Hypertensive Disorders of Pregnancy: A Systematic Review of International Clinical Practice Guidelines

Background Clinical practice guidelines (CPGs) are developed to assist health care providers in decision-making. We systematically reviewed existing CPGs on the HDPs (hypertensive disorders of pregnancy) to inform clinical practice. Methodology & Principal Findings MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, Cochrane Methodology Register, Health Technology Assessments, and Database of Abstracts of Reviews of Effects (Ovid interface), Grey Matters, Google Scholar, and personal records were searched for CPGs on the HDPs (Jan/03 to Nov/13) in English, French, Dutch, or German. Of 13 CPGs identified, three were multinational and three developed for community/midwifery use. Length varied from 3–1188 pages and three guidelines did not formulate recommendations. Eight different grading systems were identified for assessing evidence quality and recommendation strength. No guideline scored ≧80% on every domain of the AGREE II, a tool for assessing guideline methodological quality; two CPGs did so for 5/6 domains. Consistency was seen for (i) definitions of hypertension, proteinuria, chronic and gestational hypertension; (ii) pre-eclampsia prevention for women at increased risk: calcium when intake is low and low-dose aspirin, but not vitamins C and E or diuretics; (iii) antihypertensive treatment of severe hypertension; (iv) MgSO4 for eclampsia and severe pre-eclampsia; (v) antenatal corticosteroids at <34 wks when delivery is probable within 7 days; (vi) delivery for women with severe pre-eclampsia pre-viability or pre-eclampsia at term; and (vii) active management of the third stage of labour with oxytocin. Notable inconsistencies were in: (i) definitions of pre-eclampsia and severe pre-eclampsia; (ii) target BP for non-severe hypertension; (iii) timing of delivery for women with pre-eclampsia and severe pre-eclampsia; (iv) MgSO4 for non-severe pre-eclampsia, and (v) postpartum maternal monitoring. Conclusions Existing international HDP CPGs have areas of consistency with which clinicians and researchers can work to develop auditable standards, and areas of inconsistency that should be addressed by future research.


Introduction
The hypertensive disorders of pregnancy (HDPs) are common, complicating up to 6-8% of all pregnancies. As such, the HDPs are a leading cause of maternal and perinatal mortality and morbidity, worldwide. It is anticipated that this situation will only worsen, given the rising prevalence of obesity and metabolic syndrome among women of childbearing age [1][2].
Many national and international clinical practice guidelines (CPGs) have been published on the diagnosis, evaluation, and management of the HDPs. Although many such CPGs have quoted the same research papers, the between-guideline variability in specific recommendations has been highlighted by international multicentre research endeavours, such as the CHIPS Trial (Control of Hypertension In Pregnancy Study) [3]. However, no analysis of CPG quality and consistency has been achieved as for other conditions [4][5][6][7][8][9]. In addition, substandard care of women with pregnancy hypertension, especially failures related to diagnosis, evaluation, and management, continues to be recognised as a contributor to maternal death in well- [10][11] and less-resourced settings [12].
We sought to review published CPGs covering the diagnosis, evaluation, and management of the HDPs, in order to inform practicing clinicians about the consistency of the recommendations and the quality of the source guidelines.

Eligibility
Included were multi-disciplinary CPGs that were: (i) published within the last 10 years , and the then accepted 2014 SOGC guideline (now published) [30,31], (ii) covered the diagnosis, assessment and management of one/more of the HDPs in human pregnancy, and (iii) were written in English, French, Dutch or German (i.e., languages understood by the review authors). Excluded were CPGs that: were adapted for local use from an existing CPG, had no references, or were not regional/national/international in scope.

Literature search
A comprehensive literature review was undertaken by a librarian (KM) of the College of Physicians and Surgeons of British Columbia, in consultation with the principal authors of this article. Key words, related to hypertension, pregnancy, and guidelines, were used to search MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, Cochrane Methodology Register, Health Technology Assessments, and Database of Abstracts of Reviews of Effects using the Ovid interface (Appendix S1). As not all CPGs may be available on bibliographic databases cited above, additional sources were searched including personal records. Grey Matters, a tool for evidence-based searching on the internet developed by the Canadian Agency for Drugs and Technologies in Health, was used to locate online grey literature sources, which were searched using key words such as ''hypertensive disorders of pregnancy'', ''gestational hypertension'', ''hypertension during pregnancy'', ''pregnancy induced hypertension'', and ''hypertension gestationnelle''; this site includes the National Guidelines Clearinghouse. Similar search terms combined with ''guideline'' or ''recommendation'' were entered on Google Scholar and the first 100 results were screened, considering most relevant results appear first. Finally, the national websites of societies of obstetrics and gynaecology of the main French-, English-, Dutch-or German-speaking countries were searched.

