Original Article Gynecol Obstet Invest 2014;78:201–207 DOI: 10.1159/000364871
Received: February 26, 2014 Accepted after revision: May 27, 2014 Published online: September 9, 2014
Serial Ultrasonographic Examination of the Fetal Thymus in the Prediction of Early Neonatal Sepsis in Preterm Premature Rupture of Membranes Orkun Cetin a Ipek Dokurel Cetin b Seyfettin Uludag c Cihat Sen c Fatma Ferda Verit d Onur Guralp e a
Maternal and Fetal Medicine, Obstetrics and Gynecology, Yuzuncu Yuzyil University, Van, b Pediatrics, and Division of Maternal and Fetal Medicine, Obstetrics and Gynecology, Cerrahpasa School of Medicine, Istanbul University, d Obstetrics and Gynecology, Suleymaniye Obstetrics and Gynecology Research and Training Hospital, Istanbul, and e Obstetrics and Gynecology, Bozova State Hospital, Sanliurfa, Turkey c
Key Words Preterm premature rupture of membranes · Fetal thymus transverse diameter · Early neonatal sepsis · Interleukin-6 · Tumor necrosis factor-α
Abstract Background/Aims: To evaluate the diagnostic accuracy of fetal thymus transverse diameter (FTTD) in predicting fetal infection in preterm premature rupture of membranes (PPROM) and compare its accuracy with cord blood tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6). Methods: Forty consecutive pregnancies complicated with PPROM between 261/7 and 366/7 gestational weeks were evaluated prospectively. Serial fetal ultrasonography follow-ups with 3-day intervals were performed beginning on the admission day. The FTTD was recorded on every ultrasonographic examination. Cord blood TNF-α and IL-6 values were measured after delivery. Results: FTTD was decreased below 5% according to nomograms compared to the initial measurement in 45% of all PPROM cases. Decreased FTTD had a sensitivity of 100%, specificity of 73%, positive predictive value of 55%, and negative predictive value of 100% in predicting early neonatal sepsis. Cord blood TNF-α had
© 2014 S. Karger AG, Basel 0378–7346/14/0783–0201$39.50/0 E-Mail [email protected] www.karger.com/goi
a sensitivity of 80% and specificity of 90%, whereas IL-6 had a sensitivity of 90% and specificity of 63.3% in predicting early neonatal sepsis. Conclusions: Assessment of the decrease in FTTD by serial ultrasonographic examinations is a promising ‘prenatal’ method for the early detection of early neonatal sepsis. © 2014 S. Karger AG, Basel
Introduction
Preterm premature rupture of membranes (PPROM) is an important risk factor for early neonatal sepsis which remains a feared cause of morbidity and mortality in the neonatal period. The host response to infection triggers an inflammatory cascade which in turn acts synergistically with microbial impact to induce preterm labor and fetal damage. Clinical signs and laboratory findings of neonatal sepsis are often nonspecific, and this complexity requires a high degree of suspicion for early diagnosis. Some laboratory parameters might be beneficial for the screening of neonatal sepsis, but none of these are specific and sensitive enough to be used alone. Onur Guralp, Consultant, MD Akademiler Sit. B/37 Kat. 8 Zincirlikuyu TR–34340 Istanbul (Turkey) E-Mail dronur @ hotmail.com
Prediction of neonatal sepsis from the intrauterine period has long been challenged. Gram staining, leukocyte count, glucose, and interleukin-6 (IL-6) measurement in amniotic fluid are used to detect chorioamnionitis in pregnant women with high risk of intrauterine infection [1–4]. These tests are invasive since they all require the yield of amniotic fluid. Invasive tests are more sensitive and specific compared to noninvasive tests; however, there are certain risks such as preterm labor and dissemination of infection. Moreover, it is impractical to repeat invasive tests in long-term follow-up. The noninvasive tests that are used in the detection of intrauterine infection are maternal serum C-reactive protein (CRP), white blood cell (WBC) count, and IL-6 measurements. These