Sex-related differences and hormonal effects in the Dry Eye Assessment and Management (DREAM) study

Abstract

Background/aims

To compare dry eye disease (DED) signs and symptoms between men and women, as well as between premenopausal and postmenopausal women, in the Dry Eye Assessment and Management (DREAM) study.

Methods

434 women and 101 men self-reported prior medical history and underwent a standardised DED assessment using the Ocular Surface Disease Index, Brief Pain Inventory, Tear Break-Up Time (TBUT) (s), Schirmer’s test 2 (mm/5 min), National Eye Institute-graded lissamine conjunctival staining, corneal staining, meibomian gland dysfunction evaluation and tear osmolarity (mOsms/L) at baseline, 6 months and 12 months. Multivariable linear regression models were used to compare these scores.

Results

Women experienced significantly worse DED signs than men with lower Schirmer’s test scores (9.27 vs 12.16; p<0.001), higher corneal staining scores (3.59 vs 2.70; p=0.006) and worse composite DED sign scores (0.52 vs 0.40; p<0.001). Postmenopausal women experienced significantly worse DED signs than premenopausal women with higher corneal staining scores (3.74 vs 2.58, p<0.001), higher conjunctival staining scores (2.80 vs 2.22, p<0.001), higher tear osmolarity (304 vs 299, p=0.004), lower TBUT (3.37 vs 3.93, p=0.047), worse meibomian gland dysfunction (3.05 vs 2.62, p=0.04) and worse composite DED sign scores (0.54 vs 0.42, p<0.001). There were no significant differences in DED symptoms between sex and between premenopausal and postmenopausal women (all p≥0.08).

Conclusion

In the DREAM study, women experienced more severe DED signs than men. Further, postmenopausal women presented with more severe DED signs than premenopausal women. Elucidating these differences may improve DED diagnosis and provide future direction in understanding sex-related differences in DED.

Trial registration number

NCT02128763.

INTRODUCTION

Dry eye disease (DED) is a multifaceted disorder that impacts the ocular surface and tear homoeostasis.^{1 2} With associated symptoms including eye dryness, discomfort, visual disturbance and sensitivity to light, DED imposes an increasing burden on both individuals and society,³ with a global prevalence ranging from 5% to 50%.⁴ Demographically, DED impacts more women and older individuals,⁴ with large-scale studies estimating that 3.25 million US women over the age of 50 are affected by DED.^{5 6} In addition, there is evidence that women experience more severe DED symptoms and have worse signs than men.^{5 7}

The aetiology behind sex-related differences in DED is still unclear. Several small-scale randomised controlled trials have been conducted to investigate the potential effects of sex hormones on DED and yielded varying results.^8-13 For example, Golebiowski et al found that oestrogen supplementation in postmenopausal women worsens DED symptoms but testosterone supplementation had no impact.¹³ In addition, the use of aromatase inhibitors and hormonal contraceptives has been shown to affect DED signs and symptoms in women.^{8 14 15} For example, He et al found that hormonal contraceptive users experienced a greater risk of developing DED.¹⁶ In addition to biological differences, iatrogenic factors such as increased contact lens usage, serotonin-norepinephrine reuptake inhibitors and selective serotonin reuptake inhibitors in women may also be related to increased prevalence of DED.^17-19

Currently, there is a lack of broad and well-defined characterisations of the differences in DED symptoms and signs between men and women, as well as between premenopausal and postmenopausal women. This study seeks to provide insight into these clinical findings through the analysis of rich data from a large, well-characterised cohort of dry eye participants in the Dry Eye Assessment and Management (DREAM) clinical trial. The DREAM study was a large, multicentre, prospective, double-masked, randomised clinical trial designed to assess the efficacy and safety of oral omega-3 supplementation for the treatment of DED.²⁰ As part of the study, participants underwent a comprehensive assessment of symptoms and signs following standardised protocols. The present report is an in-depth analysis of sex-related differences and the impact of menopause on DED symptoms and signs in the DREAM study cohort.

METHODS

Full details on the inclusion and exclusion criteria as well as outcome measures can be found in the previously published main results paper of the DREAM study.^{20 21} The study was approved by the institutional review board/ethics committee at each centre (centres listed in Credit Roster for the DREAM study, available at www.aaojournal.org), followed the tenets of the Declaration of Helsinki and written informed consent was obtained from all patients.

