Menstrual cycle impacts lung structure measures derived from quantitative computed tomography

Abstract

Objective:

Quantitative computed tomography (qCT) is being increasingly incorporated in research studies and clinical trials aimed at understanding lung disease risk, progression, exacerbations, and intervention response. Menstrual cycle-based changes in lung function are recognized, however, the impact on qCT measures is currently unknown. We hypothesize that the menstrual cycle impacts qCT derived measures of lung structure in healthy women and that the degree of measurement change may be mitigated in subjects on cyclic hormonal birth control.

Methods:

Thirty-one non-smoking, healthy women with regular menstrual cycles (16 of which were on cyclic hormonal birth control) underwent pulmonary function testing and qCT imaging at both menses and early luteal phase timepoints. Data were evaluated to identify lung measurements which changed significantly across the two key timepoints and to compare degree of change across metrics for the sub-cohort with versus without birth control.

Results:

The segmental airway measurements were larger and mean lung density was higher at menses compared to the early luteal phase. The sub-cohort with cyclic hormonal birth control did not have less evidence of measurement difference over the menstrual cycle compared to the sub-cohort without hormonal birth control

Conclusions:

This study provides evidence that qCT derived measures from the lung are impacted by the female menstrual cycle. This indicates studies seeking to use qCT as a more sensitive measure of cross sectional differences or longitudinal changes in these derived lung measurements should consider acquiring data at a consistent time in the menstrual cycle for pre-menopausal women and warrants further exploration.

Introduction

Chronic inflammatory disorders of the lung such as asthma and chronic obstructive pulmonary disease (COPD), are characterized by airflow limitation. Although the underlying pathologies of the diseases are different, they both induce structural change and deficits in pulmonary function. Much is still not understood about the intra- and inter-subject differences in the regional relationships between structural and functional metrics assessed via emerging imaging methodologies, but certainly, gender plays an important role [1; 2]. A large percent of the female population undergoes cyclic biological variability associated with the menstrual cycle; with the average age of menarche (first menstrual cycle) being 12.5 years and menopause is 51 years [3]. These cyclic variations have been associated with changes in lung function, particularly when superimposed on a background of lung disease [4-6]. Farha et al. explored the increased estrogen hormonal effect on lung function as measured by pulmonary function testing (PFT) in healthy, never-smoking women and asthmatics [4; 7]. These studies reported a significant decrease in both populations in forced expiratory volume (FEV₁%), forced vital capacity (FVC), and diffusing capacity for carbon monoxide (DL_CO) from menses (week 1) to the early luteal phase (week 3). The trends were consistent among women with and without cyclic birth control. However, PFTs have well-known limitations in that they measure average obstruction over the whole lung and reveal only limited information about regional distribution or spatial heterogeneity of obstructive disease.

Computed tomographic (CT) imaging is a mainstay for clinical diagnostic evaluation of the chest. Spatial resolution achievable is within the sub-acinar anatomic scale, density resolution provides sensitivity down to 2-3 Hounsfield units and radiation exposures are on the order of one millisievert [8-10]. As we seek to utilize quantitative computed tomography (qCT) to carefully sub-phenotype individuals in various lung disease cohorts [11; 12], studies focus on various age ranges, smoking histories, and disease histories. Care is taken to coach individuals to the desired lung volumes (usually total lung capacity and residual volume). CT scanner protocols are now carefully defined [13], and image analysis methodologies continue to advance. However, pre-menopausal women are brought into the study at a random point within the month, despite cyclic menstrual cycle variation, potentially affecting the measurements used to characterize disease progression or response to intervention. If the characteristics of the lung are changing within the menstrual cycle, this will limit the sensitivity of the qCT metrics in detecting short term changes, weather these are changes due to interventions, assessment of short term disease progression rates, prediction of exacerbation risks, or disease susceptibility. It is important to understand the impact of menstrual cycle fluctuations on qCT derived lung metrics that can capture these early changes in the lung without subjective bias, and in a manner that incorporates spatial context. Quantifying menstrual cycle-associated changes in qCT measures is vital to ensure women are studied with the same precision as men. However, no studies have yet demonstrated that qCT cam measure menstrual-cycle associated changes in lung structure. In this study, we seek to determine the menstrual-cycle associated changes in qCT derived parenchymal and airway metrics in non-smoking women without lung disease. Further we seek to clarify if qCT metric variation over the menstrual cycle is mitigated with cyclic hormonal birth control.

