregnancies.

The following recommendations apply to women having oligohydramnios as defined by amniotic fluid volume of less than 5 cm and gestational age between 37 and 41 weeks.

Initial assessment

1. Assess for premature rupture of membranes with a thorough history and a sterile speculum exam
2. Reassess dating as oligohydramnios in post-dates pregnancy (>41 weeks) is an indication for induction (SOR: C)³⁵
3. Perform a nonstress test to assess fetal wellbeing
4. Assess for IUGR with an ultrasound for estimated fetal weight and for the ratio of head circumference (HC) to abdominal circumference (AC). A comparison with prior ultrasounds can aid in assessing interval growth. An estimated fetal weight below the 10%, an elevated HC/AC ratio, or poor interval growth would suggest IUGR
5. Arrange for an ultrasound anatomic survey for fetal anomalies, if not done previously
6. Determine if preeclampsia, chronic hypertension, diabetes, or other maternal conditions associated with uteroplacental insufficiency are present.

Action steps

With any positive findings in the initial evaluation, proceed to labor induction, as the patient does not have isolated, term oligohydramnios (SOR: C). If the initial assessment is unremarkable and the AFI is less than 5, consider hydration with oral water and repeating the AFI 2 to 6 hours later (SOR: B).

Persistent oligohydramnios at term, particularly with a ripe cervix, may lead you to consider labor induction. Continued expectant management of isolated term oligohydramnios with twice weekly fetal surveillance may also be a reasonable option due to the paucity of evidence that oligohydramnios is associated with an adverse outcome in this scenario (SOR: C). Normal results with umbilical artery Doppler flow studies have been used to decrease the need for induction in high-risk pregnancies with oligohydramnios, and this technique may eventually have a role in isolated term oligohydramnios.³⁶

It is essential that patients receive counseling and give informed consent regarding the risks and benefits of observation or induction for isolated term oligohydramnios. The ease of induction based on parity and cervical ripeness should be considered.

A primiparous woman with an unfavorable cervix who strongly desires a spontaneous, vaginal birth could be told that, although there may be a small risk for her baby, no study has demonstrated any increased long-term morbidity or mortality associated with low fluid in her situation and that labor induction may double her chance of cesarean delivery.^37,38 In such a situation, an acceptable approach for mother and clinician may be rehydration followed by a repeat AFI and close follow-up with testing for fetal well-being according to the algorithm (Figure). In a practical sense, rehydration with 2 liters of oral water for oligohydramnios may be done whether or not immediate induction is chosen, as this is a safe measure that has been shown to significantly increase AFI. Alternatively, the preferred approach for a multiparous woman with a ripe cervix by Bishop score may be labor induction.

As adverse fetal outcomes have not been demonstrated in women with isolated term oligohydramnios, there is no rationale for routinely inducing labor based on an isolated finding of a so called “borderline” amniotic fluid index in the 5-to-8 range. In this situation it is appropriate to perform the initial assessment described above and may be reasonable to repeat the amniotic fluid index in 3 to 4 days to determine if true oligohydramnios has developed.

FIGURE
Assessment of the pregnant woman with oligohydramnios at term

Acknowledgments

We appreciate the assistance of George Gilson MD, Lauren Plante MD, and William Rayburn MD in manuscript review.

Corresponding author
Lawrence Leeman, MD, MPH, University of New Mexico Depts of Family and Community Medicine, Obstetrics and Gynecology, 2400 Tucker NE, 3rd floor, Albuquerque, NM 87131. E-mail: [email protected].

The availability of perinatal care in rural communities produces better pregnancy outcomes than do perinatal systems that require rural women to seek maternity care in distant urban areas.^1-3 Unfortunately, rural maternity care has been affected by the loss of physicians who offer these services and by the closing of many rural hospitals’ maternity care units. Maintaining 24-hour operative obstetric capabilities is difficult in rural areas because they have an insufficient population base to support a physician trained in operative obstetrics. Another barrier is the lack of anesthesia services and operating room personnel.

The Guidelines for Perinatal Care developed by the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) state, “All hospitals offering labor and delivery services should be equipped to perform emergency cesarean delivery.” ⁴ Nevertheless, not all rural obstetric units can offer cesarean delivery and must transfer patients to a referral hospital for operative needs. Advisory panels in the United States and Canada have recommended similar models of rural perinatal care.^5-8 A Canadian panel estimated that 125 Canadian hospitals offer obstetric care without surgical facilities.

Studies of rural hospitals in Canada, Australia, and the United Kingdom that lack continuous on-site cesarean capability are limited by the small number of deliveries.^9-12 Most such studies are hospital based rather than population based and lack data on women who are transferred to outlying hospitals. The only population-based study that we identified found no evidence of adverse events caused by the lack of cesarean facilities; the sample size, however, was limited to 286 births.⁹

We studied all pregnancies occurring in a predominantly Native American region of New Mexico over a 5-year period to ascertain the safety of rural perinatal care based in a hospital without cesarean capability. Population-based and hospital-based outcomes are presented. This is the first study from a US community using this model of care.

Methods

We conducted an outcomes study using a historical cohort study design of all pregnancies beyond 20 weeks of gestation in the Zuni Pueblo and Ramah Navajo communities of northwestern New Mexico from 1992 to 1996. The perinatal services based at the Zuni-Ramah Indian Health Service (IHS) Hospital are the focus of this study. This 37-bed community hospital, staffed by family physicians and a part-time nurse-midwife, is part of an integrated perinatal system. The birthing unit has access to obstetrician-gynecologist (OBG) consultants at the Gallup Indian Medical Center (GIMC), 33 miles to the north, and perinatology and neonatology care in Albuquerque, 147 miles to the east. GIMC, the primary referral hospital and closest surgical facility, has an obstetric unit staffed by OBGs, family physicians, and nurse-mid-wives. Transportation time is 40 minutes by ground ambulance to GIMC or by fixed wing aircraft to Albuquerque.

The Zuni-Ramah Hospital limits intrapartum care to women designated as at low or moderate risk by criteria established by Zuni-Ramah family physicians and reviewed by GIMC OBGs. Criteria mandating transfer included prior cesarean, malpresentation, multiple gestation, intrauterine growth restriction, severe preeclampsia, placenta previa, significant vaginal bleeding, major fetal anomalies, anticipated preterm delivery (< 36 weeks), nonreassuring fetal heart tones (NRFHTs), and need for labor induction or augmentation with oxytocin. Women with gestational or type 2 diabetes who were well controlled could give birth at Zuni-Ramah unless they had end-organ damage or the fetus had known macrosomia. Physicians successfully completed the Advanced Life Support in Obstetrics (ALSO, ®American Academy of Family Physicians, 4th ed., 2000) course, attended weekly high-risk obstetric rounds, and performed quarterly reviews of obstetric complications. The family physicians performed vacuum-assisted deliveries, utilized amnioinfusion, and used continuous or intermittent fetal monitoring.

