Pregnancy Hypertension 18 (2019) 179–187

Contents lists available at ScienceDirect

Pregnancy Hypertension
journal homepage: www.elsevier.com/locate/preghy

Review article

First-line antihypertensive treatment for severe hypertension in pregnancy: T

A systematic review and network meta-analysis
Sepand Alavifarda,b,c, Rebecca Chased, Ghayath Janoudia,e, Andréanne Chaumontd,
Andrea Lanesa,b,d,f, Mark Walkera,b,f, Laura Gaudetb,g,
⁎

a
Obstetrics & Maternal Newborn Investigation (OMNI) Research Group, Department of Obstetrics and Gynecology, University of Ottawa, Ottawa, ON, Canada
b
Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada
c
Faculty of Medicine, University of Toronto, Toronto, ON, Canada
d
Department of Family Medicine, University of Ottawa, Ottawa, ON, Canada
e
Canadian Agency for Drugs and Technologies in Healthcare, Ottawa, ON, Canada
f
BORN Ontario, Children’s Hospital of Eastern Ontario, ON, Canada
g
Department of Obstetrics and Gynecology, Queen’s University, Kingston, ON, Canada

ARTICLE INFO ABSTRACT

Keywords: Background: Hydralazine, labetalol, and nifedipine are the recommended first-line treatments for severe hy-
Pregnancy-induced hypertension pertension in pregnancy. While all three are effective, there is a lack of sufficient evidence regarding their
Antihypertensive agents comparative safety and efficacy.
Hydralazine Objective: To determine the comparative safety and efficacy of the first-line treatment options for severe hy-
Labetalol
pertension in pregnancy.
Nifedipine
Methods: A systematic search of Medline, Embase, and Cochrane Central Register of Controlled Trials up to May
Network meta-analysis
31, 2018 was conducted. RCTs in pregnancy comparing a first-line antihypertensive agent to another first-line
agent for the treatment of severe hypertension in pregnancy. Screening, data abstraction, and quality assessment
were done by two independent reviewers. To estimate relative effects from all available evidence, a Bayesian
network meta-analysis with vague priors was conducted.
Main Results: Of the 1330 publications identified, 17 RCTs comprised of a total of 1591 women met our selection
criteria. For successful treatment of severe hypertension, nifedipine was found to be superior to hydralazine (OR
4.13 [95% CrI 1.01–20.75]) but not labetalol (OR 3.43 [95% CrI 0.94–19.95]). This was not associated with an
increased risk for caesarean delivery or maternal side effects. There was no significant difference between la-
betalol and hydralazine.
Conclusions: Given the results of this systematic review and network meta-analysis, maternity care providers
should feel comfortable initiating management of severe hypertension in pregnancy using oral nifedipine.

1. Introduction a number of serious adverse events for mother and child. It increases
the risk of maternal morbidities, including: stroke, maternal-end organ
It is estimated that hypertensive disorders of pregnancy (HDP) damage, abnormal clotting, postpartum hemorrhage, and placental
complicate 5–10% of pregnancies worldwide and account for 18% of abruption. In addition, the fetus is at increased risk of intrauterine
maternal deaths [1,2]. HDPs are responsible for 25% of maternal deaths growth restriction, premature delivery, and low Apgar score [1,5–7].
in Latin America and the Caribbean, 16% in developed countries, and Another serious concern is the development of severe hypertension into
9% in Asia and Africa [3]. As the leading contributor to maternal and preeclampsia, the most dangerous of HDPs [8].
neonatal mortality and morbidity across the world [1], optimal treat- For severe hypertension in pregnancy, there is general consensus
ment of HDPs represents an important avenue for scientific inquiry. supporting immediate intervention with antihypertensive agents [4,8].
Severe hypertension in pregnancy is defined as systolic blood Current clinical practice guidelines (CPGs) recommend intravenous (IV)
pressure (sBP) ≥160 mmHg and/or diastolic blood pressure (dBP) hydralazine, IV labetalol, and oral nifedipine as the first-line anti-
≥110 mmHg [4]. Severe hypertension in pregnancy is associated with hypertensive treatments [5,6,9–11]. While all three are effective

⁎
Corresponding author at: Kingston Health Sciences Centre, 76 Stuart Street, Victory 4, Kingston, ON K7L 2V7, Canada.
E-mail address: Laura.Gaudet@kingstonhsc.ca (L. Gaudet).

https://doi.org/10.1016/j.preghy.2019.09.019
Received 19 April 2019; Received in revised form 16 September 2019; Accepted 27 September 2019
2210-7789/ © 2019 International Society for the Study of Hypertension in Pregnancy. Published by Elsevier B.V. All rights reserved.
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

