The Cavernous Body of the Human Efferent Tear Ducts Contributes to Regulation of Tear Outflow

Abstract

purpose. To test the hypothesis that the surrounding vascular plexus of the lacrimal sac and the nasolacrimal duct contributes to the regulation of tear outflow.

methods. Experiments in 30 probands aged between 15 and 37 years were performed in both nasolacrimal systems of each subject by observing with an endoscope the transit time of an applied tear drop containing fluorescein dye until its entry into the inferior meatus of the nose. Four different experiments were performed to determine the median transit time under normal conditions and the influence on transit time of a decongestant drug, a foreign body on the ocular surface, and a decongestant drug applied together with a foreign body on the ocular surface. Comparisons were made between the right and left nasolacrimal system, in males and females, eyeglass wearers and non–eyeglass wearers, and the different experiments and the results statistically analyzed.

results. The tear transit time was independent of side (right or left), gender, or eyeglass wear. It showed great individual variability. Application of a decongestant drug or placement of a foreign body on the ocular surface both prolonged the dye transit time significantly. Application of a decongestant drug simultaneously with placement of a foreign body shortened the dye transit time significantly compared with the effect of the decongestant drug alone but revealed no significant difference compared with application of a foreign body alone.

conclusions. The cavernous body of the lacrimal sac and nasolacrimal duct plays an important role in the physiology of tear outflow regulation. It is subject to autonomic control and is integrated into a complex neuronal reflex feedback mechanism starting with the dense innervation of the cornea. Moreover, its function can be pharmacologically influenced.

Many factors have been hypothesized to be involved in tear outflow, reflecting the unique anatomic configuration of the efferent tear ducts. These include an active lacrimal pump mechanism that functions by contraction of the orbicularis eye muscle; a “wringing out” mechanism governed by a system of helically arranged fibrillar structures; the action of epithelial secretion products; and physical factors, such as capillarity, gravity, respiration, evaporation, and absorption of tear fluid through the lining epithelium of the efferent tear ducts (for review, see Ref. ¹ ). 

As early as 1866, Henle ² described a vascular plexus surrounding the lumen of the lacrimal sac and the nasolacrimal duct. This network of large vessels is connected caudally with the cavernous body of the nasal inferior turbinate. ^{3 4 5 6} Recent morphologic investigations have shown that the vascular system is embedded in the wall of the lacrimal sac and the nasolacrimal duct and fills more than two thirds of the bony canal between the orbit and inferior nasal duct (Fig. 1) . ⁷ Specialized types of blood vessels have been distinguished inside the vascular tissue that surrounds the lumen of the lacrimal passage: barrier arteries, capacitance veins, arteriovenous anastomoses, and throttle veins. ⁷ It has been hypothesized that the surrounding vascular plexus is comparable to a cavernous body. In addition to regulating the blood flow, it is thought the specialized blood vessels permit opening and closing of the lumen of the lacrimal passage, effected by the bulging and subsiding of the cavernous body, and at the same time regulate tear outflow. ⁷ Swelling occurs when the barrier arteries are opened and the throttle veins closed. Filling of the capacitance veins occurs at the same time as closure of the lumen of the lacrimal passage. By contrast, closure of the barrier arteries and opening of the throttle veins reduce the blood stream to the capacitance veins, simultaneously allowing blood outflow from these veins with resultant shrinkage of the cavernous body and dilatation of the lumen of the lacrimal passage. Arteriovenous anastomoses provide for direct blood flow between arteries and venous lacunae. Thus, the subepithelially situated capillary network can be avoided, and rapid filling of capacitance veins is possible when the shunts of the arteriovenous anastomosis are open. ⁷ Moreover, it was demonstrated that the specialized blood vessels of the cavernous body are densely innervated. ⁸ It has been speculated that by means of this innervation the cavernous body of the efferent tear ducts acts to protect the ocular surface against foreign bodies and is also involved in epiphora related to emotional response. ⁸  

The presumable impact of the cavernous body of the efferent tear ducts on tear outflow regulation ^{1 7 8} led us to a detailed analysis of its physiological function in the healthy state and an investigation of potential differences under experimental conditions. 

