CawV

Onal

Ssrecvt

R8thc

Edtnca2fin

FGC

POVCD

Ra

RAa

0©d

Residents’ Papers www.AJOG.org

e

hange in quantitative human chorionic gonadotropinfter manual vacuum aspiration in womenith pregnancy of unknown location

eronica Rivera, MD; Phuong H. Nguyen, MD, MMM; Anita Sit, MD, MPH

BJECTIVE: To determine whether the change in human chorionic go-adotropin after manual vacuum aspiration is predictive of an earlybnormal intrauterine pregnancy in women with pregnancy of unknownocation.

TUDY DESIGN: This is a prospective cohort study of 23 clinicallytable patients with an early abnormal pregnancy who had abnormallyising human chorionic gonadotropins and absence of sonographicvidence of an intrauterine pregnancy or ectopic pregnancy. Thehange in human chorionic gonadotropin within 24 hours after manualacuum aspiration was compared with the pathologic diagnosis andhe ultimate clinical diagnosis.

ESULTS: Ten patients had � 50% decrease (mean, 74%; range, 58-0%) in human chorionic gonadotropin after manual vacuum aspira-ion with confirmed chorionic villi on pathology results. Two patientsad a � 50% drop in human chorionic gonadotropin but absence of

proved by the Santa Cloi: 10.1016/j.ajog.2008.10.013

56 American Journal of Obstetrics & Gynecology MAY 2009

ion. The remaining 10 patients who had either rising or � 50% de-rease in human chorionic gonadotropin post manual vacuum aspira-ion all had no chorionic villi on pathology results. The sensitivity,pecificity, positive predictive value, and negative predictive value of a �0% decrease in human chorionic gonadotropin after manual vacuum as-iration in predicting an abnormal intrauterine pregnancy were 92% (95%onfidence interval [CI], 0.62-0.99), 100% (95% CI, 0.62-1.0), 100%95% CI, 0.70-1.0), and 90% (95% CI, 0.54-0.99), respectively.

ONCLUSION: A � 50% decrease in human chorionic gonadotropinithin 24 hours after manual vacuum aspiration is predictive of anbnormal intrauterine pregnancy, thereby excluding an ectopic preg-ancy and expediting the management of women with pregnancy ofnknown location.

ey words: ectopic pregnancy, human chorionic gonadotropin,anual vacuum aspiration, miscarriage, pregnancy of unknown

horionic villi, clinically consistent with complete spontaneous abor- location

ctopic pregnancy is a major cause ofmorbidity and mortality in repro-

uctive-aged women. When a gesta-ional sac is not visualized on transvagi-al ultrasound with quantitative humanhorionic gonadotropin (hCG) levelbove the discriminatory zone (1500-000 mIU/mL),1 the pregnancy is classi-ed as an abnormal intrauterine preg-ancy (IUP) or an ectopic pregnancy

(EP). A suction uterine dilatation andcurettage (D&C) can be performed to as-sist with the diagnosis, because the pres-ence of chorionic villi confirms an IUP.2

Avoiding uterine evacuation and pre-suming a diagnosis of an EP can be inac-curate in almost 40% of cases.3

Manual vacuum aspiration (MVA)has been shown to be as efficacious, safe,and tolerable as a D&C for first-trimesterand incomplete abortions.4 Moreover,waiting for pathologic diagnosis afteruterine evacuation may delay the diag-nosis and treatment of EP.

A decrease of greater than 48% inquantitative hCG within 24 hours aftera medical abortion is highly predictiveof a successful expulsion of an IUP.5,6

We hypothesize that a decrease of atleast 50% in hCG after MVA is predic-tive of an abnormal IUP in patientswith a pregnancy of unknown location(PUL).

MATERIALS AND METHODSThis is a prospective cohort study ap-

Center Institutional Review Board. Pa-tients with PUL as defined by the follow-ing inclusion criteria were recruited: (1)hemodynamically stable and (2) quanti-tative serum hCG level above 2000mIU/mL or if under 2000 mIU/mL withabnormally rising levels less than 66% in48 hours and no evidence of IUP or EPon transvaginal ultrasound by the Radi-ology Department. Exclusion criteriawere as follows: sonographic evidence ofEP or IUP including a gestational sac,heavy vaginal bleeding suggestive of pas-sage of “tissue,” or inability to tolerate anoutpatient MVA.

