OB JOURNAL CLUBFebruary 20th, 2019

Alexandra Meloccaro, MD

PGY-2

EFFECT OF IBUPROFEN VS ACETAMINOPHEN ON POSTPARTUM HYPERTENSION IN PREECLAMPSIA WITH SEVERE FEATURES: A DOUBLE-MASKED, RANDOMIZED CONTROLLED TRIAL

Blue NR, Murray-Krezan C, DrakeLavelle S, et al

AJOG June 2018

Hypertensive disorders in pregnancy

■ Important contributor to maternal morbidity & mortality

■ Women diagnosed with a hypertensive disorder of pregnancy are more likely to:

– Meet criteria for antihypertensive therapy

– Require increased length of stay postpartum

– Increased risk of cesarean section

Chronic NSAID Use

■ Increases risk of development of hypertension in healthy, nonpregnant women

■ Antagonizes the effect of some antihypertensives after use for just a few days

– Alteration of aldosterone metabolism

– Sodium retention

– Inhibition of prostaglandin-mediated vasodilation

– Production of vasoactive metabolites of arachidonic acid via cytochrome-P450

induction

NSAIDs vs Acetaminophen

NSAIDs

■ More effective in alleviating

postpartum perineal pain

■ Decrease opioid use post cesarean

Acetaminophen

■ Decreases opioid use post

cesarean

■ Contraindicated with severe, acute

elevation of LFTs

Opioids

■ Can lead to dependence

■ Slow CNS function to infant via breastmilk

Hypothesis

■ Study performed to evaluate the effect of ibuprofen vs acetaminophen on

postpartum BP control among women with preeclampsia with severe features

■ Hypothesis: Ibuprofen will increase the duration of severe range hypertension in

women with preeclampsia with severe features

Study Endpoints

■ Primary outcome: Duration of severe range hypertension

– Defined as time from delivery to the last BP ≥ 160/110

■ Secondary outcomes

– Time from delivery to last BP ≥ 150/100

– Postpartum mean arterial pressure

– Any BP ≥ 160/110

– Need for antihypertensive medication at discharge

– Prolongation of hospital stay for BP control

– Length of postpartum stay

– Postpartum use of short acting antihypertensives for acute BP control

– Need for opioid use stratified by postpartum day

– Time to BP control in with severe range hypertension

Materials & Methods

■ Double-masked, randomized control trial

■ Women meeting criteria were approached about study prior to delivery or within 6

hours after delivery

■ Eligibility

– Age > 18yo, admitted to study center Oct 29, 2016 – Nov 1, 2017

– Diagnoses

■ Preeclampsia with severe features

■ Chronic hypertension with superimposed preeclampsia with severe features

■ HELLP

■ Included women with these diagnoses even if presentation did not include severe

range hypertension

Materials & Methods

■ Exclusion criteria

– ALT or AST > 200mg/dL

– Cr > 1.0 mg/dL

– Chronic kidney disease

– Chronic liver disease

– Infectious hepatitis

– GERD or peptic ulcer disease

– Bleeding disorder

– Known sensitivity to NSAID or acetaminophen

– Current incarceration

Randomization & Follow Up

■ Randomized within 6hrs postpartum

– Either 600mg Ibuprofen q6hrs OR 650mg acetaminophen q6hrs

– Started immediately after delivery and continued throughout hospitalization

■ Cesarean delivery

– Neither group administered ketorolac

– Both groups administered 1000mg IV acetaminophen in PACU

■ Opioid analgesics for breakthrough pain PRN, upon request

Randomization & Follow Up

■ Preeclampsia with severe features

– 24hrs postpartum MgSO4

– Observed inpatient 72hrs postpartum

– Not discharged until > 24hrs after SBP ≥ 160 or DBP ≥ 110

– BP measured qhourly x 24hrs immediately postpartum & q4hrs thereafter

– Scheduled antihypertensive therapy for more than 1 BP ≥ 150/100

■ PO Labetalol OR Nifedepine XR

– Acutely treated ANY BP ≥ 160/110

■ PO labetalol or short acting nifedepine

– Daily labs followed

– Any changes in labs or new neuro findings led to automatic withdrawal

RESULTS

Discussion

■ The primary outcome measure of duration of severe-range HTN was chosen

because it conveys the effect of elevated BP on a woman’s postpartum care more

accurately than other outcome measures.

■ The clinical significance and implications of a BP ≥ 160/110 mm Hg occurring 6

hours after delivery is different from when it occurs 72 hours after delivery, as the

woman with severe range HTN 72 hours after delivery is considered to have more

persistent, severe disease, and is more likely to require prolonged observation and

treatment with scheduled antihypertensive medication.

Strengths

■ Double masked, randomized control trial

■ Intention to treat analysis

– Includes every subject who is randomized according to randomized treatment

assignment. It ignores noncompliance, protocol deviations, withdrawal, and

anything that happens after randomization.

■ Q6hr dosing rather than PRN limited potential for type 2 error based on ibuprofen

dosing schedule

■ Generalizable results due to proportion of population with clinically significant

postpartum hypertension

Weaknesses

■ Small study size limits rare complications

■ Superiority design does not allow to infer equivalence between groups

– Chosen because objective was to determine a difference rather than prove

equivalence

– Noninferiority studies require larger study size

■ Possible limitation in generalizability due to unique racial and ethnic population of

study

– High proportion of Hispanic and Native American participants

– Low proportion of black participants

Conclusion

■ Unable to identify any detrimental effect of ibuprofen on postpartum BP

■ Further studies needed for diagnosis of preeclampsia without severe features

■ Larger trial needed to determine noninferiority and exclude differences in rare,

adverse outcomes

Study Power

■ Type I Error: Reject the null hypothesis when the null hypothesis is true (α)

■ Type II Error: Failure to reject the null hypothesis when the null hypothesis if false (β)

■ Power = 1 – β … Stated many different ways

– Probability of NOT making a type 2 error

– Probability of rejecting the null hypothesis when the null hypothesis is false

– Probability of making a correct decision (to reject null) when null is false

– Probability that a test of significance will pick up on an effect that is present

– Probability that a test of significance will detect a deviation from the null hypothesis, should such a deviation exist.

