Primary external dacryocystorhinostomy versus primary endonasal dacryocystorhinostomy: a review

Authors


Dr Seng Chee Loon, National University Hospital, Lower Kent Ridge Road S119074, Singapore. Email: ploonsc@yahoo.com

Abstract

This paper aims to compare the efficacy, mean operative time and adverse effects of primary external dacryocystorhinostomy (DCR) versus primary endonasal DCR in treating acquired nasolacrimal duct obstruction. Searches were performed for studies comparing the two procedures. Two reviewers independently extracted data from 14 eligible studies. A random effects model was used to analyse the studies. Outcome measures were defined as patency of the nasolacrimal canal and mean operative time, and adverse effects as cutaneous scarring and bleeding. Both procedures were comparable in efficacy in terms of full success, partial success and anatomic patency. Subgroup analysis showed no significant difference between prospective and retrospective studies as well as between non-laser endonasal DCR versus external DCR and laser endonasal DCR versus external DCR. Endonasal DCR had a significantly shorter mean operative duration, be it laser endonasal DCR (mean difference: 37.65 min, 95% confidence intervals: 3.54–71.75 min, P: 0.03) or non-laser endonasal DCR (mean difference: 19.22 min, 95% confidence intervals: 2.15–36.28 min, P: 0.03). The odds of postoperative bleeding was not significantly different between the two procedures, whereas postoperative cutaneous scarring was unique to external DCR and occurred in 50 out of 402 (12.44%) external DCRs recorded. Endonasal DCR has comparable success rates with external DCR and has a shorter operative time and no cutaneous scar. However, drawbacks include the steep learning curve and higher costs.

Introduction

Acquired nasolacrimal duct (NLD) obstruction is a fairly common disorder that occurs more frequently in females than in males.1 A study showed that the incidence of NLD obstruction was 20.24 per 100 000 in Olmsted County, Minnesota from 1976 to 2000.2 The most common symptoms of acquired NLD obstruction are epiphora and acute or chronic dacryocystitis. Epiphora caused by lacrimal duct obstruction is a common ophthalmologic problem and accounts for approximately 3% of all ophthalmologic clinic visits.3 Dacryocystorhinostomy (DCR) has been established as the prevailing procedure for acquired NLD obstruction.4 DCR is a surgical method that allows the direct drainage of tears from the lacrimal sac into the nasal cavity, bypassing the blocked NLD. There are two main types of DCR, namely external DCR and endonasal DCR.

External DCR was first described in 1904 by Toti and the modified version by Dupuy-Dutemps and Bourguet has remained the gold standard in the treatment of acquired NLD obstruction.5 The procedure entails a skin incision and drilling or rongeuring the bone of the anterior lacrimal crest and lacrimal sac fossa to reach the lacrimal sac.6

Caldwell first described endonasal DCR in 1893 and West and Halle later modified it.7 These techniques were limited in use due mainly to the difficulty in visualizing the endonasal anatomy during the operation. The development of the operating microscope and rigid endonasal endoscope caused this procedure to gain popularity.7 Endonasal DCR can be carried out in various ways, either with or without the help of an endoscope; and with the use of different equipment such as rongeur, drill, chisel and various types of lasers.8,9 McDonogh and Meiring first used endoscopes in transnasal DCR in 1989.10

Studies have been done comparing external DCR and endoscopic DCR, with a common notion that external DCR has higher success rates. A literature search revealed that a review has yet to be done. The aim of this paper is to compare the efficacy, mean operative time and adverse effects between primary external DCR and primary endonasal DCR. Revision DCRs are not included in this review as revision DCR has been known to have a lower success rate than primary DCR.11

Methods

Selection criteria

All randomized and non-randomized, prospective and retrospective studies that directly compare primary external DCR with primary endonasal DCR that were published in English were taken into consideration. We accepted studies using various types of endonasal DCR as long as they were primary DCRs and directly compared them with primary external DCR. All studies included were required to provide the success rates of both primary external DCR and primary endonasal DCR and preferably state other comparative characteristics such as adverse events and time taken for the surgery. We did not accept any papers from which we could not obtain the success rates for only primary DCRs. The date range for the papers included was June 1998 to March 2009.

Search strategy

The databases that were searched were Medline via PubMed, The Cochrane Library, OvidSP and ScienceDirect. The Medline search strategy can be seen in Appendix 1. The authors were contacted to clarify the results of their studies whenever possible to determine if there were any additional unpublished data. No unpublished data were included in this review.

Data extraction

Two reviewers (DL and CC) independently extracted data from the included studies to gather characteristics of the papers, such as authors, locations, sample size, patient characteristics, interventions, outcomes, adverse effects and measurement methodology. Further arbitration was done by a third reviewer (SCL) regarding the choice of papers included and the quality of the papers being selected.

