One of the commonest causes of dacryocystorhinostomy (DCR) failure is the closure of the ostium due to healing of the mucosal edges, formation of synechiae, or presence of granulation tissue [1]. Another postulated cause of failure is the sump syndrome where a remnant of the inferior lacrimal sac acts as a nondraining reservoir [2]. These complications can be minimized by creating an adequately sized and appropriately placed ostium.

The ideal dimensions for the bony ostium in DCR remain unclear. Many authors believe the ostium should be long enough to allow opening of the entire lacrimal sac from the fundus to the junction with the nasolacrimal duct. It should be wide enough to allow creation of lacrimal sac flaps that lie apposed to the nasal mucosal flaps to promote primary intention healing. Similarly, complete marsupialization of the sac into the lateral nasal cavity is thought to achieve the lowest rate of ostial closure [3–5]. Others accept a smaller opening of the sac with a correspondingly smaller bony ostium.

Argin et al. attempted to define the exact dimensions for a bony ostium in the belief that creation of a large ostium will prevent closure [6]. Ben Simon et al. found a positive correlation between intraoperative osteotomy size and postoperative ostium measurements [7]. In contrast, other studies reported that the initial ostium size does not necessarily correlate with the final size [8–10]. Many authors also found that the intraoperative and final ostium size were not predictive of success [7, 8]. However, Ezra et al. observed a correlation between ostial size at 2 weeks and a successful outcome [9].

In a prospective study of 161 endoscopic DCRs where the entire sac was marsupialized, Chan and Selva found the majority of ostial shrinkage occurred within the first 4 weeks. The average ostial measurement 12 months postoperatively was 64.7 % of the initial bony osteotomy [8]. In another prospective study of 49 endoscopic DCR procedures, Mann and Wormald showed very similar results with the ostium measuring 77 % of the intraoperative size after 4 weeks and very little change after that [11]. Similar results have been observed for external DCR [7, 9, 10].

Some patients have features that make the endoscopic DCR surgery easier, such as a large lacrimal sac, thin lacrimal bone, small middle turbinate, posterior uncinate process, or an internal common opening that is situated more inferiorly in the lacrimal sac. Others have more challenging anatomy, for example, small lacrimal sac, thick frontal process of the maxilla, ethmoidal air cells significantly overlying the lacrimal sac, or a high internal common opening, which may predispose them to a higher failure rate [12]. Konuk identified large middle turbinates and severe septal deviation as causative factors in 14 % of failed cases [13].

Planning the best location and size of the bony osteotomy in endonasal DCR is dependent on awareness of the variations in anatomical landmarks of the lateral nasal wall. The most commonly used landmarks are the maxillary line and the axilla of the middle turbinate [14]. Based on cadaveric dissections, Orhan found that the maxillary line overlapped the lacrimal sac in 18/20 cadaveric specimens and that the lacrimal sac was located posterior to the maxillary line in the other two specimens [15]. However, Ali et al. [16] in their cadaveric study found that the spatial relationship of maxillary line and head of middle turbinate is not constant and hence should not be solely relied upon during surgery. There is also considerable variability in the location of another important landmark, the lacrimo-maxillary suture (LMS). Shams et al. found that the LMS was centrally located in the fossa in 25 % of Caucasian orbits, while in 32 % of orbits it was located closer to the posterior lacrimal crest indicating predominance of the thicker maxillary bone [17]. In contrast, a study of Indian orbits noted a centrally located LMS in 79 % of specimens and a maxillary dominant fossa in only 8.3 % [18]. A study based on CT findings in Asian orbits found that the lacrimal fossa was formed predominantly by the frontal process of maxilla in 79 % of patients [19].

Factors other than the size of the bony osteotomy may also affect the ultimate size of the ostium. Studies have shown that in cases of failed surgery due to osteotomy closure, healing occurs predominantly by fibrosis with very little new bone formation [20]. Especially in adults, bone growth would not be expected across a mucosal anastomosis or in the absence of periosteum, which is removed during surgery.

Once an ostium is created, the mucosal and bony edges will trigger an inflammatory response. The extent of the inflammation depends on the size of the defects between the raw edges and the individual’s innate healing response. While it seems intuitive that approximation of mucosal edges would lead to less granulation and scarring, several authors have described comparable success rates regardless of the number of created flaps [21–23]. Khalifa et al. conducted a prospective randomized controlled trial comparing endoscopic DCR with double posteriorly based nasal and lacrimal flaps to a technique in which the nasal and lacrimal mucosa are removed without creation of flaps. Although there was a better healing profile with fewer debridement sessions in the double flap group, this did not lead to a statistically significant increase in success rate (92 % vs. 87 %) [21]. However, several authors reported anatomical patency rates of more than 95 % in endoscopic DCR with the double flap technique, which allows complete marsupialization of the lacrimal sac into the lateral nasal wall [8, 24]. For external DCR, comparable success rates were achieved between groups where both anterior and posterior flaps were sutured and groups where only the anterior flaps were sutured and the posterior flaps were left either unsutured [25] or were excised [26]. Baldeschi et al. compared different patterns of mucosal dissection resulting in different number and extent of unsutured mucosal margins in external DCR. They found that the length of the margins did not adversely affect the success rate [27].

Anatomical variations in the lacrimal fossa and location of ethmoidal air cells mean that the apposition of mucosal edges may be achieved with different flap designs in different individuals. One may postulate that mucosal apposition rather than a standardized flap design may influence healing and hence the success rate. Despite the lack of evidence regarding the need for mucosal apposition, it is the authors’ preference to achieve apposition where possible to minimize secondary intention healing and the associated fibrosis.

Numerous studies have evaluated the antifibrotic properties of adjunctive Mitomycin-C (MMC) in an attempt to modulate the healing process during the initial stage of soft tissue granulation and thereby reduce the rate of anatomical failure.

Studies comparing the postoperative osteotomy size have found significantly larger ostia in patients treated with intraoperative MMC compared to control group both for external [28] and endoscopic DCR [29, 30]. A recent meta-analysis suggested that intraoperative MMC application may reduce the closure rate of osteotomies and enhance the success rate in external DCR [31] and both primary and revision endo-DCR [32]. However, several studies have failed to show any beneficial effect [33–36] No adverse effects from MMC were found in any of the DCR studies.

MMC dosage has ranged between 0.2 and 0.5 mg/ml and the exposure time from 2 to 30 min [32, 37]. Ali et al. [38] in their in vitro study attempted to address this issue and found that a concentration of 0.02 % for 3 min was the right dose. The MMC is generally applied topically but circumostial injection can also be used [39]. In their randomized controlled study, You and Feng found no significant difference in patency rate and ostium size between the groups receiving topical MMC in concentrations of 0.2 or 0.5 mg/ml and both groups had better outcomes compared to the control group [36]. In contrast, intraoperative or postoperative use of 5-Fluorouracil does not appear to influence the ostial size or the success rates [40, 41]. Wu et al. reported significant improvement of ostial patency for endoscopic DCR with use of Merogel, a hyaluronic acid derivative thought to promote epithelial healing and reduce scarring but there have been no other studies with this agent [42].

Surgeons have also utilized steroids both topically and in the form of injections into the tissue adjacent to the ostium but again there remains no evidence base on the effect this might have on patency rates.