Nasolacrimal duct obstruction following radioactive iodine 131 therapy in differentiated thyroid cancers: review of 19 cases.

BACKGROUND: Radioactive iodine 131 ((131)I) therapy has long been used in the treatment of differentiated thyroid cancers (DTC). While salivary and lacrimal glandular complications secondary to (131)I therapy are well documented, there is little in the literature addressing nasolacrimal duct obstruction (NLDO). We aimed to evaluate the frequency of (131)I therapy-acquired NLDO, its correlation to (131)I therapy doses, and the surgical treatment outcome of this rare side effect.
METHODS: From 2000-2012, a retrospective review of 864 among 1,192 patients with confirmed DTC who were treated with (131)I therapy was performed to examine the frequency of NLDO, its causative factors, as well as imaging, surgical intervention, and outcomes.
RESULTS: Nineteen (2.2%) patients were identified with NLDO. The mean age was 51.9±10.5 years (range: 39-72 years). Fifteen (78.9%) were female and four were male (21.1%). The mean individual (131)I doses were 311.1±169.3 millicurie (mCi) (range: 150-600 mCi). The mean duration between the date of (131)I therapy and the occurrence of NLDO was 11.6±4.1 months (range: 6.5-20). Fourteen (73.7%) patients had bilateral epiphora. Computed tomography dacryography allowed for the detection of all NLDO. Eighteen (94.7%) patients underwent dacryocystorhinostomy. Complete recovery was obtained in 14 (73.7%) patients. Age >45 years and (131)I therapy doses >150 mCi were significantly correlated with NLDO (P=0.02 and P=0.03, respectively).
CONCLUSION: NLDO is an underestimated complication of (131)I therapy in DTC patients. Clinicians should be aware of this rare complication for prompt intervention.

Introduction

The use of post-thyroidectomy adjuvant radioactive iodine 131 (131I) therapy is associated with dramatic decreases in locoregional recurrences, distant metastasis, and disease-related mortality in patients with differentiated thyroid carcinomas (DTC).1 Although generally well tolerated, 131I therapy is associated with substantial, albeit rare, side effects. While gastrointestinal problems, salivary and lacrimal gland complications, gonadal dysfunction, and second malignancies are well documented side effects of 131I therapy, there is scant literature addressing nasolacrimal duct obstruction (NLDO).2,3 Different studies have mentioned that the incidence of NLDO was around 3% among DTC patients treated with 131I therapy. The common presentation is an excessive overflow of tears (epiphora). Other signs and symptoms are the medial canthus mass, recurrent conjunctivitis, or infection (dacryocystitis).4

In the Kingdom of Saudi Arabia, DTC has become the second most common malignancy, behind only breast cancer, accounting for more than 10% of all cancers among women.5 The majority of these patients require adjuvant 131I therapy at some point; only a small proportion of these patients can be at risk of 131I therapy-acquired NLDO, which was the aim of our study. We also aimed to determine the correlation between NLDO and 131I therapy along with other clinical and treatment parameters in DTC patients receiving 131I therapy in our population.

Methods

After formal approval from the institutional ethics committee, the medical records of 864 patients, among a total of 1,192 patients, with confirmed DTC who were treated with 131I therapy in two major tertiary care hospitals during the period from July 2000–December 2012 were reviewed using a computer-based institutional database system. Patients with NLDO were retrieved in the following manner:

Demographic information (age at diagnosis and sex), as well as data regarding symptomatology, clinical procedures (anterior rhinoscopy and Jones fluorescein dye test), and 131I therapy doses in millicurie (mCi) were collected.

Information on the different diagnostic imaging modalities used, including dacryoscintigraphy, computed tomography (CT) dacryography of the neck, and fluorodeoxyglucose positron emission tomography, was collected.

Data regarding the surgical intervention used, including balloon dilation, stenting, dacryocystorhinostomy (DCR), type of DCR (external or endoscopic endonasal), symptoms relief, and duration of relief were also recorded.

