Menstrual cycle patterns following concussion in adolescent patients

Abstract

Objective

To describe menstrual cycle patterns in concussed adolescents and investigate whether menstrual cycle phase at injury influenced post-concussion cycle pattern changes or concussion symptoms.

Study Design

Data were collected prospectively from patients aged 13-18 years presenting to a specialty care concussion clinic for an initial visit (≤28 days post-concussion) and, if clinically indicated, at a follow-up visit 3-4 months post-injury. Primary outcomes included menstrual cycle pattern change since injury (change/no change), menstrual cycle phase at time of injury (calculated using date of last period before injury), and symptom endorsement and severity, measured by Post-Concussion Symptom Inventory (PCSI). Fisher’s exact tests were used to determine the association between menstrual phase at injury and change in cycle pattern. Multiple linear regression was used to determine if menstrual phase at injury was associated with PCSI endorsement and symptom severity, adjusting for age.

Results:

Five hundred and twelve post-menarchal adolescents were enrolled (age=15.2±1.4 years), with 111(21.7%) returning for follow-up at 3-4 months. Menstrual pattern change was reported by 4% of patients at initial visit and 10.8% of patients at follow-up. At 3-4 months, menstrual phase at injury was not associated with menstrual cycle changes (p=0.40) but was associated with endorsement of concussion symptoms on the PCSI (p=0.01).

Conclusions:

At 3-4 months post-concussion, one in ten adolescents experienced a change in menses. Menstrual cycle phase at injury was associated with post-concussion symptom endorsement. Leveraging a large sample of post-concussion menstrual patterns, this study represents foundational data regarding potential menstrual cycle effects of concussion in female adolescents.

Pediatric traumatic brain injury (TBI) remains a significant public health concern affecting more than 3 million children globally each year.¹ Of these, an estimated 80% are considered to be mild TBI or concussions.¹ Sex differences in both physiological and functional outcomes are evident following TBI, with females generally demonstrating worse outcomes than males across all severities of TBI, and worse outcomes following concussion relative to moderate and severe TBI.² Recent literature posits that these differences may be due to hormonal differences, suggesting that female sex hormones, including estrogen and progesterone, may play a role in the neural consequences following brain injury.^2,3

Neuroendocrine dysfunction is common following TBI and can negatively affect patient quality of life.⁴ A recent systematic review estimated that 15-46.6% of sport-related adult TBI patients (mean age range 20-48 years) experience post-injury pituitary dysfunction.⁵ Patients with post-TBI neuroendocrine dysfunction often present with a broad spectrum of nonspecific symptoms, many of which overlap with common post-TBI symptoms (eg, fatigue, headache, difficulty concentrating, etc.)^6,7 Current research suggests that the pituitary gland may be vulnerable to both primary injury (direct impact or shear stress) and secondary injury (hypoxia, edema, etc.) associated with TBI.^4,8 Estrogen and progesterone are produced by the hypothalamic-pituitary-gonadal axis which, if disrupted, may result in dysregulation of the menstrual cycle.⁹ In studies of adult women with moderate to severe TBI, a majority of patients experienced menstrual cycle pattern changes following injury,¹⁰ with injury severity predicting presence and duration of amenorrhea.^10,11

Research assessing pituitary hormone dysfunction following concussion has primarily focused on male or adult populations (eg, boxing/kickboxing,¹² retired professional football athletes, ¹³ and military personnel^14-16). There are currently few studies assessing the menstrual cycle implications of concussion in female patients, and even fewer in adolescent female patients. When examining the intersection between the menstrual cycle and concussion, research, mainly on adults,^3,17-19 has examined both how menstrual cycle phase at injury influences clinical presentation of concussion and menstrual cycle pattern changes following injury. One theory suggests that TBIs occurring in settings of high progesterone (ie, luteal phase) lead to sudden drops in progesterone and therefore worse clinical outcomes relative to settings of low progesterone (ie, follicular phase).²⁰ Similar to studies in moderate and severe TBI,²¹ Chen et al¹⁷ found higher circulating progesterone concentrations to have a protective effect following concussion in college-aged athletes, while another study of college-aged patients showed menstrual cycle phase at time of injury was not associated with performance during a standard concussion test battery (neurocognitive testing, postural stability, and symptom reporting).¹⁸ In patients aged 16-60 years old, other literature has described lower quality of life scores 1-month post-injury in those injured during the luteal phase than those injured during the follicular phase.¹⁹ One of the only studies to date focusing on adolescents and young women found that concussions were associated with experiencing 2 or more subsequent abnormal bleeding patterns;³ however, associations with menstrual cycle phase and/or clinical outcomes were not examined and that study did not focus exclusively on adolescents but included ages 12-21. The neuroendocrine effects of concussion on adolescent patients specifically, and associations with clinical outcomes remain unclear.

