Postpartum Depression in Patients with Chronic Migraine: A Retrospective,Observational Analysis

Introduction

Postpartum depression (PPD) occurs in approximately 14% of women. ¹ Numerous risk factors, spanning social, physical, and psychological domains, contribute to the incidence of PPD, including estrogen withdrawal, high-risk pregnancies, history of depression, and limited social support. ^1,2 Recognizing these risk factors is essential for physicians in screening for PPD and facilitating early intervention.

Few studies have investigated the link between migraine and PPD in women of reproductive age. The association between PPD and migraine is complex, with proposed theories including shared neurobiological factors such as the hypothalamic-pituitary-adrenal (HPA) axis, the immune system, dysfunction within the dopaminergic and serotonergic systems, and fluctuations in sex hormones. ³ Of these, estrogen fluctuations are particularly noteworthy; the rapid decline in estrogen in the late luteal phase can serve as a potent trigger for attacks of migraine, and the postpartum period is characterized by a significant decrease in estrogen levels following childbirth. ⁴

In one study, a comprehensive multivariate analysis including 222 women diagnosed with comorbid major depressive disorder (MDD) revealed a specific correlation between the lifetime presence of migraine and PPD within 6 weeks of delivery, with an odds ratio (OR) of 2.55; however, its exclusive focus on women with a prior MDD diagnosis limits its generalizability. ⁵ A cross-sectional study in Lima, Peru, demonstrated that pregnant women with a history of migraine had an adjusted OR of 2.23 for severe depressive symptoms compared to pregnant women without a migraine history. ⁶ We aimed to extend these findings by examining the association between chronic migraine and PPD relative to non-migraine controls and to assess the impact of comorbid conditions on this association.

Methods

All data were obtained through the Stanford Research Repository (STARR) and this project received approval from the Stanford Institutional Review Board. Groups for assessment were built on different criteria but were drawn from female patients, aged 18–45 years, who were seen in the Stanford Headache Clinic, a tertiary headache center, or the Comprehensive Neurology Clinic in the Department of Neurology between January 1, 2017, to January 24, 2024. Demographics for each group were determined including race, ethnicity, and average age.

Patient groups were then built based on International Classification of Disease, Tenth Edition (ICD-10) codes that were assigned during the observed time period.

To build our initial group of patients, we examined PPD in migraine and its subtypes relative to age and sex-matched controls drawn from the neurology clinic and who never received an ICD-10 code diagnosis of G43 “Migraine” or higher resolution. As a tertiary care center, we elected to draw our controls from our broader neurology clinic and not an obstetrics prenatal clinic. This was out of a concern that there could be a significant selection bias in seeing a tertiary neurology clinic that would not be present if controls were drawn from it as well. Additionally, we believe that the integrity of a diagnosis of migraine and higher resolution codes would be more likely to be valid in a tertiary neurology clinic. PPD was defined as ever having an ICD-10 code of F53.0 “Postpartum Depression” in the chart diagnoses. Migraine groups included Episodic Migraine, defined as the group with a G43.0 “Migraine without Aura” or G43.1 “Migraine with Aura” ICD-10 code but who had never received a G43.7 “Chronic Migraine” diagnosis; Episodic Migraine without Aura, defined as the group with a G43.0 ICD-10 code but who had never received a code of G43.1 or G43.7; Episodic Migraine with Aura, defined as the group with a G43.1 ICD-10 code who had never received a code of G43.7; Chronic Migraine, defined as the group with a G43.7 ICD-10 code; Chronic Migraine without aura, defined as the group with a G43.7 ICD-10 code but who had never received a diagnosis of G43.1; and Chronic Migraine with aura, defined as the group who had received both G43.7 and G43.1 ICD-10 codes.

To then control for the presence of comorbidities that may independently contribute to risk for PPD, we assessed for the presence of these comorbidities through additional ICD-10 codes; F32 and F33 for depression; F41 for anxiety; F43.7 for post-traumatic stress disorder (PTSD); N80 for endometriosis; O14 for preeclampsia; O24.4 for gestational diabetes; F32.81 for pre-menstrual dysphoric disorder (PMDD). Advanced maternal age was also assessed by dividing the population of chronic migraine into two groups divided by median age: 18–34 and 35–45.

The primary outcome was the OR for PPD in migraine, episodic and chronic as well as with and without aura, relative to women without migraine. Secondary outcomes included the OR for PPD in women with chronic migraine who also had depression, anxiety, PTSD, endometriosis, preeclampsia, gestational diabetes, PMDD, and older maternal age relative to those without. This is the primary analysis of this data.

Results

A total of 6248 patients with migraine and 4154 patients without migraine met inclusion criteria. Demographics and presence of comorbid conditions for patients with episodic migraine, chronic migraine, and those without migraine are presented in Table 1.