This is a secondary analysis of the data from the DREAM study. The DREAM study (ClinicalTrials.gov) enrolled 535 patients ≥18 years of age from 27 clinical centres in the USA. With the aim to include a wide spectrum of patients with symptomatic moderate-to-severe DED, the study enrolled patients who met the eligibility criteria for reported dry eye-related ocular symptoms for at least 6 months prior to the screening visit, use or desired use of artificial tears two times per day in the 2 weeks prior to the screening visit and an Ocular Surface Disease Index (OSDI) score of 25–80 at the screening visit and 21–80 at the baseline visit. Patients were randomised to groups in a 2:1 fashion to receive either a daily oral dose of 3000 mg omega-3 fatty acid supplement (n-3 eicosapentaenoic and docosahexaenoic acids) or a placebo refined olive oil pill.

Dry eye outcome measures taken at baseline, 6 months and 12 months follow-up included the OSDI score (scale of 0–100), with higher scores indicating more severe symptoms as well as the Brief Pain Inventory (BODI) and BODI Pain Score, with higher scores indicative of more severe symptoms. Conjunctival staining scores (ranging from 0 to 6) and corneal staining scores (0 to 15) per eye were measured, with higher scores indicating more severe damage to the ocular surface. The lissamine green conjunctival staining used a modified version of the National Eye Institute (NEI)/industry-recommended guidelines. Specifically, the entire temporal and the entire nasal section of each eye are graded on a scale of 0 to 3 (0: no staining and 3: severe staining) for a total possible score of 6 in each eye. Corneal staining scores were according to fluorescein corneal staining using the NEI/industry-recommended guidelines. Further, Tear Break-Up Time (TBUT) (seconds between a blink and tear film gaps) and Schirmer’s test 2 with anaesthesia after a drop of anaesthetic (length of wetting of paper strips placed on the lower eyelid) were measured, with lower values indicative of more severe DED. Osmolarity was measured from both eyes at baseline, 6 months and 12 months at the 19 of the 27 clinical centres that had the device TearLab Osmolarity System (OcuSense, San Diego, California, USA).

Each patient was administered the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36) V.2.0 at baseline, 6 months and 12 months follow-up. Two summary scores were generated from SF-36: the physical component summary and mental component summary (MCS). The MCS is scored between 0 and 100, with higher scores indicating greater psychological well-being. The recommended cut-off of 42 in MCS score has sensitivity of 73.7% and specificity of 80.6% for identifying clinical depression.²²

Statistical analysis

We compared the demographics and comorbidities between men and women, as well as between premenopausal and postmenopausal women using the two-sample t-test for continuous variables and the χ² or Fisher’s exact test for categorical variables. For the comparison of dry eye symptoms and signs between men and women, we used generalised linear models with and without an adjustment for age, race, smoking status and several comorbidities that were previously found to be associated with the severity of dry eye symptoms and signs in the DREAM study including facial rosacea, rheumatoid arthritis, peripheral artery disease and Sjögren’s syndrome²³ and depression defined as mental component summary score ≤42.²⁴ We performed these analyses using the combined data of baseline, 6 months and 12 months to improve statistical power. The correlation of repeated measures between visits and between eyes within the same participant was accounted for by using generalised estimating equations. Similar analyses were performed for each time point separately to check the consistency of results over time. Because omega-3 supplementation did not show a significant effect on DED signs and symptoms compared with placebo,²⁰ all these evaluations were based on the data from the two treatment groups combined.

For the comparison of each DED sign between men and women, or between premenopausal and postmenopausal women, signs from both eyes at the same time point were used. In addition, we adapted a method from previous studies^{20 23} to calculate a composite dry eye severity score of signs by transforming the individual score of each of the five dry eye signs (TBUT, Schirmer testing, corneal staining, conjunctival staining and meibomian gland dysfunction) to a common unit severity score between 0 (no DED signs) and 1 (most severe signs). A composite signs severity score was then calculated by averaging severity scores of the six individual DED signs. The composite sign severity score ranges from 0 to 1, with 1 indicating the most severe dry eye signs. This composite score allows for one continuous metric of assessing objective DED sign severity based on five separate dry eye sign measures.^{25 26}

Statistical analyses were performed using SAS V.9.4 (SAS Institute, Cary, North Carolina, USA), and two-sided p<0.05 was considered statistically significant.

