To perform an external DCR, a 15-mm skin incision is made medial to the medial canthus. A skin-muscle flap is formed to reveal the anterior limb of the medial canthal tendon. This is divided and the periosteum opened. The periorbita is elevated to displace the lacrimal sac and duct laterally. A 3-mm up-biting right-angled Kerrison rongeur is used to break through the thin bone of the lacrimal fossa and a bony osteotomy is formed, initially proceeding anteriorly, inferiorly, and then posteriorly. An osteotomy of at least 15 mm in diameter is created. The lacrimal sac is then probed and opened longitudinally. Any grossly suspicious mucosa should be biopsied and submitted for pathologic review. The nasal mucosa is incised in a similar longitudinal fashion, with relieving incisions at either ends forming an “H” shape. A silicone stent is inserted and tied loosely to prevent cheese-wiring of the canaliculi. The posterior lacrimal sac flap is sutured to the posterior nasal flap, typically with a single continuous 6/0 Vicryl suture. Three sutures are then used to appose the anterior nasal mucosal and anterior lacimal sac flaps. Where possible, these are suspended by attachment to the overlying orbicularis. The anterior limb of the medial canthal tendon is reapproximated and the skin is typically closed with a 6-0 polypropylene suture.

When carrying out endonasal, non-endoscopic DCR, surgeons often utilize a 20-gauge disposable vitrectomy light pipe threaded through the upper canaliculus to guide placement of the osteotomy [6]. After decongestion, an elliptical nasal mucosa incision down to bone, centered over the transilluminated light target is made. Mucosa is stripped from underlying bone and peeled away. An osteotomy is fashioned with an attempt to rongeur sufficient bone superiorly and anteriorly to easily visualize the entire width and most of the length of the lacrimal sac and duct. Care is taken to remove sufficient bone superiorly to ensure that the target light pipe when held horizontally across the common canaliculus can be visualized tenting the lacrimal sac within the nose. A posteriorly hinged U-shaped oval flap is made and reflected posteriorly and the lacrimal system is usually intubated.

Standard functional endoscopic sinus surgery (FESS) scopes were commonly used, in addition to keratomes, standard blades, Freers elevators, Blakesley forceps, and up-biting Kerrison rongeurs. Lacrimal probes or a light pipe passed into the lacrimal sac were often used to guide placement of the osteotomy. The nasal mucosa was incised and excised overlying the planned osteotomy site before carrying out the osteotomy with up-biting Kerrison ronguers. The inferior two-thirds (or less) of lacrimal sac was often all that was exposed. The sac was incised and its mucosa either reflected anteriorly and posteriorly, or trimmed. Silicone tubes were passed and the nose was often temporarily packed [15, 20, 21].

After decongestion, the nasal mucosa is usually infiltrated with 2 ml of lignocaine 2 % with 1:80,000 epinepherine using a dental syringe above and anterior to the middle turbinate. A mucosal incision with a small-angled crescent blade is made on the lateral nasal wall, 2–3 mm posterior to the maxillary line, starting 8 mm above the insertion of the middle turbinate and extending vertically down to a level just below the body of the middle turbinate. Using a number 15 scalpel blade, two horizontal incisions are made, 8 mm above the insertion of the middle turbinate and just below the body of the middle turbinate, respectively. This creates the posterior nasal mucosal flap, which is reflected using a Freer elevator, exposing the junction of the hard frontal process of the maxilla and the thin lacrimal bone. The lacrimal bone is removed off the inferior half of the sac using a Freer elevator or a forward-biting up-cutting 40° Kerrison rongeur. The frontal process of the maxilla, overlying the anterior and inferior portions of the lacrimal sac is removed and the osteotomy continued superiorly until it is no longer possible using the standard Kerrison. A burr or drill is utilized at this stage, exposing the fundus of the sac. The agger nasi air cell (the anterior most ethmoid cell) is often exposed as the fundus extends above the axilla of the middle turbinate [22].

The medial wall of the sac is then tented with a probe to ensure that all bone at least 5–10 mm above the common canalicular opening has been removed. The medial wall of the sac is incised vertically with a crescent blade to create large anterior and smaller posterior flaps. Small additional relieving incisions allow the flaps to be reflected onto the lateral nasal wall and sit “flat.” Good mobility and marsupialization of the lacrimal mucosal flaps has been associated with better outcomes [23]. A silicone stent can be passed and tied loosely to protect the internal ostium.

A posterior [12] or inferiorly hinged [13] nasal mucosal flap is formed along the frontal process of the maxilla. The mucosal flap (which will form the anteriornasal mucosal flap) is elevated using a Freer elevator, maintaining the tip of the elevator on the bone, reflecting it out of the surgical field. The technique proceeds the same as powered endoscopic DCR until the osteotomy can no longer be continued superiorly using a standard Kerrison rongeur. A modified bone nibbler may be used at this time to aid bone clearance at the fundus of the sac [13].

In order to achieve an absolute cure, a large fistula between the lacrimal sac and the nose is required leaving the canaliculi as the only zone of residual tear resistance [24]. It is generally agreed that exposure of the inferior and superior parts of the lacrimal sac should be accomplished, usually requiring an osteotomy of at least 15 mm, even approaching 20 mm [22, 24, 25]. Whether the new soft–tissue ostium of the entire marsupialized lacrimal sac remains stable in size beyond the first few postoperative weeks is unclear. It appears to reduce a small amount, with one endoscopic study measuring an average soft–tissue ostium size 12 months after surgery of 10.1 by 6.6 mm [26]. This is most in keeping with our own experience. Some have suggested that the soft–tissue ostium may shrink by 50 % at 6 months or even smaller [27, 28]. Others have found no significant relationship between bony ostium size and outcomes of surgery [23].