tests have relatively low sensitivity rates between 67 and 81% [5, 6]. Therefore, a highly sensitive and specific noninvasive prenatal test is needed to predict fetal infection in PPROM pregnancies. Neonatal infection does not occur in every PPROM case. If only we could accurately foresee the neonatal infection, we would choose active management instead of expectant management to prevent fetal morbidities. However, at present, we only have postnatal tests such as IL-6 and tumor necrosis factor-α (TNF-α) measurements in cord blood. The fetal thymus was first visualized by ultrasonographic examination by Felker et al. [7] in 1989. In the axial section, the fetal thymus is in front of the three-vessel view. It is located in the upper mediastinum and anterior to the ductal and aortic arches. It is readily distinguished as an oval-shaped mass between the fetal lungs; and it is more hypoechogenic compared to the surrounding tissue (fig. 1). Echogenic punctuations may be observed. The echogenic difference can be better observed if the insonation angle is well adjusted [8, 9]. Thymus size is known to increase as the gestational age increases [10]. The transverse diameter of the fetal thymus (FTTD) below the fifth percentile for the gestational age was found to be associated with histologic chorioamnionitis and funisitis [11]. If FTTD can predict histologic chorioamnionitis and funisitis, it may also predict early neonatal sepsis. In this study, we aimed to investigate the diagnostic accuracy of FTTD, as a noninvasive prenatal marker, in predicting fetal infection in PPROM pregnancies to decide when to choose active management in order to decrease fetal morbidity and mortality rates. We also compared the predictive values of two important promising postnatal markers, IL-6 and TNF-α, with FTTD for predicting neonatal sepsis. 202
Gynecol Obstet Invest 2014;78:201–207 DOI: 10.1159/000364871
Fig. 1. Ultrasonographic appearance of the fetal thymus in a 332/7
GW fetus. In the axial section, the fetal thymus is in front of the three-vessel view. It is located in the upper mediastinum and anterior to the ductal and aortic arches. It is readily distinguished as an oval-shaped mass between the fetal lungs; and it is more hypoechogenic compared to the surrounding tissue. FTTD of 3.36 cm is within the normal limits for 33 GW (mean: 3.3 cm; 95% reference range 2.6–3.9 cm [10]) Ao = Aorta; PA = pulmonary artery; Sv = superior vena cava.
Materials and Methods Forty consecutive pregnancies complicated with PPROM between 261/7 and 366/7 gestational weeks (GW), admitted to the Department of Maternal Fetal Medicine, Istanbul University, Cerrahpasa School of Medicine, between January 2009 and July 2011, were evaluated in a prospective study. The study was approved by the Istanbul University Clinical Research Ethics Committee. It was financed by the Istanbul University Scientific Research Projects Fund. Informed consent was obtained from all patients. STAR-D guidelines were followed. Any potential maternal or fetal problems which might influence the IL-6, TNF-α levels and fetal thymus size, or cause biologic stress or increase the risk of infection were excluded. Therefore, patients with uncontrolled diabetes mellitus, preeclampsia, placental abruption, multiple pregnancy, polyhydramnios, acute febrile illnesses, chorioamnionitis, severe intrauterine growth restriction, fetal demise, and fatal fetal anomalies were excluded. In order to minimize the potential impact of previous chorioamniotic infections on the fetal thymus, women who had rupture of membranes for longer than 12 h on admission were excluded. The diagnosis of PPROM was performed by visualization of active fluid discharge on dry speculum examination, followed by confirmation by Amnisure® (AmniSure International LLC, Boston, Mass., USA). Gestational age was determined according to the last menstrual period and measurement of crown-rump length in the first trimester.