RESULTS

Men vs women

Table 1 shows the comparison of baseline characteristics between men and women in the DREAM study. Compared with men, more women were white (76% vs 67%, p = 0.049), non-Hispanic or Latino (87% vs 83%, p = 0.03) and had Sjogren’s syndrome (11.1% vs 4.0%, p = 0.001). More women were taking antidepressants (24% vs 12%, p = 0.01), and were treated for dry eye with artificial tears (81% vs 71%, p = 0.03), cyclosporine drops (22% vs 9%, p=0.003) and other dry eye treatments beyond artificial tears or gel, cyclosporine drops, warm lid soaks, lid scrubs or baby shampoo (36% vs 18%, p<0.001). Fewer women were current smokers (4% vs 10%, p = 0.02) or self-reported diabetics (9% vs 18%, p = 0.03). Men and women were similar in the mental health component score of SF-36 (p = 0.31) and depression defined as mental health score 42 or less (15.7% and 15.8%, respectively, p=0.97).

Table 1.

Comparison of baseline characteristics between women and men

Baseline characteristics	Women (N=434)	Men (N=101)	P value*
Age (years)			0.24
Mean (SD)	58.3 (12.5)	56.6 (15.8)
Range	(20.0–85.0)	(18.0–87.0)
Ethnicity			0.03
Hispanic or Latino	54 (12.4%)	14 (13.9%)
Not Hispanic or Latino	379 (87.3%)	84 (83.2%)
Unable to answer	1 (0.2%)	3 (3.0%)
Race			0.049
White	330 (76.0%)	68 (67.3%)
Black or African American	50 (11.5%)	14 (13.9%)
Asian	13 (3.0%)	6 (5.9%)
American Indian or Alaskan Native	1 (0.2%)	2 (2.0%)
More than one race	9 (2.1%)	0 (0.0%)
Unable to answer	31 (7.1%)	11 (10.9%)
Cigarette smoking			0.02
Never	305 (70.3%)	62 (61.4%)
Former	113 (26.0%)	29 (28.7%)
Current	16 (3.7%)	10 (9.9%)
Sjögren’s syndrome: yes (%)†	48 (11.1)	4 (4.0)	0.001
Self-reported medical history
Rosacea (facial): yes (%)	91 (21.0)	18 (17.8)	0.48
Peripheral artery disease: ongoing (%)	39 (9.0)	8 (7.9)	0.81
Thyroid dysfunction: ongoing (%)	83 (19.1)	11 (10.9)	0.07
Hypertension: ongoing (%)	114 (26.3)	35 (34.7)	0.21
Diabetes: ongoing (%)	39 (9.0)	18 (17.8)	0.03
Rheumatoid arthritis: ongoing (%)	40 (9.2)	7 (6.9)	0.71
Irritable bowel: ongoing (%)	38 (8.8)	6 (5.9)	0.10
Osteoarthritis: ongoing (%)	117 (27.0)	17 (16.8)	0.07
Hypercholesterolaemia: ongoing (%)	136 (31.3)	35 (34.7)	0.50
Depression: ongoing (%)	74 (17.1)	13 (12.9)	0.45
Taking antidepressants: yes (%)	105 (24.2)	13 (12.9)	0.01
Score for measures of mental health
Mean (SD)	52.5 (9.4)	51.5 (9.5)	0.31
≤42, n (%)	68 (15.7)	16 (15.8)	0.97
Treatment for dry eye disease, yes (%)‡
Artificial tears or gel	352 (81.1)	72 (71.3)	0.03
Cyclosporine drops	96 (22.1)	9 (8.9)	0.003
Warm lid soaks	97 (22.4)	17 (16.8)	0.22
Lid scrubs or baby shampoo	71 (16.4)	12 (11.9)	0.26
Other	158 (36.4)	18 (17.8)	<0.001

^*Participants can use more than one treatment for dry eye disease at baseline.

^†Meeting 2012 American College of Rheumatology criteria for Sjögren syndrome.

