To the Editor:
We disagree with the assertion that Omalu et al.’s (6) recent article actually reports a case of “chronic traumatic encephalopathy in a National Football League (NFL) player.” We base our opinion on two serious flaws in Omalu et al.’s article, namely a serious misinterpretation of their neuropathological findings in relation to the tetrad characteristics of chronic traumatic encephalopathy and a failure to provide an adequate clinical history.
Neuropathology of Chronic Traumatic Encephalopathy
We disagree with Omalu et al.’s statements that the neuropathological findings in this case are “consistent with chronic traumatic encephalopathy” and “met criteria for chronic traumatic encephalopathy.” These statements are based on a complete misunderstanding of the relevant medical literature on chronic traumatic encephalopathy of boxers (dementia pugilistica). A review of the relevant medical literature, including that cited by Omalu et al., in the chronological order in which it was published demonstrates the flaws in Omalu et al.’s assertions.
The seminal paper on chronic encephalopathy of boxers was published by Corsellis et al. (2). In this landmark study, Corsellis et al. reported a specific pattern of neuropathological findings in the brains of 15 retired boxers. The results were based on gross and standard histological examinations of the autopsy material. Corsellis et al. determined the hallmark neuropathology of chronic traumatic encephalopathy to be: 1) abnormalities of the septum pellucidum (cavum, fenestrations); 2) cerebellar scarring on the inferior surface of the lateral lobes, most marked in the tonsillar region and loss of Purkinje cells in these areas; 3) degeneration of the substantia nigra with loss of pigmentation, neurofibrillary changes, and no Lewy bodies; and 4) widespread neurofibrillary tangles in the cerebral cortex and brainstem, most prominently in the medial temporal lobe gray matter, in contrast to “the sparsity, or in most cases, the total absence of senile plaques.”
None of the later references cited by Omalu et al. or others ever argued that the findings reported by Corsellis et al. (2) do not define chronic traumatic encephalopathy in boxers. The relevant references suggested a possible qualifier to the fourth characteristic (widespread neurofibrillary tangles and sparse or no senile plaques), but never took issue with the first three characteristics.
Roberts (7) performed immunocytochemical studies on brain material on eight of Corsellis et al.’s 15 cases. He reported that the neurofibrillary tangles in these cases were stained by a battery of antisera that also stained the neurofibrillary tangles of 15 cases of Alzheimer’s disease brains. Roberts thus suggested that head injury and Alzheimer’s disease may share some pathogenetic factors. Nothing in Roberts’ study suggested that Corsellis et al.’s original definition of chronic traumatic encephalopathy was incorrect. In fact, the material examined in the study by Roberts was defined as originating in dementia pugilistica brains because it was so designated by Corsellis in his earlier paper (7).
Adams and Bruton (1) examined the brains of 22 former boxers and found that 17 of the 22 brains showed histological evidence of recent or past hemorrhages. Thirteen of these brains were those originally studied and reported on by Corsellis et al. (2). Twelve of these brains “showed neuropathological evidence of brain damage of the type described in the “punch-drunk state” (1). The brains of five additional retired boxers were examined; three of these “showed the typical stigmata of dementia pugilistica” (1). What was the neuropathological pattern or stigmata that defined the punch-drunk or dementia pugilistica state according to Adams and Bruton? A “tetrad” of pathological processes originally described by Corsellis et al. (1). Clearly, Adams and Bruton did not dispute the fact that Corsellis et al.’s criteria defined the neuropathology of chronic traumatic encephalopathy.
Roberts et al. (9) reported a case of dementia in a punch-drunk wife. The neuropathology of the case revealed a large fenestrated cavum septum pellucidum, neurofibrillary tangles on routine staining and immunoreactive tangles and diffuse plaques on immunocytochemical studies with specific antisera (9). The authors stated that the brain of this “repeatedly battered woman” resembled that seen in dementia pugilistica. Roberts et al. suggested that the immunocytochemical picture might be a part of this condition, but mainly suggested that the case provided evidence that “head injury can be followed by Alzheimer type degeneration” (9). Roberts et al. (9) cited Corsellis et al’s (2) study as the source of his description of dementia pugilistica. Furthermore, the brain reported by Roberts in this letter had the septal abnormalities on gross examination and neurofibrillary tangles on routine histological staining that are integral to the neuropathological description of dementia pugilistica.