Methods

Subjects

Non-smoking women (<10-year lifetime exposure) with no history of lung disease and with regular menstrual cycles were recruited for the study. Exclusion criteria included; subjects who were pregnant, breastfeeding, body mass index >30kg/cm², post-menopausal, diabetic, were taking long-term non-cyclic birth control (i.e. implant, patch, or injected), or had a hysterectomy. All study subjects gave written informed consent and data was collected with approval from the University of Iowa Institutional Review Board.

Data Collection

Women tracked their menstrual cycle via temperature and ovulation kit for 1-2 months, before completing two study visits coinciding with menses (M) and early luteal phase (ELP). At each visit subjects had; peripheral blood collection, PFTs, and ultra-low-dose qCT imaging. Peripheral blood analysis of hormone levels including estradiol, progesterone as well as global measures of angiogenesis via circulating CD34⁺CD133⁺ proangiogenic hematopoietic progenitors, historically referred to as endothelial progenitor cells [14-19]. A V6200 Body Plethysmograph (Sensor Medics) was used for the collection of PFTs including pre-bronchodilator spirometry and vital capacity measurements. Standardized PFT reporting guidelines from the American Thoracic Society were followed [20].

Ultra-low dose chest qCT imaging was performed using a SOMATOM Force CT Scanner (Siemens Healthcare) with; 120 kV, 36 ref mAs, CareDose on, pitch 1.2, 0.5 sec rotation time, 0.5mm slice thickness, Qr40 (admire 3, <5 sec scan. The CTDIvol for one scan at 36mAs was approximately 2.4mGy or 1.0mSv. qCT image data was collected at coached full inspiratory (total lung capacity, TLC) and coached full expiratory (residual volume, RV) lung volumes.

Airway Analysis

The VIDA∣Vision software program (VIDA Diagnostics) was used to determine the qCT quantitative airway measurements, focusing on the segmental bronchi from five selected airway pathways passing through RB1, RB4, RB10, LB1 and LB10 and extending two generations distally (sub-segmental and sub-sub-segmental bronchi). Measurements from each generation were averaged for each subject’s menses (M) and early luteal phase (ELP) scans. Using data from all subjects, the airway measurement outliers were identified as those ±3 standard deviations away from the mean for parametric data or three times the interquartile range (IQR) for nonparametric data, as previously described by Seo [21]. Outlier airway measurements and airway branches too short for the VIDA∣Vision program to get accurate quantitative measurements were excluded from the statistical calculations.

Parenchyma Analysis

QCT lung segmentation and histogram analyses were performed using Pulmonary Analysis Software Suite (PASS)[22]. For both inspiratory and expiratory lung volumes, measures from PASS include average pixel values within the lung region, dubbed mean lung density (MLD), total lung volume, and total tissue volume. Total lung volume is calculated by multiplying the number of voxels in the segmentation by the CT scan voxel volume. Tissue volume is defined as the total volume minus the air volume, with the air content of each voxel is calculated from its relative Hounsfield Unit (HU): (65-HU)/1065 and where 65HU represents pure tissue. Low attenuation areas (LAA) were defined as the percentage of lung volume less than −950 HU (LAA₉₅₀) in inspiratory CT, and less than −856HU (LAA₈₅₆) in expiratory CT. Also calculated were the ratio of expiratory/inspiratory for MLD (MLD E/I Ratio) and the value of the 15^th percentile of the attenuation curve (Perc15).

Statistics

Following an Anderson-Darling test for normality, the measurements at menses and early luteal phases were compared for the whole cohort using paired Student’s t-test (parametric) or Wilcoxon signed rank test (nonparametric). The measurement bias (early luteal phase – menses) between the cohort with birth control versus without was tested for normality and compared using unpaired Students t-test (parametric) or Mann Whitney test (nonparametric). An alpha value of 0.05 was used to establish significance on a statistical level. Statistical analyses were performed with GraphPad Prism (GraphPad Software, Inc). Pearson’s correlation coefficient was used to examine the relationship between the change in circulating CD34₊CD133₊ proangiogenic hematopoietic progenitors over the menstrual cycle and radiological measures from the lung.

