Conjunctival goblet and epithelial cell changes in antidepressant-associated dry eye disease: a case–control study Authors: Rekha Chotaliya, Neeta Mishra, Neelam Prasad, Archana Buch, Parisha Kelkar, Veshal Madan Published: 2026 ABSTRACT PurposeAntidepressants influence tear film stability, but their effects on goblet cell density (GCD) and conjunctival epithelium remain underexplored. This study evaluated GCD and histomorphological changes in antidepressant users versus controls.MethodsIn this case-control study, 72 antidepressant users (cases), and 70 matched controls (aged 20–40 years) underwent tear break-up time (TBUT), Schirmer testing, Ocular Surface Disease Index (OSDI) assessment, corneal/conjunctival staining evaluation, and conjunctival impression cytology. GCD, nuclear-cytoplasmic (NC) fraction, cell-to-cell contact, and nuclear changes were evaluated at 400× magnification. ImageJ software was used for goblet cell counting and measuring NC fraction. Commonly prescribed antidepressants such as selective serotonin reuptake inhibitors, selective norepinephrine reuptake inhibitors and tricyclic antidepressants were included in the study.ResultsCases showed reduced GCD (183 ± 157 vs. 366 ± 164 cells/mm², P < 0.001), NC fraction (0.34 ± 0.09 vs. 0.41 ± 0.08, P < 0.001), TBUT (6.7 ± 3.1 vs. 10.8 ± 4.1 s, P < 0.001), and Schirmer wetting (5.8 ± 2.4 vs. 7.6 ± 3.6 mm/5 min, P < 0.001), with higher OSDI scores (25.6 ± 17.6 vs. 12.3 ± 7.3, P < 0.001), corneal staining (OR 13.0, 95% CI 4.7–36.1), and conjunctival staining (P = 0.009). Morphological grades worsened in cases (all P < 0.001). GCD correlated positively with TBUT (r = 0.41, P < 0.001) and Schirmer (r = 0.43, P < 0.001). No differences existed across SSRI/SNRI/TCA subgroups.ConclusionsAntidepressant use associates with goblet cell loss, epithelial alterations, and dry eye disease (DED), warranting routine ocular monitoring. Introduction The ocular surface system is a highly integrated structure consisting of cornea, conjunctiva, tear film, and associated glands to maintain clear vision and ocular comfort. Three layers of tear film —lipid, aqueous, and mucin—secreted by the meibomian glands, lacrimal glands, and conjunctival goblet cells respectively. The outermost lipid layer is the thinnest and minimizes tear evaporation; the middle aqueous layer lubricates the ocular surface and supplies nutrients; and the mucin layer maintains tear film stability, renders the ocular surface hydrophilic, and resists infection [1]. Disruption of ocular film homeostasis causes dry eye disease (DED), characterized by eye discomfort, visual disturbance, and tear instability, followed by inflammation and tear film hyperosmolarity [2]. Several systemic factors contribute to DED, including autoimmune disease, sleep disorders, malnutrition, and systemic medication use [3–5]. Antidepressants such as selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), and tricyclic antidepressants (TCAs) have been associated with ocular dryness by affecting tear secretion and ocular surface physiology [6–9]. The estimated prevalence of depressive disorders is 4.5% of the total population in India [10]. According to the National Mental Health Survey, the prevalence of depressive disorder was 5.25% [11], and it increased from 23.5% [12] to 25% after the pandemic [13]. Given the growing prevalence of depression and the widespread use of antidepressants, understanding their ocular side effects is clinically relevant. Previous studies have indicated a greater incidence of DED among psychiatric patients [14, 15], but more exploration of the underlying histopathological alterations, particularly their effects on goblet cells is needed. This study aimed to quantify goblet cell density (GCD) and evaluate histomorphological changes in the conjunctival epithelium among patients using antidepressants, and to correlate GCD with routine dry eye investigations. Methodology This case-control study (October 2023-December 2024) followed Helsinki Declaration principles. The study was approved by Institutional Ethics Sub-Committee of Dr. D.Y. Patil Medical College, Hospital and Research