A review of the delivery and transfer records of the Zuni-Ramah Hospital and GIMC obstetric services revealed that there had been 1132 births of 1137 infants during the study period. The authors used a data collection form to review prenatal and newborn records from every birth. We reviewed intrapartum records for all births at the Zuni-Ramah and GIMC hospitals. We obtained discharge summaries from tertiary-care sites. We interviewed perinatal coordinators, public health nurses, and pediatric care providers to obtain information about patients who had received perinatal care outside of the IHS system.

The outcomes measured included perinatal mortality, neonatal morbidity, obstetric emergencies, intrapartum and antepartum transfers, and cesarean delivery rate. All obstetric emergencies originating at Zuni-Ramah Hospital were reviewed to determine whether the lack of surgical facilities had resulted in adverse outcomes. The physician’s notes were used to differentiate a NRFHT pattern requiring observation at a hospital with operative facilities from a truly worrisome pattern that required urgent intervention for fetal distress.

Births were defined as deliveries of infants at 20 weeks or more of estimated gestational age. We analyzed each birth in a multiple gestation individually. The population-based perinatal mortality rate was calculated from 20 weeks’ estimated gestational age to the 28th neonatal day. The Zuni-Ramah Hospital perinatal mortality rate was calculated by inclusion of all women delivered at Zuni-Ramah Hospital or transferred during labor. Approval for the study was obtained from the IHS Institutional Review Board and the Zuni Tribal Council.

Results

Study population

We identified 1137 births to 945 women between 1992 and 1996. Zuni and Navajo births were 66.9% and 30.8%, respectively; 30% of women were primiparous and 70%, multiparous. We found that 10.4% of births had occurred in women older than 35 years and 7.8% in women younger than 18 years. Regarding prenatal care, 3.9% of women had received none; 43.0% had established prenatal care in the first trimester; 40.4%, in the second trimester; and 12.8%, in the third trimester.

Delivery sites and maternal transfers

The majority of women (64.4%, n = 732) gave birth at the Zuni-Ramah Hospital (Figure) or at GIMC (29.6%, n = 337). A small number (2.2%, n = 25) gave birth at a private hospital with surgical facilities in Gallup. Albuquerque tertiary-care hospitals were the sites of 3.2% (n = 36) of deliveries. Primary indications for tertiary care were prematurity and fetal anomalies. Seven (0.6%) deliveries occurred at other sites, including home and ambulance.

The antepartum transfers (Table 1) were required primarily for pregnancy complications requiring labor induction. Preeclampsia, diabetes, nonreassuring antepartum testing, and post dates patients accounted for 56.8% of the 290 transfers. The 107 intrapartum transfers were made predominantly for labor induction or augmentation (64.5%, n = 69), a concerning fetal heart tracing (15.9%, n = 17), or fetal malpresentation diagnosed during labor (8.4%, n = 9).

FIGURE
PREGNANCIES AT ZUNI-RAMAH HOSPITAL

TABLE 1
ANTEPARTUM AND INTRAPARTUM TRANSFERS FROM ZUNI-RAMAH HOSPITAL

Obstetric interventions

The total cesarean delivery rate (7.3%) was approximately one third the nationwide rate of 20.7% in 1996. The primary cesarean delivery rate (number of cesareans in women without prior cesarean divided by the number of women who have never had a cesarean) of 5.3% compares with a nationwide primary rate of 14.6%. The cesarean rate was 22.1% for antepartum transfers and 17.8% for intrapartum transfers. Operative vaginal delivery occurred in 5.4% of births, well below the nationwide rate of 9.4%. The induction rate of 13.8% is lower than the nationwide rate of 16.9%. The oxytocin augmentation rate of 7.7% is well below the nationwide rate of 16.9% in 1996.¹³ Parenteral narcotics were available at Zuni-Ramah; however, 81.4% of women elected to receive no labor analgesia. Epidural anesthesia was not available at Zuni-Ramah Hospital.

Perinatal mortality

The perinatal mortality rate for the population was 11.4 per 1000 births (95% CI, 5.1-17.8 by Poisson distribution), comparable to the 1991 nationwide peri-natal mortality rate of 12.8/1000.¹⁴ Nine of the 13 neonatal deaths were caused by intrauterine fetal demise before labor (Table 2). The Zuni-Ramah Hospital–based perinatal mortality rate of 1.2/1000 was comparable with the 1.3/1000 perinatal mortality rate for women in the National Birth Center study even though Zuni-Ramah Hospital accepts higher-risk patients.¹⁵

TABLE 2
PERINATAL MORTALITY IN ZUNI-RAMAH POPULATION

Neonatal morbidity

Measures of neonatal morbidity are summarized in Table 3. The frequency of 5-minute Apgar scores below 7, low birthweight, and prematurity compares favorably with 1996 US rates.¹³ The rate of assisted ventilation (intubation or bag-mask) for the entire population (4.6%, n = 52) is greater than the 1996 nationwide rate (2.9%), although the difference is of questionable clinical significance, since international studies have demonstrated a range for assisted ventilation of 1% to 10% of hospital births.¹⁶ Neonatal Intensive Care Unit (NICU) transfer occurred in 27 (2.4%) of deliveries from non-tertiary-care sites. Thirteen (1.8%) babies born at Zuni-Ramah were transferred to Albuquerque for NICU care because of respiratory distress (n = 10) or neonatal anomalies (n = 3). The 3 cases of low Apgar scores at Zuni-Ramah were attributed to pneumothorax, respiratory distress syndrome of prematurity, and sepsis with meconium aspiration.

TABLE 3
NEONATAL MORBIDITY IN ZUNI-RAMAH POPULATION, BASED ON LIVE BIRTHS

Obstetric risk factors

The study population had a greater incidence of pregnancy-induced hypertension (14.5% vs 2.6% by 1996 ACOG criteria¹⁷), chronic hypertension (2.7% vs 0.7%¹⁵), and diabetes (9.2% vs 2.6%¹⁵) than the average US obstetric population. Gestational diabetes was diagnosed according to National Diabetes Data Group criteria:¹⁸ 7.1% had gestational diabetes (class A₁ and A₂ ) and 2.1% had type 2 antepartum diabetes (classes B and C).