An MVA was performed outpatientin a standardized fashion and hCG lev-els were obtained before and afterMVA (within 24 � 4 hours) procedure.The change in hCG from before andafter MVA were compared wth thepathologic diagnosis and the ultimateclinical diagnosis. We calculated thesensitivity, specificity, and positive andnegative predictive values of � 50%decrease in hCG compared with the

rom the Department of Obstetrics andynecology, Santa Clara Valley Medicalenter, San Jose, CA.

resented at the American College ofbstetrics and Gynecology Annual District III,III, IX Meeting 2007, Victoria, Britisholumbia, Canada, August 11, 2007, as theistrict IX Award Paper.

eceived July 29, 2008; revised Oct. 1, 2008;ccepted Oct. 7, 2008.

eprints: Anita Sit, MD, MPH, 751 S. Bascomve., San Jose, CA 95128.nita.sit@hhs.co.santa-clara.ca.us.

002-9378/free2009 Mosby, Inc. All rights reserved.

tcts5pc(

Cwanu

Km

ara Valley Medical a

bsence or presence of chorionic villi

aa

RBtc

ptyporc

T1pnwdhzah

dmr(tu5vct1d�vctbcihdtosItaSDp

a

hoR

www.AJOG.org Residents’ Papers

nd the ultimate clinical diagnosis ofn abnormal IUP.

ESULTSetween July 2006 and June 2008, 23 pa-

ients were enrolled. One patient was ex-luded because of failure to follow up for

FIGUREResults and follow-up of patients a

N = 23

N = 22

hCG drop < 50%

or rise N = 10

Surgery

N = 3

MTX

N = 6 a

Continued pregnancy

N = 1

No POC N = 10

Four patients had an increase in post-MVA hCCG, human chorionic gonadotropin, IUP, intrauterine pregnancyf conception; SAB, spontaneous abortion.ivera. Change in hGC after MVA in women with PUL. Am

TABLE 1Sensitivity, specificity, PPV, and NPVariable POC�

� 50% decrease in hCG 10

...................................................................................................................

� 50% decrease in hCG 0

...............

Sensit(95%

...................................................................................................................

CI, confidence interval; hCG, human chorionic gonadotropina Two patients had no chorionic villi, suggestive of passage

Rivera. Change in hGC after MVA in women with PUL. Am J

ost-MVA hCG. The average age of par-icipants was 29 years (range, 21-42ears) with the following ethnicity: His-anic (87%) and white (13%). The rangef gestational age by last menstrual pe-iod was from 5-11 weeks. The MVAannula sizes used ranged from 4-8 mm.

r manual vacuum aspiration

N = 1 Lost to follow-

up (POC+)

No POC N = 2

Abnormal

IUP

+ POC N = 10

hCG drop ≥ 50%

N = 12

Complete

SAB

d 2 had a � 50% decrease in post-MVA hCG., methotrexate; MVA, manual vaccum aspiration; POC, products

stet Gynecol 2009.

of a > 50% decrease in hCG comparPOC�

2a

.........................................................................................................................

10

.........................................................................................................................

� 10/10 (100%)0.66-1.0)

Specificity � 1(95% CI, 0.51-

.........................................................................................................................

, intrauterine pregnancy; NPV, negative predictive value; POC, p

sue of clinical abnormal IUPs as confirmed by the subsequent

Obstet Gynecol 2009.

MAY 2009 Am

he pre-MVA hCG levels ranged from72-14,485 mIU/mL. Seventeen (73%)atients had serial hCG levels that ab-ormally rose � 66% in 48 hours,hereas 6 (27%) patients had no evi-ence of IUP or EP based on pre-MVACG levels above the discriminatoryone. Time interval between before andfter MVA hCG levels ranged between 16ours and 47 hours.Among the 12 patients with � 50%

ecrease in hCG, 10 (46%) had a confir-ation of chorionic villi on pathology

esults, with an average decline of 74%median, 70%; range, 58-80%), consis-ent with an early abnormal IUP (Fig-re). The other 2 patients, despite a �0% decrease in hCG, had no chorionicilli, suggestive of passage of tissue oflinical abnormal IUPs as confirmed byhe subsequent hCG decline. Among the0 patients who did not have a � 50%ecrease in hCG (range, -1.5% to40%), none of them had chorionic

illi. Three of these patients (hCGhange: -8.7%, �9%, and �40%) werereated with laparoscopic salpingectomyased on symptoms with pathologiconfirmation of EP. Among the remain-ng 6 patients, 4 had a rise in post-MVACGs (range, 1.5-16%), whereas, 2 had arop in hCGs of 9% and 17%, respec-ively. All of them received a single dosef methotrexate for presumed EP, withubsequent appropriate decline in hCG.nterestingly, 1 patient on follow-up ul-rasound showed a continued pregnancyfter MVA because of evacuation failure.he eventually underwent a suction&C for an undesired pregnancy. Thisatient had a pre-MVA hCG of 4535

with pathology (n � 22)Values

PPV � 10/12 (83%)(95% CI, 0.51-0.97)