^‡From two-sample t-test for comparison of means and χ² test or Fisher’s exact test for comparison of proportions.

Table 2 compares men and women’s dry eye symptoms and signs using combined data from baseline, 6 months and 12 months. In the multivariable analysis adjusted by age, race, smoking status, Sjogren’s syndrome, facial rosacea, rheumatoid arthritis and peripheral artery disease, women demonstrated significantly more severe dry eye signs in terms of lower Schirmer’s test scores (9.3 vs 12.2, p<0.001), higher corneal staining scores (3.6 vs 2.7, p = 0.006) and higher composite dry eye severity scores based on signs (0.52 vs 0.40, p<0.001). Women also had a shorter average TBUT that was statistically significant in the univariable analysis (3.45 vs 3.93 secs, p = 0.03) but only borderline significant (p=0.07) in the multivariable analysis. Women tended to have more meibomian gland abnormalities than men in both the univariable analysis (3.0 vs 2.7, p = 0.08) and multivariable analysis (p = 0.08). There was no statistically significant difference between men and women in conjunctival staining score (multivariable analysis p = 0.21). Women had more dry eye symptoms as measured by the OSDI score that was borderline significant in both univariable analysis (35.7 vs 32.7, p = 0.054) and multivariable analysis (p = 0.08). Results were consistent when data were analysed for baseline, 6 months and 12 months separately (online supplemental table S1). There were no significant differences between men and women in changes of symptoms and signs from baseline at 6 months and 12 months combined (online supplemental table S2) or when analysed separately (online supplemental table S3).

Table 2.

Comparison of dry eye symptoms and signs between female and male participants by combining baseline, 6 months and 12 months

Dry eye symptoms and signs	Female			Male			P value	Adjusted p value*
	N	Mean (SD)	Min, max	N	Mean (SD)	Min, max
OSDI total score	1225	35.69 (18.48)	0.00, 100.00	279	32.69 (18.24)	0.00, 87.50	0.054	0.08
BODI score	1225	25.87 (18.19)	0.00, 98.18	279	25.84 (17.21)	0.00, 72.73	0.94	0.82
BODI pain score	1224	37.53 (21.26)	0.00, 100.00	279	35.56 (21.67)	0.00, 90.00	0.29	0.25
Tear Break-up Time (s)	2450	3.45 (2.30)	0.00, 32.67	558	3.93 (2.81)	0.13, 30.00	0.03	0.07
Schirmer′s test (mm in 5 min)	2434	9.27 (6.58)	0.00, 40.00	558	12.16 (7.96)	0.00, 40.00	<0.001	<0.001
Corneal staining score	2450	3.59 (2.99)	0.00, 15.00	558	2.70 (2.43)	0.00, 15.00	<0.001	0.006
Conjunctival staining score	2450	2.72 (1.62)	0.00, 6.00	558	2.48 (1.50)	0.00, 6.00	0.08	0.21
Meibomian gland abnormality	2448	2.99 (1.90)	0.00, 6.00	558	2.69 (1.91)	0.00, 6.00	0.07	0.08
Composite dry eye severity score based on Signs	2450	0.52 (0.29)	0.002, 1.00	558	0.40 (0.27)	0.002, 0.996	<0.001	<0.001
Tear osmolarity (mOsms/L)	1763	303.4 (16.96)	200.00, 397.00	402	301.4 (18.15)	223.00, 399.00	0.22	0.50

^*Adjusted by visit, adjust by age, race, smoking status, Sjögren syndrome, facial rosacea, rheumatoid arthritis, peripheral artery disease and depression defined by measure of mental health score ≤42.

BODI, Brief Pain Inventory; OSDI, Ocular Surface Disease Index.

