Risk of Ocular Adverse Events With Taxane-Based Chemotherapy

This study aims to quantify the risk of epiphora, cystoid macular edema, and optic neuropathy with taxane-based chemotherapy agents by undertaking a large pharmacoepidemiologic study.

Key Points

Question

Is the use of taxane-based chemotherapy associated with an increased risk for epiphora, cystoid macular edema (CME), and optic neuropathy?

Findings

In this cohort study, those using taxane-based chemotherapy, such as paclitaxel and docetaxel, were at an increased risk of epiphora, CME, and optic neuropathy when compared with tamoxifen users.

Meaning

Results of this study suggest that individuals who are prescribed paclitaxel or docetaxel might be at a higher risk of epiphora, CME, and optic neuropathy; ophthalmologists and oncologists should suspect these drugs as potential culprits for these adverse events in patients with cancer who are receiving these drugs.

Abstract

Importance

Taxane-based chemotherapy agents, such as docetaxel and paclitaxel, are used for treating a wide range of cancers. Although much has been published on adverse events related to taxanes, data on ocular outcomes with these very important drugs are scant.

Objective

To quantify the risk of 3 mutually exclusive ocular adverse events of epiphora, cystoid macular edema (CME), and optic neuropathy with taxane-based chemotherapy agents by undertaking a large pharmacoepidemiologic study.

Design, Setting, and Participants

This retrospective cohort study design used a private health-claims database from the US that captures health information of more than 150 million enrollees. The study team created a cohort of new users of women with cancer who were taking taxane-based chemotherapy (docetaxel or paclitaxel) and new users of tamoxifen as controls. Study members were observed to the first incidence of each of the 3 mutually exclusive outcomes. An analysis of taxane-only users was also undertaken.

Exposure

Tamoxifen (unexposed) and taxanes (ie, paclitaxel and docetaxel) as the exposed.

Main Outcomes and Measures

First diagnosis of (1) epiphora, (2) cystoid macular edema (CME), or (3) optic neuropathy ascertained using International Statistical Classification of Diseases and Related Health Problems, Ninth Revision or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision.

Results

Among the 18 219 users in the epiphora analysis and optic neuropathy analysis, there were 1824 taxane users (paclitaxel and docetaxel) (age, mean [SD], 62.1 [12.7] years) and 16 395 tamoxifen users (age, mean [SD], 54.6 [12.8] years), respectively. The crude hazard ratio (HR) for epiphora was 5.55 (95% CI, 2.99-10.29) and adjusted HR was 5.15 (95% CI, 2.79-9.54). For optic neuropathy, the crude HR was 4.43 (95% CI, 1.10-17.82) and the adjusted HR was 4.44 (95% CI, 1.04-18.87). Among the 18 433 users in the CME analysis, there were 1909 taxane users (paclitaxel and docetaxel) (age, mean [SD], 62.5 years) and 16 524 tamoxifen users (age, mean [SD], 54.6 years). The crude HR for CME comparing taxane users with tamoxifen users was 1.37 (95% CI, 0.72-2.60) and adjusted HR was 1.33 (95% CI, 0.70-2.53). The HRs for epiphora and CME in the taxane cohort during the time of exposure compared with the period prior to use of the drugs were 2.86 (95% CI, 1.11-7.39) and 2.27 (95% CI, 0.68-7.54), respectively.

Conclusions and Relevance

In a cohort of women who were using taxane chemotherapy agents, there was an association with elevated risk for epiphora, CME, and optic neuropathy. Ophthalmologists and oncologists should be aware of these adverse events in women with breast cancer who receive these drugs.

