Neonatal echocardiogram in duodenal obstruction is unnecessary after normal fetal cardiac imaging

Accepted Manuscript Neonatal echocardiogram in duodenal obstruction is unnecessary after normal fetal cardiac imaging

Caroline Q. Stephens, Stephanie Dukhovny, Kathryn J. Rowland, Nicholas A. Hamilton PII: DOI: Reference:

S0022-3468(18)30313-0 doi:10.1016/j.jpedsurg.2018.04.039 YJPSU 58681

To appear in: Received date: Revised date: Accepted date:

28 November 2017 14 April 2018 28 April 2018

Please cite this article as: Caroline Q. Stephens, Stephanie Dukhovny, Kathryn J. Rowland, Nicholas A. Hamilton , Neonatal echocardiogram in duodenal obstruction is unnecessary after normal fetal cardiac imaging. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Yjpsu(2018), doi:10.1016/ j.jpedsurg.2018.04.039

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Title Page Title: Neonatal Echocardiogram in Duodenal Obstruction is Unnecessary After Normal Fetal Cardiac Imaging Authors: Caroline Q. Stephens BAa; Stephanie Dukhovny MDb; Kathryn J. Rowland MD, MPHSa;

T

Nicholas A. Hamilton MDa

Oregon Health & Science University, Department of Surgery, Division of Pediatric Surgery,

Doernbecher Children’s Hospital, Portland, OR, USA Stephanie Dukhovny, M.D.b -- [email protected] b

CR

a

IP

Caroline Q. Stephens, B.A.a [email protected]

US

Oregon Health & Science University, Department of Obstetrics and Gynecology, Portland OR

Kathryn J. Rowland, M.D., M.P.H.S.a -- [email protected]

Oregon Health & Science University, Department of Surgery, Division of Pediatric Surgery,

AN

a

Doernbecher Children’s Hospital, Portland, OR, USA

Oregon Health & Science University, Department of Surgery, Division of Pediatric Surgery,

ED

a

M

Nicholas A. Hamilton, M.D.a – [email protected]

Corresponding Author:

PT

Doernbecher Children’s Hospital, Portland, OR, USA

CE

Caroline Q. Stephens, B.A.

Oregon Health & Science University

AC

3181 SW Sam Jackson Park Road Mail Code CDW7

Portland, OR 97239

Business Phone: 503-494-8871 Fax: 503-494-6467 [email protected] Conflicts of Interest: None IRB Number: STUDY00015885

1

ACCEPTED MANUSCRIPT Abstract Background: Duodenal obstruction (DO) is associated with congenital cardiac anomalies that may complicate the delivery of anesthesia during surgical repair. As most infants undergo fetal ultrasounds that identify cardiac anomalies, our aim was to determine the utility of obtaining preoperative neonatal echocardiograms in all DO patients.

T

Methods:

IP

We conducted a retrospective cohort study of all DO patients treated at two tertiary care children’s

CR

hospitals between January 2005 and February 2016. Prenatal ultrasounds were compared to neonatal echocardiograms to determine concordance. Binomial exact analyses were used to estimate the negative predictive value (NPV) of prenatal imaging.

US

Results:

We identified 65 infants with DO. The majority of patients (93.8%) had prenatal ultrasounds, including

AN

twenty patients that underwent fetal echocardiogram. Fourteen (21.5%) were diagnosed with cardiac lesions in utero, and neonatal echocardiograms confirmed 12 lesions, without identifying any new

M

lesions. No changes to anesthetic management were made because of cardiac lesions. The NPV of prenatal imaging was 100% (95% Confidence Interval: 91.0-100.0).

ED

Conclusions:

Neonatal echocardiogram is unlikely to identify new cardiac lesions in DO patients with negative fetal

PT

imaging and delays in surgical care are unwarranted.