antihypertensive medications, there is a lack of sufficient evidence re- adverse events (SAE). Secondary fetal outcomes included: perinatal
garding their comparative safety and efficacy [8,12,13]. death, neonatal death, stillbirth, neonatal intensive care unit (NICU)
Based on the current body of evidence, the decision of which first- admission, premature delivery, Apgar < 7 at 5 min, small for gesta-
line antihypertensive treatment to use is based on physician experience, tional age (SGA), and fetal arrhythmia.
convenience of use, local availability, and cost [8,12–14]. Since effec- During data abstraction, patient and trial characteristics as well as
tive management of severe hypertension is vital for safe progression of outcome definitions were recorded and later reviewed as possible
pregnancy, a more evidence-based approach is necessary. Network sources of clinical or methodological heterogeneity. Statistical hetero-
meta-analyses (NMA) use direct and indirect evidence to strengthen geneity evaluated for the standard pairwise meta-analysis using I2. For
treatment comparisons and allow for effect estimates between inter- each outcome, network inconsistency was evaluated using an incon-
ventions that have not been compared head-to-head and for the esti- sistency plot.
mation of the relative efficacy and safety of the interventions under
investigation [15]. 3. Data synthesis
Therefore, we conducted a systematic review and NMA to determine
the comparative effectiveness and safety of hydralazine, labetalol, and To estimate relative effects from all available evidence (direct and
nifedipine for severe hypertension in pregnancy. indirect), a Bayesian network meta-analysis with vague (non-in-
formative) priors and a homogenous variance structure was conducted
for outcomes with sufficient evidence. Relative effects were calculated
2. Methodology
as odds ratios (ORs) with 95% credible intervals (CrI) using both a fixed
and random effects model. The analysis was done using 5000 burn-ins
2.1. Systematic search
and 40,000 iterations with NetMetaXL – a Bayesian network meta-
analysis tool that uses WinBUGS from within Microsoft Excel [18].
We conducted a comprehensive systematic search of Medline (ovid),
NetMetaXL was also used to generate NMA network diagrams, forest
Embase (ovid), and Cochrane central register of controlled trials
plots, and rankograms. Convergence was assessed using the Brooks-
(CCRT) up to October 20, 2015, then updated until May 31, 2018. The
Gelman-Rubin method and visual analysis of trace plots. Deviance in-
search strategies used are outlined in Appendix S1. The Medline search
formation criterion (DIC) and inconsistency plots were used to assess
strategy was verified by comparing the results with 10 RCTs known to
model fit [18,19]. For each outcome, surface under cumulative ranking
be relevant and peer reviewed by an information specialist using PRESS
curve (SUCRA) was used for treatment ranking.
[16].
To estimate the relative effect of all direct pairwise comparisons, a
standard pairwise meta-analysis using random effects model was con-
2.2. Study selection, data extraction, and study quality assessment ducted [20]. Comparative effect sizes were calculated as ORs with 95%
confidence intervals (CI).
Screening, data abstraction, and assessment of the quality of each In October 2015, this systematic review and network meta-analysis
study was done by paired independent reviewers (SA, RC or SA, AC). was registered on PROSPERO (CRD42015025839).
The quality of trials that met the selection criteria were assessed using
the Cochrane Collaboration’s ‘Risk of bias’ tool for randomized con- 4. Results
trolled trials [17]. Disagreements were resolved by consensus or a third
independent reviewer (LG). 4.1. Study characteristics
Included in this review were all randomized controlled trials (RCTs)
for treatment of severe hypertension or hypertensive emergencies in Of the 1330 publications identified, 17 RCTs comprising a total of
pregnancy using any of the three agents of interest (labetalol, hy- 1591 women met the selection criteria (Fig. 1) [21–37].
dralazine, nifedipine), regardless of dosage or administration protocol. Overall, there were 595 patients (37.4%) receiving labetalol, 598
Although severe hypertension is formally defined by International (37.6%) receiving hydralazine, and 398 (25%) receiving nifedipine. All
Society for the Study of Hypertension in Pregnancy as sBP of labetalol and hydralazine arms had an intravenous (IV) route of ad-
≥160 mmHg and/or dBP ≥110 mmHg, we broadened our criteria to ministration. Of the 11 trials with a nifedipine arm, nine orally ad-
also include studies that investigated hypertensive emergencies. ministered nifedipine, one sublingually [21], and one used sublingual
Therefore, we included studies that had minimum blood pressure cri- route for the first three doses but any further dose was given orally
teria of sBP of ≥160 mmHg and/or dBP ≥105 mmHg. The studies had [23]. There were no trials with more than two arms; six trials (802
to also report at least one outcome of interest. Studies were excluded if women) that compared labetalol to hydralazine [22,24,25,32,33,37],
they did not have an accompanying full-text in English or French, had six trials (422 women) that compared hydralazine to nifedipine
combination interventions, or were restricted to only postpartum [21,23,26,28,34,35], and five trials (367 women) that compared labe-
management of severe hypertension. talol to nifedipine [27,29–31,36]. The pairwise meta-analyses can be
The primary outcome of interest was successful treatment of hy- found in Appendix S2.
pertension, as defined by the study. The secondary outcomes of interest Selected characteristics of included trials are presented in Table 1.
were divided into maternal and fetal outcomes. Secondary maternal Of the trials that had a labetalol arm, three had a maximum adminis-
outcomes included: persistent hypertension, maternal hypotension, tration dosage of 140 mg while the remaining had a maximum dosage
composite maternal morbidity*, proteinuria, preeclampsia, eclampsia, of 300 mg. The hydralazine and nifedipine maximum dosages varied
placental abruption, emergency termination of pregnancy, caesarean between 40–90 mg and 15–45 mg, respectively. Details of trial dosage
delivery, admission to ICU, composite maternal side effects†, adverse protocols and target blood pressures are presented in Appendix S3.
events (AE), withdrawal due to adverse events (WDAE), and serious Most trials had at least one component of sBP ≥160 mmHg or dBP
≥110 mmHg as part of their inclusion criteria. There were two trials
* that studied “Hypertensive Emergencies” which were defined as some
Composite of non-fatal cerebrovascular, cardiovascular, and cardio-
pulmonary events (stroke, myocardial infarction, angina, ischemia, ar- variation of “sustained sBP ≥170 mmHg and/or dBP ≥105 mmHg”
rhythmias, congestive heart failure, pulmonary edema), kidney failure, liver [28,30] . These trials had dosage protocols and target blood pressures
failure, and permanent eyesight loss. within the variation between most other trials.
†
Composite of dizziness, nausea, vomiting, headache, tachycardia, brady- There were four trials that investigated patients with severe pre-
cardia. eclampsia. Of these, three compared hydralazine to nifedipine