The primary purpose of this study was to leverage a large registry of prospectively collected data in concussed adolescents to describe menstrual cycle pattern changes within 28 days and 3-4 months following concussion. Our secondary purpose was to investigate whether menstrual cycle phase at time of injury was associated with menstrual cycle changes or symptom reporting within 28 days and 3-4 months post-concussion.

Methods

This study queried prospectively collected data from the Minds Matter Concussion Registry using electronic health records for patients seen for concussion within the Children’s Hospital of Philadelphia pediatric network. The population for this analysis was limited to patients aged 12-18 years, presenting between January 1, 2018, and June 30, 2021, for their initial visit to the specialty care concussion program within 28 days of injury. Patients were diagnosed by a sports medicine physician using the definition of concussion set forth in the Consensus Statement on Concussion in Sport.²² Patients were included in the follow-up visit analysis if they were seen again 90-120 days post-concussion, based on clinical need. The derivation of the study sample is described in Figure 1 (online). The study was approved by the Children’s Hospital of Philadelphia Institutional Review Board (IRB# 19-016019) with a waiver of consent/assent for use of the data collected in the electronic health record for the Minds Matter Concussion Registry. The research was limited to existing data and involved no more than minimal risk to the patients.

Figure 1.

Flow diagram of patient population

Prior to the initial visit, patients completed a questionnaire that included demographic information and patient-reported medical history, including contraceptive use. Body mass index (BMI), which can be a confounder for menstrual irregularity,²³ was recorded and calculated as a z-score. Menstrual cycle data were collected through the following questions: 1) What were the patient’s menstrual patterns before the injury? (monthly, less frequently than monthly, more frequently than monthly, irregular, not applicable/has not had period); 2) Has the patient had a change in menstrual patterns since the injury? (no change, less frequent than before the injury, more frequent than before the injury, more irregular than before the injury, not applicable/has not had period); and 3) On what date did the patients’ last period before the injury start? Patients were excluded if they reported an invalid or missing date of injury or date of first concussion (eg, clinician or patient failed to document date; missing month, day, or year; or documentation error [date of injury or first concussion documented is date of birth or a date in the future]), or had incomplete or missing Post-Concussion Symptom Inventory (PCSI) or menstrual cycle pattern data, invalid date of last period prior to concussion, or invalid self-reported contraceptive use. Menstrual cycle phase at time of injury was calculated by subtracting the date of injury from the date of last period prior to injury. For the purposes of this study, a menstrual cycle was defined as 28 days and the phases were categorized as follows: menstruation (1-4 days); follicular phase (5-14 days); early luteal phase (15-21 days); and late luteal phase (22-28 days).²⁴ Patients reporting a longer menstrual cycle (29-35 days prior to injury) (n=26, 5.1%) were excluded only from analyses examining menstrual cycle phase at time of injury. At follow-up, patients were asked if their menstrual pattern had changed since the last visit (no change, more irregular, more frequent, less frequent). Concussion-related symptoms were self-reported using the PCSI, which is a 22-item scale measuring symptom endorsement (range=0-22) and symptom severity on a Likert-scale from 0-6 (range=0-132).

