Lacrimal Gland Volume Changes in Unilateral Primary Acquired Nasolacrimal Obstruction

Abstract

Purpose: To investigate whether unilateral primary acquired nasolacrimal duct obstruction (PANDO) causes Schirmer score and lacrimal gland volume changes in the contralateral non-PANDO eye.

Methods: Sixteen unilateral female PANDO and 16 female controls were enrolled in the study. Exclusion criteria were orbital trauma, inflammation, infection, tumor involvement or infiltrative diseases, history of ocular surgery, ocular surface disorder, systemic drug use that interferes with tear secretion, and chronic use of topical eye drops. Bilateral lacrimal gland volumes were measured in computed tomography (CT) images of the participants. A Schirmer test without anesthesia was also performed on each participant.

Results: As there was no significant difference between the right and left eye values for Schirmer and gland size (P > 0.05), both eyes of the control group were enrolled in the study. The groups were age matched and the mean lacrimal gland was 0.479 cm³ in the PANDO as well as the contralateral non-PANDO side, which was statistically smaller compared to the control eyes (0.580 cm³) (P = 0.04). The mean Schirmer scores in the same order were 18.5 ± 7.1 mm, 13.2 ± 8.9 mm, and 21.3 ± 10.5 mm, respectively. Non-PANDO side Schirmer scores were lower compared to the other two groups, but statistical significance was present for the control group (P = 0.04).

Conclusions: Compared to the control group, lacrimal gland volumes were bilaterally smaller in unilateral PANDO patients. Schirmer scores were statistically lower in the contralateral non-PANDO side compared to the controls.

The lacrimal gland is located in the lacrimal fossa lying in the superolateral part of the orbit and its function is to secrete and excrete tears into the superior fornix. The lacrimal gland, cornea, conjunctiva, meibomian, and accessory lacrimal glands have been defined as the lacrimal functional unit (LFU).¹ If any part of the LFU is compromised, lacrimal gland support to the ocular surface is impeded. Tear secretion is regulated by the neural response originating from the sensory afferent nerves of the ocular surface, which is processed in the lacrimal nucleus of the brain along with other sensory input (i.e., emotional). The efferent signal is then transmitted to the lacrimal gland, accessory lacrimal glands, goblet cells, and meibomian glands. The stimulus is graded according to the amount of tears needed to lubricate the ocular surface.² 

Primary acquired nasolacrimal duct obstruction (PANDO), mostly seen in postmenopausal females, is one of the most common causes of epiphora. It is associated with chronic inflammation and fibrosis of the nasolacrimal duct.³ It is possible that delayed tear clearance in PANDO patients may result in decreased neural discharge from the ocular surface, and decreased lacrimal gland stimulation and involution. 

Lacrimal gland size increases in certain situations like idiopathic orbital inflammation, Wegener granulomatosis, sarcoidosis, thyroid eye disease, tumoral involvement and decreases in senile atrophy, and when there is a lack of trophic factors.^4–7 Various studies have been performed to determine the normal dimensions and volume of the lacrimal gland using ultrasound, computed tomography (CT), and magnetic resonance imaging (MRI).^8–11 However, to the best of our knowledge, the lacrimal gland volume in PANDO patients has not been published previously. 

The current study investigates the effect of unilateral PANDO on Schirmer scores and lacrimal gland volumes of both the PANDO and non-PANDO side and compares them with controls. 