180
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

Fig. 1. PRISMA Flow Diagram – Literature Search and Screening Results.

[21,23,26] and one compared labetalol to nifedipine [36]. Of those studies showed high risk of attrition bias [21–23]. Based on the risk of
comparing hydralazine to nifedipine, two administered the agent until bias assessment, the included trials were of poor to fair study quality.
time of delivery (no maximum dosage) [21,23] and one had an unclear The severe preeclampsia trials (Table 1) and variation in drug ad-
dosage protocol [26]. ministration regimens (Appendix S3) between trials were identified as
Of the 17 studies, two recruited patients in the post-partum period the primary source of clinical heterogeneity. When statistical hetero-
[25,30]. Vermillion et al. enrolled during the intrapartum or within 24 h geneity was possible to assess, I2 was < 30% in all cases except for
postpartum but did not report the number of patients in each category. labetalol vs nifedipine for composite maternal side effects (I2 = 80%)
Mabie et al. enrolled patients during pregnancy or in the puerperium, 19 and labetalol vs hydralazine for caesarean delivery (I2 = 63%)
of 60 received the randomized medication before delivery; 13/40 in the (Appendix S2).
labetalol arm and 6/20 in the hydralazine arm.
For the network meta-analysis, the performance of both the fixed 4.3. Blood pressure control
and random effects models were evaluated. The random effects models
had a greater number of studies with a contribution to deviance close to The network meta-analysis for successful treatment of severe hy-
one and therefore was the model used for analysis (Appendix S4). pertension consisted of eight trials with a total of 790 patients – 356 in
the labetalol arm, 300 in the hydralazine arm, and 134 in the nifedipine
4.2. Quality assessment arm (Appendix S6).
As presented in Fig. 2A, nifedipine was found to be superior to
Amongst the included trials, there was a median sample size of 60 hydralazine (OR 4.13 [95% CrI 1.01–20.75]) but not labetalol (OR 3.43
(range 30–275). The majority of studies had low or unclear risk of bias [95% CrI 0.94–19.95]). There was no significant difference between
for most criteria (Appendix S5); ten out of 17 studies had a high risk of labetalol and hydralazine.
bias for both performance and detection bias, however, the effect on the There were 15 trials that reported on maternal hypotension and five
results is unclear as most outcomes were objective in nature. Three trials that reported on persistent hypertension. Due to the low event

181
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

Table 1
Selected Characteristics of Included Trials.

rate for both outcomes, an NMA was not performed (Table 2). In of pregnancy, and WDAE were not reported by any trials.
summary, the pooled incidence of hypotension was highest in the hy-
dralazine group (7.55%) when compared to labetalol (1.68%) and ni- 4.7. Fetal outcomes
fedipine (0.57%). For persistent hypertension, nifedipine had the
highest incidence (9.09%) when compared to hydralazine (6.64%) and There was a low event rate for perinatal death, neonatal death,
labetalol (2.93%). stillbirth; Apgar < 7 at 5 min, and fetal arrhythmia also had low event
rates (Tables 4 and 5) and therefore a network meta-analysis was not
4.4. Caesarean delivery conducted.
Premature delivery and SGA were not reported by any of the trials.
The NMA consisted of ten trials with a total of 725 patients – 259 in There was one trial that reported IUGR (fetal growth restriction) with
the labetalol arm, 292 in the hydralazine arm, and 174 in the nifedipine 10 events (in 103 pregnancies) in the labetalol group and 8 events (in
arm (Appendix S6). The analysis showed no significant difference in 102 pregnancies) in the hydralazine group [24].
caesarean delivery between any of the treatment comparisons (Fig. 2B).
4.8. Treatment rankings
4.5. Maternal side effects (composite)
Appendix S7 presents the rankogram and SUCRA for the reported
The NMA consisted of 14 trials with a total of 1403 patients – 500 in outcomes. For successful treatment of hypertension, nifedipine is shown
the labetalol arm, 534 in the hydralazine arm, and 369 in the nifedipine to have the highest probability of being ranked the first best (95.3%)
arm (Appendix S6). The analysis showed no significant different in while hydralazine has the highest probability of being ranked as the
compose maternal side effects between any of the treatment compar- third best treatment (62.0%). The SUCRA for maternal side effects
isons (Fig. 2C). (composite) suggests hydralazine to have the lowest probability of
being ranked as the best treatment (6.9%). There was no clear finding
4.6. Other maternal outcomes for caesarean delivery.