Outcomes of obstetric emergencies at zuni-ramah hospital

We reviewed all cases of placental abruption, uterine inversion, umbilical cord prolapse, and fetal distress at Zuni-Ramah Hospital to identify potentially preventable adverse outcomes caused by lack of operative facilities (Table W1). Umbilical cord prolapse and uterine inversion each occurred once and were appropriately managed, with excellent outcomes. In 3 of the 4 cases of placental abruption, there were clearly no adverse outcomes caused by lack of on-site operative facilities, as patients were expectantly managed upon arrival to the referral hospital (cases 3 and 4) or presented to Zuni-Ramah Hospital as an intrauterine demise (case 5).

The fourth patient with placental abruption (case 6) presented at Zuni-Ramah with vaginal bleeding, severe variable decelerations, and a 10-point drop from baseline hematocrit. She was scheduled to labor at GIMC because of a history of prior cesarean but presented to the Zuni-Ramah emergency room with vaginal bleeding. She was transferred to GIMC for an anticipated cesarean delivery; however, on arrival the patient rapidly progressed and gave birth to an infant vaginally with Apgar scores of 3 and 9. Her infant had a neonatal seizure and magnetic resonance imaging evidence of sagittal sinus thrombosis. The infant had a normal neurologic evaluation, developmental assessment, and electroencephalogram at 15 months.

We reviewed 5 cases of urgent transfer for fetal distress. These were differentiated from the 8 intrapartum transfers for NRFHTs based on the severity of fetal heart monitor tracings. Four of the 5 women who had been transferred for fetal distress gave birth to healthy infants vaginally more than 2 hours after arrival at the referral institution. One patient who was urgently transferred for repetitive late decelerations is discussed below.

Cesarean deliveries for fetal distress at referral hospitals

We reviewed all cases of cesarean deliveries for fetal distress (n = 10) at referral institutions to determine whether outcomes for any of the patients could potentially have been improved by having their cesarean deliveries earlier if operative facilities had been available at Zuni-Ramah Hospital or by being transferred earlier (Table W1). Seven of the 10 patients were transferred for preeclampsia, NRFHTs, or failure to progress. All had their cesareans for fetal distress many hours after arrival at the referral institution.

Two cases of cesarean delivery for fetal distress after transfer because of abruption were previously described. A patient presented in early labor with repetitive late decelerations and was urgently transferred to GIMC, where she underwent an immediate cesarean delivery. Her infant had Apgar scores of 1 and 7, an unremarkable neonatal course, and a normal 15-month developmental screen.

Discussion

Our outcomes demonstrate that with the use of appropriate screening criteria, childbirth can safely occur in institutions that lack surgical suites. The population-based perinatal mortality rate was similar to the nationwide rate. A review of obstetric emergencies and low Apgar scores among the 839 women laboring at Zuni-Ramah Hospital failed to identify adverse outcomes that might have been prevented if the hospital had had operative facilities. Cesarean rates were approximately one third the nationwide rate even though Zuni-Ramah patients had a higher prevalence of such risk factors as diabetes and preeclampsia.

Although they represented a high-risk obstetric population, 65% of women were able to give birth at Zuni-Ramah Hospital through use of the perinatal screening criteria. The 35% rate of transfer was caused largely by the need for oxytocin augmentation or induction. Only 21.6% of the women who were transferred for dysfunctional labor or premature rupture of membranes ultimately had a cesarean delivery. Oxytocin has not been permitted at Zuni-Ramah Hospital because of the ACOG guideline permitting oxytocin use only if “a physician capable of performing a cesarean delivery is readily available.”¹⁹ There are no studies addressing the safety of labor induction or augmentation without on-site cesarean capability.

Canadian guidelines for rural maternity care do not prohibit the use of prostaglandins or oxytocin at hospitals without operative facilities. A Consensus Conference on Obstetric Services in Rural or Remote Communities addressed the issue of labor induction or augmentation in hospitals without cesarean capability by stating, “If caring for a woman in labour is appropriate in the community, then caring for her during an augmented/induced labour is equally appropriate when there is support by trained local staff and resources.”²⁰ We concur that use of oxytocin in rural hospital units without operative facilities should be considered under well-defined clinical guidelines or research protocols.

Limitations

Our study’s limitations include lack of long-term neonatal outcomes, small size of the Zuni-Ramah population, an almost exclusively Native American population, and lack of examiner blinding during record review. Transfer rates may be increased in populations with higher rates of cesarean delivery or epidural anesthesia use. Alternatively, the high incidence of preeclampsia, chronic hypertension, and diabetes in these communities may have resulted in a higher proportion of induction. Umbilical cord pro-lapse and significant placental abruption are routinely treated by urgent cesarean delivery; therefore, obstetric literature on outcomes without immediate operative intervention is limited.^21,22 A larger study would be required to determine the potential increased neonatal morbidity or mortality resulting from delayed intervention.

Conclusions

The ACOG/AAP guideline requiring on-site surgical facilities and the ability to initiate a cesarean in 30 minutes is not based on evidence. Four small retrospective studies of emergency cesarean deliveries delayed for more than 30 minutes did not demonstrate adverse neonatal outcomes.^23-26 In our study population, no adverse outcomes (none in 839 births) were determined to have been caused by a lack of surgical facilities. Despite these excellent outcomes, the possibility always exists for outcomes that can be prevented by doing a rapid emergent cesarean delivery. Women deciding to give birth in facilities without operative capabilities should receive information regarding the risks and benefits of delivering there and should have access to other facilities. Provider discretion and patient choice must be respected to ensure community support of these birthing units. Practitioners at the rural units must have assurance that any patients who require an urgent transfer will be readily accepted.

Rural communities, medical providers, and health care facilities need to consider the overall effect of maintaining local maternity care units, as the loss of rural maternity care can increase the risk of adverse perinatal outcomes.^1-3 We concur with the Canadian panel that although maintenance of rural surgical and anesthesia capabilities is desirable, “good outcomes can be sustained within an integrated risk management system without local access to operative delivery.”⁸ Guidelines should be developed to permit rural hospitals without cesarean capability to provide maternity care as part of integrated perinatal systems with well-developed transport protocols and supportive referral institutions. Women living in rural areas should have the option to give birth near their homes in such units if they so desire.

Acknowledgments

The authors thank Robert Rhyne, MD, for editorial assistance in manuscript preparation and Betty Skipper, PhD, for statistical assistance. Current Zuni-Ramah Obstetric Guidelines are available at http://hsc.unm.edu/fcm/research/zuni.