Descriptive statistics were used to describe patient demographics in patients at initial visit and those who returned for a follow-up, and to compare demographic information for patients excluded for incomplete, invalid, or missing data with patients with complete data. Chi-square tests were used for categorical variables (race/ethnicity, sport-related concussion), independent samples t-tests were used for continuous variables (age, BMI z-score), and Mann-Whitney U tests were for number of lifetime concussions due to non-normal distribution. For our primary analysis, descriptive statistics were used to describe menstrual cycle pattern changes following concussion in pediatric patients at initial visit and 3–4-month follow-up. To analyze the relationships between menstrual cycle phase at time of injury and changes in menstrual cycle pattern, changes in menstrual cycle pattern at initial visit were categorized as “change” (less frequent, more frequent, more irregular), or “no change.” Generalized linear models with a log-link function were used to determine the odds of experiencing a change in menstrual cycle pattern at initial and follow-up visits based on menstrual cycle phase at time of injury, while adjusting for age. We explored menstrual cycle change while adjusting for BMI z-score; however, the results were not significant and therefore BMI z-score was removed from our models. Multiple linear regression was used to determine if menstrual phase at injury was associated with PCSI symptom severity at initial visit and follow-up, adjusting for age. Alpha was set to 0.05 a priori. For our secondary analysis, we completed pairwise comparisons with Bonferroni corrections (α=0.008) to explore differences in symptom endorsement and severity between menstrual cycle phases at time of injury. Data were analyzed using SAS statistical software, version 9.4 (SAS Institute Inc., Cary, NC).

Results

A total of 512 patients (age 15.2 ± 1.4 years; BMI z-score 0.6 ± 0.8) were included in our final analysis (Table I). Their initial visit was 12.5 ± 7.2 days post injury. Of these, 111 (21.7%) completed a clinical follow-up visit 3-4 months after concussion, a proportion consistent with previously reported rates of persistent post-concussive symptoms.²⁵ At initial visit, most patients presented with a sport-related concussion (n=286, 55.9%), and 50.6% reported having a history of previous concussion (n=259, lifetime median=2, IQR[1-2]).

Table 1.

Patient demographics at initial visit and follow-up visit.

	Initial Visit n=512	Follow-Up Visit n=111	p-value
Age, years, mean(SD)	15.2 (1.4)	15.4 (1.4)	0.25
Race/Ethnicity, n(%)			0.72
Non-Hispanic White	378 (73.8)	77 (69.4)
Non-Hispanic Black	46 (9.0)	13 (11.7)
Hispanic or Latino	25 (4.9)	7 (6.3)
Non-Hispanic Asian/Asian Pacific	63 (12.3)	14 (12.6)
Islander/other/multiple race/unknown
BMI Z-Score	0.6 (0.8)	0.6 (0.8)	0.34
Lifetime Concussions, median(IQR)	2 (1,2)	1 (1,3)	0.74
Days Since Injury, mean(SD)	12.5 (7.2)	104.2 (9.1)	<0.001
Sport-Related Concussion, n(%)	286 (55.9)	36 (32.4)	0.2

SD: standard deviation

BMI: body mass index

IQR: interquartile range

Initial Visit (within 28 days after concussion)

At initial visit, 82.4% (422/512) of patients reported having monthly menstrual cycles prior to the concussion, with 11.5% (59/512) having irregular cycles, 4.3% (22/512) less frequently than monthly, and 1.8% (9/512) more frequently than monthly. Four percent (18/512) reported changes in menstrual cycle pattern post-concussion, while 84.8% (434/512) reported no change and 11.7% (60/512) had not yet had their period since injury. Of the patients reporting changes, 44.4% (8/18) reported their menstrual pattern cycle as “more irregular,” 44.4% (8/18) reported their menstrual pattern cycle as “less frequent,” and 11.1% (2/18) reported their menstrual pattern cycle as “more frequent.” Of patients reporting with non-monthly cycles prior to concussion (n=90, 17.6%), 8.9% (8/90) reported changes in menstrual cycle pattern post-concussion (50.0% [4/8] “more irregular” and 50.0% [4/8] “less frequent”), 78.9% (71/90) reported no change, and 12.2% (11/90) had not yet had their period since injury. Most patients reported having sustained their concussion during the follicular phase (n=234, 48.0%), followed by early luteal phase (n=91, 18.7%), menstruation (n=89, 18.2%), and late luteal phase (n=74, 15.2%). Among patients who had had their period since injury (n=452), menstrual cycle phase at time of injury was not associated with experiencing any change in menstrual cycle pattern (p=0.57), after adjusting for age (Table II). Additionally, menstrual cycle phase at time of injury was not associated with either self-reported concussion symptom endorsement (p=0.24) or severity (0.67) at initial visit, adjusting for age (Table III).