Study Power

■ Increase power by

– Increased sample size >>> narrower distribution area

– Decrease variability in patient population

– Increase α >>> increases probability of Type I error

Discussion

Sources

Blue NR, Murray-Krezan C, DrakeLavelle S, et al. Effect of ibuprofen vs acetaminophen

on postpartum hypertension in preeclampsia with severe features: a double-masked,

randomized controlled trial. Am J Obstet Gynecol 2018;218:616.e1-8.

Evidence Based Medicine. (n.d.). Retrieved January 19, 2019,

https://www.med.uottawa.ca/sim/evidence_based_medicine_e.html

Gupta SK. Intention-to-treat concept: A review. Perspect Clin Res 2011;2:109-12

Original Research ajog.org

OBSTETRICS

Effect of ibuprofen vs acetaminophen on postpartumhypertension in preeclampsia with severe features:a double-masked, randomized controlled trial

Nathan R. Blue, MD; Cristina Murray-Krezan, MS; Shana Drake-Lavelle, BS; Daniel Weinberg, MD;Bradley D. Holbrook, MD; Vivek R. Katukuri, MD; Lawrence Leeman, MD, MPH; Ellen L. Mozurkewich, MD, MS

BACKGROUND: Nonsteroidal antiinflammatory drug use has been postpartum ibuprofen or acetaminophen for first-line pain control.

shown to increase blood pressure in nonpregnant adults. Because of this,

the American College of Obstetricians and Gynecologists suggests

avoiding their use in women with postpartum hypertension; however,

evidence to support this recommendation is lacking.

OBJECTIVE: Our goal was to test the hypothesis that nonsteroidal

antiinflammatory drugs, such as ibuprofen, adversely affect postpartum

blood pressure control in women with preeclampsia with severe features.

STUDY DESIGN: At delivery, we randomized women with pre-

eclampsia with severe features to receive around-the-clock oral dosing

with either 600 mg of ibuprofen or 650 mg of acetaminophen every 6

hours. Dosing began within 6 hours after delivery and continued until

discharge, with opioid analgesics available as needed for breakthrough

pain. Study drugs were encapsulated in identical capsules such that pa-

tients, nurses, and physicians were masked to study allocation. Exclusion

criteria were serum aspartate aminotransferase or alanine aminotrans-

ferase >200 mg/dL, serum creatinine >1.0 mg/dL, infectious hepatitis,

gastroesophageal reflux disease, age<18 years, or current incarceration.

Our primary outcome was the duration of severe-range hypertension,

defined as the time (in hours) from delivery to the last blood pressure

�160/110 mm Hg. Secondary outcomes were time from delivery to last

blood pressure �150/100 mm Hg, mean arterial pressure, need for

antihypertensive medication at discharge, prolongation of hospital stay for

blood pressure control, postpartum use of short-acting antihypertensives

for acute blood pressure control, and opioid use for breakthrough pain. We

analyzed all outcome data according to intention-to-treat principles.

RESULTS:We assessed 154 women for eligibility, of whom 100 met

entry criteria, agreed to participate, and were randomized to receive

Cite this article as: Blue NR, Murray-Krezan C, Drake-Lavelle S, et al. Effect of ibuprofen vs acetaminophen

on postpartum hypertension in preeclampsia with severe

features: a double-masked, randomized controlled trial.

Am J Obstet Gynecol 2018;218:616.e1-8.

0002-9378/$36.00ª 2018 Elsevier Inc. All rights reserved.https://doi.org/10.1016/j.ajog.2018.02.016

Related editorial, page 547.

616.e1 American Journal of Obstetrics & Gynecology JUNE 2018

Seven patients crossed over or did not receive their allocated study

drug, and 93 completed the study protocol in their assigned groups.

We found no differences in baseline characteristics between groups,

including mode of delivery, body mass index, parity, race, chronic

hypertension, and maximum blood pressure prior to delivery. We did

not find a difference in the duration of severe-range hypertension in

the ibuprofen vs acetaminophen groups (35.3 vs 38.0 hours, P ¼ .30).

There were no differences between groups in the secondary outcome

measures of time from delivery to last blood pressure �150/100 mm

Hg, postpartum mean arterial pressure, maximum postpartum systolic

or diastolic blood pressures, any postpartum blood pressure �160/

110 mm Hg, short-acting antihypertensive use for acute blood pres-

sure control, length of postpartum stay, need to extend postpartum

stay for blood pressure control, antihypertensive use at discharge, or

opioid use for inadequate pain control. In a subgroup analysis of pa-

tients who experienced severe-range hypertension, the mean time to

blood pressure control in the acetaminophen group was 68.4 hours

and ibuprofen group was 56.7 hours (P ¼ .26). At 6 weeks post-

partum, there were no differences between groups in the rates of

obstetric triage visits, hospital readmissions, continued opioid use, or

continued antihypertensive use.

CONCLUSION: The first-line use of ibuprofen rather than acetamin-

ophen for postpartum pain did not lengthen the duration of severe-range

hypertension in women with preeclampsia with severe features.