Cetin /Dokurel Cetin /Uludag /Sen /Verit / Guralp
All patients were hospitalized and checked for the vital signs and presence of uterine tenderness. A nonstress test was performed daily. Complete blood count, urine analysis, colony counting in urine samples, CRP, blood urea nitrogen, creatinine, alanine aminotransferase, aspartate aminotransferase, total protein, albumin, conjugated bilirubin, total bilirubin, and glucose measurements were performed on admission. Ampicillin 4 g/day was administered to all patients. The antibiotics were changed according to the antibiogram results if >100,000 colonies were detected in urine sample. Antibiotics were continued for 7 days. Betamethasone 12 mg intramuscular injections were performed twice with 12-hour intervals in pregnancies under 34 GW to achieve fetal lung maturity. The latent period was defined as the time interval between membrane rupture and onset of labor. Fetal ultrasonographic examination (by the ultrasound Toshiba Xario, Japan and Voluson 730, General Electric Company, USA) was performed in all patients by the same physician (O.C.). Estimated fetal weight, amniotic fluid index, fetal presentation, and FTTD were recorded. Serial fetal ultrasonographic measurements with 3-day intervals were performed beginning on the admission day. Ultrasonographic measurements were performed by the same sonographer (O.C.). FTTD was measured three times, and the mean value was recorded. To prevent any potential bias, a double blinding procedure was performed. The sonographer did not reveal the measurements until the end of the study and did not take part in obstetric management. Amniotic fluid volume and the FTTD were recorded on every ultrasonographic examination. The fetal thymus size-gestational age nomograms described by Cho et al. [10] were used. In order to investigate intraobserver variability, the diameter was measured twice in 20 patients who were selected arbitrarily. A Bland-Altman plot of agreement with 95% limits of agreement was plotted for the intraobserver variability. ‘Decreased FTTD’ was defined as a decrease in the FTTD below 5% according to the nomograms of Cho et al. [10] at ultrasonographic examinations performed after admission. Expectant management was stopped either when active labor initiated spontaneously or in the presence of fetal distress or findings of chorioamnionitis [maternal fever >38 ° C, uterine tenderness, foul-smelling vaginal discharge, maternal (>100 bpm) or fetal (>160 bpm) tachycardia, WBC >15,000 leukocytes/μl, increased CRP]. The route of delivery was chosen according to the obstetrical indications. After delivery of the baby, cord blood samples were obtained and centrifuged, and then stored at –33 ° C. IL-6 (ELISA DIA Source, Anogen, Ont., Canada) and TNF-α (ELISA DIA Source, Adipo Bioscience, Santa Clara, Calif., USA) were measured with the ELISA DIA Source kit. The epidemiologic data (age, parity, gestational age at PPROM occurrence and the duration after PPROM, route of delivery, use of induction of labor, indications of operative deliveries, birth weight, Apgar scores, and gender) were recorded after the delivery. Fetal CRP, fetal blood and gastric washing cultures were obtained from all cases.
The threshold for TNF-α positivity was accepted at 22 pg/ml, and the threshold for IL-6 positivity was accepted at 11 pg/ml. Statistical Analyses Descriptives were defined as percent, rates, mean and standard deviation for parametric variables, and median and range for nonparametric variables. The homogeneity of the parameters was tested by the Kolmogorov-Smirnov and Levene’s tests. Comparative analyses were performed by the χ2 test for nonparametric variables, and the t test was used for parametric variables. Correlation analyses were performed by Pearson’s correlation test for parametric variables and Spearman’s test for nonparametric variables. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated by the receiver operating characteristic (ROC) curve. The SPSS® (Statistical Package for Social Sciences, version 14) was used for statistical calculations. The Bland-Altman plot was performed in MedCalc® for evaluation of the intraobserver variability. p < 0.05 was accepted as significant.
Results
The Reference Standard The reference standard for the diagnosis of early neonatal sepsis is clinical diagnosis, which is based on the presence of clinical findings (pale appearance, lethargy, irritability, apnea, respiratory stress, bradycardia, tachycardia, hypotension, vomiting, and fever) and/or positive blood and gastric washing cultures.
