Ocular side-effects in breast cancer patients treated with tamoxifen and toremifene: a randomized follow-up study

Authors


Dr Minna Parkkari
Department of Ophthalmology
Tampere University Hospital
PO Box 2000
33501 Tampere
Finland
Tel: + 358 3 247 51 11
Fax: + 358 3 247 43 65
Email: minna.parkkari@uta.fi

Abstract.

Purpose: 3Tamoxifen and toremifene are non-steroidal anti-oestrogens widely used in the treatment of advanced breast cancer and as adjuvant therapy following surgery in early stage disease. Tamoxifene has also been approved for use in reducing the incidence of breast cancer amongst high risk women. However, certain well documented adverse effects, mainly involving the reproductive organs, have been reported amongst users of both drugs. The aim of this study was to monitor the ocular side-effects of both of these commonly used anti-oestrogens.

Methods: Sixty postmenopausal (age range 50–79 years) breast cancer patients were randomized into adjuvant tamoxifen or toremifene therapy groups for 3 years. Prior to commencement of medication, a thorough ocular examination was undertaken. The first follow-up visit took place after 6 months and the remaining three at 12-month intervals thereafter.

Results: Sixteen patients had cataract at the first visit (seven in the tamoxifen group and nine in the toremifene group). Ten patients developed cataract during the study period (five in each group), giving annual cataract rates of 6.8% and 6.2% in the tamoxifen and toremifene groups, respectively. Three patients had macular crystals at the first visit (one in the tamoxifen group and two in the toremifene group). The crystals remained stable throughout the follow-up. Macular drusen were diagnosed in five patients at the first ophthalmological check-up (two in the tamoxifen and three in the toremifene group). Two patients in the toremifene group developed drusen maculopathy during follow-up visits. Yellowish spots in the macular area were found in one tamoxifen-treated patient at the second visit. At the final visit after 3.5 years' follow-up the spots had disappeared. No abnormal corneal findings or keratopathy were documented during the follow-up.

Conclusion: We observed no serious ocular side-effects among the 60 breast cancer patients treated with tamoxifen or toremifene over a 3.5-year period.

Introduction

Tamoxifen is a non-steroidal anti-oestrogen commonly used in the treatment of advanced breast cancer and as adjuvant therapy following surgery in early stage disease (Kramer et al. 1993). Its use has also recently been approved for reducing the incidence of breast cancer amongst high risk women (Fisher et al. 1998).

Tamoxifen is in general a well tolerated hormonal agent without serious side-effects. However, certain well documented adverse effects, mainly involving the reproductive organs (the greatest source of concern being carcinogenicity for the endometrium), have been reported (International Agency for Research on Cancer 1996). The first cases of ocular toxicity due to tamoxifen were reported by Kaiser-Kupfer & Lippman (1978) and referred to women receiving extremely high doses of tamoxifen (240–320 mg/day) for metastatic breast cancer. Since then, a number of studies have suggested that the use of regular, low dose tamoxifen (20–40 mg/day) may also be associated with abnormalities in visual function (visual acuity) and ocular structures (refractile crystalline deposits in the retina, macular oedema, corneal opacities, lens changes and optic neuritis) (Nayfield & Gorin 1996).

Toremifene is a newer anticancer drug. It is closely related to tamoxifen but differs from it in the substitution of a chlorine atom for a hydrogen atom in the ethyl group (Pappas et al. 1999). For advanced breast cancer and in an adjuvant setting, toremifene has been found to be as effective and well tolerated as tamoxifen (Holli et al. 2000; Mäenpää et al. 2000).

The pathophysiology of retinal changes associated with tamoxifen and toremifene is unclear. Kaiser-Kupfer et al. (1981) suggested that the formation of crystalline retinal deposits in tamoxifen users may be related to axonal degeneration. The intracellular location of the retinopathic lesions in the nerve fibre and inner plexiform layers of the retina, their similarity to nerve synapses as seen in electron microscopy, and the demonstration by histochemical methods of glycosaminoglycans in the deposits support this hypothesis. It has also been postulated that tamoxifen binds with polar lipids, inhibiting normal catabolism of the lipids and resulting in the accumulation of drug−lipid complexes in lysosomes. A generalized lipidosis has been demonstrated in rats treated subchronically with very high dose tamoxifen (Lullmann & Lullmann-Rauch 1981).