Table 2.

Menstrual pattern change since injury by menstrual cycle phase at time of injury for initial (≤28 days post-injury) and follow-up visit (3-4 months post-injury). P-values reflect models adjusting for age.

	Menstrual Cycle Phase at Time of Injury
Menstrual Pattern Since Injury	Menstruation (days 0-4)	Follicular (days 5-14)	Early Luteal (days 15-21)	Late Luteal (days 22-28)	p-value
Initial Visit, n(%)					0.57
No change	74 (83.2)	201 (85.9)	73 (80.2)	62 (83.8)
Change	4 (4.5)	6 (2.6)	5 (5.5)	3 (4.1)
n/a (has not had period)	11 (18.3)	27 (45.0)	13 (14.3)	9 (12.2)
Follow-Up Visit, n(%)					0.40
No change	19 (100.0)	50 (92.6)	19 (86.4)	9 (75.0)
Change	0 (0.0)	4 (7.4)	3 (13.6)	3 (25.0)

Table 3.

Symptom reporting by menstrual cycle phase at time of injury for initial (≤28 days post-injury) and follow-up visit (3-4 months post-injury). P-values reflect models adjusting for age.

	Menstrual Cycle Phase at Time of Injury
Post-Concussion Symptom Inventory	Menstruation (days 0-4)	Follicular (days 5-14)	Early Luteal (days 15-21)	Late Luteal (days 22-28)	p-value
Initial Visit, mean(SD)
Total Endorsement	12.9 (6.4)	14.3 (5.5)	13.9 (5.9)	14.6 (5.1)	0.24
Symptom Severity	41.3 (29.5)	45.9 (27.0)	44.9 (28.3)	46.2 (24.0)	0.67
Follow-Up, mean(SD)
Total Endorsement	12.9 (5.9)	8.7 (6.0)	12.9 (6.2)	9.9 (7.1)	0.01
Symptom Severity	30.4 (20.1)	22.4 (21.8)	32.3 (23.3)	25.0 (28.5)	0.31

Follow-Up Visit (3-4 months after concussion)

Among patients seen at follow-up, 10.8% (12/111) reported experiencing a change in menstrual cycle pattern, with 58.3% (7/12) reporting more irregular, 33.3% (4/12) reporting less frequent, and 8.3% (1/12) reporting more frequent menstrual cycle patterns. Menstrual cycle phase at time of injury was not associated with experiencing any change in menstrual cycle pattern at follow-up (p=0.40) (Table II) or self-reported symptom severity (p=0.31) but was associated with self-reported symptom endorsement (p=0.01). (Table III). For our secondary analysis, patients injured in the follicular phase trended toward lower symptom endorsement at follow-up relative to those injured in menstruation (p=0.05) or early luteal phase (p=0.07); however, these differences did not maintain statistical significance after adjusting for multiple comparisons (Table IV).

Table 4.

Pairwise comparisons between each menstrual cycle phase at time of concussion for symptom endorsement and symptom severity at follow-up visit.

Menstrual Cycle Phase Comparison (A vs. B)	Mean Difference (A-B)	P-valuea	95% Confidence Limits
			Lower Bound	Upper Bound
PCSI Symptom Endorsement
Menstruation vs. Follicular	4.4	0.05	0.0004	8.79
Menstruation vs. Early Luteal	0.37	1	−4.82	5.57
Menstruation vs. Late Luteal	3.22	0.94	−2.86	9.3
Follicular vs. Early Luteal	−4.02	0.07	−8.21	0.16
Follicular vs. Late Luteal	−1.18	1	−6.43	4.08
Early Luteal vs. Late Luteal	2.84	1	−3.07	8.76
PCSI Symptom Severity
Menstruation vs. Follicular	8.32	1	−7.94	24.59
Menstruation vs. Early Luteal	−0.96	1	−20.17	18.25
Menstruation vs. Late Luteal	5.92	1	−16.58	28.42
Follicular vs. Early Luteal	−9.28	0.66	−24.76	6.2
Follicular vs. Late Luteal	−2.4	1	−21.85	17.04
Early Luteal vs. Late Luteal	6.88	1	−15	28.76

PCSI: Post-Concussion Symptom Inventory

^aBonferroni adjusted p-value = 0.008

Discussion

This study leveraged a large, prospective clinical concussion registry to examine menstrual cycle patterns following concussion in adolescent patients, contributing foundational data to sparse literature on this topic. Our findings indicate that menstrual cycle phase at time of injury was not associated with menstrual cycle pattern changes within 28 days of concussion or 3-4 months following injury; however, menstrual cycle phase at time of injury may be associated with self-reported symptoms 3-4 months post-concussion.

