Lacrimal Drainage System

Introduction

Epiphora is defined as an abnormal overflow of tears down the cheek. This condition may be congenital or acquired.

Congenital epiphora occurs in 2% to 6% of full-term newborns with most obstructions opening spontaneously in 4 to 6 weeks after birth. A 90% probability exists for spontaneous resolution in the first 12 months of life. The most common cause of epiphora is nasolacrimal duct obstruction secondary to a canal that fails to perforate the nasal mucosa at its distal end, near the vault of the anterior portion of the inferior nasal meatus. Seven other variations of congenital malformations of the lower end of the nasolacrimal duct have been described.¹ For the cases that do not resolve with conservative treatment, probing of the nasolacrimal duct has an overall success rate of approximately 100% when performed in the first 13 months of life, decreasing significantly to 30% when performed after 2 years of age.²

Acquired epiphora may be due to hypersecretion, pseudoepiphora (tear deficiency syndrome), or outflow obstruction. Congenital epiphora should be considered when examining adults with excretory problems, because true congenital nasolacrimal duct obstruction rarely carries over into adulthood. Hypersecretion is most commonly seen in patients with allergic conjunctivitis, meibomianitis, trichiasis, thyroid eye disease, and aberrant regeneration after facial nerve palsy. Pseudoepiphora is most often associated with patients with dry eye syndrome, typically postmenopausal women with or without arthritis and patients who may be taking medications such as beta-blockers or antihistamines. Excretory outflow problems may be caused by obstruction between the punctum, canaliculus, sac, or nasolacrimal duct and its opening into the inferior meatus. Outflow obstruction may be caused by puncti or canalicular obstruction secondary to the use of antiviral medications such as idoxuridine or systemic or topical 5-fluorouracil, and scarring after viral conjunctivitis (herpes simplex, vaccinia) or mucosal diseases, such as Stevens-Johnson syndrome or ocular cicatricial pemphigoid (OCP). Outflow obstruction may also be caused by trauma, such as canalicular lacerations or fractures extending into the lacrimal bone and nasolacrimal duct area, or iatrogenic surgical trauma after cosmetic rhinoplasty or endoscopic nasal sinus surgery for deviated septum, sinusitis, or tumor. Primary acquired nasolacrimal duct obstruction is the most common cause of outflow obstruction with an increased prevalence in women. Women have been found to have a smaller bony diameter at the level of the inferior lacrimal fossa and the middle nasolacrimal duct presenting a quantitative anatomical observation to explain the high prevalence of primary acquired nasolacrimal duct obstruction in women. Other explanations for this situation have included inflammatory and hormonal changes involving estrogen and thyroid levels.³

Anatomy

The lacrimal system has a specific anatomy. The drainage system starts at the puncti. The lower punctum is located 2 mm more lateral than the upper punctum opening into the apex of an avascular space creating a papilla. This leads into the vertical portion of the canaliculus called the ampulla, which measures approximately 2 mm. The horizontal portion of the canaliculus measures approximately 8 mm, converging on the majority of cases into a common canaliculus, which empties into the lacrimal sac. However, these canaliculi may enter the lacrimal sac individually approximately 3 mm to 5 mm from the apex of the sac. The punctal os generally has a diameter of approximately 0.3 mm. This is an important reference, as one chooses different sizes of punctal plugs when treating patient with dry eyes. Nonetheless, the canaliculus may reach a size of 1 mm to 2 mm in diameter. The lacrimal sac is usually 15 mm in height, of which 3 mm to 5 mm extend over the medial canthal tendon. The intraosseous portion of the nasolacrimal duct is approximately 12 mm emptying at the inferior meatus. In children 1 year of age or younger, the distance from the punctum of either canaliculus to the nasolacrimal duct is approximately 12 mm, whereas to the floor of the nose, the distance is roughly 20 mm.

With the advent of endoscopic lacrimal system examination and surgery, it is imperative to understand the intranasal anatomy. This also enables the ophthalmologist to better understand the causes of epiphora and the significance of the diagnostic tests. The lateral nasal wall has three turbinates projecting into the nasal cavity: the inferior, middle, and superior. At times, a fourth, supreme turbinate, is present. The ostium of the nasolacrimal duct is located underneath the inferior turbinate with its opening emptying approximately 15 mm above the floor of the nose. This opening is located 4 mm to 6 mm behind the anterior portion of the inferior turbinate. The anterior portion of the middle turbinate overlies the posterior portion of the lacrimal sac fossa. At times, the most anterior air cells of the ethmoid sinus, the agger nasi cells, have the lacrimal bone as their lateral wall. The ethmoidal bulla is the largest and most constant air cell of the anterior ethmoid. When present at this location, these cells must be recognized and removed when performing an external dacryocystorhinostomy.