Centre, Pimpri, Pune. (Reference no. I.E.S.C./223/2022). For cases, 72 participants aged 20–40 years, taking antidepressants, (SSRIs, SNRIs and TCAs) from psychiatric outpatient department of Dr. D.Y. Patil Medical College, Hospital and Research Centre, and 70 age/sex-matched controls who visit Dr. D.Y. Patil Institute of Optometry and Visual Sciences, Pimpri, Pune for refractive error evaluation were recruited by convenience sampling. All the participants (case/controls) were recruited irrespective of presence of dry eyes. Patients having any systemic disease, alcohol consumption, and ocular surface disease such as allergic conjunctivitis, pterygium, blepharitis, meibomitis, meibomian gland dysfunction, and keratoconus were excluded. Informed consent was obtained from the patient as well as his/her attendant that preceded examinations such as- slit-lamp biomicroscopy, Ocular Surface Disease Index (OSDI) questionnaire, the tear break-up time (TBUT), and the Schirmer test (post-0.5% proparacaine), prior to conjunctival impression cytology specimen collection. The measurements were recorded as follows: OSDI: A 12-item questionnaire assessing symptoms and visual function [16]. Scores 0–12 were normal, 13–22 were mild, 23–32 were moderate, and 33 were severe [17].TBUT: Measured as the interval between the last blink and onset of first dry spot under cobalt blue light. A TBUT < 10 s was considered abnormal [2].Schirmer test: Using Whatman 41 filter paper; <6 mm wetting for 5 min indicated abnormalities [18]. OSDI: A 12-item questionnaire assessing symptoms and visual function [16]. Scores 0–12 were normal, 13–22 were mild, 23–32 were moderate, and 33 were severe [17]. OSDI: A 12-item questionnaire assessing symptoms and visual function [16]. Scores 0–12 were normal, 13–22 were mild, 23–32 were moderate, and 33 were severe [17]. TBUT: Measured as the interval between the last blink and onset of first dry spot under cobalt blue light. A TBUT < 10 s was considered abnormal [2]. TBUT: Measured as the interval between the last blink and onset of first dry spot under cobalt blue light. A TBUT < 10 s was considered abnormal [2]. Schirmer test: Using Whatman 41 filter paper; <6 mm wetting for 5 min indicated abnormalities [18]. Schirmer test: Using Whatman 41 filter paper; <6 mm wetting for 5 min indicated abnormalities [18]. Conjunctival impression cytology Conjunctival impression cytology used mixed cellulose ester filter paper of pore size 0.45-µm and 13 mm diameter. It was cut in to half, placed on superior bulbar conjunctiva and gently pressed with applanation tonometer head. A filter paper removed gently with forceps after 10 s and fixed in 95% ethanol, and stained with periodic acid-Schiff (Table 1). Table 1Periodic acid-Schiff staining stepwise [19]StepSolutionTimePurpose195% ethanol10 minFixation*270% ethanol1 minRehydration31% Periodic acid2 minfor oxidation of glycol groups to aldehydes.4Wash in tap water30 s-5Schiff reagent10 minStains goblet cells to magenta colour.6Wash in tap water1 min-7Harris hematoxylin50 sTo stain the nucleus8Wash in tap water30 s-995% ethanol4 minDehydration**10Absolute alcohol3 minDehydration**11Xylene10–15 dipsTo make the filter paper transparent*12Mounting by DPX (Dibutylphthalate Polystyrene Xylene) and coverslip on the glass slide*More exposure of a sample to alcohol or xylene deteriorates the view of cellular morphology and quality of the stained sample under the microscope** Stirring required Table 1Periodic acid-Schiff staining stepwise [19]StepSolutionTimePurpose195% ethanol10 minFixation*270% ethanol1 minRehydration31% Periodic acid2 minfor oxidation of glycol groups to aldehydes.4Wash in tap water30 s-5Schiff reagent10 minStains goblet cells to magenta colour.6Wash in tap water1 min-7Harris hematoxylin50 sTo stain the nucleus8Wash in tap water30 s-995% ethanol4 minDehydration**10Absolute alcohol3 minDehydration**11Xylene10–15 dipsTo make the filter paper transparent*12Mounting by DPX (Dibutylphthalate Polystyrene Xylene) and coverslip on the glass slide*More exposure of a sample to alcohol or xylene deteriorates the view of cellular morphology and quality of the stained sample