The availability of perinatal care in rural communities produces better pregnancy outcomes than do perinatal systems that require rural women to seek maternity care in distant urban areas.^1-3 Unfortunately, rural maternity care has been affected by the loss of physicians who offer these services and by the closing of many rural hospitals’ maternity care units. Maintaining 24-hour operative obstetric capabilities is difficult in rural areas because they have an insufficient population base to support a physician trained in operative obstetrics. Another barrier is the lack of anesthesia services and operating room personnel.

The Guidelines for Perinatal Care developed by the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) state, “All hospitals offering labor and delivery services should be equipped to perform emergency cesarean delivery.” ⁴ Nevertheless, not all rural obstetric units can offer cesarean delivery and must transfer patients to a referral hospital for operative needs. Advisory panels in the United States and Canada have recommended similar models of rural perinatal care.^5-8 A Canadian panel estimated that 125 Canadian hospitals offer obstetric care without surgical facilities.

Studies of rural hospitals in Canada, Australia, and the United Kingdom that lack continuous on-site cesarean capability are limited by the small number of deliveries.^9-12 Most such studies are hospital based rather than population based and lack data on women who are transferred to outlying hospitals. The only population-based study that we identified found no evidence of adverse events caused by the lack of cesarean facilities; the sample size, however, was limited to 286 births.⁹

We studied all pregnancies occurring in a predominantly Native American region of New Mexico over a 5-year period to ascertain the safety of rural perinatal care based in a hospital without cesarean capability. Population-based and hospital-based outcomes are presented. This is the first study from a US community using this model of care.

Methods

We conducted an outcomes study using a historical cohort study design of all pregnancies beyond 20 weeks of gestation in the Zuni Pueblo and Ramah Navajo communities of northwestern New Mexico from 1992 to 1996. The perinatal services based at the Zuni-Ramah Indian Health Service (IHS) Hospital are the focus of this study. This 37-bed community hospital, staffed by family physicians and a part-time nurse-midwife, is part of an integrated perinatal system. The birthing unit has access to obstetrician-gynecologist (OBG) consultants at the Gallup Indian Medical Center (GIMC), 33 miles to the north, and perinatology and neonatology care in Albuquerque, 147 miles to the east. GIMC, the primary referral hospital and closest surgical facility, has an obstetric unit staffed by OBGs, family physicians, and nurse-mid-wives. Transportation time is 40 minutes by ground ambulance to GIMC or by fixed wing aircraft to Albuquerque.

The Zuni-Ramah Hospital limits intrapartum care to women designated as at low or moderate risk by criteria established by Zuni-Ramah family physicians and reviewed by GIMC OBGs. Criteria mandating transfer included prior cesarean, malpresentation, multiple gestation, intrauterine growth restriction, severe preeclampsia, placenta previa, significant vaginal bleeding, major fetal anomalies, anticipated preterm delivery (< 36 weeks), nonreassuring fetal heart tones (NRFHTs), and need for labor induction or augmentation with oxytocin. Women with gestational or type 2 diabetes who were well controlled could give birth at Zuni-Ramah unless they had end-organ damage or the fetus had known macrosomia. Physicians successfully completed the Advanced Life Support in Obstetrics (ALSO, ®American Academy of Family Physicians, 4th ed., 2000) course, attended weekly high-risk obstetric rounds, and performed quarterly reviews of obstetric complications. The family physicians performed vacuum-assisted deliveries, utilized amnioinfusion, and used continuous or intermittent fetal monitoring.

A review of the delivery and transfer records of the Zuni-Ramah Hospital and GIMC obstetric services revealed that there had been 1132 births of 1137 infants during the study period. The authors used a data collection form to review prenatal and newborn records from every birth. We reviewed intrapartum records for all births at the Zuni-Ramah and GIMC hospitals. We obtained discharge summaries from tertiary-care sites. We interviewed perinatal coordinators, public health nurses, and pediatric care providers to obtain information about patients who had received perinatal care outside of the IHS system.

The outcomes measured included perinatal mortality, neonatal morbidity, obstetric emergencies, intrapartum and antepartum transfers, and cesarean delivery rate. All obstetric emergencies originating at Zuni-Ramah Hospital were reviewed to determine whether the lack of surgical facilities had resulted in adverse outcomes. The physician’s notes were used to differentiate a NRFHT pattern requiring observation at a hospital with operative facilities from a truly worrisome pattern that required urgent intervention for fetal distress.

Births were defined as deliveries of infants at 20 weeks or more of estimated gestational age. We analyzed each birth in a multiple gestation individually. The population-based perinatal mortality rate was calculated from 20 weeks’ estimated gestational age to the 28th neonatal day. The Zuni-Ramah Hospital perinatal mortality rate was calculated by inclusion of all women delivered at Zuni-Ramah Hospital or transferred during labor. Approval for the study was obtained from the IHS Institutional Review Board and the Zuni Tribal Council.

Results

Study population

We identified 1137 births to 945 women between 1992 and 1996. Zuni and Navajo births were 66.9% and 30.8%, respectively; 30% of women were primiparous and 70%, multiparous. We found that 10.4% of births had occurred in women older than 35 years and 7.8% in women younger than 18 years. Regarding prenatal care, 3.9% of women had received none; 43.0% had established prenatal care in the first trimester; 40.4%, in the second trimester; and 12.8%, in the third trimester.

Delivery sites and maternal transfers

The majority of women (64.4%, n = 732) gave birth at the Zuni-Ramah Hospital (Figure) or at GIMC (29.6%, n = 337). A small number (2.2%, n = 25) gave birth at a private hospital with surgical facilities in Gallup. Albuquerque tertiary-care hospitals were the sites of 3.2% (n = 36) of deliveries. Primary indications for tertiary care were prematurity and fetal anomalies. Seven (0.6%) deliveries occurred at other sites, including home and ambulance.

The antepartum transfers (Table 1) were required primarily for pregnancy complications requiring labor induction. Preeclampsia, diabetes, nonreassuring antepartum testing, and post dates patients accounted for 56.8% of the 290 transfers. The 107 intrapartum transfers were made predominantly for labor induction or augmentation (64.5%, n = 69), a concerning fetal heart tracing (15.9%, n = 17), or fetal malpresentation diagnosed during labor (8.4%, n = 9).