Key words: blood pressure control, ibuprofen, nonsteroidalantiinflammatory drugs, postpartum pain control, preeclampsia

IntroductionHypertensive disorders are importantcontributors to maternal morbidity andmortality, and women diagnosed with ahypertensive disorder of pregnancy arelikely to meet criteria for postpartum

antihypertensive therapy as well asrequire several days of postpartumobservation.1-3 In addition, women withhypertensive disorders of pregnancy areat increased risk for cesarean delivery,with its attendant need for postoperativepain relief.4 Chronic use of nonsteroidalantiinflammatory drugs (NSAIDs),particularly cyclooxygenase inhibitors, isknown to increase the risk of develop-ment of hypertension (HTN) in healthy,nonpregnant women as well as toantagonize the effects of some antihy-pertensive drugs in hypertensive patientsreceiving treatment after just a few daysof NSAID use.5-13 The hypothesizedmechanisms for this effect include the

NSAID-mediated alteration of aldoste-rone metabolism,14 sodium retention,15

inhibition of prostaglandin-mediatedvasodilation,16 and production of vaso-activemetabolites of arachidonic acid viacytrochrome-P450 induction.17

Because of this concern, the AmericanCollege of Obstetricians and Gynecol-ogists (ACOG) suggests avoiding thepostpartum use of NSAIDs in womendiagnosed with preeclampsia who havepostpartum HTN, though evidence insupport of this recommendation islimited.18 Animal studies are limited to asingle investigation in a rat model ofpreeclampsia, which showed that indo-methacin had no effect on blood

AJOG at a Glance

Why was this study conducted?We sought to generate evidence for or against American College of Obstetriciansand Gynecologists (ACOG) recommendation to avoid nonsteroidal antiin-flammatory drugs in women with postpartum hypertension (HTN).

Key findingsPostpartum ibuprofen use did not adversely affect any metric of postpartumblood pressure control, including the duration of severe-range HTN, need forantihypertensive medication at discharge, or need to extend hospital stay forblood pressure control.

What does this add to what is known?The ACOG recommendation to avoid nonsteroidal antiinflammatory use inwomen with postpartum HTN is not based on substantive clinical evidence.

ajog.org OBSTETRICS Original Research

pressure (BP) while rats were still preg-nant.19 In human beings, the limiteddata on the influence of NSAIDs onpostpartum BP among women withpreeclampsia are conflicting.3,20-23

NSAIDs are ideally suited for post-partum pain control and are still widelyused for postpartum and postcesareandelivery pain management in non-hypertensive women because of theireffectiveness.24-26 They are more effec-tive than acetaminophen to alleviate painfrom obstetric perineal injury, and havealso been shown to decrease opioid useafter cesarean delivery.21,24,25,27-29 Addi-tionally, the analgesic alternatives tocyclooxygenase inhibitors have signifi-cant risks. The use of postpartum opi-oids may predispose to ongoing opioiddependence and is associated withneonatal central nervous systemdepression during breast-feeding.Although acetaminophen is an addi-tional nonopioid alternative that maydecrease the use of postcesareanopioids its use is contraindicated in thesetting of severe, acute elevation of liverenzymes, a common occurrence amongpatients with severe variants of pre-eclampsia.30-32

It remains unclear whether the theo-retical adverse effect of short-termNSAID use on postpartum BP is clini-cally significant and justifies forgoingtheir analgesic properties. We conductedthis study to evaluate the effect ofibuprofen vs acetaminophen on post-partum BP control among women withpreeclampsia with severe features. We

hypothesized that ibuprofen would in-crease the duration of severe-range HTNin women with preeclampsia with severefeatures.

Materials and MethodsDetails of ethics approvalThis study was approved by HumanResearch Protections Office of the Uni-versity of New Mexico (Albuquerque,NM) and was registered at clinicaltrials.gov prior to study onset (study identi-fier: NCT02911701). All enrolled par-ticipants provided written informedconsent.

EligibilityWe performed a double-masked, ran-domized controlled trial of ibuprofen vsacetaminophen for postpartum paincontrol. We assessed for eligibility allwomen aged at least 18 years who wereadmitted to the University of NewMexico Hospital from Oct. 29, 2016,through Nov. 1, 2017, with any of thefollowing diagnoses: preeclampsia withsevere features, chronic HTN withsuperimposed preeclampsia with severefeatures, or HELLP syndrome (hemoly-sis, elevated liver function tests, and lowplatelets), as defined by ACOG.18 Weincluded women with the above di-agnoses even if their presentation did notinclude severe-range HTN as thesewomen are still at risk of significantpostpartum HTN, and make our resultsmore generalizable. Women had to beable to give informed consent in eitherEnglish of Spanish. Exclusion criteria

JUNE 2018 Ameri

were alanine aminotransferase (ALT) oraspartate aminotransferase (AST) >200mg/dL, serum creatinine >1.0 mg/dL,chronic kidney disease, chronic liverdisease, infectious hepatitis, gastro-esophageal reflux disease, peptic ulcerdisease, bleeding disorder, knownsensitivity or allergy to NSAIDs or acet-aminophen, or current incarceration.Women were approached prior to de-livery or within 6 hours after delivery.

Randomization and follow-upWithin 6 hours after delivery, we ran-domized study participants to receivescheduled, around-the-clock dosingwith either 600 mg of ibuprofen or 650mg of acetaminophen every 6 hours,starting immediately after delivery andcontinuing for the duration of theirpostpartum hospitalization. Randomi-zation was performed by our investiga-tional pharmacy using a blockrandomization scheme with block sizesof 10. Patients, nurses, and physicianswere masked to study allocation. Theinvestigational pharmacy executed studydrugmasking by either encapsulating thetablet form of the study drug or bymeasuring and encapsulating theappropriate quantity of ibuprofen oracetaminophen bulk powder andmicrocrystalline cellulose in unmarked,identical capsules. Ibuprofen tabletswere manufactured by Ascend Labora-tories LLC (Parsippany, NJ) and pur-chased from Cardinal Health Inc(Dublin, OH), and acetaminophen tab-lets were manufactured by Major Phar-maceuticals (Livonia, MI) andpurchased from Cardinal Health Inc.Powdered forms of both study drugsalong with the capsule shells andmicrocrystalline cellulose were pur-chased from Fagron Inc (St Paul, MN).Women undergoing cesarean delivery inboth study arms did not receive ketor-olac but were given 1000 mg of intrave-nous acetaminophen in the recoveryroom, and their study drug was initiated6 hours after delivery. All study partici-pants were given opioid analgesics uponrequest for breakthrough pain.