Methods

Both nasolacrimal systems of 30 healthy volunteers were examined. All subjects had full visual acuity and no recent trauma, eye or nasal infections, or diseases potentially involving or affecting ocular or lacrimal function (including the efferent tear duct system); none wore contact lenses. They comprised 15 women and 15 men, aged 15 to 37 years (mean age, 25 years; Table 1 ) of which eight females and seven male wore eyeglasses (Table 1) for slight myopia or hypertropia. All volunteers were carefully interviewed, to determine a possible history of tearing, lacrimal gland infection, dacryocystitis or dacryostenosis, and symptoms of mucoid discharge in the medial canthal region. After anamnestic exclusion of facial surgery or trauma, allergies, family history of tearing, external eyelid disease, topical eye medication, eyelid malposition, and periocular neoplasm were excluded by a careful ocular examination at the Department of Ophthalmology, Christian Albrecht University (Kiel, Germany). The lacrimal passage was assessed by inspection and palpation, and a nasal endoscopic examination was performed in each subject. Moreover, in all subjects, Schirmer I test results at 1 and 5 minutes were recorded, but in a session separate from the one used for the experiments. None of the volunteers had less than 20 mm of moistened Schirmer strip after 5 minutes, which would have resulted in exclusion from the study. All participants were advised of the nature and purpose of the examination, and informed consent was obtained from each person according to the provisions of the Declaration of Helsinki. All experiments were performed with the permission of the Ethics Committee, Christian Albrecht University of Kiel and a proband liability policy was taken out with DBV-Winterthur. 

The investigations applied to each proband were subdivided into four single experiments for each nasolacrimal system, each of which was performed at a minimum interval of 24 hours, so that each volunteer had eight different experimental sessions (Table 2 ; excluding the session in which Schirmer I test was performed). Additional experiments were performed in some of the subjects either for control purposes or because the volunteer had a common cold when tested. The experiments, as shown in Table 2 , were all performed by one of the authors (MA). 

The basic protocol used for each experiment was a variation of the Jones primary dye test as described by Bosshard ⁹ and Becker. ¹⁰ For experiments 1R and 1L, the subject placed her or his head against the support rest of an examination seat and was asked to look straight ahead and not to move. One drop of fluorescein (standardized amount, 0.025 mL; 0.2 g fluorescein in 10 mL 0.9% NaCl) was placed in the conjunctival sac of the right or left eye without anesthesia, and the subject was then asked to make 10 complete blinks and subsequently to keep his or her eyes closed. Timing began with fluorescein application. Beginning at this point, an endoscope (model BA 600; Storz, Tuttlingen, Germany) equipped with a blue light (Storz) and a fluorescein filter (Storz) as well as 0° 2.7-mm rigid optics (Storz) was used. The objective was placed in the anterior part of the right or left inferior meatus of the nose just below the head of the inferior turbinate, and time was stopped when fluorescence became visible. Insertion of the endoscope into the anterior part of the right or left inferior meatus of the nose was performed without touching the nasal mucosa, to prevent increase of tear secretion through a neuronal reflex arch. 

The protocol was varied for experiments 2R and 2L (Table 2) , in that one drop of xylometazoline (xylometazoline HCl, Otriven; Novartis, Basel, Switzerland; an α-sympathomimetic drug; standardized amount 0.025 mL) was placed into the conjunctival sac 5 minutes before the basic protocol, and the subject was then asked to make 10 complete blinks. The xylometazoline drop was administered with the idea of causing the cavernous body of the lacrimal sac and the nasolacrimal duct to subside and thus to make the lumen of the lacrimal passage more open. 

For experiments 3R and 3L (Table 2) a 3 × 3-mm segment of an electro-oculogram (EOG) strip was placed in the conjunctival sac of the right or left eye without anesthesia for 2 minutes. Afterward, the EOG strip segment was removed from the conjunctival sac and the basic protocol was repeated. The EOG strip segment was supposed to imitate a foreign body at the ocular surface, the idea being to cause the cavernous body of the lacrimal sac and the nasolacrimal duct to bulge and thus to make the lumen of the lacrimal passage more closed. 