under the microscope** Stirring required Periodic acid-Schiff staining stepwise [19] *More exposure of a sample to alcohol or xylene deteriorates the view of cellular morphology and quality of the stained sample under the microscope ** Stirring required Microscopic analysis Microscopic evaluation under 400x magnification, was carried out to assess the morphology of epithelial cells which included the GCD, NC fraction, cell-to-cell contact, type of nuclear changes and frequency of nuclear changes [20]. The GCD and NC fractions were measured via ImageJ software [21–23] and classified on the basis of Nelson’s grading system [24]. To count goblet cells, five random areas of the microscope under 400x magnification were selected from the PAS-stained samples. The goblet cells in each area were counted, and the mean number of cells in the five areas was divided by the field area of 0.20mm2 to obtain the GCD in cells/mm2 [19]. The NC ratio was measured by dividing the diameter of the nucleus and the diameter of cytoplasm of individual epithelial cells [19]. Here, the NC ratio was calculated as a decimal, which was then referred to as the NC fraction. All the morphological parameters were graded as shown in Table 2. Table 2Grading of ocular surface histomorphological parametersVariableGrade 0Grade 1Grade 2Grade 3AbnormalityGCD [24]500 cells/mm²350–500 cells/mm²100–350 cells/mm²< 100 cells/mm²Grade 2 and 3NC fraction [24]0.41–0.500.31–0.400.21–0.300.11–0.20Grade 1 to 3Cell-to-cell contact [20]DenseSlightly loosenedLoosenedSingle cellsFrequency of nuclear changes [20]NoneSporadicSeveralAbundantTypes of nuclear changes [20]Nonelow grade- dissociation of chromatin, chromatin condensation1 + snake-like chromatin2 + nuclear fragmentation Table 2Grading of ocular surface histomorphological parametersVariableGrade 0Grade 1Grade 2Grade 3AbnormalityGCD [24]500 cells/mm²350–500 cells/mm²100–350 cells/mm²< 100 cells/mm²Grade 2 and 3NC fraction [24]0.41–0.500.31–0.400.21–0.300.11–0.20Grade 1 to 3Cell-to-cell contact [20]DenseSlightly loosenedLoosenedSingle cellsFrequency of nuclear changes [20]NoneSporadicSeveralAbundantTypes of nuclear changes [20]Nonelow grade- dissociation of chromatin, chromatin condensation1 + snake-like chromatin2 + nuclear fragmentation Grading of ocular surface histomorphological parameters Sample size The sample size was calculated via Openepi software [25] by referencing a study by E Shrestha et al. [26] They assessed 114 dry eye cases and 72 controls and reported that 49.2% of the cases and 26.3% of the controls had abnormal GCD. To obtain sample size the confidence interval was kept at 95%, the power at 80%, the ratio of controls to cases at 1, the proportion of controls with exposure at 26.3%, and the proportion of cases with exposure at 49.2%. The obtained total sample size was 140; 70 cases and 70 controls. Statistical analysis WinPepi and Jamovi software were used for statistical analysis [27, 28]. Normality was tested with the Shapiro–Francia test. Non-normal distribution was analysed using Mann–Whitney U, Kruskal–Wallis, and Bonferroni–Dunn post hoc tests. Pearson’s correlation assessed the relationships between GCD, with TBUT, and Schirmer test, and spearman’s correlation coefficient was used to correlate GCD with cell-to-cell contact, type of nuclear change and frequency of nuclear changes. Significant predictors were identified by multiple linear regression analysis. A P-value < 0.05 was considered statistically significant results. Result A total of 142 participants, 72 cases and 70 controls, were enrolled in the study. The cases included 42 males (58.33%) and 30 females (41.66%) with a mean age of 30.68 ± 5.37 years, and the age matched controls (P = 0.141) included 39 males (55.71%) and 31 females (44.28%) with a mean age of 32.14 ± 6.36 years. All baseline parameters of cases and controls are detailed in Table 3. Among the 72 cases, the common psychiatric disorders involved depression (33.33%) and anxiety (26.38%), followed by obsessive-compulsive disorder (9.72%), schizophrenia (8.33%), and mixed anxiety and depression (6.94%). Another 15% included somatoform disorder, delusion