Brain magnetic resonance imaging findings in postpartum headache

Abstract

Objectives

The objectives were to determine the frequency of abnormal magnetic resonance imaging findings in patients with postpartum headache and related factors.

Methods

A total of 102 patients with postpartum headache underwent brain magnetic resonance imaging study. The images were examined by a blinded radiologist. The related demographic, obstetric and headache-related factors were recorded. The multiple logistic regression model was used to determine the predictive factors.

Results

Abnormal magnetic resonance imaging findings were observed in 42 of 102 patients (41.2%, 95% confidence interval = 31.6 to 50.7%). The most common finding was sinusitis (10 of 42 patients, 23.8%, 95% confidence interval = 15.5 to 32%). Then, posterior reversible encephalopathy syndrome (six of 42 cases, 14.2%, 95% confidence interval = 7.4 to 20.9%), cerebral venous thrombosis (four of 42 cases, 9.5%, 95% confidence interval = 3.8 to 15.1%), and subarachnoid hemorrhage (four of 42 cases, 9.5%, 95% confidence interval = 3.8 to 15.1%) were most prevalent findings. Convulsions (odds ratio of 3.39) and initiation of headache earlier than 5 days postpartum (odds ratio of 0.29) were significant predictive factors.

Conclusion

Abnormal brain magnetic resonance imaging findings were seen in a considerable number of patients with postpartum headache. When headache starts in the first 5 days postpartum and accompanied by convulsions, there are likely to be abnormal magnetic resonance imaging findings.

Introduction

Postpartum headache is described as headache, neck pain or shoulder pain that occurs during the first 6 weeks following delivery.¹ Headache is one of the most common postpartum symptoms, and up to 39% of women experience headaches during the first week after childbirth.¹ Some reports noted an incidence of postpartum headache as high as 80%.²

Headaches that occur after delivery are divided into primary and secondary groups. Primary headaches account for about 90% of postpartum headaches. They are often benign and include migraine headaches, cluster headaches, tension headaches, etc.³ But another group known as secondary headaches is much more important and can sometimes be life-threatening.⁴ Secondary headaches occur due to complications of anesthesia, complications of delivery, and neurological diseases. They include post-dural puncture headache, cerebrovascular diseases such as stroke, increased intracranial pressure, brain tumor, reversible cerebrovascular vasoconstriction syndrome, venous sinus thrombosis, subarachnoid hemorrhage, and complications of pregnancy such as preeclampsia and eclampsia.²

The puerperal period is associated with unique hormonal and physiological changes such as hypercoagulability and endothelial dysfunction. These factors can increase the incidence of secondary headaches such as those that arise due to cerebrovascular diseases.⁵

Earlier diagnosis of secondary headaches and their differentiation from primary headaches, which are mostly benign, is important and can be very difficult. In a previous study, secondary headache was reported in 53% of postpartum women.² In addition to clinical findings in differentiating primary versus secondary headaches, imaging techniques are of great importance. Computed tomography (CT) scanning and magnetic resonance imaging (MRI) have both been suggested as preferred diagnostic imaging methods for investigating postpartum headache. However, MRI is more sensitive than a CT (computerized tomography) scan in postpartum period.⁶

So far, limited studies have been published on the use of MRI and the observed findings in patients with postpartum headache. Because there is not enough evidence of the role of MRI in postpartum headache, the aim of this study was to determine the frequency of abnormal findings detected on brain MR images of women with postpartum headaches. In addition, the association of clinical variables with abnormal findings was investigated.

Materials and methods

Study design and population

In this cross-sectional study, the study population consisted of all women who were admitted to the obstetrics department of our university hospital in 2016 to give birth via natural delivery or cesarean section. Those patients who had a new-onset headache after delivery and neurology consultation determined the necessity for brain imaging with MRI and had no contra-indications for this imaging were eligible to be included. Imaging is considered necessary in the case of progressive headache, characteristics of intracranial hypertension, focal neurological signs, altered quality or severity of headache, or the presence of background clinical conditions including immunosuppression, taking anti-coagulants, or previous history of hypercoagulation (such as deep vein thrombosis). As per protocols implemented and followed strictly in the hospital, all patients who experience postpartum headache must consult a neurologist. The inclusion criteria were occurrence of postpartum headache at a time interval of 6 weeks after delivery and requirement for neuroimaging by MRI. Exclusion criteria were previous history of migraine headaches, previous history of cranial surgery, and receiving epidural spinal anesthesia for childbirth.