The main cause of failure in DCR surgery is fibrosis of the intranasal soft-tissue ostium, both in external DCR and endonasal DCR [11]. For the surgery to be successful, the mucosa of the lacrimal sac must anastamose to the nasal mucosa with the fistula remaining patent. The natural response from a surgical insult means granulation tissue can grow over the surgical ostium, rendering the procedure a failure. In successful surgery, once the lacrimal and nasal epithelium have healed together, the signal for secondary intention healing is turned off [14]. In a recent article looking at 20 failed DCRs, all had rhinostomy sites that were closed with fibrous tissue. None had canalicular or common internal ostium obstructions before undergoing revision surgery [29]. Presuming we should aim for anatomic surgery, we can maximize the success of DCR surgery by any means that helps tip the balance toward primary intention healing of the mucosa and away from secondary intention granulation [14]. The benefits of anatomic surgery may be difficult to prove, with many studies comparing different techniques or simple flap removal, but the concept should be sensible to any contemplative surgeon. Many authors have found that creation of mucosal flaps does not seem to increase the success rate of endoscopic DCR and can be technically challenging or time-consuming [5]. Others have described successful results with simple flap removal [30, 31]. It is only when endonasal DCR began to emulate the approach of external DCR that success rates improved [22].

The evidence base either in favor of or against the practice of routine intubation remains lacking [32]. Certainly, in experienced hands it does not appear to be necessary to intubate every patient, but until very recently, the majority of surgeons still routinely did [18, 32–35]. Silicone tubes are inserted with the aim of reducing the risk of fibrosis of the internal ostium of the common canaliculus while epithelial migration and repair takes place. In the absence of definitive canalicular disease, there is no clear evidence that intubation in routine DCR is superior to nonintubation. In the setting of canalicular disease, nonintubation may not be appropriate [36]. Other situations prompting intubation, but for which evidence is also currently lacking, include previous acute dacryocystitis, poor flap creation, revision surgery, excessive bleeding, inflammatory disease, and small lacrimal sacs [32].

A retrospective study has attempted to compare surgical outcomes in a group of 48 endonasal laser DCR procedures without MMC to outcomes in a group of 123 consecutive procedures in which MMC (0.5 mg/ml) was applied to the intranasal ostium for 5 min. The success rate in the MMC-treated group was statistically significantly greater than that of the controls (99 % vs. 90 %) [37]. Assessment of outcomes with or without MMC further blurs true differences. MMC cannot always deliver success from a poor procedure and should not be regarded as the solution for poor primary surgery. “The MMC Dilemma” chapter in this volume has analyzed in depth the usefulness of MMC in DCR surgery.

It may be fair to say that in experienced hands, there is no significant difference in the time taken to perform a successful DCR. Any technique that inadequately removes bone and incompletely excises mucosa would be faster, hence endoscopic laser DCR is arguably the quickest surgery [21].

Success rates for external DCR have historically been quoted as over 90 % [38], and often over 95 % [11, 16]. These high success rates are similar for both anatomical patency and resolution of patient symptoms. Early mechanical endoscopic DCR could not match these figures: in an early series of 123 patients, 83 % success was reported [39]. Subsequent smaller series claimed to have improved upon this (86–100 %) [40, 41].

Due to the perceived inferiority and technical complexities of endoscopic DCR, it remained unpopular with Ophthalmologists when compared to external DCR [42]. The development of surgical lasers was thought to hold the key, as a less invasive form of lacrimal surgery that would improve success. Despite early promise (100 % success in ten patients) [43], it became accepted that success was still lower than conventional surgery (77–83 %) [42, 44]. The high failure rate of endoscopic laser DCR was attributed to scarring (nasal and medial lacrimal sac mucosa was excised or obliterated) and the small size of the bony osteotomy. It is not possible to remove the thick bone of the frontal process of the maxilla with most lasers, leading to a small and inadequate osteotomy [45]. This led others to focus on mechanical means of creating a larger osteotomy, with slightly greater success (86 %) [46, 47].

Modern endoscopic DCR respects anatomic surgical principles key to all successful DCR surgery. A large osteotomy is created with preservation of mucosa so that flaps can be fashioned to achieve a mucosal anastomosis with the lacrimal sac, minimizing secondary intention healing and scarring response [22]. Endoscopic anatomical success could finally be achieved and replicated at other centers in 95 % (or more) of cases [10, 13].

Long-term analyses have reported 91 % success with external DCR (437 cases, average follow-up 71 months) [48]. Long-term studies of endoscopic DCR describe 82–94 % success (108 and 165 cases, average follow-up 49 and 92 months) [49, 50]. Grouping endoscopic DCR as a single entity, one can see is unhelpful. It does not distinguish between types of endoscopic techniques, nonstandard osteotomy, or flap formation. There are many individual variations. Published success rates, therefore, do not allow direct comparison of techniques. Success itself is a loosely applied term. Subjective dependence on symptoms is unreliable and some early papers based their outcomes on this [36]. It is rare for symptoms to completely resolve in elderly patients, yet these papers report a high level of symptom “resolution.”

Attempts to be more objective by incorporating syringing into the assessment, does not necessarily provide a straightforward “black or white” success or failure. Syringing is not physiological and papers that report “obstructed” or “completely patent” may either have excluded those with a small (10–20 %) degree of regurgitation on syringing or are ignoring subtleties before or after surgery. Other objective tests such as fluorescein dye retention testing or functional endoscopic dye test have been inconsistently utilized. Patient selection is not standard. It is easy to offer and predict a good outcome for patients with complete obstruction, but less so for those with partial obstruction, canalicular disease, or that overused and loosely defined term, the group with “functional epiphora.” [51] The lack of agreed or standardized outcome measures or even duration of follow-up, highlights how difficult comparisons actually are.