Premenopausal versus postmenopausal women

We grouped women into premenopause and postmenopause. The premenopause group (n=59) included women who had not undergone menopause, tubal ligation or a hysterectomy procedure; likewise, the postmenopause group (n=375) included women who have undergone menopause, tubal ligation or a hysterectomy procedure. Table 3 shows comparisons of baseline characteristics in premenopause versus postmenopausal women. Beyond significant differences in age (38 vs 62 years, p<0.0001), significantly more premenopausal women were Hispanic or Latino (19% vs 12%, p = 0.03) and current smokers (6.8% vs 3.2%, p = 0.02). More postmenopausal women self-reported ongoing hypertension (30% vs 3%, p<0.001), ongoing osteoarthritis (30% vs 9%, p = 0.001), ongoing hypercholesterolaemia (36% vs 3%, p<0.001), higher scores of mental health (53.2 vs 48.3, p<0.001) and concurrent use of cyclosporine drops for DED treatment (p = 0.02), but there were no other significant differences in the use of artificial tears or gel, warm lid soaks, lid scrubs, baby shampoo or other treatments for DED.

Table 3.

Comparison of baseline characteristics between premenopause and postmenopause women

Baseline characteristics	Premenopause (N=59)	Postmenopause (N=375)	P value*
Age (years)			<0.001
Mean (SD)	37.9 (9.0)	61.5 (9.6)
Range	(20.0–54.0)	(30.0–85.0)
Ethnicity			0.03
Hispanic or Latino	11 (18.6%)	43 (11.5%)
Not Hispanic or Latino	47 (79.7%)	332 (88.5%)
Unable to answer	1 (1.7%)	0 (0.0%)
Race			0.07
White	43 (72.9%)	287 (76.5%)
Black or African American	3 (5.1%)	47 (12.5%)
Asian	3 (5.1%)	10 (2.7%)
American Indian or Alaskan Native	0 (0.0%)	1 (0.3%)
More than one race	3 (5.1%)	6 (1.6%)
Unable to answer	7 (11.9%)	24 (6.4%)
Cigarette smoking			0.02
Never	47 (79.7%)	258 (68.8%)
Former	8 (13.6%)	105 (28.0%)
Current	4 (6.8%)	12 (3.2%)
Sjögren’s syndrome: yes (%)†	8 (13.6)	40 (10.7)	0.78
Self-reported medical history
Rosacea (facial): yes (%)	13 (22.0)	78 (20.8)	0.86
Peripheral artery disease: ongoing (%)	5 (8.5)	34 (9.1)	0.50
Thyroid dysfunction: ongoing (%)	6 (10.2)	77 (20.5)	0.06
Hypertension: ongoing (%)	2 (3.4)	112 (29.9)	<0.001
Diabetes: ongoing (%)	3 (5.1)	36 (9.6)	0.29
Rheumatoid arthritis: ongoing (%)	6 (10.2)	34 (9.1)	0.31
Irritable bowel: ongoing (%)	5 (8.5)	33 (8.8)	0.32
Osteoarthritis: ongoing (%)	5 (8.5)	112 (29.9)	0.001
Hypercholesterolaemia: ongoing (%)	2 (3.4)	134 (35.7)	<0.001
Depression: ongoing (%)	13 (22.0)	61 (16.3)	0.38
Taking antidepressants: yes (%)	16 (27.1)	89 (23.7)	0.62
Score for measures of mental health
Mean (SD)	48.3 (9.8)	53.2 (9.1)	<0.001
≤42, n (%)	10 (16.9)	58 (15.5)	0.85
Treatments used for dry eye disease* ‡
Artificial tears or gel	44 (74.6%)	308 (82.1%)	0.21
Cyclosporine drops	6 (10.2%)	90 (24.0%)	0.02
Warm lid soaks	12 (20.3%)	85 (22.7%)	0.87
Lid scrubs or baby shampoo	8 (13.6%)	63 (16.8%)	0.70
Other	16 (27.1%)	142 (37.9%)	0.15

^*Participants can use more than one treatment for dry eye disease at baseline.

^†Meeting 2012 American College of Rheumatology criteria for Sjögren syndrome.

^‡From two-sample t-test for comparison of means and χ² test or Fisher exact test for comparison of proportions.

Table 4 presents results from analysis of combined data from baseline, 6 months and 12 months for comparison between premenopausal and postmenopausal women. In the multivariable analysis, postmenopausal women demonstrated significantly more severe dry eye signs in terms higher corneal staining scores (3.7 vs 2.6, p<0.001), higher conjunctival staining scores (2.8 vs 2.2, p<0.001), higher meibomian gland abnormality scores (3.1 vs 2.6, p = 0.04), higher tear osmolarity (304 vs 299, p = 0.004), shorter TBUT (3.4 vs 3.9, p = 0.047) and higher composite severity scores based on dry eye signs (0.54 vs 0.42, p<0.001). These significant differences in signs were consistent when data was analysed separately for baseline, 6 months and 12 months (online supplemental table S4).

