Should Patients With Radiation-Induced Optic Neuropathy Receive Any Treatment?
2011; Lippincott Williams & Wilkins; Volume: 31; Issue: 1 Linguagem: Inglês
10.1097/wno.0b013e31820d5361
ISSN1536-5166
Autores Tópico(s)Ocular Oncology and Treatments
ResumoClinical Presentation: Michael S. Lee Rarely, a patient treated with radiation therapy develops optic nerve or chiasmal injury. This typically occurs several months to years after completion of radiation (mean onset: 18 months) most frequently for paranasal sinus tumors or other skull-based lesions. While the most common site of injury involves the chiasm and/or retrobulbar optic nerve, anterior radiation optic neuropathy (RON) may also occur (1-5). The risk of RON increases with total radiation doses of more than 50 gray (Gy) or a single fraction of greater than 2 Gy using fractionated radiotherapy (1-5). Radiosurgery (eg, Gamma Knife and Cyberknife) doses of greater than 8 Gy also raise the risk of RON (6). Patients with diabetes mellitus or those receiving concomitant chemotherapy may have an increased risk for RON. The threshold for RON may be lower following radiation for pituitary tumors (42-50 Gy) possibly from the additive effect of chiasmal compression (7). Patients typically suffer painless vision loss in one or both eyes, with bilateral disease developing in 75% of cases either simultaneously or sequentially (8). Symptoms may progress over weeks to months from mild visual loss to complete blindness. Overall, the prognosis is very poor with a final visual acuity of 20/200 or worse in 85% of the affected eyes (9). At the time of examination, the optic disc may appear normal (posterior RON), swollen (anterior RON), or pale (previous optic atrophy). Brain and orbit CT demonstrate normal results, but MRI with gadolinium during the acute phase of RON reveals enhancement of the optic nerves and/or chiasm. Interestingly, an MRI performed for other reasons may show visual pathway enhancement prior to visual loss (2). Spontaneous recovery has been reported but is uncommon (2,10). Pathophysiology The pathophysiology of RON is unknown. It is presumably ischemic because of the demyelination, reactive astrocytes, and obliterative endarteritis seen on histopathologic specimens (1). One study showed that the optic nerves exposed to radiation plaques for ocular melanoma treatment had significant loss of vascular endothelial cells, supporting a vascular injury (11). This observation also helps explain the obvious gadolinium enhancement seen on MRI with RON. However, vascular injury alone, does not adequately explain the long-time delay between radiation therapy and visual loss. One hypothesis includes a critical role for replicating glial cells in axonal conduction. Because of the slow turnover rate of these cells, several months may pass prior to the onset of visual loss (2,5). PRO-Patients with radiation-induced optic neuropathy should be treated with hyperbaric oxygen: François Xavier Borruat, MD As Dr. Lee notes, RON is an unpredictable and dramatic complication of radiotherapy (1,8,26). Supporting an ischemic mechanism from vascular endothelial cell damage, magnetic resonance spectroscopy (MRS) in cerebral radionecrosis shows elevated lactate levels associated with severe morphologic changes. There was no MRS support for a primary demyelinating process as no elevation of choline was found within the injured tissues (30). In addition, pathology studies of human irradiated optic nerves showed a depletion of vascular endothelial cells (11). Other changes found in irradiated tissues are consistent with a state of chronic hypoxia without any signs of spontaneous microvascular revascularization (previously irradiated bones) (31). Given this evidence, a vascular/ischemic process appears to be the most important component in the development of RON. Can RON be prevented? Ideally, irradiated patients at risk should be medically protected to prevent the development of RON. Ramipril, an angiotensin-converting enzyme inhibitor, has led to modest improvement in optic nerve function and anatomy in an animal model of RON when given soon after irradiation (32,33). With controlled irradiation and ramipril doses, 4 of 7 rats were protected, while 3 of 7 lost optic nerve function (33). Large-scale application of such a potentially protective agent to all irradiated patients at risk of developing RON raises several questions. The safety and the costs of such a treatment have to be considered. Also, the timing of such treatment is uncertain given the variable time interval between radiotherapy and onset of RON. Finally, if the results of the animal model are extrapolated to humans, nearly 50% of treated patients would still develop RON. Can RON be treated and which therapy is adequate? In humans, therapeutic options include corticosteroids, anticoagulants, antiaggregants, and hyperbaric oxygen (HBO). There has been no randomized double-blind study for any treatment of RON. A retrospective literature review found no favorable effect of either corticosteroids or anticoagulation (4), and cases of radiation damage and visual loss have been reported to develop despite anticoagulation (19,20). Hyperbaric oxygen therapy The evidence of a vascular/ischemic etiology for RON suggests a possible role for increasing oxygenation of the irradiated tissues using hyperbaric oxygen therapy. HBO therapy consists of breathing O2 at 100% concentration at a pressure higher than 1 atmosphere (ATA). HBO induces a steep oxygen gradient between the healthy and the irradiated hypoxic tissues. This enhances both the proliferation of fibroblasts and the synthesis of collagen, creating supportive tissue for the proliferation of new blood vessels. Revascularization of hypoxic hypocellular irradiated tissues has been demonstrated not only in experimental animal models but also in humans (34,35). Oxygenation of previously hypoxic irradiated tissues has been demonstrated to persist in vivo in humans (34). The favorable effect of HBO requires both a minimal threshold of oxygen pressure and a minimum number of sessions (35). Using HBO at 2.0 ATA instead of 2.4 ATA led to a lower rate of success for treatment of osteoradionecrosis (34). That might explain why some authors using HBO at 2.0 ATA failed to report success in treating RON patients (3). Based on experimental and clinical studies, the following HBO protocol is recommended: 30 sessions of HBO therapy, breathing 100% oxygen at 2.4 ATA for 90 minutes per session. HBO therapy is a widely used, safe, and proven technique for a variety of conditions (Table 1). With these settings, systemic complications are rare (Table 2). The incidence of convulsions varies between 2.4 and 7 per 100,000 (Table 2) (28,35,36). Ocular side effects are also rare including transient myopia and cataract. One case of transient blindness was reported in a patient with a previous history of optic neuritis who underwent an unusually long course of HBO (6-hour session at 2.0 ATA) (37).TABLE 1: Approved indications for hyperbaric oxygen therapyTABLE 2: Complications of hyperbaric oxygen therapySpontaneous improvement of retrobulbar RON is exceedingly rare, and only 1 such case has been reported (2). The poor natural history and the lack of efficacy for corticosteroids, antiaggregants, or anticoagulants make HBO a reasonable treatment option (5,35-37). HBO is the only therapy that has favorably influenced the visual outcome of patients with RON. Guy and Schatz (4) were the first to report visual improvement in 2 patients with RON treated with HBO within 72 hours of visual loss. Their 2 other patients did not improve but were treated later at 15 days and 6 weeks after the onset of visual loss. Several additional cases of visual improvement following HBO have been reported in the literature (38). Among 5 personal cases of RON, 3 were treated with HBO and 1 markedly improved from 20/200 to 20/30 following HBO (unpublished data). Others have reported their experience with HBO in treating RON patients with no cases of visual improvement (2,3,9). Possible reasons for the unpredictability of visual improvement after HBO in RON patients include: 1) HBO is applied too late after the onset of RON, 2) HBO parameters (oxygen pressure and number of sessions) are inadequate, and 3) visual loss in some patients might be mainly from axonal loss rather than vasculopathy. None of the 13 patients with RON treated with HBO showed any visual improvement, but they were all treated with oxygen at 2.0 ATA instead of 2.4 ATA (3). Further, none were treated within 15 days of onset of visual loss, and 75% were given fewer than 30 sessions of HBO. There is no scientific study providing the appropriate timing for initiating HBO in RON, but common sense implies the sooner the better. To give patients with RON a chance to recover or stabilize vision, they should be treated with HBO as early as possible after the onset of visual loss. HBO therapy is a safe procedure, and current guidelines suggest a treatment protocol of 30 sessions at a pressure of 2.4 ATA. CON-Patients with radiation-induced optic neuropathy should not be treated with hyperbaric oxygen: Michael S. Lee, MD Proposed therapies No randomized, masked, controlled clinical trials (class 1 evidence) to treat RON exist for any proposed therapy of RON. Previous studies have reported the use of intravenous corticosteroids (12,13), anticoagulation (14), intravitreal vascular endothelial growth factor inhibitors (15), optic nerve sheath fenestration (ONSF) (16), and hyperbaric oxygen therapy with scattered success (8). Unfortunately, these reports used retrospective data collection from routine clinical charts. These case reports and small case series suffer from potential bias. The examiner and patient were not masked to therapy, and both may develop expectations of outcome based on treatment vs no treatment. Patients are rarely refracted in a standardized fashion using a uniform chart at each routine visit, which makes it difficult to assess true improvement in acuity. Finally, patients with central visual field loss may demonstrate improved performance as acuity, and visual fields are serially tested. Scanning techniques or increased familiarity with perimetry might result in improved visual fields. For example, in the Ischemic Optic Neuropathy Decompression Trial (IONDT), there were 245 eyes with visual field and acuity follow-up at 12 months (17). Regardless of treatment, there were 75 eyes with improvement of ≥3 lines of acuity. Of those that improved, only 38% demonstrated visual field improvement, suggesting that eyes with stable or even worsening visual fields learned to read the eye chart better (17). Spontaneous improvement in RON has been reported albeit infrequently. Three of the 4 patients experienced anterior RON, while the other suffered posterior RON (2,10). While the paucity of reports may reflect the true natural history of the