For experiments 4R and 4L (Table 2) , xylometazoline was instilled into the conjunctival sac, and the subject was then asked to make 10 complete blinks. After 5 minutes, a 3 × 3-mm segment of EOG strip was placed in the conjunctival sac of the right or left eye without anesthesia for 2 minutes, and then the basic protocol was repeated. Xylometazoline and the EOG strip segment were applied together, the idea being to induce contrary effects in the cavernous body of the lacrimal sac and the nasolacrimal duct (i.e., bulging [EOG strip] and subsiding [xylometazoline]). 

Two additional experiments were performed in six volunteers (three men and three women; age range, 23–31 years): To exclude the possibility that application of any other fluid might have the same effect as xylometazoline on the cavernous body of the lacrimal sac and the nasolacrimal duct, one drop of NaCl (isotone NaCl solution 0.9% diluted in 10 mmol HEPES [pH 7.4]; standardized amount 0.025 mL) was applied instead of xylometazoline and experiments 2R and 2L and 4R and 4L were repeated. These experiments were labeled C1R and C1L, and C2R and C2L (Table 2) . 

Statistical Analysis

Statistical analysis of the data was performed by applying the matched pair signed rank test to compare differences between the right and left nasolacrimal systems, males and females, and eyeglass wearers and nonwearers. The Wilcoxon rank sum test was performed to compare differences between the individual experiments. The level of significance was set at a probability below 0.05. All data are presented as box plots. 

Results

The present data are always expressed as the median transit time for both eyes together. Median transit times for an individual site are given in brackets and also are demonstrated in the figures. The transit time was independent of the side (right or left), gender, or eyeglass wear. 

Effect of Variables on Transit Time

Comparison of results between the right and left efferent tear duct systems (P > 0.05; Fig. 2 ), males and females (P > 0.05; Fig. 3 ), and eyeglass wearers and nonwearers (P > 0.05; Fig. 4 ) showed no significant differences among all four experimental designs. 

Variability of Dye Transit Time

The median transit time of a tear drop containing dye was 177 seconds (right 193 seconds and left 161 seconds; Fig. 2 ). Great individual variability was observed, with minimum values of 15 seconds on the right side and 22 seconds on the left and maximum values of 1103 seconds on the right and 578 seconds on the left (Fig. 2) . 

Effect of Xylometazoline

Application of the decongestant xylometazoline before fluorescein significantly prolonged the median dye transit time from 177 seconds to 375 seconds (right from 193 seconds to 404 seconds and left from 161 seconds to 346 seconds; P < 0.05; Fig. 2 ). The effect of xylometazoline in comparison to experiment 1 is shown in Figure 5 . The control experiment (C1) in six volunteers in which NaCl (0.9%) was applied instead of xylometazoline verified that the effect was caused by xylometazoline (Table 3) . A comparison of fluorescein alone (experiment 1) with application of NaCl (0.9%) 5 minutes before fluorescein showed no significant effect (P > 0.05; Table 3 ). 

Effect of a Foreign Body

Putting a foreign body (a 3 × 3-mm piece of EOG strip) in the conjunctival sac before fluorescein prolonged the dye transit time significantly from 177 seconds to 302 seconds (right from 193 seconds to 295 seconds and left from 161 seconds to 310 seconds; P < 0.05; Fig. 2 ). The effect of a foreign body in comparison to experiment 1 is shown in Figure 6 . 

