A fatal and mysterious illness

Michael D. Shulman
Philadelphia, Pennsylvania, United States



 Scans, like  those above, showing reflux and stasis of cerebrospinal fluid
(CSF) in patients with dialysis dementia prompted placement of a CSF
shunt in at least one patient before intriguing new diagnostic clues emerged.

In late 1972, a flurry of letters began to appear in the British medical journal The Lancet which captured the alarm, the bafflement, and the intense professional curiosity aroused by a mysterious new illness. The illness was unique to patients receiving hemodialysis, typically those who had been on dialysis for many months. Some of the very first cases were discovered in Chicago, where Dr. George Dunea served as chairman of the Renal Division of Cook County Hospital. His literary bent is evident in the way his case report leaned poignantly into portraiture.1 The patient was a middle-aged woman whose life seems to have involved one grief after another. Her husband had been sent to prison after a conviction for drug-dealing. This meant she had been left to care for “numerous” children on her own, a task that became immeasurably more difficult when her kidneys failed and she required regular dialysis–six hours, three times a week. The dialysis unit knew her to be “unhappy,” an indication, perhaps, that she was clinically depressed. Then misfortune struck again. She began to suffer bizarre behavioral changes–stuttering, twitching, grimacing, jerking of the limbs. To those around her, it seemed the emotional strains that had hounded her for so long had conspired to darken her life still more.

But it was not emotional stress that was driving the patient’s peculiar transformation. Over the next few weeks, her condition worsened. She became confused, then incoherent. She suffered generalized seizures and hallucinations. Finally she slipped into a coma and died. It had been only three months since she exhibited the first ambiguous signs of a problem with speech and memory. No longer was there any question that “a fatal and mysterious illness” had claimed her. The very strangeness of her illness may have given her fellow patients a foreboding that they had not seen the last of it.

If the other patients did indeed have an intimation of miseries to come, it was well-founded. Reports of similar neurologic disorders in dialysis patients, it was soon learned, had been issued for at least two years.2,3 Indeed, the illness had already been described in detail by Dr. Allen Alfrey, a practitioner and researcher who oversaw a dialysis unit in Denver.4 These reports were followed by eerily similar ones from places as remote as South Africa and Australia. For the dialysis units in the Chicago area, the moribund woman with the unhappy home life proved to be a bellwether. In 1978, Dr. Dunea tallied the cases that had been observed over a period of four years–cases that reproduced the same pattern of confusion, involuntary movements, and disturbed speech. There were twenty of them, the patients in every case falling into a vegetative state and dying within nine months of diagnosis.5

As the number of victims multiplied, the dialysis units of Cook County fell more and more under a pall of apprehension. Stuttering in a patient who had previously spoken fluently, often the first sign of the illness, set off jolting alarms in the nurses.6 Vague signals of the growing disaster somehow reached consumers’ groups. Having no way to bring material relief, they limited themselves to dispatching representatives to the units to impart the depth of their concern. Patients arriving for dialysis began to look at each other with furtive suspicion “wondering whose turn would come next.”7

It was clear what needed to be done. It was time for the mysterious new illness to be taken up by clinical researchers who could devote the resources needed to investigate it, but now the illness needed a name, and selecting one called for skill of a different kind. A suitable name would not only capture the prominent features of the disease but would also be an aide-mémoire. And there would be a further payoff that would operate almost covertly. To assign a name to the outbreak was to gather its manifestations, diverse as they were, into a coherent portrait, and thus to signal to clinicians that they were not dealing merely with a hodgepodge of complaints in a population that was often unwell. Attaching a name to the syndrome meant it was a discrete illness, with a unique origin and perhaps a discoverable cure. Dr. Dunea briefly considered calling the novel illness “dialysis madness,” a term that was discarded, perhaps because it suggested a derangement of the kind associated with leather restraints in Victorian asylums. He opted instead, wisely and influentially, for “dialysis dementia.” The name was “catchy and euphonious,” and it stuck.

By 1973, dialysis dementia was widely recognized as a terrible new illness. But like readers embarking on a dense detective novel, clinicians looking for a culprit found it difficult to keep track of all the suspects. First to consider were complications arising from kidney failure. Patients who are under dialyzed can suffer mental status changes. So, too, can patients whose blood pressure soars because the kidneys are not performing their many regulatory functions. These possibilities were discarded quickly. Much interest was aroused when it was shown that the dynamics of the cerebrospinal fluid were altered in patients with dialysis dementia, creating a picture resembling that of normal pressure hydrocephalus.8 But this seemed more likely to be secondary to some other unidentified neurologic disorder than to be the primary cause of the unexplained illness.