SAE was reported by three studies. two comparing labetalol to ni- 5. Discussion
fedipine and one comparing hydralazine to nifedipine [21,29,36]. The
incidence of events was zero for all groups. Similarly, eclampsia was 5.1. Main findings
reported by two studies, both with no incidence [23,24]. Other ma-
ternal outcomes that were had insufficient evidence for NMA are pre- We have performed a network meta-analysis to assess the safety and
sented in Table 3. Preeclampsia as an outcome, emergency termination efficacy of three widely used antihypertensive treatments in the context

182
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

Fig. 2. Forest Plots for Network Meta-Analysis.

of severe hypertension in pregnancy. Through the use of both direct and analysis as it was published after our search end-date.
indirect evidence, this systematic review suggests that oral nifedipine A network meta-analysis by Sridharan et al. has also studied phar-
improves successful treatment of hypertension with no observed in- macotherapy management of severe hypertension in pregnancy. Their
crease in risk for caesarean delivery or maternal side effects relative to inclusion criteria were different than our review in that they did not
IV hydralazine and with no significant difference to IV labetalol. limit their analysis to first-line treatment options, and included trials
A recent trial comparing oral nifedipine to IV labetalol, consistent that compared combination treatment (eg. Ding et al. comparing com-
with our conclusion, shows no difference between the two drugs in bination of nifedipine and resveratrol dual-therapy to nifedipine as
successful treatment of severe hypertension [38]. It supports the use of monotherapy [39]). While their analysis included more interventions,
nifedipine as it reports target BP to be achieved more rapidly and with they were unable to make conclusions for comparisons outside of hy-
fewer number of doses by nifedipine. This study was not included in our dralazine, labetalol, and nifedipine. Similar to our quality assessment,

183
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

Table 2 Table 4
Maternal Hypotension and Persistent Hypertension (Raw Data). Perinatal Death, Neonatal Death, and Still Birth (Raw Data).
Author (Publication Year) # of Events / # of Patients Author (Publication Year) # of Events / # of Patients

Labetalol Hydralazine Nifedipine Labetalol Hydralazine Nifedipine

Maternal Hypotension Perinatal Death

Mabie (1987) 0/40 1/20 - Fenakel (1991) – 2/27 1/26
Vermillion (1999) 0/25 - 1/25 Shekhar (2013) 2/30 – 0/30
Aali (2002) - 0/61 0/65 Morris (2016) 0/15 1/14 –
Vigil-De Gracia (2006) 0/100 2/100 –
Neonatal Death
Rezaei (2011) – 0/25 1/25
Vigil-De Gracia (2006) 2/103 2/102 –
Raheem (2012) 0/25 – 0/25
Shekhar (2013) 2/30 – 0/30
Shekhar (2013) 0/30 – 0/30
Shi (2016) 0/73 – 0/74
Delgado De Pasquale (2014) 0/131 0/130 –
Sharma (2017) – 0/30 0/30
Dhananjaya (2015) 0/30 – 0/30
Morris (2016) 0/15 0/14 – Stillbirth
Sabir (2016) – 2/50 0/50 Martins-Costa (1992) 0/30 – 0/30
Shi (2016) 0/73 – 0/74 Shekhar (2013) – 0/17 2/20
Patel (2017) 0/76 1/76 – Patel (2017) 3/76 4/76 –
Sharma (2017) – 1/30 0/30 Total (Percentage) 9/357 (2.52) 9/266 (3.38) 3/240 (1.25)
Tariq (2017) 10/50 35/50 –
Total (Percentage) 10/595 (1.68) 42/556 (7.55) 2/354 (0.57)

Persistent Hypertension Table 5

Vigil-De Gracia (2006) 5/100 5/100 – Apgar Score < 7 at 5 min and Fetal Arrhythmia (Raw Data).
Rezaei (2011) – 11/25 5/25
Delgado De Pasquale (2014 2/131 6/130 – Author # of Events / # of Patients
Patel (2017) 2/76 2/76 –
Sharma (2017) – 0/30 0/30 Labetalol Hydralazine Nifedipine
Total (Percentage) 9/307 (2.93) 24/361 (6.64) 5/55 (9.09)
Apgar < 7 at 5 min
Fenakel (1991) – 2/27 1/26
Rezaei (2011) – 0/25 0/25
Sridharan et al. found the evidence to be sub-optimal in quality pri- Shekhar (2013) 2/30 – 1/30
marily due to inclusion of studies with high risk of bias. Vermillion (1999) 2/14 – 1/15
Similar to our findings, Sridharan et al. found no significant differ- Vigil-De Gracia (2006) 4/103 2/102 –
ence when comparing hydralazine and nifedipine to labetalol. For Shi (2016) 12/73 – 10/74
Patel (2017) 1/76 0/76 –
success of treatment, our results favored nifedipine compared to hy- Sharma (2017) – 1/30 2/30
dralazine (OR 4.13 [95% CrI 1.01–20.75]) while theirs found no dif- Total (Percentage) 21/296 (7.09) 5/260 (1.92) 15/200 (7.50)
ference (OR 2.1 [95% CI 0.9–5.2]). The reliability of both findings is
Fetal Arrhythmia
unclear given the wide interval estimates. Fenakel (1991) – 0/27 1/26
While Sridharan et al. had the same primary outcome as our study, Vermillion (1999) 1/14 – 2/15
their secondary outcomes also included number of doses required and Vigil-De Gracia (2006) 6/103 8/102 –
Rezaei (2011) – 3/25 1/25
the time taken for successful treatment. The study found target BP to be
Sabir (2016) – 3/50 1/50
achieved with fewer doses when using nifedipine compared to hy- Patel (2017) 12/76 21/76 –
dralazine with a mean difference (95% CI) of −0.4 (−0.8, −0.1) with Sharma (2017) – 0/30 0/30
no significant difference in time for achieving target blood pressure. Total (Percentage) 19/193 (9.84) 35/310 (11.3) 5/146 (3.42)
Their study’s trial sequential analysis concluded sufficient evidence
for the primary outcome when comparing hydralazine and nifedipine to
labetalol, but insufficient evidence for the comparison between hy- The results of our study with respect to hydralazine and nifedipine
dralazine and nifedipine. are consistent with previous meta-analyses. The most recent Cochrane
Collaboration review on drugs for treatment of very high blood pressure