Effect of Xylometazoline Applied with a Foreign Body

Applying the decongestant xylometazoline and putting a foreign body (a 3 × 3-mm piece of Schirmer strip) in the conjunctival sac without anesthesia, both before instillation of fluorescein, shortened the median dye transit time significantly from 375 seconds to 301 seconds (right from 404 seconds to 326 seconds and left from 346 seconds to 275 seconds; P < 0.05; Fig. 7 ) compared with application of xylometazoline alone (experiment 2) but revealed no significant difference compared with application of a foreign body alone (experiment 3; P > 0.05; Fig. 8 ). The control experiment (C2) in six subjects, in which NaCl (0.9%) was applied instead of xylometazoline but together with a foreign body, showed that xylometazoline had an effect. Comparison between application of a foreign body followed by analysis of the dye transit time (experiment 3) and applying NaCl (0.9%) before introduction of the foreign body followed by analysis of the dye transit time revealed no significant effect (P > 0.05; Table 3 ). 

Discussion

Tear drainage involves several different mechanisms. Beside epithelial secretion products (i.e., mucins and trefoil factor family [TFF] peptides), ^{11 12} a decisive role is played by capillary attraction ^{13 14 15 16 17 18 19} aided by contraction of the lacrimal part of the orbicularis muscle with blinking ^{3 16 20 21 22 23} and distension of the sac, as well as a passive wringing out of the sac because of its medial attachment and helically arranged fibrillar structures. ³ The results of the present study suggest that the cavernous body of the lacrimal passage acts as a further mechanism in tear outflow, is under neuronal control by involvement in a complex neuronal reflex feedback mechanism, and can be influenced pharmacologically. 

The Jones primary dye test ²⁴ is a commonly used test of lacrimal outflow. However, some clinicians find it of limited practical significance because of variable outcome and relatively low sensitivity in documenting normal lacrimal excretory function. ^{25 26} Tucker and Codere ²⁷ showed that the volume of fluorescein used may be an important factor affecting variability in the outcome of the primary dye test. They used a single drop of fluorescein in their experiments, in which the median dye transit time was 8 minutes, compared with 1.4 minutes using multiple drops. 

Our results in experiment 1 are largely in accordance with the findings of Tucker and Codere. ²⁷ The dye transit time was somewhat shorter (4.55 minutes) in our experiment, using a single drop of fluorescein, probably because we, in contrast to Tucker and Codere, applied fluorescein without anesthetizing the ocular surface. Thus, some reflex tearing of the lacrimal gland was initiated that increases lacrimal fluid volume, which in turn shortens the dye transit time, as shown by Tucker and Codere. Further, our results show intraindividual variability similar to the findings of Tucker and Codere, with a standard deviation of 3.23 minutes and minimum and maximum values between 15 seconds and more than 18 minutes in one case. 

There may be several factors that determine the high level of intraindividual variability in dye transit time: fluctuations within a single individual over time, family predisposition, emotional status, the fluid balance, blink rate, basal tear film production, atmospheric conditions of testing, tear pump efficiency, and hormonal status. Murai et al. ²⁸ found that blinking frequency is an important factor. This factor was excluded in our study by having each volunteer make 10 complete blinks and then close the eyes after application of fluorescein. Another factor discussed in this context is the osseous nasolacrimal canal. It is said to be longer and narrower in women than in men and longer in whites than in other subgroups. ^{29 30} However, an impact of the osseous nasolacrimal canal in terms of sex difference appears rather unlikely, as there were no significant differences observed between males and females in our study. Because of ethics committee restrictions, only volunteers between 15 and 40 years of age were analyzed, and thus no statement was possible regarding the influence of more advanced age (>60 years) or changes in hormonal status (postmenopausal women). 