Processing data
The AGREE II tool was used to assess the methodological quality of all included CPGs [13]. Using the standardised AGREE II methodology, scores of 1 to 7 were given both overall and to each of 23 items in six domains related to standard methodology. Percentages of maximum possible scores were calculated for each domain. Also, each reviewer responded to the following question, ''I would recommend this guideline for use'' with 'yes', 'yes with modifications' and 'no'. For this review, 'yes' was given to a guideline considered useful as a reference document for busy clinicians as such, 'no' was assigned to CPGs that did not formulate recommendations (but consisted of text only). Two authors (TG, AP, and/or LM) rated each CPG independently and discrepancies were resolved by consensus.
Descriptive analyses were undertaken to present the general characteristics of each CPG, including the grading system used to assess the quality of evidence and strength of recommendations. For all CPGs recommended for use, we examined: (i) criteria for diagnosis and HDP classification, using information from the tables and text as diagnostic criteria do not lend themselves well to recommendations, and (ii) recommendations about 'actionable items' related to prevention of preeclampsia or management of any HDP, that were reported commonly (by at least three CPGs) and/or designated to have a high rating for quality of evidence and strength of recommendation.

Guideline characteristics
Level of development

AGREE II scoring
The AGREE II scores for each CPG are presented in Table 4. The highest scores ( §80%) were obtained for the domains 'scope and purpose' ( No CPG achieved §80% of the maximal score for all six domains, but the WHO and NICE guidelines did so for 5/6 domains. Four guidelines did not obtain one score §80% in any domain [ASH, DGGG, ESC, SOMANZ]; these same CPGs were also rated as not being clinically useful. As such, the HDP classification and recommendations regarding prevention and treatment are described for the remaining nine guidelines.
'Severe' pre-eclampsia is defined by most (7/9) CPGs, but there is little consistency. Heavy proteinuria is included by some (N53) [WHO, NVOG, AOM], but specifically excluded by others (N52) [ACOG, SOGC]. Five CPGs define end-organ complications of severe pre-eclampsia; the most common maternal are: headache/visual symptoms (N55 CPGs), right upper quadrant/ epigastric abdominal pain (N54), severe hypertension (N55), eclampsia (N52), pulmonary oedema (N53), low platelets (N54), renal insufficiency (N53), and elevated liver enzymes (N53); these mirror the diagnostic criteria used in some guidelines. Fetal manifestations of pre-eclampsia are specified by three CPGs, all of which list stillbirth and none of which specify abruption without evidence of fetal compromise; IUGR is included by WHO and SOGC, but specifically excluded by ACOG. The SOGC 'severity' criteria are indications for delivery, and include some features that in other CPGs: (i) define pre-eclampsia but not severe pre-eclampsia (e.g., stroke), (ii) define both pre-eclampsia and severe pre-eclampsia (e.g., eclampsia, pulmonary oedema, platelet count ,100610 9 /L, and acute kidney injury), or (iii) define neither pre-eclampsia nor severe preeclampsia but are widely regarded as indications for delivery (e.g., uncontrolled severe hypertension).
In the three CPGs that specify that proteinuria is mandatory to define preeclampsia [ Prediction (Table S3) Screening only by clinical risk markers is recommended (N53 CPGs, 0 high rating), with no guideline recommending routine use of biomarkers or ultrasonography. The actual risk markers used were not reviewed.  For women with chronic hypertension who are taking antihypertensive therapy and planning pregnancy, it is recommended that preconceptual counselling be undertaken (N54) [NICE, QLD, NVOG, SOGC] and that this include discussion of alternatives to ACE inhibitors and ARBs which should be stopped if inadvertently taken in early pregnancy (N54) [NICE, NVOG, ACOG, SOGC].  (Table S5).

Findings
We identified 13 CPGs that published recommendations about the diagnosis, classification, prevention and treatment of the HDPs. Four CPGs were assessed as 'not useful' to busy clinicians, based on text-only publication [DGGG, SOMANZ, ASH] or limited text with only a few focused recommendations [ESC]. Analysis of the nine remaining CPGs revealed few consistencies and/or high rating recommendations. Consistency was seen for the definitions of hypertension, proteinuria, chronic and gestational hypertension. Consistency and high ratings (by at least one CPG) was seen for: (i) the preventative strategies of calcium (in the setting of low intake) and low-dose aspirin for women at increased risk of preeclampsia, and neither vitamins C and E or diuretics; (ii) antihypertensive treatment of severe hypertension; (iii) MgSO 4 for eclampsia and severe preeclampsia; (iv) antenatal corticosteroids at ,34 wks when delivery is probable within the next seven days; (v) delivery for women with severe pre-eclampsia who do not yet have a viable fetus and for those with any pre-eclampsia at term; and (vi) active management of the third stage of labour with oxytocin.
Notable inconsistencies, illustrative of a lack of consensus, were in areas well reported by CPGs that differed nevertheless in their recommendations: (i) definitions of pre-eclampsia and in particular, severe pre-eclampsia and superimposed pre-eclampsia that reflect our evolving understanding of the multisystem nature of the disease; (ii) target BP among women with non-severe hypertension, regardless of the HDP; (iii) timing of delivery for women with pre-eclampsia and severe pre-eclampsia; (iv) MgSO4 for non-severe pre-eclampsia, and (v) postpartum monitoring for maternal safety and improvement of long-term cardiovascular health. These are areas requiring further research and consensus-building for optimising management of a high risk group of women.
Some guidelines covered areas neglected by others, and those CPGs could be useful sources of specific information. Notable examples include post-delivery discharge planning for transfer of care [NICE] and obstetric anaesthesia for the HDPs [SOGC].