The sampling method was a census. All postpartum women who met the inclusion criteria during the study period were included. In the tertiary care university-affiliated medical center where the study was performed, about 6,000 women give birth annually. All postpartum women who complain of headache during their stay at the hospital or after discharge are seen by a neurologist.

Variables and imaging

After the neurologists referred the eligible patients to the imaging department, the patients were visited by the research team members and their medical records were reviewed to collect the required data. The data extracted were demographic characteristics (age, body mass index (BMI)), pregnancy-related variables (type of delivery: normal delivery or cesarean section), number of births, history of preeclampsia and seizure associated with headache) and headache-related variables (time that headache started and headache pattern).

The MRI of the brain was then performed using a Phillips MRI machine with a magnetic field intensity of 1 Tesla. The imaging was performed with T1-weighted, T2-weighted and FLAIR sequences with additional sequences of gradient and diffusion-weighted (DW) imaging and magnetic resonance venography. MRI images were interpreted by a board-certified radiologist (ES, with 15 years of experience as the faculty member of the university) who was blind to the history and clinical findings of the patients, and abnormal findings were recorded.

Statistical analysis

Normal distribution of the continuous data was determined by the Kolmogorov-Smirnov test. Age did not have a normal distribution. Descriptive indices such as mean, standard deviation (SD), median, frequency and percentage were used to describe the variables. To compare the findings of brain MRI based on age, BMI, history of preeclampsia, seizure, type of delivery and number of deliveries, a Chi-square test or Fisher’s exact test was used. To determine the effect of age, BMI, history of preeclampsia, occurrence of seizures, type of delivery and number of deliveries on the presence of abnormal MRI findings, a multivariate logistic regression analysis with calculation of odds ratio (OR) and its 95% confidence intervals (CI) were used. The significance level was considered as 0.05. The analyses were done using SPSS software (Ver. 22.0, IBM, USA).

Ethics

Firstly, the objectives of the study were described to the patients. Then informed written consent was obtained from the patients. The protocol of this study was approved by the ethics committee of our medical university.

Results

Demographic variables

In the study period, a total of 6,000 women gave birth at our hospital. There were about 3,600 cesarean section deliveries and about 2,400 natural deliveries. More than 95% of cesarean deliveries were carried out under spinal anesthesia and were excluded. Of the remaining patients, 102 patients developed postpartum headache that required brain imaging by MRI. The age range of the patients was 19 to 35 years with a median of 29 years. The mean (±SD) BMI was 22.66 (±2.22) kg/m² (range, 19 to 27).

Pregnancy- and childbirth-related variables

Overall, 18 of 102 patients (17.6%, 95% CI = 10.7 to 24.4%) had cesarean delivery and 84 of 102 patients (82.4%, 95% CI = 75 to 89.7%) had normal delivery. We found 14 of 102 patients (13.7%, 95% CI = 7 to 20.3%) had preeclampsia history in their previous pregnancies; 16 of 102 cases (15.7%, 95% CI = 8.6 to 22.7%) had experienced seizures. In total, 22 of 102 patients (21.6%, 95% CI = 13.6 to 29.5%) had no previous history of delivery, 26 of 102 cases (25.5%, 95% CI = 17 to 33.9%) had a single delivery history, 38 of 102 subjects (37.3%, 95% CI = 27.9 to 46.6%) had a history of two deliveries, 10 of 102 cases (9.8%, 95% CI = 4 to 15.5%) had a history of three deliveries, and six of 102 patients (5.9%, 95% CI = 1.3 to 10.4%) had a history of four deliveries.

Headache-related variables

Mean (±SD) time interval from delivery to the onset of postpartum headache was 5.09 (±3.21) days (range: 1 to 14 days). The clinical pattern of headache in 62 patients of 102 cases (60.8%, 95% CI = 51.3 to 70.2%) was of a tension headache, in 24 of 102 patients (23.5%, 95% CI = 15.2 to 31.7%) was of a migraine headache, and in 16 of 102 patients (15.7%, 95% CI = 8.6 to 22.7%) was of a cluster type.