Introduction

Chemotherapy remains the mainstay therapy for most cancers, including breast cancer, lung cancer, genitourinary cancers, and head and neck cancers.¹ Although much is known about adverse events related to taxanes, such as docetaxel and paclitaxel, information on the frequency of ocular adverse events, including their incidence, is scant. A number of ocular adverse events secondary to paclitaxel or docetaxel usage have been published, which have mainly focused on (1) epiphora,^2,3,4 (2) cystoid macular edema (CME),^5,6 and (3) optic neuropathy.⁷ Epiphora is a significantly debilitating condition that may be irreversible and greatly affects quality of life, in particular outdoor activities,⁸ and research has shown that if treatable, correction of epiphora can vastly improve overall health and quality of life.⁹ CME and optic neuropathy are also serious adverse events that can significantly decrease vision and quality of life, and their prompt diagnosis and treatment can affect patient outcomes. Most of the data on these 3 adverse events with paclitaxel and docetaxel usage have been mainly reported as case reports, which cannot quantify a true incidence or provide an effect size as to the magnitude of these events.

In 2020, more than 2 million women around the world were diagnosed with breast cancer.¹⁰ Approximately 60% of women with breast cancer are diagnosed with invasive breast cancer, which usually requires chemotherapy that includes paclitaxel or docetaxel.¹¹ Therefore, ophthalmologists and oncologists need to be familiar with the frequency and type of ocular adverse events secondary to taxane chemotherapy. Recognizing these ocular adverse events early will potentially prevent complications and improve patient outcomes. Therefore, we sought to quantify the risk of these adverse events by undertaking a large pharmacoepidemiologic study.

Methods

Data Sources

We used the PharMetrics Plus database (IQVIA) as the primary data source for the study. The PharMetrics database is a private health-claims database from the US that captures health information of more than 150 million enrollees in the US.¹² The data captured included demographics (including the age and sex of patients), physician diagnosis (through International Statistical Classification of Diseases and Related Health Problems, Ninth Revision [ICD-9] or International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10]), and prescription drug information (drug name, day supply, and dosage). We had access to a random sample of this database from 2006 to 2020 on 16 424 887 enrollees who had at least 1 year of enrollment. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational cohort studies were followed. Approval for the study was obtained by the University of British Columbia’s Clinical Research Ethics Board, which said individual patient consent was not needed as this was a health-claims database study.

Cohort Description

We created a cohort of women who were new users of taxanes defined as use of docetaxel or paclitaxel, the 2 most prescribed taxanes in women. New users were defined as those who were not using either drug and did not have any of the study outcomes in the year prior to cohort entry. We also included women who were new users of tamoxifen for the treatment of estrogen-positive breast cancer who were not using paclitaxel, docetaxel, or fluorouracil, another chemotherapy drug known to increase the risk of epiphora. Tamoxifen was chosen as a reference drug because it is used for similar indications (breast cancer) as taxanes, and it has shown to increase the risk of CME¹³ and optic neuropathy,¹⁴ although it has not been associated with epiphora. However, since the mechanism of action of taxanes and tamoxifen are different, we decided to use tamoxifen as a comparator against taxanes and compare the ocular safety of taxanes with that of paclitaxel or docetaxel. Cohort members were followed up to the first of the following mutually exclusive events: (1) epiphora, which also included stenosis of lacrimal canaliculi and obstruction of nasolacrimal ducts (ICD-9 codes 375.20, 375.53, and 375.55 and ICD-10 codes H04.54 and H04), (2) cystoid macular edema (ICD-9 codes 362.53 and 362.83 and ICD-10 codes H35.351, H35.352, H35.353, and H35.81), and (3) optic neuropathy (ICD-9 codes 377.34 and 377.41 and ICD-10 codes H47.013, H47.012, H47.011, H47.019, and H46.3). Since current data show that taxanes might cause both ischemic optic neuropathy and toxic optic neuropathy,⁷ both types were included. We also censored users who switched from a taxane to tamoxifen or vice versa.