Table 4.

Comparison of dry eye symptoms and signs between premenopause and postmenopause females by combining baseline, 6 months and 12 months

Dry eye symptoms and signs	Premenopause			Postmenopause			P value	Adjusted p value*
	N	Mean (SD)	Min, max	N	Mean (SD)	Min, max
OSDI total score	162	37.93 (18.67)	0.00, 77.27	1063	35.35 (18.44)	0.00, 100.00	0.23	0.41
BODI score	162	29.40 (16.47)	0.00, 71.82	1063	25.33 (18.38)	0.00, 98.18	0.03	0.13
BODI #3 pain score	162	40.37 (20.00)	0.00, 80.00	1062	37.10 (21.42)	0.00, 100.00	0.15	0.29
Tear Break-up Time (s)	324	3.93 (3.21)	0.27, 32.67	2126	3.37 (2.12)	0.00, 19.13	0.06	0.047
Schirmer test (mm in 5 min)	318	10.35 (7.54)	0.00, 35.00	2116	9.11 (6.41)	0.00, 40.00	0.15	0.16
Corneal staining score	324	2.58 (2.24)	0.00, 10.00	2126	3.74 (3.06)	0.00, 15.00	<0.001	<0.001
Conjunctival staining score	324	2.22 (1.43)	0.00, 6.00	2126	2.80 (1.64)	0.00, 6.00	<0.001	<0.001
Meibomian gland abnormality	324	2.62 (1.89)	0.00, 6.00	2124	3.05 (1.89)	0.00, 6.00	0.052	0.04
Composite dry eye severity score based on Signs	324	0.42 (0.28)	0.002, 0.994	2126	0.54 (0.29)	0.002, 1.00	<0.001	<0.001
Tear osmolarity (mOsms/L)	232	299.2 (14.70)	268.00, 381.00	1531	304.0 (17.19)	200.00, 397.00	0.006	0.004

^*Adjusted by visit, race, smoking status, Sjögren syndrome, facial rosacea, rheumatoid arthritis, peripheral artery disease and depression defined by measure of mental health score ≤42.

BODI, Brief Pain Inventory; OSDI, Ocular Surface Disease Index.

In the multivariable analysis of combined data from baseline, 6 months and 12 months, there was no significant difference between premenopausal and postmenopausal women in dry eye symptoms including OSDI score (p = 0.41) or BODI score (p = 0.13). When data from baseline, 6 months and 12 months were analysed separately, premenopausal women had worse symptoms at baseline with higher OSDI scores (46.9 vs 41.9, p = 0.04) and higher BODI scores (37.9 vs 29.8, p = 0.002). However, these significant differences were not maintained at 6 months and 12 months (all p>0.50, online supplemental table S4). When compared with baseline, premenopausal women experienced a higher reduction in BODI score (−12.6 vs −6.0, p = 0.009, online supplemental table S5) at 6 months than postmenopausal women, and showed a borderline reduction, though not statistically significant (−12.7 vs 7.6, p = 0.06) at 12 months.

DISCUSSION

Men vs women

In the present study, we found that women in the DREAM study cohort experienced more severe DED signs compared with men, independent of their age, race, smoking status and comorbidities. Specifically, compared with men, women had significantly lower Schirmer’s test scores, worse corneal staining and worse composite dry eye severity scores based on signs. However, we did not find any significant difference in dry eye symptoms between men and women as measured by the OSDI and BODI.

Broadly, our findings are consistent with a study by Vehof et al which found that women overall experienced more severe DED signs than men.⁷ Similar to our study, Vehof et al found that women had significantly worse corneal staining, conjunctival staining, TBUT and tear osmolarity. However, in contrast to our study, they did not find that Schirmer scores and meibomian gland dysfunction significantly differed between men and women.