Our study contributes novel data exclusively on adolescents to previous research examining self-reported menstrual cycle patterns following concussion, which has produced mixed results and has primarily been conducted in college-aged and adult populations. When examining menstrual cycle phase at time of injury in a small sample of college-aged athletes (n=16), 50% of concussions occurred during the late luteal phase.²⁴ In contrast, our results, based on over 500 female adolescents, indicate that most concussions occurred during the follicular phase. Based on the number of days in each defined phase relative to a 28-day cycle, we expected 35.7% of concussions to occur in the follicular phase, followed by 25% in both the early and late luteal phase, and 14.3% in menstruation. Our results found a larger proportion occurred in both the menstruation and follicular phase and a smaller proportion occurred in early and late luteal phases relative to expected. Although we used the same operational definition for menstrual cycle phase and self-report method of days between most recent menstruation and concussion as previous work,²⁴ our differing results may be due to our larger sample size or possibly a younger age group which may have less experience estimating the date of their last period and is more likely to have variable menstrual cycle lengths.

When assessing the effect of menstrual cycle on clinical and patient-reported outcomes following concussion, there are also mixed results. Our findings indicate that patients with concussions during early luteal phase may experience greater symptom burden at follow-up relative to those who sustained their concussion during the follicular phase. While differences in symptom endorsement between concussions occurring during the follicular, early luteal, and menstruation phases did not reach significance after adjusting for multiple comparisons, the mean differences may be clinically meaningful (as evidenced by the significant main effect of phase on symptoms) and warrant continued investigation. Compared with those injured in the follicular phase, patients injured during the menstruation or early luteal phase reported an average of 4.4 or 4.0 more symptoms, respectively. Additionally, patients injured in the early luteal phase, compared with those injured in the follicular phase, reported a symptom severity score that was on average 9.3 higher. Recent literature described a symptom severity score cutoff of 7 to differentiate significantly those athletes concussed from uninjured controls with 89.2% accuracy,²⁶ and a symptom change of 3 or more (increase in severity or endorsement) as clinically meaningful symptom exacerbation in exercise tolerance protocols after concussion.²⁷ In this context, our findings demonstrate greater group differences in both symptom endorsement and severity than these previously published clinically relevant differences in symptom burden, indicating that these differences are likely clinically important. In a broader age range (16-60 years), Wunderle et al²⁰ found that women injured during the luteal phase demonstrated worse concussion symptoms and quality of life one month after injury relative to women injured during the follicular phase while Mihalik et al¹⁸ found that menstrual cycle phase at time of injury had no effect on post-concussion clinical outcomes in college-aged females including neurocognitive testing, post-concussion symptom score, or postural control. Our findings align with these studies in that we found no associations between menstrual cycle phase at time of injury and symptoms within 28 days of concussion, but that menstrual cycle phase at time of injury had an effect on symptom reporting at follow-up 3-4 months later. This suggests that if menstrual cycle pattern changes occur after concussion, they may not present clinically in the acute or subacute phase. This finding may be due to timing of assessment, method of measurement used in these studies (eg, relying on self-report, defining a 28-day cycle length), or may reflect actual differences in the severity or manifestations of concussion.

There are fewer studies examining the influence of concussion on menstrual cycle patterns in the adolescent population. This could be due to the variability of ovulatory menstrual cycles in adolescent patients. In the US, menarche occurs at an average age of 12.8 years²⁹ and onset of regular ovulatory cycles occurs in just 18-45% of adolescent girls within 2 years of menarche,²⁸ making the study of cycle pattern changes between the ages of 12-14 more difficult. Among our patients completing a follow-up visit, 71% (79/111) were 15-18 years old, indicating that a majority of our sample were beyond the range of variable ovulatory cycles. Snook et al³ investigated abnormal menstrual cycle patterns following sport-related injury in adolescent and young adult athletes (aged 12-21 years), finding that concussed athletes had a higher risk of experiencing 2 or more abnormal cycles relative to athletes with orthopedic injuries. Our study adds to this very limited literature by focusing exclusively on adolescents, finding that changes in menstrual cycle patterns in concussed adolescents were not associated with the menstrual cycle phase at time of injury, suggesting that if patients experience menstrual cycle pattern changes after concussion, these changes may not be attributable to the menstrual phase in which the concussion occurs.