A further possibility was that hemodialysis itself might have precipitated the fatal decline of the affected patients. While hemodialysis is an invaluable therapeutic instrument and undoubtedly keeps patients alive, it is not a perfect substitute for ailing kidneys, and it can produce neurologic complications of its own. These include heparin-induced bleeding into the brain and a peculiar form of disequilibrium that occurs when dialysis is begun too aggressively. Infection is another commonplace occurrence in the dialysis unit. To investigate an infectious process, brain matter harvested from a deceased patient with dialysis dementia was injected into monkeys. When the animals remained healthy four years later, it was concluded that dialysis dementia was unlikely to be due to infection with a slow virus.6

Excluding all these possibilities was an important step forward, but the cause of dialysis dementia was now up for debate. And the debate, much of it taking place in the Letters section of The Lancet, was vigorous, occasionally stepping very close to the margins of collegiality. Some of the theories offered by physicians were improbable or had been ruled out in the original reports of the disorder–recurrent hypoglycemia, for example.9 Some theories were torpedoed into oblivion only weeks after they were put forward. An example was the proposal that patients with dialysis dementia suffered a deficiency of the amino acid arginine, a possibility immediately refuted by evidence that arginine levels during dialysis do not fall.10,11 A suggestion that the disease was one of dopamine deficiency faltered when it was shown that repletion with dopamine was of no clinical benefit.12,13

But it was M. M. Platts, in yet another Lancet letter, who dropped the stone with the most widely propagated ripples.14 Dialysis dementia had been observed in Sheffield, England, but only in patients who received dialysis using untreated water. No patient being dialyzed with “soft” water–water from which dissolved metals had been removed–had ever been afflicted with dialysis dementia. Could something found in “hard” water be implicated in the genesis of the lethal neurologic disorder?

Any of the metals found in the tap water used to perform dialysis might be delivered to patients while they received treatment. The list was a long one. It included, at a minimum, lead, copper, zinc, iron, manganese, silicon, nickel, chromium, and cadmium.15 Probing the mystery further would mean addressing three questions: Did affected patients harbor deposits (or suffer deficiencies) of any of the metals on the list? Did the onset and severity of dialysis dementia parallel the intake of the suspected toxin, keeping in mind the oral route as well as the possible delivery of the toxin via dialysis? And most critically, did ending intake of the suspected toxin affect the incidence of the disease?

In its very early stages, every clinical investigation is hit-or-miss, and the investigation of dialysis dementia was no different. Inevitably, much time and capital were spent exploring promising avenues that turned out to be cul-de-sacs. It was noted that the brains of patients dying of the disease contained reduced concentrations of rubidium, a finding that went nowhere because rubidium deficiency did not, as far as anyone knew, produce symptoms.16 Moreover, giving patients with dialysis dementia rubidium did not result in improvement.8 Deposits of tin were also found in the post-mortem brains of affected patients. These were most extensive in the patients who had been treated with dialysis the longest. But larger brain deposits of tin did not distinguish patients dying with dialysis dementia from patients dying of other causes.

Then in 1976, came a study that in some quarters had been anticipated for years. Dr. Allen Alfrey reported in The New England Journal of Medicine that the gray matter in the brains of patients dying from dialysis dementia contained elevated levels of aluminum–four times the level found in dialysis patients dying of other causes, and ten times the level of patients dying without kidney failure.16As important as this finding was in its own right, it assumed greater importance knowing that Alfrey’s decision to focus on aluminum had not been made ex nihilo. The possibility of aluminum toxicity in dialysis patients had been an issue–a contentious one–in the past. At stake was the optimal treatment of a common disorder in dialysis patients–the elevated levels of phosphate in the blood that result from  an inability to excrete the phosphate contained in a normal diet.

In hundreds of dialysis units in America and abroad, the primary treatment for an elevated blood phosphate had been an aluminum salt. Aluminum, taken therapeutically in oral form, forms a complex with ingested phosphate, which then passes harmlessly through the gastrointestinal tract. Blood levels of phosphate are thus prevented from rising. But a concern lingered: some fraction of the aluminum administered in this way might not form a complex with phosphate, leaving it free to be absorbed into the bloodstream.