Results

Thirty-one women, ranging from 21 to 51 years of age, were recruited for the study. Fifteen of the recruited women were not on any form of hormonal birth control (no birth control), 16 had a cyclic hormonal birth control with 0.02 to 0.035mg of estradiol. The women on cyclic hormonal birth control tended to have less change in estradiol, progesterone, and proangiogenic progenitor cells (CD34⁺ CD133⁺) over the menstrual cycle than did the cohort of women without cyclic birth control (Table 1).

Table 1.

Demographics

Parameter (±SD)	ALL	No Birth Control	Birth Control
	ALL	NBC	BC
N	31	15	16
Age (yr)	29.04 (8.66)	31.80 (9.37)	23.88 (3.68)
BMI (kg/cm2)	25.2 (3.5)	25.0 (3.8)	23.3 (3.1)
Race (White:Asian:Hispanic)	27:3:1	14:0:1	13:3:0
Estradiol (pg/mL): M	36.2 (25.8)	40.4 (25.3)	32.4 (26.5)
Estradiol (pg/mL): ELP	102.2 (138.2)	101.9 (62.9)	102.5 (183.0)
Progesterone (ng/mL): M	0.7 (1.2)	1.1 (1.6)	0.4 (0.2)
Progesterone (ng/mL): ELP	2.0 (3.2)	3.7 (4.1)	0.4 (0.2)
CD34+CD133+ (%): M	0.20 (0.10)	0.23 (0.13)	0.19 (0.04)
CD34+CD133+ (%): ELP	0.19 (0.11)	0.18 (0.11)	0.18 (0.13)

M - menses, ELP - early luteal phase

Airway Measurements

There were 22 measurements across all the airways investigated that were calculated to be outliers. Of those outlier measurements, 68% of them occurred during the early luteal phase, 14% during the menses time point, and 18% for the same measurement during both time points. These outlier values were excluded from the statistical evaluation as they were likely caused by local motion artifact (cardiogenic) or a change in orientation upon re-scan impacting partial volume effects.

In Table 2 we focus on qCT derived airway measures comparing all subjects at the segmental, subsegmental and sub-subsegmental level. At the segmental level, there was a significant difference in the inner area, wall thickness, minor outer diameter, major outer diameter, outer area, and inner perimeter outer perimeter, with all measurements being larger at menses. Bland Altman Plots in Supplemental Figure 1 provide visual representation of the consistent trend for larger airway measures at menses compared to luteal phase for the significant segmental airway measurements. At the subsegmental level, likely because of resolution limits, the only significant difference was reflected in the wall area fraction. For the measures without significant change, we report them in Table 2 because it is instructive to note the measurement repeatability.

Table 2.