The osteomeatal complex is compromised by the uncinate process, the hiatus semilunaris (two-dimensional space), and the infundibulum (three-dimensional space). Beneath the central segment of the middle turbinate, the uncinate process is seen as a thin leaflet running from anterosuperior to posteroinferior. This forms the immediate posterior communication of the agger nasi cells. All of the paranasal sinus ostiums drain into the ethmoidal infundibulum except the posterior ethmoid air cells, which open into the superior meatus, and the sphenoid sinus ostium, which opens into the sphenoethmoidal recess. Drainage and ventilation of the sinuses are to a large extent dependent upon the patency and function of the osteomeatal complex. Patients with sinusitis problems frequently have nasolacrimal duct obstruction. This frequently occurs in children with intractable nasolacrimal duct obstructions.

Clinical Findings

Effective tear drainage depends on many factors, including the volume of tear secretion, eyelid position, and the lacrimal drainage and intranasal anatomy. Symptomatic tearing may result from a normal lacrimal drainage system overwhelmed by hypersecretion or from a drainage system that is anatomically compromised and unable to handle normal tear production. On the other hand, a patient with partial nasolacrimal duct obstruction may have a concomitant reduction in tear production and be completely asymptomatic, or even suffer from dry eye syndrome. Therefore, while epiphora is defined as an abnormal overflow of tears down the cheek, it is determined by the balance between tear production and tear drainage, and not by the absolute function or dysfunction of either one.

In evaluating patients with epiphora, it is important to obtain a good history and to carefully observe the patient. Patients may lead you to believe they have epiphora by stating that they are constantly wiping their eyes with a tissue. Nonetheless, when you observe them, you never see gross epiphora, and in instances there is a slightly positive tear film at best. These patients have tear deficiency syndrome with reflex tearing when their corneas become dry and irritated; rarely, do patients with dry eyes have gross epiphora.

Acquired lacrimal duct obstruction may manifest, as mentioned earlier, with epiphora. Nonetheless, it may also present as mucoid discharge from the punctum or an infectious dacryocystitis. A combination of all three forms may occur.

Of note is the complaint or observation of serosanguinous tears, which may be indicative of a lacrimal sac tumor.

Diagnostic Tests

A complete eye examination is essential, with emphasis on the lacrimal sac and punti, eyelid, conjunctiva, and cornea status. A finger palpation of the lacrimal fossa for an enlarged lacrimal sac is essential. Mucoid or mucopurulent reflux on gentle pressure of the lacrimal sac establishes the diagnosis of dacryocystitis. A dilated pouting punctum, with or without some erythema and induration, along with the appearance of creamy pus when the pressure is applied over the canaliculus, may lead to a diagnosis of canaliculitis. This is usually caused by the atypical mycobacteria, Actinomyces israelii. A gritty sensation on gentle probing and expression of yellowish concretions confirm this diagnosis.

The eyelids should be evaluated for blepharitis, meibomianitis, and trichiasis. Lid tone is assessed by the "snap back test," to rule out lacrimal pump dysfunction. The conjunctiva should be evaluated, foremost, on its superior tarsal surface, to identify papillary or giant papillary reaction pointing to an allergy or to an inciting irritative factor that may be producing a hypersecretion state. The conjunction should be evaluated for blunting or symblepharon formation, which may occur with cicatricial pemphigoid or medications such as pilocarpine, epinephrine, timolol, idoxuridine, echothiophate iodide, and demecarium bromide.⁶

Tear film breakup time is an important test for the evaluation of tear film stability, which depends on basal mucous layer secretion from goblet cells. A drop of fluorescein is placed in the eye and the patient is asked to blink once. The tear film on the cornea is observed with a cobalt blue light until the film breaks up, exposing the dry spots on the cornea. Normal breakup time is 15 to 30 seconds; therefore, a tear film breakup time of less than 10 seconds is abnormal. The presence or absence of a tear meniscus, or its height, is a useful finding in ascertaining dry or wet eyes.

