Our results firstly show that in a cohort of patients with epiphora where alternate causes have been clinically excluded, a negative syringing (100% patency) failed to detect NLDS in 44% and FNLDD in 54% of cases. These results confirm that lacrimal syringing is unreliable for ruling out NLD stenosis or functional delay. Furthermore, full patency (or a negative syringing) is used by many to be a criterion for the diagnosis of non-anatomical FNLDD [13,14,15,16]. Our results suggest that significant NLD stenosis could likewise be found to be fully patent on syringing and in a similarly high proportion of cases. They further show that reflux on syringing may be found in close to half of FNLDD cases and cannot be used to differentiate it from NLDS.

The results also suggest that while any reflux may indicate pathology, it is also found in “normal” systems as defined by imaging. That is, the specificity of syringing was found to be limited, as approximately one-third of the cases that had normal NLD drainage on imaging were noted to have some degree of reflux on syringing. On the other hand, no patency on syringing was highly specific, with only one false-positive case. Taken together, no patency on syringing would suggest NLD impaired drainage with high confidence (98%), while partial (but not full) patency would suggest impaired drainage with lower confidence (65%).

The detection of NLD stenosis and non-anatomical functional delay is particularly challenging in the clinical setting, and the role of syringing in these cases has not been clearly characterized to date [7]. Syringing is the most frequently used, sometimes stand-alone test in clinical practice [1, 2]. On the other hand, lacrimal imaging studies are used infrequently [10]. Clinicians could thus benefit from knowing the strengths and limitations of syringing in the context of NLD drainage impairment, especially when patency (or partial patency) is demonstrated. This may guide consideration of further investigation and have implications for consenting patients regarding intervention success rates.

Whereas other studies have analyzed the relationship of syringing to DCG and DSG findings separately, to our knowledge, this study is the first to correlate lacrimal syringing to the combined (and complementary) findings on DCG and DSG [4,5,6, 12, 15]. The combination of these imaging studies currently provides the closest “gold standard” to diagnose the specific type and degree of NLD drainage impairment (before intraoperative confirmation). Namely, it could differentiate complete anatomical obstruction (NLDO), from partial obstruction (NLDS), from non-anatomical “functional” delay (FNLDD) [7, 8]. FNLDD was historically defined as “incomplete blockage” based on negative Jones 1 and a positive Jones 2 test [17, 18]. Nonetheless, this methodology probably incorporates stenosis and non-anatomical (functional) delay in the same cohort [8]. Furthermore, the Jones test is less commonly used in the lacrimal clinic than syringing [1].

The correlation between syringing and lacrimal imaging (DCG or DSG, separately) has been previously investigated; however, various definitions for full patency or a normal syringing were used. Some authors defined a fully patent syringing as less than 20% reflux [4, 12], others as 0% reflux [3, 5], while some did not clearly state their definition of normal syringing [6]. Our results suggest that any reflux should be considered a positive test, and the utility of less stringent criteria (such as > 20% reflux) is probably limited due to significantly diminished sensitivities. Furthermore, these previous studies did not stratify the results by the site of imaging abnormality (presac or postsac), whereas the current analysis focused on post sac (NLD) impairment.

These methodological discrepancies notwithstanding, our reported syringing general specificity (65%) and sensitivity for NLDO (91%) compare favorably with the figures reported by Nixon et al. [6] based on their comparison to DCG (53% and 86%, respectively). Our reported sensitivities for FNLDD (46%) and NLDS (56%) are similar to those reported by Kim et al. [5] In their study, reflux on syringing had a sensitivity of 50% when the DSG showed delay. Arguably, their results based only on DSG capture both stenosis and functional delay, thus falling within the range of our reported figures.

One previous study compared syringing to DCG and DSG separately [4]. The authors defined less than 20% reflux on syringing as a “freely patent” result. Peter and Pearson’s [4] study yielded a syringing sensitivity of 28.6% for anatomic abnormality on DCG and 25% for delay on DSG. While in the current analysis, the same cut-off of syringing (20% reflux) yielded higher sensitivity in NLDS (50%) and FNLDD (37%), both studies underscore the limitation of syringing in detecting these impairments.

The strengths of this study include a large number of patients and the use of comprehensive imaging (DCG and DSG), allowing analysis of homogenous etiologies of drainage impairment (NLDO, NDLS, NLDD). This study’s limitations firstly include its retrospective nature. Another limitation is using DCG and DSG as the diagnostic reference, as these modalities are themselves confined by investigative sensitivity and specificity constraints [6, 19, 20]. Nevertheless, combining both modalities may increase the sensitivity to 98% [15], and they are considered complimentary [8, 21]. Noteworthy, dacryoendoscopy can directly visualize an obstruction’s degree, level, and nature and perhaps be a better diagnostic gold standard [22, 23]. Nonetheless, it is a lacrimal procedure, often with simultaneous treatment of obstruction, and requires injection of local anesthesia (possibly with the addition of sedation in selected cases). It also entails special equipment (dacryoendoscope) that is not widely available yet.

Last, since syringing is a crude test, inter-tester variations in the exact pressure and subjective estimation of reflux (proportions) are possible. Nevertheless, in the current study, oculoplastic surgeons performed syringing, all trained under the last author (DS), who supervised and validated the technique. Thus, adherence to the same testing standards could be assumed for the entire study period. Furthermore, we categorized the estimated reflux to four cut-offs (100%, < 50%, < 20%, 0%) which were scrutinized separately. Thus, grouping into categories (ranges) and not relying on “exact” proportions should minimize inter-tester bias.

In conclusion, full patency on syringing was unreliable for ruling out NLD stenosis or functional delay. Hence, in patients with troublesome epiphora who are fully patent on syringing, imaging may determine the presence of NLD impairment and exclude those with normal systems which would be unlikely to benefit from intervention.

We believe that based on the results, any reflux should be considered an abnormal test in the context of epiphora. Finally, a positive syringing may be associated with functional NLD delay and cannot reliably differentiate it from stenosis, and this may have implications for consenting patients regarding success rates of intervention.