We were surprised to register that application of the decongestant xylometazoline significantly prolonged the dye transit time. Assuming application of xylometazoline before fluorescein (experiment 2) leads to shrinkage of the cavernous body (probably mediated by sympathetic nerves through an α-adrenergic effect of xylometazoline on the specialized barrier arteries of the cavernous body with reduction of blood inflow and with no effect on the diameter of the canaliculi) and thus to a more open lumen of the lacrimal passage, it can be concluded from this experiment that the cavernous body is in a distinctly swollen state under resting conditions, enabling optimal flow of lacrimal fluid through the lacrimal passage (Fig. 9) . It is known from morphologic investigations that the lumens of the lacrimal sac and nasolacrimal duct are not circular but are slitlike under resting conditions, with several mucosal folds in the lacrimal sac (Fig. 1) . ^{1 6 7} Experiment 2 suggests that this slitlike lumen is caused by the swelling of the cavernous body. The epithelia of both sites of the slit are in close contact, and we suggest that this close contact is needed to enable optimal flow of lacrimal fluid under resting conditions. Slight swelling and subsiding of the cavernous body with slight dilation and narrowing of the lumen of the lacrimal passage appears to be a physiological process that regulates the varying flow of tear volumes through the lacrimal sac and nasolacrimal duct. Such differences are seen, for example, during sleep and awake states. In the awake state, there is more tear fluid to be drained than during sleep, when tear fluid production nearly comes to a halt. ³¹ A higher amount of fluid at the ocular surface (for example after instillation of several drops of fluorescein), resulting in subsiding of the cavernous body, enables storage of the fluid in the lacrimal sac which then acts as a reservoir. In this case, the mucosal folds are spread. 

With no knowledge of the cavernous body, Amanat et al. ³² suggested that the lacrimal sac does not empty with each blink but rather that tear fluid accumulates in it until the volume is sufficient to open the valve of Hasner and drain into the nose. However, it now appears more likely that it is rather the cavernous body of the nasolacrimal duct that creates resistance to drainage of lacrimal fluid from the sac instead of the valve of Hasner, as the opening into the nose is very variable, often consisting of nothing but a direct opening. ³³ If the inflow of lacrimal fluid through the canaliculi into the filled lacrimal sac is higher than the outflow capacity through the nasolacrimal duct, epiphora result due to the limited drainage capacity of the lacrimal outflow system. After application of xylometazoline, the cavernous body subsides, and the lumen of the lacrimal sac and nasolacrimal duct are dilated (Fig. 9) . This status is disadvantageous for lacrimal flow and leads to the prolongation of the dye transit time. 

The pathophysiology of functional lacrimal drainage insufficiency (i.e., patients with epiphora despite patent lacrimal passages determined by irrigation) can also be explained by this mechanism: Malfunctions of the cavernous body may lead to disturbances in the tear outflow cycle, ocular congestion, or total occlusion of the efferent tear ducts. Such malfunctions may be caused by acute diseases, such as allergic conjunctivitis, hay fever, rhinitis, or dacryolithiasis. Further, in most patients, persistent epiphora after dacryocystorhinostomy can be explained by destruction of the surrounding cavernous body. ³⁴  

The possibility has been discussed that tear fluid is absorbed by the epithelial lining before it reaches the nose, ^{5 6} and recently it has been shown in an animal model that lipophilic substances are absorbed from tear fluid by the epithelium of the nasolacrimal ducts. ³⁵ In this context, the cavernous body could play a role in drainage of the reabsorbed fluid. Moreover, the present results suggest that contact times between tear fluid and mucosa may be regulated by the swelling of the cavernous body within a distinct range. 

Experiment 3 (insertion of a 3 × 3-mm piece of EOG strip into the conjunctival sac) showed that the cavernous body acts to protect against foreign bodies that have entered the cornea or conjunctiva. When the net outflow of blood from the cavernous body is less than the inflow, the mucosa expands and functionally decreases the tear outflow through the efferent tear duct system (Fig. 9) . Not only is tear fluid production increased by the lacrimal gland, tear outflow is also significantly prolonged or completely interrupted by the swelling of the cavernous body to flush out the foreign body and protect the efferent tear ducts themselves. Moreover, the experiment suggests that the protection system functions on the basis of a complex neuronal reflex feedback mechanism, starting with the dense innervation of the cornea and ending with the innervation of the cavernous body inside the lacrimal tear ducts. Whether a person wears eyeglasses or not seems to have no influence, in this context. 

Besides increased tear secretion from the lacrimal gland and accessory lacrimal glands, epiphora related to emotions such as sorrow or happiness can be explained by the action of the cavernous body. Cholinergic innervation of specialized blood vessels ⁸ in the cavernous body can lead to swelling of the submucosal cavernous tissue with closure of the lacrimal passage. By contrast, adrenergic innervation of the blood vessels ⁸ of the cavernous tissue may trigger a mechanism involving relaxation of the submucosal swelling and improvement of lacrimal tear passage, although this latter mechanism is still a matter of controversy. 