Roberts et al. (8) also reported the results of immunocytochemical staining of 20 former boxers’ necropsy brains with an antibody to β-protein amyloid. Fourteen of the cases were from Corsellis et al.’s (2) original material. One was from a boxer who died during a bout. The other five consisted of three additional professional and two amateur boxers. Nineteen of the 20 cases had a cavum septum pellucidum. On routine staining, 13 had neurofibrillary tangles with minimal or no plaque formation, four had equal amounts of tangle and plaque formation, and three had no neurofibrillary tangles. Thus, most of the cases fit the definition of dementia pugilistica as described by Corsellis et al. (2) and most, in fact, were from Corsellis’ original material (8).
Roberts et al. (8) reported that 19 of the cases exhibited “substantial numbers of hitherto invisible diffuse β-protein immunoreactive plaques” on immunocytochemical analysis. They suggested that “the present neuropathological description of dementia pugilistica (tangles, but no plaques) should be altered to acknowledge the presence of substantial β-protein deposition (plaques).” The authors did not indicate that the gross brain findings or microscopic findings on routine histology reported by Corsellis et al. (2) were incorrect, but rather offered a qualifier to the fourth criterion of the tetrad, made possible by the development of new immunocytochemical techniques that were not available to Corsellis et al. (2).
Dale et al. (3) reported the results of immunocytochemical staining of 16 former boxers’ brains with antisera to ubiquitin (a component of neurofibrillary tangles in Alzheimer’s disease) and to BF-10 (a component of tangles in Alzheimer’s disease and dementia pugilistica). Eleven of these cases had dementia pugilistica, as defined by the Corsellis et al. (2) criteria. In fact, 10 of the cases were from Corsellis’ original material. Nine of these 10 cases were also included in Roberts et al’s (8, 9) study. The results demonstrated that ubiquitin (a component of Alzheimer’s disease tangles) is also present in neurofibrillary tangles in dementia pugilistica and that more neurofibrillary tangles in dementia pugilistica are stained with BF-10 than with antiubiquitin.
The authors proposed that Alzheimer’s disease and dementia pugilistica neurofibrillary tangle formation “share many common features” (3). At no point, however, did the authors imply or state that Corsellis et al.’s (2) criteria for dementia pugilistica were incorrect; and in fact, the authors used Corsellis’ criteria to identify the cases of dementia pugilistica for their study.
Tokuda et al. (10) reexamined eight former boxers’ brains “using immunohistochemistry with antibodies to amyloid β-protein and τ-protein.” Seven of the eight cases were taken from Corsellis et al.’s (2) original material. Tokuda et al. (10) reported that the neurofibrillary tangles in dementia pugilistica are similar to those seen in Alzheimer’s disease. They also found other τ-immunoreactive structures (neuropil threads and degenerating neurites) in the periphery of degenerating plaques in dementia pugilistica. The authors also found β-protein cerebral amyloid angiopathy in three of the eight cases. Tokuda et al. (10) did not comment on any of the other Corsellis et al. criteria, did not indicate any disagreement with those criteria and in fact used Corsellis et al.’s original material (brains that meet the Corsellis et al. criteria for a diagnosis of dementia pugilistica) as brains defined as exhibiting dementia pugilistica.
Geddes et al. (5) reported the autopsy results of a 23-year-old boxer who died in the ring. They found widespread neurofibrillary tangles in all neocortical areas with sparing of the medial temporal lobe. No Alzheimer’s disease changes and none of the other characteristics findings of dementia pugilistica (as defined by Corsellis et al.) were reported (5). The authors suggested that the development of neurofibrillary tangles may be the earliest change in the development of chronic encephalopathy of boxers. Nothing in Geddes’ study challenged Corsellis’ criterion for the diagnosis of dementia pugilistica (5).