disorder, panic disorder, adjustment disorder, bodily distress disorder, conversion disorder, mood disorder and acute and transient psychiatric disorder. The mean duration of taking medicine was 12.52 ± 7.23 (median 12, IQR 6 to 16) months. Majority of the participants were taking SSRIs (75%), followed by SNRIs (11.11%), and TCAs (13.88%). Table 3Frequency of normal and abnormal ocular surface paramaeters of the participants among case and controlsVariablesGradeCase (n = 72)Controls (n = 70)Frequency (n)Percentage (%)Frequency (n)Percentage (%)GCD (in cells/mm2)Normal (> 500)34.11825.7Slightly abnormal (350–500)79.722434.2Abnormal (100–350)3345.82535.7Significantly abnormal (< 100)2940.234.3NC fraction0.41–0.501926.44462.90.31–0.402940.31515.70.21–0.301520.81121.40.19–0.20912.500Cell-to-cell contactDense (0)1622.24057.1Slightly loosened (1)3041.62231.4Loosened (2)2433.3811.4Single cells (3)22.7800Frequency of nuclear changesNone (0)2636.14767.1Sporadic (1)1419.41318.5Several (2)1520.834.29Abundant (3)1723.6710.0Types of Nuclear changesNone (0)2433.34767.1Low graded: chromatin condensation (1)1723.61115.71 + snakelike chromatin (2)1419.434.292 + nuclear fragmentation (3)1723.6912.8TBUT(in seconds)Normal (> 10)1216.74158.5Mild (8–10)1419.41318.5Moderate (5–7)2433.3108.5Severe (< 5)2230.6614.2Schirmer test (in mm)Normal (> 6)3650.05274.3Abnormal (< 6)3650.01825.7OSDI scoreNormal (0–12)2027.84057.1Mild (13–22)1419.42433.3Moderate (23–32)1622.2457.1Severe (33–100)2230.622.8Corneal stainingAbsent (0)3650.06592.8Present (1)3650.057.1Conjunctival stainingAbsent (0)3548.65071.4Present (1)3751.42028.6 Table 3Frequency of normal and abnormal ocular surface paramaeters of the participants among case and controlsVariablesGradeCase (n = 72)Controls (n = 70)Frequency (n)Percentage (%)Frequency (n)Percentage (%)GCD (in cells/mm2)Normal (> 500)34.11825.7Slightly abnormal (350–500)79.722434.2Abnormal (100–350)3345.82535.7Significantly abnormal (< 100)2940.234.3NC fraction0.41–0.501926.44462.90.31–0.402940.31515.70.21–0.301520.81121.40.19–0.20912.500Cell-to-cell contactDense (0)1622.24057.1Slightly loosened (1)3041.62231.4Loosened (2)2433.3811.4Single cells (3)22.7800Frequency of nuclear changesNone (0)2636.14767.1Sporadic (1)1419.41318.5Several (2)1520.834.29Abundant (3)1723.6710.0Types of Nuclear changesNone (0)2433.34767.1Low graded: chromatin condensation (1)1723.61115.71 + snakelike chromatin (2)1419.434.292 + nuclear fragmentation (3)1723.6912.8TBUT(in seconds)Normal (> 10)1216.74158.5Mild (8–10)1419.41318.5Moderate (5–7)2433.3108.5Severe (< 5)2230.6614.2Schirmer test (in mm)Normal (> 6)3650.05274.3Abnormal (< 6)3650.01825.7OSDI scoreNormal (0–12)2027.84057.1Mild (13–22)1419.42433.3Moderate (23–32)1622.2457.1Severe (33–100)2230.622.8Corneal stainingAbsent (0)3650.06592.8Present (1)3650.057.1Conjunctival stainingAbsent (0)3548.65071.4Present (1)3751.42028.6 Frequency of normal and abnormal ocular surface paramaeters of the participants among case and controls TBUT (in seconds) Comparison of ocular surface parameters among cases and controls Comparison by Mann-Whitney U test (Table 4a, 4b) revealed that the GCD (P < 0.001) and NC fractions (P < 0.001) were statistically significantly reduced among cases. Additionally, in terms of morphological parameters, significantly increased grades of cell-to-cell contact (P < 0.001), frequency of nuclear changes (P < 0.001), and type of nuclear changes (P < 0.001) were observed (Figs. 1 and 2). The result of routine dry eye investigations such as TBUT (P < 0.001) and the Schirmer test (P = 0.001) were also statistically significantly reduced, with the OSDI score being higher among the cases (P < 0.001) than among the control group, with a greater prevalence of positive corneal staining (P < 0.001) and conjunctival staining (P = 0.009) among the cases. Table 4aComparison of ocular surface parameters among cases and controlsSr. no.VariablesCase(n = 72)Control(n = 70)ZUP (two tailed)1.GCD182.99 ± 157.33366.14 ± 163.626.076493< 0.0012.NC fraction0.339 ± 0.090.406 ± 0.004.676124.5< 0.0013.Cell-to-cell contact1.17 ± 0.800.54 ± 0.694.593951.0< 0.0014.Frequency of nuclear changes1.32 ± 1.190.57 ± 0.973.954112.0< 0.0015.Type of Nuclear changes1.33 ± 1.170.63 ± 1.053.884122< 0.0016.TBUT6.67 ± 3.0810.83 ± 4.125.836432< 0.0017.Schirmer test5.77 ± 2.387.59 ± 3.593.335817.50.0018.OSDI25.55 ± 17.5812.3 ± 7.314.943795< 0.001P < 0.05 significant Table 4aComparison of ocular surface parameters among cases and controlsSr. no.VariablesCase(n = 72)Control(n = 70)ZUP (two tailed)1.GCD182.99 ± 157.33366.14 ± 163.626.076493< 0.0012.NC fraction0.339 ± 0.090.406 ± 0.004.676124.5< 0.0013.Cell-to-cell contact1.17 ± 0.800.54 ± 0.694.593951.0< 0.0014.Frequency of nuclear changes1.32 ± 1.190.57 ± 0.973.954112.0< 0.0015.Type of Nuclear changes1.33 ± 1.170.63 ± 1.053.884122< 0.0016.TBUT6.67 ± 3.0810.83 ± 4.125.836432< 0.0017.Schirmer test5.77 ± 2.387.59 ± 3.593.335817.50.0018.OSDI25.55 ± 17.5812.3 ± 7.314.943795< 0.001P < 0.05 significant Comparison of ocular surface parameters among cases and controls P < 0.05 significant Table 4bComparison of corneal and conjunctival staining among cases and controlsVariablesChi-squaredfP valueCorneal staining29.6911< 0.001Conjunctival staining6.77010.009P < 0.05 significant Table 4bComparison of corneal and conjunctival staining among cases and controlsVariablesChi-squaredfP valueCorneal staining29.6911< 0.001Conjunctival staining6.77010.009P < 0.05 significant Comparison of corneal and conjunctival staining among cases and controls P < 0.05 significant Fig. 1GCD and cell-to-cell contact according to severity grades. The photomicrograph shows (a) Grade 0- Small, dense, and round epithelial cells with normal nuclei and abundant goblet cells. (b) Grade 1-Slightly loosened epithelial cells with reduced GCD. (c) Grade 2 -Loosened sheet of enlarged polygonal epithelial cells with significantly reduced goblet cells (d) Grade 3-Single and folded polygonal epithelial cells with pyknotic nuclei and the absence of goblet cells. (PAS,400X) GCD and cell-to-cell contact according to severity grades. The photomicrograph shows (a) Grade 0- Small, dense, and round epithelial cells with normal nuclei and abundant goblet cells. (b) Grade 1-Slightly loosened epithelial cells with reduced GCD. (c) Grade 2 -Loosened sheet of enlarged polygonal epithelial cells with significantly reduced goblet cells (d) Grade 3-Single and folded polygonal epithelial cells with pyknotic nuclei and the absence of goblet cells. (PAS,400X) Fig. 2Conjunctival smear reveal different types of nuclear changes in epithelial cells according to severity. Photomicrograph shows (a) Grade 0- Small, round cuboidal epithelial cells with basophilic nuclei. (b) Grade 1- Dissociation of nuclear chromatin indicated by a yellow arrow (c) Grade 2- squamous cells with snake-like nuclear chromatin. (d) Grade 3- Apoptic cell showing nuclear fragmentation. (PAS,400X) Conjunctival smear reveal different types of nuclear changes in epithelial cells according to severity. Photomicrograph shows (a) Grade 0- Small, round cuboidal epithelial cells with basophilic nuclei. (b) Grade 1- Dissociation of nuclear chromatin indicated by a yellow arrow (c) Grade 2- squamous cells with snake-like nuclear chromatin. (d) Grade 3- Apoptic cell showing nuclear fragmentation. (PAS,400X) Analysis of odds ratios for antidepressant-related ocular adverse events The increase in the odds ratio (Fig. 3) revealed that all the morphological parameters and results of routine dry eye investigations were more likely to be affected by the use of antidepressants. Among all the parameters, GCD (odds ratio = 9.3, 95% CI = 4.09–21.14) and TBUT (odds ratio = 7.0, 95% CI = 3.236–15.44) were more likely to be affected by the use of antidepressants. The presence of corneal staining (odds ratio = 13, CI 4.687–36.05) was more likely with the use of antidepressants. Fig. 3Forest plot showing odds ratios for antidepressant-related ocular adverse events Forest plot showing odds ratios for antidepressant-related ocular adverse events Comparison across different classes of antidepressants SSRIs, SNRIs, and TCAs Kruskal-Wallis test revealed no significant differences in ocular surface parameters (all P > 0.05) among SSRIs, SNRIs, and TCAs. Clinically lower values of the GCD, NC fraction, TBUT