MRI variables

MR image examinations were normal in 60 patients of 102 patients (58.8%, 95% CI = 49.2 to 68.3%). However, abnormal MRI findings were observed in 42 out of 102 patients (41.2%, 95% CI = 31.6 to 50.7%). The most common findings were sinusitis (10 of 42 patients, 23.8%, 95% CI = 15.5 to 32%), posterior reversible encephalopathy syndrome (PRES) (six of 42 patients, 14.2%, 95% CI = 7.4 to 20.9%), cerebral venous thrombosis (CVT) (four of 42 patients, 9.5%, 95% CI = 3.8 to 15.1%) (Figures 1 and 2), and subarachnoid hemorrhage (four of 42 patients, 9.5%, 95% CI = 3.8 to 15.1%). Other abnormal findings, which were observed in two separate patients of 42 subjects (4.7%, 95% CI = 0.5 to 8.8%) included intracranial hypertension, multiple sclerosis plaques, small vessel ischemic disease, neurofibroma, sinus polyps, extra-axial hydrocephalus, posterior cerebral artery infarct, meningioma, and lenticular nucleus infarcts.

Figure 1.

T1-weighted sagittal magnetic resonance imaging of a 38-year-old pregnant woman with postpartum headache and cerebral venous thrombosis showing high signal intensity thrombus (arrows) in the posterior half of the superior sagittal sinus.

Figure 2.

Axial T2-weighted (a) and FLAIR images (b) of the same patient, which shows high signal intensity thrombus (arrows) in the superior sagittal sinus.

Comparison of abnormal MRI findings based on the studied variables

Table 1 presents comparison of the frequency of normal and abnormal MRI findings according to the studied variables. As observed, there was no significant difference between normal and abnormal MRI findings groups regarding age, BMI, delivery mode, number of previous deliveries, and history of previous preeclampsia. In contrast, patients who had experienced seizures were more likely to have abnormal brain MRI findings. Abnormal brain MRI findings were significantly more prevalent among those patients whose postpartum headache had started in the first 5-day period after delivery. In addition, more patients with abnormal brain MRI findings had cluster headaches.

Table 1.

Comparison of the frequency of normal and abnormal brain MR images findings among patients with postpartum headache according to demographic, pregnancy, and headache variables.

	Normal brain MRI (N = 60)	Abnormal brain MRI (N = 42)	P value, 95% CI
Age
<30 years	34 (58.6%)	24 (41.4%)	0.962, –0.191 to 0.199
≥30 years	26 (59.1%)	18 (40.9%)
Body mass index
<25	44 (57.9%)	32 (42.1%)	0.82, –0.141 to 0.199
≥25	16 (61.5%)	10 (38.5%)
Delivery mode
Normal delivery	46 (54.8%)	38 (45.2%)	0.112, –0.0009 to 0.276
Cesarean delivery	14 (77.8%)	4 (22.2%)
Number of previous deliveries
0	14 (63.6%)	8 (36.4%)	0.789, –0.116 to 0.202
1–2	36 (56.3%)	28 (43.8%)
3–4	10 (62.5%)	6 (37.5%)
Pre-eclampsia	6 (42.9%)	8 (57.1%)	0.191, –0.05 to 0.23
Seizures	6 (33.3%)	12 (66.7%)	0.019, 0.029 to 0.342
Headache onset time, day
<5	24 (46.2%)	28 (53.8%)	0.009, 0.078 to 0.455
≥5	36 (72%)	14 (28%)
Headache pattern
Tension	38 (61.3%)	24 (38.7%)	0.006, 0.068 to 0.369
Migraine	18 (75%)	6 (25%)
Cluster	4 (25%)	12 (75%)

CI: confidence interval; MRI: magnetic resonance imaging.

Predictors of abnormal brain MRI

The independent variables included in the binary logistic regression model were age, BMI, history of preeclampsia, seizure, delivery mode, number of previous deliveries, postpartum headache onset time, and headache pattern. The dependent variable was the presence or absence of abnormal MRI findings in the brain (binary variable). Table 2 shows the results of the regression analysis. As observed, occurrence of seizures and headache onset times fewer than 5 days following delivery remained the statistically significant predictors for abnormal brain MRI. In other words, these factors significantly contributed to the classification of patients with abnormal and normal MRI findings and the prediction of abnormal MRI findings in women with postpartum headache.

Table 2.

Binary logistic regression analysis to determine predictors of abnormal MRI findings among women with postpartum headache.