Materials and Methods

Patient Selection

The study was approved by the institutional ethics committee and was conducted with adherence to the tenets of the Helsinki declaration. The study was conducted between May 2014 and February 2015. The diagnosis of PANDO in patients who presented to the ophthalmology clinic with epiphora was confirmed with a lack of patency on lacrimal irrigation. Unilateral PANDO patients were consulted to the otorhinolaryngology department preoperatively, and the ones in whom the otorhinololaryngologist requested paranasal CT imaging underwent Schirmer test evaluation bilaterally. Control patients consisted of subjects who attended the otorhinolaryngology day clinic and in whom paranasal CT imaging was requested. The control patients had also undergone a bilateral Schirmer test evaluation before CT imaging was performed. Exclusion criteria applicable to all participants were possible diseases that might affect the size of the lacrimal gland such as orbital trauma, inflammation, infection, tumoral involvement, and infiltrative diseases. History of ocular surgery, ocular surface disorder, ocular diseases like glaucoma that necessitates a continuous use of chronic topical medications, and the use of systemic medications that may affect tear secretion were the other exclusion criteria. Finally, 16 eligible female PANDO patients and 16 female control subjects were included in the study. 

Image Analysis

Non-enhanced CT exams were performed with a 64-slice scanner (Toshiba Aquilion 64, Otawara, Japan). Continuous nonoverlapping sections of CT scan were obtained with acquisition parameters of 0.5 mm slice thickness, 100 to 120 kV, and 150 to 200 mAs. The computed tomography data sets were stored as Digital Imaging and Communication in Medicine (DICOM) files. Axial images in a soft tissue algorithm were used for segmentation processing. For segmentation, the lacrimal glands were outlined with a free-hand technique using the pencil tool in all consecutive images in which the lacrimal gland was present (Fig. 1). After outlining the gland borders, a three-dimensional (3D) image of the lacrimal gland was reconstructed using segmented images on the workstation (Aquarious iNtuition, version 4.6; TeraRecon, San Mateo, CA, USA). The volumes of the lacrimal gland were automatically calculated by the workstation with a 3D reconstruction in cubic centimeters. Each lacrimal gland volume was measured separately. To determine interobserver variations, two blinded observers (EB and BY) measured the lacrimal glands in CT images. After reliability was determined, the measurements of one investigator (EB) were used for statistical analysis. The intraobserver reliability was also assessed. 

Figure 1

 

Lacrimal gland measurement on CT.

Schirmer Test

A stimulated Schirmer test (Mandhu Instruments, New Delphi, India) was performed without anesthetic with the eye closed for 5 minutes after the strip was inserted at the junction of the lateral and middle third of the lower conjunctival sac, avoiding contact with the cornea, and the length in millimeters was recorded after 5 minutes. 

Statistical Analysis

Shapiro-Wilk test was used to evaluate normality of distribution. Normal distribution was found in patients and controls for lacrimal gland volumes but not for age and Schirmer measurements. Mann-Whitney U test was used for age comparison between PANDO and control subjects. One-way ANOVA and post hoc least significant difference tests were used to compare volumes of lacrimal gland in the PANDO side, the contralateral non-PANDO side, and the control group. Kruskall-Wallis test and Mann-Whitney U test with bonferroni correction was performed for comparison in terms of Schirmer measurements. Intraclass correlation coefficient (ICC) was calculated with the ICC interpretations described by Landis and Koch.¹² The post hoc power calculation was performed to detect the significance of the difference between lacrimal gland volumes of patients and controls. 

Results

The mean age was 49.31 ± 7.46 years in PANDO patients and 46.63 ± 5.57 years in control subjects. There was no significant difference between groups in terms of age (P = 0.46). The mean right and left gland volumes and Schirmer values in the control group were not statistically different; therefore, 32 eyes of 16 control subjects were accepted as the control eyes (P = 0.62 and 0.74, respectively). Finally, the study included 16 PANDO and 16 contralateral non-PANDO and 32 control eyes. The gland volume measurements of the two observers showed excellent agreement with an ICC of 0.837. The intraobserver reliability was also excellent with an ICC of 0.916. The mean lacrimal gland volume of PANDO patients was 0.479 ± 0.156 cm³ in the PANDO side and 0.479 ± 0.141 cm³ in the contralateral non-PANDO side, and both were statistically smaller compared to the control lacrimal gland volume, which was 0.580 ± 0.174 cm³ (P < 0.05) (see Fig. 2). Comparison of patients and control group gland volumes achieved a power of 86% in detecting a difference of 0.3 between the null hypothesis that both group means are 0.5 and the alternative hypothesis that the mean of group 2 is 0.1 with estimated group standard deviations of 0.6 and 0.2 and with a significance level (α) of 0.05. The Schirmer scores were 18.5 ± 7.1 mm, 13.2 ± 8.9 mm, and 21.3 ± 10.5 mm for PANDO, non-PANDO side, and control eyes, respectively (see Fig. 3). The non-PANDO side Schirmer scores were statistically lower compared to control eyes (P = 0.04). The PANDO side Schirmer scores were also lower compared to the non-PANDO and difference was close to the statistically significant level (P = 0.06). The Table summarizes the results. 

