NEW ORLEANS — Peripartum stroke occurs more frequently than has been assumed and has debilitating consequences, according to investigators who reported here in February at the American Stroke Association Annual Meeting.
“Although stroke during childbirth is relatively uncommon, it is the sixth largest cause of maternal death, and 22 percent of those who survive these events do not go home immediately after discharge,” said principal investigator Cheryl Bushnell, MD, in an interview.
Because the costs of such a stroke are so devastating for the affected family, which must arrange for childcare while the mother convalesces, and for society, Dr. Bushnell encouraged neurologists to collaborate with obstetricians, nurse-midwives, and family physicians to address maternal hypertension and to educate pregnant women on the early warning signs of stroke. Dr. Bushnell is Assistant Professor of Medicine at the Duke Center for Cerebrovascular Disease of Duke University Medical Center in Durham, NC.
COVERING THE FIELD
She and her co-investigators, including neurologists and obstetricians, conducted the study to ascertain the incidence of stroke in pregnant women in the US, as well the mortality and disability. They obtained data from the Nationwide Inpatient Sample for the years 2000 through 2001. Within this sample, they analyzed all pregnancy-related discharge International Classification of Diseases-9 (ICD-9) codes: antepartum, cesarean or vaginal delivery, and postpartum. Among the pregnancy-related discharges, they searched the ICD-9 codes for strokes: hemorrhagic, ischemic, cerebral vein thrombosis, and pregnancy-related cerebrovascular event. The investigators then analyzed discharge destination data as a way to gauge disability associated with such events and compared the destinations of women with and without strokes.
Among 8,328,927 discharges with a pregnancy, delivery, or postpartum code, 2,850 had stroke codes, for a rate 34.2 per 100,000 deliveries. Ischemic stroke was coded in 9.2 per 100,000 deliveries; hemorrhagic in 8.5 per 100,000 deliveries; and cerebral vein thrombosis in 0.6 per 100,000 deliveries. Interestingly, the least specific code, pregnancy related cerebrovascular event, was the most common, occurring in 15.9 of 100,000 deliveries.
Among these discharges, the investigators documented 117 stroke-related deaths, or 1.4 per 100,000 deliveries. The case fatality rate was 4.1 percent. Among the fatal strokes, 72.6 were hemorrhagic. Twenty-two percent of stroke survivors were discharged to a skilled nursing facility, but only 3 percent of all pregnancy-related discharges were sent to such facilities.
“The findings show the need for better coding for strokes in pregnant women,” Dr. Bushnell said. “For example, although cerebral venous thrombosis is a relatively common event, we found few in the ICD-9 codes. Venous syndromes may have been coded as pregnancy-related cerebrovascular events.”
Although the neurologist typically sees patients after the stroke occurs, they often see pregnant patients who may have increased risk of stroke, she said. For example, pregnant patients with headache, or those with a history of transient ischemic attack, may consult a neurologist. In those patients, physicians should consider managing blood pressure more aggressively. For patients with any thrombotic risk factors, low-molecular weight heparin should be considered.
As with other headache patients, pregnant women with such histories need to be educated to report atypical headaches, especially in the third trimester. Pre-eclampsia can cause headaches, and fortunately, the treatment of pregnancy-related hypertension can resolve headache, Dr. Bushnell said.
NEXT STEP: UNDERSTANDING MECHANISMS
The next step is to understand the specific mechanisms involved in pregnancy-related stroke, she said. “In the past, many women died from hemorrhage during pregnancy,” she said. “Because those who did not have a hemorrhage survived, there may be a natural selection in favor of thrombosis.” Investigators will want to discern whether stroke in pregnant women is related to pre-eclampsia, or hypertension, or to a separate mechanism, such as PFO or thrombotic factors.
“We also need to know which women are having reversible ischemia related to pre-eclampsia versus stroke,” she said. “Such lesions resolve completely, although they look terrible on imaging studies. Therefore, we need to know how reversible vasculopathies are coded.”
She and her co-investigators are currently studying pre-eclampsia in women who have cardiovascular disease after pregnancy. They are also designing a large prospective, multicenter study regarding pregnancy and stroke.
RECOMMENDATION FOR CASE-CONTROL STUDY
Although pregnancy-related stroke is still a relatively rare event, the findings are concerning and warrant further research, said S. Claiborne Johnston, MD, PhD, Director of the Stroke Service at the University of California-San Francisco Medical Center, in a phone interview.
“We need to understand the etiology of stroke in this setting and find ways to reduce the risks,” said Dr. Johnston. “For example, a case-control study could help us understand what is different about women who have stroke during pregnancy in contrast to women who don't. We need to know if they have, for example, a genetic predisposition to clotting, uncontrolled hypertension, or other stroke risk factors.”
He pointed out that drawing conclusions from the Nationwide Inpatient Sample has limitations, since the entries are “only as good as the reliability of the coding, which is done by non-physicians.” The person doing the coding has to interpret the physician's notes in the chart and then assign an ICD-9 code. Some anomalies can therefore appear, he continued, like the surprisingly disproportionately large number of pregnancy-related cerebrovascular events. “The accuracy of the code depends on what the physician wrote and the data entry worker's ability to interpret what the physician wrote,” Dr. Johnston said.
However, he stressed that the Nationwide Inpatient Sample is useful for defining the overall impact of a condition, like pregnancy-related stroke, on the health care system. “The study reminds us that stroke can occur in a population that is not at elevated risk for stroke,” he said. “Trying to identify the pathogenesis is important because stroke in pregnancy has a different pattern of risk than stroke in other at-risk populations, such as older patients.” Further research, such as a case-control study, may help identify the causes behind that pattern and in turn help identify young women at risk before they become pregnant, he said.
ARTICLE IN BRIEF
✓ An analysis of the Nationwide Inpatient Sample for the years 2000 through 2001 found that peripartum stroke was more common than thought, but that there were anomalies in the coding for specific types of stroke related to pregnancy.