..................................................................................................................

NPV � 10/10 (100%)(95% CI, 0.66-1.0)

..................................................................................................................

2 (83%))

..................................................................................................................

cts of conception; PPV, positive predictive value.

decline.

fte

G an, MTX

J Ob

V ed

......... .........

......... .........

ivityCI,

0/10.97

......... .........

; IUP rodu

of tis hCG

erican Journal of Obstetrics & Gynecology e57

maM

pc(spCv1pvMiC1CTi

COecpotucMadncbsg

wChmwttmwirtMear

sttttcredelpcmnidmcwp

whwottsmwbabswracc

Mittswfirpl

R1Pen2p

Residents’ Papers www.AJOG.org

e

IU/mL with a lack of visualization ofn IUP or EP. Her hCG rose 25% afterVA to 7035 mIU/mL.A � 50% decrease in hCG when com-

ared with the pathologic diagnosis ofhorionic villi had a sensitivity of 100%95% confidence interval [CI], 0.66-1.0),pecificity of 83% (95% CI, 0.51-0.97),ositive predictive value of 83% (95%I, 0.51-0.97), and a negative predictivealue of 100% (95% CI, 0.66-1.0) (Table). The sensitivity, specificity, positiveredictive value, and negative predictivealue of a � 50% decrease in hCG afterVA in predicting a clinically abnormal

ntrauterine pregnancy were 92% (95%I, 0.62-0.99), 100% (95% CI, 0.62-1.0),00% (95% CI, 0.70-1.0), and 90% (95%I, 0.54-0.99), respectively (Table 2).he prevalence of clinical abnormal IUP

n this cohort was 54%.

OMMENTur study presents a novel approach to

valuating patients with PUL by using ahange in 24-hour post-MVA hCG toredict an abnormal IUP. First, insteadf a traditional D&C, our protocol useshe less expensive, outpatient MVA forterine evacuation. Second, using thehange in hCG within 24 hours after

VA allows a more expedient method ofssessing the pregnancy location. In ouresign, the MVA pathologic results wereot compared with D&C curetting, be-ause the MVA has been documented toe just as effective.4 Our data are furthertrengthened by the exclusion of sono-

TABLE 2Sensitivity, specificity, PPV, and NPcompared with clinical diagnosis o

Variable Clinic

� 50% decrease in hCG 12

...................................................................................................................

� 50% decrease in hCG 1a

............

Sens(95%

...................................................................................................................

CI, confidence interval; hCG, human chorionic gonadotropina This patient had a continued IUP after manual vacuum aspRivera. Change in hGC after MVA in women with PUL. A

raphic evidence of a gestational sac, d

58 American Journal of Obstetrics & Gynecology

hich can be mistaken for a pseudo sac.7

onsistent with the literature,3 our co-ort confirmed the finding of approxi-ately 40% of women with PUL whoould have been treated with metho-

rexate if MVA were not performed. Fur-hermore, the majority of patients had

ore than 2 consecutive hCG values,hich failed to have risen minimally us-

ng the new redefined hCG curves (53%ise over 2 days).8,9 Finally, we calculatedhe diagnostic characteristics of post-

VA hCG change in predicting the pres-nce or absence of chorionic villi as wells clinical abnormal IUP, which is moreelevant from a clinical standpoint.

Limitations of our study are themall sample size, the variations in theime interval (range, 16-47 hours) be-ween before and after MVA hCGs, andhe potential of incomplete removal ofrophoblastic tissue. One patient had aontinued pregnancy diagnosed after aise in post-MVA hCG as a failure tovacuate an early IUP. This finding un-erscores the importance of serial hCGvaluations and close clinical fol-ow-up after MVA, regardless of theathologic results. Furthermore, thisase also highlights the potential ofisclassifying a developing IUP as ab-

ormal when an IUP was not visual-zed at the currently recommendediscriminatory zone of 1500-2000IU/mL.10 Moreover, the rate of de-

line of hCG after an induced abortionas highly variable, depending on there-MVA hCG levels. A longer median

of a > 50% decrease in hCGn abnormal IUP (n � 22)

abnormal IUPPresumed ectpregnancy

0

.........................................................................................................................