Rose bengal is an iodinated fluorescein dye that stains devitalized epithelial cells. Increased staining of the conjunctiva and the corneal epithelium is a sensitive indicator of inadequate tear function.

To further assess patients with tear deficiency syndrome, the Schirmer test offers a means to specifically quantitate the amount of tears that are produced in the eye. Information can be gathered by a Schirmer test with anesthesia, which specifically targets the basal tear secretion produced by the lacrimal glands of Wolfring and Krause. They are most responsible for the constant and adequate lubrication of the corneal surface. A normal Schirmer test with anesthesia should wet the filter paper 10 mm or more in 5 minutes. Less than 10 mm is considered abnormally low, consistent with a deficiency in tear production. It is important to allow the anesthetic to be fully absorbed in the eye. Any excess anesthetic should be blotted from the cul-de-sac with a tissue, so it is not absorbed by the filter strip, leading to an erroneous interpretation. The Schirmer test may be highly variable, and at times must be repeated to verify the initial results. For the most part, this test should be used with other clinical findings to arrive at a final conclusive diagnosis, rather than as an isolated absolute test.

As one concentrates on the outflow system, two tests are most practical and useful - the dye disappearance test and the lacrimal irrigation test. The former is a functional test while the latter is a structural test. The dye disappearance test is performed by placing 1 drop of 2% fluorescein into the lower fornix of each eye. Asymmetrical clearance of the dye in one eye points to a probable outflow obstruction on that side. A delay in both eyes after 5 minutes may point to a bilateral obstruction. Although usually unnecessary, this test is especially useful in children where the diagnosis of lacrimal obstruction is questionable and an in office irrigation test is usually impossible. Children with congenital lacrimal duct obstruction present in the majority of cases with visual clues that support their diagnosis such as an increased tear lake, eye secretions, crusting and matting of the eyelashes, and erythema of the inferior palpebral skin secondary to contact dermatitis from the tear components.

Lacrimal irrigation is performed by dilating the punctum after placing topical 4% Xylocaine (lidocaine HCl, Astra Zeneca) on the punctum for 1 minute. A 23-gauge blunt lacrimal needle is placed within the canaliculus, which is then irrigated with saline. There should be little or no reflux from the upper system when the lower system is gently irrigated. Forceful irrigation overwhelms the system and a significant amount of reflux may occur, which is misleading. If there is significant reflux from the lower to the upper system on gentle irrigation, along with saline into the throat, then there is most likely a partial obstruction in the system. The potential obstruction sites, secondary to reflux, may be at the common canaliculus or distal to the sac. A gentle probing with a #0 or #00 Bowman probe reaching the lacrimal bone will rule out a common canalicular obstruction. For patients who have no canalicular obstruction and do not obtain saline in the nose upon irrigation, the examiner should block the upper canaliculus by pressing it with a cotton swab against the supraorbital bone while irrigating the inferior canaliculus. If fluid is obtained in the nose, a severe partial obstruction is present. If no fluid is able to pass into the nose and a severe back pressure is encountered, then a complete nasolacrimal obstruction is creating the problem.

Finally, if there is no reflux and easy irrigation in the throat, a functional block may be diagnosed, specifically if the lid tone is poor (assessed by pulling the eyelid away from the globe and seeing how quickly it snaps back against the globe).

When there is difficulty differentiating a functional blockage from eyelid or other etiological problems, a dacryoscintigraphy can be performed. This test uses gamma-emitting radionuclides such as technetium 99 to evaluate the physiologic flow or tears. The test is performed by instilling 1 drop of technetium pertechnetate (Te 99) in saline or technetium sulfur colloid in the conjunctival cul-de-sac, followed by imaging of the lacrimal system with a gammagram. This test works best for those patients who show a contradictory or inconsistent result with lacrimal drainage system irrigation, but have a strong and symptomatic epiphora. This is a test that is easily performed, providing less radiation exposure to the lens when compared to a contrast dacryocystography, which works better in providing the fine anatomic details of the lacrimal excretory apparatus. When performing a dacryocystography, a radiopaque dye is injected into the canalicular system on one or both sides simultaneously followed by immediate imaging. Films can also be taken at 10 minutes to show any delay in drainage. This test provides information on the level of blockage by better defining the anatomy or extent of a lacrimal sac malformation in congenital cases. When this test is combined with computerized digital substraction, a better imaging of the system is obtained. Radiographic and nuclear analysis of the lacrimal system is a good resource, but seldom must it be utilized. There is a sound application in planning surgery to excise a "lacrimal anlage duct." This is a congenital fistula of the lacrimal sac resulting from a canalization of a strand of epithelial cords that extends from the sac to the skin surface. The opening is located inferior and medial to the lower punctum.