As per our institutional protocol andthe ACOG guidelines, women diagnosedwith preeclampsia with severe features

can Journal of Obstetrics & Gynecology 616.e2

FIGUREEnrollment, randomization, and follow-up

b

b

a

a Women not approached due to speaking language other than English or Spanish, limited availabilityof study personnel, or lack of prepared study drug. b Both patients requesting early discharge wereconsidered to have completed study as they had both received �8 doses of study drug and hadalready achieved blood pressure (BP) control.

Blue et al. Ibuprofen and postpartum blood pressure control in preeclampsia. Am J Obstet Gynecol 2018.

Original Research OBSTETRICS ajog.org

received intravenous magnesium sulfatefor 24 hours after delivery, were observedin-hospital for a minimum of 72 hourspostpartum, and were not dischargeduntil at least 24 hours after having eithera systolic BP�160 or a diastolic BP�110mm Hg. BP was measured hourly forthe first 24 hours postpartum and every4 hours thereafter. We initiated sched-uled oral antihypertensive therapy witheither labetalol or extended-releasenifedipine for >1 BP �150/100 mmHg in a 24-hour period, and acutelytreated any BP �160/110 mm Hg witheither oral labetalol or oral short-actingnifedipine, at the discretion of the careprovider. Daily laboratory samples weredrawn, and women with new-onsetcreatinine >1.1 mg/dL, AST/ALT >250mg/dL, seizure activity, stroke, orfocal neurologic findings requiring

616.e3 American Journal of Obstetrics & Gynecol

neuroimaging were automatically with-drawn and their physicians unmasked totheir study allocation. A data safetymonitoring board was created toconvene in the event of any of thefollowing sentinel events: stroke, seizure/eclampsia, posterior reversible enceph-alopathy syndrome, or death. Partici-pants who received at least 8 doses of thestudy medication were considered tohave completed the study intervention.

Study endpoints and sample sizeconsiderationsOur primary outcome was duration ofsevere-range HTN, defined as the time(in hours) from delivery to the last BP�160/110 mm Hg before discharge. Forour sample size assumptions, weanalyzed a pilot sample from our peri-natal database of 31 gravidas meeting the

ogy JUNE 2018

same inclusion criteria as our studypopulation who had both postpartumibuprofen and acetaminophen madeavailable on an as-needed basis. Wefound the duration of severe-range HTNwas exponentially distributed, with anexponential mean of 35 hours (95%confidence intervals [CI], 25e51 hours).To have 80% power to detect the clini-cally significant difference of 24 hoursbetween study arms with a type I errorrate of 5%, 46 women in each groupwere required for our study. In additionto its previous use in comparing time-driven postpartum outcomes, we chosea difference of 24 hours between groupsbecause this represents 1 additional dayof hospitalization, which we consideredto be generalizable in its clinical signifi-cance.2 To account for study groupcontamination, we obtained approval torecruit 100 women.

Secondary outcomes included timefrom delivery to last BP �150/100 mmHg, postpartum mean arterial pressure(MAP), any BP �160/110 mm Hg, needfor antihypertensive medication atdischarge, prolongation of hospital stayfor BP control, length of postpartumhospital stay, postpartum use of short-acting antihypertensives for acute BPcontrol, need for opioid use stratified bypostpartum day, and time to BP controlin those who experienced severe-rangeHTN. At 6 weeks after delivery, partici-pants were contacted and a question-naire collected to determine the rates ofobstetric triage visits, hospital read-missions, opioid medication use, andoral antihypertensive medication use.

Analysis planWe differentiated between the primaryendpoint of “duration of severe-rangeHTN,” which included all randomizedpatients (including those who neverexperienced severe-range HTN), and thesecondary endpoint of “time to BPcontrol,” which was only analyzed in thesubset of patients who experiencedpostpartum severe-range HTN. Thedistribution of the primary endpoint wasassessed with the Shapiro-Wilk test forexponentiality. Mean durations ofsevere-range HTN and their 95% CIfrom the exponential distribution were

TABLE 1Demographic characteristics

Baseline characteristicIbuprofenn ¼ 50

Acetaminophenn ¼ 50 P value

Maternal age, y, mean (SD) 31.9 (5.9) 30.5 (6.4) .3a

BMI, mean (SD) 36.5 (8.0) 36.4 (7.5) 1.0

Parity, n (%)

Nulliparous 18 (36) 19 (38) .8b

Multiparous 32 (64) 31 (62)

Ethnicity and race, n (%)

Hispanic 20 (40) 20 (40) .8c

White 13 (26) 11 (22)

Black 1 (2) 2 (4)

Asian 0 (0) 2 (4)

Native American 12 (24) 9 (18)

Otherd 4 (8) 6 (12)

Mode of delivery, n (%)

Vaginal 25 (50) 32 (64) .4c

Operative vaginal 1 (2) 1 (2)

Cesarean 24 (48) 17 (34)

Chronic HTN requiringtreatment, n (%)

8 (16) 7 (14) .8a

Need for IV antihypertensivesbefore delivery, n (%)

37 (74) 42 (84) .2b

Maximum SBP beforedelivery, mean (SD)

181 (16) 183 (14) .5a

Maximum DBP beforedelivery, mean (SD)

107 (11) 106 (11) .7a

BMI, body mass index; DBP, diastolic blood pressure; HTN, hypertension; IV, intravenous; SBP, systolic blood pressure.

a Independent 2-sample t test; b c2 Test; c Fisher exact test; d Includes women identifying as biracial, or another race orethnicity not otherwise accounted for.

