Association of Conventional Cardiovascular Risk Factors With Cardiovascular Disease After Hypertensive Disorders of Pregnancy

Key Points Question Women with history of hypertensive pregnancy disorders have a higher risk of cardiovascular disease, but how much of their excess cardiovascular risk is associated with conventional cardiovascular risk factors? Findings In this cohort study, blood pressure and body mass index were associated with up to 77% of the excess risk of cardiovascular disease in women with history of hypertensive pregnancy disorders, while glucose and lipids were associated with smaller proportions. Meaning The association of conventional risk factors, in particular blood pressure and body mass index, with the development of cardiovascular in women with history of hypertensive pregnancy disorders indicate that preventive efforts aimed at decreasing the levels of these risk factors could reduce cardiovascular risk in women with history of hypertensive pregnancy disorders.


Study Population
HUNT is an ongoing longitudinal population-based study that since the 1980s has invited all residents aged 20 years and above of Nord-Trøndelag county in Norway to undergo comprehensive health assessments with clinical measurements, blood sampling, questionnaires and interviews. To date, three surveys have been conducted; HUNT1 1 1984-86, HUNT2 1995-97 2 and HUNT3 2006-08 3 . Participation rates for women were 89.9% in HUNT1 1 , 75.5% in HUNT2 2 and 58.7% in HUNT3 3 . The population in Nord-Trøndelag county is predominantly White and considered to be representative of Norway as a whole 2 .
MBRN 4 has recorded all births in a mandatory registry since 1967 and contains information on child and maternal health including details of pregnancy complications. In total, 31 364 women from HUNT had at least one birth registered in MBRN between 1967 and 2012. Since we defined our exposure based on history of pregnancy complications by age 40, we excluded 454 women who had their first birth after age 40 and 3901 women who turned 40 after 31 st December 2012, the end of the MBRN follow-up. We excluded 3133 births (and consequently 227 women) which either 1) were multiple pregnancies, 2) had a gestational length < 20 weeks, 3) were preceded by maternal chronic hypertension, 4) produced offspring with a birth weight <350 grams or 5) lacked information on birth weight and gestational length. Further, 1593 women were excluded due to incomplete information on smoking, educational level or family history of coronary heart disease. Additionally, we excluded 1012 women who moved out of Nord-Trøndelag county and 292 women with cardiovascular events before start of follow-up, resulting in a final study population of 23 885 women.

Exposure
Diagnoses of preeclampsia and gestational hypertension in the MBRN were based on internationally recommended diagnostic criteria 5 : Gestational hypertension was generally defined as de novo hypertension (≥140 mmHg systolic and/or ≥90 mmHg diastolic) after 20 weeks of gestation, and preeclampsia also required proteinuria (300 mg/24 hours or ≥1+ on the dipstick test). Previous studies 6,7 have validated the MBRN diagnoses of preeclampsia and gestational hypertension within the HUNT study population finding predictive positive values of 88% and 68% for preeclampsia and gestational hypertension, respectively.

Cardiovascular risk factors
All serum analyses were performed in fresh non-fasting samples at the Central Laboratory, Levanger Hospital, Nord-Trøndelag Hospital Trust using a Hitachi 911 Autoanalyzer in HUNT2 and Architect cSystems ci8200 in HUNT3. Height and weight were measured with the person wearing light clothes and no shoes and were rounded to the nearest cm (height) and half kilo (weight). BMI was calculated as weight (in kg) divided by the squared value of height (in m). Blood pressure in HUNT1 was measured manually two times at 1-minute intervals using a sphygmomanometer after the person had come to rest, and we used the mean value of these two measurements in our analysis. In HUNT2 and HUNT3, blood pressure was measured three times at 1-minute intervals using an automatic oscillometric method (Dinamap, Critikon, Florida) after the person had come to rest, with cuff size adjusted to arm circumference. We used the mean of the second and third measurement, except for 2153 women in HUNT3 who lacked the third measurement due to sick leave amongst staff; for them, we used the second measurement only. Serum total and high density lipoprotein (HDL) cholesterol and triglycerides were analyzed using enzymatic colorimetric methods (Boeheringer Mannheim, Germany) in HUNT2. In HUNT3 HDL cholesterol was measured with an accelerator selective detergent methodology, total cholesterol was analyzed by a cholesterol esterase methodology by equipment from Abbott, Clinical Chemistry, USA. Non-HDL cholesterol was calculated as the difference between total and HDL cholesterol. In HUNT1 capillary glucose was measured at the examination stations in participants above 40 years (Reflocheck-Glucose, Boehringer Mannheim, Germany), and for the analysis of mean glucose levels, we transformed capillary levels to equate serum values (in mmol/L) by multiplying with 1.11 8 . In HUNT2 and HUNT3 serum glucose was measured for all persons using an enzymatic hexokinase method.