Although the Jones test was first described in 1961 and is always mentioned in the evaluation of the lacrimal system, it will not be mentioned in this tutorial; it is not very useful and practical in an office or resident clinical setting. The tests mentioned above plus a history and physical have been satisfactory in making the correct diagnosis.

Finally, a nasal exam for rhinitis, polyps, or tumors should be performed. It is at times necessary to order a sinus screen with plain X rays (Waters view) or paranasal sinus computed tomography scan to rule out sinusitis or eroding lesions that may be indicative of carcinoma of the sinuses.

Treatment

Because of the high rate of spontaneous resolution, congenital lacrimal duct obstruction is treated conservatively during the first months of life. This includes the chronic use of topical antibiotics to suppress the ongoing mucous discharge and infection. This regimen should, nonetheless, be kept to a minimum until the lacrimal duct spontaneously opens or probing occurs. This may be complemented with the Criggler massage as described by Kushner which some advocate as beneficial. The ideal age to perform a probing is controversial. In practical terms, I advise parents to perform probings as early as 6 months of age, if the nasolacrimal duct problem persists at this time. I involve parents in the decision making process by providing them with a clear explanation about the natural course of the disease; the details, benefits, and risks of the procedure; and anesthesia. Parents observe their child daily, and can best assess how much this condition is creating a problem, and how fast they want to solve it. Some parents may elect to have their child undergo a lacrimal duct probing earlier than at 6 months of age. On the other hand, some ophthalmologists elect to perform in-office probings, especially at an early age.⁷ This is a somewhat risky maneuver, as the surgeon may face an oculocardiac response that may prompt an arrhythmic situation in an uncontrolled medical environment. The clinician must also consider that this procedure may cause significant pain in a child. In this uncontrolled circumstance, selecting an appropriate probing method may also be difficult.

Probings are best performed under general anesthesia with or without endotracheal intubation. A Rudemann dilator is preferred to dilate the punctum and canaliculus by first directing it straight vertically and then horizontally toward the medial canthus while tractioning on the eyelid laterally with the opposite hand. This is repeated with an #0 or #00 Bowman probe which is advanced until a hard stop is encountered. At times, a larger #1 Bowman probe may be required. At this time, the probe is then rotated superiorly up to the supraorbital notch at the superior orbital rim. It is then directed posteriorly and laterally advanced down the nasolacrimal duct. To identify a successful procedure, the probe can be verified in the nose by obtaining a metal on metal touch with a larger Bowman placed underneath the inferior turbinate. This obviates the need for irrigation of saline mixed with fluorescein, which adds to the preparation of the procedure and may be successful, even when the probe has not passed out the nasolacrimal duct end into the nose. This is the reason why pediatric and general ophthalmologists, as well as oculoplastic surgeons, should be well familiarized with the intranasal anatomy.

Some surgeons also accompany the probing procedure with an infracture of the inferior turbinate.⁸ A true impacted inferior turbinate rarely occurs, and infracturing it may potentially create scarring which may aggravate the lacrimal duct drainage problem. Lacrimal duct probing has an extremely high success rate to warrant this additional step.

For patients whose nasolacrimal duct obstruction failed to resolve with probing or patients who are older than 18 months of age when the success rate of probing drops dramatically, silicone intubation is indicated. Different types of techniques and intubation sets have been described as are the Crawford (Jedmed, Inc., St. Louis, MO) and, most recently, Ritleng (FCI Ophthalmics, Marshfield Hills, Mass.) bicanalicular silicone tubes. The Ritleng is a silicone tube attached to a prolene monofilament suture that facilitates its retrieval from the nose with less mucosal trauma. It requires a reusable Ritleng metal probe where the prolene is introduced.¹⁰ Taking cost effectiveness into account, the Crawford silicone tubes tend to work well. The tubes have a double metal end with an olive tip, which helps in retrieving both ends out of the nose with a Crawford hook. Some keys to success are appropriate lighting with a fiberoptic headlight, vasoconstriction of the nasal mucosa, and patience. Again, it should be known that the metal end of the tube will exit underneath the inferior turbinate 4 mm to 6 mm behind its anterior end. Once both ends have exited the nose, the tubes are ligated with a 6-0 silk suture using a single square knot that allows its removal through the canalicular system. Nonetheless, this may provoke a child to pull on them jeopardizing the success of the procedure. In this regard, I have elected to ligate both tubes together with a 6-0 prolene suture on a P1 needle and then suture them to the anterior inner portion of the lateral nasal wall.

