A Contemporary Risk Analysis of Iatrogenic Transmission of Creutzfeldt-Jakob Disease (CJD) via Corneal Transplantation in the United States

With a mortality rate of 90% within 1 year of diagnosis [1], Creutzfeldt-Jakob disease (CJD) is a rare, rapidly progressive, neurodegenerative disease caused by the abnormal accumulation of misfolded prion protein in the brain [2]. First described by German neuropathologists Hans Gerhard Creutzfeldt and Alfons Maria Jakob in 1920 [3], CJD is classified as a transmissible spongiform encephalopathy (TSE), and is characterized mainly by progressive dementia, neuron loss [4], spongiform change on histopathological sections [5], and visual disorders, with a higher prevalence in individuals over 50 years of age. Creutzfeldt-Jakob disease can be transmitted through an autosomal-dominant hereditary pattern (gCJD), a sporadic pattern (sCJD), or iatrogenically (iCJD) [6]. A variant form of CJD (vCJD) thought to be caused by the agent responsible for past bovine spongiform encephalopathy outbreaks has also been identified in humans, potentially resulting from contact with contaminated meat products [7]. The latency period of the various forms of CJD has been found to have a longitude of up to 60 years, with an estimated average of 10–11 years for all forms of CJD [8]. While the majority of cases of CJD are acquired either sporadically (85%) or through inheritance (10%), there have been various reports of cases being transmitted iatrogenically through blood transfusions [9], human pituitary growth hormone (HGH) administration [10, 11], dura mater allografts [11, 12], surgical instrumentation [13], and transplantation of solid organs, including corneas. Though the likelihood of disease transmission through corneal transplantation has previously been calculated to be minute, risk factors and the prevalence of latent CJD in the transplant-intended donor pool have not been estimated since the analysis performed by Kennedy et al. [14] for the year 1997.

The lethality of CJD in conjunction with rising trends in corneal transplant volume, instances of multiple transplants, overall deaths due to CJD, and an aging U.S. population all warrant a contemporary risk analysis of CJD in the U.S. cornea donor pool, as performed in this article. With the development of effective screening methods of CJD in progress, the identification of donor and recipient risk factors can aid in distinguishing susceptible populations and in reporting potential cases of iatrogenic transmission of CJD through corneal transplantation globally.

Between 1961 and 2018, 1,954,600 corneal transplants have been performed with corneas from U.S. eye banks [16]. Corneal transplantation procedures are becoming more routine than ever, with a 60% increase in endothelial keratoplasty (EK) procedures, including Descemet stripping endothelial keratoplasty (DSEK) and Descemet membrane endothelial keratoplasty (DMEK), within just the last 10 years. As the largest exporter of corneal tissue, the United States exported over one third of donated cornea tissue to countries across the globe in 2018 [16], exemplifying the globalization of corneal transplantation. With the number of corneal transplants rising each year in the United States, and with dozens of countries using U.S. corneal tissue for transplantation, an understanding of the potential risk of iatrogenic transmission from donor to host through corneal tissue is essential, given its life-threatening implications. Creutzfeldt-Jakob disease is a debilitating, fatal prion disease that has been shown to cause infection in organ tissues, with reported cases of suspected iatrogenic transmission through neurosurgery [23], kidney transplants [24], and corneal transplantation in human beings. Although case reports of CJD transmission via corneal transplantation have been sporadic and infrequent, the severity of this disease, the greatly varying latency period, the possibility of unreported cases, and increased screening measures for CJD transmission risk all warrant a proper analysis of estimating the number of corneas in the annual U.S. donor pool from individuals with undiagnosed, latent CJD.

Both our 2018 analysis and our historical analysis from 1979 to 2018 yielded the same result: an estimated 1 in every 30,000 transplants in the U.S. will be performed with a cornea from a donor between the ages of 31 and 80 with latent CJD. Although we estimated 3.8 corneas from donors with latent CJD entering the U.S. donor pool in 2018, a prior analysis performed by Kennedy et al. in 1997, when the number of cornea donations was approximately half the present number (Table 4), yielded an estimate of 2.4 corneas from donors with latent CJD entering the U.S. donor pool annually.

