Intended for healthcare professionals

Rapid response to:

Practice Uncertainties

Which antihypertensive treatment is better for mild to moderate hypertension in pregnancy?

BMJ 2022; 376 doi: https://doi.org/10.1136/bmj-2021-066333 (Published 18 January 2022) Cite this as: BMJ 2022;376:e066333

Rapid Response:

Re: Which antihypertensive treatment is better for mild to moderate hypertension in pregnancy?

Dear Editor:

We write in response to the Uncertainties article about antihypertensive therapy in pregnancy, published by Ashworth et al.(1), to provide what we feel are important clarifications.

First, mild to moderate hypertension in pregnancy is defined as 140-159/90-109mmHg; the definition provided (140-169/90-109mmHg) includes some women defined by NICE guidance as having severe hypertension in pregnancy (i.e., ≥160/110mmHg). Severe hypertension is a surrogate for adverse pregnancy outcome, similar to pre-eclampsia, and warrants antihypertensive therapy in hospital and under observation.

Second, we believe that it is misleading to suggest that national and international guidelines most commonly recommend only labetalol and nifedipine; they also recommend methyldopa, at least as frequently as nifedipine(2). We emphasise this because it is important not to unnecessarily restrict choice of agents. As the authors rightly point out, some drugs may be less appropriate for individual women, such as labetalol for women with severe asthma, nifedipine for women with migraine, or methyldopa for women with depression. Also, when one medication is used chronically in high-dose (e.g., labetalol) and breakthrough severe hypertension occurs, another agent with a different pharmacological profile is more likely to successfully lower blood pressure (BP).

Third, we feel it is important to clarify that only taken together were beta-blockers (that included, but were not restricted to, labetalol, a combined alpha- and beta-receptor blocker) and calcium channel blockers (that included nifedipine) more effective than methyldopa. The data are limited, but neither beta-blockers (risk ratio 0.59, 95% confidence interval [CI] 0.33, 1.05; 9 trials, 592 women) nor calcium channel blockers (risk ratio 0.23, 95% CI 0.04, 1.22; 2 trials, 46 women) taken alone, were more effective than methyldopa, although the 95% Cis were more consistent with greater (not less) effectiveness(3). The subgroup analysis in the 2018 Cochrane review referenced used a fixed effects model to generate the effect estimate quoted, rather than the random effects model used in the 2014 Cochrane review and more appropriate given the between-trial heterogeneity in outcome (i.e., I2=63%)(4).

Finally, we would like to highlight the findings of our recently-published network meta-analysis of antihypertensive therapy in pregnancy, in which we uniquely, analysed data on labetalol distinct from other beta-blockers(5) . In this analysis that used all available evidence to inform drug vs. drug comparisons, similar effectiveness was seen for prevention of severe hypertension, for labetalol, nifedipine, and methyldopa compared with placebo/no therapy. Also, labetalol was superior to: methyldopa for severe hypertension, methyldopa or calcium channel blockers (including nifedipine) for proteinuria/preeclampsia, and placebo/no therapy for perinatal mortality. An associated trial sequential analysis indicated that 2500-10,000 women/arm would be required to show similar effectiveness between agents for severe hypertension or safety outcomes, with the exception of fetal/newborn death that would require >15,000/arm. Further, it is highly unlikely that future trials will be of sufficient size to better estimate effects in important subgroups (such as ethnicity). As such, using alternative data sources to validate and generalise trial findings to other populations will be of increasing importance(6).

We agree that GiantPANDA will inform clinical care in this area, and like the Uncertainties authors, we include a co-investigator on the trial. However, we must not unnecessarily restrict choice of antihypertensive agents in pregnancy, and we must acknowledge the need for non-randomised study designs to complement randomised trials, to inform care in this area, particularly for perinatal death and other safety outcomes. All of this information will be required so that clinicians can individualize antihypertensive therapy in pregnancy, based on maternal history, physical, and laboratory assessment, and the medication side effect profile.

Sincerely,

Magee LA, Bone J, Sandhu A, Abalos E, Khalil A, Singer J, Prasad S, Omar S, Vidler M, von Dadelszen P

REFERENCES
1. Ashworth D, Battersby C, Green M, et al. Which antihypertensive treatment is better for mild to moderate hypertension in pregnancy? BMJ 2022;376:e066333.
2. Scott G, Gillon TE, Pels A, von Dadelszen P, Magee LA. Guidelines-similarities and dissimilarities: a systematic review of international clinical practice guidelines for pregnancy hypertension. Am J Obstet Gynecol 2020;10.1016/j.ajog.2020.08.018.
3. Abalos E, Duley L, Steyn DW. Antihypertensive drug therapy for mild to moderate hypertension during pregnancy. Cochrane Database Syst Rev 2014:CD002252.
4. Abalos E, Duley L, Steyn DW, Gialdini C. Antihypertensive drug therapy for mild to moderate hypertension during pregnancy. Cochrane Database Syst Rev 2018;10:CD002252.
5. Bone JN, Sandhu A, Abalos E, Khalil A, Singer J, Prasad S, Omar S, von Dadelszen P, Magee LA. Oral antihypertensive therapy for non-severe pregnancy hypertension – systematic review, network meta- and trial sequential analyses Hypertension 2022 (MS# HYPE/2021/18415R1, in press)
6. Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial Is not available. Am J Epidemiol 2016;183(8):758–64

Competing interests: LA Magee is a co-investigator on the GiantPANDA trial and CI of the WILL trial on which Pollyanna Hardy and Lucy Chappell are co-investigators. LA Magee is from the same department as D Ashworth and L Chappell.

03 February 2022
Laura A. Magee
Professor of Women's Health
J Bone, A Sandhu, E Abalos, A Khalil, J Singer, S Prasad, S Omar, M Vidler, P von Dadelszen
King's College London
London