Any discrepancies in data acquisition were resolved by directly questioning the patients and treating physicians (ear, nose, and throat, oncology, ophthalmology, and nuclear medicine).

However, no attempt to systematize other associated sinonasal problems (potential confounders) of the patients was performed.

Statistical analysis

The primary endpoint was the frequency of NLDO. All descriptive data were reported as the mean ± standard deviation, and percentages. The secondary endpoint was the correlation of NLDO with 131I therapy (as tested by Pearson’s correlation coefficients). Efforts were made to minimize the effects of potential confounders, which were controlled using multivariate analysis. Statistical analyses were performed using the computer program SPSS version 16.0. The Kaplan–Meier method was used to predict the cumulative risk of NLDO in our series.

Results

Among the 864 DTC patients who were treated with 131I therapy, 19 (2.2%) were identified with NLDO (Table 1). No event of NLDO was seen in any patient who was untreated with 131I therapy or was treated with 131I therapy doses below 150 mCi. The mean age of the study population was 51.9±10.5 years (range: 39–72 years), and our population consisted of 15 females (78.9%) and four males (21.1%), with a female-to-male ratio of 3.7:1. The mean individual 131I dose was 311.1±169.3 mCi (range: 150–600 mCi). The mean duration between the date of 131I therapy and the occurrence of apparent NLDO was 11.6±4.1 months (range: 6.5–20 months).

Table 1

Patients’ characteristics

Patient	Age (years)/sex	Diagnosis	131I dose (mCi)	Symptoms	Onset from last 131I treatment (months)	DSG	CT dacryography	Treatment	Symptom control
1	49/F	PTCT3N1M0	300	Bilateral epiphora	8	+	+	E-DCR	Complete
2	41/F	PTCT3N1M1	600	Right eye epiphora	13	+	+	Observation	Partial
3	65/F	FTCT3N1M1	400	Bilateral epiphora	7	+	+	E-DCR	Complete
4	53/F	PTCT4N1M0	150	Right eye epiphora	10	+	+	E-DCR	Complete
5	69/F	FTCT2N0M1	450	Bilateral epiphora	12	+	+	E-DCR	Complete
6	47/F	PTCT4N0M0	150	Left eye epiphoraBlurring of vision	11	+	+	EE-DCR	Complete
7	51/F	PTCT3N1M1	550	Bilateral epiphora	20	+	+	EE-DCR	Partial
8	39/F	PTCT2N1M0	150	Bilateral epiphora	6.5	−	+	E-DCR	Complete
9	47/M	PTCT3N1M0	300	Bilateral epiphoraRight eye dacryocystitis	12.5	+	+	E-DCR	Complete
10	72/F	PTCT3N1M0	300	Right eye epiphora	8	+	+	E-DCR	Complete
11	42/F	PTCT3N1M1	550	Bilateral epiphora	15	+	+	E-DCR	Partial
12	42/M	PTCT2N1M0	150	Left eye epiphora	7.3	−	+	E-DCR	Complete
13	58/F	PTCT2N1M1	210	Bilateral epiphora	14	+	+	E-DCR	Complete
14	45/M	PTCT3N1M1	350	Bilateral epiphora	14	+	+	E-DCR	Complete
15	46/F	PTCT2N1M0	300	Bilateral epiphora	9	+	+	E-DCR	Partial
16	66/F	PTCT3N1M0	300	Bilateral epiphora	12	−	+	EE-DCR	Complete
17	39/F	PTCT2N1M1	200	Bilateral epiphora	6.5	+	+	E-DCR	Complete
18	56/M	PTCT2N1M1	600	Bilateral epiphora	19	+	+	Left eye observationRight eye E-DCR	Complete
19	60/F	PTCT4N0M1	600	Bilateral epiphora	17	+	+	E-DCR	Complete

Abbreviations: 131I, radioactive iodine 131; mCi, millicurie; DSG, dacryoscintigraphy; CT, computed tomography; F, female; PTC, papillary thyroid carcinoma; E-DCR, external dacryocystorhinostomy; FTC, follicular thyroid carcinoma; EE-DCR, endoscopic endonasal dacryocystorhinostomy; M, male.