In 1970, two years before the first case of dialysis dementia was reported, there had been warnings that the extent of this absorption had been underestimated. Blood levels of aluminum were in fact shown to be elevated in patients given aluminum as a phosphate-binder.17 The observation that all the patients with dialysis dementia described by Alfrey had been taking oral aluminum thus stirred the embers of a debate that had never been extinguished. Now, with Alfrey’s study wielded like a scourge, the author of those earlier warnings returned to the pages of The Lancet. His letter was brief, but its bitter tone echoed as if the author had slammed the door on his way out of a heated conference room: “Had our advice [to discontinue oral aluminum] been taken some years ago,” wrote the author, “some patients might not have developed dialysis dementia.”18

After Alfrey’s study was published, no metal on the list of possible toxins received anything like the attention of aluminum. And yet, as late as 1976, an editorial in The Lancet could still argue that the link between dialysis dementia and aluminum was unproven.13 After all, many dialysis units prescribing aluminum for their patients reported no cases of progressive dementia at all. Perhaps the aluminum deposits noted by Dr. Alfrey were a nonspecific response to brain injury produced by some other unsuspected toxin. But in Europe, where aluminum was often added to the drinking water to remove particulate matter, the link between dialysis dementia and aluminum began to look ironclad. It was shown that dialysis dementia in Scotland occurred only in regions were the public water supply was treated with aluminum. Neighboring regions using treated water were unaffected.19 An outbreak of dialysis dementia in the Netherlands followed promptly when a faulty electrode in the water heater used for dialysis began to leak aluminum.20

Meanwhile in the United States, Dr. Dunea and his colleagues made a related discovery. A startling graph showed that that the first cases of dialysis dementia in the Chicago area had occurred only after the city began adding aluminum to its water supply.5 Thereafter, spikes in the incidence of dialysis dementia coincided closely with spiking aluminum levels in city water. Then, in 1975, the Chicago dialysis units installed deionizers, greatly reducing the level of aluminum in the water flowing through the dialysis machines. If a debate among medical professionals is anything like a high-stakes game of bridge, and if there is an analogy to playing trump, this was such an occasion. The last case of dialysis dementia was observed just as the deionizing devices were being introduced. After that, dialysis dementia simply vanished.

Doubts about the cause of dialysis dementia continued to be expressed, based largely on the failure to demonstrate aluminum toxicity in dogs and rats, and on the discovery of elevated brain aluminum in patients who had no neurologic symptoms.21 Additionally, the way dialysis dementia chose its victims, targeting some dialysis units while sparing others nearby, could not be explained.22 And there was continuing uncertainty, which persists to the present day, regarding whether aluminum in oral form or in the water supply posed the greater danger. Such reservations and qualifications had merit, but they were overwhelmed by evidence that the illness could be prevented altogether if dialysis units fired both barrels—withdrawing oral aluminum and purifying their water supply.23 Both measures were soon adopted as the standard of care. By 2001, Dr. Dunea was able to write of dialysis dementia unambiguously as an “aluminum encephalopathy.” Much more importantly, he was able to memorialize it as “an epidemic that came and went.”7 If puzzles remained–particularly about the exact mechanism of aluminum toxicity24–they were of a kind familiar to clinicians, who know that medicine is not like pure mathematics, and that papers on medical topics are not apt to conclude authoritatively with a quod erat demonstrandum that routs every gainsayer.

As things turned out, Dr. Dunea’s optimistic postscript was fully justified: dialysis dementia by 2001 had indeed come and gone. But this schematic way of summing up events perhaps bespoke an excessive modesty in its author. Between its coming and its going, dialysis dementia had been scrupulously documented, debated, investigated, and subjected to several provisional (and ineffective) therapies. Only after great effort, pointing at last to its cause, was it possible to relegate dialysis dementia to the annals of medical history. A well-tuned clinical antenna, sound thinking, collective effort, and the right tools had all been part of the winning formula.

Medical breakthroughs as definitive as this one are not, of course, made routinely. Even today, we are only just beginning to apply science to clinical studies, relying on technologies that offer new insight into the critical biochemical and genetic mechanisms that underlie disease.  The unraveling of dialysis dementia, in fact, relied heavily on new technology.  An undramatic but indispensable aspect of the story was the role played by improved methods for assaying aluminum.  But it is not a technological breakthrough that makes the above account of general interest.  It is its intimation of how much is medically possible because there is an abiding conviction in the medical community—and not exempting the community’s many critics– that science makes anything possible.  In the battle against disease, the modern idea of progress, much maligned, has not lost its meaning.   Looking back, the eradication of dialysis dementia is a reminder that the mysteries in medicine, even those that seem to spring forth from the vacuum, lie tantalizingly within reach of human understanding.



