Lacrimal System Anatomy and Physiology

The precorneal tear film lubricates the eye, delivers nutrients to the underlying cornea, and provides an optimal refractive surface. The tear film is composed of an inner mucoprotein layer covering the corneal epithelium, an intermediate aqueous layer, and a superficial lipid layer that stabilizes the tear film and prevents evaporation of underlying layers. Baseline tear production involves goblet cells, meibomian glands, and subconjunctival accessory glands.3 Reflexive tearing is primarily generated by the lacrimal gland located within the superolateral orbit.4 In instances such as crying or an ocular foreign body, the lacrimal gland generates tearing by augmenting the intermediate aqueous layer of the tear film.

The lacrimal drainage system transports tears from the ocular surface into the nasal cavity. The lacrimal puncta can be found 5 to 7 mm lateral to the medial canthus, with the lower punctum slightly more lateral than the upper punctum.3 The punctal os generally sits at the apex of a small mound known as the papilla. Tears enter the punctum, then travel through the canalicular system. The canaliculus has a short vertical section, measuring ~2 mm, followed by a longer horizontal section.

The horizontal components of the upper and lower canaliculus join medially to form the common canaliculus. The common canaliculus enters the lacrimal sac roughly 3 to 5 mm inferior to its apex.4 The lacrimal sac measures 15 mm in height and is situated between the bony anterior and posterior lacrimal crests. Tears flow from the sac into the intraosseous nasolacrimal duct, which travels though the maxilla and finally exits into the nose via the Hasner valve at the inferior meatus.

The drainage of tears is an active process driven primarily by contraction and relaxation of the orbicularis oculi muscle.4 Movement of the eyelids distributes the tear film evenly over the surface of the eye. When the eye is open, tears collect in lakes adjacent to the lacrimal puncta. Eyelid closure milks tears from medial to lateral toward the puncta, shortens the canaliculi, and creates negative pressure within the lacrimal sac. Subsequent eyelid opening allows tears to be drawn into the sac and propelled through the nasolacrimal duct toward the nasal cavity. Although there are controversies regarding the lacrimal pump, it is clear that a functional lacrimal pump mechanism depends on normal lid-globe apposition, active muscle movement, and a patent lacrimal drainage system.

Endoscopic nasolacrimal anatomy relies on two important landmarks. The first landmark is the maxillary line ( Fig. 37.1 ).5 The maxillary line appears as a curved eminence along the lateral nasal wall at roughly the level of the anterior head of the middle turbinate. The line extends from the superior attachment of the middle turbinate (axilla) to the root of the inferior turbinate. Anatomically, this line represents the vertical suture between the frontal process of the maxilla anteriorly and the lacrimal bone posteriorly. The uncinate process inserts onto the lateral nasal wall just posterior to the maxillary line. The lacrimal sac is located deep (lateral) to the upper aspect of the maxillary line with approximately two-thirds of the sac anterior and one-third of the sac posterior to the line. Inferiorly, the nasolacrimal duct is generally located just anterior to the line. The second important landmark is the axilla of the middle turbinate— the anterosuperior insertion site of the middle turbinate onto the lateral nasal wall. Anatomical studies have shown that the apex of the lacrimal sac extends an average of 8 mm above the axilla of the middle turbinate.6

Patient Selection

Epiphora, or excess tearing, occurs when there is an imbalance between tear production and drainage. The differential diagnosis for epiphora is broad and includes lid malposition, infectious/inflammatory blepharitis, and foreign body reaction ( Table 37.1 ).7 DCR is primarily indicated in the management of lacrimal system obstruction at or distal to the lacrimal sac. Nasolacrimal duct obstruction is most commonly idiopathic, but it may also be seen secondary to recurrent dacryocystitis, facial trauma, or complications of sinonasal surgery.

Examination of the lacrimal system begins with inspection of the lacrimal puncta and medial canthal region. In instances of chronic infection and stasis, palpation of the medial canthus may result in reflux of mucupurulent discharge from the punctal os. Patency of the lacrimal system is most commonly assessed by probing and syringing.8 To carry out this diagnostic procedure, the inferior punctum is dilated, and a Bowman lacrimal probe is inserted into the canaliculus vertically and then horizontally toward the medial canthus. A hard stop is noted when the probe abuts the sac and adjacent medial bony wall. If obstruction is within the canalicular system, a soft stop will be noted. Next, saline is gently instilled with a syringe. A hard stop followed by reflux of saline via the opposite punctum indicates obstruction of the lacrimal sac or duct. Flow of saline into the nasal cavity indicates a patent lacrimal system, although subtle narrowing or functional obstruction is not fully ruled out.