Increases in cornea donation and size of the U.S. population have not significantly altered the risk of tissue recipients being exposed to CJD. Over the last decade, the CJD incidence rate has remained relatively stagnant; as such, employing routine molecular screening modalities to capture undetected CJD in banked donor corneas is still not recommended, despite the improved efficacy and cost of these assays. As a result, the most pragmatic course of action is to identify risk factors for iatrogenic transmission of CJD through corneal transplants from historical data. These risk factors could aid in more focused and rapid investigations into potential cases.

Based on the results of our analysis, age was identified as a primary donor risk factor for a deceased individual with latent CJD contributing a cornea to the transplant-intended U.S. donor pool. From 2007 to 2017, there was a 34% increase in the population of individuals over 65 years of age in the U.S. [25]. As a result, more elderly individuals are donating corneas, receiving cornea transplants, and dying of CJD, as reflected by the doubling of the number of CJD mortalities within the past 20 years. Of the 3.8 corneas from donors with latent CJD estimated to enter the transplant-intended U.S. donor pool in 2018, 84% (3.2) were contributed by individuals aged 61–80 (Table 3). Our findings are congruent with those of Kennedy et al., who found that the highest-risk donor age group for iatrogenic transmission of CJD was 60–69 years. This can be accounted for by the fact that an estimated 92% of deaths among individuals between 31–80 with latent CJD occur in individuals over 60 years old. For other non-PKP uses of corneal tissue, such as DMEK, the average age of donors is 70 [26]. Recipients of corneas from donors over 60 years of age are at greater risk of transmission of CJD than those receiving corneas from donors under the age of 60.

Thus, of the donors reported in the 10 global case reports, two donors were under the age of 20 (Table 1), which significantly diminishes the likelihood that they died of a prion disease. Based on this, it is highly unlikely that these younger donors transmitted CJD to the recipients through corneal transplant; these cases could likely be due to coincidental sporadic CJD in corneal transplant recipients. For the two global cases in which both the donor and the recipient died of CJD, the donors were 55 and 63 years of age, the latter of which was at greater risk of transmitting CJD through the cornea than younger donors due to old age (Table 1).

A contributing risk factor in recipients is having received multiple transplants in a lifetime, especially if their donors were elderly. In 2018, 12.3% of all corneal transplants performed, or 5966 procedures, were repeat transplants [16]. It is worth noting that individuals who have had multiple corneal transplants will have an increased risk of undergoing a transplant with a cornea from an individual with latent CJD. For example, if a patient undergoes a single corneal transplant in their lifetime, based on 2018 data, the overall probability that the cornea used in their transplant was from a donor with latent CJD between the ages of 31 and 80 is 0.0034% (3.8/111,703), or one cornea from a donor with latent CJD of every 29,500 corneas intended for transplant. However, if this patient undergoes a repeat corneal transplant, the overall likelihood that one of their donors was affected with latent CJD is doubled and becomes 0.0068%, or a 1 in 14,700 likelihood. A third corneal transplant would increase this likelihood to 0.00102%, or a 1 in 9800 likelihood. The risk becomes 0.005% for one transplant if the donor is over 60 years of age; for multiple transplants, this probability increases accordingly. These risk factors can be utilized to more easily identify candidates in which the likelihood of iatrogenic transmission may be heightened. It is also worth noting that penetrating keratoplasties were the main form of corneal transplant performed among these 10 cases; procedures such as DMEK and DSEK, which are becoming increasingly popular, do not the involve transplant of the entire cornea, and may result in reduced transmissibility rates, as less infected tissue is being used. Thus, penetrating keratoplasties may potentially pose a greater risk of transmission than other transplant procedures; however, there is insubstantial evidence to confirm this hypothesis as a definite risk factor.

Although the prevalence of CJD in the corneal donor pool is exceedingly low, the global reporting of a mere 10 cases of potential CJD transmission through corneal donation since 1974, with no cases documented after 2006, is lower than expected. Maddox et al. previously estimated that a case of CJD in corneal transplant recipients would be expected to occur once every 1.5 years [19]. At this rate, we would estimate there to have been eight potential cases since the last reported case in 2006. Thus, underreporting of cases of overlapping histories of death by CJD and prior corneal transplantation is likely occurring.