CE

Key Words: Duodenal obstruction, duodenal atresia, neonatal echocardiogram

AC

Levels of Evidence: Study of Diagnostic Test-Level II

2

ACCEPTED MANUSCRIPT Manuscript 1. Introduction

Congenital duodenal obstruction (duodenal atresia or stenosis, DO1) occurs in 1 in 6,000 live births and manifests as bilious emesis in the newborn period with a classic “double-bubble” on fetal

IP

T

ultrasound or abdominal x-ray.[1] Over 50% of patients have other associated anomalies.[2,3] Congenital heart disease (CHD) is common in this population, impacting 24% of those with normal

CR

karyotypes and 62% of those with Down syndrome.[4] As CHD causes significant morbidity in these

US

patients,[3] cardiac lesions may substantially impact the neonatal course of DO patients. In 2013, the American Institute of Ultrasound in Medicine (AIUM), in collaboration with the

AN

American College of Obstetricians and Gynecologists, the Society for Maternal Fetal Medicine and the

M

American Society of Echocardiography, recommended that all fetuses diagnosed with extra-cardiac

ED

anomalies on fetal ultrasound receive a fetal echocardiogram (ECHO).[5] These recommendations build off the increasing evidence that fetal ECHOs are effective at detecting the majority of critical congenital

PT

heart defects.[6] As DO is often diagnosed prenatally,[7] an increasing number of these patients receive 1

AC

CE

Abbreviations: DO: Duodenal Obstruction (Atresia or Stenosis) CHD: Congenital Heart Disease ECHO: Echocardiogram AIUM: American Institute of Ultrasound in Medicine GA: Gestational Age LOS: Length of Stay PFO: Patent Foramen Ovale PDA: Patent Ductus Arteriosus IQR: Interquartile Range PPV: Positive Predictive Value NPV: Negative Predictive Value ASD: Atrial Septal Defect AVC: Atrioventricular Canal TOF: Tetralogy of Fallot VSD: Ventricular-septal defect ROC: Receiver operating characteristic 3

ACCEPTED MANUSCRIPT fetal echocardiograms and are diagnosed with cardiac lesions prior to birth. After birth, these patients typically also undergo pulse oximetry evaluation, another screening tool that is effective at diagnosing critical cardiac defects.[8–10] Despite the use of two screening methods, operative repairs of duodenal obstruction are often

T

delayed until a postnatal ECHO confirms the absence of cardiac disease.[11,12] While awareness of an

IP

associated cardiac abnormality may impact intra-operative management of infants with DO, for those

CR

that do not have cardiac disease, the neonatal ECHO often confirms the prior screening exam

US

findings.[13] As a result, recent studies have called into question the regular use of neonatal ECHOs for infants with surgical diseases and negative prior clinical/radiologic exams.[13,14] We hypothesized that

AN

preoperative neonatal ECHOs are unnecessary in DO patients without previously identified congenital cardiac lesions on prenatal imaging. Our aim was to determine the utility of neonatal ECHOs in patients

2. Materials and Methods

ED

M

with duodenal obstruction.

PT

We conducted at retrospective cohort study of all patients with DO (duodenal atresia, web or

CE

symptomatic stenosis) treated at two tertiary care children’s hospitals between January 2005 and February 2016. IRB approval was obtained from the coordinating hospital (STUDY00015885).

AC

Demographic information, including gestational age (GA), surgical technique, age at surgical repair, and other congenital comorbidities were analyzed. Infants were categorized as small for gestational age using the gender-specific classification system for tenth percentile of birth weight.[15] Other clinical course markers were collected, including length of stay (LOS), ventilation status, and post-operative complications, such as administration of antibiotics for presumed infection. All fetal ultrasounds, fetal ECHOs and neonatal ECHOs were reviewed. While fetal ECHOs evaluate more cardiac views than fetal ultrasounds,[5,16] fetal ultrasound findings may be used in those

4

ACCEPTED MANUSCRIPT instances where infants did not receive fetal ECHOs. As a result, the cardiac findings from both modalities (fetal ultrasound and fetal ECHO) were collapsed into one category termed “prenatal imaging findings,” and a positive finding was recorded if either modality noted a cardiac lesion. Prenatal imaging findings were then compared to neonatal ECHOs to determine concordance and discordance between

T

these methods.