	B	SE	P value	OR	95% CI
Age	−0.102	0.536	0.849	1.108	0.387–3.169
BMI	−0.775	0.585	0.185	0.461	0.147–1.449
History of pre-eclampsia	0.324	0.671	0.629	1.383	0.371–5.156
Seizures	1.221	0.615	0.047	3.39	1.016–11.307
Delivery mode	−1.299	0.69	0.06	0.273	0.071–1.056
Number of previous deliveries	0.178	0.449	0.692	1.195	0.495–2.882
Headache onset time	−1.21	0.501	0.016	0.298	0.112–0.795
Headache pattern	0.431	0.321	0.179	1.539	0.82–2.889
Constant	2.63	1.613	0.103	13.93

BMI: body mass index; SE: standard error; OR: odds ratio; CI: confidence interval; MRI: magnetic resonance imaging. P values of significant predictors are in bold.

Discussion

According to the results, the occurrence of seizure and the onset of postpartum headache in the first 5 days following delivery were significantly associated with observation of abnormal brain MRI findings. Headache due to intracranial damage is a common neurologic complaint in pregnancy and the postpartum period. Most of them are of primary headaches without clinically significant risk. However, there are several secondary headaches.³ About 40% of women experience postpartum headaches,⁷ which usually occur on the first day after delivery.⁸ The results of this study are consistent with those reported by previous studies. Tension and migraine headaches are the most common headaches after delivery.^2,4,9,10

Postpartum headache generally occurs with various diagnoses.¹¹ Several brain pathologies result in postpartum headaches. The causes of postpartum headaches can be due to physiological changes that occur during pregnancy such as eclampsia encephalopathy, postpartum cerebral angiopathy, subarachnoid hemorrhage, and cortical vein thrombosis with or without parenchymal changes. In addition, conditions not related to pregnancy such as stroke, infections, gliotic foci, and primary or metastatic brain tumors can also contribute to secondary headaches.

MRI imaging can be a useful tool for the correct diagnosis and identification of the exact cause of postpartum headache. Usually, proper diagnosis is achieved with a delay in clinical practice because the headache is initially considered eclampsia due to headaches being a common symptom. Early diagnosis of these conditions and appropriate treatment improves outcome and prognosis.⁵ MRI imaging plays an important role in detecting these lesions and is helpful in the future management of headaches. Apart from all other causes, eclampsia encephalopathy and cortical vein thrombosis are important causes of postpartum headache.¹² In a previous study,¹³ in agreement with what we observed here, 55.3% of patients with postpartum neurological findings had normal brain MRI findings. Eclampsia encephalopathy (33.3%) was the most common cause of abnormal brain MRI findings, followed by CVT (23.8%), PRES (17.4%), and ischemic stroke (12.7%).¹³ Subarachnoid hemorrhage (6.3%), pituitary apoplexy (4.8%), and pituitary adenoma (1.6%) had the lowest frequency.¹³ In another study, comparable to the current observations, 22.5% of patients with postpartum headache and seizures did not have any abnormality on brain on MRI or CT scans. CVT was the most common (42.5%) abnormal finding followed by PRES (27.5%). Infection, infection and gliosis were the least frequent findings reported.⁶ A study by Shah et al. reported that 55% of women with neurological postpartum symptoms had seizures that were accompanied by headache for more than 1 day. The abnormal MRI findings included white matter lesions and the most abnormalities were in the frontal and parietal lobes.¹⁴ In another study, the most common headache patterns were migraine and tension headaches (47%). Brain imaging findings were abnormal in 68% of patients.²

Brain MRI is more sensitive than CT scans in detecting different types of lesions. When patients have CVT, MRI findings show parenchymal changes that can be classified (using a DW sequence) as parenchymal abnormalities with cytotoxic edema or vasogenic edema or cytotoxic edema without any bleeding.⁶ These patterns can overlap in MR imaging. The abnormalities can even be seen better using MR venography with or without using contrast.¹⁵

Limitations

We were not able to follow the patients after the study and the relationship between the observed abnormalities with prognosis of the patients cannot be discussed.

Conclusion

Abnormal brain MRI findings were observed in 41.2% of patients with postpartum headache. The most common abnormal findings were sinusitis followed by PRES, CVT, and subarachnoid hemorrhage. Also, occurrence of seizures as well as postpartum headache onset in the first 5 days following delivery was recognized as a significant predictor of an abnormal brain MRI.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Conflict of interest

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.