Factors Related to Fetal Death in Pregnant Women with Cholera, Haiti, 2011–2014

We assessed risk factors for fetal death during cholera infection and effect of treatment changes on these deaths. Third trimester gestation, younger maternal age, severe dehydration, and vomiting were risk factors. Changes in treatment had limited effects on fetal death, highlighting the need for prevention and evidence-based treatment.

We assessed risk factors for fetal death during cholera infection and effect of treatment changes on these deaths. Third trimester gestation, younger maternal age, severe dehydration, and vomiting were risk factors. Changes in treatment had limited effects on fetal death, highlighting the need for prevention and evidence-based treatment.
C holera infections during pregnancy are associated with high rates of fetal death, especially when women are severely dehydrated (1)(2)(3)(4)(5)(6)(7). In Haiti in 2011, pregnant women with clinical signs of cholera who sought treatment from Médecins Sans Frontières (MSF) in Port-au-Prince were sent to a general cholera treatment center (CTC). In April 2012, MSF established a CTC to improve fetal outcomes in pregnant women by facilitating intensive follow-up for dehydration and rapid access to obstetric and neonatal services. In June 2013, a more aggressive rehydration protocol was implemented (online Technical Appendix Table 1, http:// wwwnc.cdc.gov/EID/article/22/1/15-1078-Techapp1.pdf). To assess the effects of cholera infection, establishment of a specialized CTC, and the new rehydration protocol, we conducted a retrospective cohort analysis of pregnant women with suspected cholera admitted to MSF's CTCs during September 1, 2011-December 31, 2014.

The Study
This study was approved by the National Bioethical Committee of the Ministry of Public Health and Population of Haiti. A cholera case-patient was defined as someone who passed >3 liquid stools with or without vomiting or dehydration in the previous 24 hours or within 6 hours of seeking treatment. Women of childbearing age were asked whether they were pregnant. Urine dipstick tests were conducted in cases of uncertainty or early stages of pregnancy. Gestational age was determined by date of the woman's last menstruation, uterine height, or ultrasound. Fetal status was assessed at admission and hourly by using fetal stethoscope or ultrasonic fetal Doppler. Women who had a miscarriage at home and were not bleeding at admission were not classified as pregnant. Dehydration status was determined according to World Health Organization categories (online Technical Appendix Table 1) (8). We assigned women into 3 treatment groups (TGs) according to whether they were treated in the general or specialized CTC and whether they were given the original or new protocol (online Technical Appendix Table 1).
We analyzed fetal outcome for all pregnant women by initial signs and symptoms, TG, and clinical evolution. Multiple logistic regression modeling was used for adjusted analyses. All analyses used Stata software version 12.0 (StataCorp LP, College Station, TX, USA).
There was no modification effect of TG on postadmission fetal death. Weak evidence of a difference in effect of severe dehydration on postadmission fetal death between TGs (p = 0.09) ( Table 2) was potentially due to a lower rate among severely dehydrated women in TG2 (OR 0.4. 95% CI 0.1-1.7; data not shown). However, there was insufficient power to detect these differences, and the final model did not require adjustment. Although the proportion of postadmission fetal deaths within a TG decreased with each protocol change, the proportion in TG3 (8.5%, 27/317) was not different from TG1 (10.0%, 12/120) or TG2 (9.5%, 38/399).

Conclusions
Fetal death occurred in 141 (16%) of 900 pregnancies. Risk factors were third trimester, younger maternal age, severe dehydration, and vomiting. Treatment in a specialized CTC and aggressive rehydration did not prevent fetal death though the trend was toward improved outcomes.
Women <20 years of age were twice as likely as older women to experience fetal death. Although the relationship between fetal death and maternal age during cholera has not been documented, younger age is associated with increased risk for other adverse pregnancy outcomes (10).
The risk for fetal death was highest in the third trimester, even after controlling for maternal age, dehydration level, and vomiting. The relationship between fetal death and trimester of pregnancy is unclear (1,3,6).
Determination of dehydration status of pregnant women is difficult in later stages of pregnancy (2,11). Misclassification of dehydration status could affect fetal outcome due to placement of patients under the wrong treatment protocol. In addition, increased placental blood flow with gestational age may increase the effect of dehydration (12). Even after we controlled for dehydration level, we determined that fetal death was twice as likely in women experiencing vomiting, potentially due to electrolyte changes in amniotic fluid (2,7,(13)(14)(15).
Lack of effect of a specialized CTC on fetal death could result from a detection bias in that establishment of the specialized CTC led to an increased likelihood of detection of fetal deaths. In addition, 45% of fetal deaths occurred before women sought treatment. Fetal death may occur early in a pregnant woman's illness with cholera (6), and more than half the women sought treatment >24 hours after symptom onset, likely contributing to poor fetal outcomes. Likewise, the effect of the new treatment protocol may have been limited by fetal death occurring before the women sought treatment or by women being assigned the incorrect protocol due to difficulty in determining dehydration status.
Limitations include lack of laboratory-confirmed diagnoses. Data were collected for programmatic rather than research purposes and lack electrolyte levels, amniotic fluid composition, maternal blood group, and fetal cause of death. Some first-trimester pregnancies may have been missed. Pregnancies in women who completed miscarriage at home were not counted, potentially underestimating overall risk for fetal death. Because there was no follow-up of women after discharge, some early-term fetal deaths might have been missed. In addition, the longterm effect of treatment on fetal well-being could not be determined. TG outcomes also may have been affected by differences in factors such as women's access to health services over time.
Although the implementation of a specialized CTC did not decrease fetal deaths, specialized CTCs play a vital role in preserving patients' dignity and providing patient-centered care. Determining the mechanism of fetal death in cholera infection would enable development of evidence-based treatment protocols. Because many fetal deaths occurred before women sought treatment, the importance of cholera prevention and the risk for poor fetal outcomes should be emphasized.

Acknowledgments
We thank all staff working on cholera in Port-au-Prince, especially the midwives, nurses, and obstetricians who provided quality patient-centered care for pregnant women with cholera.
We thank Sarah Venis for editing assistance. E.S. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. E.S. and C.A. analyzed the data, and E.S. drafted the manuscript. A.L. supervised the epidemiologic analysis, and L.B. and R.S.D. coordinated the data collection. All authors contributed to the conception and design of the study and to the interpretation of the data. All authors critically revised the manuscript for intellectual content and approved the final draft for submission. The study design, data collection, analysis, decision to publish, and preparation of the manuscript was approved by Médecins Sans Frontières and