How our findings fit with the published literature
We are aware of only one review of CPGs for pregnancy hypertension [45]. The AGREE II instrument was used to evaluate methodological quality; CPGs scored highest in ''clarity of presentation'' and lowest in ''editorial independence'', consistent with our findings. Between-CPG differences in the number and extensiveness of recommendations were identified, but recommendation content, similarities and differences between guidelines were not explored.
The 2014 position statement from the International Society for the Study of Hypertension in Pregnancy (ISSHP) endorses areas of consistency within published CPGs [46]. Of note, ISSHP endorses a definition of pre-eclampsia that does not require proteinuria, but can be made based on maternal end-organ involvement and/or fetal IUGR. Uniquely, ISSHP does not support a distinction between severe and ''mild'' pre-eclampsia which, ''…should be considered as one that is at anytime capable of being severe and life-threatening for mother and baby''. Superimposed pre-eclampsia should not be diagnosed based on a rise in BP alone. Gestational proteinuria is mentioned specifically as potentially signifying evolving pre-eclampsia or underlying renal disease. All women with pre-eclampsia should be admitted to hospital, at least initially. MgSO 4 is advocated for all women with pre-eclampsia in low-and-middle-income countries. Particular emphasis is placed on the importance of recognising 'white coat' hypertension, the promising future of biomarkers as diagnostic and/or prognostic tools [47], and the importance of each unit having its own written policies to promote uniform care, the outcomes of which can be monitored.
Two of the CPGs presented 'auditable' standards [33,43], but their complexity in one would be difficult to operationalise [NICE] (http://www.nice.org.uk/ guidance/qs35) and only one criterion is presented in the other CPG [WHO]. Our review suggests that where there is consistency between CPGs, there is the potential for standardisation of both: definitions that will support research efforts [48,47], and quality of care criteria, particularly if the between-CPG differences in quality of evidence/strength of recommendation can be resolved.
The AGREE II tool is the standard for assessing the quality of published CPGs [13]. However, it has never been shown to improve guideline uptake or implementation [8], and use of the AGREE II presents some difficulties. First, AGREE II lists many criteria and few CPGs in our review scored highly on some or most domains, which may reflect space limitations in the journal of publication, rather than guideline quality. Second, AGREE II scores do not reflect important usability issues, such as the length of the CPG document and appendices/evidence tables (extensive for the NICE guideline), number of formulated recommendations, and presentation of the grading of the evidence relative to the recommendation, or lack of assessment of the strength of the recommendation (absent from AGREE II), all of which must be considered when evaluating how easy guidelines would be to use clinically. Although we did not exclude any CPGs based on their AGREE II domain scores, the four CPGs deemed 'not useful' for busy clinical practice did receive the lowest scores [ASH, DGGG, ESC and SOMANZ].

Strengths and limitations
The strengths of our review include: a comprehensive literature search by an information technology specialist, inclusion of CPGs published in the last 10 years and in any of four languages, and the systematic summary of the diagnostic criteria and stated recommendations for the prevention of pre-eclampsia and treatment of the HDPs, incorporating quality of evidence and strength of recommendations. Although it is recommended that guidelines be updated every 3-5 years [49][50][51], the time consuming work of developing CPGs could mean that some guidelines are not updated as frequently as recommended and in our review, limiting to publication within the last 5 years would have excluded 4/13 CPGs.
Our review has limitations. A potential selection bias exists in the fact that only CPGs written in English, German, French and Dutch were included; however, we excluded only two CPGs (in Spanish [28,29]) for this reason; the French CPG excluded was because it was available only by purchase for a significant sum [26]. The CPGs were assessed by two appraisers; this approach meets the minimum number of appraisers advised by AGREE II, but some authors have used more. As our focus was on the clinician, we did not extract and compare information available only from guideline text, some of which runs to 288 pages (even without appendices or evidence tables) [NICE] and none of which has associated 'strength of recommendation' that would aid the reader in deciding whether to comply or not. Finally, discrepancies in grading of evidence between different systems for the same recommendation were noted.

Conclusions
The existing CPGs that inform care for women with a HDP have areas of consistency with which clinicians and researchers can work to develop auditable standards, and areas of inconsistency that should be addressed by future research.