qCT derived airway measurements for all subjects

Airway qCT	Feature	Mean (±SD)		p-value
Parameter (±SD)		Menses	Early Luteal Phase
Segmental	Minor Inner Diameter (mm)	3.87 (0.29)	3.83 (0.31)	0.12
	Major Inner Diameter (mm)	5.00 (0.40)	4.91 (0.36)	0.06
	Inner Area (mm2)	15.82 (2.37)	15.40 (2.17)	0.05
	Wall Thickness (mm)	1.33 (0.06)	1.32 (0.06)	0.01*
	Minor Outer Diameter (mm)	6.52 (0.36)	6.46 (0.39)	0.03*
	Major Outer Diameter (mm)	7.79 (0.44)	7.69 (0.41)	0.05*
	Outer Area (mm2)	40.61 (0.44)	39.75 (4.06)	0.03*
	Inner Perimeter (mm)	14.32 (4.36)	14.10 (0.99)	0.04*
	Outer Perimeter (mm)	22.79 (1.23)	22.52 (1.16)	0.02*
	Wall Area Fraction	0.62 (0.02)	0.62 (0.02)	0.18
Subsegmental (G1)	Minor Inner Diameter (mm)	2.75 (0.32)	2.76 (0.28)	0.63
	Major Inner Diameter (mm)	3.66 (0.40)	3.71 (0.36)	0.15
	Inner Area (mm2)	8.43 (1.80)	8.54 (1.62)	0.44
	Wall Thickness (mm)	1.15 (0.07)	1.15 (0.06)	0.33
	Minor Outer Diameter (mm)	5.04 (0.44)	5.03 (0.38)	0.91
	Major Outer Diameter (mm)	6.14 (0.54)	6.19 (0.48)	0.25
	Outer Area (mm2)	25.17 (4.11)	25.23 (3.60)	0.83
	Inner Perimeter (mm)	10.46 (1.08)	10.58 (0.98)	0.18
	Outer Perimeter (mm)	17.90 (1.46)	17.99 (1.30)	0.42
	Wall Area Fraction	0.68 (0.02)	0.67 (0.02)	0.04*
Sub-subsegmental (G2)	Minor Inner Diameter (mm)	2.19 (0.21)	2.23 (0.23)	0.21
	Major Inner Diameter (mm)	3.09 (0.00)	3.07 (0.26)	0.74
	Inner Area (mm2)	5.69 (1.00)	5.76 (1.07)	0.66
	Wall Thickness (mm)	1.02 (0.06)	1.02 (0.05)	0.83
	Minor Outer Diameter (mm)	4.19 (0.34)	4.24 (0.32)	0.37
	Major Outer Diameter (mm)	5.34 (0.35)	5.33 (0.37)	0.94
	Outer Area (mm2)	18.29 (2.56)	18.47 (2.63)	0.62
	Inner Perimeter (mm)	8.83 (0.68)	8.81 (0.74)	0.91
	Outer Perimeter (mm)	15.51 (1.00)	15.51 (1.08)	0.99
	Wall Area Fraction	0.70 (0.01)	0.70 (0.01)	0.70

^*Significance of difference (p) determined using a paired t-test for normally distributed data (Anderson-Darling test), and otherwise with the Wilcoxon Signed Rank test.

When separating subjects with birth control versus no birth control, the difference in segmental airway measurements across the menstrual cycle had an overall negative trend for both sub-cohorts and no statistically significant difference. This indicates a consistency in the impact of the menstrual cycle on the segmental airways regardless of hormonal birth control (Supplemental Table 1). For the subsegmental and sub-subsegmental airways the cohort with without hormonal birth control had a less consistent trend for larger measurements at menses compared to the cohort on birth control. At the sub-subsegmental level, the change in Minor Outer Diameter (average (SD) shift from luteal to menses in cohort with birth control: −0.06 (0.23) mm vs no birth control: 0.15 (0.27) mm) and Inner Perimeter (birth control: −0.22 (0.51) mm vs no birth control: 0.20 (0.61) mm) were statistically significantly different between the sub-cohorts.

The Pi10 measures is a mixed metric reflecting both airway wall thickness and luminal dimension changes. Across all subjects, the mean Pi10 for the whole lung and the major airway paths was stable with no significant change between values at menses and early luteal phase. There also were minimal differences found in the Pi10 for the whole lung and major airway paths between the sub-cohorts with and without birth control. Only one significant difference (p = 0.03) in the bias was found for the Pi10 calculated along the LB10 pathway between the birth control cohort (average (SD) change from luteal phase to menses of −0.02 (0.11)) and the cohort without birth control (change of 0.06 (0.08)).

Parenchymal Measures

Table 3 summarizes the changes in qCT histogram-based measures for all subjects comparing menses to the early luteal phase. Results demonstrate a statistically significant higher mean lung density (MLD) and mean lung density ratio (MLD E/I Ratio) at menses compared to the early luteal phase when assessed on expiration. Statistical significance for differences in tissue volume and LAA₈₅₆ were borderline (p = 0.06 and p = 0.07 respectively), suggesting a trend of higher tissue volume and increased air trapping during menses. At full inspiration there were no significant differences in the parenchymal measurements. Supplemental Figure 2 demonstrates high consistency between the qCT lung volumes at menses and early luteal phase for both expiratory and inspiratory lung volumes.