When the clinical picture remains unclear, diagnostic imaging may be done. A dacryocystogram involves injection of contrast into the lacrimal system followed by either plain films or computed tomography (CT) scans. Magnetic resonance imaging (MRI) may also be done in select cases, especially if lacrimal neoplasm is suspected. A lacrimal scintigram involves placing of a radioisotope in the conjunctival fornix and following its progress through the lacrimal system over time ( Fig. 37.2 ). This modality is particularly helpful if the lacrimal system is patent on syringing, but a functional obstruction is suspected.8

Endoscopic Dacryocystorhinostomy Technique

Endoscopic DCR (see Video 58, Endoscopic Dacryocystorhinostomy ) offers several advantages over external DCR, including avoidance of external incisions, superior intranasal visualization, and the ability to address any concurrent sinonasal pathology, such as septal deviation or chronic sinusitis.9 Endoscopic DCR is generally contraindicated in instances of known malignancy of the lacrimal apparatus, with relative contraindications including large lateral sac diverticuli, common canalicular stenosis, or retrieval of large lacrimal stones.7

Endoscopic DCR may be performed under either local or general anesthesia, although general anesthesia is preferred by most patients and surgeons. The nasal cavity is decongested with 0.05% oxymetazoline spray as well as pledgets soaked with 4% cocaine solution or a similar vasoconstrictive agent. A 4-mm 0- or 30-degree scope can be used throughout the case, depending on anatomy and surgeon preference. Injections of 1% lidocaine with 1:100,000 epinephrine are performed along the lateral nasal wall, as well as adjacent to the axilla of the middle turbinate. The procedure begins by inspecting the nasal cavity to identify the presence of septal deviation or signs of chronic sinusitis that may need to be addressed by ancillary procedures. The location of critical structures, including the maxillary line and the axilla of the middle turbinate, should be positively identified.

Surgical dissection is begun by removing the uncinate process located just posterior to the maxillary line. Although this step is omitted by some surgeons, it optimizes exposure of the sac and can be accomplished quickly and safely. Next, an incision is made with a sickle knife in the lateral nasal wall mucosa overlying the sac ( Fig. 37.3 ). This incision is generally brought above the axilla of the middle turbinate and anterior to the maxillary line so as to fully expose the sac. If unsure of anatomy, sac localization can be facilitated through transillumination with a 20-gauge fiberoptic endoilluminator or with the aid of a surgical navigation system ( Fig. 37.4 ). The mucosa is widely elevated from the underlying bone and removed with through-cutting forceps.

To expose the lacrimal sac, the bony lacrimal fossa must be uncovered. Generally, the lacrimal bone over-lying the posterior sac is soft and can be removed with spoon curettes. The frontal process of the maxilla over-lying the anterior sac is much denser and generally requires bone rongeurs or powered instrumentation to remove. We prefer a high-speed drill with a cutting bur, although use of a diamond bur may decrease the likelihood of tearing of the underlyIng sac wall ( Fig. 37.5 ). Care should be taken to avoid traumatizing the adjacent middle turbinate mucosa while drilling, as this can create raw surfaces that might later form synechiae. Adequate exposure of the sac is confirmed by ballotting of the medial canthus or movement of a lacrimal probe that has been passed by the assistant through the canaliculus into the sac. A well-placed, generous bony rhinostomy of at least 8 to 12 mm facilitates a successful outcome.

After removal of the overlying bone, the wall of the lacrimal sac is incised with a sickle knife ( Fig. 37.6 ). It is often helpful for the assistant to “tent out” the medial wall of the sac with the previously inserted lacrimal probe(s).

The medial wall of the sac is removed with through-cutting instruments and submitted for histopathologic examination ( Fig. 37.7 ). After removal of the medial sac wall, a bicanalicular Silastic stent is placed ( Fig. 37.8 ). This is accomplished by intubating both canaliculi, with subsequent retrieval of the probes from the rhinostomy site using Blakesley forceps. The tubing is then tied at the nasal vestibule to create a closed-loop stent. Care is taken to ensure that the stent is loose enough at the medial canthus so as not to cheese-wire the puncta, but not so loose that the stent retracts out excessively. The use of intranasal packing material to prevent bleeding is left to the discretion of the surgeon.

Patients are generally discharged the same day with either oral or topical antistaphylococcal antibiotics for 1 week. Saline irrigations are encouraged twice daily with follow-up in 1 week for nasal endoscopy and débridement. Postoperative endoscopic inspection of the lateral nasal wall will show the stent in place. Movement of the stent with blinking often occurs and is considered a positive prognostic sign. Stents are usually removed in 6 to 12 weeks, although there is little evidence to support this practice ( Fig. 37.9 ).

