The success of DCR can be gauged by functional success and anatomical success. Functional success refers to lack of tearing 3 months after surgery, a good indicator of successful surgery as suggested by the Royal College of Ophthalmologists guidelines. Anatomical success can be confirmed by patency on lacrimal irrigation, visualization of ostium on nasal endoscopy, positive functional endoscopic dye test, and scintillography or contrast dacryocystography.

In most cases, the causes of failed DCR can be determined by lacrimal probing and nasal endoscopy. Syringing will be nonpatent in cases of failed DCR and probing should be performed to identify the site of obstruction. For a scarred internal ostium, a negative endoscopic dye test will be observed. The use of imaging studies like scintillography or dacryocystography (DCG) with plain films may provide further information in delineation of the lacrimal drainage tract and determination of the exact site and nature of obstruction, helping to formulate a surgical strategy for revision. DCG can also be performed with computed tomography (CT) or magnetic resonance imaging (MRI). However, cost and availability may be an issue. Typically, a patent fistulous tract confirmed by lacrimal probing and irrigation gives a characteristic “Y-on-its-side” configuration of the soft tissue on CT. Occlusion of osteotomy by soft tissue corresponds to a mucocele-like soft tissue density with a central lucency and soft tissue obstruction. Occlusion of osteotomy by inadequately excised bone is evident by bone in the region of osteotomy [11]. Using the spiral technique, CT-dacryocystography (CT-DCG) of high resolution allows measurement of the diameter of the osteotomy window, evaluation of osteotomy position relative to the lacrimal sac and reveals abnormal finding around the osteotomy-like extension of ethmoidal air cells medial to the lacrimal sac, concha bullosa, nasal polyp, and any medial canthal mass that might contribute to the failure of DCR (Fig. 26.6) [18]. In the study by Choi et al. [19], preoperative evaluation of obstruction level using DCG was helpful in predicting surgical outcome of endoscopic DCR. Among all, treatment of sac–duct junction obstruction with DCR had the highest success rate, followed by NLDO, common canalicular obstruction. Saccal obstructions carried the worst prognosis [19]. Finally, dacryoendoscopy, if available, may be used to delineate the intraluminal pathology within the lacrimal system.

Before deciding on revision surgery, patient with a failed primary DCR must be reexamined to determine the etiology of symptoms, especially to rule out other causes of tearing, such as blepharitis, trichiasis, lid malpositions like lower-lid laxity, entropion, ectropion, punctal abnormalities, and canalicular obstructions. Systemic inflammatory disease should also be excluded. Standard preoperative evaluation includes dye-disappearance testing, lacrimal irrigation and probing, and endoscopic evaluation of the internal ostium and nasal cavity are essential. If the ostium is found to be patent on irrigation and with a nasal endoscopy, the diagnosis of functional NLDO should be considered. Functional NLDO is defined as delayed tear clearance on scintillography or dacryocystography in the absence of anatomic obstruction. It is thought to be caused by a narrowing of the nasolacrimal duct or failure of the pump mechanism [20, 21]. Functional NLDO has a greater incidence of surgical failure, and patients may experience persistence of symptoms despite adequate surgery. Revision surgery in this subset of patients has been shown to be of little value [22].

In managing a failed DCR, the options are mainly surgical, though some patients may opt for observation. A failed primary external DCR can be revised externally or endoscopically; likewise, a failed primary endoscopic DCR can be amended endoscopically or externally. Various adjunctive measures including intraoperative application of Mitomycin-C (MMC) and intubation with silicone stents have been proposed to enhance the success rate of revision surgery. Recently, balloon dacryoplasty has been suggested as a less traumatic alternative to salvage a failed primary surgery. In cases of common canalicular obstruction, the revision can be as well performed by external or endoscopic approaches [9].

