Symptoms attributable to lacrimal outflow obstruction were, according to the renowned ophthalmic historian Julius Hirschberg,¹ described in the papyrus documents of the ancient Egyptians. Lacrimal sac abnormalities were mentioned in the Talmud.² The term epiphora dates back to ancient Greece and is based on the Greek word epifora, which is in turn derived from the root words ferein (to bring) and epi (upon).³ Hirschberg⁴ noted that the school of Hippocrates recognized the relationship between epiphora and aging. Lacrimal sac abscess (agchiloph) and fistula formation (aigiloph) were reported by the ancient Greeks, and the Roman Celsus wrote about surgical management of these problems.⁵ Wafai⁶ wrote that the prominent Arabian ophthalmologist Ali Ibn Isa described dacryocystitis in his encyclopedic ophthalmic reference published 900 years ago.

The relationship between lacrimal fistula and underlying inflammation was reported by Stahl in Germany approximately 300 years ago.^7,8 Shortly thereafter, Dominique Anel of France ushered in the era of contemporary lacrimal evaluation utilizing lacrimal system probing and irrigation.⁹ Current surgical approaches to the correction of acquired lacrimal outflow obstruction date back to the descriptions of intranasal and external dacryocystorhinostomy in 1893 and 1904 by Caldwell¹⁰ and Toti,¹¹ respectively. In 1920, Dupuy-Dutemps and Bourguet¹² described their modified technique of external dacryocystorhinostomy (utilizing flaps of lacrimal sac and nasal mucosa), which remains a mainstay of treatment of acquired nasolacrimal duct obstruction today.

While ocular disorders including herpes simplex¹⁹ and/or treatments including glaucoma medications²⁰ have been associated with the development of lacrimal obstruction, the frequency of these potential contributory factors in a defined population of patients with lacrimal obstruction has not been determined. In addition, numerous systemic disorders,^21–23 including environmental allergy,²⁴ sarcoidosis,²⁵ Wegener granulomatosis,²⁶ and lymphoma,²⁷ may be associated with lacrimal obstruction, but the frequency of these associated disorders is similarly unclear.

• To define the age-specific incidence of SALOO in Olmsted County, Minnesota, during intervals from 1976–2000.
• To test the hypothesis that the incidence of SALOO has increased from the beginning of the study period (during the interval 1976–1979) to the end of the study period (during the interval 1995–2000).
• To describe the demographics and presenting signs and symptoms of SALOO in this population.
• To estimate the incidence of selected ocular disorders in this patient population, including glaucoma, herpes simplex and zoster virus infection, dry eye syndrome, ectropion, entropion, and ocular cicatricial pemphigoid.
• To determine the incidence of the selected ocular treatments in this patient population, including topical glaucoma medications, topical antiviral medications, and previous punctual occlusion.
• To determine the incidence of selected medical disorders in this patient population, including sinusitis, environmental allergy, sinonasal neoplasm or surgery, Wegener granulomatosis, sarcoidosis, asthma, hypertension, diabetes mellitus, leukemia, lymphoma, other systemic neoplasms, systemic chemotherapy, other systemic medications including hormonal therapy, and cigarette smoking.
• For patients with acquired dacryostenosis, to define location(s) of obstruction, treatment option(s) elected, and treatment outcomes.
• For patients undergoing dacryocystorhinostomy (DCR) surgery, to define risk factors for failure, including age, gender, use of topical medications, associated ocular or systemic disorders, and initial clinical presentation (epiphora, conjunctivitis, dacryocystitis, or a combination thereof).
• For patients undergoing failed DCR, to define the time course to failure.

From the epidemiologic viewpoint, each provider uses a unit medical record system whereby all data collected on an individual are assembled in one place. The Mayo Clinic unit record, for example, contains the details of every inpatient hospitalization at its 2 large affiliated hospitals, every outpatient visit to the office, clinic, or emergency room, every physician house or nursing home visit, every laboratory result, and correspondence. The Mayo Clinic system now contains over 6,000,000 histories, and the record storage system has allowed an extremely high percentage of record retrieval.