Figure 2

 

Lacrimal gland volume distribution among groups.

Figure 3

 

Schirmer measurements among groups.

Table

 

Comparison of Lacrimal Gland Volumes and Schirmer Scores Among Groups

 

Discussion

Tear secretion is normally controlled by LFU that is composed of the ocular surface (cornea, conjunctiva, and meibomian glands), main and accessory lacrimal glands, and the sensory and motor neurons that connect them.¹³ Ocular surface innervation is carried out by afferent fibers of trigeminal nerve and inputs from these nerves form the basis of the reflex arc, which adjusts tear secretion to meet daily demands. Compatible with this mechanism, studies have found that tear secretion was reduced in nasolacrimal obstruction cases and improved after the obstruction was treated.¹⁴ This might be a good example to show that the reflex arc acts loosely to stimulate lacrimal glands for tear secretion in situations where tears do not drain and persist in the ocular surface longer than usual. 

Lacrimal gland size measurements may be performed with different imaging modalities.^8–11 In our study, images from patients who the otorhinolaryngologist had requested paranasal CT imaging from were used. Therefore, we only had the opportunity to perform the study with CT imaging. To date there is only one recent study that measured lacrimal gland volume with CT imaging. In this particular study, Bingham et al.⁹ evaluated a Caucasian population without lacrimal pathology and found a mean lacrimal gland volume of 0.680 cm³ in male and 0.662 cm³ in female orbits. Our control group was similar to theirs and had a mean age of 46 years, but the lacrimal gland volume was 0.580 cm³, which is approximately 12% smaller than their mean female lacrimal gland volume. Bukhari et al.¹⁰ studied lacrimal gland volumes in different ethnicities using MRI and also performed a comparison of CT and MRI in a small sample of the same study group. The mean lacrimal gland volume in females was 0.809 cm³, which is higher compared to our study and the study by Bingham et al.⁹ However, the mean age of their study population was 32 years, and the younger age might be the reason for their larger lacrimal gland volumes, as age and volume have an inverse relationship. They also stated that there is a significant difference between different ethnicities, which might be another explanation for the lower gland volumes in our control group compared to the mentioned studies. 

Both the PANDO and the control group showed an inverse correlation between age and gender in our study with a greater correlation coefficient in the PANDO group. Bingham et al.⁹ also documented a moderate inverse correlation of age and gland volume, whereas Tamboli et al.¹⁵ found a decrease in the length but not the width. Lorber¹⁶ has also declared a decrement in the lacrimal gland mass with increasing age. Atrophy and fibrosis of the gland with increasing age was also demonstrated in autopsy studies.^17,18 

Lorber¹⁶ performed a review of the lacrimal glands gross characteristics, and he stated that lacrimal gland is not a geometric organ with straight contours rather it curves and tapers superiorly and inferiorly, from side to side, and anteroposteriorly. Therefore, it is not wise to make assumptions about the gland volume from simple multiplication of the sizes since it may conclude a falsely higher result. We outlined the gland in all axial images from the superior end to the inferior end. The software automatically produced the 3D image and calculated the volume (see Fig. 4). The 3D image of the gland resembles what Lorber has stated, wider in the cranial side tapering caudally with a curve, which might be a clue for our volumetric measurements accuracy. 