Blue et al. Ibuprofen and postpartum blood pressure control in preeclampsia. Am J Obstet Gynecol 2018.

ajog.org OBSTETRICS Original Research

calculated for the 2 study arms andcompared with a maximum likelihoodtest. An exponential regression modelwas fitted to duration of severe-rangeHTN that included arm, mode of de-livery, and an interaction between them.Continuous secondary endpoints wereassessed for normality. Means and SDwere calculated for continuous data andcompared between study arms with ttests with the exception of time variables,for which the mean and CIs for expo-nentially distributed data were calcu-lated. Time to BP control among thesubset with severe-range HTN wasassessed by fitting an accelerated failure

time model to compare arms, adjustedfor mode of delivery and included aninteraction between group and mode ofdelivery. Frequencies and percentageswere calculated for categorical data andstudy arms were compared with c2 orFisher exact tests, as appropriate. Theopioid use data were skewed, so the log-transformation of opioid use inmorphine equivalents was used foranalysis. Repeated measures logisticregression was used to compare anyopioid use between arms over post-partum days 0-2, and repeated measureslinear regression was used to comparemean morphine equivalents between

JUNE 2018 Ameri

arms over postpartum days 0-2. Weanalyzed all outcome data according tointention-to-treat principles but the perprotocol analysis set was also analyzed toassess for any bias in the primaryoutcome. All analyses were performed inSAS 9.4 (SAS Institute Inc, Cary, NC)and R 3.4 (R Foundation for StatisticalComputing, Vienna, Austria). The studyprotocol and outcomes were reportedaccording to CONSORT guidelines.

ResultsWe assessed 154 women for eligibility. Ofthese, 100 women met all entry criteriaand consented to participate. They wererandomly assigned to receive ibuprofenor acetaminophen for postpartum paincontrol. Of the randomized participants,93 subjects completed the study ac-cording to the protocol for their assignedgroup: 46 in the ibuprofen group and 47in the acetaminophen arm. Of those whodid not complete the protocol or whocrossed over, details are as follows: in theibuprofen arm, 1 woman did not receiveany study drug because of inability toswallow capsules, and 3 women with-drew to ensure they were receivingibuprofen and were also later givenacetaminophen. In the acetaminophenarm, 1 woman was automatically with-drawn for postpartum ASTand ALT thatincreased >250 mg/dL, and 2 otherswithdrew voluntarily to ensure they werebeing given ibuprofen, later receivingboth ibuprofen and acetaminophen. Ineach group, 1 patient requesteddischarge on the day before the standardpostpartum observation period of 72hours had elapsed. Both patients hadreceived �8 study drug doses and hadalready achieved adequate BP control, sowere considered to have completed thestudy. Per our intention-to-treat plan, allwomen who withdrew or crossed overwere included in the final analysis. Thisis summarized in the Figure. None of thesentinel safety events occurred, so thedata safety monitoring board did notconvene.

Among study participants, the meanmaternal age was 31.2 years (SD 6.1).The ethnic/racial distribution was asfollows: Hispanic, 40%; White, 24%;Native American, 20%; Black, 3%;

can Journal of Obstetrics & Gynecology 616.e4

TABLE 2Postpartum outcomes

OutcomeIbuprofenn ¼ 50

Acetaminophenn ¼ 50 P value

Duration of severe-range HTN, h,mean (95% CI)a

35.3 (27.2e47.5) 38.0 (29.4e51.3) .3b

Time from delivery to last BP �150/100mm Hg, h, mean (95% CI)

58.3 (48.0e68.6) 57.1 (45.8e68.5) .9b

Time to BP control, h, mean (95% CI)c 56.7 (45.1e71.2) 68.8 (54.0e86.7) .3d

Postpartum MAP, mean (SD) 97.6 (6.2) 97.3 (9.1) .9e

Maximum postpartum SBP, mean (SD) 168 (16) 165 (15) .3e

Maximum postpartum DBP, mean (SD) 99 (10) 96 (10) .2e

Any postpartum BP �160/110 mm Hg, n (%) 34 (68) 31 (62) .5f

Any postpartum meds for acute BP control, n (%) 30 (60) 26 (52) .4f

Postpartum stay, d, mean (SD) 3.8 (1.4) 4.0 (1.3) .5e

Postpartum stay extended for BP control, n (%) 18 (36) 23 (46) .3f

On antihypertensives at discharge, n (%) 33 (66) 31 (62) .7f

BP, blood pressure; CI, confidence interval; DBP, diastolic blood pressure; HTN, hypertension; MAP, mean arterial pressure; SBP, systolic blood pressure.

a Time from delivery to last BP �160/110 mm Hg; b Maximum likelihood test for exponential means; c Subgroup analysis in only patients who experienced any severe-range HTN, n ¼ 31 inacetaminophen group and n ¼ 34 in ibuprofen group; d Wald c2 test; e Independent 2-sample t test; f c2 Test.

Blue et al. Ibuprofen and postpartum blood pressure control in preeclampsia. Am J Obstet Gynecol 2018.

Original Research OBSTETRICS ajog.org

Asian, 2%; other, 11%. In all, 41 weredelivered by cesarean, 2 by operativevaginal delivery, and 57 vaginally. Be-tween study arms, there were no differ-ences in baseline demographic orobstetric characteristics (Table 1).

For the primary outcome of durationof severe-range HTN, we did not detectany difference between women assignedto receive ibuprofen and those assignedto acetaminophen (35.3 vs 38.0 hours,respectively; P¼.30). There were also nodifferences between groups in secondaryoutcome measures of BP control: timefrom delivery to last BP �150/100 mmHg, postpartum MAP, maximum post-partum systolic BP, maximum diastolicBP, any postpartum BP �160/110 mmHg, need for short-acting antihyperten-sives for acute BP control, length ofpostpartum stay, need to extend post-partum stay for BP control, or require-ment of antihypertensives at discharge(Table 2). To assess the impact of modeof delivery on duration of severe-rangeHTN, we fitted an exponential regres-sion model that included study arm,mode of delivery, and the interactionbetween the two, which found no inter-action (F ¼ 0.43, P ¼ .51). The primary