disease, it may also represent lack of reporting. We learned a lesson from the IONDT: 43% of the observation group improved by 3 or more lines of acuity at 6-months follow-up, which came as a surprise to the neuro-ophthalmic community (18). In the absence of a large observation group of patients with RON, it is impossible to know the true natural history or how often spontaneous improvement occurs after RON. Corticosteroids High-dose corticosteroids have been used in RON but with limited success. Girkin et al (12) reported 4 patients with RON, all of whom received systemic corticosteroids and 1 received adjuvant hyperbaric oxygen therapy. One of the patients who received corticosteroids alone had acuity improved from 20/70 to 20/30. Lee et al (13) reported 3 cases of RON, and 1 stabilized on corticosteroids alone. Meanwhile, Borruat et al (8) reviewed the existing literature and did not find that corticosteroids affected the outcome of RON. Although uncommon, corticosteroid use carries a risk of avascular necrosis, gastrointestinal bleeding, and psychosis. I personally have not observed any success with corticosteroids for RON. Anticoagulation Neurologic deficits due to radionecrosis in other parts of the central nervous system have responded to anticoagulation (14). However, this has not translated into therapy for RON. There are several cases of patients developing RON while taking warfarin, suggesting anticoagulation may not be helpful (19,20). However, in the absence of a control group, it is unknown whether anticoagulation affects the outcome of RON. Other therapies There is 1 report of a patient with anterior RON who received intravitreal bevacizumab. The visual acuity improved from 20/32 to 20/20, and optic disc edema resolved (15). Another report described 3 patients with anterior RON who underwent ONSF. Each enjoyed substantial improvement (16). These patients suffered anterior RON, which likely has a different course than posterior RON. Hyperbaric oxygen therapy HBO involves placing the patient in a chamber at nearly 100% oxygen at more than 1 atmosphere of pressure. The Food and Drug Administration-approved uses for HBO (Table 1) include delayed radiation injury to soft tissue and bone but not the nervous system. In fact, several meta-analyses have shown no beneficial effect of HBO in neurologic disorders. A systematic review of 6 randomized clinical trials did not find evidence that HBO improves clinical outcomes in acute stroke (21) with similar results for traumatic brain injury (22,23) and multiple sclerosis (24). A review of 8 randomized clinical trials found that HBO was of benefit for radiation damage to bone and soft tissues of the head and neck, radiation proctitis, and prevention of osteoradionecrosis following tooth extraction. However, no advantage was observed for late radiation injury to either peripheral (brachial plexopathy) or central neural tissue (cognitive deficits) (25). Regarding optic neuropathy, a retrospective case-controlled study of HBO showed no benefit among 22 eyes with anterior ischemic optic neuropathy (AION) compared to 27 controls (26). No randomized clinical trials of HBO in RON have been conducted. In 1996, Borruat et al (8) reported that 2 of 4 patients with RON who received HBO enjoyed improvement in visual function. Their review of the literature found no cases of spontaneous improvement. For HBO to be effective, they recommended that patients receive HBO within 3 days of vision loss at ≥2.4 ATM for 30 dives of 90 minutes each. More than a dozen patients with RON were treated with this protocol at Johns Hopkins, and none of them improved (5). I have sent 1 patient with RON for HBO. Her visual acuity was 20/80 in the right eye, with a centrocecal field defect and no light perception in the left eye. After 26 dives (100% O2, 2.4 ATM, 90 minutes each), her acuity improved to 20/30 in the right eye, but her visual field defect remained unchanged. Unfortunately, within 3 weeks of completing HBO, she became encephalopathic from new-onset widespread radionecrosis of the brain. Since her visual field defect remained unchanged, the ability to read at 20/30 may have been due to improved test taking ability rather than any recovery from HBO. Interestingly, the HBO did not prevent the onset of progressive more widespread radiation necrosis. The cost of HBO in a private facility is approximately $110 per dive (26). In 2008, Medicare paid approximately $400 per dive for the facility fee and a $125 professional fee. The average length of approved treatments from Medicare is 20 dives (27). Borruat et al (8) described 30 dives, which would equate to almost $16,000, but this does not include the cost of travel, time off work, childcare, parking, or other costs necessary to receive this therapy. Since RON is not an approved indication for HBO, insurance may deny coverage. Additionally, HBO carries some risk (Table 2). Common minor side effects include dry eye and transient myopic shift. Uncommon but severe adverse events include ruptured tympanic membrane, lung and paranasal sinus barotraumas, and seizures. One patient reportedly died from suffering a seizure while receiving HBO (28). History has shown a lack of effective treatment for nonarteritic anterior ischemic optic neuropathy (NAION) and a much higher rate of spontaneous recovery than previously thought (18). It is possible that RON has a better natural history than the literature suggests. One could argue that we