FIGURE
PREGNANCIES AT ZUNI-RAMAH HOSPITAL

TABLE 1
ANTEPARTUM AND INTRAPARTUM TRANSFERS FROM ZUNI-RAMAH HOSPITAL

Obstetric interventions

The total cesarean delivery rate (7.3%) was approximately one third the nationwide rate of 20.7% in 1996. The primary cesarean delivery rate (number of cesareans in women without prior cesarean divided by the number of women who have never had a cesarean) of 5.3% compares with a nationwide primary rate of 14.6%. The cesarean rate was 22.1% for antepartum transfers and 17.8% for intrapartum transfers. Operative vaginal delivery occurred in 5.4% of births, well below the nationwide rate of 9.4%. The induction rate of 13.8% is lower than the nationwide rate of 16.9%. The oxytocin augmentation rate of 7.7% is well below the nationwide rate of 16.9% in 1996.¹³ Parenteral narcotics were available at Zuni-Ramah; however, 81.4% of women elected to receive no labor analgesia. Epidural anesthesia was not available at Zuni-Ramah Hospital.

Perinatal mortality

The perinatal mortality rate for the population was 11.4 per 1000 births (95% CI, 5.1-17.8 by Poisson distribution), comparable to the 1991 nationwide peri-natal mortality rate of 12.8/1000.¹⁴ Nine of the 13 neonatal deaths were caused by intrauterine fetal demise before labor (Table 2). The Zuni-Ramah Hospital–based perinatal mortality rate of 1.2/1000 was comparable with the 1.3/1000 perinatal mortality rate for women in the National Birth Center study even though Zuni-Ramah Hospital accepts higher-risk patients.¹⁵

TABLE 2
PERINATAL MORTALITY IN ZUNI-RAMAH POPULATION

Neonatal morbidity

Measures of neonatal morbidity are summarized in Table 3. The frequency of 5-minute Apgar scores below 7, low birthweight, and prematurity compares favorably with 1996 US rates.¹³ The rate of assisted ventilation (intubation or bag-mask) for the entire population (4.6%, n = 52) is greater than the 1996 nationwide rate (2.9%), although the difference is of questionable clinical significance, since international studies have demonstrated a range for assisted ventilation of 1% to 10% of hospital births.¹⁶ Neonatal Intensive Care Unit (NICU) transfer occurred in 27 (2.4%) of deliveries from non-tertiary-care sites. Thirteen (1.8%) babies born at Zuni-Ramah were transferred to Albuquerque for NICU care because of respiratory distress (n = 10) or neonatal anomalies (n = 3). The 3 cases of low Apgar scores at Zuni-Ramah were attributed to pneumothorax, respiratory distress syndrome of prematurity, and sepsis with meconium aspiration.

TABLE 3
NEONATAL MORBIDITY IN ZUNI-RAMAH POPULATION, BASED ON LIVE BIRTHS

Obstetric risk factors

The study population had a greater incidence of pregnancy-induced hypertension (14.5% vs 2.6% by 1996 ACOG criteria¹⁷), chronic hypertension (2.7% vs 0.7%¹⁵), and diabetes (9.2% vs 2.6%¹⁵) than the average US obstetric population. Gestational diabetes was diagnosed according to National Diabetes Data Group criteria:¹⁸ 7.1% had gestational diabetes (class A₁ and A₂ ) and 2.1% had type 2 antepartum diabetes (classes B and C).

Outcomes of obstetric emergencies at zuni-ramah hospital

We reviewed all cases of placental abruption, uterine inversion, umbilical cord prolapse, and fetal distress at Zuni-Ramah Hospital to identify potentially preventable adverse outcomes caused by lack of operative facilities (Table W1). Umbilical cord prolapse and uterine inversion each occurred once and were appropriately managed, with excellent outcomes. In 3 of the 4 cases of placental abruption, there were clearly no adverse outcomes caused by lack of on-site operative facilities, as patients were expectantly managed upon arrival to the referral hospital (cases 3 and 4) or presented to Zuni-Ramah Hospital as an intrauterine demise (case 5).

The fourth patient with placental abruption (case 6) presented at Zuni-Ramah with vaginal bleeding, severe variable decelerations, and a 10-point drop from baseline hematocrit. She was scheduled to labor at GIMC because of a history of prior cesarean but presented to the Zuni-Ramah emergency room with vaginal bleeding. She was transferred to GIMC for an anticipated cesarean delivery; however, on arrival the patient rapidly progressed and gave birth to an infant vaginally with Apgar scores of 3 and 9. Her infant had a neonatal seizure and magnetic resonance imaging evidence of sagittal sinus thrombosis. The infant had a normal neurologic evaluation, developmental assessment, and electroencephalogram at 15 months.

We reviewed 5 cases of urgent transfer for fetal distress. These were differentiated from the 8 intrapartum transfers for NRFHTs based on the severity of fetal heart monitor tracings. Four of the 5 women who had been transferred for fetal distress gave birth to healthy infants vaginally more than 2 hours after arrival at the referral institution. One patient who was urgently transferred for repetitive late decelerations is discussed below.

Cesarean deliveries for fetal distress at referral hospitals

We reviewed all cases of cesarean deliveries for fetal distress (n = 10) at referral institutions to determine whether outcomes for any of the patients could potentially have been improved by having their cesarean deliveries earlier if operative facilities had been available at Zuni-Ramah Hospital or by being transferred earlier (Table W1). Seven of the 10 patients were transferred for preeclampsia, NRFHTs, or failure to progress. All had their cesareans for fetal distress many hours after arrival at the referral institution.

Two cases of cesarean delivery for fetal distress after transfer because of abruption were previously described. A patient presented in early labor with repetitive late decelerations and was urgently transferred to GIMC, where she underwent an immediate cesarean delivery. Her infant had Apgar scores of 1 and 7, an unremarkable neonatal course, and a normal 15-month developmental screen.

Discussion

Our outcomes demonstrate that with the use of appropriate screening criteria, childbirth can safely occur in institutions that lack surgical suites. The population-based perinatal mortality rate was similar to the nationwide rate. A review of obstetric emergencies and low Apgar scores among the 839 women laboring at Zuni-Ramah Hospital failed to identify adverse outcomes that might have been prevented if the hospital had had operative facilities. Cesarean rates were approximately one third the nationwide rate even though Zuni-Ramah patients had a higher prevalence of such risk factors as diabetes and preeclampsia.