616.e5 American Journal of Obstetrics & Gynecol

outcome results were the same in the perprotocol analysis, which excluded 7 pa-tients who crossed over or never receivedstudy drug (see description of excludedpatients in the first paragraph of “Re-sults” section).For the subgroup analysis of time to

BP control, there were 30 (60%) patientsin the acetaminophen group who expe-rienced severe-range HTN and 33 (66%)in the ibuprofen arm. We excluded fromthis analysis the 2 study participants (1from each group) who were deliveredvaginally with forceps or vacuumbecause their small number woulddisproportionately influence the test forinteraction between mode of deliveryand study arm. There were no differen-tial effects in time to BP control betweenthe arms when accounting for mode ofdelivery (accelerated failure time model:interaction between study group andmode of delivery Wald c2 0.50, P ¼ .48).The adjusted mean time (95% CI) to BPcontrol in the acetaminophen group was68.4 hours (54.0e86.7), and 56.7 hours(45.1e71.2) in the ibuprofen arm.Though it was not a planned outcome

measure, we also analyzed the frequencyof at least 1 postpartum BP �150/100

ogy JUNE 2018

mm Hg to more directly compare ourresults with findings published after ourtrial was initiated.23 We also found nodifference between groups for this BPcutoff (ibuprofen 92% vs acetamino-phen 84%, P ¼ .22). Comparisons ofpostpartum opioid use also revealed nodifference between groups overall orover time (Table 3). At 6 weeks post-partum, there was no difference in therates of obstetric triage visits, hospitalreadmissions, or continued use of opi-oids or antihypertensive medications(Table 4).

CommentPrincipal findingsThe main finding of our study is thatcompared to acetaminophen, ibuprofendid not extend the duration of severe-range HTN in women with preeclamp-sia with severe features. Our findings donot lend support to the recommendationput forth ACOG’s 2013 Hypertension inPregnancy Task Force monograph toavoid NSAIDS in women with pre-eclampsia with severe features.18 Whencomparing women randomized toreceive ibuprofen rather than acetamin-ophen for first-line pain control, we did

TABLE 3Pain control and opioid use

Any opioid useIbuprofenn ¼ 50

Acetaminophenn ¼ 50 P value

Postpartum d 0, n (%) 22 (44) 20 (40) .2a

Postpartum d 1, n (%) 27 (54) 25 (50)

Postpartum d 2, n (%) 26 (52) 18 (36)

Opioid requirement (in MEQ),among those requiring opioids

Ibuprofen Acetaminophen

Mean (SD) Median Mean (SD) Median

Postpartum d 0, mg 22.1 (25.1) 15.0 22.6 (13.9) 18.8 .9b

Postpartum d 1, mg 27.5 (18.7) 30.0 32.3 (17.5) 30.0

Postpartum d 2, mg 28.9 (18.9) 22.5 45.4 (32.3) 41.3

Total MEQ dose 77.4 (64.8) 60.0 88.4 (108.1) 30.0 .6c

MEQ, morphine equivalent.

a c2 Test for interaction effect of group and postpartum day on any opioid use from repeated measures logistic regression; b F test for interaction effect of group and postpartum day onlog-transformed opioid MEQs from repeated measures linear regression; c t Test for 2 independent samples on log-transformed total MEQ dose.

Blue et al. Ibuprofen and postpartum blood pressure control in preeclampsia. Am J Obstet Gynecol 2018.

ajog.org OBSTETRICS Original Research

not identify a difference in any metric ofpostpartum BP control, including thetime to BP control among those whoexperienced postpartum severe-rangeHTN. Additionally, we did not find anydifference in the need for antihyperten-sive use either at hospital discharge or at6 weeks postpartum.

We chose the primary outcome mea-sure of duration of severe-range HTNbecause it conveys the effect of elevatedBP on a woman’s postpartum care moreaccurately than other outcome mea-sures. Alternative outcomes, such as thepresence of any BP �160/110 mm Hg,MAP, or rate of persistent postpartumBP �150/100 mm Hg, treat a woman

TABLE 4Outcomes at 6 weeks postpartum

Outcome, n (%)Ibun ¼

Continued antihypertensive use 15

Continued opioid use 2

OB triage visits after discharge 7

Hospital readmission 3

Readmission related to preeclampsia 1

OB, obstetric.

a c2 Test; b Fisher exact test.

Blue et al. Ibuprofen and postpartum blood pressure control

with the specified degree of HTN for afew hours postpartum as categoricallyequal to a woman with the same degreeof HTN for 48 or 72 hours after delivery.In the management of women withpostpartum HTN, the role of timing isimportant. The clinical significance andimplications of a BP �160/110 mm Hgoccurring 6 hours after delivery isdifferent from when it occurs 72 hoursafter delivery, as the woman with severe-range HTN 72 hours after delivery isconsidered to have more persistent, se-vere disease, and is more likely to requireprolonged observation and treatmentwith scheduled antihypertensivemedication.

profen43

Acetaminophenn ¼ 34 P value

(34.9) 7 (20.6) .2a

(4.7) 0 (0) .3b

(16.3) 5 (15.7) .6b

(7.0) 0 (0) .2b

(2.3) 0 (0) .6b

in preeclampsia. Am J Obstet Gynecol 2018.

JUNE 2018 Ameri

Strengths and weaknessesOur study had several strengths. It was adouble-masked, randomized controlleddesign with an intention-to-treat anal-ysis. Additionally, we chose outcomemeasures that were clinically relevantand included additional measures thatwere reported in other studies to facili-tate comparison. The use of around-the-clock study drug dosing helped ensurethat the inherent potential for type IIerror would be related to limited powerrather than underestimation of NSAIDeffect because of the less frequent dosingdue to PRN dosing. Lastly, the highproportion of our study population withclinically significant postpartum HTNmakes our findings generalizable towomen with persistent postpartumHTN, the population among whichACOG initially advised caution.