Table 3.

qCT derived histogram measures for all subjects

Parameter (±SD)	Menses	Early Luteal Phase	p-value
Inspiratory (TLC) qCT
Lung Volume (mL)	5214.4 (621.9)	5208.7 (678.2)	0.85
Tissue Volume (mL)	732 (95.6)	725 (92.4)	0.15
MLD (HU)	−850 (12.5)	−851 (13.9)	0.54
LAA950 (%)	2.9 (1.6)	3.1 (2.1)	0.65
Perc15 (HU)	−921.1 (9.9)	−921.8 (10.4)	0.56
Expiratory (RV) qCT
Lung Volume (mL)	1680.0 (313.0)	1690.3 (313.3)	0.37
Tissue Volume (mL)	694.8 (86.8)	686.4 (82.1)	0.06
MLD (HU)	−552.9 (51.5)	−560.8 (50.5)	0.03*
LAA856 (%)	0.2 (0.4)	0.4 (0.7)	0.04*
MLD E/I Ratio	0.65 (0.06)	0.66 (0.05)	0.05*

^*Significance of difference (p) determined using a paired t-test for normally distributed data (Anderson-Darling test), and otherwise with the Wilcoxon Signed Rank test.

TLC - total lung capacity, RV- residual volume, MLD - mean lung density, LAA - low attenuation area, Perc15 - 15^thpercentile of the attenuation curve, E/I - RV/TLC Ratio

The majority of quantitative parenchymal measurements for subjects without birth control (No Birth Control) have a larger degree of change over the menstrual cycle, compared to those subjects on cyclic hormonal birth control: As reflected in the larger average magnitude of change (ELP-M) in Table 4. Expiratory tissue volume change was statistically different between the birth control cohort (change from luteal phase to menses of 1.08mL) and the cohort without birth control (change of −15.86mL). In the sub-cohort without birth control, the change (ELP-M) in circulating CD34⁺CD133⁺ proangiogenic hematopoietic progenitors showed a moderate inverse correlation (r = −0.6) with the change in expiratory tissue volume, but no correlation (r = 0.2) in the sub-cohort with birth control (Supplemental Figure 3).

Table 4.

Change in qCT derived histogram measures (ELP-M) between cohorts with and without birth control

Parameter (±SD)	Birth Control	No Birth Control	p-value
Inspiratory (TLC) qCT
Lung Volume (mL)	−42.59 (139.20)	43.30 (142.27)	0.09
Tissue Volume (mL)	−8.86 (24.07)	−2.63 (24.23)	0.95
MLD (HU)	−0.15 (7.14)	−1.61 (6.16)	0.56
LAA950 (%)	0.27 (1.41)	0.11 (0.93)	0.95
Perc15 (HU)	−0.16 (6.19)	−1.07 (4.82)	0.66
Expiratory (RV) qCT
Lung Volume (mL)	14.67 (44.35)	−0.65 (57.96)	0.37
Tissue Volume (mL)	1.08 (12.35)	−15.86 (22.95)	0.02*
MLD (HU)	−2.92 (12.5)	−10.10 (18.6)	0.25
LAA856(%)	0.08 (0.35)	0.15 (0.34)	0.04*
MLD E/I Ratio	0.003 (0.01)	0.01 (0.03)	0.22

^*Significance of difference (p) determined using an unpaired t-test for normally distributed data (Anderson-Darling test), and otherwise with the Mann Whitney test

TLC - total lung capacity, RV- residual volume, MLD - mean lung density, LAA - low attenuation area, Perc15 – 15^thpercentile of the attenuation curve, E/I - RV/TLC Ratio

Pulmonary function testing

Although we did not detect statistically significant differences in PFTs in this population (Table 5) over the menstrual cycle, a trend for increased diffusing capacity for carbon monoxide (DL_CO) and the DL_CO divided by the alveolar volume (DL_CO/VA) at menses compared to the early luteal phase was found in accordance with prior published findings [4].

Table 5.

PFT derived measures for all subjects

Parameter (±SD)	Menses	Early Luteal Phase	p-value
Pulmonary Function Test (PFT)
TLC (%)	96.29 (11.87)	96.27 (10.71)	0.88
RV (%)	105.50 (40.91)	102.80 (34.00)	0.92
FEV1 (%)	95.13 (13.07)	96.06 (11.22)	0.16
FVC (%)	100.37 (11.38)	100.80 (11.44)	0.52
FEV1/FVC	95.33 (9.75)	95.55 (8.00)	0.68
DLCO (%)	92.97 (13.00)	90.27 (17.32)	0.19
DLCO/VA (%)	99.67 (12.83)	96.12 (16.97)	0.08

^*Significance of difference (p) determined using a paired t-test for normally distributed data (Anderson-Darling test), and otherwise with the Wilcoxon Signed Rank test.