Although they represented a high-risk obstetric population, 65% of women were able to give birth at Zuni-Ramah Hospital through use of the perinatal screening criteria. The 35% rate of transfer was caused largely by the need for oxytocin augmentation or induction. Only 21.6% of the women who were transferred for dysfunctional labor or premature rupture of membranes ultimately had a cesarean delivery. Oxytocin has not been permitted at Zuni-Ramah Hospital because of the ACOG guideline permitting oxytocin use only if “a physician capable of performing a cesarean delivery is readily available.”¹⁹ There are no studies addressing the safety of labor induction or augmentation without on-site cesarean capability.

Canadian guidelines for rural maternity care do not prohibit the use of prostaglandins or oxytocin at hospitals without operative facilities. A Consensus Conference on Obstetric Services in Rural or Remote Communities addressed the issue of labor induction or augmentation in hospitals without cesarean capability by stating, “If caring for a woman in labour is appropriate in the community, then caring for her during an augmented/induced labour is equally appropriate when there is support by trained local staff and resources.”²⁰ We concur that use of oxytocin in rural hospital units without operative facilities should be considered under well-defined clinical guidelines or research protocols.

Limitations

Our study’s limitations include lack of long-term neonatal outcomes, small size of the Zuni-Ramah population, an almost exclusively Native American population, and lack of examiner blinding during record review. Transfer rates may be increased in populations with higher rates of cesarean delivery or epidural anesthesia use. Alternatively, the high incidence of preeclampsia, chronic hypertension, and diabetes in these communities may have resulted in a higher proportion of induction. Umbilical cord pro-lapse and significant placental abruption are routinely treated by urgent cesarean delivery; therefore, obstetric literature on outcomes without immediate operative intervention is limited.^21,22 A larger study would be required to determine the potential increased neonatal morbidity or mortality resulting from delayed intervention.

Conclusions

The ACOG/AAP guideline requiring on-site surgical facilities and the ability to initiate a cesarean in 30 minutes is not based on evidence. Four small retrospective studies of emergency cesarean deliveries delayed for more than 30 minutes did not demonstrate adverse neonatal outcomes.^23-26 In our study population, no adverse outcomes (none in 839 births) were determined to have been caused by a lack of surgical facilities. Despite these excellent outcomes, the possibility always exists for outcomes that can be prevented by doing a rapid emergent cesarean delivery. Women deciding to give birth in facilities without operative capabilities should receive information regarding the risks and benefits of delivering there and should have access to other facilities. Provider discretion and patient choice must be respected to ensure community support of these birthing units. Practitioners at the rural units must have assurance that any patients who require an urgent transfer will be readily accepted.

Rural communities, medical providers, and health care facilities need to consider the overall effect of maintaining local maternity care units, as the loss of rural maternity care can increase the risk of adverse perinatal outcomes.^1-3 We concur with the Canadian panel that although maintenance of rural surgical and anesthesia capabilities is desirable, “good outcomes can be sustained within an integrated risk management system without local access to operative delivery.”⁸ Guidelines should be developed to permit rural hospitals without cesarean capability to provide maternity care as part of integrated perinatal systems with well-developed transport protocols and supportive referral institutions. Women living in rural areas should have the option to give birth near their homes in such units if they so desire.

Acknowledgments

The authors thank Robert Rhyne, MD, for editorial assistance in manuscript preparation and Betty Skipper, PhD, for statistical assistance. Current Zuni-Ramah Obstetric Guidelines are available at http://hsc.unm.edu/fcm/research/zuni.

The availability of perinatal care in rural communities produces better pregnancy outcomes than do perinatal systems that require rural women to seek maternity care in distant urban areas.^1-3 Unfortunately, rural maternity care has been affected by the loss of physicians who offer these services and by the closing of many rural hospitals’ maternity care units. Maintaining 24-hour operative obstetric capabilities is difficult in rural areas because they have an insufficient population base to support a physician trained in operative obstetrics. Another barrier is the lack of anesthesia services and operating room personnel.

The Guidelines for Perinatal Care developed by the American College of Obstetricians and Gynecologists (ACOG) and the American Academy of Pediatrics (AAP) state, “All hospitals offering labor and delivery services should be equipped to perform emergency cesarean delivery.” ⁴ Nevertheless, not all rural obstetric units can offer cesarean delivery and must transfer patients to a referral hospital for operative needs. Advisory panels in the United States and Canada have recommended similar models of rural perinatal care.^5-8 A Canadian panel estimated that 125 Canadian hospitals offer obstetric care without surgical facilities.

Studies of rural hospitals in Canada, Australia, and the United Kingdom that lack continuous on-site cesarean capability are limited by the small number of deliveries.^9-12 Most such studies are hospital based rather than population based and lack data on women who are transferred to outlying hospitals. The only population-based study that we identified found no evidence of adverse events caused by the lack of cesarean facilities; the sample size, however, was limited to 286 births.⁹

We studied all pregnancies occurring in a predominantly Native American region of New Mexico over a 5-year period to ascertain the safety of rural perinatal care based in a hospital without cesarean capability. Population-based and hospital-based outcomes are presented. This is the first study from a US community using this model of care.

Methods

We conducted an outcomes study using a historical cohort study design of all pregnancies beyond 20 weeks of gestation in the Zuni Pueblo and Ramah Navajo communities of northwestern New Mexico from 1992 to 1996. The perinatal services based at the Zuni-Ramah Indian Health Service (IHS) Hospital are the focus of this study. This 37-bed community hospital, staffed by family physicians and a part-time nurse-midwife, is part of an integrated perinatal system. The birthing unit has access to obstetrician-gynecologist (OBG) consultants at the Gallup Indian Medical Center (GIMC), 33 miles to the north, and perinatology and neonatology care in Albuquerque, 147 miles to the east. GIMC, the primary referral hospital and closest surgical facility, has an obstetric unit staffed by OBGs, family physicians, and nurse-mid-wives. Transportation time is 40 minutes by ground ambulance to GIMC or by fixed wing aircraft to Albuquerque.

The Zuni-Ramah Hospital limits intrapartum care to women designated as at low or moderate risk by criteria established by Zuni-Ramah family physicians and reviewed by GIMC OBGs. Criteria mandating transfer included prior cesarean, malpresentation, multiple gestation, intrauterine growth restriction, severe preeclampsia, placenta previa, significant vaginal bleeding, major fetal anomalies, anticipated preterm delivery (< 36 weeks), nonreassuring fetal heart tones (NRFHTs), and need for labor induction or augmentation with oxytocin. Women with gestational or type 2 diabetes who were well controlled could give birth at Zuni-Ramah unless they had end-organ damage or the fetus had known macrosomia. Physicians successfully completed the Advanced Life Support in Obstetrics (ALSO, ®American Academy of Family Physicians, 4th ed., 2000) course, attended weekly high-risk obstetric rounds, and performed quarterly reviews of obstetric complications. The family physicians performed vacuum-assisted deliveries, utilized amnioinfusion, and used continuous or intermittent fetal monitoring.

