Maternal infection is a leading cause of severe maternal morbidity and was responsible for 12.7% of reported maternal deaths from 2011 to 2013.1,2 Maternal intrapartum fever is often attributed to chorioamnionitis, an infection or inflammation of the amnion, chorion, or both.3,4 The prevalence of chorioamnionitis ranges from 1 in 30 for term deliveries to 1 in 4 for preterm deliveries.5,6 A clinical suspicion for chorioamnionitis triggers a wide range of diagnostic workup and antibiotic treatment for the mother and newborn.
In January 2015, the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) invited a panel of experts to address knowledge gaps surrounding the diagnosis and management of intrapartum fever.4 This panel recommended replacing the term chorioamnionitis in favor of “intrauterine inflammation or infection or both” or “triple I” and established criteria to improve the specificity of the diagnosis of triple I to decrease overtreatment of intrapartum women and low-risk newborns (Table 1).4 Antibiotic therapy would be recommended for women meeting criteria for suspected triple I only. Although improved diagnostic specificity may decrease unnecessary intervention, early recognition and treatment of clinical infection is key to decreasing adverse infectious outcomes including sepsis.7
Our objective was to assess the test characteristics of the suspected triple I criteria to predict confirmed triple I or an adverse clinical infectious outcome and to examine the incidence of adverse outcomes among febrile women who do and do not meet the NICHD classification for suspected triple I.
This was a retrospective cohort study of women with intrapartum fever, defined as oral or axillary temperature 100.4°F or greater (38.0°C) at a single academic tertiary care center from June 2015 to September 2017. Axillary temperatures were rarely taken and only when a patient had recently consumed a cold beverage or ice. Since 2009, the study site has had a protocol for the evaluation and treatment of febrile intrapartum women. This protocol emphasizes early evaluation of intrapartum women with a temperature of 100.4°F (38.0°C) or greater with a complete blood count, blood cultures, and urine cultures before initiating a standardized regimen of intravenous antibiotics for all patients, as previously described.8,9 The duration of antibiotic therapy is determined by mode of delivery and other clinical factors.
Febrile women were identified by examination of the medical record among women admitted to labor and delivery with at least one temperature of 100.4°F (38.0°C) or greater. Among this group of women, we queried the hospital microbiology database of blood cultures, which catalogs the provenance of every culture ordered. Women with at least one temperature 100.4°F or greater (38.0°C) during labor or within 1 hour postpartum and with blood culture data available met initial inclusion criteria. Women with a fever within 1 hour postpartum were included because the source is presumed to be uterine and would still require evaluation and treatment of the newborn.4
We excluded women with gestational age less than 24 weeks, intrauterine fetal demise, fetal anomaly, chronic immunosuppression, documented nonobstetric sources of infection, and expectantly managed preterm prelabor rupture of membranes. Although at higher risk of infectious outcomes of interest, women with preterm prelabor rupture of membranes were excluded because their management includes intravenous and oral antibiotic therapy for latency and steroids for fetal lung maturity, interventions that may affect findings on outcomes such as pathology. Women with nonobstetric infectious diagnoses, based on discharge documentation by the attending obstetrician, and women taking antibiotics for nonobstetric indications before developing an intrapartum fever were excluded. Women with positive Group B streptococci screening received intravenous antibiotics in accordance with the Centers for Disease Control and Prevention guidelines and were not excluded from the study.10,11
The NICHD recommendations describe criteria to diagnose intrapartum fever with additional clinical characteristics that should raise suspicion for suspected triple I. Per the NICHD, a documented fever is defined as a single temperature 102.2°F or greater (39.0°C) or a temperature 100.4°F or greater (38.0°C) but less than 102.2°F (39.0°C) sustained over two measurements 30 minutes apart (Table 1).4 For this study, we defined documented fever similarly, but allowed an interval of 45 minutes within which a repeat temperature could be measured. Our institutional evaluation of an intrapartum fever does not mandate the assessment of a repeat temperature, which is left to the discretion of the clinical team. With the time needed for the initial clinical evaluation, initiation of intravenous antibiotic treatment, and continued labor management, most repeat temperature measurements clustered within 30–45 minutes of the initial fever in our data set. Women with an initial temperature 102.2°F or greater (39.0°C) were included, regardless of whether a repeat temperature was measured within 45 minutes. If the initial temperature was 100.4°F or greater (38.0°C) but less than 102.2°F (39.0°C), only women with any repeat temperature measurement within 45 minutes of the index temperature were included in the analysis, because the evaluation of a repeat temperature value is an important component of evaluation according to the NICHD criteria.
