Cobalt Toxicity Presenting as Bilateral Optic Neuropathy

DISCUSSION

In 2010, Tower first used the term arthroprosthetic cobaltism to describe the phenomenon of systemic cobalt toxicity stemming from damaged prosthetic joints.⁸ The problem is increasingly being reported over time.⁹ To our knowledge, 16 cases documenting ocular cobalt toxicity originating from damaged prosthetic joints have been published. Tables 1 and 2 summarize the findings from these case reports.^4,6-8,10-21

Males and females were affected equally (8:8), and patient ages ranged from 39 to 71 years. Visual acuity data were not reported in 8 cases.^{8,10,14-16,19,20} Of the cases that reported objectively measured visual acuity, improvement in vision was documented in 4 cases.^4,6,7,12 One case reported worsening vision,¹⁷ and 3 cases reported stability.^13,18,21

Apel et al described retinopathy only.¹² Ng et al reported chorioretinal degeneration.¹³ Optic neuropathy without retinopathy was described in 7 reports,^{6-8,11,17,19,20} while 4 cases described concurrent optic neuropathy and retinopathy.^4,10,18,21 Rattay et al described extraocular movement abnormalities.⁷

Serum cobalt levels in the patients in these studies ranged from 35 μg/L⁸ to 1,461 μg/L.²⁰ At present, US Food and Drug Administration guidelines do not specify cutoff levels of cobalt in the blood for detection of arthroprosthetic cobaltism, citing a lack of evidence.²² The Medicines & Healthcare products Regulatory Agency (MHRA) in the United Kingdom suggests that whole blood cobalt levels ≥7 ppb (equivalent to 119 nmol/L or 7 μg/L) may indicate wear or corrosion of the joints. The MHRA also suggests that this threshold may be lower for stemmed total hip replacements.²³ A serum cobalt level of 0.19 μg/L is considered normal in the general population.^5,24 Concentrations ≥1.0 μg/L indicate possible environmental or occupational exposure.²⁵

ERG is a diagnostic test that measures the electrical responses of the retina to light stimulation. Six of the case reports included ERG results.^{4,6,12,13,18,21} Garcia et al reported that multifocal ERG showed decreased tracing amplitudes in both eyes.⁴ Grillo et al and Ng et al reported normal ERG.^6,13 ERG studies in the case described by Apel et al suggested an inner retinal pathology with persistent abnormalities in all phases of ERG.¹² Weber et al reported that full-field ERG showed severe cone dysfunction.¹⁸ Shaikh et al reported depressed photopic b-wave and flicker ERG amplitudes.²¹

VEPs are electrical signals generated in the brain in response to a visual stimulus and can be measured using electrodes placed on the scalp. VEP results were described in 5 cases.^6,11-13,19 Grillo et al described multifocal VEP with decreased central amplitudes.⁶ Rizzetti et al described irregular cortical visual responses.¹¹ VEP was normal in the cases reported by Apel et al¹² and Ng et al.¹³ Samargandi et al reported an abnormal VEP but did not describe the abnormality.¹⁹

Rizzetti et al¹¹ and Ho et al¹⁷ described contrast enhancement of the optic nerves on MRI. Lecoanet et al reported T1 hyperintensity and T2 hypointensity in the basal ganglia and caudate on MRI, corresponding to cobalt deposition.²⁰ Shaikh et al described perineural enhancement of the optic nerves after contrast administration.²¹ We did not administer contrast to our patient because of her poor renal function.

To our knowledge, this case is the first report of ocular cobalt toxicity associated with a shoulder prosthesis. The cases summarized in Tables 1 and 2 describe ocular cobalt toxicity related to hip prostheses. Khan et al reported a case of cutaneous titanium metallosis caused by a damaged shoulder prosthesis,²⁶ but in our review of the literature, we could not find any cases of systemic heavy metal toxicity associated with shoulder prostheses.

Studies have shown that patients who underwent shoulder joint replacement have lower whole blood levels of cobalt than patients who underwent hip replacement.²⁷ Consistent with this pattern, the serum cobalt concentration in our patient was lower than the levels reported in the other studies. However, there is conflicting evidence about the relationship between systemic cobalt levels and the occurrence of symptoms.²⁸ Finley et al concluded that the earliest symptoms of toxicity were noted at blood cobalt levels >300 μg/L.²⁹ A review by Gessner et al found that 17% of the reported cases of arthroprosthetic cobaltism had blood cobalt levels <20 μg/L.⁹ Despite this inconsistency, the risk of whether an individual will develop systemic toxicity depends on multiple factors that determine susceptibility, such as disease states and genetic polymorphisms.^5,30

Metal ions in the blood can exist in bound and unbound forms. In bound form, metal ions are less bioavailable; hence, the toxicologically active form of cobalt is free divalent Co (II) ions.^5,31 Strong cobalt binding sites in the serum include albumin, transferrin, and reduced glutathione.^5,30 Conditions that can alter the binding of the cobalt ions to these molecules—such as end-stage renal disease, genetic polymorphisms, hypercholesterolemia, and hypoalbuminemia—can affect how much free cobalt is available in the blood.^5,30 Our patient had renal failure, which is associated with higher than normal levels of ischemia-modified albumin (a form of albumin that has a modified N-terminus and thus reduced cobalt binding capacity), as well as decreased cobalt clearance.^5,32 Both of these factors should lead to elevated levels of free cobalt. Despite these factors, however, the serum cobalt level in our patient was not very high.

In their review of cobalt toxicity, Paustenbach et al report that the earliest effects of systemic cobalt toxicity are thyroid dysfunction and defective erythropoiesis, with symptoms occurring when the concentration of cobalt in the blood is >300 μg/L. Cardiac and neurologic effects are seen at much higher concentrations of >700 μg/L.⁵

Thyroid function and complete blood count results were normal in our patient. However, she may have had relative polycythemia, as anemia would be expected in the setting of poor renal function. Our patient had noteworthy peripheral neuropathy; however, she also had diabetes, and her neurology team did not definitively ascribe a cause to the neuropathy. Some unknown factors may have made her susceptible to the toxic effects of cobalt, even at low serum concentrations. Catalani et al suggested that neurologic symptoms associated with cobalt toxicity may be a form of mitochondrial neuropathy resulting from cobalt-induced mitochondrial injury.³⁰ Unknown mitochondrial genetic polymorphisms could have predisposed our patient to an enhanced sensitivity to cobalt.

Even though our patient's clinical setting was atypical for cobalt toxicity, all evidence points to cobalt toxicity as the cause of her optic neuropathy because visual acuity, color vision, and visual fields improved after her serum cobalt levels were managed. The damaged shoulder prosthesis was clearly the source of the cobalt ions because the cobalt levels in the synovial fluid were 35 times higher than normal, and the patient's serum cobalt concentration began to trend downward after the surgery.