The toxicity of tamoxifen and toremifene has been investigated in the human retinal pigment epithelium in vitro. Tamoxifen and toremifene have been shown to reduce the activities of two lysosomal enzymes, N-acetyl-β-glucosaminidase and cathepsin D, in retinal pigment epithelial cells (Toimela et al. 1998). Further, the phagocytosis of rod outer segments is inhibited by tamoxifen and toremifene in retinal pigment epithelial cell cultures (Mannerström et al. 2001). No statistical difference in toxicity between tamoxifen and toremifene has been noted (Toimela et al. 1998). Karlsson et al. (1996) found no oculotoxic properties in a 2-year dietary carcinogenicity study of the anti-oestrogen toremifene administered to Sprague-Dawley rats.

The purpose of the present study was to monitor the ocular side-effects of both anti-oestrogens, tamoxifen and toremifene, in a randomized adjuvant setting.

Patients and Methods

A total of 60 postmenopausal (age range 50–79 years) breast cancer patients in Tampere University Hospital were included in this follow-up study. They were randomized to receive toremifene 40 mg/day or tamoxifen 20 mg/day for 3 years. The study was approved by the ethical committee of Tampere University Hospital. It represented a substudy in a larger, multicentre adjuvant trial co-ordinated by the Finnish Breast Cancer Group (FBCG) (Holli et al. 2000).

Subjects had to be postmenopausal women who had had their last natural menstrual period more than 6 months prior to the diagnosis of breast cancer. Inclusion criteria required the subjects to have histologically verified invasive breast carcinoma with one or more metastatic axillary lymph nodes and to have undergone mastectomy or segmental resection and axillary dissection less than 10 weeks before randomization, and to have had no other treatment than radiotherapy for breast cancer. Their World Health Organization (WHO) performance status had to be less than 2. Male patients and patients with distant metastases at the time of the diagnosis were excluded, as were patients with inflammatory breast cancer, cancer other than in-situ cervical carcinoma, or a history of deep venous or arterial thrombosis. Patient characteristics are shown in Table 1.

Table 1.  Characteristics of patients, disease and primary treatment in the randomized adjuvant tamoxifen versus toremifene FBCG ocular study.
CharacteristicsTamoxifenToremifene
 group (n = 28)group (n = 32)
 n (%)n (%)
Mean age, years ± SD61 ± 861 ± 6
Hypertension6 (21)6 (19)
Diabetes2 (7)0 (0)
Tumour size
 T119 (68)25 (78)
 T27 (25)4 (13)
 T32 (7)3 (9)
Primary treatment
 Mastectomy15 (54)11 (34)
 Segmental resection13 (46)21 (66)
Radiotherapy given28 (100)32 (100)

Randomization was carried out centrally in the Finnish Cancer Registry. After randomization for the main protocol, and upon informed consent to the ocular subprotocol, patients were sent to the Department of Ophthalmology at Tampere University Hospital without indication as to the treatment group they had been allotted to. All consecutive patients treated in Tampere University Hospital were admitted to the ocular subprotocol. Eleven patients refused to take part in this subprotocol, the main reason being the distance to the hospital.

This study was designed to monitor ocular side-effects of adjuvant anti-oestrogen therapy. Prior to commencement of medication, a thorough ocular examination was made (Table 2). The first follow-up visit took place after 6 months and the remaining three at 12-month intervals thereafter. The total planned follow-up time was 3.5 years per patient. The final visit took place 6 months after finishing the therapy. Recurrence of breast cancer during follow-up was also considered as an end-point.

Table 2.  Ophthalmological findings at baseline prior to medication in adjuvant tamoxifen (11 findings in nine patients) versus toremifene (15 findings in 13 patients).
FindingTamoxifenToremifene
 group (n = 28)group (n = 32)
Cornea, arcus senilis1
Lens, cataract79
Macula
 Crystals12
 Drusen23
Retina, papilla oedema1

The main outcome measures were slit-lamp findings in the cornea, lens, macular area and retina. At each visit, data were also collected on the following items: ocular and other illnesses, medication and symptoms. Visual acuity (VA), intraocular pressure and slit-lamp findings were registered. The experienced physician (A-M. P) who interviewed and examined the subjects was blinded to their study treatment.

Results

Patient characteristics and disease did not differ between the treatment groups (Table 1).