9

........................................................................................................................

y � 12/13 (92%)0.62-0.99)

Specificity � 9(95% CI, 0.62-

.........................................................................................................................

, intrauterine pregnancy; NPV, negative predictive value; PPV, p

n requiring a suction dilatation and curettage for undesired preObstet Gynecol 2009.

isappearance time was associated s

MAY 2009

ith higher initial level.11 In our co-ort, the pre-MVA hCG levels rangedidely, from 172-14,485 mIU/mL. An-ther potential weakness is that pa-ients with a � 50% drop in hCG werereated with methotrexate for pre-umed EP without a pathologic confir-

ation. Under this protocol, treatingomen with single-dose methotrexateased on lack of villi, the ultimate di-gnosis of EP or abnormal IUP cannote obtained. We acknowledge the pos-ibility of overtreating those 2 patientsith falling post-MVA hCG levels. Se-

ial hCG evaluations would have beenpreferred treatment option, espe-

ially in a population of stable andompliant patients.

Our study has demonstrated a post-VA hCG drop of � 50% within 24 hours

s predictive of an abnormal IUP, thus po-entially decreasing the proportion of pa-ients receiving methotrexate for pre-umed EP. A larger clinical trial isarranted to validate our findings. If con-rmed, our algorithm will add to the cur-ent armamentarium in the evaluation ofatients at risk for EP when the pregnancy

ocation is indeterminate. f

EFERENCES. Barnhart K, Mennuti MT, Benjamin I, et al.rompt diagnosis of ectopic pregnancy in anmergency department setting. Obstet Gy-ecol 1994;84:1010-5.. Gracia CR, Barnhart KT. Diagnosing ectopicregnancy: decision analysis comparing six

cValues

PPV � 12/12 (100%)(95% CI, 0.70-0.1.0)

..................................................................................................................

NPV � 9/10 (90%)(95% CI, 0.54-0.99)

..................................................................................................................

(100%))..................................................................................................................

ve predictive value.

cy.

Vf a

alopi

......... .........

.......... .........

itivitCI,

/91.0

......... .........

; IUP ositi

iratio gnanm J

trategies. Obstet Gynecol 2001;97:464-70.

3sG4tpS5mfS

6cmG7pS8Su

9por21p1Pa

www.AJOG.org Residents’ Papers

. Barnhart KT, Katz I, Hummel A, et al. Pre-umed diagnosis of ectopic pregnancy. Obstetynecol 2002;100:505-10.. Hemlin J, Moller B. Manual vacuum aspira-ion, a safe and effective alternative in earlyregnancy termination. Acta Obstet Gynecolcand 2001;80:563-7.. Barnhart KT, Bader T, Huang X, et al. Hor-one pattern after misoprostol administration

or a non-viable first-trimester gestation. Fertil

teril 2004;81:1099-105. s

. Creinin MD. Change in serum beta-humanhorionic gonadotropin after abortion withethotrexate and misoprostol. Am J Obstetynecol 1996;174:776-8.. Ahmed AA, Tom BD, Calbrese P. Ectopicregnancy diagnosis and the pseudo-sac. Fertilteril 2004; 81:1225-8.. Barnhart KT, Sammel MD, Rinaudo PF, et al.ymptomatic patients with an early viable intra-terine pregnancy: HCG curves redefined. Ob-

tet Gynecol 2004;104:50-5. 1

MAY 2009 Am

. Seeber BE, Sammel MD, Guo W, et al. Ap-lication of redefined human chorionic gonad-tropin curves for the diagnosis of women atisk for ectopic pregnancy. Fertil Steril006;86:454-9.0. Seeber B, Barnhart K. Suspected ectopicregnancy. Obstet Gynecol 2006;107:399-413.1. Aral K, Gurkan Zorlu C, Gokmen O.lasma human chorionic gonadotropin levelsfter induced abortion. Adv Contracept 1996;

2:11-4.

erican Journal of Obstetrics & Gynecology e59