Table 3
Other Maternal Outcomes (Raw Data).
Author (Publication Year) # of Events / # of Patients

Outcome Labetalol Hydralazine Nifedipine

Adverse Events (Minor) Aali (2002) – 10/61 11/61

Adverse Events Shekhar (2013) 0/30 – 0/30
Adverse Events Shi (2016) 14/73 – 11/74
Chest Pain Shi (2016) 0/73 – 0/74
Convulsion Patel (2017) 1/76 3/76
HELLP Dhananjaya (2015) 0/30 – 1/30
ICU Admission Raheem (2012) 2/25 – 0/25
Kidney failure Fenakel (1991) – 1/25 2/24
Kidney failure Dhananjaya (2015) 0/30 – 1/30
Kidney failure and pulmonary edema Vigil-De Gracia (2006) 1/100 0/100 –
Palpitations Raheem (2012) 0/25 – 1/25
Palpitations Dhananjaya (2015) 2/30 – 10/30
Placental abruption Vigil-De Gracia (2006) 1/100 2/100 –
Proteinuria Dhananjaya (2015) 26/30 – 22/30
Shortness of Breath Shi (2016) 0/73 – 0/74

184
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

during pregnancy found calcium channel blockers (nifedipine and ni- 5.2. Strengths & limitations
cardipine) to reduce the risk of persistent hypertension when compared
to hydralazine [8]. There are several strengths associated with our study. Notably, this
Ultimately, the available literature suggests no greater harm with study uses direct and indirect evidence to evaluate the safety and effi-
any one of the recommended first-line treatments. There is some evi- cacy of antihypertensive agents for severe hypertension in pregnancy.
dence suggesting nifedipine’s superiority and therefore should be fur- In addition, we have included a number of recent trials that have not
ther investigated in a well-designed randomized controlled trial. previously been appraised or synthesized. Our study was also con-
One of the commonly reported adverse effects of hydralazine is ducted in accordance to the Cochrane Handbook for Systematic
hypotension which is concerning due to the association between feto- Reviews of Interventions and the PRISMA-NMA statement [43].
placental hypoperfusion [5]. Our study provides up-to-date evidence of The quantity of the available evidence and data did not facilitate
the frequency of hypotension among the addressed interventions and conducting a robust network meta-analysis; as evident by the wide
confirms hydralazine as the intervention with the highest frequency; credible intervals demonstrated and the inability to perform a network
Labetalol 1.7%, Hydralazine 7.6%, Nifedipine 0.6%. We did not per- meta-analysis on a number of pre-planned outcomes. The quality of the
form a meta-analysis due to the low event rates (Table 2). evidence informing our study is subject to a number of limitations as
In a systematic review which included one trial of 50 women, discussed in the results section. Our results are subject to poor to fair
caesarean delivery risk was shown to be non-significant between those study quality and trials. In addition, we are unable to make a strong
receiving nifedipine compared to labetalol [12]. This is consistent with assessment of the safety profile of the treatments due to low event rates
the findings of our study which included direct evidence from a total of or poor reporting of safety outcomes.
three trials with 139 women as well as indirect evidence from seven Based on the information reported, the assessment of sources of
other trials. clinical heterogeneity is limited. While most studies focused on women
In our study, there was no significant difference between maternal with severe gestational hypertension, four of the 18 studies limited
side effects of the interventions. This is in contrast to the meta-analysis their inclusion criteria to those with severe preeclampsia. Although
of three trials and a total of 207 patients by Shekhar et al. which found severe hypertension (sBP ≥160 and/or dBP ≥110) is a component of
nifedipine to be associated with fewer maternal side effects than labe- both clinical conditions, preeclampsia is associated with significantly
talol [13]. Two of the three trials were included in our analysis. The more adverse pregnancy outcomes than isolated gestational hyperten-
third trial was not included as no full-text was identified. Our analysis sion and thus, this may account for clinical heterogeneity. While most
included three trials that were not included in that meta-analysis. It is trials in the meta-analysis had a maximum number of dosages as part of
important that not all trials reported on all outcomes included in the their protocol, three of the four trials studying preeclampsia had no
composite. A core outcomes set for hypertension in pregnancy, similar maximum dosage (Table 2). The findings of the individual trials were