Utilization of the medical records has been facilitated by the maintenance at the Mayo Clinic of extensive indices based on clinical, radiologic, and histologic diagnoses and surgical procedures since the early 1900s. The Rochester Epidemiology Project has developed a similar index for the records of all other providers of medical care to Rochester and Olmsted County residents. These include the Olmsted Medical Group, the Olmsted Community Hospital, the University of Minnesota Hospitals and VA Medical Center in Minneapolis, community hospitals in the surrounding counties, and the few sole medical practitioners in Olmsted County. Through the Rochester Epidemiology Project, there is thus effective linkage of medical records from essentially all sources of medical care to the Olmsted County population. The Rochester Epidemiology Project has served as the basis for a large number of population-based studies.

Following preliminary case identification, the complete inpatient and outpatient medical record of each potential subject was obtained and reviewed to confirm the diagnosis of acquired lacrimal outflow obstruction. Diagnostic criteria included at least one of the following for each patient: a visibly elevated tear meniscus or visible epiphora, impaired fluorescein dye clearance, punctal stenosis, reflux upon lacrimal sac compression, acute dacryocystitis, abnormalities upon lacrimal probing or irrigation, and abnormalities on dacryocystography or dacryoscintigraphy. The year of initial diagnosis was used to establish the time period for incidence calculation purposes; each patient was therefore counted only once in this regard to ensure true calculation of incidence as opposed to prevalence.

Once the diagnosis of lacrimal obstruction was confirmed, each record was abstracted to ascertain age at diagnosis, signs and symptoms at presentation, pertinent past ocular history, past medical history, medical or surgical treatment, and subsequent clinical course, using the Data Form presented in Appendix B. Past ocular and past medical historical factors were listed as positive for a given patient if present at or before the time lacrimal obstruction was diagnosed. In this study, surgical success was defined as resolution or improvement of symptoms to the point that no further treatment would be contemplated as documented in the medical record.

Upon chart review, 525 charts did not meet the diagnostic criteria listed above for lacrimal obstruction or Olmsted County residency. Analysis of the remaining 587 charts served as the basis for the current study.

TABLE 1 STATISTICAL ANALYSIS OF INTERVAL INCIDENCE DIFFERENCES OF ALL FORMS OF DACRYOSTENOSIS DURING STUDY PERIOD

TABLE 9 SUMMARY AGE- AND GENDER-ADJUSTED ANNUAL INCIDENCE RATES FOR NASOLACRIMAL DUCT OBSTRUCTION, 1976–2000*†

TABLE 10 SITE(S) OF LACRIMAL OBSTRUCTION IN 587 PATIENTS

As seen in Figure 1, there was a statistically significant increase in incidence in diagnosis of all forms of lacrimal obstruction between the periods 1976–1980 and 1991–1995, with an apparent decrease in incidence during the period 1996–2000. The overall increase in incidence from 1976–1979 to 1996–2000 was, however, statistically significant (P=.01, Student t test). Statistical analysis of differences in incidence between other intervals during the study period is shown in Table 1. As shown in Figures 1 through 4, the incidence of all forms of lacrimal obstruction and of nasolacrimal duct obstruction alone was higher in females than males in all age ranges and throughout all time intervals during the study. The incidence of nasolacrimal duct obstruction and of all forms of lacrimal obstruction increased slowly beginning at age 40, with a more rapid rate of increase beginning at age 60. Over one-half of all patients with nasolacrimal duct obstruction in this study were older than the age of 65 years.

FIGURE 1 Adjusted annual incidence rates of all forms of dacryostenosis for 5-year intervals from 1976 through 2000.

FIGURE 4 Annual incidence rates of nasolacrimal duct obstruction for the period of 1976–2000 by age.

Two hundred nine of all patients (35.6%) demonstrated only left eye involvement, 176 (30%) only right eye involvement, and 202 (34.4%) bilateral involvement. One hundred fifty-seven (39.6%), 132 (33.3%), and 102 (27.2%) patients with nasolacrimal duct obstruction demonstrated left, right, and bilateral involvement, respectively. Ninety-eight patients (24.6%) with nasolacrimal duct obstruction were diagnosed with partial obstruction.