Outcome measures

One of the outcome measures that we used was efficacy. This was defined as full success, partial success and anatomic patency. Anatomic patency of the NLD can be determined by observing minimal or no reflux through the opposite canaliculus during lacrimal irrigation or through a negative functional endoscopic dye test. Full success was defined as resolution of epiphora together with anatomic patency of the NLD and partial success was defined as the improvement of epiphora with anatomic patency of the NLD. The outcome measures were selected based on the most commonly available data across the studies we included.

Another outcome measure that we have reviewed was the difference in the mean operative time between primary external DCR and primary endonasal DCR.

We also reviewed the most common adverse events that were reported which were postoperative complications, bleeding and cutaneous scarring.

Quantitative data synthesis and analysis

We extracted the data from each study and used Review Manager 5 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, 2008) to gather summary estimates. Outcome anatomic patency and adverse events were reported as odds ratio (OR) with 95% confidence intervals (CI). Mean operative duration was reported as mean difference (MD) with 95% CI.

The Q-test was used to determine inter-study statistical heterogeneity. A random effects model was used to calculate the results outcome, as it weights the studies relatively more equally than a fixed effects model in the presence of heterogeneity.12 The calculated I2 indicates the variability in effect estimates that is associated with heterogeneity rather than chance.

Subgroup analysis

Subgroup analysis was performed for the outcome asymptomatic and anatomic patency. The subgroups analysed were prospective and retrospective studies as well as non-laser endonasal DCR and laser endonasal DCR compared with external DCR.

Results

Literature search

An initial search revealed 988 papers, of which we obtained the full papers of 24 publications and narrowed them down 14 papers for this review (Fig. 1). We merged the data of two of the studies by Hartikainen et al. as both used the same population of patients that underwent external DCR but had different populations for the non-laser endonasal DCR group7 and the laser endonasal DCR group.13 We also merged the data of the papers by Lester et al.14 and Mirza et al.15 as both papers used the same population of external DCRs but had different populations of non-laser endonasal DCR and laser endonasal DCR, respectively.

Figure 1.

Flow chart of selection of studies for review. DCR, dacryocystorhinostomy.

Characteristics of papers

There were 1680 patients in total, with 1782 DCRs being performed. Out of the 1782 DCRs performed, 360 (20.2%) of them were done on male patients, 908 (51.0%) of them were done on female patients. The genders of the remaining 28.8% were unable to be determined as these papers reported gender based on the number of patients instead of the number of DCRs performed. The other characteristics of the papers are summarized in Table 1.

Table 1.  Characteristics of papers
PublicationYearLocationNo. of patientsNo. of DCRsMFMean ageMean f/u period (months)Commercial support
  1. DCR, dacryocystorhinostomy; NS, not stated.

Leong et al.242009UK5160NSNS627No
Sharma222008Nepal30230265237359No
Lester et al.142007UK10010639675910No
Ajalloueyan et al.182007Iran210244NSNS4218No
Yigit et al.102007Turkey103104NSNS5712No
Tsirbas et al.252004Australia5055NSNS6113No
Malhotra et al.192003UK4851NSNS648No
Dolman62003Canada3493541182366012No
Mirza et al.152002UK727636406012No
Watts et al.162001UK414110317417No
Ibrahim et al.172001UK148163561076732No
Cokkeser et al.232000Turkey115130161143925No
Hartikainen et al.71998Finland606413516312Yes
Hartikainen et al.131998Finland31327256512Yes

Efficacy

We compared external DCR and endonasal DCR with regards to efficacy, defined as full success, partial success and anatomic patency. No statistically significant difference was found between the two techniques in terms of full success (OR: 1.28, 95% CI: 0.85–1.95, P: 0.24) (Fig. 2), partial success (OR: 1.54, 95% CI: 0.60–3.96, P: 0.37) (Fig. 3) and anatomic patency (OR: 2.79, 95% CI: 0.86–9.07, P: 0.09) (Fig. 4).

Figure 2.

Full success – asymptomatic and anatomic patency at final follow-up. ‡Combined data from Hartikainen, Antila 1998 and Hartikainen, Grenman 1998. §Combined data from Mirza 2002 and Lester 2007.

Figure 3.

Partial success – symptomatic improvement and anatomic patency at final follow-up. §Combined data from Mirza 2002 and Lester 2007.

Figure 4.

Anatomic patency at final follow-up. ‡Combined data from Hartikainen, Antila 1998 and Hartikainen, Grenman 1998.

Watts et al.16 was not included in the analysis for full success as the definition for success was anatomic patency. Although they did mention that the patients were asked for symptomatic improvement during follow-up, we were unable to determine the number of full successes and partial successes.