and Schirmer test measurements (Table 5a) and a higher prevalence of corneal and conjunctival staining (Table 5b) among the TCA group were observed, but were statistically insignificant. Table 5aComparison of ocular surface parameters across different classes of antidepressants (SSRIs, SNRIs, and TCAs)VariableSSRI(n = 54)SNRI(n = 8)TCA(n = 10)SSRI(n = 54)SNRI(n = 8)TCA(n = 10)H(df)P valueMean ± SDMedianGCD181.20 ± 148.78203.13 ± 256.72176.50 ± 115.1113597.50187.500.39(2)0.822NC fraction0.34 ± 0.090.35 ± 0.110.29 ± 0.090.360.360.293.73(2)0.155Cell-to-cell contact1.13 ± 0.781.25 ± 0.711.30 ± 1.061.001.001.500.43(2)0.807Frequency of nuclear changes1.28 ± 1.171.63 ± 1.411.30 ± 1.251.002.001.000.48(2)0.787Types of nuclear changes1.33 ± 1.201.50 ± 1.201.20 ± 1.141.001.501.000.27(2)0.875TBUT6.86 ± 3.356.81 ± 2.105.50 ± 1.876.256.255.751.05(2)0.592Schirmer test5.94 ± 2.376.19 ± 2.534.55 ± 2.106.006.004.003.21(2)0.201OSDI27.02 ± 18.8220.84 ± 14.3621.37 ± 11.9522.5018.7523.750.82(2)0.663P < 0.05 significant Table 5aComparison of ocular surface parameters across different classes of antidepressants (SSRIs, SNRIs, and TCAs)VariableSSRI(n = 54)SNRI(n = 8)TCA(n = 10)SSRI(n = 54)SNRI(n = 8)TCA(n = 10)H(df)P valueMean ± SDMedianGCD181.20 ± 148.78203.13 ± 256.72176.50 ± 115.1113597.50187.500.39(2)0.822NC fraction0.34 ± 0.090.35 ± 0.110.29 ± 0.090.360.360.293.73(2)0.155Cell-to-cell contact1.13 ± 0.781.25 ± 0.711.30 ± 1.061.001.001.500.43(2)0.807Frequency of nuclear changes1.28 ± 1.171.63 ± 1.411.30 ± 1.251.002.001.000.48(2)0.787Types of nuclear changes1.33 ± 1.201.50 ± 1.201.20 ± 1.141.001.501.000.27(2)0.875TBUT6.86 ± 3.356.81 ± 2.105.50 ± 1.876.256.255.751.05(2)0.592Schirmer test5.94 ± 2.376.19 ± 2.534.55 ± 2.106.006.004.003.21(2)0.201OSDI27.02 ± 18.8220.84 ± 14.3621.37 ± 11.9522.5018.7523.750.82(2)0.663P < 0.05 significant Comparison of ocular surface parameters across different classes of antidepressants (SSRIs, SNRIs, and TCAs) P < 0.05 significant Table 5bComparison of corneal and conjunctival staining across different classes of antidepressants SSRIs, SNRIs, and TCAsVariableFisher’s PCorneal staining0.151Conjunctival staining0.141P < 0.05 significant Table 5bComparison of corneal and conjunctival staining across different classes of antidepressants SSRIs, SNRIs, and TCAsVariableFisher’s PCorneal staining0.151Conjunctival staining0.141P < 0.05 significant Comparison of corneal and conjunctival staining across different classes of antidepressants SSRIs, SNRIs, and TCAs P < 0.05 significant Correlation of GCD with ocular surface parameters among cases A significant moderate positive correlation was observed between the GCD with TBUT (r = 0.413) and Schirmer test (r = 0.432). The duration of antidepressant intake did not affect GCD (r = -0.03). Additionally, cell-to-cell contact (rho= -0.530), the frequency of nuclear changes (rho = -0.575) and the type of nuclear changes (rho = -0.526) were negatively associated with the GCD (Table 6). The presence of conjunctival staining was also significantly associated with abnormal GCD grades (P = 0.015) (Table 7). Table 6Correlation of the GCD with the duration of antidepressant use and other ocular surface parameters among casesGCDCorrelationP valueDurationPearson (r)-0.030.905TBUTPearson (r)0.41< 0.001Schirmer testPearson (r)0.43< 0.001Cell-to-cell contactSpearman (rho)-0.53< 0.001Frequency of nuclear changesSpearman (rho)-0.58< 0.001Type of nuclear changesSpearman (rho)-0.53< 0.001P < 0.05 significant Table 6Correlation of the GCD with the duration of antidepressant use and other ocular surface parameters among casesGCDCorrelationP valueDurationPearson (r)-0.030.905TBUTPearson (r)0.41< 0.001Schirmer testPearson (r)0.43< 0.001Cell-to-cell contactSpearman (rho)-0.53< 0.001Frequency of nuclear changesSpearman (rho)-0.58< 0.001Type of nuclear changesSpearman (rho)-0.53< 0.001P < 0.05 significant Correlation of the GCD with the duration of antidepressant use and other ocular surface parameters among cases P < 0.05 significant Table 7Association between GCD grades with presence of conjunctival staining among antidepressant usersGCD rangesConjunctival staining(n)Fisher exact