TABLE 11 FREQUENCY OF SELECTED OCULAR CONDITIONS PRECEDING OR NOTED AT TIME OF DIAGNOSIS OF LACRIMAL OBSTRUCTION IN 587 PATIENTS

TABLE 13 FREQUENCY OF SELECTED SYSTEMIC DISORDERS PRECEDING OR NOTED AT TIME OF DIAGNOSIS OF LACRIMAL OBSTRUCTION IN 587 PATIENTS

TABLE 14 FREQUENCY OF SELECTED SYSTEMIC NEOPLASMS PRECEDING OR NOTED AT TIME OF DIAGNOSIS OF LACRIMAL OBSTRUCTION

TABLE 15 SURGICAL TREATMENTS PERFORMED FOR MANAGEMENT OF LACRIMAL OBSTRUCTION

Kaplan-Meier survival curve analysis was performed for patients undergoing punctoplasty and DCR and is presented in Figures 5 and 6. An attempt was made to use a proportional hazards analysis to identify factors associated with DCR success. Neither age, gender, nor history of cigarette smoking appeared to be associated with differences in DCR success. There was a trend toward a higher success rate of surgery performed in patients with a history of acute dacryocystitis (with no surgical failures in 18 patients with such a history undergoing DCR) than in patients without a history of acute infection. This trend, however, did not achieve statistical significance (P=.30). It should be noted, however, that due to small numbers of surgical failures, this study lacked the statistical power to detect a difference between these 2 groups. It was not possible to calculate a risk ratio with confidence interval due to the lack of DCR failures in the acute dacryocystitis group. Similarly, due to the small number of DCR failures in this series, it was not possible to identify other factors that were associated with a higher risk of surgical failure. As noted in Figure 6, DCR failures in this series occurred within 3 years of surgery.

FIGURE 5 Kaplan-Meier survival curve analysis for punctoplasty success.

FIGURE 6 Kaplan-Meier survival curve analysis for dacryocystorhinostomy (DCR) success.

Lacrimal sac pathology studies were performed on 17 of 109 patients undergoing DCR surgery in this series. In 13 (76%) of these 17 patients, chronic inflammation was noted. Dacryolith formation was noted in 3 patients, and inverted papilloma was noted in 1 patient with previously diagnosed inverted papilloma involving the nasal cavity.

Two of 6 patients undergoing conjunctivodacryocystorhinostomy (CDCR) with Jones tube placement were considered to represent surgical successes after the initial procedure. Two additional patients underwent successful tube revision and replacement surgery, respectively. The remaining 2 patients were considered surgical failures and underwent tube removal.

Several investigators have studied the incidence of congenital nasolacrimal duct obstruction. Kendig and Guerry¹⁶ and Noda and colleagues¹⁷ noted incidence rates of 5.7% and 8% in series of American and Japanese infants, respectively, studied prospectively. These studies were performed on large consecutive series of neonates in the authors’ institutions, enhancing the accuracy of the resulting data. It is more difficult to determine the true incidence of acquired lacrimal obstruction, as this would entail prospective lacrimal evaluation of every individual in the population under study. The study performed by Dalgleish¹⁸ did entail lacrimal irrigation of every patient presenting for intraocular surgery at one hospital, but this population was not, therefore, necessarily representative of the population at large. In addition, this study did not provide information concerning the site(s) of obstruction or associated clinical symptoms.

An ideal study design to determine the incidence of both symptomatic and asymptomatic lacrimal obstruction would entail prospective evaluation of every patient in the population under study (including the performance of a history and physical examination incorporating lacrimal system irrigation and probing). The performance of lacrimal system irrigation and probing on asymptomatic patients for purely research purposes would not be approved according to current guidelines of the IRB at the author’s institution.

In view of the difficulties inherent in prospectively studying the incidence of lacrimal obstruction in the general population, we elected to study the incidence of symptomatic acquired lacrimal outflow obstruction (ie, lacrimal obstruction resulting in symptoms prompting medical evaluation). Even symptomatic acquired lacrimal outflow obstruction is challenging to study relative to congenital dacryostenosis due to difficulties in identifying a well-defined patient population in which disease incidence may be assessed.

A search by a certified medical reference librarian of multiple databases, including MEDLINE (back through 1957), EMBASE, Web of Science/Current Concepts, Journals@Ovid, and the Cochrane Library did not reveal previous population-based studies evaluating the incidence of symptomatic lacrimal obstruction in adult patients. This study revealed an overall age- and gender-adjusted incidence rate for all forms of symptomatic acquired lacrimal obstruction during the study period of 30.5 per 100,000 population (SE, 1.3; 95% confidence interval [CI], 28.0–33.0). The most common form of acquired dacryostenosis in this study, nasolacrimal duct obstruction, occurred with a frequency of 20.2 per 100,000 population (SE, 1.0; 95% CI, 18.2–22.3). The incidence of all forms of SALOO demonstrated a statistically significant increase (P = .01, Student t test) from the interval 1976–1979 to the interval 1996–2000.