There are several possible reasons why we found significant differences in DED signs between men and women. One possible reason may be attributed to hormonal differences. Various studies have reported results quantifying blood serum levels or the local presence of sex hormones in ocular tissues or tears. Most consistently, associations of dry eye with high oestrogen levels,¹¹ both in the context of sex differences and specific hormonal changes in women (menstrual cycle, pregnancy, menopause), and low androgen levels have been reported.^{27 28} There have been several studies^{27 28} reporting that androgen deficiency promotes meibomian gland dysfunction and is associated with DED. While not reaching a level of significance in our study, women showed more meibomian gland abnormalities than men.

Interestingly, hormones also appear to play a role in other inflammatory systemic diseases presenting with significant sex differences, such as systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA).²⁹ For example, for SLE there have been mixed findings on whether oestrogen or hormone replacement therapy in women increases the risk of developing the disease even though these medications do not place women at a greater risk for severe flares if already diagnosed with SLE.²⁹ Yet, the higher incidence of disease is still conferred with the biological levels of oestrogen and failures in X chromosome inactivation, in addition to increased Toll-like receptor gene products and changes in microRNA function.^{29 30} Further emphasis on these autoimmune conditions theorise that innate and adaptive immune systems interact with hormonal cues that may be related to pathogenesis in such diseases.³¹ Castagnetta et al found that compared with androgen levels, oestrogen levels were significantly elevated in the synovial fluid in patients with RA of both sexes, which was thought to be due to the high aromatase activity from local inflammatory cytokines.³² Furthermore, there is evidence that RA onset and severity is inversely associated with androgen levels exists.^33-36 This phenomenon tentatively explains the increased RA incidence in men >55 years as well as the relatively lower RA severity in men compared with women.^{35 36} It is possible that similar hormonal mechanisms are at play in DED.

Beyond hormone-related factors, other contributing factors to previously reported sex differences in the severity of DED signs include alcohol use,³⁷ diabetes mellitus (DM), migraines, depression and antidepressant medication.³⁸ Specifically, alcohol use is a DED risk factor for women but not for men, where there is a mild protective effect.³⁷ In addition, one previous study found that diabetes increases the risk for symptomatic DED in women but not men for patients <65 years.³⁹ While we did not record alcohol use in this study, in our study, significantly more men had DM; however, DM is not associated with severity of signs and symptoms in the DREAM study.²³ In addition, in our study, more women were taking antidepressant medications compared with men, despite no significant difference in mental health scores. As antidepressant medications are associated with decreased lacrimal secretion, this may have contributed to the more severe DED signs in women seen in our study.^40-42

In this study, women and men are similar in age (58.3 years in women and 56.6 years in men, p = 0.24) and our comparisons in signs and symptoms between women and men were still adjusted by age to minimise any residual confounding effect from age. Because increasing age is associated with higher prevalence of dry eye disease and may be associated with severity of dry eye, it is of interest to evaluate the age effect on dry eye signs and symptoms in both men and women.

In contrast with our study, which did not find a significant difference between sexes for OSDI symptom scores, previous studies have reported that women experience more severe DED symptoms. These contrasting findings could partially be attributed to the different questionnaires used to measure symptom severity for studies that did not use OSDI. In particular, in addition to using the OSDI, our study also included the BODI score, which has not previously been used in other dry eye studies. Further, Schaumberg et al²⁶ used the Symptom Assessment iN Dry Eye (SANDE) assessment as well as a subset of the OSDI to assess DED severity differences between men and women, while Vehof et al used the OSDI and found significant differences between men and women.⁷ While both OSDI and SANDE metrics are validated tools for measuring DED severity, there is a consistent tendency for SANDE to score symptoms higher than OSDI.^{43 44} Schaumberg et al hypothesised that a tendency for women to score higher in SANDE than men, resulting in a significant difference in symptom severity.²⁶

Another possible reason why our results regarding sex differences in DED symptoms differ from other previously published studies is that the baseline DED severity of patients may have been different across studies. For example, the Groningen Longitudinal Sicca Study cohort from Vehof et al demonstrated that in the mild-to-moderate DED group, women had worse symptoms compared with men, but in the severe DED cohorts there was no significant difference.⁷ As the DREAM study enrolled patients with moderate-to-severe DED,²⁰ our finding of no significant difference in symptoms between men and women is consistent with the hypothesis that sex differences in symptom severity may decrease with increased severity of DED.