We classified women into one of two groups: 1) suspected triple I, defined as women with documented fever plus clinical signs of intrauterine infection; and 2) isolated maternal fever, defined as women with at least one temperature 100.4°F or greater (38.0°C) who did not meet criteria for suspected triple I (Table 1; Fig. 1). Clinical signs of intrauterine infection included any of the following findings: maternal leukocytosis greater than 15,000 per mm3, fetal tachycardia greater than 160 beats per minute, and purulent amniotic fluid.4 Our primary outcome was confirmed triple I, defined as women with suspected triple I and placental pathology diagnostic of infection. Placental pathology diagnostic of infection was defined as moderate or severe acute chorioamnionitis, umbilical cord or chorionic plate vascular involvement, acute villitis, and funisitis.12–14 We did not include amniocentesis criteria for confirmation of triple I because amniocentesis is seldom performed at our institution in women with intrapartum fever. Data on demographic, obstetric, and clinical exposures were abstracted from the electronic medical record by two obstetricians (S.O., G.W.). Blood culture data were obtained from the microbiology database. Two pathologists (Z.O.S., D.J.R.) reviewed a randomly generated subset of abstracted data from the pathology records to assess accurate interpretation and otherwise adjudicated placental pathology reports for correct classification as appropriate. Study data were collected and stored in REDCap, a secure electronic record system provided by Partners Healthcare.
We denoted our secondary outcome to be a composite adverse clinical infectious outcome, defined as maternal bacteremia, maternal intensive care unit admission, maternal fever greater than 24 hours postpartum, postpartum hemorrhage, maternal readmission for obstetric infection, maternal death resulting from obstetric infection, maternal additional procedures (postpartum dilation and curettage, interventional radiology drain placement, opening of an infected surgical site, exploratory laparotomy, hysterectomy), neonatal bacteremia or infection, neonatal intubation for greater than 2 days, intraventricular hemorrhage, hypoxic–ischemic encephalopathy, and neonatal death resulting from infection.6,15
We compared sociodemographic, medical, and obstetric characteristics for women in the isolated maternal fever group and the suspected triple I group using χ2 or Fisher exact tests for categorical variables and Wilcoxon rank-sum test for nonparametric continuous variables. Statistical significance was defined as a two-sided P<.05. We analyzed the sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio, of suspected triple I using isolated maternal fever as the test negative group to predict confirmed triple I (primary outcome) and the composite adverse clinical infectious outcome (secondary outcome).
All analyses were performed using SAS 9.4, and test characteristics were calculated using MedCalc online calculator software (https://www.medcalc.org/calc/diagnostic_test.php). The study was approved by the Partners Human Research Committee (Protocol #2017P001453).
During the study period, there were 10,138 total deliveries; 1,073 (10.6%) women had a diagnosis of fever intrapartum or immediately postpartum. Of these, 848 (79%) had blood cultures sent, and 339 women were included in the final analysis (Fig. 1). The most common reasons for exclusion were occurrence of index fever greater than 1 hour postpartum and lack of repeat temperature assessment within 45 minutes for women with initial fever 100.4°F or greater (38.0°C) but less than 102.2°F (39°C).