testNegativePositiveTotal< 100131629 0.015 100–350132033350–500707500213Total3537 72 P < 0.05 significant Table 7Association between GCD grades with presence of conjunctival staining among antidepressant usersGCD rangesConjunctival staining(n)Fisher exact testNegativePositiveTotal< 100131629 0.015 100–350132033350–500707500213Total3537 72 P < 0.05 significant Association between GCD grades with presence of conjunctival staining among antidepressant users P < 0.05 significant A multiple linear regression analysis was performed to identify predictors of GCD. The overall model was statistically significant (F(7, 134) = 19.6, P < 0.001) and accounted for approximately 48% of the variance in GCD (Adjusted R2 = 0.480) (Table 8). The use of antidepressants was a significant factor, demonstrating significantly lower GCD among antidepressant users (β = 95.17, SE = 29.16, P = 0.001) (Table 9). Several other clinical markers such as lower TBUT scores (β = 11.91, P < 0.001) and lower Schirmer’s test results (β = 20.92, P < 0.001) were also emerged as significant independent predictors of GCD. Conversely, patient age, OSDI scores, and both corneal and conjunctival staining did not significantly contribute to the model (all P > 0.05), suggesting these factors were not independent predictors of GCD when controlling for other variables. Table 8Model fit measures R R²Adjusted R²Overall Model TestFdf1df2 P 0.7110.5060.48019.67134< 0.001 Table 8Model fit measures R R²Adjusted R²Overall Model TestFdf1df2 P 0.7110.5060.48019.67134< 0.001 Model fit measures Table 9Model coefficients - GCDPredictorEstimateSEt P Intercept44.8766.470.6750.501Age-1.991.624-1.2240.223TBUT11.913.2353.681< 0.001Schirmers test20.923.9395.310< 0.001OSDI score1.240.8631.4370.153AntidepressantsNo-Yes95.1729.1633.2640.001Corneal staining-22.4333.546-0.6690.505Conjunctival staining-38.4228.151-1.3650.175 Table 9Model coefficients - GCDPredictorEstimateSEt P Intercept44.8766.470.6750.501Age-1.991.624-1.2240.223TBUT11.913.2353.681< 0.001Schirmers test20.923.9395.310< 0.001OSDI score1.240.8631.4370.153AntidepressantsNo-Yes95.1729.1633.2640.001Corneal staining-22.4333.546-0.6690.505Conjunctival staining-38.4228.151-1.3650.175 Model coefficients - GCD Antidepressants No-Yes Discussion The study was conducted to examine the tear film mucin layer by quantifying the GCD, and associated epithelial cells integrity to determine the etiology of dry eye in antidepressant users. This understanding is crucial for targeting the deficient layer of the tear film. This study focused on the morphology of conjunctival epithelial cells in patients taking antidepressants. The study parameters included histomorphological parameters such as GCD, the NC fraction, cell-to-cell contact, the frequency of nuclear changes, the type of nuclear changes. Simultaneously, routine investigations for dry eye such as TBUT, the Schirmer test, the presence or absence of conjunctival staining, and corneal staining were also evaluated. In the present study population, the age group of 20–40 years was considered to exclude the chances of dry eye associated with aging [29]. The depression and anxiety were the common psychiatric illness observed among cases. In our study, it was observed that, antidepressant use significantly altered ocular surface morphology and tear film stability, primarily via reduced GCD and epithelial changes. Notably, GCD and TBUT exhibited the greatest changes, linking antidepressant exposure to tear instability and ocular surface inflammation. In this study, the exclusion of meibomian gland abnormalities prevented the likelihood of evaporative dry eye due to lipid layer deficiency, indicating that evaporative dry eye due to mucin deficiency was associated with goblet cell loss among cases. In routine dry eye investigations, a lower TBUT, and Schirmer test, a higher OSDI score and more prevalent corneal and conjunctival staining indicate more dryness among cases. This aligns with previous research linking antidepressants to DED [6–8]. All antidepressant classes (SSRIs, SNRIs, TCAs) produced comparable alterations in clinical and morphological ocular surface parameters. Reduced