Premenopausal versus postmenopausal women

In the current study, we found that postmenopausal women presented with more severe DED signs compared with premenopausal women in the DREAM study cohort. However, we did not find any significant difference between the two groups regarding dry eye symptoms.

Similar to the significant differences in DED signs between men and women, more severe DED signs in postmenopausal women could stem from hormonal changes that occur after menopause. This would be consistent with previous studies examining hormonal effects on DED within this population. For example, Gagliano et al reported that postmenopausal women with severe evaporative dry eye had an inverse relationship between 17-b-estradiol, estrone and total testosterone with tear film osmolarity.⁴⁵ Further, Pieragostino et al found that steroids in tears collected from a Schirmer’s strip showed significantly decreased levels of cortisol, 4-androstene-3,17-dione and 17a-hydroxyprogesterone levels in women with DED compared with women without DED.⁴⁶ Ablamowicz et al found that postmenopausal women had a positive correlation between increased meibomian gland dropout rate and higher serum testosterone levels, with a significant difference between mild and severe meibomian gland dropout groups.⁴⁷ There may also be a temporal component affecting DED signs between premenopausal and postmenopausal women, with demonstrated associations between the oestrogen peak of the follicular phase with exacerbated ocular surface dryness, inflammation and increased impairment of tear production and stability.⁴⁸

From an endogenous oestrogen perspective, men and postmenopausal women should be more similar to one another, yet we found that postmenopausal women had relatively worse signs compared with premenopausal women. Thus, how the hormonal pathway contributes to DED sign severity may have additional factors at play compared with differences between sexes. This is further supported by a study by Golebiowski et al,¹³ which demonstrated that oestrogen supplementation increased DED severity while testosterone therapy exhibited no effect on DED symptoms.

It is also possible that changes in feedback mechanisms in postmenopausal women could relate to the increased severity in signs. For example, Gagliano et al reported that in postmenopausal women with severe DED that oestrogen and testosterone levels are lower than in those without severe DED.⁴⁵ In addition, the authors found that tear osmolarity was negatively correlated with these sex hormones. While Ablamowicz et al found that oestrogen and testosterone were both increased at insignificant levels in women with DED compared with women without DED, their study found that in postmenopausal women, there was a decrease in meibomian glands despite the increase in testosterone levels, thus pointing towards a possible mechanism in feedback loop activity resulting in increased severity in DED signs in postmenopausal women.⁴⁷ The increase in testosterone levels at a later point in life, as also demonstrated by Golebiowski’s study, may not have an effect on DED severity.¹³

Previous studies have compared DED symptoms in postmenopausal women with or without DED,⁴⁹ but studies comparing symptoms between premenopausal and postmenopausal women are lacking. In these studies of postmenopausal women, DED symptom severity has been reported with both low and high oestrogen levels.⁵⁰ There have also been various studies regarding the effects of oestrogen and androgen therapy for DED symptoms.^{5 50 51} Jensen et al found postmenopausal women using hormone replacement therapy reported significantly fewer issues compared with postmenopausal women not taking hormone replacement therapy⁵¹ though overall benefit may be limited.⁵⁰ Further, transdermal androgen patches in women who have low serum testosterone demonstrated benefit of DED symptoms.^{50 52} However, large studies directly comparing DED symptoms between premenopausal and postmenopausal women are needed.

CONCLUSION

In this study, we found that women experienced more severe DED signs in comparison to men. In addition, postmenopausal women experienced greater severity of DED signs in comparison to premenopausal women. Interestingly, there was no significant difference in DED symptoms between men and women, or between premenopausal and postmenopausal women. Elucidating the underlying cause of these differences may improve DED diagnosis and provide future direction in understanding sex-related differences and hormonal effects in DED.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• More women than men are affected by dry eye disease (DED) but these differences in DED are not well-characterised.

WHAT THIS STUDY ADDS

• This is a large study from a well-defined cohort that demonstrates how clinical signs differ between men and women and between premenopausal and postmenopausal women.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• By better understanding the sex-related differences in DED signs and symptoms, future research can further focus on mechanisms of disease and treatment.

Data availability statement

Data are available upon reasonable request.