Two hundred twelve women (62.5%) met criteria for the suspected triple I group and 127 women (37.5%) met criteria for the isolated maternal fever group (Fig. 1; Table 2). No significant differences were noted in baseline demographic or clinical characteristics between the groups. Overall, the median age of women in the cohort was 31.0 years (interquartile range 26.5–33.9) with a median gestational age of 39.9 weeks (interquartile range 39.1–40.6). Approximately 80% of the women in the cohort were nulliparous, and 68% had a vaginal delivery. Almost all had neuraxial analgesia (99.4%).
Placental pathology was available for 86.4% of the overall cohort. The frequency of placental pathology diagnostic of infection was 37.6% in the suspected triple I group and 26.2% in isolated maternal fever group (P=.045; Table 3). The incidence of adverse clinical infectious outcomes was similar: 11.8%in the suspected triple I group and 9.5% in the isolated maternal fever group (P=.50; Table 2).
The sensitivity and specificity of the suspected triple I criteria to predict confirmed triple I were 71.4% (95% CI 61.4–80.1%) and 40.5% (95% CI 33.6–47.8%), respectively. The sensitivity and specificity of the suspected triple I criteria to predict an adverse clinical infectious outcome were 67.6% (95% CI 50.2–82.0%) and 38.1% (95% CI 32.6–43.8%), respectively. The corresponding positive likelihood ratio and negative likelihood ratio for confirmed triple I were 1.20 (95% CI 1.01–1.42) and 0.71 (95% CI 0.49–1.01), respectively. The positive likelihood ratio and negative likelihood ratio for adverse clinical infectious outcomes were 1.09 (0.86–1.39) and 0.85 (0.52–1.39), respectively.
Our analysis demonstrates that the triple I criteria may miss a large proportion of febrile women at risk for adverse clinical infectious outcomes. Considering the potential consequences of untreated maternal or neonatal infection, our data suggest caution in universal implementation of the triple I criteria to guide clinical management of febrile women in the intrapartum period.
In our study, the triple I criteria have a low sensitivity, specificity, a positive likelihood ratio, and a negative likelihood ratio to predict pathologic placental changes diagnostic of infection. Prior studies have also shown an inconsistent relationship between clinical suspicion for chorioamnionitis and histopathologic findings.16–19 Because concern for adverse clinical outcomes drives decision-making in the management of febrile women in labor, we used adverse clinical infectious outcomes as our secondary outcome and still found low sensitivity and specificity for the triple I criteria with this clinically driven approach.
Our second finding is that febrile women not meeting suspected triple I criteria, and thus not meeting criteria for intrapartum antibiotic treatment, may still be at risk for maternal and fetal morbidity, as evidenced by our high rates of adverse clinical infectious outcomes. Although a prior study at our institution indicated that a temperature greater than 102.2°F (39.0°C) was predictive of bacteremia,8 a recent study of these bacteremic women demonstrated that 37.5% of women with adverse infectious outcomes had a fever below 102.2°F (39.0°C).9 This is consistent with our findings that “low-risk” febrile women per NICHD criteria are still at risk for adverse outcomes. Of note, our reported morbidities occur in the context of universal treatment for fever in labor. Although we believe that prioritizing early treatment is essential, our approach to intrapartum fever limits our ability to fully assess the triple I criteria as well as the generalizability of our findings. Had we not treated women with isolated maternal fever, an unknown proportion of women may have developed additional characteristics that meet suspected triple I criteria and therefore warrant treatment. However, waiting to meet additional criteria before initiating treatment may be associated with higher morbidity. This is an area to consider for future research.