Our study also had several limitations.The superiority design does not allow usto infer equivalence between treatmentarms, and the relatively small sample sizelimited our ability to assess rare adverseoutcomes. We chose a superiority designbecause our objective was to identify adifference between groups rather thanprove equivalence. Equivalence or non-inferiority designs are powered to ruleout a narrower margin of difference be-tween groups with more confidence, andso require much larger sample sizes.

can Journal of Obstetrics & Gynecology 616.e6

Original Research OBSTETRICS ajog.org

Lastly, the unique racial and ethnicmakeup of our study population, withhigh proportions of Native Americanand Hispanic participants and a lowproportion of Black participants,may limit the generalizability of ourfindings.

Comparison with existing literatureOur findings are consistent with 2 of 3previous publications addressing thisresearch question. The 2 studies consis-tent with our findings were retrospectivein design and did not identify any asso-ciation between postpartum NSAID useand persistent BP �150/100 mm Hg,MAP, antihypertensive use, or length ofpostpartum hospital stay.3,22

Our findings are not consistent withthe 1 previously published open-label,randomized controlled trial, which re-ported that women with preeclampsiawith severe features who were givenibuprofen after vaginal delivery had atleast 1 BP �150/100 mm Hg at morethan twice the rate of those given acet-aminophen (63.1 vs 28.6%, P ¼ .01).23

While such a difference is impressive,this outcome measure may not havesignified clinically relevant postpartumhypertensive morbidity. The in-vestigators did not report whether theHTNwas persistent and whether it led toantihypertensive use or extended post-partum hospital stay.

Significant differences in study meth-odology include their use of open-labelrandomization, limitation of the studypopulation to those having a vaginaldelivery, use of a primary outcome thatwould not affect clinical managementaccording to ACOG clinical guidelines,and less frequent BPmonitoring (every 4hours during the first 24 hours post-partum and every 12 hours thereafter).In contrast, our study employed adouble-masked design, included allmode of delivery types, used a clinicallymeaningful primary outcome thataccurately assessed the difficulty of HTNcontrol, and more frequent BP moni-toring (hourly for first 24 hours post-partum and every 4 hours thereafter).

A principle difference between thestudies is that in their randomizedcontrolled trial, the rate of HTN in the

616.e7 American Journal of Obstetrics & Gynecol

acetaminophen group was only 28.6%compared to 84% in our cohort, whichmay be related to differences in studydesign or attributable to their cohorthaving less severe disease. This maysuggest that patients with less severedisease are more susceptible to NSAID-induced changes in BP. By comparison,our study’s primary and secondary out-comes were specifically chosen toaddress the question of whether anyNSAID-associated increase in BP wouldalter care. The above randomizedcontrolled trial is the only study weidentified that supports the recommen-dation to avoid NSAID use in womenwith postpartum HTN.

ConclusionIn conclusion, we were unable to identifyany detrimental effect of ibuprofen onpostpartum BP. Our study does notsupport the hypothesis that postpartumuse of NSAIDs adversely affects BPcontrol in women with preeclampsiawith severe features. A larger trial is stillrequired to establish noninferiority andexclude differences in rare, adverse out-comes. Additionally, these findings arelimited to women diagnosed with pre-eclampsia with severe features, thereforea trial to assess the effect of NSAIDs onpostpartum BP in the setting of pre-eclampsia without severe features iswarranted. n

AcknowledgmentThe authors would like to thank to Dr Eve Espeyand Dr Rebecca Rogers for their guidance andsupport; Dr Luis Izquierdo for his assistancewiththe institutional review board; as well as DrHeather Riese for her contribution to patientrecruitment.

References

1.Matthews Z. World health report 2005: makeevery mother and child count. World Health2005;33:409-11.2. Sharma KJ, Greene N, Kilpatrick SJ. Orallabetalol compared to oral nifedipine for post-partum hypertension: a randomized controlledtrial. Hypertens Pregnancy 2017;36:44-7.3. Viteri OA, England JA, Alrais MA, et al. Asso-ciation of nonsteroidal antiinflammatory drugsand postpartum hypertension in women withpreeclampsia with severe features. ObstetGynecol 2017;130:1.4. Kim LH, Cheng YW, Delaney S, Jelin AC,Caughey AB. Is preeclampsia associated with

ogy JUNE 2018

an increased risk of cesarean delivery if labor isinduced? JMatern Fetal NeonatalMed 2010;23:383-8.5. Dedier J, Stampfer MJ, Hankinson SE,WillettWC, Speizer FE, CurhanGC. Nonnarcoticanalgesic use and the risk of hypertension in USwomen. Hypertension 2002;40:604-8.6. Curhan GC, Willett WC, Rosner B,Stampfer MJ. Frequency of analgesic use andrisk of hypertension in younger women. ArchIntern Med 2002;162:2204-8.7. White WB, Kent J, Taylor A, Verburg KM,Lefkowith JB, Whelton A. Effects of celecoxib onambulatory blood pressure in hypertensive pa-tients on ACE inhibitors. Hypertension 2002;39:929-34.8. Morgan TO, Anderson A, Bertram D. Effect ofindomethacin on blood pressure in elderly peo-ple with essential hypertensionwell controlled onamlodipine or enalapril. Am J Hypertens2000;13:1161-7.9. Klassen DK, Jane LH, Young DY,Peterson CA. Assessment of blood pressureduring naproxen therapy in hypertensive pa-tients treated with nicardipine. Am J Hypertens1995;8:146-53.10.Whelton A, White WB, Bello AE,Puma JA, Fort JG. Effects of celecoxib androfecoxib on blood pressure and edema inpatients > or ¼65 years of age with systemichypertension and osteoarthritis. Am J Cardiol2002;90:959-63.11. Izhar M, Alausa T, Folker A, Hung E,Bakris GL. Effects of COX inhibition on bloodpressure and kidney function in ACE inhibitor-treated blacks and Hispanics. Hypertension2004;43:573-7.12. Floor-Schreudering A, De Smet PAGM,Buurma H, et al. NSAID-antihypertensive druginteractions: which outpatients are at risk for arise in systolic blood pressure? Eur J Prev Car-diol 2015;22:91-9.13. Johnson AG, Nguyen TV, Day RO. Dononsteroidal anti-inflammatory drugs affectblood pressure? Ann Intern Med 1994;121:289-300.14. Knights KM, Mangoni AA, Miners JO. Non-selective nonsteroidal anti-inflammatory drugsand cardiovascular events: is aldosterone thesilent partner in crime? Br J Clin Pharmacol2006;61:738-40.15.Whelton A, Schulman G, Wallemark C, et al.Effects of celecoxib and naproxen on renalfunction in the elderly. Arch Intern Med2000;160:1465-70.16. Narumiya S, Sugimoto Y, Ushikubi F. Pros-tanoid receptors: structures, properties, andfunctions. Physiol Rev 1999;79:1193-226.17. Rahman M, Wright JT, Douglas JG. The roleof the cytochromeP450-dependentmetabolitesof arachidonic acid in blood pressure regulationand renal function. Am J Hypertension1997;7061:356-65.18. American College of Obstetricians and Gy-necologists; Task Force on Hypertension inPregnancy. Executive summary: hypertension inpregnancy. Obstet Gynecol 2013;122:1122-31.