Subgroup analysis

The results for subgroup analysis between prospective and retrospective studies were similar. No statistically significant difference was found between the prospective (OR: 1.12, 95% CI: 0.48–2.61, P: 0.79) and retrospective studies (OR: 1.44, 95% CI: 0.96–2.17, P: 0.08) in terms of full success (Fig. 2).

Subgroup analysis also found that the success rates were not significantly different between laser endonasal DCR and external DCR (OR: 2.15, 95% CI: 0.93–5.00, P: 0.08) as well as between non-laser endonasal DCR and external DCR (OR: 1.20, 95% CI: 0.79–1.82, P: 0.40) (Fig. 2).

Mean operative duration

Endonasal DCR was consistently found to have a significantly shorter mean operative duration compared with external DCR, be it laser endonasal DCR (MD: 37.65 min, 95% CI: 3.54–71.75 min, P: 0.03) (Fig. 5) or non-laser endonasal DCR (MD: 19.22 min, 95% CI: 2.15–36.28 min, P: 0.03) (Fig. 6).

Figure 5.

Mean difference in operative duration between laser endonasal dacryocystorhinostomy (DCR) and external DCR.

Figure 6.

Mean difference in operative duration between non-laser endonasal dacryocystorhinostomy (DCR) and external DCR.

Adverse effects

No significant difference was found between the occurrence of external DCR and endonasal DCR for the postoperative complication, bleeding (OR: 1.00, 95% CI: 0.50–2.01, P: 0.99) (Fig. 7).

Figure 7.

Postoperative complication (Cx). ‡Combined data from Hartikainen, Antila 1998 and Hartikainen, Grenman 1998. §Combined data from Mirza 2002 and Lester 2007.

Postoperative cutaneous scarring was unique to the external DCR procedure and occurred in 50 out of 402 (12.44%) external DCRs recorded.

Publication bias

The funnel plot for the outcome of full success showed no evidence of publication bias (Fig. 8).

Figure 8.

Funnel plot for publication bias.

Discussion

Outcome measures

Anatomic patency is the most common measure of success of DCR. However, anatomic success does not always translate to functional success. Hence, we analysed full success, partial success and anatomic patency separately.6,17

Summary of key findings

Our findings have shown that there was no significant difference between primary endonasal DCR and primary external DCR in terms of full success, partial success and anatomic patency. Subgroup analysis showed no significant difference between prospective and retrospective studies as well as between laser endonasal DCR versus external DCR and non-laser endonasal DCR versus external DCR in terms of efficacy. Mean operative time was found to be significantly shorter in endonasal DCR than in external DCR. The occurrence of postoperative bleeding was not significantly different between the two procedures. Postoperative cutaneous scarring was unique to the external DCR procedure.

External validity

The patients from all the trials were derived from eight countries (Australia, Canada, Finland, Iran, Nepal, Turkey, UK and USA). Majority of the studies were concentrated in countries with Caucasian populations, hence providing a limited analysis of other races. This poses as a potential limitation due to the different scarring and healing outcomes of various races.

The lasers used in the studies that had laser endonasal DCR were endocanalicular diode laser,18 holmium:YAG laser,19 KTP laser15 and CO2-Nd:YAG laser,13 whereas the efficacy of other types of lasers was not explored.

Clinical implications

One of the major concerns with the use of endonasal DCR for managing NLD obstruction is the lower success rates as compared with external DCR. Our review has shown that both non-laser endonasal DCR and laser endonasal DCR had no significant difference in success rates when compared with external DCR. However, the OR for laser endonasal DCR versus external DCR is higher than that for non-laser endonasal DCR versus external DCR even though both subgroup analyses favour external DCR and are not significant. Maini et al. compared surgical endoscopic endonasal DCR and laser endoscopic endonasal DCR and reported that endonasal laser DCR had better symptomatic success rates at 3 months but lower symptomatic success rates at 12 months. However, this difference is not statistically significant. The authors concluded that the findings suggested that there was a deterioration of results for the endonasal laser group even though the change was not statistically significant.20 There are several ways to carry out non-laser endonasal DCR, which use different materials to expose the lacrimal sac, such as rongeour, drill and chisel, with various advantages and disadvantages.10 Dolman suggested that the use of laser may generate char around the ostium site, requiring postoperative wound cleansing and potentially accounting for the poorer success rates.6

Our studies have a relatively short follow-up period, ranging from 7 to 25 months. Differences in the outcomes for the different DCR procedures may only be apparent in a longer follow-up. It would be interesting to compare results for the various methods of DCR over a longer follow-up period.

Another suggested advantage of endonasal DCR is the shorter operative time. Our review has also shown that endonasal DCR had significantly shorter operative time as compared with external DCR. Moreover, local anaesthesia can be used for endonasal DCR,4 leading to a higher number of cases that can be done in each theatre session,17,21 as well as being able to be done as day cases.4,21 Local anaesthesia expands the population eligible for this procedure, as it allows older and less fit patients21 as well as inpatients for whom hypotensive anaesthesia is not advisable to safely undergo surgery.16 However, it should be noted that local anaesthesia can be used for external DCR as well.16 The use of local anaesthesia may be limited to specific types of procedures and local practices.