FEV₁ - forced expiratory volume in 1 sec, FVC - forced vital capacity, DL_CO - diffusion capacity for carbon monoxide, VA - alveolar volume

The sub-cohorts demonstrated a very similar magnitude of change in PFT parameters (Table 6), including DL_CO and DL_CO/VA, reflecting higher values at menses as seen for all subjects in Table 5. A significant difference between the groups was reflected in the alternate direction of change in total lung capacity (TLC), with the birth control group having a slight increase in TLC at menses while the no birth control group had a minor decrease in TLC at menses.

Table 6:

Change in PFT derived measures (ELP-M) between cohorts with and without birth control

Parameter (±SD)	Birth Control	No Birth Control	p-value
Pulmonary Function Test (PFT)
TLC (%)	−2.12 (5.71)	2.86 (6.87)	0.04*
RV (%)	−9.35 (38.61)	7.29 (26.77)	0.37
FEV1 (%)	0.93 (3.43)	0.71 (6.21)	0.78
FVC (%)	0.42 (3.34)	0.36 (3.84)	0.93
FEV1/FVC	−0.24 (4.52)	0.79 (7.17)	0.51
DLCO (%)	−2.62 (11.23)	−2.21 (9.62)	0.85
DLCO/VA (%)	−3.38 (10.60)	−3.00 (9.98)	0.99

^*Significance of difference (p) determined using an unpaired t-test for normally distributed data (Anderson-Darling test), and otherwise with the Mann Whitney test

M - menses, ELP - early luteal phase, FEV₁ - forced expiratory volume in 1 sec, FVC - forced vital capacity, DL_CO - diffusion capacity for carbon monoxide, VA - alveolar volume

Discussion

The results from this study indicate that there are significant changes in qCT metrics over the menstrual cycle that may affect how disease is measured and longitudinally followed in females. Results also indicate that the patterns of lung structure change across the menstrual cycle may be impacted by cyclic hormonal birth control. Quantitative measurements of lung structure derived from qCT are being increasingly utilized to understand disease etiology, track changes over time, and objectively evaluate intervention response. This study highlights the impact the female menstrual cycle has on these quantitative, CT derived lung measurements. It is instructive to observe that, while others have observed menstrual cycle related differences in PFTs [4; 7], in this small study cohort qCT was more sensitive than PFTs in detecting menstrual cycle changes.

Objective, quantitative image analysis revealed increased segmental airway dimensions including diameter, perimeter, and area as well as airway wall thickness occurs at menses compared to the early luteal phase in this population of healthy women (Table 2). The change in segmental airway measurement over the menstrual cycle was not statistically different between the sub-cohort taking hormonal birth control compared to those without hormonal birth control. These results indicate further study is warranted to determine the impact of the menstrual cycle in asthmatic women. Some asthmatic women experience worsening asthma symptoms during menses: perimenstrual asthma (PMA) [23-25]. More recent studies have investigated whether taking hormonal birth control could help prevent PMA with varying results. In a study of women with mild to moderate asthma, not taking any form of hormonal birth control, significantly improved symptoms were found from estradiol administration during menses [26]. Similarly, Nwaru et al. found that hormonal birth control may decrease asthma exacerbations [27]. However, other studies have found the opposite to be true, especially for obese women [27-29].