Following the publication of the triple I criteria, the American College of Obstetricians and Gynecologists issued a statement recommending consideration of treatment of women who fall under the umbrella of isolated maternal fever as defined by the NICHD, and our study supports this recommendation.20 In addition, application of the triple I criteria has important implications for neonatal management. There has been significant correlation reported between clinical chorioamnionitis and cerebral palsy at all gestational ages.18 Several studies have shown that maternal antibiotic use for chorioamnionitis is associated with a decrease in culture-proven early-onset sepsis, neonatal intensive care unit admission, and associated costs among late preterm to term newborns, suggesting a neonatal benefit to maternal treatment.21–23 Recent data have explored decreasing unnecessary antibiotic use among newborns delivered in women with suspected chorioamnionitis.4,22,23 The published NICHD criteria were another step in this direction; however, our findings suggest caution in uniform implementation of these guidelines before further evaluation.
Strengths of our study include the use of a unique, large database with physiologic, laboratory, and microbiological parameters. Furthermore, in assessing for other possible confounders, we were able to assess exposure to factors that could be associated with intrapartum fever, including prostaglandin or neuraxial analgesia use, and found no significant differences.17,24–26
We note several limitations to our study as well. First, we included only those women who had blood cultures sent at initial fever, because this allowed us to identify febrile women with the most complete fever workup to allow assessment of adverse clinical infectious outcomes. However, only 80% of febrile patients had blood cultures sent despite the institutional protocol requiring blood cultures as part of the initial evaluation of an intrapartum fever. Our labor and delivery comprises the resident practice, several private practices, and a midwifery service, and the implementation of the protocol is self-enforced, particularly in the nonresident practices. Of note, an important consideration in interpreting our defined morbidity of maternal bacteremia is our inability to differentiate transient compared with pathologic bacteremia. Additionally, because repeating the temperature was not part of our clinical protocol to evaluate an intrapartum fever, 37% of febrile women were excluded because they did not have a repeat temperature taken within 45 minutes of the first febrile temperature. We acknowledge that these limitations create a selection bias, because requiring a second temperature and sending blood cultures may have led to selecting a sicker group of women in whom a clinical decision was made to send blood cultures and reevaluate the temperature. However, this limitation reflects the applicability of the proposed criteria to real-life clinical settings in which fevers may be treated without repeat temperatures being measured in the recommended timeframe. Lastly, only 86% of women had placental pathology sent. Although we had incomplete outcome assessment for this variable, we did not exclude women based on available placenta pathology because our outcome assessment was not solely based on placenta pathology.
The need for an improved method of diagnosis for intraamniotic infection remains based on our assessment of the utility of the NICHD criteria. A simple bedside minimally invasive marker of infection may be ideal. Future research should aim to create a clinically relevant algorithm to identify and treat intraamniotic infection with the addition of such a bedside test. If clinicians adopt the currently proposed triple I criteria for the evaluation and management of intrapartum fever, close monitoring of maternal and neonatal outcomes should be a high priority.
1. Berg CJ, Callaghan WM, Syverson C, Henderson Z. Pregnancy-related mortality in the United States, 1998–2005. Obstet Gynecol 2010;116:1302–9.
2. Creanga AA, Syverson C, Seed K, Callaghan WM. Pregnancy-related mortality in the United States, 2011–2013. Obstet Gynecol 2017;130:366–73.
3. Blanc WA. Pathways of fetal and early neonatal infection: viral placentitis, bacterial and fungal chorioamnionitis. J Pediatr 1961;59:473–96.
4. Higgins RD, Saade G, Polin RA, Grobman WA, Buhimschi IA, Watterberg K, et al. Evaluation and management of women and newborns with a maternal diagnosis of chorioamnionitis: summary of a workshop. Obstet Gynecol 2016;127:426–36.
5. Johnson CT, Adami RR, Farzin A. Antibiotic therapy for chorioamnionitis to reduce the global burden of associated disease. Front Pharmacol 2017;8:97.
6. Ericson JE, Laughon MM. Chorioamnionitis: implications for the neonate. Clin Perinatol 2015;42:155–65, ix.
7. Rhodes A, Evans LE, Alhazzani W, Levy MM, Antonelli M, Ferrer R, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Intensive Care Med 2017;43:304–77.