Geddes et al. (4) later reported autopsy results of five young men who experienced mild chronic head injury. Two were boxers, one was a soccer player, one was an epileptic who frequently hit his head during seizures, and one was a developmentally disabled man with a long history of head banging.
Geddes et al. (4) performed routine staining and immunostaining with antisera to β-amyloid and τ-proteins. All of the cases demonstrated neocortical neurofibrillary tangles and τ-positive neuropil threads. There was no abnormal staining with antisera to β-amyloid. No abnormal plaques were seen (4). The authors concluded that neocortical neurofibrillary tangle and τ-positive neuropil thread formation without β-amyloid deposition is the earliest neuropathological finding in repetitive head injury in young adults. This report is thus very consistent with Corsellis et al.’s original findings and is, in fact, inconsistent with Omalu et al.’s (6) assertions. Omalu et al.’s case demonstrated diffuse amyloid plaques, sparse neuritic threads, and sparse neurofibrillary tangles which is the opposite of Geddes et al.’s findings of widespread neurofibrillary tangles without β-amyloid plaque formation (4, 6).
This chronological review of the development of the neuropathological description of dementia pugilistica indicates the following: 1) in 1973, Corsellis et al. (2) described the tetrad of findings that still define the neuropathological picture of dementia pugilistica; 2) this tetrad includes abnormalities of the septum pellucidum, cerebellar scarring, degeneration of the substantia nigra, and widespread neurofibrillary tangles with minimal or no plaque formation on routine histological staining; 3) in the late 1980s and 1990s, the availability of new techniques of immunocytochemical analysis further described the nature of neurofibrillary tangles in dementia pugilistica and revealed that plaques were indeed present when these new techniques were applied; and 4) a qualifier to Corsellis et al.’s fourth criterion (neurofibrillary tangles without plaques on routine histological examination) can be added: immunocytochemical studies reveal that plaques are indeed present in these brains.
None of the references reviewed argued, in any way, that Corsellis et al.’s (2) original tetrad of findings was not correct or do not define the syndrome of dementia pugilistica. In fact, all of the more recent studies cite Corsellis et al.’s (2) paper, and many of them actually used brain material from Corsellis et al.’s original subjects (defined as being dementia pugilistica brains) for their newer analyses. These studies would not have done so if their authors did not agree that Corsellis et al’s description accurately classifies the neuropathology of chronic traumatic encephalopathy. None of the references reviewed ever suggested that Corsellis et al.’s neuropathological description be abandoned or that immunocytochemical criteria alone should replace Corsellis et al.’s tetrad in defining chronic traumatic encephalopathy.
This historical review clearly demonstrates the error of Omalu et al.’s (6) statements that their case is “consistent with chronic traumatic encephalopathy” and “met criteria for chronic traumatic encephalopathy.” The following compares Omalu et al.’s reported case findings with the tetrad neuropathological criteria of Corsellis et al. (2): 1) Omalu et al.’s case did not have a cavum septum pellucidum or a fenestrated septum pellucidum; 2) Omalu et al.’s case did not have scarring of the inferior surface of the cerebellum; 3) Omalu et al.’s case did not have widespread neurofibrillary tangles with minimal or no plaque formation on routine histological examination; and 4) Omalu et al.’s case did have depigmentation of the substantia nigra.
Omalu et al’s case demonstrated only one of the four of Corsellis et al.’s (2) tetrad criteria, certainly not enough to be consistent with or meet criteria for chronic traumatic encephalopathy of boxers. On immunohistochemical staining, Omalu et al.’s case had frequent amyloid plagues, but only “sparse neuritic threads and neurofibrillary tangles.” Close review of the literature indicates that these findings share some similarities to Roberts et al.’s (8) results, but are not exactly the same as those reported by Tokuda et al. (10), who found significant, not sparse, neurofibrillary tangles and significant, not sparse, amounts of neurofibrillary tangles and neuropil threads), or Geddes et al. (4). Even if one accepts that these immunohistochemical results are similar to those reported in the references (this is not at all certain), this consistency with the qualifier to Corsellis et al.’s fourth criterion is still not enough to be consistent with a diagnosis of chronic traumatic encephalopathy of boxers.