Fourteen (73.7%) patients had bilateral epiphora. Dacryoscintigraphy allowed for the detection of 16 NLDO with a sensitivity of 84.2% (95% confidence interval [CI]: 60.4–96.4) and a specificity of 99.6% (95% CI: 98.9–99.9), while CT dacryography detected all cases of NLDO with a sensitivity of 100% (95% CI: 82.2–100) and a specificity of 100% (95% CI: 82.2–100).

DCR was performed in 32 completely stenosed nasolacrimal ducts of 18 (94.7%) patients – either bilaterally (63.2%) or unilaterally (26.3%). Three patients (15.8%) had endoscopic endonasal DCR. Observation was planned for one patient; but the patient had partial relief and was considered for DCR. Among the DCR patients, complete recovery was obtained in 14 (73.7%) patients, and partial relief was obtained in three (15.8%) patients. The success rates of external DCR and endoscopic endonasal DCR were 85.7% and 66.7%, respectively.

A significant correlation between NLDO and 131I therapy doses above 150 mCi was seen (R=0.92; R2=0.86; P=0.03). Multivariate analysis showed that 131I therapy doses above 150 mCi and an age >45 years were important prognosticators (Figure 1).

Figure 1

Kaplan–Meier curve of the cumulative hazard risk of 131I-acquired nasolacrimal duct obstruction.

Abbreviation: 131I, radioactive iodine 131.

Discussion

Acquired NLDO secondary to 131I therapy is an extremely rare complication in patients with DTC. In our cohort, the frequency of NLDO of 2.2% is consistent with other reported results (Table 2).6–10 The incidence rate of NLDO in the general population is lacking, despite a detailed medical literature search. In the majority of cases, the diagnosis is made upon the basis of spontaneous reporting of epiphora by the patients.6 This diagnostic delay leads to complete stenosis of the nasolacrimal ducts, and thus renders these patients to aggressive treatment.6 In the clinical setting, the Jones fluorescein dye test is of significant value in NLDO patients. One Australian study has demonstrated that a negative single-drop Jones fluorescein dye test is also predictive of symptomatic improvement after DCR in patients with NLDO.11

Table 2

Different case studies and reports of 131I therapy-acquired NLDO reported in the literature

Study	NLDO N	Total patients treated with 131I therapy	Age (years)	Mean individual 131I dose (mCi)	Onset from last131I therapy (months)	Symptoms	Treatment	Symptoms relief
Kloos et al3	10 (2.6%)	390	47.8 (range: 13–74)	180±15	6.5±1.4 (range: 3–16)	Epiphora, bilateral (50%)Epiphora, unilateral (50%)	Observation (30%)Balloon dilation (30%)DCR (40%)	Partial (60%)Complete (40%)
Sun and Hatipoglu4	1	–	23	150	6	Epiphora, bilateral	DCR	Complete
Fonseca et al6	17	–	30–80	571 (range: 200–1,200)	13.2 (range: 4–30)	Epiphora, bilateral (64.7%)Epiphora, unilateral (35.3%)	DCR (100%)	Complete (82.4%)Partial (17.6%)
Shepler et al7	1	–	50	150	6	Epiphora, unilateral	Observation (5 years),DCR	Complete
Burns et al8	26 (4.6%)	563	–	525 (range: 200–951)	13	–	Balloon dilation, DCR	–
Brockmann et al9	1	–	54	700	6	Epiphora, bilateral	DCR	Complete
Savage et al10	2	–	–	–	–	Epiphora, bilateral	–	–
Present study	19 (2.2%)	864	51.9 (range: 39–72)	311.1±169.3	11.6±4.1 (range: 65–20)	Epiphora, bilateral (73.7%)Epiphora, unilateral (26.3%)	Observation (5.3%)DCR (94.7%)	Partial (21.1%)Complete (78.9%)

Note: Data are presented as mean or mean ± standard deviation, except where otherwise indicated.