The mean age and median parity of 40 PPROM cases were 31.2 ± 5.3 years and 1 (0–3), respectively. Mean gestational age on admission was 32.4 ± 3.32. Twenty-three (57.5%) women had induction of labor, 5 (12.5%) women had spontaneous vaginal delivery, and 12 (30%) underwent cesarean section. Cesarean indications were fetal stress in 41.7%, previous cesarean section in 33.3%, and cephalopelvic disproportion or arrest of labor in 25%. Mean birth weight was 2,184 ± 757 g (range 400–3,280 g). Mean duration of follow-up was 4.8 ± 2.6 days (range 3–8 days). Twenty women were
Received: February 26, 2014 Accepted after revision: May 27, 2014 Published online: September 9, 2014
Serial Ultrasonographic Examination of the Fetal Thymus in the Prediction of Early Neonatal Sepsis in Preterm Premature Rupture of Membranes Orkun Cetin a Ipek Dokurel Cetin b Seyfettin Uludag c Cihat Sen c Fatma Ferda Verit d Onur Guralp e a
Maternal and Fetal Medicine, Obstetrics and Gynecology, Yuzuncu Yuzyil University, Van, b Pediatrics, and Division of Maternal and Fetal Medicine, Obstetrics and Gynecology, Cerrahpasa School of Medicine, Istanbul University, d Obstetrics and Gynecology, Suleymaniye Obstetrics and Gynecology Research and Training Hospital, Istanbul, and e Obstetrics and Gynecology, Bozova State Hospital, Sanliurfa, Turkey c
Key Words Preterm premature rupture of membranes · Fetal thymus transverse diameter · Early neonatal sepsis · Interleukin-6 · Tumor necrosis factor-α
Abstract Background/Aims: To evaluate the diagnostic accuracy of fetal thymus transverse diameter (FTTD) in predicting fetal infection in preterm premature rupture of membranes (PPROM) and compare its accuracy with cord blood tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6). Methods: Forty consecutive pregnancies complicated with PPROM between 261/7 and 366/7 gestational weeks were evaluated prospectively. Serial fetal ultrasonography follow-ups with 3-day intervals were performed beginning on the admission day. The FTTD was recorded on every ultrasonographic examination. Cord blood TNF-α and IL-6 values were measured after delivery. Results: FTTD was decreased below 5% according to nomograms compared to the initial measurement in 45% of all PPROM cases. Decreased FTTD had a sensitivity of 100%, specificity of 73%, positive predictive value of 55%, and negative predictive value of 100% in predicting early neonatal sepsis. Cord blood TNF-α had
© 2014 S. Karger AG, Basel 0378–7346/14/0783–0201$39.50/0 E-Mail [email protected] www.karger.com/goi
a sensitivity of 80% and specificity of 90%, whereas IL-6 had a sensitivity of 90% and specificity of 63.3% in predicting early neonatal sepsis. Conclusions: Assessment of the decrease in FTTD by serial ultrasonographic examinations is a promising ‘prenatal’ method for the early detection of early neonatal sepsis. © 2014 S. Karger AG, Basel
Introduction
Preterm premature rupture of membranes (PPROM) is an important risk factor for early neonatal sepsis which remains a feared cause of morbidity and mortality in the neonatal period. The host response to infection triggers an inflammatory cascade which in turn acts synergistically with microbial impact to induce preterm labor and fetal damage. Clinical signs and laboratory findings of neonatal sepsis are often nonspecific, and this complexity requires a high degree of suspicion for early diagnosis. Some laboratory parameters might be beneficial for the screening of neonatal sepsis, but none of these are specific and sensitive enough to be used alone. Onur Guralp, Consultant, MD Akademiler Sit. B/37 Kat. 8 Zincirlikuyu TR–34340 Istanbul (Turkey) E-Mail dronur @ hotmail.com