cell-to-cell contact, with increased frequency and types of nuclear changes indicating the loss of epithelial cohesion, increased condensation of nuclear chromatin, and fragmentation—point toward accelerated apoptosis. Inflammatory mediators such as IL-1β, IL-6, and TNF-α, which are activated via the NF-κB and MAPK pathways, may underlie these changes among SSRI users [30, 31]. More pronounced ocular surface changes in the TCA group could be the result of the anticholinergic property of TCAs [32]. TCAs likely contribute to GCD loss via potent anticholinergic activity, which inhibits the muscarinic signalling required for goblet cell maintenance and mucus secretion. The positive correlation of the GCD with the TBUT and Schirmer test results reinforces the integrated function of the tear film layers. Goblet cell loss reduces mucin availability, promoting tear hyperosmolarity and perpetuating the inflammatory cycle of DED [33–36]. A significant positive association between presence of conjunctival staining and severity of GCD depletion revealed that the presence of conjunctival staining is a biomarker of goblet cell loss. Multiple linear regression analysis revealed that alteration in GCD was strongly associated with antidepressant use. The strong positive associations between TBUT, Schirmer’s test, and GCD suggesting antidepressant-associated dry eye contribute to ocular surface damage by altering the mucin and aqueous secretion. This mechanistic insight aligns with prior observations of antidepressant-induced dry eye [33]. Conversely, OSDI symptoms and age showed no independent association after adjustment, suggesting that altered GCD reflects structural rather than symptomatic or age-related compromise. These results support targeted screening for antidepressant use in contact lens users. Routine ocular surface evaluation is essential for patients initiating the course of antidepressants. Early recognition of ocular side effects allows timely intervention with lubricants, mucin-stimulating agents, and interprofessional collaboration between psychiatrists and eye care providers. Limitations This study has several limitations. First, the monocentric design may limit the generalizability of our findings to more diverse patient populations. Because we used non-probability sampling hence lack of randomization may have induced a selection bias. This means our participants’ characteristics and baseline disease severity might not perfectly represent the general population. Furthermore, the inclusion of different antidepressants such as SSRIs, SNRIs, and TCAs introduces pharmacological heterogeneity. Since these drugs have different mechanisms of action and systemic side effects, we cannot determine if the observed effects were drug-specific. This issue was compounded by a small sample size and an unequal distribution of patients across medication classes. Finally, the lack of incorporating a standardized grading system for the cornea and conjunctiva may have affected the consistency of our objective clinical assessments. Strength of the study The effect of antidepressants on ocular surface morphology has been overlooked in previous research, which fills a significant gap in the present study [6–8, 15]. By investigating how antidepressants influence ocular surface morphology, this study opens a new avenue for understanding the broader implications of antidepressant therapy. Additionally, the findings underscore the importance of regular eye examinations for patients undergoing psychiatric treatment. As changes in the GCD may adversely affect ocular health, this study emphasizes the need for healthcare providers to remain vigilant in monitoring their patients’ visual health as part of comprehensive psychiatric care. Conclusion Antidepressant therapy is associated with significant ocular surface alterations, including reduced goblet cell density, epithelial damage, and decreased tear stability, all of which contribute to DED. Regular ocular examinations, patient education, and coordinated management between psychiatrists and eye care professionals are essential to mitigate ocular complications in antidepressant users.