Supplemental Validation Information
From the electronic patient administrative system at the two primary hospitals in Nord-Trøndelag county, Levanger Hospital and Namsos Hospital (Nord-Trøndelag Hospital Trust), we obtained information on all study participants registered with at least one of the following cardiovascular diagnoses between September 1, 1987 (the first date of electronic recording) and April 24, 2015: ICD-9,402, 404, 410-414, 425, 427.5, 428, 430-438, 440-448; ICD-10, G45, I11, I13, I20-I25, I42, I46, I50, I60-I67, I71, I72, I74. For each identified patient, one of two experienced cardiologists (B.K., H.D.), who were unaware of the pregnancy history of the participants, examined the medical records to determine the first validated occurrence of each of the following cardiovascular events: myocardial infarction, heart failure, cardiac arrest, cerebral infarction, transient ischaemic attack (TIA), intracranial hemorrhage, aortic dissection or aneurysm, dissection or aneurysm of peripheral arteries, and embolism or thrombosis in peripheral arteries. Myocardial infarction diagnosed using ESC/ACCF/AHA/WHF criteria 9 and classified as STEMI or NSTEMI. Heart failure was diagnosed based on echocardiography findings combined with symptoms/signs according to ESC guidelines, 10 and was classified as systolic or diastolic. Ejection fraction was recorded or, if possible, estimated based on the echocardiography description in cases where no value was reported. If echocardiography had not been performed, the diagnosis was made using the clinically based Framingham criteria. 11 Right ventricular failure without left ventricular failure (e.g. cor pulmonale following pulmonary embolism or COPD) was not included. Cardiac arrest was diagnosed based on a documented or highly probable cardial cause of pulselessness and lack of breathing requiring cardiopulmonary resuscitation at hospital or before hospital admission. Cerebral infarction was diagnosed based on typical symptoms and signs combined with radiological evidence from computer tomography (CT) or magnetic resonance imaging (MRI) scans. In some cases, the diagnosis of cerebral infarction was based on typical symptoms and signs in absence of radiological evidence, but only if CT or MRI scan had been performed, and no alternative explanation for the clinical presentation was found. TIA was diagnosed based on typical symptoms/signs combined with absence of radiological signs of cerebral infarction on CT or MRI scan. Events probably caused by other diseases were excluded. TIA occurring after intracerebral infarction was not recorded. Possible cerebrovascular events accompanying intracerebral malignancy, infection or inflammation were not included. Transitory global amnesia and migraine with aura were not considered as valid cerebrovascular diagnoses. Intracranial hemorrhage was classified based on radiological evidence as subarachnoid, intracerebral, subdural or epidural, and hemorrhage secondary to trauma or intracerebral malignancy was not included. For a diagnosis of aortic aneurysm, an aortic diameter >30 mm in abdomen or >40 mm i thorax was required. Corresponds to the direct effect of HDP on CVD in mediation analysis terminology. § Estimates are adjusted for age, maternal birth year, highest obtained educational level, ever smoked daily, parity before age 40 and family history of coronary heart disease. ¶ The proportion mediated will exceed 100% when the direct effect of HDP go below 1 and appear protective.