Of the 47 estimated CJD-affected corneas entering the U.S. donor pool since 1979 (Table 4), five of them, or 10.6%, were reported to have potentially caused CJD infection in recipients after transplantation (Table 1). Based on this small number of reported cases, and the lack of sufficient donor data for most of these reported cases, it is likely that transmission through the cornea does not occur in 100% of cases of corneal transplantation from a donor with latent CJD, which is supported in an article written by Tullo et al. [27] It is more likely that if these recipients did develop the infection from CJD-affected corneas, they may still be in the latency phase of the infection, which can continue for decades. Thus, they may begin to display symptoms in years to come, and because of this, there is no way of truly estimating an accurate transmissibility risk of CJD through the cornea other than through histopathological examination. Because the rate of transformation between receiving a CJD-affected cornea and developing latent CJD is unknown, it is difficult to distinguish whether the low number of reported cases is driven primarily by a low disease conversion rate, underreporting, or a combination of both.

Additionally, there is an apparent lack of geographical diversity among reported cases. Of the four internationally reported cases, two were in Japan and two were in Germany (Table 1), both countries with heightened awareness of CJD due to past history. Because of the wide transmission of CJD through German Lyodura brand dura mater cadaveric grafts in the 1970s and 1980s, attention was called to iatrogenic transmission of CJD in Germany. Similarly, Japan has reported an unusually high prevalence of dura mater-related CJD transmission, primarily due to the use of these German Lyodura grafts before 1987 [11, 12]. Cases of iatrogenic transmission of CJD through corneal transplantation may have been present in other countries. However, due to a lack of awareness of iatrogenic transmission, these cases most likely have not been reported, perhaps because cornea transplantation history of individuals who died of CJD was likely not investigated.

However, the reality that CJD was confirmed in both the donor and host in only 2 out of 10 globally reported cases (Table 1) lends credence to the former hypothesis of low transmissibility. To simply report overlapping histories of death by CJD and prior transplantation, although important for capturing possible cases, in no way evidences a causal link between the two. In fact, the molecular mechanisms underlying how CJD might spread from an implanted cornea to the central nervous system of the host remain entirely unclear. Experiments in guinea pigs wherein CJD-affected corneal tissue was implanted directly into the anterior chamber have shown the potential for misfolded prion proteins to diffuse through the vitreous to cause an ascending CNS infection [28]. However, these experimental conditions do not fully model the incorporation of live tissue as in a human corneal transplant.

Past studies of sporadic CJD in humans determined that prion seeds were present in ocular tissues such as the choroid, sclera, optic nerve, vitreous, lens, muscle, and the cornea [29]. Retrograde transmission from brain to cornea has been hypothesized to occur by the transport of misfolded prion protein by cranial nerves; the prions spread from the brain into the retina by axonal transport in the optic nerve, accumulating in small nerves in the cornea [29]. The transmission of iatrogenic CJD from the cornea to the brain would follow an anterograde pattern, potentially through axonal transport of exogenous misfolded prion protein to the midbrain via retinal ganglion cells [30, 31]. The overall mechanism may be that misfolded prion proteins from the CJD-affected transplanted cornea could spread to the retina of the recipient through vitreous humors; there, it accumulates in retinal ganglion cells, and axonal transport by these retinal ganglion cells to the midbrain would result in CJD transmission.