In the setting of revision DCR, the keys to success include:

The nasal mucosal flap is incised slightly more anteriorly over the frontal process of maxilla than for primary cases (Fig. 26.7). This allows the mucosal incision to be made onto bone and when this flap is elevated off the bone, it allows the correct surgical plane to be established for dissection of the mucosal flap off the underlying scar tissue. The osteotomy is then enlarged with ronguer until the lacrimal sac is completely exposed. If the sac is relatively normal in size, standard mucosal flaps are fashioned. In cases when the sac is scarred and contracted and it is difficult to fashion mucosal flaps, the mucosal apposition between the nasal and lacrimal mucosa can be obtained by trimming correspondingly less of the nasal mucosal flap (Fig. 26.8). One may also consider using the agger nasi mucosa as a free graft to create functional mucosa surrounding the common canaliculus–sac junction [23]. It is very important to clear the area around common canaliculus and expose it well (Fig. 26.9). Correction of nasal pathologies such as deviated nasal septum or turbinoseptal synechiae around the ostium might be required to allow adequate surgical exposure [24]. The rest of the procedure is similar to primary cases. Intraoperative MMC (Fig. 26.10) and silicone intubation (Fig. 26.11) should ideally be used. Additional use of Sisler’s canalicular trephines for distal canalicular obstructions and balloon dacryoplasty to dilate the ostial stenosis can be combined with revision endoscopic DCR as a multimodal management for selected cases.

Apart from the established advantages for endoscopic primary DCR, namely the avoidance of cutaneous scar and preservation of pump action of the orbicularis muscle, some authors suggested that the endonasal approach is well-catered for revision surgery given its direct access to the residual lacrimal sac through the previously created bony ostium, improved visualization of osteotomy position relative to the lacrimal sac, easier hemostasis, and the ability to address concurrent intranasal pathologies [25]. However, like the primary procedure, the potential drawbacks are steep learning curve and high equipment cost.

A skin incision is made through the original scar (Fig. 26.12). Orbicularis is separated at the junction of orbital and palpebral portions (Fig. 26.13). Bowman lacrimal probes are used to check the patency of both inferior and superior canaliculi. The medial palpebral tendon, if present, is divided. A combination of sharp and blunt dissection is used to separate the scar above and below the probes. Great care is taken not to enter the lumen of the common canaliculus. The anterior edge of the original rhinostomy is identified and periosteum is freed from bone for approximately 4 mm anterior to the bony edge of the original rhinostomy to expose the uncut bone all around (Fig. 26.14). The bony ostium is enlarged (Fig. 26.15) to allow adequate exposure of the sac and to expose virgin nasal mucosa (Fig. 26.16). A trapdoor incision based in the newly exposed virgin nasal mucosa is cut so that the lateral free edge is close to the previously identified common canaliculus and the upper and lower edges are next to the edges of the newly enlarged rhinostomy. The nasal mucosal flap is reflected anteriorly with traction sutures. The interior of the rhinostomy is then examined for any intervening bone, ethmoid air cell, synechiae, dacryoliths, or simple cicatrix and appropriately removed. The virgin lacrimal sac flaps if any are fashioned as usual but in cases of intense fibrosis, careful elevation is mandatory (Fig. 26.17). Anterior alone or both anterior and posterior mucosal flaps are sutured with fine absorbable suture like Dexon or Vicryl (Fig. 26.18). If common canalicular obstruction is present, the area can be trephined to remove the cicatrix or an endocanaliculotomy can be done. It is important that the flaps be sutured under slight tension so that they do not adhere internally and form obstruction. Mitomycin-C (Fig. 26.19) and silicone stents (Fig. 26.20) have been found to be beneficial in revision DCRs specially if there is canalicular pathology or if the sac is small, scarred, or inflamed. Lastly, orbicularis and tendon are repositioned with an absorbable suture and the skin is closed with an interrupted nylon suture [9].

The success rates for revision endoscopic DCR are in the range of 76–100 % [23, 26–29], comparable to that reported for external revision DCR, which varied from 80 to 90 % [9, 30]. Tsirbas A et al. [23] showed a direct comparison of the endoscopic and external techniques and showed that revision endoscopic DCR surgery was successful in 77 % of cases and external revision was successful in 85 %, yet this difference was not statistically significant. In another study by Paik et al. [31], which included 82 endoscopic revisions, a success rate of 84 % was achieved for those with failed primary external DCR and 81 % for those who failed a primary endoscopic DCR. In preoperative nasal endoscopy, more of those who underwent primary external DCR exhibited a hypertrophic middle turbinate or severe septal deviation, whereas more of those who underwent primary endoscopic DCR exhibited a small ostium.