This detailed review of the relevant literature including that cited by Omalu et al. shows clearly that Omalu et al.’s description of chronic traumatic encephalopathy is completely wrong. The characteristic neuropathological findings for chronic traumatic encephalopathy, “especially in boxers” are not sparse to many neurofibrillary tangles in the cortex, neuropil threads and neocortical diffuse amyloid plaques as stated by Omalu et al. (6). It is clear that Omalu et al.’s case did not meet the accepted criteria for chronic traumatic encephalopathy as described by Corsellis et al. (2) with only one qualification uncovered by newer immunocytochemical techniques in the late 1980s and early 1990s.
The dearth of clinical information seriously compromises Omalu et al.’s (6) assertions that this case is one of chronic traumatic encephalopathy. The diagnosis of a chronic condition requires a medical history indicating a long-standing nature of the illness. Clinical information regarding the onset and course of the illness over time is invaluable in this regard. Such a history is completely lacking in Omalu et al.’s (6) report.
The diagnosis of “encephalopathy” requires a detailed neurological history and neurological examination indicating clear evidence of clinical brain dysfunction. Postmortem telephone interviews with surviving relatives, which reveal “a neuropsychiatric history that resembled a dysthymic disorder” do not qualify as an adequate detailed neurologic and/or psychiatric history and clinical examination. One sentence indicating “a deficit in memory and judgment as well as Parkinsonian symptoms” is too vague and nonspecific. Did this information also come from postmortem telephone interviews with surviving relatives?
Results of medical-neurological clinical evaluations of the patient while he was alive are necessary to determine if he had any brain dysfunction. In the absence of adequate clinical evidence of any type of brain dysfunction, one cannot make a diagnosis of encephalopathy.
To state that a condition is “traumatic” in origin, one needs to present a history of significant injury (in this case, head injury) and rule out other possible etiological factors. Omalu et al. (6) indicate that their subject had a long career in professional football as an offensive lineman. They correctly point out that offensive linemen have a low incidence of mild traumatic brain injury (MTBI) compared with other position players. In the premortem history of this case, Omalu et al. do not mention any history of cerebral concussion, MTBI, or removal from play after a blow to the head.
In their discussion, Omalu et al. (6) indicate that “he sustained numerous episodes” of MTBI and/or concussive injury, as evidenced by the “histological evidence of remote hemorrhages.” This is circular reasoning. Although there is no history of head injury, the subject must have had multiple head injuries because of the autopsy findings.
The absence of an adequate medical history cannot be compensated for by reporting autopsy results. Omalu et al. (6) go on to state that “there was no known history of brain trauma outside professional football.” In fact, there was no known history of brain trauma inside professional football either. Omalu et al. also overlooked the patient’s participation in both high school and college football, in which head injuries have been known to occur. Omalu et al. have also ignored other possible etiologies of brain dysfunction (if there was indeed brain dysfunction) in this case.
The patient had severe heart disease (atrial fibrillation, coronary artery disease, cardiomyopathy) and obesity. Certainly, he was a good candidate for hypertension, diabetes, hyperlipidemia, and numerous other metabolic abnormalities. He was likely on multiple medications for treatment of his medical conditions. Omalu et al. (6), however, present us with no information regarding these clinical issues. Furthermore, they present no information regarding possible alcohol, steroid, or illicit drug abuse in this subject. All of these substances can cause brain dysfunction and thus deserve mention, at least with a simple statement about the presence or absence of such abuse in the medical history.
For all of these reasons, one cannot state that head trauma was a primary cause of his purported brain dysfunction. It is our contention that there is inadequate clinical evidence that the subject had a chronic neurological condition, a traumatically induced brain condition or, in fact, a clinical encephalopathy. The term “chronic traumatic encephalopathy” thus should not be applied to their case.
We have demonstrated that Omalu at al.’s (6) case does not meet the clinical or neuropathological criteria of chronic traumatic encephalopathy. We, therefore, urge the authors to retract their paper or sufficiently revise it and its title after more detailed investigation of this case.
Ira R. Casson
Elliot J. Pellman
David C. Viano
New York, New York