Experiment 4 showed that the complex neuronal reflex feedback mechanism starting with the dense innervation of the cornea and ending with the innervation of the cavernous body inside the lacrimal tear ducts comprises a very important protection reflex. This reflex has a strong effect on the specialized blood vessels of the cavernous body of the efferent tear ducts. Shrinkage of the cavernous body function by xylometazoline seems not to influence the reflex, as there was no significant difference compared with response to application of a foreign body alone. In contrast, the reflex counteracts the effect of xylometazoline significantly compared with application of xylometazoline alone (Fig. 9) . 

It can be concluded that the cavernous body of the lacrimal passage plays an important role in tear outflow. Further investigations are needed to evaluate the function of the cavernous body in different pathologic conditions of the efferent tear ducts, especially in dry eye syndrome. It will be interesting to find out whether decongestants such as xylometazoline will become useful tools in the management of early cases of primary acquired dacryostenosis by counteracting hyperemia and swelling of the cavernous body. It will also be interesting to find out whether wearing contact lenses has an effect on the corneal–cavernous body reflex, as it has been shown that daily contact lens wear is associated with loss of corneal sensitivity. ³⁶  

 Supported by German Research Foundation Grant Pa 738/1-2 and Pa 738/1-4.

 Submitted for publication May 20, 2003; revised June 29, 2003; accepted July 23, 2003.

 Disclosure: M. Ayub, None; A.B. Thale, None; J. Hedderich, None; B.N. Tillmann, None; F.P. Paulsen, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Histologic horizontal section of a human lacrimal sac. The lumen (l) of the lacrimal passage is slitlike and is surrounded by a cavernous body rich in specialized blood vessels (CV, capacitance veins) with wide lumina. e, epithelium; arrows: mucosal folds. Goldner staining; magnification, ×59. (For material and methods used for the figure, see Ref. ⁷ or ³⁵ ).

Table 1.

 

Proband Data

Proband	Age	Sex	Eyeglasses
1	27	F	X
2	24	F	X
3	16	F
4	15	F
5	24	F	X
6	37	F
7	26	F	X
8	24	F
9	25	F
10	22	F	X
11	20	F
12	27	F
13	21	F	X
14	20	F	X
15	22	F	X
16	22	M	X
17	26	M
18	27	M
19	27	M
20	27	M
21	26	M
22	34	M	X
23	34	M	X
24	22	M	X
25	25	M	X
26	23	M	X
27	21	M
28	34	M
29	26	M
30	25	M	X

Table 2.

 

Experiments

Efferent tear duct system
1	R	L	Determination of dye transit time
2	R	L	Determination of the dye transit time after introduction of xylometazoline eye drop into the conjunctival sac
3	R	L	Determination of the dye transit time after introduction of a foreign body into the conjunctival sac
4	R	L	Determination of the dye transit time after introduction of a xylometazoline eye drop and a foreign body into the conjunctival sac
Control experiments
C1	R	L	Determination of the dye transit time after introduction of a drop of NaCl into the conjunctival sac
C2	R	L	Determination of the dye transit time after introduction of a drop of NaCl and a foreign body into the conjunctival sac

Figure 2.

 

Dye transit time in seconds as determined by the four different experiments for the right (R) and left (L) nasolacrimal system (n = 30). Data are presented as box plots. Horizontal lines: median (50th percentile, R1 to 93.0 seconds, L1 to 121.5 seconds, P > 0.05; R2 to 241.5, L2 to 276.0 seconds, P > 0.05; R3 to 205.0 seconds, L3 to 250.5 seconds, P > 0.05; R4 to 226.5 seconds, L4 to 205.0 seconds, P > 0.05); rectangles: 25th and 75th percentiles; bars: the shortest and longest times measured, respectively (i.e., no outliers; ○ outliers at >1.5 × the box length distance from the 75th percentile, and *extreme values at >3 × the box length distance from the 75th percentile).