Statistical Analysis

We used descriptive statistics to examine demographics and covariate distribution between the exposed groups (paclitaxel or docetaxel) and controls (tamoxifen users). Our study was controlled by age and sex (the 2 main confounders for these study questions). Age was stratified and sex was controlled for by design. We further constructed a Cox regression model to compute crude and adjusted hazard ratios (HRs) for the 3 mutually exclusive outcomes comparing docetaxel or paclitaxel users with tamoxifen users. For the model looking at epiphora, we adjusted for the following covariates¹⁵ (mainly risk factors), keratoconjunctivitis, dry eye syndrome, keratitis, blepharitis, and lid malposition, and the following aromatase inhibitors, anastrozole, exemestane, and letrozole. For the model looking at CME, we adjusted for previous cataracts and diabetes.¹⁶ For the model examining optic neuropathy, we also adjusted for hypertension, coronary artery disease, diabetes, and number of ophthalmologist visits.¹⁷ We computed incident rates and adjusted hazard ratios for all 3 outcomes. To further control for confounding by cancer severity, we undertook 2 additional analyses. First, we restricted the cohort to only women with a diagnosis of breast cancer. Second, we performed an all-taxanes cohort analysis. We followed up the cohort during the exposed period (after exposure to PTS or docetaxel) and censored patients if they experienced 1 of the 3 mutually exclusive outcomes or existed in the database. For the unexposed period, we followed up the users in a similar fashion during from the time of the first paclitaxel or docetaxel to cohort entry time. Age-stratified HRs were computed for the 3 mutually exclusive outcomes in the 1 year after paclitaxel or docetaxel compared with the year prior to use. All analyses were performed using SAS (version 9.4; SAS Institute).

Results

For the epiphora analysis, there were 1824 taxane users (paclitaxel or docetaxel) (age, mean [SD], 62.1 [12.7] years) and 16 395 tamoxifen users (age, mean [SD], 54.6 [12.8] years) (Table 1). The incidence of epiphora among taxane users was 55.6/10 000 person-years vs 7.9/10 000 person-years for tamoxifen users (Table 1). The crude HR for this association was 5.55 (95% CI, 2.99-10.28) with an adjusted HR of 5.15 (95% CI, 2.79-9.54) (Table 2). For the CME analysis, there were 1909 taxane users and 16 524 tamoxifen users (Table 1). The incidence of CME among taxane users was 34.8/10 000 person-years vs 16.8/10 000 person-years users in the tamoxifen users (Table 1). The crude HR for CME comparing taxane users with tamoxifen users was 1.37 (95% CI, 0.72-2.60) and the adjusted HR was 1.33 (95% CI, 0.70-2.53) (Table 2). For optic neuropathy, there were 1913 taxane users and 16 566 tamoxifen users (Table 1). The incidence of optic neuropathy for taxane users was 10.6/10,000 person-years and 1.2/10 000 person-years for tamoxifen users (Table 1). The crude HR was 4.43 (95% CI, 1.10-17.82) and the adjusted HR was 4.44 (95% CI, 1.04-18.87) (Table 2). When the study team restricted the analysis to only women with breast cancer, results were similar (Table 3). The adjusted HRs for epiphora, CME, and optic neuropathy were 6.42 (95% CI, 3.13-13.15), 1.37 (95% CI, 0.57-3.31), and 10.73 (95% CI, 1.07-107.63), respectively (Table 3). Among the taxane users in the epiphora analysis, 60% had a breast cancer diagnosis, 11.7% had a uterine cancer diagnosis, 15.6% had a ovarian cancer diagnosis, and 12.7% had diagnoses for other cancers.

Table 1. Characteristics of Tamoxifen and Docetaxel/Paclitaxel Users.