Figure 4

 

Three-dimensional reconstruction of the lacrimal gland.

The mean gland volumes in PANDO and the contralateral non-PANDO sides were similar, and both lacrimal glands were statistically smaller than the control glands. Consistently Schirmer scores were lower in the non-PANDO side compared to the PANDO and the control eyes where statistical significance was present at the level of 0.06 and 0.04, respectively. In fact this finding is similar to what Yen et al.¹⁹ found when they performed a temporary punctual occlusion in normal subjects. They found that both tear secretion and ocular surface sensation was reduced bilaterally in unilateral punctal occlusion. They proposed that reduction of tear secretion might be due to a reduction in ocular surface sensation. They claimed that this finding might be due to the crossed sensory stimulation with decreased trigeminal nerve stimulation in one eye resulting in decreased sensory stimulated tear secretion in both eyes. The documented lacrimal gland involution in our study might be the next level of the decreased crossed sensory stimulation phenomenon. Denervation atrophy is a known fact and has been demonstrated for secretory organs in previous studies.^20,21 It is possible that epiphora and delayed tear clearance in our PANDO patients resulted in decreased sensory stimulation from the ocular surface to the brain in the occluded side and caused bilateral denervational involution of the lacrimal glands. Not only the involution of the glands but the decrement in Schirmer scores especially evident in contralateral non-PANDO side might be an additional supportive finding to the crossed sensory stimulation phenomenon described by Yen et al.¹⁹ The non-PANDO side Schirmer scores might be reconsidered as the baseline tear secretion, and the higher Schirmer score in the PANDO side is due to the obstruction and delayed tear clearance. 

Another possible explanation for smaller lacrimal glands in the PANDO group may lie in the pathophysiology of nasolacrimal obstruction. Inflammation and fibrosis are the main reasons for obstruction of the lacrimal drainage pathway. They may be triggered by different factors like infection, allergy, preservatives in eye drops, nasal pathologies, and menopause.²² PANDO is typically seen in elderly postmenopausal females who are devoid of hormonal support, which is very critical for viability of the lacrimal gland as necrosis and atrophy of the gland in the absence of hormonal support was demonstrated in previous studies.^23,24 Therefore, it is possible that both PANDO and lacrimal gland involution might have developed simultaneously in postmenopausal status without any causation. In this situation, the reduced Schirmer scores might be the result of fibrotic and involuted lacrimal gland that is not related to the crossed sensory stimulation phenomenon but to the fibrosis and atrophy induced by the lack of hormonal support. 

All our PANDO patients were female, so we missed the chance to see if the involution was present also among males. If so, then the involution of the lacrimal gland might better be attributable to the neural stimulus deficiency rather than the hormonal stimulus deficiency. This might be a limitation in our study. The second limitation might be the imaging modality. Magnetic resonance imaging, with improved soft tissue contrast, is generally considered to be superior in evaluation of orbital structures compared to CT.²⁵ However, contrary to general belief, Bukhari et al.¹⁰ did not find a significant difference in terms of lacrimal gland measurements of CT and MRI. 

In conclusion, compared to the control group, the lacrimal gland volume was statistically smaller bilaterally in unilateral PANDO patients, and Schirmer scores were bilaterally reduced with statistical significance in the non-PANDO side. While the entities where the lacrimal gland size tends to be smaller were specified, PANDO might be a cause to remember. This is a preliminary study and further prospective large scaled randomized studies in different ethnicities with a greater study power are needed to reach a definitive conclusion. 

Acknowledgments

Disclosure: A. Yazıcı, None; E. Bulbul, None; H. Yazıcı, None; E. Sari, None; N. Tiskaoglu, None; B. Yanik, None; S. Ermis, None