IP

Cardiac disease was classified as significant if imaging had evidence of a ductal-dependent lesion, or

CR

a non-ductal-dependent cyanotic lesion. Small-to-moderate patent foramen ovale (PFO) and small-to-

US

moderate patent ductus arteriosus (PDA) were classified as normal, given that these defects are part of normal fetal circulation, are often present in the early neonatal period, and typically resolve without

AN

medical intervention.[17] All medications to manage cardiac lesions were documented and charts were

M

examined for cardiac surgery both during the neonatal period and after discharge. All analyses were completed in STATA 14.0 and numerical variables were summarized by median

ED

and interquartile range (IQR). Binary variables were summarized by percent, and binomial exact

PT

analyses were completed to determine sensitivity, specificity, positive predictive value (PPV), and

3. Results

CE

negative predictive value (NPV), with neonatal ECHO findings used as an indication of true disease.

Table 1

AC

We identified 65 newborns diagnosed with DO. There were 30 females and 35 males. The median gestational age at birth was 37 weeks (Table 1). Twelve patients (19.7%) had associated cardiac lesions. Forty-one patients (63.1%) had non-cardiac related congenital anomalies in addition to their diagnosis of DO, including 5 patients (7.7%) with malrotation. Twenty (30.8%) patients had Down syndrome, 5 of whom (25%) had associated cardiac lesions (Tetralogy of Fallot [TOF], atriovenous canal [AVC], atrial septal defect [ASD] with aneurysmal atrial septum, and ventriculoseptal defect [VSD] x 2).

5

ACCEPTED MANUSCRIPT Sixty-one patients (93.8%) received prenatal ultrasounds, 16 (26.2%) of which were performed at referring hospitals (Figure 1). Of these, 80.3% were abnormal, with the most common findings being double bubble (67.2%) and polyhydramnios (52.5%). Ten ultrasounds demonstrated cardiac anomalies. Twenty patients (30.7%) then underwent fetal ECHO, all because of cardiac or extra-cardiac

T

abnormalities identified in the screening prenatal ultrasound. New cardiac lesions were identified on

IP

fetal ECHO in 4 patients. In total, 14 patients had cardiac lesions diagnosed prenatally. Overall, prior to

CR

the AIUM recommendations[5], 40% of tertiary center patients received a fetal ECHO following findings of abnormalities on the screening ultrasound exam. After the 2013 recommendation release,

US

compliance increased to 63.6%.

AN

Postnatally, 56 (91.8%) patients underwent ECHO, 3 of whom had no prior prenatal imaging. Of these 44 (78.6%) were negative for any cardiac anomaly and 12 were positive for cardiac lesions. All

M

patients found to have CHD on neonatal ECHO were identified as having cardiac lesions on prenatal

ED

ultrasound or prenatal ECHO. Six patients who had normal fetal ECHOs underwent postnatal ECHOs;

PT

this imaging identified no new cardiac disease. Surgical correction of DO was performed in all patients. Median age at day of repair was 2 days (IQR:

CE

1, 3). Fifteen patients were repaired after day 3 of life; 13 of these delays to surgery were related to a delay in diagnosis, one was due to the patient also having a tracheoesophageal fistula (which was

AC

Figure 1

repaired first and the patient allowed to recover) and one was 400 grams at birth and surgical repair was delayed until she was 800 grams. No delays to surgery were related to cardiac lesions. Seventeen patients underwent successful laparoscopic repair, including 3 with cardiac anomalies. Three additional patients had attempted laparoscopic repairs that were converted to open operations (one air embolus, one concern for bleeding and one poor visualization); none of these patients had cardiac anomalies. The surgery of the patient with a giant omphalocele was complicated by subcapsular

6

ACCEPTED MANUSCRIPT injury to the liver and intraoperative hemorrhage and hypotension, which resulted in administration of dopamine during intraoperative resuscitation, and was continued post-operatively. Postoperatively, 13 patients (20%) had complications. Two patients developed a wound infection, one patient had an anastomotic leak and one patient developed Bell’s class I necrotizing enterocolitis. Three patients had

T

mechanical line related complications (infiltration, thrombus and dislodgement) and 6 patients were

IP

treated with antibiotics for documented or presumed sepsis. There was one mortality in this group; a

CR

patient who also had a giant omphalocele, a large VSD and significant pulmonary hypertension who

were discharged after a median 21 days (IQR: 15, 29).