As noted above, to the author’s knowledge there are no studies in other patient populations assessing the incidence of symptomatic lacrimal outflow obstruction in adult patients. It is of interest, however, to compare the incidence of SALOO to that of other ophthalmic disorders in the same population. Schoff and associates²⁸ noted an overall age- and gender-adjusted annual incidence rate of open-angle glaucoma of 14.5 per 100,000 population in the Olmsted County population during the period 1965–1980. The rates increased with age from 1.6 in the fourth decade of life to 94.3 in the eighth decade. There was no significant difference in glaucoma incidence by gender. Similarly, while it is difficult to estimate the incidence of cataract per se in the Olmsted County population during the defined time period, Baratz and colleagues²⁹ found the age- and gender-adjusted incidence of cataract extraction to be 470 per 100,000 population in Olmsted County in 1994. The incidence of all forms of acquired lacrimal obstruction and of nasolacrimal duct obstruction per se is thus somewhat higher than (but the same order of magnitude as) the incidence of glaucoma but significantly lower than the incidence of cataract extraction in the Olmsted County population.

This study also confirms the frequently made statement²⁰ that acquired lacrimal obstruction is most commonly encountered in middle-aged or older female patients. The mean and median ages of patients with all forms of obstruction were 61 and 67 years, respectively. In this study, 69% of patients with all forms of obstruction and 73% of patients with nasolacrimal duct obstruction were female. It has been speculated that the smaller diameter of the inferior bony lacrimal fossa and lacrimal canal in females may contribute to the increased prevalence of nasolacrimal duct obstruction in female patients.^30,31

As noted above, this study reveals an increase in the incidence of diagnosis of lacrimal obstruction from the interval 1976–1979 through the interval 1996–2000. It is difficult to determine whether this increase represented a true increase in incidence of obstruction, a greater tendency for patients to seek or be referred for medical evaluation of lacrimal obstruction as the study period progressed, more accurate diagnosis of obstruction, or some combination thereof. It is possible to comment superficially on at least the changes in potential availability of ophthalmic care during the study period. In specific, based on (1) Olmsted County population figures above age 5 years of 81,506³² in 1975 and 115,387³³ in 2000 and (2) practicing ophthalmologist numbers in Olmsted County of 14³⁴ and 34³⁵ in 1975 and 2000, the number of ophthalmologists per 10,000 serving this population increased 71%, from 1.72 to 2.94, over the study interval. It remains difficult, however, to determine the degree, if any, to which the increasing number of ophthalmologists per capita in Olmsted County over the study period would have contributed to an increased frequency of diagnosis of lacrimal obstruction in the absence of data regarding other factors that could have influenced the frequency of diagnosis. Such factors could have included improved education with increased awareness of the symptoms and signs of lacrimal obstruction among referring physicians and/or patients over the study period and a resultant increase in physician- or self-referral for ophthalmic evaluation. The decrease in apparent incidence of obstruction during the period 1996–2000 compared with the period 1991–1995 would be consistent with a plateau in those factors reflecting increased case capture (as opposed to a true increase in incidence), such as more active patient referral or enhanced diagnostic accuracy. Ongoing studies would be required to determine whether the rates noted during the 1996–2000 period represent true current baseline incidence rates.

Several ocular disorders, including herpes simplex and herpes zoster infection,^18,36,37 have been associated with acquired dacryostenosis. These disorders were noted in only 6 patients and 1 patient, respectively, in this series. Of the 6 patients with a history of herpes simplex infection, 3 had nasolacrimal duct obstruction, 2 had functional lacrimal outflow obstruction, and 1 had common canalicular obstruction. Idoxuridine, also associated with lacrimal obstruction,³⁸ was used by only 1 patient in this cohort. Over 5% of patients in this study had a history of glaucoma, with a number of patients using topical medications, including pilocarpine hydrochloride and echothiophate iodide, which have been implicated in lacrimal obstruction.¹⁹ Three patients had a history of prior punctal plug placement, which may also be associated with outflow obstruction.^39,40 This population-based study failed to reveal patients with lacrimal obstruction in association with ocular allergy, as has been previously reported.²³