to that which is being prepared for preeclampsia, would help improve similar to others comparing the same two interventions (Appendix S2).
the interpretability of results from meta-analyses on this topic [40]. Another source of clinical heterogeneity was introduced by the two
One advantage of conducting a network meta-analysis is the ability trials that studied “Hypertensive Emergencies”. The blood pressure
to use SUCRA to provide comparative treatment ranking between three criteria for these trials (sustained sBP ≥170 mmHg and/or dBP
or more interventions. SUCRA provides a measure of the probability of ≥105 mmHg) falls within the criteria reported by ACOG Committee
a treatment to be ranked, compared to other treatments examined, as Opinion #767 (“acute-onset, severe hypertension [sBP of ≥160 mmHg
the most effective, second most effective, third most effective, etc. and/or dBP ≥110 mmHg] that…is persistent for 15 min or more”) but
When examining the probability of being ranked number 1 (most ef- differs from the definition. Given the overlap in definition, a priori we
fective treatment) we found that nifedipine had the highest probability broadened our inclusion criteria to include these trials. Importantly,
to be the most effective at 95.3%. A potential interpretation can be these trials had dosage protocols and target blood pressures comparable
phrased that nifedipine was shown to be most effective treatment in to most other trials. The findings were similar to other trials (Appendix
almost all of the Markov chain Monte Carlo simulations. A hypothesized S2).
clinical translation of this finding can be viewed as if the treating
clinician has access to all the reviewed agents, providing nifedipine will 5.3. Interpretation
likely provide the highest effectiveness in more than 95% of the pa-
tients given the medication. This can be contrasted to the decision to This systematic review provides a comprehensive summary of the
give labetalol or hydralazine where the expectation would be for them evidence on first-line management of severe hypertension in pregnancy.
to be effective in 2–3% of the patients receiving this drug. Although that While our network meta-analysis showed oral nifedipine to have better
the difference between the probability to be ranked the number one efficacy than IV hydralazine, a similar study did not find the relation-
drug (nifedipine) and the other two drugs is drastically different, as ship to be significant. In terms of safety, the current evidence, while
described previously by Wang et al., it is important to be cautious in limited, suggests a comparable safety profile between labetalol, hy-
interpreting SUCRA findings as they can be misleading and imprecise dralazine, and nifedipine; nifedipine has shown to have the lowest in-
[41]. Therefore, considering that the SUCRA does not provide a mea- cidence of hypotensive episodes. As an inexpensive oral agent that does
sure of potential uncertainty, it is difficult to ascertain if this difference not require special storage, nifedipine also provides an economic ad-
is sufficiently and clinically meaningful. vantage, and is more accessible for community care and resource-con-
Although we excluded trials exclusively studying the postpartum strained settings. While maternity care providers should feel comfor-
period, only one previous trial was identified that focused on the table using oral nifedipine for management of severe hypertension, the
management of severe hypertension in the postpartum period relationship between nifedipine and hydralazine should be further in-
[12,35,42]. This trial randomized 82 women with to IV labetalol or IV vestigated.
hydralazine and found no significant difference in treatment of severe
hypertension or maternal side-effects. The 2019 ACOG Committee 6. Conclusion
Opinion on management of acute-onset postpartum severe hyperten-
sion recommends IV labetalol and hydralazine as first-line treatments Given the results of this systematic review and the network meta-
with extrapolation of safety of oral nifedipine based on trials that re- analysis, maternity care providers should feel comfortable initiating
cruited in pregnancy. The National Institute for Health and Care Ex- management of severe hypertension in pregnancy using oral nifedipine.
cellence 2019 recommendations propose treatment of hypertension in Further efforts are needed on the direct comparison between nifedipine
the postpartum period as an avenue for further research. and hydralazine as well as educational and translational studies, and