In the unadjusted model, both episodic migraine and chronic migraine, with and without aura, were associated with statistically significant increased ORs for PPD (Table 2). However, after adjusting for demographic factors and comorbid medical conditions, only chronic migraine, with and without aura, had a statistically significant increase in OR for PPD. Notably, chronic migraine without aura had the highest OR for PPD (OR: 2.13; 95% CI: 1.29 to 3.53, p = 0.003).

The impact of chronic migraine comorbidities, and advanced maternal age, on PPD were then assessed (Table 3). After controlling for demographic factors, preeclampsia was associated with the largest OR for PPD in patients with chronic migraine (OR: 7.91; 95% CI: 4.86 to 12.90, p < 0.001), followed by depression (OR: 4.03; 95% CI: 2.56 to 6.36, p < 0.001), gestational diabetes (OR: 3.87; 95% CI: 2.30 to 6.52, p < 0.001), and PMDD (OR: 2.89; 95% CI: 1.30 to 6.47, p = 0.009). Anxiety, advanced maternal age, endometriosis, and PTSD were not associated with increased OR for PPD in patients with chronic migraine.

Discussion

Our analysis demonstrated a significant association between chronic migraine, with and without aura, and PPD when compared to non-migraine controls but not episodic migraine. Our findings indicate that certain comorbid conditions, notably preeclampsia, depression, gestational diabetes, and PMDD independently elevate the risk of PPD among patients with chronic migraine.

Previous studies have examined the link between chronic migraine and PPD; however, these studies focused on women with a previous diagnosis of MDD. ⁵ Moreover, the relationship between chronic migraine and PPD in women with MDD was only observed during the initial 6 weeks following childbirth. Our analysis extends this to show migraine as an independent risk factor for the lifetime occurrence of PPD in individuals without a history of diagnosed comorbid depression.

The association is more pronounced in patients with preeclampsia and chronic migraine, exhibiting a high OR of 7.91. Previous meta-analyses have underscored this correlation, linking preeclampsia not only to an increased risk of PPD but also to heightened severity of depressive symptoms. This correlation may be attributed to shared etiologies of both conditions, including elevated pro-inflammatory cytokines, endothelial dysfunction affecting neurotransmitter function, and genetic alterations. ⁷ Migraine also independently elevates the risk for preeclampsia, as both share underlying mechanisms such as abnormal vascular reactivity. ⁸ Similarly, a meta-analysis identified an independent link between gestational hypertension and PPD, with a relative risk of 1.59 (95% CI: 1.22–2.07). This relationship may stem from disruptions in the HPA axis, inflammatory processes, and serotonergic system dysregulation. ⁹ Preeclampsia and gestational hypertension may increase PPD risk through shared biological pathways and their status as high-risk pregnancy complications, potentially causing trauma for the mother during pregnancy. ¹⁰

It was not surprising that depression was an independent risk factor for PPD among women with chronic migraine. The bidirectional relationship between depression and chronic migraine, well-established in the literature, suggests that depression may arise from the psychological response to debilitating headache episodes, or that both conditions may share common underlying causes. ³ Notably, depression can exacerbate migraine symptoms, leading to refractory-to-treatment and medication-overuse headaches. ¹¹ Even after controlling for depression, chronic migraine remained an independent risk factor for PPD. This association may be attributed to various social, biological, and psychological factors, including shared hormonal etiologies, disruptions in the HPA axis, and the psychological strain of managing a chronic condition amid the inherent stress of the postpartum period.

Moreover, PMDD has been robustly associated with an increased risk of both chronic migraine and PPD. Migraine attacks are often triggered by estrogen withdrawal during the late luteal phase of the menstrual cycle, corresponding with the onset of PMDD symptoms. ¹² Similarly, PMDD is a recognized independent risk factor for PPD, possibly because PMDD is caused by an abnormal response to hormonal fluctuations during the menstrual cycle, as well as an underlying serotonin transporter polymorphism, leading to tryptophan depletion and triggering PPD. ^13,14

Unlike prior studies, we did not observe that advanced maternal age independently influenced the risk of PPD in individuals with migraine. Previous population-based investigations have demonstrated a statistically significant rise in the incidence of PPD among younger mothers (aged 15–24 years) and older mothers (aged >36 years) lacking a history of depression. ¹⁵ Notably, this risk was subject to modification by the presence of prior depression. Numerous studies corroborate the relationship between maternal age and PPD risk, including community-based inquiries conducted in Canada and Japan, revealing that woman aged 40–44 years faced a heightened risk of PPD compared to their younger counterparts. ^16,17

Additionally, the effects of anxiety, PTSD, and endometriosis on the risk of PPD in chronic migraine was minimal; there was no increase in the OR for PPD in patients with those comorbidities. The lack of a significant association suggests that these conditions may not independently contribute to the risk of experiencing PPD in this population. These findings diverge from prior research suggesting that anxiety and PTSD serve as independent risk factors for PPD. ¹⁸ The inclusion of women with endometriosis and chronic migraine in the analysis provided a comparable chronic pain group; existing literature indicates that endometriosis leads to an escalation in depressive symptoms, particularly among individuals experiencing substantial pain. ¹⁹ Yet, in patients with chronic migraine and endometriosis, the presence of endometriosis did not contribute an additional risk toward PPD. Future research endeavors should delve into the intricate interplay between these conditions and their collective impact on the risk of PPD.