Characteristic	Tamoxifen	Docetaxel/paclitaxel
Epiphora
No.	16 395	1824
Age, y, mean (SD)	54.6 (12.8)	62.1 (12.7)
Follow-up, y, mean (SD)	3.5 (3.1)	2.0 (1.9)
Events, No.	45	20
Rate	7.9/10 000 person years	55.6/10 000 person years
Keratoconjunctivitis, %	0.05	0.11
Dry eye syndrome, %	0.29	2.25
Keratitis, %	2.07	3.95
Blepharitis, %	2.16	4.61
Eyelid malposition, %	0.55	1.35
Cystoid macular edema
No.	16 524	1909
Age, y, mean (SD)	54.6 (12.8)	62.5 (12.8)
Follow-up, y, mean (SD)	3.5 (3.1)	2.0 (1.9)
Events, No.	96	13
Rate	16.8/10 000 person years	34.8/10 000 person years
Cataract, %	0.6	2.1
Diabetes, %	8.0	14.3
Optic neuropathy
No.	16 395	1824
Age, y, mean (SD)	54.6 (12.8)	62.5 (12.8)
Follow-up, y, mean (SD)	3.5 (3.1)	2.0 (1.9)
Events, No.	7	4
Rate	1.2/10 000 person years	10.6/10 000 person years
Hypertension, %	6.0	38.2
Diabetes, %	9.9	17.4
Coronary artery disease, %	5.7	18.6

Table 2. Hazard Ratios (HRs) for Cases of Epiphora, Cystoid Macular Edema, and Optic Neuropathy With Exposure to Tamoxifen and Docetaxel/Paclitaxel.

Characteristic	HR (95% CI)
	Crude	Adjusted
Epiphora
Tamoxifen	1 [Reference]	1 [Reference]
Docetaxel/paclitaxel	5.55 (2.99-10.28)	5.15 (2.79-9.54)a
Cystoid macular edema
Tamoxifen	1 [Reference]	1 [Reference]
Docetaxel/paclitaxel	1.37 (0.72-2.60)	1.33 (0.70-2.53)b
Optic neuropathy
Tamoxifen	1 [Reference]	1 [Reference]
Docetaxel/paclitaxel	4.33 (1.10-17.82)	4.44 (1.04-18.87)c

^a Adjusted for age, keratoconjunctivitis, dry eye syndrome, keratitis, blepharitis, and eyelid malposition.

^b Adjusted for age, diabetes, and cataracts.

^c Adjusted for age, diabetes, hypertension, and coronary artery disease.

Table 3. Risk of Epiphora, Cystoid Macular Edema, and Optic Neuropathy Among Women With Breast Cancer Taking Tamoxifen or Taxane Chemotherapy.

Characteristic	HR (95% CI)
	Crudea	Adjusted
Epiphora
Tamoxifen	1 [Reference]	1 [Reference]
Docetaxel/paclitaxel	6.60 (3.30-13.21)	6.42 (3.13-13.15)
Cystoid macular edema
Tamoxifen	1 [Reference]	1 [Reference]
Docetaxel/paclitaxel	1.45 (0.60-3.48)	1.37 (0.57-3.31)
Optic neuropathy
Tamoxifen	1 [Reference]	1 [Reference]
Docetaxel/paclitaxel	11.18 (1.34-93.35)	10.73 (1.07-107.63)

Abbreviation: HR, hazard ratio.

^a Hazard ratios were adjusted for aromatase inhibitors (anastrozole, exemestane, and letrozole), keratoconjunctivitis, dry eye syndrome, keratitis, blepharitis, eyelid malposition, cataracts, diabetes, and coronary artery disease.

The HRs for epiphora and CME in the taxane-only cohort during the time of exposure compared with the period prior to use of the drugs were 2.86 (95% CI, 1.11-7.39) and 2.27 (95% CI, 0.68-7.54). There were only 3 cases of optic neuropathy after using taxanes compared with 0 cases in the period before use of these drugs; thus, an HR was not computed.

Discussion

The results of our study suggest that taxane-based chemotherapy agents are associated with an increased risk of epiphora and optic neuropathy when compared with a cohort of women taking tamoxifen and not taking taxanes. There was also an association with elevated risk with CME, but results were not precise and 95% CIs were wide, ranging from a 30% protective effect to 153% increase in risk (HR, 1.33; 95% CI, 0.70-2.53). The taxanes-only analysis and stratification among patients who had breast cancer only showed similar results.