need to “do something” for RON instead of “nothing” because of the poor prognosis and the potential benefit of HBO with a low-risk profile. I used to have that same attitude with regard to high-dose intravenous corticosteroids for indirect traumatic optic neuropathy. I thought there was very little risk to a young healthy individual and great possible benefit. Then, the CRASH trial (29) randomized patients with a closed head injury and Glasgow Coma Score of 14 or less to receive intravenous corticosteroids vs placebo. Patients and investigators were masked to treatment. The trial halted enrollment at 10,000 patients when they found that more patients in the steroid group died in the first 2 weeks. This and other studies open the possibility that HBO could cause more harm than good for the treatment of RON. It is simply unknown until a randomized controlled clinical trial is performed. At this point, there is no evidence that HBO is beneficial for RON, NAION, or any neurologic disease for that matter, and I would not recommend HBO or any therapy routinely for the treatment of RON. Rebuttal: Dr. Francois Xavier Borruat Dr. Lee rightfully questions the rate of spontaneous visual improvement in RON and, based on reported cases of both anterior and retrobulbar RON, suggests that spontaneous recovery might be more frequent than currently reported (2,10). Indeed, the clinical course might differ between anterior and retrobulbar RON. For retrobulbar RON, there is only 1 published case of spontaneous improvement suggesting that the true incidence of spontaneous visual recovery is very low (2). Brown et al (10) reported a series of 6 patients who developed anterior RON. Of the 3 patients who were followed, 2 improved spontaneously. Anterior RON might then carry a better visual prognosis than posterior RON (10). In addition, Dr. Lee suggests that some of the reported cases of visual improvement attributed to HBO may have been artifact with visual improvement related to nonstandardized testing methods and learning effect than actual functional recovery. His argument is based on the results of the IONDT, where 75 of 245 patients with NAION had improvement in visual acuity but only 38% of the 75 patients had improved visual fields (17). However, improved vision in NAION can result from other mechanisms, such as resorption of subretinal fluid, detected by optical coherence tomography (39). Further, some of the RON patients treated with HBO had improvement in both acuity and visual fields (4,8). The assumption that visual improvement from HBO therapy results from a learning process is not necessarily correct. Dr. Lee raises the possibility of dangerous effects of HBO including seizures and death. Cerebral oxygen toxicity results mainly from the formation of reactive oxygen species resulting in oxidative cell membrane damage, whereas in the lungs, capillary endothelial damage and pulmonary edema can be present (37). In a large series of HBO-treated patients, there were only 2 cases of seizures among 80,679 patient-treatments (rate: 2.4 per 100,000 patient-treatments) (28). One of the 2 patients died in status epilepticus. This 22-year-old man had benefited of 30 HBO sessions when he developed seizures during the last session. After a 45-minute postictal stage, he recovered consciousness and received anticonvulsant therapy. On the fifth day, he again went into status epilepticus and eventually died. Autopsy was not carried out (28). In another series of more than 50,000 HBO patient-treatments, 1 sudden death was reported in a 72-year-old man who presented respiratory arrest during his tenth HBO session (40). Autopsy was not performed. Hence, the incidence of sudden death during HBO therapy can be estimated at 1 per 50,000 to 1 per 100,000 patient-treatments. Two other cases of death during HBO have been reported: myocardial infarction in a 80-year-old woman and pulmonary edema in a patient with aortic stenosis (40). In these cases as well, a causal role of HBO is uncertain. With careful evaluation and screening of patients prior to HBO therapy, this treatment modality can be safely given to individuals with RON. Dr. Lee is correct that questions remain regarding the natural history of RON, and no class 1 evidence exists regarding treatment. An international registry of RON cases should be created to expand and share our knowledge of the clinical profile of this visually devastating disorder. Rebuttal: Dr. Michael S. Lee The strongest argument for a benefit of HBO in RON depends upon a hypoxic environment in the absence of irreversible axonal loss. Yet, this pathophysiologic mechanism remains unproven. Dr. Borruat extrapolates from studies of HBO for radiation injury to nonneurologic tissues to RON. I agree that revascularization of hypoxic hypocellular irradiated tissues occurs in the mandible and its surrounding soft tissue, but I would caution against making this assumption in the anterior visual pathways. There is no definitive evidence that this happens in RON. The data do not exist to support routine HBO in RON. While it is impressive that Dr. Borruat has successfully treated 3 of 5 patients with RON, these numbers are not sufficient to generate broad generalizations about the treatment. I would advocate for a controlled trial, but given the rarity of this disorder, this seems unlikely to happen. Acknowledgment Dr. Michael Lee thanks Dr. Tariq Bhatti for his assistance with the literature review.
Referência(s)