www.elsevier.com/locate/ajem

American Journal of Emergency Medicine (2012) 30, 1627–1629

Brief Report

Accuracy of emergency physicians using ultrasoundmeasurement of crown-rump length to estimate gestationalage in pregnant females☆

Caitlin Bailey MDa,⁎, Jennifer Carnell MDb, Farnaz Vahidnia MD, PhDa,Sachita Shah MDc, Michael Stone MDa, Mickeye Adams MDd, Arun Nagdev MDa

aDepartment of Emergency Medicine, Alameda County Medical Center, Oakland, CA 94602, USAbDepartment of Emergency Medicine, Baylor College of Medicine, Houston, TX, USAcDivision of Emergency Medicine, Harborview Hospital, Seattle, WA, USAdDepartment of Obstetrics and Gynecology, Alameda County Medical Center, Oakland, CA, USA

Received 2 September 2011; revised 28 November 2011; accepted 2 December 2011

AbstractStudy objective: The objective of this study is to evaluate the accuracy of emergency providers (EPs) ofvarious levels of training in determination of gestational age (GA) in pregnant patients using bedsideultrasound measurement of crown-rump length (CRL).Methods: We conducted a prospective, cross-sectional, observational study of patients in obstetricalcare at an urban county hospital. We enrolled a convenience sample of women at 6 to 14 weeksgestation as estimated by last menstrual period. Emergency providers used ultrasound to measure theCRL. Repeat CRL measurements were performed by either an obstetrical ultrasound technician orsenior obstetrician and used as the criterion standard for true GA (TGA).Results: One hundred five patients were evaluated by 20 providers of various levels of training. Theaverage time required to complete the CRL measurement was 85 seconds. When CRL measurementsperformed by EPs were compared with the TGAs, the average correlation was 0.935 (0.911-0.959).Using standard accepted variance for CRL measurements at different GAs according to the obstetricsliterature (±3 days for 42-70 days and ±5 days for 70-90 days), correlation between EP ultrasound andmeasured TGA was 0.947 (0.927-0.967).Conclusions: Emergency providers can quickly and accurately determine GA in first-trimesterpregnancies using bedside ultrasound to calculate the CRL. Emergency providers should consider usingultrasound to calculate the CRL in patients with first-trimester bleeding or pain because this estimatedGA may serve as a valuable data point for the future care of that pregnancy.© 2012 Elsevier Inc. All rights reserved.

☆ Equipment used: Sonosite M-turbo, Bothell Washington.⁎ Corresponding author.E-mail addresses: Caitlin.h.bailey@gmail.com,

caitlin_bailey@post.harvard.edu (C. Bailey).

0735-6757/$ – see front matter © 2012 Elsevier Inc. All rights reserved.doi:10.1016/j.ajem.2011.12.002

1. Introduction

1.1. Background

Despite the common application of the ultrasonographiccrown-rump length (CRL) measurement by obstetricians and

1628 C. Bailey et al.

ultrasound technicians to calculate gestational age (GA),there are no published studies assessing the ability ofemergency providers (EPs) to accurately measure CRL.Given the rising application of ultrasound in emergencypractice, validation of this skill within this field is vital forquality assurance.

1.2. Significance

Vaginal bleeding in pregnancy, particularly in the firsttrimester, is an extremely common chief complaint [1]. Inunderserved areas with minimal health care access, patientsoften present to the emergency department (ED) withcomplaints of vaginal bleeding without any previous prenatalcare. In addition, many patients are unsure of the timing oftheir last menstrual period. Even if it is known, estimation ofGA based on last menstrual period dating can be inaccurate[2]. Accurate fetal dating during the initial ED visit may becrucial for the care of that patient because many patients donot return for routine prenatal care until later in theirpregnancy. A CRL performed in early pregnancy (up to 12-14 weeks) is more accurate than sonographic dating methodsafter 14 weeks. If a pregnant patient subsequently suffers acomplication of pregnancy, other medical disease, or traumalater in her pregnancy, knowledge of the GA is critical indetermining the safety of early fetal delivery or drug therapy.Accurate dating including the use of ultrasound was alsoshown to significantly reduce the number of pregnanciesconsidered postterm in 1 large retrospective study [3]. Thiscould result in fewer inductions for postdate pregnancies,with associated cost savings for the hospital. Prioremergency ultrasound studies of pregnant patients haveshown that EPs can accurately identify free fluid in theabdomen of the pregnant patient with blunt trauma [4] andestimate GA in later pregnancy [5]. However, no previousstudies have demonstrated the ability of EPs to accuratelyand rapidly measure GA in the first trimester via CRL.

1.3. Goals of investigation

We hypothesized that EPs can rapidly and accuratelyassess GA in early pregnancy using ultrasonographiccalculation of the CRL.

Fig. 1 Measurement of CRL.

2. Methods

2.1. Study design

We conducted a prospective, blinded observational studyof women in their first trimester of pregnancy presenting toour hospital's obstetrical clinic for a routine sonogrambetween March 10, 2010, and September 17, 2010. Writteninformed consent was obtained in all patients. The study wasapproved by our hospital's institutional review board.

2.2. Setting

The study was performed at a large, urban teachinghospital with an emergency medicine residency andultrasound fellowship.

2.3. Participants

A convenience sample of women older than 18 years oldwas enrolled on days when study EPs and an ultrasoundtechnician or senior obstetrician (OBGYN) were available.The 20 participating emergency sonographers includedsenior medical students, emergency medicine residentswho had completed a 1-month required rotation inemergency ultrasound, 3 emergency ultrasound fellows,ED physician assistants, and ultrasound-credentialed EDattending physicians. All providers viewed a 5-minutestandardized video demonstrating CRL measurements pro-duced by the senior investigator (AN). The participatingultrasound technician was Registered Diagnostic MedicalSonographer (RDMS) certified, and the senior OBGYN hadmore than 10 years of ultrasound experience.