Figure 3.

 

Dye transit time in seconds as determined by the four different experiments for the right (R) and left (L) nasolacrimal system of females (f, n =15) and males (m, n = 15). Data presented as box plots. Horizontal lines: median (50th percentile, P > 0.05). Remaining symbols are as in Figure 2 .

Figure 4.

 

Dye transit time in seconds as determined by the four different experiments for the right (R) and left (L) nasolacrimal systems of non–eyeglass wearers (ng, n = 15) and eyeglass wearers (g, n = 15). Data are as described in Figure 3 .

Figure 5.

 

Effect of xylometazoline on the dye transit time (n = 20). Data presented as box plots. The median obtained in experiment 1 is set as 0 seconds, and the data are compared with those in experiment 1. Horizontal lines: represent the median (50th percentile; right, 119.0 seconds; left, 163.0 seconds, P < 0.05); rectangles: 25th and 75th percentiles; bars the shortest and longest times measured, respectively (i.e., no outliers; ○ outliers at >1.5 × the box-length distance from the 25th and 75th percentiles).

Table 3.

 

Results of Control Experiments

		All Volunteers			Male Volunteers			Female Volunteers			Eyeglass Wearers			Non–Eyeglass Wearers
				R	L	R	L	R	L	R	L	R	L
1	Median dye transit time	86.5	82.8	111	76	83.3	134	65	109	97.3	69.8
C1	Influence of NaCl (0.9%) on the median dye transit time	65.8	86.2	67.6	82.0	64.0	90.3	70.0	111.5	68.8	73.5
2	Influence of xylometazoline on the median dye transit time	251.3	207.2	231.3	152.3	271.3	262.0	315.5	273.0	220.3	174.3
3	Influence of a foreign body on the median dye transit time	92.0	127.8	79.0	99.7	105.0	156.0	137.0	207.0	69.5	88.3
C2	Influence of NaCl (0.9%) together with a foreign body on the median dye transit time	141.0	185.3	168.0	103.3	114.0	167.3	180.0	334.0	133.2	95.6
4	Influence of xylometazoline and a foreign body on the median dye transit time	116.0	127.0	159.7	79.0	94.0	153.0	195.5	86.5	92.8	130.8

Figure 6.

 

Effect of a foreign body (3 × 3-mm piece of EOG strip) on the dye transit time (n = 24). Data presented as box plots. The median obtained in experiment 1 (177 seconds) is set as 0 seconds, and the data are compared with those in experiment 1. Horizontal lines represent the median (50th percentile; right, 105.0 seconds; left, 101.0 seconds, P < 0.05); rectangles: 25th and 75th percentiles; bars: the shortest and longest times measured.

Figure 7.

 

Effect of xylometazoline and a foreign body (3 × 3-mm piece of Schirmer strip) on the dye transit time (n = 20). Data presented as box plots. The median obtained in experiment 2 (375 seconds) is set as 0 seconds, and the data are compared with experiment 2. Horizontal lines: median (50th percentile; right, −51.0 seconds; left, −72.0 seconds, P < 0.05). The remaining symbols are as described in Figure 5 .

Figure 8.

 

Effect of xylometazoline and a foreign body (3 × 3-mm piece of EOG strip) on the dye transit time (n = 26). Data presented as box plots. The median obtained in experiment 3 (301 seconds) is set as 0 seconds, and the data are compared with those in experiment 3. Horizontal lines: represent the median (50th percentile; right,−34.5 seconds; left −33.0 seconds, P > 0.05); rectangles: 25th and 75th percentiles; bars the shortest and longest times measured, respectively (i.e., no outliers).

Figure 9.

 

Schematic anatomic model of the state of the cavernous body and lacrimal passage in the resting state and under the different experimental conditions (left to right), indicating the specific swelling and compression of the cavernous body and how it permits or restricts tear drainage.

The authors thank Clemens Franke for the schematic illustration in Figure 9 and Michael Beall for editing the English.