Prion diseases have usually been diagnosed through post-mortem histopathological examination of brain tissue, which has been neither cost- nor time-effective [32]. Because genetically screening the entire donor population may be infeasible, it is important to focus on ongoing development of cost-effective, efficient screening measures for prion disease. Besides genetic screening for familial CJD, screening for prion disease, including all forms of CJD, should be routine in donor examination, or general wellness exams, especially for potential donors over the age of 60. Recently, there have been efforts to improve the efficacy of post-mortem CJD detection. In 2017, Gregori et al. [33] proposed a method of rapid testing for CJD in cornea donors, using commercial testing combined with rapid sample collection, which proved to be able to detect even low concentrations of prion protease-resistant abnormal prion protein, or PrP^TSE. PrP^TSE testing could potentially be used to detect CJD even during its early stages in the cornea, enhancing the safety of corneal transplants. Strides have been made in the innovation of pre-mortem detection of prion disease as well. As of 2018, diagnostic tests utilizing protein misfolding cyclic amplification (PMCA) for pre-mortem detection of prion disease, specifically vCJD, through the blood were being developed. Pre-mortem tests, if applicable to multiple variants of CJD, including sporadic CJD, could assist in more accurate diagnosis of individuals with pre-symptomatic, latent CJD, potentially eliminating the possibility of iatrogenic transmission through organ tissues [34].

It may be worthwhile to further investigate the transmission rate of CJD through corneal transplantation from a latent, CJD-affected donor into a healthy recipient and confirm the presence of CJD along the ocular route to the brain through a histopathological examination, rather than by waiting for symptoms to present. Although previous studies have been conducted with less sensitive immunohistochemical assays, with the highly sensitive, rapid techniques of prion protein detection present nowadays, this study could be conducted with higher accuracy and serve to better illustrate the mechanisms by which CJD may be transmitted from the cornea to the brain.

Our methodology had several limitations. We utilized CDC WONDER as our sole mortality database for both CJD and all-cause mortality. There exist multiple databases that report varying numbers of CJD deaths and we acknowledge that CDC WONDER may underreport the number of annual CJD cases; however, we selected CDC WONDER due to the use of this database in past analyses. This analysis does not include individuals with symptomatic undiagnosed CJD. Because the average time to diagnosis after symptom onset is approximately 6 months [35], this would contribute a minimal number of donors.

For the global case report summary, there were cases in which the recipient had received multiple past corneal transplants, leading us to estimate ranges for the average incubation period, rather than an exact average. Many donors were unable to be identified; thus, the cause of death of donors was not confirmed.

For all calculations, we used an incubation period of 10 years. The CJD incubation period has been found to last as long as 60 years; because there has been no statistical analysis performed to determine the incubation period for all forms of CJD confidently, we made an estimate based on literature review and values used in other analyses. We also could not conduct any analyses for donors younger than 31 because there was insufficient mortality data from CDC WONDER. We were given the total number of corneas intended for transplant, but we were not given age-stratified data for this number, so we estimated this based on overall percent contributions of each age group to the total U.S. donor pool in 2018.

For the historical data analysis, we were unable to calculate values for the donor age group 31–80, because prior to 1999, CDC WONDER reported mortality data only in 10-year age stratifications starting at age 25. Thus, we used the age group 35–84. We were also unable to find the number of corneas intended for transplant in the U.S. prior to 2011, which is why we used a linear regression and trend analysis to estimate that number. Because EBAA only reports data in 10-year age intervals (example: 31–40), we assumed that the total number of corneas intended for transplant from donors aged 31–80 would be similar to the number from donors aged 35–84.

Although contemporary and past statistical analyses have shown that the prevalence of CJD-affected corneas entering the U.S. donor pool is low, we conclude that potential cases of CJD occurrence in both the donor and recipient of corneal transplants are underreported globally. For the sake of advancing neurological research in this field, all mortalities from CJD in the United States, and even globally, should be investigated for past corneal transplant history. If it is discovered that an individual who dies from CJD also has a history of corneal transplants, the cause of death of their cornea donor(s) should be investigated and reported. Similarly, if recipients of corneal transplants are diagnosed with CJD sometime within their lifetime, this should be reported immediately, and any potential recipient of the other cornea from the same donor pair should be promptly alerted and tested for any sign of prion disease. The advanced age of donors and multiple transplant history of recipients both present increased risk for iatrogenic transmission of CJD through corneal tissue. Advances in screening methods, especially among the elderly, could result in more cost-effective, rapid detection of prion disease in latent donors prior to transplantation. With corneal transplants and the number of deaths caused by CJD on the rise, further research should be conducted to investigate the pathology and detection methods behind the transmission of Creutzfeldt-Jakob disease through corneal tissue, as it can lead to a potentially life-threatening outcome of transplantation.