Revision Endoscopic Dacryocystorhinostomy Technique

Failure of primary DCR with return of epiphora occurs in ~5 to 15% of cases.2 Placement of fluorescein dye in the conjunctival fornix followed by endoscopic visualization serves as an excellent test of ostial patency. Probing and syringing may also be done as needed. Revision endoscopic DCR is favored over external techniques because it allows careful endonasal inspection for factors that may have led to failure of the initial procedure.10 This often involves synechiae formation between the lateral nasal wall and middle turbinate or septum. However, in many instances fibrosis of the rhinostomy site is found with no evident cause. After the nasal cavity is decongested and local anesthesia injected, a probe is inserted and used to explore the rhinostomy site. Probing can determine the precise location of the lacrimal sac, as well as the extent of prior bone removal. Fibrous tissue at the site of the prior sac opening is tented into the nasal cavity and then removed sharply. Residual bone is removed with a high-speed drill to create a generous rhinostomy site. In some instances, extensive scar tissue is present, and care must be taken not to injure the medial sac wall, common canaliculus, or intraorbital structures as can occur with aggressive avulsion of soft tissues.

In instances where an obvious cause is not evident, one may choose to use mitomycin C to decrease fibrosis at a concentration of 0.4 mg/mL for 4 minutes, followed by generous saline irrigation.11 Silicone stents are placed, and the remainder of postoperative care is similar to that in primary procedures.

Outcomes

Endoscopic DCR was initially met with some skepticism, as early reports described success rates inferior to that of external DCR; however, over the last decade, reported success rates for nonlaser endoscopic DCR have consistently been equivalent to external approaches.12–15 This improvement likely represents the natural learning curve that has occurred with endoscopic sinonasal surgery in general. Due to the heterogeneity by which success is defined in the literature, it has not been possible to group studies for meta-analyses or to do direct comparisons between external and endoscopic approaches. A recent systematic review of 73 studies encompassing 4921 DCR procedures noted a success rate between 65 and 100% after external DCR and 84 and 94% after endoscopic DCR.2 This literature is primarily level 3 and 4 rated evidence, as detailed by the Oxford Center for Evidence-based Medicine.16 Outcomes of revision DCR are generally considered inferior to primary DCR but still quite good. Success rates of revision endoscopic DCR are reported between 69 and 100% in the literature, with case series much smaller than primary DCR.2 One can reasonably conclude that endoscopic DCR in experienced hands is safe, with an efficacy equal to that of external procedures.

Complications

Intraoperative complications during endoscopic DCR occur in ~1% of cases.2 To date, there have been no reported cases of permanent vision impairment or cerebrospinal fluid (CSF) leak. Bleeding is the most commonly reported intraoperative complication, followed by exposure of periorbital fat.9 One case of anterior ethmoid artery bleeding has been reported. Postoperative complications occur in ~6% of cases and include periorbital ecchymosis, synechiae, formation of granulation tissue, and periorbital emphysema.2,7

Controversies

Endocanalicular Stenting

The majority of published reports on DCR use closed-loop endocanalicular silicone stents, such as a Guibor or Crawford tubes.2 The duration of stenting varies among studies from 1 to 6 months. Stents theoretically maintain patency of the newly created rhinostomy during the healing phase. Others have suggested that the stent helps dilate the common canaliculus, which may be narrow in certain instances.8 Stents are generally well tolerated but may occasionally irritate the surrounding conjunctiva or become dislodged, requiring repositioning. Additionally, there is some evidence that silicone stents may induce granulation tissue at the rhinostomy site, prolonging the overall healing time.21 Reported case series of endoscopic DCR without stenting show success rates between 74 and 94%.2 Only a few studies have directly compared endoscopic DCR with and without stenting, with most showing no difference. A single prospective, randomized study was done by Smirnov et al.22 Forty-six consecutive patients were randomized to endoscopic DCR with or without silicone stenting for a period of 2 months. Patients were followed for 6 months postoperatively. The success rate of 100% was noted for the group without stenting versus 78% for the group with stents, a difference that just reached statistical significance. Currently, there is little evidence to suggest that endocanalicular stenting improves outcomes. However, this practice remains entrenched and is unlikely to change until additional randomized trials or large case series are presented.

Technique

The patient is positioned supinely, and topical vasoconstriction is achieved with oxymetazoline (0.05%) pledgets (see Video 59, Endoscopic Orbital Decompression for Graves Orbitopathy ). The eyes are maintained within the surgical field and protected with scleral shields. Image guidance systems may be used at the surgeon′s discretion. Local injection of lidocaine 1% with 1:100,000 epinephrine is administered along the lateral nasal wall in the region of the maxillary line. Surgery begins with an incision just posterior to the maxillary line, and the uncinate process is removed, exposing the natural ostium of the maxillary sinus. During an orbital decompression, it is critical to widely open the maxillary sinus.