ajog.org OBSTETRICS Original Research

19. Zlatnik MG, Buhimschi I, Chwalisz K,Liao QP, Saade GR, Garfield RE. The effect ofindomethacin and prostacyclin agonists onblood pressure in a rat model of preeclampsia.Am J Obstet Gynecol 1999;180:1191-5.20. Makris A, Thornton C, Hennessy A. Post-partum hypertension and nonsteroidal anal-gesia. Am J Obstet Gynecol 2004;190:577-8.21. Schachtel BP, Thoden WR, Baybutt RI.Ibuprofen and acetaminophen in the relief ofpostpartum episiotomy pain. J Clin Pharmacol1989;29:550-3.22.Wasden SW, Ragsdale ES, Chasen ST,Skupski DW. Impact of non-steroidal anti-inflammatory drugs on hypertensive disor-ders of pregnancy. Pregnancy Hypertens2014;4:259-63.23. Vigil-De Gracia P, Solis V, Ortega N.Ibuprofen versus acetaminophen as a post-partum analgesic for women with severe pre-eclampsia: randomized clinical study. J MaternNeonatal Med 2017;30:1279-82.24. Lowder JL, Shackelford DP, Holbert D,Beste TM. A randomized, controlled trial tocompare ketorolac tromethamine versusplaceboafter cesareansection to reducepainandnarcoticusage. AmJObstetGynecol 2003;189:1559-62.25. Sammour RN, Ohel G, Cohen M, Gonen R.Oral naproxen versus oral tramadol for analgesiaafter cesarean delivery. Int J Gynecol Obstet2011;113:144-7.

26. Deussen AR, Ashwood P, Martis R. Anal-gesia for relief of pain due to uterine cramping/involution after birth. Cochrane Database SystRev 2011;5:CD004908.27. Bhounsule SA, Nevreker PR, Agshikar NV,Pal MN, Dhume VG. A comparison of four an-algesics in post-episiotomy pain. Indian JPhysiol Pharmacol 1990;34:34-8.28.Wilasrusmee S, Chittachareon A,Jirasiritum S, Srisangchai P. Naproxen sup-pository for perineal pain after vaginal delivery.Int J Gynaecol Obstet 2008;102:19-22.29. El-Tahan MR, Warda OM, Yasseen AM,Attallah MM, Matter MK. A randomized study ofthe effects of preoperative ketorolac on generalanesthesia for cesarean section. Int J ObstetAnesth 2007;16:214-20.30. Towers CV, Shelton S, van Nes J, et al.Preoperative cesarean delivery intravenousacetaminophen treatment for postoperative paincontrol: a randomized double-blinded placebocontrol trial. Am J Obstet Gynecol 2018;218:353.e1-4.31. Lam J, Kelly L, Ciszkowski C, et al. Centralnervous system depression of neonatesbreastfed by mothers receiving oxycodonefor postpartum analgesia. J Pediatr 2012;160:33-7.e2.32. Altenau B, Crisp CC, Devaiah CG,Lambers DS. Randomized controlled trial ofintravenous acetaminophen for postcesarean

JUNE 2018 Ameri

delivery pain control. Am J Obstet Gynecol2017;217:362.e1-6.

Author and article informationFrom the Division of Maternal-Fetal Medicine, Depart-

ment of Obstetrics and Gynecology (Drs Blue, Weinberg,

Holbrook, Katukuri, and Mozurkewich, and Ms Drake-

Lavelle); Division of Epidemiology, Biostatistics, and

Preventive Medicine, Department of Internal Medicine

(Ms Murray-Krezan); and Departments of Family and

Community Medicine and of Obstetrics and Gynecology

(Dr Leeman), University of New Mexico, Albuquerque,

NM.

Received Dec. 8, 2017; revised Feb. 13, 2018;

accepted Feb. 26, 2018.

This study received support from the following sour-

ces: internal grant award from the Department of Ob-

stetrics and Gynecology, University of New Mexico;

University of New Mexico Clinical and Translational Sci-

ence Center Biostatistics Core (National Institutes of

Health [NIH] award UL1TR001449); and REDCap data

management funded by Department of Health and Hu-

man Services/NIH/National Center for Research Re-

sources grant 8UL1TR000041.

The authors report no conflict of interest.

Presented in the late-breaking oral session at the 38th

Annual Pregnancy Meeting of the Society for Maternal-

Fetal Medicine, Dallas, TX, Jan. 29-Feb. 3, 2018.

Corresponding author: Nathan R. Blue, MD.

nblue1297@gmail.com

can Journal of Obstetrics & Gynecology 616.e8