Numerous studies have investigated specific treatments for individuals with both chronic migraine and PPD, grounded in the understanding that these conditions share hormonal etiologies. For instance, initiating a combined oral contraceptive or estrogen supplementation regimen can prevent both migraine and PPD. ²⁰ Using combined oral contraceptives could be particularly beneficial for women with a history of PMDD, as combined oral contraceptives manage this psychiatric condition by inducing anovulation and low hormone levels. ²¹ Of note, the American College of Gynecology (ACOG), does not recommend estrogen-containing contraceptives in patients with migraine with aura due to increased risk of stroke; however, more recent studies have called the extent of that increase into question and suggest patients may be counseled regarding the risks and benefits. ^22,23 Patients with additional comorbidities may benefit from treatment with selective serotonin reuptake inhibitors or other mood stabilizers, which are generally considered safe for use during breastfeeding. ²⁰ Tricyclic antidepressants such as amitriptyline are effective in both migraine prevention and treating depressive symptoms, and may be prescribed to breastfeeding mothers. ^24,25 Selective norepinephrine reuptake inhibitors such as venlafaxine and duloxetine can also be helpful in treating patients with chronic migraine and concurrent PPD. ²⁵

Even before beginning treatment, when patients initiate care during the perinatal period, physicians ought to evaluate not just PPD alone but also accompanying factors that can affect both their headache severity and elevate the likelihood of developing PPD. The Edinburgh Postnatal Depression Scale (EPDS) is a validated screening tool routinely administered at the 6-week postpartum appointment to screen for the presence and severity of depressive symptoms. Patients with chronic migraine should undergo screening before the 6-week postpartum appointment, and during prenatal and early postpartum visits. Moreover, collaborating with a multidisciplinary team comprised of reproductive psychiatrists, OBGYNs, social workers, physical therapists, and headache specialists is crucial to address both the patient’s primary headache condition and psychiatric concerns effectively.

With migraine prevalence reaching its peak during childbearing years, the association between chronic migraine and PPD is a relevant problem. In addition, PPD can profoundly impact both new mothers and their children and is a leading cause of disability worldwide. Developing a validated scoring system to assess the relative influence of risk factors on the development of PPD will substantially aid treatments, much like the Clinical Risk Assessment for Preeclampsia tool aids in determining which patients will benefit from aspirin for preeclampsia prevention. ²⁶ Cheryl Beck has created the PPD Predictors Inventory with a list of 13 risk factors for developing perinatal depression, however, a history of migraine is not included among these. ²⁷ Further research would benefit from exploring whether the frequency of migraine (with or without aura) at specific perinatal time points (prenatal, pregnancy, labor, postpartum) differentially influences the risk for PPD. ²⁸ Understanding the interplay between migraine frequency, severity, aura presence during pregnancy with PPD can in turn influence the need for migraine treatment during various perinatal time points. Given that migraine improves among 80% of women with migraine without aura during pregnancy, migraine prophylactic treatments are often unnecessary during pregnancy. In contrast, women with a history of depression, often benefit from treatment with an SSRI to prevent PPD. ²⁹

Our study is constrained by the characteristics of the dataset, which affects the generalizability of the findings and reduces the overall statistical power of our study. Furthermore, to abstract data for the retrospective chart review, we utilized the International Classification of Disease-10th edition (ICD-10) codes. However, these codes may not fully adhere to the precise criteria for the conditions we examined. For example, patients diagnosed with PPD may not have met strict DSM-5 diagnostic criteria. We included women who had experienced any lifetime episode of PPD instead of verifying if they received this diagnosis within the 4-week onset period as specified by the DSM-5. Nonetheless, PPD is commonly diagnosed based on an EPDS score of 12 or above, and diagnostic practices vary among physicians; many physicians extend the 4-week time frame to the full year after childbirth. Therefore, the ICD-10 code serves as a close proxy for a PPD diagnosis. Similarly, the migraine group identified through the ICD-10 code may not all fulfill the criteria outlined in the International Classification of Headache Disorders-3rd edition.

One confounding variable could be the level of healthcare access and engagement among patients diagnosed with PPD who also have chronic migraine. These individuals may have increased interaction with the medical system. This heightened medical interface could result in earlier detection and diagnosis of PPD in patients with chronic migraine, whereas it might be missed in individuals with limited access to healthcare resources or a decreased need to consistently engage with medical care. To help control for this, we used a control group of age and sex-matched patients from our tertiary neurology clinic without migraine who would have a similar degree of interaction with the health care system.

Our research highlights a significant association between chronic migraine and lifetime occurrence of PPD, emphasizing the importance of comprehensive healthcare management for individuals experiencing these conditions. Future prospective trials could explore additional comorbidities that influence this relationship, providing a more nuanced understanding of the interplay between chronic migraine and PPD.