2.4. Study protocol

Each emergency sonographer measured CRL with astandard transabdominal curvilinear probe (Sonosite M-turbo; Sonosite, Bothell, WA). The measurements displayedon the screen of the ultrasound machine were covered whilethe examinations were performed. The lead author (AN)recorded the measurements and calculated GA immediatelyafter each examination. Time was measured from placementof the transducer on the patient's abdomen until the point atwhich each measurement was completed.

Immediately after the EP ultrasound, all study patientsunderwent comprehensive sonography by our ultrasoundtechnician or senior OBGYN. This measurement served asthe true GA. All emergency medicine participants as well as

1629Emergency physicians crown-rump length accuracy

the obstetrical ultrasound technician and the OBGYN wereblinded to each other's measurements.

2.5. Ultrasound measurements

A 2- to 5-MHz curvilinear abdominal transducer was usedto measure CRL in the best view possible for the provider. Instandard fashion, the yolk sac and limbs were not included inthe measurements (Fig. 1). Calipers were placed from the topof the head to the best-estimated region of the rump. The EDsonographers were informed not to include the yolk sac orlimbs during the training but were not instructed during theactual measurements. All ultrasounds for the study, includ-ing the comprehensive studies, were performed on aSonoSite M-Turbo ultrasound system.

3. Results

One hundred five patients were evaluated by 20 EPs,including fourth year medical students, residents (R1-R4),physician assistants, ultrasound fellows, and attendingphysicians. At least 5 patients were scanned by practitionersat each level of training; medical students performed 27scans, whereas fellows and attending physicians performed22 and 11 scans, respectively. The average time required tocomplete the CRL measurement was 85 seconds. When CRLmeasurements performed by EPs were compared with thetrue GA, the average correlation was found to be 0.935(0.911-0.959). Using standard variance for CRL measure-ments as published (3 days for 42-70 days and 5 days for 70-90 days [6]), correlation between EP ultrasound andmeasured true GA was 0.947 (0.927-0.967).

4. Discussion

4.1. Limitations

There are several important limitations of our study. First,our subjects were studied in a controlled clinic environment;this limits our ability to extrapolate this process to the settingof a busy ED. Second, transvaginal ultrasounds were notperformed. For fetuses younger than 8 weeks GA, providersmight choose to perform a transvaginal examination for bettervisualization. Although wemight assume that a high degree ofcorrelation would persist with images obtained in this manner,our study does not address this directly. Finally, although weincluded providers with a range of experience with ultrasound(from medical students to ultrasound fellows and attendingphysicians), a relatively small number of providers partici-pated at each level of training. Although a small number of

scans were performed by each provider type, we feel that theshort standardized training and broad range of experienceallows for a moderate level of external validity.

To the best of our knowledge, this is the first study toevaluate the accuracy of EP measurement of CRL to calculateGA in early pregnancy. The high degree of correlationbetween EP CRL measurement and those of obstetricspersonnel suggests that EP measurement of GA in the firsttrimester may be accurate. In addition, our data suggest thatthe CRL measurement can be rapidly performed, onlyminimally prolonging the examination for the busy EP.Emergency provider measurement of the CRL could becomea crucial data point in the progression of a pregnancy forpatients who present to the ED as their first portal of entry intoprenatal care. This is frequently the case in underservedpopulations who do not have their own physicians when theybecome pregnant and for patients who do not know they arepregnant at the time of their ED visit. These patients oftenhave difficulty arranging and maintaining timely obstetricsfollow-up; by the time they are evaluated by an obstetrician,they may be past the window for the use of the CRLmeasurement (N14 weeks) and their sonographic dating,therefore, less accurate. This becomes relevant later inpregnancy when complications may arise and require aconfident assessment of GA for consideration of delivery orthe need for therapeutics such as steroids. In conclusion, ourstudy found that EPs with limited training in sonographicmeasurement of the CRL were able to determine the CRLwith a high degree of accuracy in a short amount of time. Thissuggests that the EP measurement of the CRL may be used asa reliable estimate of delivery date in subsequent decisionmaking regarding a pregnancy initially assessed in ED.

References

[1] Wittels KA, Pelletier AJ, Brown DFM, Camargo Jr CA. United Statesemergency department visits for vaginal bleeding during earlypregnancy, 1993-2003. Am J Obstet Gynecol 2008;198(5):523.e1-6.

[2] Dietz PM, England LJ, Callaghan WM, et al. A comparison of LMP-based and ultrasound-based estimates of gestational age using linkedCalifornia livebirth and prenatal screening records. Paediatr PerinatEpidemiol 2007;21(Suppl 2):62-71.

[3] Mongelli M, Wilcox M, Gardosi J. Estimating the date of confinement:ultrasonographic biometry versus certain menstrual dates. Am J ObstetGynecol 1996;174(1 Pt 1):278-81.

[4] Goodwin H, Holmes JF, Wisner DH. Abdominal ultrasound examina-tion in pregnant blunt trauma patients. J Trauma 2001;50(4):689-93[discussion 694].

[5] Shah S, Teismann N, Zaia B, et al. Accuracy of emergency physiciansusing ultrasound to determine gestational age in pregnant women. Am JEmerg Med 2010;28(7):834-8.

[6] Hadlock FP, Shah YP, Kanon DJ, Lindsey JV. Fetal crown-rumplength: reevaluation of relation to menstrual age (5-18 weeks) with high-resolution real-time US. Radiology 1992;182(2):501-5.