Eight patients were excluded from the final analysis because they did not complete the planned 3.5-year follow-up time; six of them were in the tamoxifen group and two were in the toremifene group. Three of these eight patients experienced recurrence of the cancer (two in the tamoxifen group and one in the toremifene group). Two patients in the tamoxifen group moved away, one patient in each group experienced side-effects other than ocular, and one in the tamoxifen group died of causes other than breast cancer. The total follow-up time was 194 person-years.

The ocular findings seen at baseline were similar in the two groups (Table 2). Arcus senilis was diagnosed in one patient in the toremifene group before commencement of medication. Sixteen patients had cataract at the first visit (seven in the tamoxifen group and nine in the toremifene group). Ten developed cataract during the study period (five in each group) (Table 3), giving annual incidences of cataract of 6.8% and 6.2% in the tamoxifen and toremifene groups, respectively. Three patients had macular crystals at the first visit (one in the tamoxifen group and two in the toremifene group). Macular drusen were diagnosed in five patients at the first ophthalmological check-up (two in the tamoxifen group and three in the toremifene group). Two patients in the toremifene group developed drusen maculopathy during follow-up visits. Yellowish spots in the macular area were found in one tamoxifen-treated patient at the second visit. However, as VA was normal and there were no other ophthalmological abnormalities and no subjective complaints, the medication was continued. At the final visit after a 3.5-year follow-up, the spots had disappeared. One patient in the tamoxifen group had papilla oedema at the first visit. The reason for the condition remained unknown and VA remained stable. During follow-up the papilla returned to normal. In the toremifene group a preretinal haemorrhage was documented in one patient at the final visit, when the medication had already been discontinued.

Table 3.  New ophthalmological findings during the follow-up of 3.5 years in the adjuvant tamoxifen versus toremifene ocular study.
FindingTamoxifenToremifene
 group (n = 22)group (n = 30)
  • *

    In one patient yellowish spots in the macular area were found at the 1.5 year visit, but the changes disappeared during the 2-year follow-up period.

Cornea, arcus senilis
Lens, cataract55
Macula
 Crystals*
 Drusen2
Retina
 Papilla oedema
 Preretinal haemorrhage1

Discussion

This study investigated the oculotoxic properties of tamoxifen and toremifene using a randomized approach. No serious ocular side-effects were observed during the follow-up period of 194 person-years.

The main strengths of this study were represented by its prospective randomized design and the systematic repeated ophthalmic examinations carried out by an experienced ophthalmologist who was blinded to the treatment. Its limitation was the relatively small number of patients (n = 60) and follow-up years. One further limitation concerned the lack of a control group. In other words, the study design allowed only comparison between tamoxifen and toremifene adjuvant therapies. The mean age of the patients involved was 61 years. It is possible that the higher mean age may have meant the subjects were more predisposed to corneal and retinal changes. Our results do, however, show the importance of a baseline study, as retinal changes for other reasons are common amongst the older population.

Although tamoxifen has been used in the management of breast cancer since the beginning of the 1970s, its ocular toxicity has been little studied. A number of case reports have been published (Kaiser-Kupfer & Lippman 1978; Kaiser-Kupfer et al. 1981; McKeown et al. 1981; Pugesgaard & Von Eyben 1986; Griffiths 1987; Ashford et al. 1988; Gerner 1989; Costa et al. 1990; Bentley et al. 1992; Chang et al. 1992; Chern & Danis 1993; Sekhar & Nagarajan 1995; Yanyali et al. 2001). In the first cross-sectional study, Beck & Mills (1979) found no ocular changes attributable to the drug in 19 women. In another cross-sectional study, De Jong-Busnac (1989) found two cases of tapetoretinal degeneration among 20 patients, and Vinding & Nielsen (1983) found two cases of retinopathy and four cases of subepithelial corneal deposits among 17 patients. Longstaff et al. (1989) found no ocular toxicity attributable to tamoxifen in a cross-sectional study of 79 patients and 115 control subjects. Heier et al. (1994) found two cases of intraretinal crystals in a cross-sectional study involving 135 patients. Bernardo et al. (1994) found paracentral corneal opacities (12%), dots in the foveal area of the retina (5%), macular degenerations (9%) and vascular abnormalities (2.5%) in 119 patients in a cross-sectional study. Therssen et al. (1995) examined 61 patients and found two with mild crystalline deposits in the macular region, one with corneal deposits and one with optic neuritis. Gorin et al. (1998) published a cross-sectional study of 303 patients in which the participants included women who had never taken tamoxifen (n = 85), women who had taken tamoxifen for an average of 4.8 years and then come off the drug (n = 140), and women who had taken tamoxifen continuously for an average of 7.8 years (n = 78). Intraretinal crystals and posterior subcapsular opacities were more frequent in the tamoxifen-treated group. Lazzaroni et al. (1998) found refractile retinal opacities in only a low percentage of patients (3.1%) in a cross-sectional study of 129 women. As the investigators emphasized, however, they could not be certain that the refractile retinal opacities observed were really caused by tamoxifen, as a differential diagnosis excluding age-related macular degeneration with cuticolar drusen appears almost impossible. All of these studies are limited by the lack of a baseline ophthalmic examination.