US

experienced an intraoperative liver capsular tear and postoperative multisystem organ failure. Patients

AN

Table 2 describes the findings of these patients’ neonatal and fetal imaging, along with their clinical management for their cardiac lesions. Four patients were managed medically with diuretics and/or inotropes/vasopressors. All patients received diuretics outside of the perioperative period. One

M

Table 2

ED

patient received dopamine as a result intraoperative hemorrhage (as described above). All other inotropes/vasopressors were administered outside of the perioperative period. Two patients underwent

PT

surgical correction at 5 and 7 months, respectively, and 5 required no management. There were no

CE

changes in anesthesia management as a result of new findings of cardiac disease. Overall, no changes to

AC

planned anesthesia occurred because of CHD diagnosis. Prenatal imaging, ECHO and ultrasound combined, was compared to confirmed cardiac lesions based on neonatal imaging to determine the sensitivity, specificity, NPV and PPV of prenatal imaging Table 3

(Table 3). The sensitivity of prenatal imaging was 100% (95% Confidence Interval (CI): 73.5, 100.0), with a NPV of 100% (95% CI: 91.0, 100.0). The specificity of prenatal imaging was 95.0% ( 95% CI: 83.5, 99.4), with a PPV of 85.7% (95% CI: 57.2, 98.2). 4. Discussion

7

ACCEPTED MANUSCRIPT Our study demonstrates that prenatal imaging (ultrasound and ECHO) is an effective screening tool for cardiac lesions in patients with DO. These imaging modalities identified all clinically significant cardiac lesions, with most (66.7%) found on screening ultrasound during the anatomic screen in the second trimester. The remaining 33.3% were identified through fetal echocardiograms after extracardiac

T

disease was found on ultrasound. Overall, given the high sensitivity and negative predictive value of

IP

prenatal imaging, we find that use of AIUM recommendations would ensure that all patients with

CR

cardiac lesions are identified in the prenatal period. Thus, despite previous studies recommending routine postnatal ECHOs for all infants with DO,[11] our findings corroborate more recent evidence that

US

demonstrates both prenatal imaging and pulse oximetry provide sufficient screening for CHD.[8–

AN

10,13,14]

Neonatal ECHO is unlikely to identify any new patients with CHD, and may instead delay surgery 2 to

M

3 days. This delay interrupts other aspects of clinical care, such as initiation of enteral feeds. As DO

ED

patients already have adequate prenatal screening for CHD, these delays are unwarranted in clinically

PT

stable patients without evidence of cardiac disease. Similarly, neonatal ECHOs in patients with known cardiac lesions did not change the intraoperative

CE

management of any patients in our study. However, we recognize that, in patients with CHD, this

AC

imaging may provide helpful information to the anesthesiologist should an infant become unstable intraoperatively. In addition, as Table 2 demonstrates, neonatal ECHOs were more specific than prenatal imaging for the type of cardiac lesion, which may in turn impact neonatal cardiac management. As a result, we recognize that neonatal ECHOs will continue provide information vital to the care of infants with known cardiac lesions and DO. Even so, no preoperatively stable patients experienced intraoperative cardiac instability. Thus, given the questionable utility of the neonatal ECHO findings in

8

ACCEPTED MANUSCRIPT the perioperative period, the anesthesia, neonatal, surgical and cardiac teams should discuss whether preoperative ECHOs are warranted. Four patients had cardiac lesions that were not identified on fetal ultrasound but were found on fetal ECHO. This underscores the concerns raised over the user-dependent nature of both

T

ultrasonography and echocardiography. It is well known that routine obstetric ultrasonography is less

IP

sensitive for cardiac lesions than fetal echocardiography, as it only examines 3 cardiac views (a 4-

CR

chamber view of the heart and its outflow tracts) rather than the ≥9 views used in fetal ECHO.[5,6,18]