185
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

evaluation of the implementation of nifedipine as first-line treatment (2014) 1210–1218 discussion 1220.
for severe hypertension in pregnancy. [13] S. Shekhar, N. Gupta, R. Kirubakaran, P. Pareek, Oral nifedipine versus intravenous
labetalol for severe hypertension during pregnancy: a systematic review and meta-
analysis, BJOG 123 (1) (2016) 40–47.
Acknowledgements [14] L.A. Magee, C. Cham, E.J. Waterman, A. Ohlsson, P. von Dadelszen, Hydralazine for
treatment of severe hypertension in pregnancy: meta-analysis, BMJ 327 (7421)
(2003) 955–960.
Not applicable. [15] E.J. Mills, K. Thorlund, J.P. Ioannidis, Demystifying trial networks and network
meta-analysis, BMJ 346 (2013) f2914.
Disclosure of interests [16] M. Sampson, J. McGowan, C. Lefebvre, D. Moher, J.G. PRESS: Peer Review of
Electronic Search Strategies. Ottawa: Canadian Agency for Drugs and Technologies
in Health. 2008.
The authors declare no conflict of interest. [17] J.P.T. Higgins, D.G. Altman, P.C. Gøtzsche et al. The Cochrane Collaboration’s tool
for assessing risk of bias in randomised trials. Vol 3432011.
[18] S. Brown, B. Hutton, T. Clifford, T. Clifford, D. Coyle, et al., A Microsoft-Excel-based
Contribution of authors
tool for running and critically appraising network meta-analyses–an overview and
application of NetMetaXL. Syst. Rev. (2014) 2046-4053 (Electronic).
Sepand Alavifard contributed to the study conception, planning, [19] S. Dias, N.J. Welton, A.J. Sutton, D.M. Caldwell, G. Lu, A.E. Ades, Evidence
performing the systemic review, network meta-analysis, and manu- synthesis for decision making 4: inconsistency in networks of evidence based on
randomized controlled trials, Med. Decis. Making 33 (5) (2013) 641–656.
script preparation. [20] Review Manager (RevMan) [computer program]. Version Version 5.2. Copenhagen:
Dr. Rebecca Chase contributed to performing the systematic review The Nordic Cochrane Centre (2012).
and preparation of this manuscript. [21] B.S. Aali, S.S. Nejad, Nifedipine or hydralazine as a first-line agent to control hy-
pertension in severe preeclampsia, Acta Obstet. Gynecol. Scand. 81 (1) (2002)
Andréanne Chaumont contributed to performing the systematic re- 25–30.
view and preparation of this manuscript. [22] S. Delgado De Pasquale, R. Velarde, O. Reyes, K. De La Ossa, Hydralazine vs la-
Dr. Ghayath Janoudi contributed to the study conception, planning, betalol for the treatment of severe hypertensive disorders of pregnancy. A rando-
mized, controlled trial, Pregnancy Hypertens. 4 (1) (2014) 19–22.
data analysis, and manuscript preparation. [23] K. Fenakel, G. Fenakel, Z. Appelman, S. Lurie, Z. Katz, Z. Shoham, Nifedipine in the
Dr. Andrea Lanes contributed to the study conception, planning, and treatment of severe preeclampsia, Obstet. Gynecol. 77 (3) (1991) 331–337.
manuscript preparation. [24] P. Vigil-De Gracia, M. Lasso, E. Ruiz, et al., Severe hypertension in pregnancy:
hydralazine or labetalol. A randomized clinical trial, Eur. J. Obstet. Gynecol.
Dr. Mark Walker contributed to the study conception and planning, Reprod. Biol. 128 (1–2) (2006) 157–162.
provided clinical expertise, and contributed to manuscript preparation. [25] W.C. Mabie, A.R. Gonzalez, B.M. Sibai, E. Amon, A comparative trial of labetalol
Dr. Laura Gaudet contributed to the study conception and planning, and hydralazine in the acute management of severe hypertension complicating
pregnancy, Obstet. Gynecol. 70 (3 Pt 1) (1987) 328–333.
provided clinical expertise, and contributed to manuscript preparation.
[26] S. Martins-Costa, J.G. Ramos, E. Barros, R.M. Bruno, C.A. Costa, J.R. Goldin,
Randomized, controlled trial of hydralazine versus nifedipine in preeclamptic
Funding women with acute hypertension, Clin. Exp. Hypertens. - Part B Hypertens.
Pregnancy 11 (1) (1992) 25–44.
[27] I.A. Raheem, R. Saaid, S.Z. Omar, P.C. Tan, Oral nifedipine versus intravenous la-
This project was funded in part by the University of Ottawa betalol for acute blood pressure control in hypertensive emergencies of pregnancy:
Scholarship in Maternal and Neonatal Medicine. a randomised trial, BJOG 119 (1) (2012) 78–85.
[28] Z. Rezaei, F.R. Sharbaf, M. Pourmojieb, et al., Comparison of the efficacy of nife-
dipine and hydralazine in hypertensive crisis in pregnancy, Acta Med. Iran 49 (11)
Appendix A. Supplementary data (2011) 701–706.
[29] S. Shekhar, C. Sharma, S. Thakur, S. Verma, Oral nifedipine or intravenous labetalol
Supplementary data to this article can be found online at https:// for hypertensive emergency in pregnancy: a randomized controlled trial, Obstet.
Gynecol. 122 (5) (2013) 1057–1063.
doi.org/10.1016/j.preghy.2019.09.019. [30] S.T. Vermillion, J.A. Scardo, R.B. Newman, S.P. Chauhan, A randomized, double-
blind trial of oral nifedipine and intravenous labetalol in hypertensive emergencies
References of pregnancy, Am. J. Obstet. Gynecol. 181 (4) (1999) 858–861.
[31] B. Dhananjaya, R. Jamuna, Oral nifedipine versus intravenous labetalol in hy-