A review of the delivery and transfer records of the Zuni-Ramah Hospital and GIMC obstetric services revealed that there had been 1132 births of 1137 infants during the study period. The authors used a data collection form to review prenatal and newborn records from every birth. We reviewed intrapartum records for all births at the Zuni-Ramah and GIMC hospitals. We obtained discharge summaries from tertiary-care sites. We interviewed perinatal coordinators, public health nurses, and pediatric care providers to obtain information about patients who had received perinatal care outside of the IHS system.

The outcomes measured included perinatal mortality, neonatal morbidity, obstetric emergencies, intrapartum and antepartum transfers, and cesarean delivery rate. All obstetric emergencies originating at Zuni-Ramah Hospital were reviewed to determine whether the lack of surgical facilities had resulted in adverse outcomes. The physician’s notes were used to differentiate a NRFHT pattern requiring observation at a hospital with operative facilities from a truly worrisome pattern that required urgent intervention for fetal distress.

Births were defined as deliveries of infants at 20 weeks or more of estimated gestational age. We analyzed each birth in a multiple gestation individually. The population-based perinatal mortality rate was calculated from 20 weeks’ estimated gestational age to the 28th neonatal day. The Zuni-Ramah Hospital perinatal mortality rate was calculated by inclusion of all women delivered at Zuni-Ramah Hospital or transferred during labor. Approval for the study was obtained from the IHS Institutional Review Board and the Zuni Tribal Council.

Results

Study population

We identified 1137 births to 945 women between 1992 and 1996. Zuni and Navajo births were 66.9% and 30.8%, respectively; 30% of women were primiparous and 70%, multiparous. We found that 10.4% of births had occurred in women older than 35 years and 7.8% in women younger than 18 years. Regarding prenatal care, 3.9% of women had received none; 43.0% had established prenatal care in the first trimester; 40.4%, in the second trimester; and 12.8%, in the third trimester.

Delivery sites and maternal transfers

The majority of women (64.4%, n = 732) gave birth at the Zuni-Ramah Hospital (Figure) or at GIMC (29.6%, n = 337). A small number (2.2%, n = 25) gave birth at a private hospital with surgical facilities in Gallup. Albuquerque tertiary-care hospitals were the sites of 3.2% (n = 36) of deliveries. Primary indications for tertiary care were prematurity and fetal anomalies. Seven (0.6%) deliveries occurred at other sites, including home and ambulance.

The antepartum transfers (Table 1) were required primarily for pregnancy complications requiring labor induction. Preeclampsia, diabetes, nonreassuring antepartum testing, and post dates patients accounted for 56.8% of the 290 transfers. The 107 intrapartum transfers were made predominantly for labor induction or augmentation (64.5%, n = 69), a concerning fetal heart tracing (15.9%, n = 17), or fetal malpresentation diagnosed during labor (8.4%, n = 9).

FIGURE
PREGNANCIES AT ZUNI-RAMAH HOSPITAL

TABLE 1
ANTEPARTUM AND INTRAPARTUM TRANSFERS FROM ZUNI-RAMAH HOSPITAL

Obstetric interventions

The total cesarean delivery rate (7.3%) was approximately one third the nationwide rate of 20.7% in 1996. The primary cesarean delivery rate (number of cesareans in women without prior cesarean divided by the number of women who have never had a cesarean) of 5.3% compares with a nationwide primary rate of 14.6%. The cesarean rate was 22.1% for antepartum transfers and 17.8% for intrapartum transfers. Operative vaginal delivery occurred in 5.4% of births, well below the nationwide rate of 9.4%. The induction rate of 13.8% is lower than the nationwide rate of 16.9%. The oxytocin augmentation rate of 7.7% is well below the nationwide rate of 16.9% in 1996.¹³ Parenteral narcotics were available at Zuni-Ramah; however, 81.4% of women elected to receive no labor analgesia. Epidural anesthesia was not available at Zuni-Ramah Hospital.

Perinatal mortality

The perinatal mortality rate for the population was 11.4 per 1000 births (95% CI, 5.1-17.8 by Poisson distribution), comparable to the 1991 nationwide peri-natal mortality rate of 12.8/1000.¹⁴ Nine of the 13 neonatal deaths were caused by intrauterine fetal demise before labor (Table 2). The Zuni-Ramah Hospital–based perinatal mortality rate of 1.2/1000 was comparable with the 1.3/1000 perinatal mortality rate for women in the National Birth Center study even though Zuni-Ramah Hospital accepts higher-risk patients.¹⁵

TABLE 2
PERINATAL MORTALITY IN ZUNI-RAMAH POPULATION

Neonatal morbidity

Measures of neonatal morbidity are summarized in Table 3. The frequency of 5-minute Apgar scores below 7, low birthweight, and prematurity compares favorably with 1996 US rates.¹³ The rate of assisted ventilation (intubation or bag-mask) for the entire population (4.6%, n = 52) is greater than the 1996 nationwide rate (2.9%), although the difference is of questionable clinical significance, since international studies have demonstrated a range for assisted ventilation of 1% to 10% of hospital births.¹⁶ Neonatal Intensive Care Unit (NICU) transfer occurred in 27 (2.4%) of deliveries from non-tertiary-care sites. Thirteen (1.8%) babies born at Zuni-Ramah were transferred to Albuquerque for NICU care because of respiratory distress (n = 10) or neonatal anomalies (n = 3). The 3 cases of low Apgar scores at Zuni-Ramah were attributed to pneumothorax, respiratory distress syndrome of prematurity, and sepsis with meconium aspiration.

TABLE 3
NEONATAL MORBIDITY IN ZUNI-RAMAH POPULATION, BASED ON LIVE BIRTHS

Obstetric risk factors

The study population had a greater incidence of pregnancy-induced hypertension (14.5% vs 2.6% by 1996 ACOG criteria¹⁷), chronic hypertension (2.7% vs 0.7%¹⁵), and diabetes (9.2% vs 2.6%¹⁵) than the average US obstetric population. Gestational diabetes was diagnosed according to National Diabetes Data Group criteria:¹⁸ 7.1% had gestational diabetes (class A₁ and A₂ ) and 2.1% had type 2 antepartum diabetes (classes B and C).