Our study results are consistent with published case reports that have suggested an increase in the risk of these 3 important ocular adverse events.^2,5,8 The survival benefits of taxane-based chemotherapeutic agents have made these agents an integral part of drug regimens for many cancers, especially breast, ovarian, and uterine cancers. Although treatment for CME and optic neuropathy secondary to taxanes might be limited, results from this study can better inform oncologists and ophthalmologists on the risk of these 3 ocular outcomes, and where possible, allow them to choose alternative agents.

There is a paucity of evidence surrounding the potential pathophysiological mechanism of taxane-induced optic neuropathy. It has been hypothesized that the systemically administered taxanes can cause neurotoxic or ischemic damage to the retinal ganglion cells.⁷ Other research has shown that taxanes can induce maladaptive electrophysiological changes within the retina and optic nerve, which were attributed to vasospasm or vascular dysregulation.¹⁸ This hypothesis is supported by findings from case studies such that the patient presentation of taxane-induced optic neuropathy resembles that of interferon alfa therapy, which is also believed to be caused by ischemia and vascular dysregulation.⁷ Furthermore, since taxanes are known to be neurotoxic and can cause peripheral neuropathy, it is possible that the optic nerve is damaged via a similar mechanism.¹⁹

The proposed mechanism for taxane-induced epiphora is canalicular stenosis. The drug is secreted in the tear film and chronic inflammation of the canaliculi, as these tears are drained through the nasolacrimal system, results in scarring and narrowing of the passageway.^2,20 However, epiphora has also been shown to occur in the absence of canalicular stenosis and is believed to be secondary to dry eye disease caused by taxane-based chemotherapy.² The use of artificial tears may not only serve to dilute and clear the drug from the ocular surface through the nasolacrimal system, but may treat any related dry eye disease. Early recognition and discontinuation of the drug, if possible, may prevent surgical intervention.²⁰ In some cases, early detection of symptoms can lead to less invasive interventions, such as insertion of canalicular silicone stents, which can mitigate complete closure of canaliculi.

It has been postulated that taxane-induced CME can result from a disruption of the retina-blood barrier and can lead to fluid accumulation in intracellular spaces.²¹ Some have also proposed that taxanes can cause retinal cell toxicity due to taxanes’ inherent mechanism of inhibiting microtubular reorganization.²¹

There are more than 2 million women worldwide who are experiencing breast cancer¹⁰ and more than 60%¹¹ of these women are likely undergoing taxane-based chemotherapy. There is projected to be nearly $2.9 billion²² in sales of taxanes, indicating that use of these drugs will greatly increase over the next few years for both breast and other types of cancers. Therefore, the ocular adverse events studied will become more significant in the near future.

Strengths and Limitations

Our study has strengths and limitations. To our knowledge, this was the first and largest study of women who used a taxane (with aged-matched controls using tamoxifen) that examined 3 important ocular adverse events. The strengths of our study are the comparator group and adjustment for other confounders or risk factors. Although most taxane usage is for women with breast cancer, it is possible that they were used for other types of cancer. However, we do not believe different types of cancers would alter the risk of these 3 ocular adverse events. Moreover, the analysis that examined the risk among taxane users only also showed an increase in the risk with epiphora and CME. Taxane users could have also been using other chemotherapeutic agents, as cancer treatment usually involves multiple chemotherapeutic drugs. However, epiphora and CME have mainly been identified with taxanes.^2,3,5 It is possible for other chemotherapy agents combined with taxanes to also cause optic neuropathy. Additionally, our study is only generalizable to women. As with all pharmacoepidemiologic studies that use claims data, we did not have access to patients’ medical records for outcome validation.

Conclusion

In a cohort of women who were using taxane chemotherapy agents, there was an association with elevated risk for epiphora and optic neuropathy, and an association with an increase in the risk for CME (a 33% increase with wide 95% CIs that span from 0.70 to 2.53). Ophthalmologists and oncologists should be aware of these adverse events in women with breast cancer who use these drugs. Early identification of such events may help in treatment of complications and lead to better outcomes and quality of life. Future studies should confirm these results.