Case series have documented lacrimal obstruction in the context of a number of systemic disorders. Conditions such as sinusitis⁴¹ and sinonasal surgery⁴² were reported in 4.43% and 1.70%, respectively, of patients in this study. Facial trauma is a well-established risk factor for lacrimal obstruction^43,44 and was encountered in 3.07% of patients in this series. Two patients in this series had sarcoidosis, another disorder associated with obstruction,²⁴ and a total of 8 patients had leukemia or lymphoma.²⁶ Several studies have documented the relationship between systemic 5-fluorouracil (5-FU) administration and dacryostenosis,^45,46 and a history of 5-FU treatment was obtained in 6 patients in this study. In all, approximately 10% of patients in this cohort had a history of systemic disorders or treatments that have been associated with lacrimal outflow obstruction. In the absence of other associated predisposing factors, approximately 90% of patients in this study may be considered to have primary acquired nasolacrimal duct obstruction, as described by Linberg and McCormick.⁴⁷ As noted in Table 13, there were a number of other systemic conditions occurring with a frequency exceeding 5% in patients, including asthma, hypertension, diabetes, cigarette smoking, hormonal replacement therapy, and various nonhematologic neoplasms. Assessment of the frequency of these disorders in an appropriate control population would be required to determine whether the coexistence of these disorders with lacrimal obstruction is more frequent than would be expected on the basis of chance alone.

From the outcomes perspective, the success rate of DCR in this study is comparable to that reported in numerous case series.^48–50 As noted above, due to the small number of failures, we were not able to identify specific ocular or systemic risk factors associated with surgical failure, although as noted above, there appeared to be a trend to a higher success rate in association with a history of acute dacryocystitis. The outcomes of and problems associated with CDCR with Jones tube placement in the small number of patients undergoing this procedure in this cohort were similar to those noted in other reports.⁵¹

It is important to note that this study has a number of potential limitations. First, case identification is based on accurate diagnosis-based coding and retrieval of individual patient records through the Rochester Epidemiology Project. Fortunately, the infrastructure of this National Institutes of Health–funded data resource is robust with a trained coding staff, allowing the Rochester Epidemiology Project to serve as the basis for over 1,500 publications since 1966.⁵² In this retrospective study, data analysis was based only on the information that was present in each record, and in some cases these data were incomplete. In addition, the generalizability of the results of this study is limited by the fact that the demographics of Olmsted County are not necessarily representative of the United States in its entirety. A detailed comparison of the demographic characteristics of Olmsted County residents and the US population in 1990 is presented in Table 16.⁵³ Based on the 2000 census, approximately 90% of Olmsted County residents were white,⁵⁴ compared with 69% for the United States in general.⁵⁵ The mean level of educational attainment⁵⁶ and median household income⁵⁷ is higher in Olmsted County than in the nation as a whole.^56,57 Furthermore, access to ophthalmic and lacrimal subspecialty medical care (which may facilitate the diagnosis of acquired lacrimal obstruction) in Olmsted County may differ from that available nationwide. In terms of comparisons between study participants and other Olmsted County residents, study participants were more likely to be female (69% vs 51%) and older (mean/median ages 61/67 years vs 35/35 years). It is possible that study participants differed from other Olmsted County residents in other systematic ways that this retrospective study was unable to capture. Nonetheless, this study provides at least initial relatively current incidence information regarding a clinically important ophthalmic disorder. While this study suggests that the incidence of SALOO has increased from 1975–2000, it is difficult to reliably predict the future incidence of this disorder based on the findings of this study. Based on the significantly increased incidence of acquired lacrimal obstruction in female patients above the age of 60 years, the generalized aging of the US population⁵⁸ and the increased preponderance of women in the elderly population,⁵⁸ however, it appears possible that this condition may become more prevalent in the future. This likelihood supports the importance of continuing to improve our understanding of new factors that may be associated with lacrimal obstruction, such as docetaxel administration.⁵⁹ The potentially increasing frequency of acquired lacrimal obstruction also supports the continued investigation of new, possibly less invasive options (including endonasal dacryocystorhinostomy⁶⁰ and balloon dacryoplasty⁶¹) for the management of this common disorder.

Funding/Support: None. Financial Disclosures: None.

Other Acknowledgments: Robert Kennedy, MD, PhD, MBA, and George Bartley, MD, provided valuable advice regarding study design and reviewed a near-final draft of the manuscript. James Garrity, MD, and Elizabeth Bradley, MD, provided helpful suggestions regarding study design. Won Chan Park, MD, assisted in data collection. David Hodge and Gabriela Rosales performed statistical analyses.