Prediction of neonatal sepsis from the intrauterine period has long been challenged. Gram staining, leukocyte count, glucose, and interleukin-6 (IL-6) measurement in amniotic fluid are used to detect chorioamnionitis in pregnant women with high risk of intrauterine infection [1–4]. These tests are invasive since they all require the yield of amniotic fluid. Invasive tests are more sensitive and specific compared to noninvasive tests; however, there are certain risks such as preterm labor and dissemination of infection. Moreover, it is impractical to repeat invasive tests in long-term follow-up. The noninvasive tests that are used in the detection of intrauterine infection are maternal serum C-reactive protein (CRP), white blood cell (WBC) count, and IL-6 measurements. These tests have relatively low sensitivity rates between 67 and 81% [5, 6]. Therefore, a highly sensitive and specific noninvasive prenatal test is needed to predict fetal infection in PPROM pregnancies. Neonatal infection does not occur in every PPROM case. If only we could accurately foresee the neonatal infection, we would choose active management instead of expectant management to prevent fetal morbidities. However, at present, we only have postnatal tests such as IL-6 and tumor necrosis factor-α (TNF-α) measurements in cord blood. The fetal thymus was first visualized by ultrasonographic examination by Felker et al. [7] in 1989. In the axial section, the fetal thymus is in front of the three-vessel view. It is located in the upper mediastinum and anterior to the ductal and aortic arches. It is readily distinguished as an oval-shaped mass between the fetal lungs; and it is more hypoechogenic compared to the surrounding tissue (fig. 1). Echogenic punctuations may be observed. The echogenic difference can be better observed if the insonation angle is well adjusted [8, 9]. Thymus size is known to increase as the gestational age increases [10]. The transverse diameter of the fetal thymus (FTTD) below the fifth percentile for the gestational age was found to be associated with histologic chorioamnionitis and funisitis [11]. If FTTD can predict histologic chorioamnionitis and funisitis, it may also predict early neonatal sepsis. In this study, we aimed to investigate the diagnostic accuracy of FTTD, as a noninvasive prenatal marker, in predicting fetal infection in PPROM pregnancies to decide when to choose active management in order to decrease fetal morbidity and mortality rates. We also compared the predictive values of two important promising postnatal markers, IL-6 and TNF-α, with FTTD for predicting neonatal sepsis. 202
Gynecol Obstet Invest 2014;78:201–207 DOI: 10.1159/000364871
Fig. 1. Ultrasonographic appearance of the fetal thymus in a 332/7
GW fetus. In the axial section, the fetal thymus is in front of the three-vessel view. It is located in the upper mediastinum and anterior to the ductal and aortic arches. It is readily distinguished as an oval-shaped mass between the fetal lungs; and it is more hypoechogenic compared to the surrounding tissue. FTTD of 3.36 cm is within the normal limits for 33 GW (mean: 3.3 cm; 95% reference range 2.6–3.9 cm [10]) Ao = Aorta; PA = pulmonary artery; Sv = superior vena cava.
Materials and Methods Forty consecutive pregnancies complicated with PPROM between 261/7 and 366/7 gestational weeks (GW), admitted to the Department of Maternal Fetal Medicine, Istanbul University, Cerrahpasa School of Medicine, between January 2009 and July 2011, were evaluated in a prospective study. The study was approved by the Istanbul University Clinical Research Ethics Committee. It was financed by the Istanbul University Scientific Research Projects Fund. Informed consent was obtained from all patients. STAR-D guidelines were followed. Any potential maternal or fetal problems which might influence the IL-6, TNF-α levels and fetal thymus size, or cause biologic stress or increase the risk of infection were excluded. Therefore, patients with uncontrolled diabetes mellitus, preeclampsia, placental abruption, multiple pregnancy, polyhydramnios, acute febrile illnesses, chorioamnionitis, severe intrauterine growth restriction, fetal demise, and fatal fetal anomalies were excluded. In order to minimize the potential impact of previous chorioamniotic infections on the fetal thymus, women who had rupture of membranes for longer than 12 h on admission were excluded. The diagnosis of PPROM was performed by visualization of active fluid discharge on dry speculum examination, followed by confirmation by Amnisure® (AmniSure International LLC, Boston, Mass., USA). Gestational age was determined according to the last menstrual period and measurement of crown-rump length in the first trimester.