Abbreviations: 131I, radioactive iodine 131; NLDO, nasolacrimal duct obstruction; N, number; mCi, millicurie; DCR, dacryocystorhinostomy.

The postulated mechanism for 131I therapy-acquired NLDO is the direct uptake of 131I in the nasolacrimal duct membrane epithelial cells via the sodium (Na+)/iodide (I) symporter (NIS) leading to radiation-induced cell damage (inflammation and fibrosis/stenosis).9 The NIS is a membrane glycoprotein that mediates 131I uptake in the thyroid gland and several other extrathyroid tissues (salivary glands, lacrimal glands, and ciliary body of the eye).12 Recently, Morgenstern et al13 have reported the overexpression of NIS in the nasolacrimal ducts using reverse transcriptase polymerase chain reaction and immunohistochemical analyses. In addition to the presence of sinonasal disease at the time of 131I therapy, the treatment may theoretically further decrease the flow in the nasolacrimal ducts and thus increase the exposure to 131I, which is secreted through the lacrimal system.3

The diagnosis of 131I therapy-acquired NLDO is challenging. Other possible causes (allergic or viral conjunctivitis, keratoconjunctivitis sicca, trichiasis, canaliculitis, iritis, and functional causes of excessive tears) should be ruled out to reach a definitive diagnosis of 131I therapy-acquired NLDO.14 131I therapy-acquired NLDO is defined as: 1) the presence of NLDO in 131I therapy-treated patients; 2) an 131I dose–NLDO relationship; 3) immediate onset of symptoms after 131I therapy; and 4) the bilateral nature of NLDO in most of cases. CT dacryography using diluted contrast medium was found to be an easy and highly sensitive tool to evaluate NLDO in our cohort, which endorsed the findings of Garcier et al.15

Patients with incomplete NLDO have an excellent outlook with the use of balloon dilation and stenting, as complete recovery of symptoms is observed in 100% of cases.3 Although the literature supports the hypothesis of complete symptom recovery with observation alone, in a few cases of incomplete NLDO, we did not find that observation was a feasible option. The reason for this can be explained by the fact that NLDO is possibly a chronic radiation-induced injury that progresses over a period of time.16 The DCR success rate of 78.9% in our cohort is consistent with the findings of Fonseca et al.6 However, the endoscopic endonasal DCR success rate of 60% is much inferior to that reported in the literature, which warrants more personnel training in this area in our region.17

Because NLDO impairs the quality of life of a patient significantly, prevention of NLDO is a key issue. Currently, there is no mechanism available to stop the blockage of 131I uptake in the nasolacrimal ducts via the NIS during 131I therapy,3 and there is scant data available regarding the prophylactic use of 2-(S)-(3-aminopropylamino) ethylphosphorothioic acid (amifostine) during 131I therapy.18 Further, the effectiveness of topical saline drops, nasolacrimal duct massage, and delaying 131I therapy in patients with acute or chronic sinonasal problems as preventive measures needs to tested. Low 131I therapy doses, whenever possible, can reduce the incidence of NLDO.19

The limitations of our study were: 1) no systematic attempt to screen for NLDO or the presence of other potential confounders (sinonasal problems) was performed; and 2) the presence of possible selection and recall bias in our cohort.

Conclusion

In conclusion, NLDO following 131I therapy in DTC patients is an under-recognized and rare complication, especially at higher 131I doses (.150 mCi). A multidisciplinary approach (nuclear medicine/oncology, ear, nose, and throat, and ophthalmology) is very important in the prevention and treatment of NLDO.