Outcomes of obstetric emergencies at zuni-ramah hospital

We reviewed all cases of placental abruption, uterine inversion, umbilical cord prolapse, and fetal distress at Zuni-Ramah Hospital to identify potentially preventable adverse outcomes caused by lack of operative facilities (Table W1). Umbilical cord prolapse and uterine inversion each occurred once and were appropriately managed, with excellent outcomes. In 3 of the 4 cases of placental abruption, there were clearly no adverse outcomes caused by lack of on-site operative facilities, as patients were expectantly managed upon arrival to the referral hospital (cases 3 and 4) or presented to Zuni-Ramah Hospital as an intrauterine demise (case 5).

The fourth patient with placental abruption (case 6) presented at Zuni-Ramah with vaginal bleeding, severe variable decelerations, and a 10-point drop from baseline hematocrit. She was scheduled to labor at GIMC because of a history of prior cesarean but presented to the Zuni-Ramah emergency room with vaginal bleeding. She was transferred to GIMC for an anticipated cesarean delivery; however, on arrival the patient rapidly progressed and gave birth to an infant vaginally with Apgar scores of 3 and 9. Her infant had a neonatal seizure and magnetic resonance imaging evidence of sagittal sinus thrombosis. The infant had a normal neurologic evaluation, developmental assessment, and electroencephalogram at 15 months.

We reviewed 5 cases of urgent transfer for fetal distress. These were differentiated from the 8 intrapartum transfers for NRFHTs based on the severity of fetal heart monitor tracings. Four of the 5 women who had been transferred for fetal distress gave birth to healthy infants vaginally more than 2 hours after arrival at the referral institution. One patient who was urgently transferred for repetitive late decelerations is discussed below.

Cesarean deliveries for fetal distress at referral hospitals

We reviewed all cases of cesarean deliveries for fetal distress (n = 10) at referral institutions to determine whether outcomes for any of the patients could potentially have been improved by having their cesarean deliveries earlier if operative facilities had been available at Zuni-Ramah Hospital or by being transferred earlier (Table W1). Seven of the 10 patients were transferred for preeclampsia, NRFHTs, or failure to progress. All had their cesareans for fetal distress many hours after arrival at the referral institution.

Two cases of cesarean delivery for fetal distress after transfer because of abruption were previously described. A patient presented in early labor with repetitive late decelerations and was urgently transferred to GIMC, where she underwent an immediate cesarean delivery. Her infant had Apgar scores of 1 and 7, an unremarkable neonatal course, and a normal 15-month developmental screen.

Discussion

Our outcomes demonstrate that with the use of appropriate screening criteria, childbirth can safely occur in institutions that lack surgical suites. The population-based perinatal mortality rate was similar to the nationwide rate. A review of obstetric emergencies and low Apgar scores among the 839 women laboring at Zuni-Ramah Hospital failed to identify adverse outcomes that might have been prevented if the hospital had had operative facilities. Cesarean rates were approximately one third the nationwide rate even though Zuni-Ramah patients had a higher prevalence of such risk factors as diabetes and preeclampsia.

Although they represented a high-risk obstetric population, 65% of women were able to give birth at Zuni-Ramah Hospital through use of the perinatal screening criteria. The 35% rate of transfer was caused largely by the need for oxytocin augmentation or induction. Only 21.6% of the women who were transferred for dysfunctional labor or premature rupture of membranes ultimately had a cesarean delivery. Oxytocin has not been permitted at Zuni-Ramah Hospital because of the ACOG guideline permitting oxytocin use only if “a physician capable of performing a cesarean delivery is readily available.”¹⁹ There are no studies addressing the safety of labor induction or augmentation without on-site cesarean capability.

Canadian guidelines for rural maternity care do not prohibit the use of prostaglandins or oxytocin at hospitals without operative facilities. A Consensus Conference on Obstetric Services in Rural or Remote Communities addressed the issue of labor induction or augmentation in hospitals without cesarean capability by stating, “If caring for a woman in labour is appropriate in the community, then caring for her during an augmented/induced labour is equally appropriate when there is support by trained local staff and resources.”²⁰ We concur that use of oxytocin in rural hospital units without operative facilities should be considered under well-defined clinical guidelines or research protocols.

Limitations

Our study’s limitations include lack of long-term neonatal outcomes, small size of the Zuni-Ramah population, an almost exclusively Native American population, and lack of examiner blinding during record review. Transfer rates may be increased in populations with higher rates of cesarean delivery or epidural anesthesia use. Alternatively, the high incidence of preeclampsia, chronic hypertension, and diabetes in these communities may have resulted in a higher proportion of induction. Umbilical cord pro-lapse and significant placental abruption are routinely treated by urgent cesarean delivery; therefore, obstetric literature on outcomes without immediate operative intervention is limited.^21,22 A larger study would be required to determine the potential increased neonatal morbidity or mortality resulting from delayed intervention.

Conclusions

The ACOG/AAP guideline requiring on-site surgical facilities and the ability to initiate a cesarean in 30 minutes is not based on evidence. Four small retrospective studies of emergency cesarean deliveries delayed for more than 30 minutes did not demonstrate adverse neonatal outcomes.^23-26 In our study population, no adverse outcomes (none in 839 births) were determined to have been caused by a lack of surgical facilities. Despite these excellent outcomes, the possibility always exists for outcomes that can be prevented by doing a rapid emergent cesarean delivery. Women deciding to give birth in facilities without operative capabilities should receive information regarding the risks and benefits of delivering there and should have access to other facilities. Provider discretion and patient choice must be respected to ensure community support of these birthing units. Practitioners at the rural units must have assurance that any patients who require an urgent transfer will be readily accepted.

Rural communities, medical providers, and health care facilities need to consider the overall effect of maintaining local maternity care units, as the loss of rural maternity care can increase the risk of adverse perinatal outcomes.^1-3 We concur with the Canadian panel that although maintenance of rural surgical and anesthesia capabilities is desirable, “good outcomes can be sustained within an integrated risk management system without local access to operative delivery.”⁸ Guidelines should be developed to permit rural hospitals without cesarean capability to provide maternity care as part of integrated perinatal systems with well-developed transport protocols and supportive referral institutions. Women living in rural areas should have the option to give birth near their homes in such units if they so desire.

Acknowledgments

The authors thank Robert Rhyne, MD, for editorial assistance in manuscript preparation and Betty Skipper, PhD, for statistical assistance. Current Zuni-Ramah Obstetric Guidelines are available at http://hsc.unm.edu/fcm/research/zuni.