Cetin /Dokurel Cetin /Uludag /Sen /Verit / Guralp
All patients were hospitalized and checked for the vital signs and presence of uterine tenderness. A nonstress test was performed daily. Complete blood count, urine analysis, colony counting in urine samples, CRP, blood urea nitrogen, creatinine, alanine aminotransferase, aspartate aminotransferase, total protein, albumin, conjugated bilirubin, total bilirubin, and glucose measurements were performed on admission. Ampicillin 4 g/day was administered to all patients. The antibiotics were changed according to the antibiogram results if >100,000 colonies were detected in urine sample. Antibiotics were continued for 7 days. Betamethasone 12 mg intramuscular injections were performed twice with 12-hour intervals in pregnancies under 34 GW to achieve fetal lung maturity. The latent period was defined as the time interval between membrane rupture and onset of labor. Fetal ultrasonographic examination (by the ultrasound Toshiba Xario, Japan and Voluson 730, General Electric Company, USA) was performed in all patients by the same physician (O.C.). Estimated fetal weight, amniotic fluid index, fetal presentation, and FTTD were recorded. Serial fetal ultrasonographic measurements with 3-day intervals were performed beginning on the admission day. Ultrasonographic measurements were performed by the same sonographer (O.C.). FTTD was measured three times, and the mean value was recorded. To prevent any potential bias, a double blinding procedure was performed. The sonographer did not reveal the measurements until the end of the study and did not take part in obstetric management. Amniotic fluid volume and the FTTD were recorded on every ultrasonographic examination. The fetal thymus size-gestational age nomograms described by Cho et al. [10] were used. In order to investigate intraobserver variability, the diameter was measured twice in 20 patients who were selected arbitrarily. A Bland-Altman plot of agreement with 95% limits of agreement was plotted for the intraobserver variability. ‘Decreased FTTD’ was defined as a decrease in the FTTD below 5% according to the nomograms of Cho et al. [10] at ultrasonographic examinations performed after admission. Expectant management was stopped either when active labor initiated spontaneously or in the presence of fetal distress or findings of chorioamnionitis [maternal fever >38 ° C, uterine tenderness, foul-smelling vaginal discharge, maternal (>100 bpm) or fetal (>160 bpm) tachycardia, WBC >15,000 leukocytes/μl, increased CRP]. The route of delivery was chosen according to the obstetrical indications. After delivery of the baby, cord blood samples were obtained and centrifuged, and then stored at –33 ° C. IL-6 (ELISA DIA Source, Anogen, Ont., Canada) and TNF-α (ELISA DIA Source, Adipo Bioscience, Santa Clara, Calif., USA) were measured with the ELISA DIA Source kit. The epidemiologic data (age, parity, gestational age at PPROM occurrence and the duration after PPROM, route of delivery, use of induction of labor, indications of operative deliveries, birth weight, Apgar scores, and gender) were recorded after the delivery. Fetal CRP, fetal blood and gastric washing cultures were obtained from all cases.
The threshold for TNF-α positivity was accepted at 22 pg/ml, and the threshold for IL-6 positivity was accepted at 11 pg/ml. Statistical Analyses Descriptives were defined as percent, rates, mean and standard deviation for parametric variables, and median and range for nonparametric variables. The homogeneity of the parameters was tested by the Kolmogorov-Smirnov and Levene’s tests. Comparative analyses were performed by the χ2 test for nonparametric variables, and the t test was used for parametric variables. Correlation analyses were performed by Pearson’s correlation test for parametric variables and Spearman’s test for nonparametric variables. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated by the receiver operating characteristic (ROC) curve. The SPSS® (Statistical Package for Social Sciences, version 14) was used for statistical calculations. The Bland-Altman plot was performed in MedCalc® for evaluation of the intraobserver variability. p < 0.05 was accepted as significant.
Results
The Reference Standard The reference standard for the diagnosis of early neonatal sepsis is clinical diagnosis, which is based on the presence of clinical findings (pale appearance, lethargy, irritability, apnea, respiratory stress, bradycardia, tachycardia, hypotension, vomiting, and fever) and/or positive blood and gastric washing cultures.
The mean age and median parity of 40 PPROM cases were 31.2 ± 5.3 years and 1 (0–3), respectively. Mean gestational age on admission was 32.4 ± 3.32. Twenty-three (57.5%) women had induction of labor, 5 (12.5%) women had spontaneous vaginal delivery, and 12 (30%) underwent cesarean section. Cesarean indications were fetal stress in 41.7%, previous cesarean section in 33.3%, and cephalopelvic disproportion or arrest of labor in 25%. Mean birth weight was 2,184 ± 757 g (range 400–3,280 g). Mean duration of follow-up was 4.8 ± 2.6 days (range 3–8 days). Twenty women were