Prior to the present study, there has been only one prospective study where an ophthalmic check-up has been performed before initiation of tamoxifen medication (Pavlidis et al. 1992) and one prospective study where some patients were already on tamoxifen and some had not yet started tamoxifen medication at the time of their initial visit (Noureddin et al. 1999). The former group, in a prospective non-controlled study, found that four of their 63 subjects developed retinopathy involving bilateral macular oedema and yellow-white dots in the paramacular and foveal areas. In addition, one patient had subepithelial corneal opacities. Noureddin et al. (1999) reported eight patients with ocular toxicity among a 65-patient study group. Seven patients had keratopathy and three of these had concurrent symptomatic bilateral pigmentary retinopathy. One patient developed bilateral optic neuritis.

Prospectively collected information on ocular toxicity is available for 2375 women who received tamoxifen alone or in combination with cytotoxic therapy in several large randomized clinical trials of adjuvant therapy for early stage breast cancer (ICI Pharmaceuticals Group 1990). Among 926 patients who received tamoxifen, 11 reported ocular symptoms and four of these 11 had ocular findings: keratitis (n = 1), retinal abnormalities (n = 2) and cataracts (n = 1). These data would suggest that there is no difference in ocular symptomatology between patients who receive tamoxifen alone and those who receive no treatment (Nayfield & Gorin 1996).

Heier et al. (1994) observed calcified drusen, reflections off the limiting membrane and small epiretinal membranes to be the most common findings initially identified as suspicious refractile macular lesions in their large cross-sectional study of asymptomatic patients receiving tamoxifen therapy. The investigators suggested contact lens biomicroscopy to facilitate identification of these mimicking lesions.

Corneal changes similar to those described in patients taking tamoxifen have also been reported in asymptomatic carriers of Fabry's disease (Anderson-Fabry disease) (Nayfield & Gorin 1996). The characteristic corneal lesions are present in 70% of female carriers. In addition, subepithelial lines and opacities such as those reported in map-dot-fingerprint dystrophy may occur in more than 40% of the general population and 75% of persons over the age of 50 years, independently of ocular symptoms and recurrent corneal erosions (Werblin et al. 1981).

Toremifene has been available for clinical use since 1982, but its ocular toxicity has not been studied in a randomized setting. Our results show the importance of recording ophthalmic findings prior to initiation of medication, as anti-oestrogen retinopathy must be differentiated from other pathologies which can produce a similar retinal aspect, for example oxalosis (Small et al. 1990), nephrophatic cystinosis (Kaiser-Kupfer et al. 1986), retinitis punctata albescens (Smith et al. 1959), Bietti's crystalline dystrophy (Francois & Laey 1977), Alport's disease (Polak & Hogewind 1977), drusen (Nayfield & Gorin 1996), Sjögren–Larsson syndrome (Jagell et al. 1980), canthaxanthine retinopathy (Harnois et al. 1989) and talc embolism (McLane & Carroll 1986).

In studies by Fisher et al. (1998) and Paganini-Hill & Clark (2000), tamoxifen users were reported to develop cataracts more frequently than non-users. In both cases the information was based on uncontrolled self-reporting. In the present study, the proportion of new cataracts in initially cataract-free patients was 5/21 (23.8%) for tamoxifen and 5/23 (21.7%) for toremifene, giving annual rates of cataract of 6.8% and 6.2% in the respective groups. These figures are similar to those reported by Klein et al. (2002) for healthy white women of the same age in the USA.

In conclusion, our prospective randomized 3.5-year follow-up study involving 60 patients revealed no significant ocular side-effects in breast cancer patients treated with tamoxifen or toremifene. However, it showed that in careful ophthalmological examination ocular findings are common among older women.

Acknowledgement

This study was financially supported by the Medical Research Fund of Tampere University Hospital, Tampere, Finland.

Ancillary