Prior work by Kauczor et al. reported CT scans performed during full expiration are more informative of pulmonary function compared to inspiratory scans and that a decrease in expiratory MLD indicates an obstructive impairment [30]. In our study, parenchymal analysis of menses versus early luteal phase timepoints did find a significant change in expiratory MLD of about 10 HU. As the use of qCT to evaluate disease progression in lungs grows, it is important to understand how meaningful a change in lung density of 10 HU is for measurements. A research study by Mitsunobo et al. found a ~10 HU significant increase in weighted MLD measures from treating asthma exacerbations with systemic glucocorticoid therapy [31]. Studies using qCT to determine changes in lungs in smokers also found <10 HU to be significant between groups of past, current, and never smokers [32; 33]. As studies involving asthmatic subjects (and those investigating the early impact of smoking cessation or vaping) have a high potential of including pre-menopausal female subjects, care should be taken in acquiring longitudinal qCT data at the same point in the female menstrual cycle. In our study, the decrease in expiratory Tissue Volume at luteal phase compared to menses, approched significance (p = 0.06). Both the MLD and Tissue Volume change likely represent some combination of change in pulmonary capillary blood volume, permeability and extravascular fluid. While the trends in parenchymal measurements were similar at inspiratory volumes, the changes were not significant. The MLD and Tissue Volume change over the mensural cycle may be more evident at full expiratory (RV) breath hold, versus full inspiratory (TLC) breath hold, because of reduced tension on the vascular bed and extravascular compartments. These results warrant further study in a larger subject cohort.

The MLD and Tissue Volume decrease detected in this study from menses to luteal phase also supports the decrease in lung function for women during the luteal phase based on DL_CO found by previous research and attributed to changes in the pulmonary capillary blood volume measures [4; 7]. While the change in DL_CO was not statistically significant in our cohort of healthy women, the same trend of a decrease in function from a peak at menses was observed (Table 5). While the DL_CO measures were stable across the birth control cohorts (Table 6), a comparison of the qCT derived measures between these cohorts illustrated that subjects without birth control had greater variability over the menstrual cycle.

We hypothesized that the sub-cohort with cyclic hormonal birth control would have less evidence of measurement difference over the menstrual cycle compared to the sub-cohort without hormonal birth control. The qCT derived airway measurements indicated a consistent pattern of change, with larger airway measures at menses, regardless of birth control. The parenchymal analysis indicated that the sub-cohort without birth control did have a larger degree of change in expiratory Tissue Volume and LAA₈₅₆ compared to those with birth control (Table 4).

This study includes some limitations. To limit the radiation exposure associated with the study, data collection targeted two key time points in the menstrual cycle corresponding to menses and the early luteal phase. These times were selected as they correspond to the maximum difference in percentage predicted forced expiratory volume (%FEV1) and forced vital capacity (%FVC) in healthy and asthmatic women previously reported by Farha et al [4]. In addition, the early luteal phase can be reliably detected following confirmation of ovulation via ovulation test strips, aiding in standardized timing of data acquisition in females not on cyclic hormonal birth control. This study included only healthy women without lung disease to isolate the impact of the menstrual cycle on lung measurements without confounding effects of disease variability. Our findings confirm detectable differences in qCT metrics associated with the menstrual cycle and warrant further investigation in asthmatic, smoking, and vaping populations. The body mass index (BMI) of subjects in this study were limited to minimize variability in resultant qCT data resolution and noise. However, obesity is known to impact lung mechanics and is a disease risk factor and modifier for asthma, COPD, pulmonary hypertension, sleep apnea, and other cardiopulmonary conditions [34]. Hence, future work will explore if the menstrual cycle-associated changes in qCT derived metrics in the lung of obese women are consistent or different from those found in this study.

This study provides evidence that quantitative measurements of lung structure derived from qCT imaging are impacted by the female menstrual cycle. Care should be taken in studies exploring longitudinal change in qCT derived parenchymal and segmental airway measurements, to acquire data at a consistent time in the menstrual cycle for pre-menopausal women. Also, this study highlights that further exploration of the impact of cyclic hormonal birth control is warranted to better understand the impact on qCT measurements in healthy and asthmatic women.

Key points.

• Lung measurements from chest computed tomography are used in multicenter studies exploring lung disease progression and treatment response

• The menstrual cycle impacts lung structure measurements.

• Cyclic variability should be considered when evaluating longitudinal change with CT in menstruating women.

Abbreviations

COPD

chronic obstructive pulmonary disease

DL_CO

diffusing capacity for carbon monoxide

E/I Ratio

expiratory/inspiratory ratio

ELP

early luteal phase

FEV₁%

forced expiratory volume

FVC

forced vital capacity

LAA

low attenuation area

M

menses

MLD

mean lung density

PFT

pulmonary function testing

PMA

perimenstrual asthma

qCT

quantitative computed tomography

RV

residual volume

TLC

total lung capacity

VA

alveolar volume