8. Molina RL, Easter SR, Venkatesh KK, Cantonwine DE, Kaimal AJ, Tuomala RE, et al. Defining physiological predictors of peripartum maternal bacteremia. Am J Perinatol 2015;32:1342–50.
9. Easter SR, Molina RL, Venkatesh KK, Kaimal A, Tuomala R, Riley LE. Clinical risk factors associated with peripartum maternal bacteremia. Obstet Gynecol 2017;130:710–7.
10. Cape A, Tuomala RE, Taylor C, Puopolo KM. Peripartum bacteremia in the era of group B streptococcus prophylaxis. Obstet Gynecol 2013;121:812–8.
11. Verani JR, McGee L, Schrag SJ; Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention (CDC). Prevention of perinatal group B streptococcal disease—revised guidelines from CDC, 2010. MMWR Recomm Rep 2010;59:1–36.
12. Tita AT, Andrews WW. Diagnosis and management of clinical chorioamnionitis. Clin Perinatol 2010;37:339–54.
13. Buhimschi CS, Bhandari V, Hamar BD, Bahtiyar M-O, Zhao G, Sfakianaki AK, et al. Proteomic profiling of the amniotic fluid to detect inflammation, infection, and neonatal sepsis. PLoS Med 2007;4:e18.
14. Redline RW. Villitis of unknown etiology: noninfectious chronic villitis in the placenta. Hum Pathol 2007;38:1439–46.
15. Grobman WA, Bailit JL, Rice MM, Wapner RJ, Varner MW, Thorp JM, et al. Can differences in obstetric outcomes be explained by differences in the care provided? The MFMU Network APEX study. Am J Obstet Gynecol 2014;211:147.e1–16.
16. Roberts DJ, Celi AC, Riley LE, Onderdonk AB, Boyd TK, Johnson LC, et al. Acute histologic chorioamnionitis at term: nearly always noninfectious. PLoS One 2012;7:e31819.
17. Riley LE, Celi AC, Onderdonk AB, Roberts DJ, Johnson LC, Tsen LC, et al. Association of epidural-related fever and noninfectious inflammation in term labor. Obstet Gynecol 2011;117:588–95.
18. Jessop F, Sebire N. Histological chorioamnionitis: current concepts of diagnosis, classification and clinical significance. Fetal Matern Med Rev 2011;22:25–44.
19. Sharma SK, Rogers BB, Alexander JM, McIntire DD, Leveno KJ. A randomized trial of the effect of antibiotics prophylaxis on epidural related fever in labor. Anesth Analg 2014;118:604–10.
20. Intrapartum management of intraamniotic infection. Committee Opinion No. 712. American College of Obstetricians and Gynecologists. Obstet Gynecol 2017;130:e95–101.
21. Braun D, Bromberger P, Ho N, Getahun D. Low rate of perinatal sepsis in term infants of mothers with chorioamnionitis. Am J Perinatol 2016;33:143–50.
22. Puopolo KM, Draper D, Wi S, Newman TB, Zupancic J, Lieberman E, et al. Estimating the probability of neonatal early-onset infection on the basis of maternal risk factors. Pediatrics 2011;128:e1155–63.
23. Jan AI, Ramanathan R, Cayabyab RG. Chorioamnionitis and management of asymptomatic infants ≥35 weeks without empiric antibiotics. J Pediatr 2017;140. pii: e20162744.
24. Goetzl L, Cohen A, Frigoletto F, Lang JM, Lieberman E. Maternal epidural analgesia and rates of maternal antibiotic treatment in a low-risk nulliparous population. J Perinatol 2003;23:457–61.
25. Goetzl L. Epidural analgesia and maternal fever. Curr Opin Anaesthesiol 2012;25:292–9.
26. Knowles S, O’Sullivan N, Meenan A, Hanniffy R, Robson M. Maternal sepsis incidence, aetiology and outcome for mother and fetus: a prospective study. BJOG 2015;122:663–71.