Importantly, our study represents key foundational data from a large concussion registry that expands on sparse existing literature examining menstrual cycle patterns following concussion in adolescent patients. The association between menstrual cycle and sport-related injury has primarily focused on muscular, tendinous, and ligamentous injuries, with evidence that menstrual cycle phase influences tissue laxity and/or stiffness,³⁰ specifically when examining anterior cruciate ligament injuries.^31-33 In both the lower extremity injury and concussion literature, the methodologies of determining or defining menstrual cycle phase varies between the use of biomarkers (eg, serum or urine samples) and self-reported questionnaires.^{3,17,18,20,24,31,32} Studies assessing self-reported menstrual cycle length have demonstrated considerable error, such that most women tend to overestimate their cycle length,³⁴ with younger women (≤ 25 years) showing greater odds of inaccurate self-reported cycle length.³⁵ A recent abstract by Ott et al³⁶ examined self-reported menstrual cycle phase and phase determined by blood plasma concentrations in concussed adolescent patients and found 54.5% agreement identifying the follicular phase and 75% agreement identifying the luteal phase. Based on these findings, we acknowledge the limitations of categorizing menstrual phase by self-report, but in the context of the very sparse literature on this topic, consider our retrospective analyses of a large clinically collected dataset useful foundational insight into the relationship between menstrual cycle patterns and adolescent concussion. There is a need for further investigations comparing self-reported and physiological metrics of menstrual cycle phase, particularly in adolescent patients, in order to inform our conclusions regarding concussion-related menstrual cycle dysfunction in this age group with more objective data.

Our study was not without limitations. We relied on patient self-report for date of last period prior to concussion, menstrual cycle pattern before concussion, and changes in menstrual cycle patterns since concussion which is subject to patient interpretation of health history questions and available responses. We did not collect date of menarche and were unable to account for time since menarche in our analysis. Additionally, we excluded patients with a missing or invalid date of last period. These data are subject to recall bias and missing or invalid data not likely at random, which reduces the generalizability of the study. This study used a previously published definition of menstrual cycle phase²⁴ which assumes a 28-day cycle and may not be generalizable to patients with shorter or longer cycles. In the current sample, 26 (5.1%) patients were excluded due to reporting cycle length longer than 28 days. The menstrual cycle phase definition used in the current study was previously used to assess changes in phase patterns following sport-related concussion, enabling comparison of our results across other studies.²⁴ Menstrual cycle phase has been defined numerous different ways using both self-reported date of last period^3,18,24 and biomarkers^17,20,37 and future research should continue to investigate the validity of clinically applicable, self-reported metrics of menstrual cycle function relative to objective biomarkers. The absolute number of patients with menstrual changes at 3-4 months follow-up was small and could introduce errors from limited sample size. We did not examine clinical follow-up beyond 3-4 months post-injury and are unable to address the prevalence of more long-term menstrual cycle pattern changes. The patients in our sample were seen by physicians in a specialty care concussion program and 44% (226/512) of our patients sustained a non-sports related concussion, therefore our results may not be generalizable to patients seen in other health care settings (e., emergency department, primary care, urgent care) or an exclusively sports-related concussion population.

In summary, our study found that approximately one in ten adolescent patients at 3–4-month follow-up had experienced a change in menstrual cycle pattern following concussion. Additionally, menstrual cycle phase at time of injury was associated with concussion symptom endorsement, but not a change in menstrual cycle pattern. Leveraging the largest study sample to date of menstrual patterns in adolescents with concussion, this study represents key foundational data on the potential menstrual cycle effects of concussion in this age group.

Abbreviations:

TBI

Traumatic brain injury

PCSI

Post-concussion symptom inventory