Our review found that there was no significant difference between the postoperative bleeding occurrences of both external and endonasal DCR.

We have acknowledged in this review that a cutaneous scar is unique to external DCR. One of the most important advantages of endonasal DCR is the lack of a cutaneous scar and its associated wound complications, such as skin swelling and bruising.6,9,10,17,22 The absence of an external incision diminishes the risk of damage to the medial canthal ligament, orbicularis oculi muscle and pretarsal fibres and therefore maintains the lacrimal pump.10,22,23

Another reported advantage of endonasal DCR is that haemostasis can be easily achieved, leading to less intraoperative bleeding.9,10,16,19 Other advantages of endonasal DCR include direct access to the rhinostomy site,7 the possibility of a thorough inspection of the intranasal anatomy at the time of surgery,7 decreased operative morbidity,10 enhanced recovery and the ability to concurrently address other nasal and/or paranasal sinus abnormalities through the same surgical approach.4,9,10,19

A disadvantage of endonasal DCR is the steep learning curve.4,9,15,23 It also has high equipment and instrumentation costs,23 in particular relating to the use of lasers in endonasal DCR.4 A concern about endonasal DCR is the lack of the sophisticated equipment may result in endonasal DCR not being able to be performed in less developed countries. However, Dolman has shown that by using surgical knives instead of lasers and surgical loupes instead of microscopes or videoendoscopes, both the surgical time and cost can be reduced. He has also performed and taught endonasal DCR in remote clinics in northern Canada and in several developing nations.6 A study reported that endoscopic endonasal DCR generated approximately twice as much income as external DCR and hence was more cost-effective mainly due to the higher number of cases per session, day case operations and using local anaesthesia.21

Ibrahim et al. reported three patients who had a bilateral functionally successful DCR with external DCR done on one side and endonasal endoscopic laser DCR done on the other side. All preferred the laser endonasal DCR procedure, citing a shorter recovery period and lack of skin incision.17 Dolman also reported five patients who underwent both external DCR and endonasal DCR with all preferring the non-laser endonasal DCR route.6

Reported advantages of external DCR include the unimpaired view of surgical area and being able to obtain a lacrimal sac biopsy.9 External DCR has been favoured for patients with lacrimal sac neoplasm,15 evidence of a severe post-traumatic bony deformity of the lacrimal sac15,22 or canalicular pathology.22

Some papers have reported that antimetabolitic agents such as mitomycin C16,19 and 5-FU16 inhibit fibroblast proliferation and hence reduce scarring at the internal ostium, which may be a cause of failure of endonasal DCR. Watts et al. used 5-FU in their endonasal DCR technique with no apparent difference in the results compared with success rates reported in other series. However, they had no control group in their study.16 More studies could be carried out to study the effect of antimetabolitic agents on the outcome of endonasal DCR.

Limitations

Only three randomized controlled trials were included in this review, and the rest of the prospective trials were non-randomized. Half of the studies included in this review were retrospective ones. Retrospective studies face a problem of having to exclude certain cases due to patient selection bias and inadequate follow-up,22 which may alter the success rates and patients included in the studies and hence affect the comparison of external DCR and endonasal DCR. Ben Simon et al. reports a need for prospective studies, with standardization of blockage site and osteotomy size and strict definitions of improvement and failure in order to evaluate more accurately between the two procedures.9

Another limitation of this review is the heterogeneity in terms of endonasal DCR techniques. The various endonasal DCR methods included in this review used either endoscopic or non-endoscopic means to visualize the surgical site and used different means to ablate nasal mucosa and bone such as powered and mechanical means and laser. Not all the various types of laser endonasal DCR have been reviewed in comparison with external DCR. The use of different lasers in endonasal DCR may lead to varying success rates.8

Furthermore, both external DCR and endonasal DCR are surgical procedures. Hence, the outcomes are often dependent on the surgeons performing the procedures.

We conclude that endonasal DCR has comparable success rates with external DCR, be it using non-laser or laser means. It also has the added advantages of having a shorter operative time and a lack of a cutaneous scar. The steep learning curve and higher costs are drawbacks to the endonasal DCR procedure.

Appendix

Appendix I

Search strategy and results

No.Search HistoryResults
1DCR or dacryocystorhinostomy1893
2External and 1315
3Endonasal or endoscopic and 1465
4Nasolacrimal duct obstruction1119
5Nasolacrimal duct obstruction and 1517
61 and 2 and 3164
71 and 2 and 3 and 497

Ancillary