pertensive emergencies of pregnancy: a randomised trial, Res. J. Pharm., Biol.
[1] J.A. Hutcheon, S. Lisonkova, K.S. Joseph, Epidemiology of pre-eclampsia and the Chem. Sci. 6 (2) (2015) 1673–1681.
other hypertensive disorders of pregnancy, Best Pract. Res. Clin. Obstet. Gynaecol. [32] R. Morris, I. Sunesara, M. Darby, et al., Impedance cardiography assessed treatment
25 (4) (2011) 391–403. of acute severe pregnancy hypertension: a randomized trial, J. Matern. Fetal
[2] R. Lozano, M. Naghavi, K. Foreman, et al., Global and regional mortality from 235 Neonatal Med. 29 (2) (2016) 171–176.
causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the [33] P. Patel, D. Koli, N. Maitra, T. Sheth, P. Vaishnav, Comparison of efficacy and safety
Global Burden of Disease Study 2010, Lancet 380 (9859) (2012) 2095–2128. of intravenous labetalol versus hydralazine for management of severe hypertension
[3] K.S. Khan, D. Wojdyla, L. Say, A.M. Gulmezoglu, P.F. Van Look, WHO analysis of in pregnancy, J. Obstet. Gynaecol. India (2017) 1–6.
causes of maternal death: a systematic review, Lancet 367 (9516) (2006) [34] S. Sabir, S. Yasmin, G. Abbas, Comparison of oral nifedipine with intravenous hy-
1066–1074. dralazine for acute hypertensive emergencies of pregnancy, J. Postgraduate Med.
[4] M.A. Brown, L.A. Magee, L.C. Kenny, et al., The hypertensive disorders of preg- Inst. 30 (4) (2016) 328–330.
nancy: ISSHP classification, diagnosis & management recommendations for inter- [35] C. Sharma, A. Soni, A. Gupta, A. Verma, S. Verma, Hydralazine vs nifedipine for
national practice, Pregnancy Hypertens. 13 (2018) 291–310. acute hypertensive emergency in pregnancy: a randomized controlled trial, Am. J.
[5] L.A. Magee, E. Abalos, P. von Dadelszen, et al., How to manage hypertension in Obstet. Gynecol. 217 (6) (2017), https://www.ajog.org/article/S0002-9378(17)
pregnancy effectively, Br. J. Clin. Pharmacol. 72 (3) (2011) 394–401. 30965-1/fulltext.
[6] L.A. Magee, E. Abalos, P. von Dadelszen, B. Sibai, S.A. Walkinshaw, C.S. Group, [36] D.D. Shi, F.Z. Yang, L. Zhou, N. Wang, Oral nifedipine vs. intravenous labetalol for
Control of hypertension in pregnancy, Curr. Hypertens. Rep. 11 (6) (2009) treatment of pregnancy-induced severe pre-eclampsia, J. Clin. Pharm. Ther. 41 (6)
429–436. (2016) 657–661.
[7] S.J. Kilpatrick, A. Abreo, N. Greene, et al., Severe maternal morbidity in a large [37] S. Tariq, A. Shahid, T. Yousof, Comparison of maternal hypotension after admin-
cohort of women with acute severe intrapartum hypertension, Am. J. Obstet. istration of labetalol versus hydralazine in treating patients having severe preg-
Gynecol. 91215 (1) (2016) e91–e97. nancy induced hypertension, Pak. J. Med. Sci. 11 (2) (2017) 541–543.
[8] L. Duley, S. Meher, L. Jones, Drugs for treatment of very high blood pressure during [38] M. Zulfeen, R. Tatapudi, R. Sowjanya, IV labetalol and oral nifedipine in acute
pregnancy, Cochrane Database Syst. Rev. 7 (7) (2013) CD001449. control of severe hypertension in pregnancy-a randomized controlled trial, Eur. J.
[9] T.E. Gillon, A. Pels, P. von Dadelszen, K. MacDonell, L.A. Magee, Hypertensive Obstet. Gynecol. Reprod. Biol. 236 (2019) 46–52.
disorders of pregnancy: a systematic review of international clinical practice [39] J. Ding, Y. Kang, Y. Fan, Q. Chen, Efficacy of resveratrol to supplement oral nife-
guidelines, PLoS One 9 (12) (2014) e113715. dipine treatment in pregnancy-induced preeclampsia, Endocrine Connections 6 (8)
[10] G.T. Too, J.B. Hill, Hypertensive crisis during pregnancy and postpartum period, (2017) 595–600.
Semin. Perinatol. 37 (4) (2013) 280–287. [40] J.M. Duffy, J. van’t Hooft, C. Gale, et al., A protocol for developing, disseminating,
[11] L.A. Magee, A. Pels, M. Helewa, E. Rey, P. von Dadelszen, Canadian hypertensive and implementing a core outcome set for pre-eclampsia, Pregnancy Hypertens. 6 (4)
disorders of pregnancy working G. Diagnosis, evaluation, and management of the (2016) 274–278.
hypertensive disorders of pregnancy, Pregnancy Hypertens. 4 (2) (2014) 105–145. [41] Z. Wang, R.E. Carter, Ranking of the most effective treatments for cardiovascular
[12] T. Firoz, L.A. Magee, K. MacDonell, et al., Oral antihypertensive therapy for severe disease using SUCRA: Is it as sweet as it appears? Eur. J. Prev. Cardiol. 25 (8) (2018)
hypertension in pregnancy and postpartum: a systematic review, BJOG 121 (10) 842–843.

186
S. Alavifard, et al. Pregnancy Hypertension 18 (2019) 179–187

[42] P. Vigil-De Gracia, E. Ruiz, J.C. Lopez, I.A. de Jaramillo, J.C. Vega-Maleck, [43] B. Hutton, G. Salanti, D.M. Caldwell, et al., The PRISMA extension statement for
J. Pinzon, Management of severe hypertension in the postpartum period with in- reporting of systematic reviews incorporating network meta-analyses of health care
travenous hydralazine or labetalol: a randomized clinical trial, Hypertens. interventions: checklist and explanations, Ann. Intern. Med. 162 (11) (2015)
Pregnancy 26 (2) (2007) 163–171. 777–784.

187