The LacriCath (Atrion Medical Products, Inc., Birmingham, Ala.) pediatric balloon catheter dacryoplasty has been introduced in more recent years as another alternative to treat patients who have failed probings or silicone intubations. The LacriCath is a catheter with a collapsed balloon covering its distal end. After successfully probing the system with a #00 Bowman probe, the balloon catheter is introduced in a similar manner confirming its distal position intranasally. Using the reference markings, the balloon is inflated with sterile saline to 8 atm of pressure holding it for 90 seconds and 60 seconds at the distal lacrimal duct and lacrimal sac/proximal duct area. The LacriCath is seldom required for the resolution of congenital lacrimal duct obstructions. A major drawback is its cost. The LacriCath may be of use when all other alternatives have failed. This procedure may or may not be accompanied by silicone tube placement.

The balloon catheter dacryoplasty has been proposed with nominal variations in the technique as an alternative for the treatment of adult patients with partial nasolacrimal duct obstruction to avoid incisional surgery, with a high success rate.

For decades, the mainstay treatment for nasolacrimal duct stenosis or obstruction is the dacryocystorhinostomy (DCR). This treatment can be performed externally or endoscopically. The external DCR has a success rate of nearly 90%, whereas the endoscopic DCR can achieve similar results when performed by experienced surgeons. The indication for DCRs is nasolacrimal duct stenosis or obstruction which may manifest itself as epiphora, mucoid discharge, or an infection. The procedure consists of the creation of a direct passage from the lacrimal sac into the nose through an osteotomy on the lateral wall of the nose just in front of the middle turbinate. An incision is made as horizontal as possible at the nasojugal fold dissecting the orbicularis oculi muscle up to the anterior lacrimal crest. The periosteum is incised and then elevated laterally along with the lacrimal sac using a Freer instrument. The lacrimal bone is pushed and used as the opening to start an osteotomy anteriorly. The success of a DCR depends on achieving an osteotomy size between 10 mm and 15 mm in diameter. When opening the lacrimal sac to communicate it directly into the nose, the common canaliculus opening should be clear of any surrounding tissue for at least 5 mm circumferentially. The lacrimal sac is communicated into the nose by incising the nasal mucosa and uniting the anterior leaflets of each structure with double arm 5-0 vicryl sutures after eliminating the posterior leaflets. The majority of surgeons prefer to place silicone tubes in the system which can be removed after 6 weeks. It is important to remind patients to discontinue aspirin or nonsteroidal anti-inflammatory medications 2 weeks before surgery, as this may increase the chances of a bleeding complication. External DCR is best performed with local anesthesia including an anterior ethmoidal and supratrochlear nerve block along with local infiltration of anesthesia into the incision area. Nasal mucosa vasoconstriction is achieved by placing a 0.5 inch by 3 inch piece of cotton soaked with oxymetazoline hydrochloride intranasally under direct visualization.

Mitomycin C, an antifibroproliferative alkylating agent, may play a role in certain cases of failed external or endonasal DCRs. However, due to the high success rate of the procedure, mitomycin C is rarely indicated for routine use. Mitomycin C may provide some advantages when used primarily in pediatric DCRs because fibrosis and scarring tend to occur more often in pediatric cases. Mitomycin C can be administered by applying a soaked piece of cotton to the osteotomy site.⁹ Different dilutions and application times have been proposed.⁹

Endoscopic DCR can be performed as laser assisted or non-laser assisted. Non-laser assisted DCR is best performed on patients with a clear view of the anterior portion of the middle turbinate and enough room to place the instruments. After appropriate hemostasis, a 0° to 30° endoscope is placed in the nose and the middle turbinate is infractured with the blunt end of a Freer elevator. The uncinate process is the first protrusion encountered underneath the middle turbinate which is also infractured with the elevator to incise the nasal mucosa just anterior to it. The uncinate process is completed with the Blakesley forceps exposing the lacrimal bone. Using a guarded monopolar cautery, the nasal mucosa is cautherized for 10 mm and then scraped inferiorly exposing the lacrimal bone. An osteotomy is then created with a 90° rongeur starting at the posterior edge. The lacrimal sac is opened by fashioning an inferiorly based U-shaped flap, and silicone tubes are placed.