US

However, fetal ultrasonography is effective at identifying extracardiac lesions like duodenal obstruction.[7,19] Our findings mirrored these trends, as 80% of infants undergoing fetal ultrasound had

AN

abnormal findings, including 67% with duodenal obstruction. While our study examined two different tertiary care centers, one academic hospital and one private, there was little discrepancy between the

M

two hospitals’ prenatal imaging findings. In addition, approximately one-fourth of these infants

ED

underwent fetal screening ultrasounds at community hospitals, where upon finding evidence of duodenal obstruction, these institutions appropriately referred pregnant mothers to tertiary care

PT

centers for further evaluation. This underscores the importance of following the 2013 AIUM

CE

recommendations, which state that fetal echocardiograms should be performed in all patients found to have extracardiac lesions on prenatal ultrasound.[5] Despite only having 64% compliance with these

AC

recommendations, fetal ECHOs played a vital role in capturing all cardiac lesions prenatally. While fetal echocardiography is known to be user dependent,[6] new techniques, like spatiotemporal image correlation, and an increased training of ultrasonographers has significantly improved the accuracy of this imaging modality. As a result, reference ranges have been developed for valve areas, and ECHOs can now utilize volume data to diagnose lesions like heterotaxy and identify small structures like papillary muscles. When used in high-risk populations, spatiotemporal image

9

ACCEPTED MANUSCRIPT correlation was found to be 91.6% sensitive for CHD, and 4D ultrasound shown to have a sensitivity of 93% with strong inter-institution agreement.[17,20] While the added benefit these methods may be under debate, the development of these techniques demonstrate the increased standardization and accuracy of fetal echocardiography.

T

There are several limitations to our study. First and foremost, this is a retrospective study with a

IP

relatively small sample size. Due to the rare nature of the disease, we only had 65 patients with DO

CR

treated at our two hospitals over the 10 years of the study, and our findings were limited to chart review

US

and the description of care in clinical notes. Additionally, because of the relatively long duration of records reviewed, only the imaging reports could be examined, as not all of the actual images were

AN

available for review. We also did not review other adjuncts, such as chest x-ray, that could potentially be used to identify infants with CHD, nor do we know with certainty that those patients in whom we could

M

not find record of prenatal ultrasound did not have the study performed at an outside hospital.

ED

Maternal obesity, known to result in poor image quality of both fetal ultrasound and

PT

echocardiogram,[18,21,22] was not included in our analysis. Lastly, because of the user-specific nature of ultrasound, our results may be biased to findings

CE

specific to our institution based on our ultrasonographers. However, nearly one quarter of our patients

AC

had anomalies initially identified at referring hospitals, which demonstrates the reliability of obstetric ultrasonography at identifying congenital disease in the prenatal period. We encourage other centers to examine the accuracy of fetal ultrasound in their own communities and support the adoption AIUM guidelines to ensure that all patients with extracardiac disease are appropriately referred to tertiary care centers for fetal ECHO.

10

ACCEPTED MANUSCRIPT

5. Conclusions Overall, we found that fetal imaging meets the criteria of an effective screening tool both in its

T

sensitivity and ability to detect asymptomatic patients with cardiac lesions. Thus, DO patients without

IP

cardiac symptomatology in whom no cardiac lesions were identified prenatally do not need

AC

CE

PT

ED

M

AN

US

CR

preoperative evaluation by neonatal ECHO, as additional imaging is unlikely to identify new lesions.

11

ACCEPTED MANUSCRIPT Table 1. Demographics (n=65) 37 (35, 39)

Gestational Age (Weeks), M (IQR) Ventilation Days, M(IQR)

2 (1, 4)

Age at DO Repair (Days), M (IQR)

2 (1,3) 53.9%

Small for Gestational Age, %

32.3%

Other Non-Cardiac Comorbidities, %

63.1%

Trisomy 21, %

30.8%

Technique, (% open)

73.8%

IP

Gender, (% male)

T

21 (15, 29)

LOS (Days) , M (IQR)

AC

CE

PT

ED

M

AN

US

CR

*Length of stay (LOS), Median (M), Interquartile Range (IQR)