A large maxillary antrostomy optimizes access to the orbital floor and prevents blockage of the maxillary ostium from protruding orbital fat following decompression. Enlargement is performed primarily in a posterior direction, as extension of the antrostomy too far anteriorly risks damage to the nasolacrimal duct. An endoscopic sphenoethmoidectomy is then performed in standard fashion. We advocate removal of the middle turbinate during orbital decompression to optimize exposure of the medial orbital wall and facilitate postoperative cleaning. An image guidance system may be used at this point to confirm removal of all ethmoid cells along the medial orbital wall and to ensure complete dissection to the sphenoid face and posterior skull base. The skeletonized medial orbital wall is then carefully penetrated in a controlled fashion with a spoon curette or other blunt instrument ( Fig. 37.11 ). Ideally, the periorbita is kept intact during this maneuver to prevent fat herniation, which can obscure visualization.

The thin bone of the lamina papyracea is elevated while preserving the underlying periorbita. Bone fragments are removed using Blakesley forceps ( Fig. 37.12 ). Bone removal proceeds superiorly toward the ethmoid roof, inferiorly to the orbital floor, and anteriorly to the maxillary line. Bone in the region of the frontal recess is left intact; if bone is removed from this region, herniated fat may obstruct drainage of the frontal sinus, resulting in iatrogenic frontal sinusitis or a mucocele.

As dissection proceeds posteriorly, thick bone is encountered within 2 mm of the sphenoid face. This bone corresponds to the location of the anulus of Zinn, from which four of six extraocular muscles originate and through which the optic nerve passes. This landmark represents the posterior limit of a standard decompression. For patients with optic neuropathy, experienced surgeons may consider continuing the decompression posteriorly into the sphenoid sinus; however, the benefits of incorporating an optic nerve decompression into standard orbital decompression are unclear and may lead to inadvertent injury to the nerve.

Removal of the orbital floor can be technically challenging, depending on its thickness. Only that portion of the floor that is medial to the infraorbital nerve is removed beginning ~1 cm posterior to the orbital rim. A spoon curet is used to engage the orbital floor at its medial extent and the bone down-fractured ( Fig. 37.13 ). The bone of the orbital floor is thicker than that of the medial orbital wall, and significant force may be required for this maneuver. If the spoon curette is not sturdy enough for this portion of the procedure, a heavier instrument such as a mastoid curette may be used. The orbital floor may fracture in one large piece, typically with a natural cleavage plane at the canal of the infraorbital nerve, or it may fracture into several small pieces. A 30-degree endoscope and angled forceps facilitate bone removal while preserving the infraorbital canal as the lateral limit of dissection. The maxilloethmoid strut can be left in place to prevent hypoglobus. Once the lamina papyracea and medial orbital floor have been removed, the periorbita should be fully exposed. A sickle knife is then used to open this fascial layer ( Fig. 37.14 ). Care must be taken to avoid “burying” the tip of the sickle knife and potentially injuring the underlying orbital contents, such as the medial rectus muscle or optic nerve. The periorbital incision should be initiated at the posterior limit of the decompression (just anterior to the sphenoid face) and brought anteriorly so that prolapsing fat does not obscure visualization. Parallel incisions are performed along the ethmoid roof and orbital floor. The periorbita is then grasped and removed, exposing the underlying fat. A ball-tipped probe and sickle knife may be used to identify and incise remaining fibrous bands that often course superficially between lobules of orbital fat. If desired, fat lobules may be removed at this time. Upon completion of the procedure, a generous prolapse of fat into the opened ethmoid and maxillary cavities should be observed ( Fig. 37.15 ). The globe may be ballotted to encourage maximal fat herniation and confirm a decrease in retropulsive resistance.

Depending on the clinical scenario and desired degree of decompression, a subsequent lateral decompression may be performed through an external approach. When performed immediately following medial decompression, the orbital contents are easily retracted in a medial direction, allowing for excellent exposure of the lateral bony wall. Bilateral decompressions may be performed concurrently or in a staged procedure, depending on patient and surgeon preferences. Nasal packing is avoided to ensure maximal decompression and avoid compression of exposed orbital contents. The patient is discharged the morning after surgery with a prescription for oral antistaphylococcal antibiotics and instructions to begin twice daily nasal saline irrigations. At the first postoperative visit 1 week following surgery, crusts and debris are cleaned from the surgical site under endoscopic guidance ( Fig. 37.16 ).