  1. Mahurkar SD, Salta R, Smith E, Dhar S, Meyers JRL, Dunea G. Dialysis dementia. The Lancet. 1973 Jun 23;301(7817):1412-1415.
  2. SiddiquiI JY, Fitz AE, Lawton RL, KirkendallI WM: Causes of death in patients receiving long-term hemodialysis. JAMA. 1970; 212:1350-1355.
  3. Chokroverty S, Bruetman ME, Berger V, Reyes MG. Progressive dialytic encephalopathy. Journal of Neurology, Neurosurgery, and Psychiatry.1976;39:411-419.
  4. Alfrey AC, MishellI JM, Burks J, Contiguglia SR, Rudolph H, Lewin E, Holmes JH: Syndrome of dyspraxia and multifocal seizures associated with chronic hemodialysis. Transactions of the American Society of Artificial Internal Organs. 1972; 18:257-261.
  5. Dunea G, Mahurkar SD, Mamdani B, Smith EC. Role of aluminum in dialysis. Annals of Internal Medicine. 1978 Apr 1;88(4):502-504.
  6. Berks JS, Alfrey AC, Huddlestone J, Norenberg MD, Lewin E. A fatal encephalopathy in chronic haemodialysis patients. The Lancet. 1976 Apr 10;307(7963):764-768.
  7. Dunea G. Dialysis dementia: An epidemic that came and went. Dunea G. Dialysis dementia: an epidemic that came and went. Asaio Journal. 2001 May 1;47(3):192-194.
  8. Mahurkar SD, Meyers L, Cohen J, Kamath RV, Dunea G. Electroencephalographic and radionuclide studies in dialysis dementia. Kidney International. 1978;13(4):306-315.
  9. Greenblatt D. Dialysis dementia. The Lancet. 1973 Jul 21;302(7821):159.
  10. Gunale SR. Dialysis dementia: Asparagine deficiency? The Lancet. 1973 Oct 13;302(7833):847.13.
  11. Riley V. Dialysis dementia: Probably not asparagine deficiency. The Lancet. 1973 Dec 1;302(7840):1275.
  12. Wardle EN. Dialysis dementia. The Lancet. 1973 Jul 7;302(7819):47.
  13. Unsigned. Dialysis dementia: Aluminum again? The Lancet.1976 Feb 14;1:349.
  14. Platts MM, Moorhead PJ, Grech P. Dialysis dementia. The Lancet. 1973 Jul 21;302(7821):159.
  15. Lyle WH. Dialysis dementia. The Lancet. 1973 Aug 4;302(7823):271.
  16. Alfrey AC, LeGendre GR, Kaehny WD. The dialysis encephalopathy syndrome. New England Journal of Medicine.1976 Jan 22;294:184-188.
  17. Berlyne GM, Pest D, Ben-Ari J, Weinberger J, Stern M, Gilmore GR, Levine R. Hyperaluminaemia from aluminum resins in renal failure. The Lancet. 1970 Sep 5; 296(7671):494-496.
  18. Berlyne GM. Aluminum toxicity. The Lancet.1976 Mar 13;307(7959):589.
  19. Elliott HL, Dryburgh F, Fell GS, Sabet S, MacDougall A. Aluminum toxicity during regular haemodialysis. British Medical Journal.1978;1:1101-1103.
  20. Flendig JA, Kruis H, Das HA. Aluminum and dialysis dementia. Lancet.1976;1:1235.
  21. Arieff AI, Cooper JD, Armstrong D, Lazarowitz VC. Dementia, renal failure, and brain aluminum. Annals of Internal Medicine. 1979;90(5):741-747.
  22. Dunea G, Mahurkar SD, Mamdan B, Smith EC. Role of aluminum in dialysis dementia. Annals of Internal Medicine.1978;88:502-504.
  23. Davison AM, Walker JS, Oli H, Lewins AM. Water supply aluminum concentration, dialysis dementia, and effect of reverse-osmosis water treatment. The Lancet. 1982 Oct 9;320(8302):785-787.
  24. Wills MR, Savory J. Water content of aluminum, dialysis dementia, and osteomalacia. Environmental Health Perspectives.1985 Nov;63:141-147.



MICHAEL SHULMAN holds a PhD in clinical psychology from Fordham University and a medical degree from the Perelman School of Medicine at the University of Pennsylvania. He is retired from the private practice of nephrology.


Summer 2018  |  Hektorama  |  Nephrology