12

ACCEPTED MANUSCRIPT

Changes in Anesthesia Management

None

T

Table 2. Management of Neonates with Congenital Heart Disease (n=12) Neonatal Fetal Fetal Additional Echocardiogram: Ultrasound Echocardiogram Non-Cardiac Cardiac Lesions Findings Findings Comorbidities TOF, Pulmonary Artery Malformation, Polyhydramnios, Down Tricuspid Valve TOF DO, AV Canal Syndrome thickening with regurgitation AVC, ASD, Bicuspid ASD, Aorta, Large PDA, Polyhydramnios, Down Aneurysmal Tricuspid DO Syndrome Atrial Septum Regurgitation

Medical Management: Diuretic

Congenital Omphalocele, Annular Pancreas

Dopamine administration (secondary to Intraoperative Liver Injury)

Medical Management: Inotrope/ vasopressor, Diuretic

None

None

None

Surgical Correction at 7 months

None

None

None

None

None

Medical Management: Inotrope/ vasopressor, Diuretic Surgical Correction at 5 months

IP

None

DO, Cardiac defect

Heterotaxy, Double Outlet RV, Interrupted IVC, VSD

NBS1, Small RSided Spleen, Malrotation

M

ED

DO, TOF

Polyhydramnios, VSD

TOF

TOF

AC

TOF, Bilateral SVC, Large VSD

US

Malrotation

AN

Interrupted IVC

CE

TOF, VSD, Moderate ASD

DO

PT

Interrupted IVC, Small ASD Heterotaxy, Dextrocardia, Double Outlet RV, Malposed GV, Right dominating AVC, Bilateral SVC, Interrupted IVC, ASD

Small VSD

AVC, ASD, VSD, RV enlargement, Elongation LVOT

CR

Large VSD, RV Polyhydramnios, Dilation/Hypertrophy, Omphalocele, Tricuspid DO Regurgitation

Polyhydramnios, DO, AVC

AVC

CHD Management Medical Management: Inotrope/ vasopressor, Diuretic

Malrotation, Congenital Vertebral Anomaly VACTERL Association, Imperforate Anus, TEF/Esophageal Atresia

Down Syndrome

13

ACCEPTED MANUSCRIPT

N/A

VSD

Polyhydramnios, DO, VSD

VSD

VSD

Polyhydramnios, DO

VSD

VSD

DO, VSD

None

None

None

None

None

None

None

Down Syndrome

None

None

AN

US

CR

IP

T

Moderate ASD, RV Dilation

ASD/VSD, Omphalocele, Absent Right Kidney, Absent Right Arm, Small Tibia, Club Foot, Polydactyly

VACTERL Association, Omphalocele, TEF, Imperforate Anus, Bilateral Horseshoe Kidney, Club Foot, Polydactyly, Congenital Vertebral Anomaly Down Syndrome, TEF, Finger Contracture

VSD

AC

CE

PT

ED

M

*Congenital Heart Disease (CHD), Tetralogy of Fallot (TOF), Duodenal obstruction (DO), Ventricular-septal defect (VSD), Atrioventricular Canal (AVC), Atrial Septal Defect (ASD), Right Ventricular (RV), Superior Vena Cava (SVC), Inferior Vena Cava (IVC), Great Vessels (GV), Left Ventricular Outflow Tract (LVOT), Tracheoesophageal Fistula (TEF)

14

ACCEPTED MANUSCRIPT Table 3. Sensitivity of Prenatal Imaging for True Cardiac Defects (n=53) No Cardiac Lesion found on Neonatal Imaging

Measures

Positive Prenatal Imaging

12

2

PPV= 85.7% (95% CI: 57.2, 98.2)

Negative Prenatal Imaging

0

39

NPV= 100.0% (95% CI: 91.0, 100.0)

Sensitivity= 100.0% (95% CI: 73.5, 100.0)

Specificity= 95.1% (95% CI: 83.5, 99.4)

Measures

T

Cardiac Lesion found on Neonatal Imaging

AC

CE

PT

ED

M

AN

US

CR

IP

*Positive Predictive Value (PPV), Negative Predictive Value (NPV), Confidence Interval (CI)

15

Figure 1