A benefit of the endoscopic DCR is that it provides direct access to the rhinostomy site and obviates an external scar. A disadvantage is that it does not allow direct and clear visualization of the sac which may make the surgeon miss the visualization of tumors or dacryoliths. It also has a steep learning curve to master the endoscopic maneuvers and requires expensive equipment. Proponents of external DCRs state that when properly placed, external incisions leave a negligible scar.

For patients with upper system abnormalities such as punctal stenosis, one-, two-, or three-snip punctoplasty may be performed. Conversely, patients with dry eye syndrome who require tear drainage blockage may benefit from punctal occlusion either using plastic punctal plugs or permanent thermal cautherization. The latter is easy to perform but may require more than one session as surprisingly ducts tend to reopen after a period of 2 weeks. In these patients, conservative treatment should be instituted first, using preservative-free artificial tears and at times a topical nonsteroidal anti-inflammatory drug or mast cell stabilizer. In cases of "crocodile tears" secondary to facial paralysis, involuntary lacrimation has been controlled with injection of botulinum toxin directly into the lacrimal gland. Canaliculitis requires a canaliculostomy with evacuation of the yellowish concretions by compressing the canaliculus. The canaliculostomy should be performed by first dilating it with a Wilder lacrimal dilator (Bausch & Lomb Surgical E4360), and then introducing the inferior blade of sharp scissors (Bausch and Lomb Surgical, Storz Curved Stitch scissors E3592) through the ampulla across the horizontal portion of the canalicus and unroofing the passage. The canaliculus is then irrigated with either penicillin G solution (100,000 units/mL), cefazolin (100 mg/mL), or iodine 1% solution. If the patient tilts the head forward 30°, the liquid will slide forward down the nose instead of running down the throat. This can be followed with the use of topical antibiotics for 1 to 2 weeks. There is no need for stitching as the passage remains patent and unroofed without creating a functional or aesthetic problem. Recurrent cases may also require silicone intubation. Acute dacryocystitis is commonly caused by B-hemolytic Streptococci or Staphylococcus. Treatment consists of oral and topical antibiotics. If a localized abscess develops, incision and drainage are required. The incised compartment is packed opened with a Nugauze strip embedded in antibiotic and allowed to heal secondarily. DCR is the final treatment, preferably performed when the infection subsides. However, DCR can be performed during the acute process.

References

1. Jones LT, Wobig JL. Surgery of the eyelids and lacrimal system. Birmingham, Ala: Aesculapius Publishing Co; 1976.
2. Katowitz JA, Welsh MG. Timing of initial probing and irrigation in congenital nasolacrimal duct obstruction. Ophthalmology. 1987;94:698.
3. Groessl SA, Sires BS, Lemke BN. An anatomical basis for primary acquired nasolacrimal duct obstruction. Arch Ophthalmol. 1997;115:71.
4. Zide BN, Jelks GW. Surgical Anatomy of the Orbit. New York, NY: Raven Press; 1985.
5. Anderson JE. Grant's Atlas of Anatomy, 8th ed. Baltimore, MD: Williams and Wilkins; 1983.
6. Nesi FA, Lisman RD, Levine MR, Brazzo BG, Gladstone GJ. Smith's Ophthalmic Plastic and Reconstructive Surgery. 2nd ed. St Louis, MO: Mosby-Year Book, Inc; 1998.
7. Stager D, Baker JD, Frey T. Office probing of congenital nasolacrimal duct obstruction. Ophthalmic Surg. 1992;23:482.
8. Wesley RE. Inferior turbinate fracture in the treatment of congenital nasolacrimal duct obstruction and congenital nasolacrimal duct anomaly. Ophthalmic Surg. 1985;16:368.
9. Kao SCS, Liao CL, Tseng JHS, Chen MS, Hou PK. Dacryocystorhinostomy with intraoperative mitomycin C. Ophthalmology. 1997;104:86.
10. Pe MRL, Langford JD, Lindberg JV, Schwart TL, Sondhi N. Ritleng intubation system for the treatment of congenital nasolacrimal duct obstruction. Arch Ophthalmol. 1998;116:387.
11. Otis PT. Medical management of congenital nasolacrimal duct obstruction. J Pediatric Ophthalmol Strabismus. 1985;22:68.